Eur. J. Med. Chem.

34 (1999) 125135 Elsevier, Paris

125

Original article

The search for TCP analogues binding to the low affinity PCP receptor sites in the rat cerebellum
Jacques Hamona, Florence Espazea, Jacques Vignonb, Jean-Marc Kamenkaa*
a

CRBM, CNRS UPR 1086, Ecole Nationale Suprieure de Chimie, 8, rue de lEcole Normale, 34296 Montpellier cedex 5, France b INSERM U 336, Ecole Nationale Suprieure de Chimie, 8, rue de lEcole Normale, 34296 Montpellier cedex 5, France (Received 18 December 1997; revised 25 September 1998; accepted 8 October 1998)

Abstract With the aim of obtaining selective ligands of the low affinity binding sites of [3H]-1-[1-(2-thienyl)cyclohexyl]piperidine ([3H]TCP) in the rat cerebellum, oxygen and sulfur atoms were introduced in the TCP structure and derivatives to obtain analogues with a lowered lipophilicity. These compounds, and others already obtained, were assayed comparatively to determine their affinities for three sites labeled with [3H]TCP: one in the forebrain, the originally described PCP receptor, and two in the rat cerebellum. Lowering the lipophilicity and modifying the hetero-aromatic moiety yielded some ligands with increased affinity for the low affinity sites in the rat cerebellum and decreased affinity for the high affinity sites in the forebrain. Particularly, two compounds displaying both a high affinity and a good selectivity might be valuable tools to elucidate the pharmacology of the low affinity PCP sites labeled with [3H]TCP in the rat cerebellum. Elsevier, Paris TCP / TCP analogues / PCP receptor subsites / rat cerebellum

1. Introduction [3H]-5-methyl-10,11-dihydro-5H-dibenzo[a, d]cyclohepten-5,10-imine ([3H]MK-801) and [3H]-1-[1-(2thienyl)cyclohexyl]piperidine ([3H]TCP) are the most frequently used ligands for the labeling of the noncompetitive antagonists binding site within the N-methylD-aspartate (NMDA) receptor associated Ca2+ channel (the PCP receptor) [1, 2]. However, competition data analysis using a two-site model revealed that both ligands labeled at least two different binding sites in the rat brain: (i) high affinity sites, well represented in the forebrain and corresponding to the initially discovered PCP receptor and (ii) lower affinity sites, more abundant in the hindbrain and particularly in the cerebellum [38]. The nature and the role of the low affinity binding sites are poorly documented mostly because selective ligands are not available. Particularly, their possible role in neuronal protection is unknown whereas the high affinity sites is a well-established target for neuroprotective agents [6]. Thus, the nding of ligands binding selectively and
*Correspondence and reprints

potently to the low affinity sites might be crucial for their pharmacological characterization. For clarity, the different binding sites discussed here will be marked PCP1 (the [3H]TCP high affinity binding sites in the forebrain, i.e., the PCP receptor within the NMDA receptor associated ionic channel), PCP2 (the [3H]TCP high affinity binding sites in the cerebellum), and PCP3 ([3H]TCP low affinity binding sites in the cerebellum). We have previously introduced oxygen and sulfur atoms in the cyclohexyl and piperidinyl moieties of the TCP structure to obtain analogues with a lowered lipophilicity. In rat forebrain membranes, the affinity of these analogues for the PCP1 sites labeled with [3H]TCP was decreased as a function of lipophilicity: the lower lipophilicity was, the lower affinity was [7]. However, their affinity for PCP2 and PCP3 sites in rat cerebellum membranes was not studied. The possibility that lowering the lipophilicity decreased the affinity for PCP1 in the forebrain while increasing the affinity for PCP3 in the cerebellum was attractive since the lipophilic TCP and MK-801 (log P = 4.56 0.30 and 3.71 0.39 respectively) displayed very low affinities for PCP3 (table I). Thus, we have decided to investigate this hypothesis by:

126
Table I. Inhibition of [3H]TCP binding in rat forebrain and cerebellum membranes by TCP and MK-801. The mean of at least three independent determinations was analyzed according to a single-site model in the forebrain and cerebellum (IC50 a, nM, Hills number) or a two-site model in the cerebellum. The two-site model was statistically more probable in the cerebellum: the proportion of PCP2 and PCP3 sites (%) and affinities (IC50, nM) are given. SEM are in brackets. Compound Forebrain single-site model IC50 (PCP1) TCP MK-801
a

Cerebellum single-site model IC50 nH 0.56 (0.05) 0.33 (0.04)

Cerebellum two-site model IC50 (PCP2) 59 (17) 9.3 (4.6) % (PCP2) 72.6 (3.6) 45.3 (4.9)
b

nH 1.00 0.95 (0.04)

b

IC50 (PCP3)

% (PCP3)

9.3 3.67 (0.65)

188 (63) 995 (458)

3716 (1195) 29.8 (4.5) 11125 (3179) 57.2 (5.8) from [20].

IC50: concentration of unlabeled drug that inhibited 50% of specic [3H]TCP binding on specied sites;

(i) preparing TCP analogues with varied lipophilicities by means of O, S or hydroxyl substitution; (ii) using these new TCP analogues and some of those previously obtained for competition measurements in rat forebrain and cerebellum membranes labeled with [3H]TCP; (iii) determining the more probable model of interaction (single- or two-site) in the cerebellum and comparing the affinities of compounds for PCP1, PCP2, and PCP3. 2. Chemistry Two different synthetic strategies were used to obtain TCP analogues according to the presence or absence of a methyl substitution in the cyclohexyl or heterocyclohexyl moiety (X, Y-substituted ring, see table II). The unsubstituted compounds were easily obtained by means of a Bruylants reaction [9]. It consists in the replacement of a cyano group by an aryl- or hetero-aryl group by the reaction of an -aminonitrile with an arylor hetero-arylmagnesium halide (gure 1). The suitable -aminonitrile resulted from a Strecker-like synthesis in an organic or aqueous medium [10, 11]. When the heteroaryl moiety was a 2-furyl (8, 9), 4-methyl-2-thienyl (15) or a 5-methyl-2-thienyl (14) group, the Grignard reagent was best obtained by means of a magnesium/Lithium exchange reaction [12, 13] between MgBr2 and the suitable 2-Li derivative. The pure (e.e. > 99%) enantiomers of the 3-methyl-piperidine derivatives 10 and 11 were obtained according to the same pathway (gure 1) but starting from optically active 3-methyl-piperidines. The optical resolution of the 3-methyl-piperidine racemate was achieved by means of a crystallization procedure with (+)- and ()-mandelic acid in ethyl acetate resulting in enantiomeric purities up to 9899% [14, 15]. It should be noticed that we have previously described 10-() and 10-(+) obtained by a different strategy [15] but with very similar enantiomeric purities. The new compounds obtained according to gure 1 (69, 1115) were all checked by 13C-NMR spectroscopy of the hydrochloride salts (table III). In these series

indeed, the hydrochlorides solutions are stabilized in almost homogeneous conformations: the aromatic or hetero-aromatic rings are essentially restrained to the axial position. Consequently, the conformational trapping induced by the protonation allows for structural comparisons (see below) between similar (axial aromatic rings) conformations [1619]. Diastereomers 16 and 17, bearing a methyl substitution at the hetero-cyclohexyl ring (table IV), were prepared using the azide synthesis shown in gure 2 [15, 20]. The cis/trans congurations were attributed from the 13CNMR spectra of their HCl salts. The chemical shifts were attributed by comparison with the spectra of GK-11 and GK-12 hydrochlorides, two analogues whose diastereomeric congurations have been previously characterized [20, 21]. The cis (Me/Pip) conguration was attributed to compound 16 where the specic -interaction due to the axial methyl substitution causes a clear upeld shift of carbon 5 (table IV). The synthesis of non-commercial ketonic or piperidinic starting materials was required. Briey, according to reference [22], alkylation of ethyl 2-sulfanylacetate with ethyl 4-chlorobutanoate gave a 76% yield of ethyl 4-[(2-ethoxy-2-oxoethyl)sulfanyl]butanoate which was submitted to a Dieckmann cyclization to afford a 54%

Figure 1.

127
Table II. TCP derivatives and analogues unsubstituted at the cyclohexyl or heterocyclohexyl ring and their calculated log P.

Compound TCP 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
a

X CH2 S O CH2 S O SO2 CH2 CH2 S CH2 S CH2 CH2 S S

Y CH2 CH2 CH2 CH2 CH2 CH2 CH2 S CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH2

Z S S S S S S S S O O S S S S S S

K CH2 CH2 CH2 O O O CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH2 CH2

R1 H H H H H H H H H H H H H H CH3 H

R2 H H H H H H H H H H H H H H H CH3

R3 H H H H H H H H H H CH3 CH3 CH2OH CH3 H H

R4 H H H H H H H H H H H H H OH H H

log P

4.56 0.30 3.66 0.50 2.85 0.39 3.02 0.39 2.11 0.56 1.31 0.41 1.99 0.40 3.52 0.53 4.05 0.29 3.14 0.52 5.06 0.30 4.15 0.53 3.35 0.37 3.21 0.37 4.12 0.53 4.12 0.53

Calculated with the ACD/Log P program (ACD, Inc.) (95% condence, octanol/water).

yield of ethyl 3-oxotetrahydro-2H-thiopyran-2-carboxylate. A decarboxylation in a 10% sulfuric acid solution gave 54% of pure dihydro-2H-thiopyran-3(4H)one (gure 3). 3-Methyltetrahydro-4H-thiopyran-4-one
13 a b

was obtained by reacting tetrahydro-4H-thiopyran-4-one with methyl-iodide in THF in the presence of one equivalent of LDA at the temperature of 80 C. 3-Methyl-4-piperidinol was obtained essentially as its

Table III. Carbon 1 2 3 4 5 6 R CAr

a

C-NMR 6
c

chemical shifts 7

(hydrochloride in CDCl3, ppm from TMS). 8 23.4 22.5 29.8 67.6 29.8 22.5 47.0 47.0 22.3 22.3 21.7 146.6110.7
b

9 24.5 30.2 66.8 30.2 24.5 46.7 46.7 22.0 22.0 21.3 145.2110.7

11 24.9 33.4 68.9 33.2 24.9 52.3 46.2 28.2 30.5 22.1 18.9 134.5127.9

12 22.2 23.0 33.4 64.2 33.0 23.0 49.7 46.8 36.6 25.3 24.0 69.7 135.8128.0
c

13 22.5 30.8 32.8 69.0 32.8 30.8 50.6 45.3 35.4 32.8 70.0 15.1 135.1127.5

14 24.7 32.8 68.8 32.8 24.7 46.5 46.5 22.3 22.3 21.7 14.7 142.9125.8

15 25.2 33.6 69.4 33.6 25.2 47.2 47.2 22.8 22.8 21.1 15.6 138.9123.9

48.0 30.5 66.6 30.5 48.0 47.3 47.3 22.8 22.8 21.5 132.9128.4

22.7 33.4 68.0 32.4 26.8 47.0 47.0 25.5 25.5 23.9 134.6128.6

For carbon atom numbering see table IV;

italicized chemical shifts may be exchanged;

DMSO-d6.

128
Table IV. Comparative 13C- (up) and 1H-NMR (down) chemical shifts of GK11, GK12, and compounds 16, 17 (hydrochloride in CDCl3, ppm from TMS).

Figure 3.
Carbon 1 2 3 4 5 6 CH3 2 3 4 5 -CH3 J (Hz)
a

GK11 (cis) 17.8 30.2 35.4 72.9 26.5 22.6 15.8 48.9 46.7 22.3 22.1 22.5 137.1 127.6 127.2 130.1 1.6 6.9

GK12 (trans) 22.5 30.0 a 36.3 74.5 30.7 a 22.0 17.1 48.2 47.9 21.72 21.69 21.9 136.4 126.8 126.4 130.6 1.1 6.6

17 (trans) 29.8 a 34.7 73.3 29.6 a 22.7 16.8 48.6 47.2 21.9 21.6 21.2 135.8 127.3 127.0 130.3 1.2 6.2

16 (cis) 33.2 34.8 72.4 27.0 24.9 15.1 48.9 46.3 22.0 22.0 22.5 134.8 128.3 127.8 131.1 1.8 6.7

3. Pharmacology The binding assays in rat forebrain and cerebellum membranes and the data analysis are described in experimental protocols (Section 5.2). We have rst checked that a two-site model was more probable than a single-site model to describe the competition of TCP and MK-801 in the rat cerebellum membranes labeled with [3H]TCP (table I). The results were consistent with those previously reported in membranes [3, 4] as well as in cultured cerebellum cells [5]. The same treatment was applied to 21 compounds derived from the TCP structure; the results are presented in table V. 4. Results and discussion In forebrain homogenates, [3H]TCP inhibition curves were better tted to a single-site model (PCP1-sites) (table V). This result was likely since Hill numbers were most generally close to unity. However, in cerebellum homogenates, Hill numbers were mostly lower than unity and the inhibition curves were better tted to a two-site

Italicized chemical shifts may be exchanged.

trans isomer by a 4-step synthesis starting from benzamide and ethyl acrylate (gure 4) [23]. The conguration was attributed by comparison with the 13C-NMR spectra of cis- and trans-2-methylcyclohexanol [24].

Figure 2.

Figure 4.

129
Table V. Inhibition of [3H]TCP binding in rat forebrain and cerebellum membranes. The mean of at least three independent determinations was analyzed according to a single-site model in the forebrain and cerebellum (IC50 a, nM, Hills number) or a two-site model in the cerebellum (SEM in brackets). When the two-site model was more probable, the proportion of PCP2 and PCP3 sites (%) and affinities (IC50 a , nM) are given. Very low affinities precluded the two-site computation (n.d.). Compound Forebrain single-site model IC50 (PCP1) 1 2 3 4 5 6 7 8 9 10-() 10-(+) 10-() 11-() 11-(+) 11-() 12 13 14 15 16 17
a

Cerebellum single-site model IC50 nH 0.68 0.58 0.69 0.72 0.60 0.77 0.65 0.70 0.55 0.80 0.62 0.90 0.69 0.59 0.85 0.61 1.02 0.63 0.62 0.71 0.60 (0.10) (0.12) (0.09) (0.02) (0.08) (0.22) (0.02) (0.11) (0.03) (0.13) (0.08) (0.08) (0.07) (0.04) (0.13) (0.06) (0.09) (0.06) (0.06) (0.07) (0.06)

Cerebellum two-site model IC50 (PCP2) % (PCP2) 586 (96) 5840 (1690) 183 (65) n.d. n.d. n.d. 659 (206) 187 (42) 1015 (262) 47.3 (5.2) 28.3 (17.1) b 265 (101) 53.2 (15.8) b 2240 (330) b 532 (20) 772 (202) 1355 (733) n.d. 75.2 (8.8) 80.3 (6.5) 64.3 (3.4) 77.8 (3.7) 66.3 (12.0) 68.3 (2.8) 68.5 (9.6) 52.5 (10.5) 69.1 (9.8) 66.3 (1.1) 69.0 (4.0) 83 (3.5) 76.4 (2.6) 62.8 (7.3)
b

nH 1.05 0.98 0.77 0.78 _ 0.96 1.00 0.83 0.87 1.09 1.04 1.08 1.14 0.84 0.93 0.96 1.07 0.97 0.93 0.89 0.94 (0.05) (0.05) (0.02) (0.06) (0.18) (0.05) (0.06) (0.06) (0.07) (0.06) (0.09) (0.11) (0.03) (0.05) (0.09) (0.04) (0.07) (0.06) (0.01) (0.09)

IC50 (PCP3) % (PCP3) 8.5 (4.5) 35 (18) 2127 (1273) n.d. n.d. n.d. 8.0 (4.8) 5.6 (3.5) 13.8 (3.6) 1450 (590) 808 (47) b 2850 (433) 2076 (747) b 12.4 (5.7) b > 100 M 5.4 (3.4) 17.6 (6.6) n.d. 25.2 (7.2) 15.5 (5.7) 33.0 (1.1) 23.4 30.7 32.3 33.7 51.0 31.6 34.6 31.0 15.3 21.8 34.7

71.6 (10.5) 1223 (143) 294 (48) 8253 (320) > 100 M 97700 (19066) 73.4 (1.4) 47.8 (1.1) 133 (20) 5.5 (1.3) 5.2 (1.0) 158 (20) 132 (39) 24 (8.2) 529 (50) 28.4 (6.6) 2406 (218) 462 (88) 77.8 (9.3) 26.8 (4.5) 17233 (1417)

178 (40) 2396 (475) 1001 (408) 29700 (8022) 63800 (10400) 198500 (8500) 228 (43) 150 (78) 0.55 (0.03) 122 (38) 155 (10) 452 (53) 265 (80) 231 (71) 996 (314) 467 (158) 3362 (1013) 1170 (242) 762 (335) 206 (30) 50467 (2663)

(5.0) (10.0) (4.4) (8.1) (11.8) (11.1) (2.5) (5.0) (6.9) (3.8) (9.5)

IC50: concentration of unlabeled drug that inhibited 50% of specic [3H]TCP binding on specied sites; probable.

a one-site model was more

model (PCP2- and PCP3-sites) which produced a signicant reduction in the sum of squares (P < 0.05, Students t-test) and a DurbinWatson coefficient between 1.5 and 2.5 (see Experimental protocols). The results presented in table V conrmed that [3H]TCP labeled two sites in the cerebellum in a 70:30 mean relative proportion (PCP2/PCP3). Interestingly, most compounds were more or less able to interact with these binding sites. However compounds 4, 5, 6, and 17 displayed a very low affinity whatever the binding sites and thus their affinity and selectivity could not be evaluated. Compounds 10-(), 11-(), and 13 displayed Hill numbers close to unity and their interaction was best described by a single-site model. As previously shown, the affinity for PCP1-sites was reduced with the decrease of lipophilicity [7] in homogeneous groups of structures. Indeed, TCP and derivatives displayed greater affinities for PCP1 than their thio- or oxa-analogues: TCP had a higher affinity for the PCP1sites than 1 or 2 (tables I and V), similarly 10 was a higher affinity ligand than 11 whatever the chirality is (table V). The same behavior was revealed when comparing TCP and 8. Interestingly, 1 and 7 were equipotent

ligands for the PCP1-sites although their sulfur atoms (in the six atoms ring) are located in two different positions. Affinity for PCP2-sites could be determined for 16 compounds only (table V). Among them, 3, 11-(), 11(+), 13, and 14 were unable to discriminate signicantly PCP1- and PCP2-sites since they displayed statistically close affinities for both sites. At the contrary, the 12 remaining compounds had affinities signicantly different at PCP1- and PCP2-sites. These affinities were apparently linearly correlated (P < 0.004, r = 0.76, see gure 5). The lower affinity of these compounds for PCP2-sites might be related to the distribution of NMDA NR2 subunits. Indeed NR2A and NR2B subunits are highly expressed in the forebrain while NR2C subunit is mainly expressed in the cerebellum [25, 26]. NMDA receptors involving NR2C subunits are less sensitive to MK-801 blockade than those comprising NR2A and/or NR2B subunits [27]. These results are conrmed by competition experiments with [3H]MK-801 [4, 5] and consistent with MK-801 affinities in table I. Thus PCP1- and PCP2-sites are likely to represent two different states of the NMDA receptor discriminated by some of the new molecules.

130

Figure 5. Linear relationships between affinities for PCP1- and PCP2-sites (r = 0.76, p < 0.004) for 12 compounds discriminating signicantly both sites.

Figure 6. Relationships between selectivity (PCP3/PCP1) and log P (r = 0.84, p < 0.08) for molecules differing from the TCP model only by substitution of one heteroatom in the cyclohexyl ring and/or in the hetero-aromatic moiety.

The affinities for the PCP3-sites could not be correlated with those for PCP1- or PCP2-sites and they appeared very sensitive to structure. Indeed, molecules tested appeared grossly separated into two groups: (i) molecules differing from TCP only by substitution of one heteroatom in the cyclohexyl ring and/or in the hetero-aromatic moiety, (ii) molecules differing from TCP by introduction of a methyl substitution in any of the constitutive rings or by specic modications of the piperidine ring. Comparatively to the TCP model structure, lowering the lipophilicity (table II) clearly directed the rst group of molecules (1, 2, 7, 8, 9) toward the PCP3-sites since they displayed higher affinities for these sites (5.6 to 35 nM) than for PCP1- (47.8 to 1223 nM) or PCP2-sites (187 to 5840 nM) (table V).Moreover, plotting IC50 PCP3/IC50 PCP1 against log P conrmed this tendency: the lower log P was, the higher the selectivity for PCP3 was (gure 6), in line with our hypothesis. A similar tendency was found with regard to PCP2 (not shown). Finally, this group of compounds revealed interesting selectivities for PCP3 when compared to PCP2 and PCP1. 2, the less potent (35 nM) and the less lipophilic (log P = 2.85) in this group, displayed a high selectivity for the PCP3-sites when compared to the PCP1- and PCP2-sites (PCP3/PCP1 < 0.026; PCP3/PCP2 < 0.006). In the second group, changing the piperidine for a morpholine ring decreased considerably the affinities (3) or precluded the two-site computation given very low affinities (4, 5). Substitution of a methyl group in the piperidine ring gave compounds with low affinities for

PCP3-sites: 13, 10-(), 10-(+), and their thio-analogues 11-(), 11-(+). Interestingly, the affinities of 10-(), 11(), and 13 were better described by a single-site model in the cerebellum. Compound 12 bearing an hydroxymethyl substitution in the piperidine ring exhibited a high affinity for both PCP3- and PCP1-sites with a high selectivity with regard to PCP2-sites. The methyl substitution in the thiopyranyl ring gave a high affinity cis-compound (16) equipotent at the PCP1- and PCP3-sites and an inactive trans-compound (17). The difference between cis and trans diastereomers in the forebrain was higher than previously observed in the cyclohexyl homologue series [15, 20] although affinities were lower. Finally, the position of a methyl group in the hetero-aromatic moiety was crucial. In the position from the sulfur atom (14), it lowered the affinity for PCP3 and at the contrary, in the position (15), it increased the affinity for these sites. Moreover 15 displayed a good selectivity with regard to both PCP1- and PCP2-sites (PCP3/PCP1 < 0.07; PCP3/PCP2 < 0.007). This is a conrmation of the very important role played by the aromatic or hetero-aromatic ring in the arylcyclohexylamines selectivity of binding [28]. In this second group of molecule steric interactions due to substitutions are likely to inuence more selctivity than lipophilicity. It is now well admitted that the NMDA receptor is an hetero-oligomeric protein composed with 5 sub-units differently expressed in forebrain, cerebellum, and spinal cord [2931]. The resulting diversity of NMDA receptors

131 in the CNS, and consequently the diversity of the PCP receptors, is responsible for the heterogeneity of pharmacological responses [32]. The present study conrms the binding sites heterogeneity since TCP analogues were able to interact differently with subsites labeled with [3H]TCP in the rat cerebellum. In this region interestingly NMDA receptors display particular properties when compared with NMDA receptors found in the forebrain [26]. Some of the structures we have prepared and tested might be valuable tools to clarify the pharmacological role of the low affinity sites in the rat cerebellum (PCP3). Indeed, compounds 2 and 15 possess both a high affinity and a good selectivity for these sites. Such molecules might also be leads for new compounds able to interact with this specic target. 5. Experimental protocols 5.1. Chemistry Melting points (uncorrected) were determined with a BchiTottoli apparatus. Yields were not optimized. Elemental analysis was performed at the CNRS Microanalytical Section in Montpellier on the hydrochloride salts and were within 0.4% of theoretical values. 1H- and 13 C-NMR spectra were obtained on a Brucker AC 200 spectrometer at 200.13 and 50.32 MHz respectively in 5-mm sample tubes in the FT mode. For some 13C-signal assignments, a spin-echo sequence (Jmod) was used. Chemical shifts are reported in () ppm downeld from TMS. Enantiomeric purities were determined on a Shimadzu HPLC equipment (LC-1O AD pump, SPD-6A UV spectrometer), computer-controlled by the Class LC-10 program. Analysis were made on a Chiralcel-OD column (10 mm, 4.6 250 mm) (Daicel Chemical Industries) in heptane (0.6 mL /min) at 36 C. UV-detection was made at 240 nm. Typical injection volumes were 5 mL of a 30 M solution of base compound in heptane. Optical rotations were obtained in methanol with a Perkin-Elmer 241 polarimeter in a 1-dm microcell at 20 C. For NMR and in vitro experiments, compounds were used as their hydrochloride salts; salts were precipitated by adding a dry HCl ethereal solution in ether to a solution of base in ether. After ltration, the solids collected were dried in vacuum. 5.1.1. Synthesis of dihydro-2H-thiopyran-3(4H)-one 5.1.1.1. Ethyl 4-[(2-ethoxy-2-oxoethyl)sulfanyl]butanoate Sodium (4.6 g, 200 mmol, 1 eq.) was added cautiously by portion in a nitrogen atmosphere to ethanol (100 mL). After the solid was consumed, the mixture was cooled to 0 C and ethyl 2-sulfanylacetate (22 mL, 200 mmol, 1 eq.) then ethyl 4-chlorobutanoate (30.1 g, 200 mmol, 1 eq.) was added slowly. The resulting mixture was stirred for 20 h at room temperature. The NaCl precipitate was ltered, the ltrate concentrated in vacuum, the oil obtained was diluted in water (100 mL) and extracted with ether (3 80 mL). The combined organic layers were dried over MgSO4 and concentrated in vacuum. The resulting yellow oil was distilled to yield 35.6 g (76%) of a colorless oil.

5.1.1.2. Ethyl 3-oxotetrahydro-2H-thiopyran-2-carboxylate Ethyl 4-[(2-ethoxy-2-oxoethyl)sulfanyl]butanoate in anhydrous ether (200 mL) was added dropwise in a nitrogen atmosphere to a solution of sodium ethanolate (20.7 g, 0.3 mol, 2 eq.) at 0 C, stirred at 0 C for 45 min and at room temperature for 3 h. The mixture was hydrolyzed with a water/acetic acid (80:20) solution, the aqueous phase was separated and extracted with ether (3 50 mL), the combined organic layers were dried over MgSO4 and concentrated in vacuum. The resulting yellow oil obtained was distilled under reduced pressure to yield 15.3 g (54%) of a colorless liquid.

5.1.1.3. Dihydro-2H-thiopyran-3(4H)-one Ethyl 3-oxotetrahydro-2H-thiopyran-2-carboxylate (15.2 g, 80.9 mmol) was reuxed in a 15% sulfuric acid solution for 18 h then cooled to room temperature. A 10% NaOH solution in water was then added dropwise to reach pH 6. The mixture was extracted with ether (3 50 mL), the organic phases washed with water, dried over MgSO4, and concentrated in vacuum. The resulting oil was distilled under reduced pressure to yield 5.1 g (54%) of a colorless oil.

5.1.2. Synthesis of 1,1-dioxo-tetrahydro-16-thiopyran-4-one A solution of hydrogen peroxide (11.4 mL, 0.1 mol, 2 eq.) was added dropwise to a mixture of tetrahydro-4Hthiopyran-4-one(5.6 g, 48 mmol, 1 eq.) and acetic acid (25 mL) keeping the temperature below 30 C. The mixture was stirred for 4 h, the acetic acid distilled under reduced pressure, and the crystallized yellow residue was ltered and washed with ether to yield 4.7 g (67%) of white crystals.

132 5.1.3. Synthesis of 3-methyl-4-piperidinol 5.1.3.1. Ethyl 1-benzoyl-4-oxo-3-piperidinecarboxylate Sodium hydride (4 g, 0.1 mol, 1 eq.) was added in a nitrogen atmosphere to a solution of benzamide (12.1 g, 0.1 mol, 1 eq.) in toluene (200 mL), the mixture was reuxed for 1 h, cooled to 0 C and ethyl acrylate (32.6 mL, 0.3mol, 3 eq.) was then rapidly added. The solution was stirred at 60 C for 24 h, cooled to 0 C, diluted with ice-cold water (100 mL) and stirred for 0.5 h. The aqueous phase was separated, washed with ether (50 mL), acidied until pH 3 and extracted with CH2Cl2 (3 50 mL). The combined organic layers were dried over Na2SO4, ltered and concentrated in vacuum. The yellow oil obtained was puried by column chromatography (SDS Chromagel 70100 ) in ether to yield 9.4 g (34%) of a red oil. 5.1.3.2. Ethyl 1-benzoyl-3-methyl-4-oxo-3-piperidinecarboxylate A mixture of ethyl 1-benzoyl-4-oxo-3-piperidinecarboxylate (9.3 g, 33.8 mmol, 1 eq.) and sodium hydride (1.35 g, 34 mmol, 1 eq.) in dimethoxyethane (50 mL) was reuxed for 2 h then cooled to 0 C. ICH3 (5.2 mL, 84 mmol, 2.5 eq.) was added and the mixture heated at 60 C for 40 h. The mixture, after concentration in vacuum, was diluted with water (100 mL) and extracted with CH2Cl2 (3 50 mL). The combined organic layers were washed successively with a 5% NaOH water solution (50 mL), a 5% HCl water solution (50 mL), with water (50 mL), then dried over Na2SO4, and concentrated in vacuum. The resulting brown oil (9.4 g) was puried by column chromatography (SDS Chromagel 70200 ) in ether to yield 7.7 g (79%) of a slightly yellow oil. 5.1.3.3. 3-Methyl-piperidin-4-one hydrochloride A solution of ethyl 1-benzoyl-3-methyl-4-oxo-3piperidinecarboxylate (7.7 g, 26.6 mmol, 1 eq.) in a 6 N HCl aqueous solution was reuxed for 72 h, then the benzoic acid precipitate was ltered, the ltrate washed with ether (3 50 mL) and concentrated in vacuum. The brown solid obtained was crystallized in ethanol to get white crystals (2.85 g, 72%). 5.1.3.4. 3-Methyl-4-piperidinol A solution of 5% NaOH in water (7.6 mL) was added dropwise to a solution of 3-methyl-piperidin-4-one hydrochloride (2.85 g, 19 mmol, 1 eq.) in methanol (30 mL) and the mixture was stirred for 0.5 h at room temperature. A solution of NaBH4 (2.46 g, 6.5 mmol, 1.4 eq.) in methanol (30 mL) was added, the mixture was stirred for 4 h at room temperature, cooled to 0 C, made acidic with few drops of a 5% HCl aqueous solution, and concentrated in vacuum. The resulting yellow solid was dissolved in hot ethanol and precipitated at room temperature by the addition of drops of ether to yield a white solid (1.9 g, 87%). 5.1.4. Synthesis of -aminonitriles -aminonitriles were obtained by means of two different synthetic methods. Since the same method was applied to various compounds we describe only one example in each case. 5.1.4.1. Method A Acetone cyanohydrine (2.2 g, 25.8 mmol, 1 eq.) was added dropwise to a stirred mixture of tetrahydro-4Hthiopyran-4-one (3 g, 25.8 mmol, 1 eq.), anhydrous MgSO4 (9.3 g, 77.4 mmol, 3 eq.), dimethylacetamide (2.25 g, 25.8 mmol, 1 eq.) and piperidine (4.4 g, 51.6 mmol, 2 eq.). The pasty mixture was heated at 45 C for 48 h, cooled to room temperature, poured onto ice and stirred for 30 min. The aqueous mixture obtained was extracted with ether and the organic layer was washed with water until neutrality, dried over Na2SO4, ltered, and concentrated in vacuum to yield 5.3 g (98%) of an orange oil of 4-piperidinotetrahydro-2H-thiopyran-4carbonitrile. The following -aminonitriles were similarly synthesized: 4-piperidinotetrahydro-2H-pyran-4-carbonitrile (79%, F = 4647 C) from piperidine and tetrahydro-4Hpyran-4-one; 4-(3-methylpiperidino)tetrahydro-2H-thiopyran-4-carbonitrile (> 98%, oil) from 3-methylpiperidine and tetrahydro-4H-thiopyran-4-one; R-4-(3methylpiperidino)tetrahydro-2H-thiopyran-4-carbonitrile (> 98%, oil) and S-4-(3-methylpiperidino)tetrahydro-2Hthiopyran-4-carbonitrile (> 98%, oil) from R-()-3methylpiperidine and S-(+)-3-methylpiperidine [7, 8, 21], and tetrahydro-4H-thiopyran-4-one; 4-(3-methylpiperidino)tetrahydro-2H-thiopyran-4-carbonitrile (81%, solid) from 3-hydroxymethylpiperidine and cyclohexanone. 5.1.4.2. Method B 5% HCl (a few drops) was added to a stirred mixture of 3-methyl-4-piperidinol (0.8 g, 7 mmol, 2 eq.) in 10 mL of water and cyclohexanone (2.7 g, 23.6 mmol, 1 eq.) to reach pH 3. KCN (0.47 g, 7.3 mmol, 1.05 eq.) was added (pH reached 11). The mixture was stirred at room temperature for 24 h then extracted with CH2Cl2, dried over Na2SO4, ltered, and concentrated in vacuum to yield 1.43 g (93%) of 1-(4-hydroxy-3-methylpiperidino)cyclohexanecarbonitrile as a colorless oil.

133
Table VI. Purication and properties of new compounds. Purication 6 7 8 9 11-() 11-(+) a 11-() b 12 13 14 15
a

Solvent (v/v) (Al2O3) (Al2O3) (Al2O3) (Al2O3) (Al2O3) (Al2O3) (Al2O3) (Al2O3) (Al2O3) (Al2O3)
c

Fbase (C) 167 5860 Oil 7981 Oil Oil Oil Oil 120122 9395 8991
b

FHCl (C) 179183 184186 166167 171173 168169 170172 169171 176178 170172 160162 144146

Yield (%) 48 83 76 71 57 59 64 56 71 67 55

Crystallization Column chromatograpy Column chromatograpy Column chromatograpy Column chromatograpy Column chromatograpy Column chromatograpy Column chromatograpy Column chromatograpy Column chromatograpy Column chromatograpy

AcOEt Petroleum CH2Cl2 Petroleum Petroleum Petroleum Petroleum Petroleum Petroleum Petroleum Petroleum

ether/ether (40:60) ether/ether ether/ether ether/ether ether/ether ether/ether ether/ether ether/ether ether/ether (98:2) (95:5) (95:5) (95:5) (50:50) (20:80) (98:2) (90:10)

HPLC (chiral phase): Rt = 29.1 min, e.e. > 99%; [D20]base = +12 (c 1, CH3OH); [D20]base = 11 (c 1, CH3OH); c aluminium oxide 90, 23 Merck.

HPLC (chiral phase): Rt = 32.3 min, e.e. > 99%;

1,1-Dioxo-4-piperidinohexahydro-16-thiopyran-4carbonitrile (64%, F = 148 C) was similarly prepared from piperidine and 1,1-dioxo-tetrahydro-16-thiopyran4-one. 5.1.5. Synthesis of TCP analogues from -aminonitriles Compounds 1 [7], 2 [7, 33], 35 [7] as well as compounds 10-(), 10-(), 10-(+) [15] have been previously described. TCP analogues were obtained by means of two different synthetic methods. Since the same method was applied to various compounds we describe only one example to illustrate each strategy. Compounds purications and properties are given in table VI. 5.1.5.1. Method 1: TCP analogues with a thiophenyl hetero-aromatic ring (6, 7, 1113) 1-[3-(2-thienyl)tetrahydro-2H-thiopyran-3-yl]piperidine 7: An ethereal solution of 3-piperidinotetrahydro-2Hthiopyran-3-carbonitrile (1 g, 4.76 mmol, 1 eq.) was added dropwise at room temperature to a well stirred solution containing a Grignard reagent prepared from 2-bromo-thiophene (2.3 g, 14.3 mmol, 3 eq.) and Mg turnings (0.35 g, 14.3 mmol, 3 eq.). The mixture was reuxed for 20 h, cooled to room temperature and treated as follows: the mixture was poured carefully on an ice-cold saturated solution of NH4Cl, stirred for 30 min, extracted with ether, the combined ether layers were extracted 3 times with 10% HCl, 20% NH4OH was added to the aqueous phase until neutrality. The aqueous phase was extracted with ether, the organic phase washed with water, dried over Na2SO4, ltered, and concentrated in vacuum. The crude product obtained was puried as described in table VI to yield 1.05 g (83%) of 7 as a white solid.

5.1.5.2. Method 2: TCP analogues with a furanyl (8, 9) or a substituted thiophenyl (14, 15) hetero-aromatic ring 1-[4-(2-furyl)tetrahydro-2H-thiopyran-4-yl]piperidine 9: Firstly, a MgBr2 solution was prepared from 1,2dibromo-ethane (6.76 g, 36 mmol, 4 eq.) and magnesium turnings (0.88 g, 36 mmol, 4 eq.) in ether (80 mL) in a nitrogen atmosphere. Secondly, a solution of 2-furylLithium was prepared in a nitrogen atmosphere by the dropwise addition of a n-butyl-lithium solution 1.6 M in hexane (28 mL, 45 mmol, 5 eq.) to a mixture of furane (3.1 g, 45 mmol, 5 eq.) and TMEDA (5.2 g, 45 mmol, 5 eq.) in anhydrous ether (100 mL) at 20 C (without TMEDA in the synthesis of 15). The mixture was then reuxed for 2 h, cooled to room temperature, and added dropwise to the MgBr2 solution in ether. A solution of 4-piperidinotetrahydro-2H-thiopyran-4-carbonitrile (1.9 g, 9 mmol, 1 eq.) in ether was added dropwise at room temperature, the mixture reuxed for 16 h, cooled to room temperature, and treated as described above. The crude product was puried as described in table VI to yield 1.6 g (71%) of 9 as a white solid. 5.1.6. Preparation of compounds 16 and 17 5.1.6.1. 3-Methyltetrahydro-4H-thiopyran-4-one n-Butyl-lithium (37.5 mL, 60 mmol, 1 eq.) was added dropwise in a nitrogen atmosphere to a stirred solution of diisopropylamine (8.4 mL, 60 mmol, 1 eq.) in THF (74 mL) and the mixture was stirred 0.5 h at room temperature, then cooled to 80 C. After slow addition of tetrahydro-4H-thiopyran-4-one (6.96 g, 60 mmol, 1 eq.) and stirring for 0.5 h, ICH3 (5.6 mL, 90 mmol, 1.5 eq.) was added, the mixture allowed to warm up to room temperature while stirred for 5 h. The mixture was diluted with a NaCl saturated 5% solution of sodium bicarbonate

134 in water and the organic phase separated, dried over Na2SO4 and concentrated in vacuum. The orange oil obtained was puried by chromatography (SDS Chromagel 70200 ) in petroleum ether/ethyl acetate (90:10) to yield 3.2 g (41%) of a colorless oil. 5.1.6.2. Cis- and trans-3-methyl-4-(2-thienyl)tetrahydro-2H-thiopyran-4-ol 2-Thienyl-magnesium bromide was prepared by adding dropwise in a nitrogen atmosphere a solution of 2-bromo-thiophene (6.1 g, 37.2 mmol, 1.1 eq.) in ether (100 mL) to Mg turnings (0.9 g, 37.2 mmol, 1.1 eq.). After the mixture was gently reuxed for 3 h, a solution of 3-methyltetrahydro-4H-thiopyran-4-one (4.2 g, 32 mmol, 1 eq.) dissolved in ether (50 mL) was added at room temperature. The mixture was reuxed for 16 h, cooled to room temperature, poured onto a saturated NH4Cl solution in water, stirred 0.5 h. The aqueous phase was separated, extracted with ether (3 50 mL) and the combined organic layers extracted with 15% HCl (3 50 mL). 25% NH4OH was added to the aqueous phase until neutrality and the solution extracted with ether. The nal organic phase was washed with water, dried over Na2SO4, ltered, and concentrated in vacuum. The green oil obtained was puried by chromatography (SDS Chromagel 70200 ) in petroleum ether/ether (90:10) to yield 6.8 g (98%) of a slightly blue oil containing the diastereomeric alcohols. 5.1.6.3. Cis- and trans-3-methyl-4-(2-thienyl)tetrahydro-2H-thiopyran-4-azide Sodium azide (4 g, 61.7 mmol, 2 eq.) was cautiously added to a solution of trichloro-acetic acid (15.1 g, 92.4 mmol, 3 eq.) in CHCl3 (100 mL), the mixture was cooled to 10 C, and a solution of diastereomeric alcohols 3-methyl-4-(2-thienyl)tetrahydro-2H-thiopyran-4-ol(6.6g, 30.8 mmol, 1 eq.) in CHCl3 (50 mL) was added. The mixture was stirred for 72 h at 1012 C, then a 10% NH4OH solution was added until neutrality and the aqueous phase extracted with CH2Cl2 (3 100 mL). The pooledorganic phases were washed with water, dried over Na2SO4, ltered, and concentrated in vacuum. The oil obtained (6.8 g, 92%) was used in the next step without further purication. 5.1.6.4. Cis- and trans-3-methyl-4-(2-thienyl)tetrahydro-2H-thiopyran-4-amine A solution of cis- and trans-3-methyl-4-(2-thienyl) tetrahydro-2H-thiopyran-4-azide (5.5 g, 23 mmol, 1 eq.) in THF (30 mL) was added dropwise at 0 C to a solution of lithium aluminium hydride (0.87 g, 23 mmol, 4 eq.) in THF in a nitrogen atmosphere stirred for 24 h at room temperature. The strict minimum amount of NH4OH necessary to destroy lithium aluminium hydride was added slowly, the precipitate ltered, washed with CH2Cl2 (300 mL), and the ltrate concentrated in vacuum. The brown oil obtained was diluted in ether, extracted with a 10% HCl solution (3 100 mL) and 20% NH4OH added to the aqueous phase until neutrality. The aqueous phase was extracted with ether (3 100 mL) and the combined organic layers washed with water, dried over MgSO4, ltered, and concentrated in vacuum. The crude diastereomeric oily mixture was puried by chromatography (SDS Chromagel 70200 ) in petroleum ether/ether (50:50) to yield 3.7 g (75%) of a cis/trans primary amines mixture as a colorless oil. 5.1.6.5. Cis- and trans-1-[3-methyl-4-(2-thienyl)tetrahydro-2H-thiopyran-4-yl]piperidine 16, 17 Potassium carbonate (5.8 g, 42 mmol, 2 eq.) and 1,5dibromopentane (3.6 mL, 26.2 mmol, 1.25 eq.) was added in a nitrogen atmosphere to cis- and trans3-methyl4-(2-thienyl)tetrahydro-2H-thiopyran-4-amine (4.5 g, 21 mmol, 1 eq.) dissolved in HMPT (50 mL) and the resulting mixture stirred at 60 C for 48 h then cooled to room temperature. The mixture was poured onto water (200 mL) and extracted with ether (3 100 mL). The combined organic layers were extracted 3 times with 10% HCl and 20% NH4OH was added to the aqueous phase until neutrality. The aqueous phase was extracted with ether and the nal organic phase was washed with water, dried over MgSO4, ltered, and concentrated in vacuum. The resulting yellow oil was chromatographied (SDS Chromagel 70200 ) in petroleum ether/ether (90:10) to yield the pure diastereomers as white solids: 17 (2 g, 34%) and 16 (0.7 g, 12%). 5.2. Biochemistry 5.2.1. Binding assays [3H]TCP (Amersham, 48 Ci/mmol) binding to the PCP receptor subtypes was measured as previously described [3]. Briey, the rat (wistar) brain (minus the cerebellum) or the cerebellum was homogenized with an Ultraturax (Ika Werke, maximum setting) in a 50 mM Tris/HCl, pH 7.7 buffer for 20 s at 4 C. The homogenate was then centrifuged at 49000 g for 20 min. The pellet was resuspended in the same buffer and the homogenizationcentrifugation steps performed a second time. The nal pellet was resuspended in 10 volumes of a 50 mM Tris/Hepes, pH 7.7 buffer and used without further purication. The forebrain or cerebellum homogenate (0.50.8 mg protein/mL) was incubated with [3H]TCP (1 nM or 2.5 nM respectively) in a 5 mM Tris/Hepes, pH 7.7 buffer (0.5 mL or 1 mL respectively) in the absence (total

135 binding) or in the presence of the competing drug for 30 min at 25 C. The incubation was terminated by ltration over GF/B (Whatman) glass bre presoaked in 0.05% polyethyleneimine (Aldrich) with an MR24 Brandel cell harvester. The lters were rinsed three times with 5 mL 50 mM NaCl, Tris HCl 10 mM, pH 7.7 buffer and the radioactivity retained was counted in 3.5 mL ACS (Amersham) with an Excel 1410 (LKB) liquid scintillation spectrophotometer. The non-specic binding was determined in parallel experiments in the presence of 100 M unlabelled TCP. 5.2.2. Data analysis In each experiment, values are the mean of three independent determinations. Each experiment was performed 35 times. The data from competition experiments were rst analyzed by Hills representation according to a single-site model, then by a non linear regression method (MarquardtLevenberg algorithm) according to a two-site model using the Sigmaplott 4 software (Jandel). The two-site interaction was represented by: [LB] = (B1 + B2) {([I] B1/(IC501 + [I])) + ([I] B2/(IC502 + [I]))} where [LB] was the percentage of radioligand concentration specically bound, [I] the competitor concentration, B1 and B2 the percentage of each binding site, IC501 and IC502 the concentrations of unlabeled competitor that inhibited 50% of specic [3H]TCP binding on specied sites. Two constraints were xed: (i) IC501, IC502, B1, B2 > 0; (ii) 95% < (B1 + B2) < 105% because of the uncertaincy on the total binding (10%). Experimental results were submitted to an ANOVA test followed by a DurbinWatson test. The two-site model was preferred when it produced a signicant reduction in the sum of squares (P < 0.05, Students t-test) and when the DurbinWatson coefficient was closer to 2 than in the single-site model and comprised between 1.5 and 2.5. Acknowledgements This work was supported by D.R.E.T. (grant 94/141). Thanks are due to M. Michaud for her excellent technical assistance. References
[1] Vignon J., Chicheportiche R., Chicheportiche M., Kamenka J.M., Geneste P., Lazdunski M., Brain Res. 280 (1983) 194197. [2] Wong E.H.F., Kemp J.A., Priestley T., Knight A.R., Woodruff G.N., Iversen L.L., Proc. Natl. Acad. Sci. USA 83 (1986) 71047108. [3] Vignon J., Privat A., Chaudieu I., Thierry A., Kamenka J.M., Chicheportiche R., Brain Res. 378 (1986) 133141. [4] Ebert B., Wong E.H.F., Krogsgaard-Larsen P., Eur. J. Pharmacol. 208 (1991) 4952. [5] Berman F.W., Murray T.F., J. Biochem. Toxicol. 11 (1996) 217226. [6] Choi D.W., Cerebrovasc. Brain Metab. Rev. 2 (1990) 105147. [7] Hamon J., Vignon J., Kamenka J.M., Eur. J. Med. Chem. 31 (1996) 489495. [8] Vignon J., Chaudieu I., Allaoua H., Journod L., Javoy-Agid F., Agid Y., Chicheportiche R., Life Sci. 45 (1989) 25472555. [9] Bruylants P., Bruylants P., Bull. Soc. Chim. Belg. 33 (1924) 467478; Bull. Soc. Chim. Belg. 35 (1926) 139154. [10] Geneste P., Kamenka J.M., Dessapt P., Bull. Soc. Chim. Fr. 3& 4 II (1980) 186191. [11] Ilagouma A.T., Dornand J., Liu C.F., Znone F., Mani J.C., Kamenka J.M., Eur. J. Med. Chem. 25 (1990) 609615. [12] Minato A., Tarrao K., Hayashi T., Susuki K., Kumada M., Tetrahedron Lett. 52 (1981) 53195322. [13] Coderc E., Martin-Fardon R., Vignon J., Kamenka J.M., Eur. J. Med. Chem. 28 (1993) 893898. [14] Marwaha J., Palmer M., Hoffer B., Freedman R., Rice K.C., Paul S., Skolnick P., Arch. Pharmacol. 315 (1981) 203209. [15] Coderc E., Cerruti P., Vignon J., Rouayrenc J.F., Kamenka J.M., Eur. J. Med. Chem. 30 (1995) 463470. [16] Geneste P., Kamenka J.M., Org. Magn. Res. 7 (1975) 579584. [17] Kamenka J.M., Geneste P., In: Domino E.F. (Ed.), PCP (Phencyclidine): Historical and Current Perspectives, NPP Books, Ann Arbor, 1981, pp. 4782. [18] Manoharan M., Eliel E.L., Carrol F.I., Tetrahedron Lett. 24 (1983) 18551858. [19] Manoharan M., Eliel E.L., Kanan W., Tetrahedron Lett. 25 (1984) 32673268. [20] Michaud M., Warren H., Drian M.J., Rambaud J., Cerruti P., Nicolas J.P., Vignon J., Privat A., Kamenka J.M., Eur. J. Med. Chem. 29 (1994) 869876. [21] Geneste P., Kamenka J.M., Ung S.N., Herrmann P., Goudal R., Trouiller G., Eur. J. Med. Chem. 14 (1979) 301308. [22] Young T.E., Heitz L.J., J. Org. Chem. 38 (1973) 15621566. [23] Baty J.D., Jones G., Moore C., J. Chem. Soc. (C) (1967) 26452647. [24] Pouchert C.J., Behnke J. (Eds.), The Aldrich library of 13C and 1H FT NMR Spectra, Vol. 1, Aldrich Chemical Company, 1993, pp. 235236. [25] Conley E.C., In: Conley E.C. (Ed.), The Ion Channel Facts Book I, Academic Press, New York, 1996, pp. 140233. [26] Sekiguchi M., Okamoto K., Sakai Y., J. Neurosci. 10 (1990) 21482155. [27] Yamakura T., Mori Y., Masaki H., Shimoji K., Mishima M., Neuroreport 4 (1993) 687690. [28] Chaudieu I., Vignon J., Chicheportiche M., Kamenka J.M., Trouiller G., Chicheportiche R., Pharmacol. Biochem. Behav. 32 (1989) 699705. [29] Moriyoshi K., Masu M., Ishii T., Shigemoto R., Mizuno N., Nakanishi S.,, Nature 354 (1991) 3137. [30] Durand G.M., Bennett M.V.L., Zukin R.S., Proc. Natl. Acad. Sci. USA 90 (1993) 67316735. [31] Widdowson P.S., Trainor A., Lock E.A., J. Neurochem. 64 (1995) 651661. [32] Lynch D.R., Lawrence J.J., Lenz S., Anegawa N.J., Dichter M., Pritchett D.B., J. Neurochem. 64 (1995) 14621467. [33] Eiden F., Schmidt M., Buchborn H., Arch. Pharmacol. 320 (1987) 348361.

Eur. J. Med. Chem. 34 (1999) 671685 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

671

Invited review

Inhibitors of 5-lipoxygenase: a therapeutic potential yet to be fully realized?

Robert N. Young*
Department of Medicinal Chemistry, Merck Frosst Centre for Therapeutic Research, P.O. Box 1005, Pointe Claire - Dorval, Qubec H9R 4P8, Canada (Received 21 June 1999; accepted 21 June 1999)

Abstract Inhibition of leukotriene biosynthesis has been extensively studied as a potential for the development of novel therapies for inammation and respiratory diseases and, in particular, for asthma. Many compounds have been identied which inhibit the key enzyme, 5-lipoxygenase. Four distinct classes of compounds have been identied, namely, (1) redox inhibitors (alternative substrates), (2) iron chelating inhibitors, (3) competitive reversible inhibitors, and (4) inhibitors of the 5-lipoxygenase activating protein. Experience over the past two decades with redox inhibitors has been disappointing and although a number of potent compounds have been identied, they have often been associated with ancillary toxicity and non-specicity due to their redox activity. Iron chelating inhibitors have been more successful and one compound, Zileutont, has reached the market. However, more potent analogues have often encountered toxicity problems. Competitive inhibitors have been identied by a number of research groups but, as yet, none has been successful. Inhibitors of the 5-lipoxygenase activating protein (FLAP) have been identied and compounds such as MK-0591 and BaY-X-1005 have shown efficacy in asthma trials. To date, however, no clear advantage for inhibitors of lipoxygenase has been demonstrated relative to the leukotriene D4 receptor antagonists such as Singulairt and Accolatet. 1999 ditions scientiques et mdicales Elsevier SAS 5-lipoxygenase / inhibitor / 5-lipoxygenase activating protein / asthma / inammation

1. Introduction After the initial characterization of the slow-reacting substance of anaphylaxis in 1940 [1], nearly four decades passed before the characterization by Professor Bengt Samuelsson of the Karolinska Institute [2, 3] of these potent contractile substances as novel peptido-lipid hybrids, which became known as leukotrienes C4, D4 and E4. These substances and a companion lipid which became known as leukotriene B4 were proposed by Samuelsson to be derived from a common epoxide intermediate which he named leukotriene A4. Samuelssons biosynthetic detective work, which led to the proposal of the leukotriene biosynthetic pathway (gure 1) was a scientic triumph. Considering the potent pro-inammatory properties of leukotriene B4 [4] and the multiple activities of the peptido-lipid leukotrienes LTC4, D4 and E4, it became apparent that modulation of this pathway could have important implications in the development of novel therapeutics for diseases such as asthma,
*Correspondence and reprints

allergy and a host of other inammatory diseases. It was recognized in the course of this work that a key enzyme in the process, 5-lipoxygenase, could transform arachidonic acid (AA) in a two-step process to rst, 5-hydroperoxyeicosatetraenoic acid (5-HPETE), and thence through a dehydration step to leukotriene A4. This common unstable intermediate is then taken on to leukotriene B4 via leukotriene B4 synthase in certain cells, such as neutrophils, or could be converted via a specic glutathione-transferase enzyme (leukotriene C4 synthase) to provide the peptido-lipid leukotrienes C4, D4 and E4 in cells such as eosinophils [5]. In the early 1980s, a number of pharmaceutical research laboratories throughout the world recognized that novel therapeutics could be derived from the development of inhibitors of 5-lipoxygenase (which would potentially modulate the entire pathway) or from receptor antagonists at the specic leukotriene receptors. At the onset, it was not clear which of these alternative approaches would prove most successful, although, as of this date, it would appear that the latter approach, namely

672

Figure 1. Leukotriene biosynthetic pathway.

673 the development of leukotriene D4 receptor antagonists, has been most successful. Nonetheless, the efforts that have gone into the development of specic inhibitors of the 5-lipoxygenase enzyme have been extensive and have yielded at least one novel therapeutic agent, namely Zileuton. A number of reviews on the development of 5-lipoxygenase inhibitors have been published in the past [68]. This review will attempt to provide an overview of what has been learned in the course of these efforts and, hopefully, provide a useful perspective for future drug development. 2. 5-Lipoxygenase mechanism A detailed review on the 5-lipoxygenase mechanism has been published [7] and doesnt bear detailed repeating here. However, a number of key factors are pertinent in understanding the efforts to discover a novel and safe inhibitor of 5-lipoxygenase. 5-lipoxygenase was found to be a cytosolic enzyme which contained a non-haem iron atom. For maximum activity, it is necessary that the enzyme be converted from an inactive reduced state (Fe(II) state (Er)) to the active Fe(III) state (Eo) through interaction with an oxidizing agent such as fatty acid hydroperoxide, AOOH. The oxidized enzyme then interacts with the fatty acid substrate (AH) to yield the Fe(II) enzyme with a bound pentadienyl radical (A), which then reacts with molecular oxygen to yield a hydropyroxy radical (A00), which then picks up a hydrogen atom to yield the fatty acid hydroperoxide and to regenerate the oxidized enzyme. The exact mechanism whereby the 5-HPETE is then converted to leukotriene A4 is not fully understood but is catalysed in a second rapid step by the same enzyme (gure 2). The enzyme itself is subject to turnover inactivation, presumably through the generation of reactive radical biproducts. The arachidonic acid substrate is generally limiting in the system and is generated near the cell wall surface by the enzyme phospholipase A2. Processes are in place to rapidly reacylate liberated arachidonic acid, such that free levels of arachidonic acid are normally very low in the cells. Thus, in order to have optimal activity, the 5-lipoxygenase enzyme translocates under mediation of a variety of factors, such as calcium levels in the cell, to interact with the cell membrane. Thus, a variety of approaches can be considered for the development of inhibitors of 5-lipoxygenase. Considering the redox cycling nature of the enzyme, it should be possible to inhibit the enzyme by competition with an alternative substrate which would itself be oxidized through radical transfer, and thus, divert the enzyme from its normal task. Such a substrate could itself produce reactive intermediates which could facilitate turnover inactivation of the enzyme. The inhibitor radicals thus produced, if stable, could be cycled in their own right or go on to decompose. Alternatively, inhibitors which are not substrates for the enzyme could interact with either the reduced state or the oxidized state of the enzyme (or both) and form reversible dead-end inhibitor-enzyme complexes. Finally, it could be possible to inhibit the translocation of 5-lipoxygenase to the membrane where it obtains substrate or to inhibit the transfer of the substrate to the enzyme. Although the details of these possibilities were not known as research began in the early 1980s to nd inhibitors of 5-lipoxygenase, inhibitors in all three of these classes have been discovered and, in some cases, shown to be useful drugs. 3. Redox inhibitors of 5-lipoxygenase (gure 3) Many small organic compounds such as phenols, quinones, dihydroquinones, etc., can interact in redox mechanisms. Early screening to nd lipoxygenase inhibitors employed cell models such as human or rat polymorphonuclear cells (PMNs) stimulated with calcium ionophores, measuring the production of leukotriene B4. These initial efforts were rewarded by the discovery of a wide variety of inhibitors, some of which showed quite potent and apparently selective inhibition (selectivity was generally determined relative to other oxygenase enzymes such as cyclooxygenase or other lipoxygenases such as 12-lipoxygenase). Considering the lipophilic nature of the substrate, these inhibitors were generally small lipophilic molecules such as mono- and polycyclic aromatics. In our own laboratories, we were initially elated to discover the tricyclic benzothiazinone, L-615,919, which showed nanomolar potency for inhibition of leukotriene synthesis in PMN cells stimulated with calcium ionophore. The compound also showed bioavailability via the oral route and biochemical activity ex vivo. Our initial elation, however, was rapidly quenched when it was found that the compound promoted methaemoglobin information in the blood of dogs treated with the drug. It was apparent that the drug not only interfered with the redox cycling of 5-lipoxygenase but also could serve to convert iron in haemoglobin to the oxidized Fe(III) state producing profound toxicity. This was the rst indication of problems of non-specicity and ancillary activity due to redox cycling that were to plague the development of redox inhibitors of 5-lipoxygenase for many years to come. Our initial feelings were that L-615,919 was too redox-active and the product of its reduction (a dihydroaminoquinone) would be too powerful an oxidizing

674

Figure 2. A kinetic mechanism of 5-lipoxygenase: oxygenase activity (right hand side), results from the reaction of the Fe(II) enzyme (ER) with the arachidonic acid substrate (AH) to yield the Fe(II) enzyme (ER) with a bound arachidonyl radical (A), followed by binding of O2 and reaction to yield hydroperoxyl radical (AOO) bound to ER and then regeneration of EO and release of 5-hydroperoxyeicosatetraenoic acid (AOOH). Free ER can result from dissociation of A from ER(A) or from reduction of EO by various reducing agents (RH). The product of one electron oxidation of RH (R) can be reduced by thiols. Reduced enzyme (ER) can be reoxidized by fatty acid hydroperoxide (AOOH) to yield a hydroxide ion and an alkoxy radical (AO). Non-redox reversible dead end inhibitors (I) can bind to ER, EO and possibly other states of the enzyme as well.

agent, and thus shows the observed toxicity. Examination of the literature indicated a number of redox-active substances such as menadione [9] which were known to cause methaemoglobinaemia and also to produce superoxide anions which led to other manifestations of toxicity such as Heinz body formation and haemolysis of blood cells. Our efforts were then directed to produce compounds with less potent redox activity hoping that these types of toxicities could be differentiated from the inhibition of the lipoxygenase enzyme itself. Modulation of redox potential via substitution led to the discovery of L-651,392 [10] which did not show methaemoglobin

formation in dog blood. The compound was, however, a potent 5-lipoxygenase inhibitor (IC50 = 60 nM (rat PMN)). Unfortunately, the compound was only poorly soluble and poorly absorbed. It also showed a variety of toxicities including genotoxicity and therefore the compound was abandoned. Further studies on phenothiazinone analogues including the naphthyl analogue, such as L-649,927, provided compounds which were free of the tendency to form methaemoglobin, although, again, poorly absorbed. A prodrug carbonate, L-654,623, with better solubility and bioavailability which showed good in vitro activity was identied (Y. Girard, unpublished

675

Figure 3. Redox inhibitors of 5-lipoxygenase.

results). Unfortunately, again, problems were encountered in toxicological evaluation and L-654,623 was shown to cause Heinz body formation in blood cells in dogs. As this kind of toxicity could be attributed to the redox activity of this series, further work on phenothiazi-

nones was stopped. In later work, a number of hydroxybenzofurans were identied through screening. Further studies and optimization led to the identication of L-656,224 [11, 12]. Although the compound showed good in vitro and in vivo activity, toxicological evaluation

676 revealed indications of hypersensitivity reactions, apparently due to metabolic conversion to quinone-type metabolites [13], and development was suspended. Further efforts at Merck Frosst derived a second dihydrobenzofuranol, L-670,630 [14]. Unfortunately, again, signs of toxicity similar to that observed for L-656,224 were observed and investigation of redox inhibitors was abandoned at Merck Frosst in preference for a search for competitive non-redox inhibitors. Other companies have investigated a variety of redox inhibitors over the years, although none to date have been brought to market. The quinone AA-861 has been reported to be in clinical development but little recent information is available [15, 16]. Other phenolic 5-lipoxygenase inhibitors have entered development and clinical trials, such as, TMK-688 (linazolast) [17], DuP654 [18], R-68151 [19] and E-6080 [20]. None, however, have proceeded to market to this date and presumably have encountered a variety of problems. Many redoxactive heterocyclic compounds such as BW-755C [21], ICI-207968 [22] and A-53612 [23] have been reported as 5-lipoxygenase inhibitors over the years. However, studies have shown that compounds of this type are prone to one electron oxidation [24] and can cause methaemoglobin formation in blood [22] and, thus, they apparently have not been developed further. Recently, a study on a series of tetrahydro-1,2,4triazien-3-ols has been published by workers at Abbott [25] which are reported to be free of methaemoglobinaemia in rats. This suggests that it may be possible to derive such inhibitors free of the toxicity problems observed in related phenothiazinone and phenol series of 5-lipoxygenase inhibitors. The heterocyclic phenol from Jansen, R-68151, has been reported to be in clinical trials for psoriasis by the topical route and has apparently shown mild to moderate therapeutic effect [26]. A more potent analogue, R-85355, was also investigated for psoriasis but, unfortunately, was found not to show signicant clinical activity [27]. 4. Iron chelator inhibitors (gure 4) The most successful efforts to derive non-toxic redox type inhibitors of 5-lipoxygenase have been in the area of hydroxamic acids and related N-hydroxy ureas. These compounds were designed basically with the expectation that the functional groups might chelate iron and therefore inhibit the enzyme. Studies in this area have been pursued by researchers from Glaxo/Burroughs Wellcome, Abbott Laboratories, SmithKline Beecham, Wyeth Ayerst, Ciba Geigy and many others. Early studies of N-acylhydroxylamine compounds led to the discovery of BW-A4C [28] and A-63162 [29] as potent and selective 5-lipoxygenase inhibitors. However, BW-A4C was found to be rapidly metabolized in humans [30] and shown to be oxidized by 5-lipoxygenase to form nitroxide radicals [31]. It was also found that O-glucuronidization was a problem in this series, as well as the hydrolysis of the hydroxamic acid which impaired in vivo activity [30]. Researchers at Abbott continued extensive studies in this area concentrating on an analogous series of N-hydroxyureas which were more hydrolytically stable, had reduced glucuronidization and therefore superior in vivo properties. These efforts ultimately yielded A-64077 (Zileuton) which has undergone extensive clinical evaluation [32]. Zileuton has shown efficacy in chronic asthma where it provided some degree of bronchodilation and anti-inammatory and steroid sparing effects [33]. The compound was brought to market in 1996 as the rst of the new class of anti-leukotriene drugs as a 600 mg q.i.d. dose. The compound, however, has shown a variety of adverse effects including elevated liver enzymes and other hepatic toxicities as well as signicant drug interactions [32]. Zileuton has been shown to induce a variety of liver enzymes, including P450-2B and P450-4A, and induces hepatomegaly on chronic treatment in rats [34]. Recent reports have indicated that Zileuton showed efficacy in a rat model of ulcerative colitis [35]. However, a subsequent clinical trial at a dose of 600 mg q.i.d. for six months failed to show activity signicantly better than placebo, and showed less efficacy when compared with mesalazine [36]. Efforts by Abbott researchers have derived a number of more potent analogues of Zileuton with better pharmacokinetics which promise the potential for a q.d. drug with doses lower than required for Zileuton. Studies derived an optimized compound, A78773, and its more potent R(+) enantiomer, A79175 [37] which was reported to have entered clinical trials [38, 39]. A phenol analogue, A-76745 (fenleuton) has been evaluated as potential treatment for allergic and inammatory disorders in horses [40]. The compound has been reported to have some efficacy in a horse model of chronic obstructive pulmonary disease [41]. A thiophene analogue (ABT761) (atreleuton) has apparently proceeded to Phase III trials as a second generation hydroxyurea 5-lipoxygenase inhibitor and is reported to show potent biochemical efficacy for nine hours after a single oral dose of 200 mg. The compound showed efficacy in blocking bronchoconstriction in asthmatics following a challenge by exercise [42]. ABT-761 has a long elimination half-life of about 15 hours consistent with once-a-day dosing [43]. However, the compound has been reported to show some

677

Figure 4. Iron chelator inhibitors of 5-lipoxygenase.

drug interaction with oral contraceptive steroids [44]. It remains to be seen whether atreleuton will reach the market. Structure activity studies on the research which derived atreleuton have been published [45]. SmithKline Beecham have studied a series of dihydrobenzofuran-N-hydroxy ureas (which have derived SB-210661 and SB-202235) [46, 47]. These compounds have not yet been reported to have reached clinical trials. Ciba-Geigy/Novartis have reported on a series of heterocyclic hydroxy ureas including CGS-26529 [48] and CGS 23885 [49, 50], the latter compound has been

reported to have been in Phase I clinical trials. Wyeth Ayerst researchers have reported structure-activity studies on a series of azophenoxyhydroxy ureas [51], however, the developmental stages of this series is unknown. Although the N-hydroxy ureas have been designed as iron chelators, it has been shown that these compounds can be turned over by the 5-lipoxygenase enzyme to derive radical by-products. In particular, Zileuton has been shown to be metabolized to form nitroxides and, as such, serves as a reducing substrate for 5-lipoxygenase [52]. If the degree of turnover by the enzyme

678 correlates with potency of these compounds, it is possible that potent analogues of Zileuton will continue to be plagued by toxicities derived from production of radicals and radical by-products. 5. 5-Lipoxygenase inhibitors with dual activities (gure 5) Many companies have reported over the last decades on compounds that inhibit both 5-lipoxygenase and other enzymes involved in inammation as approaches to treat a variety of inammatory diseases. BW-B70C has been reported to have dual 5- and 15-lipoxygenase inhibitor activity [53]. The compound showed some efficacy in allergic models in guinea-pigs. The phenolic benzothiazole analogue, E-3040, from Eisai, was reported to have potent dual 5-lipoxygenase and thromboxane A2 synthase inhibitory activity. The compound has been evaluated in models of experimental ulcerative colitis [54] and has been reported to be in Phase II clinical trials. Another quinonoid dual inhibitor of 5-lipoxygenase and thromboxane A2 synthase (CV-6504) has been reported to show anti-tumour activity in murine models of adenocarcinoma [55]. Most efforts, however, have been directed towards development of dual 5-lipoxygenase/cyclooxygenase inhibitors. The compound tepoxaline, or RWJ-20485, has received extensive biochemical and clinical evaluation by

Figure 5. Inhibitors of 5-lipoxygenase with dual activities.

679 Johnson & Johnson as an anti-inammatory agent. However, clinical trials indicated potent cyclooxygenase inhibition at doses from 25800 mg p.o., whereas only weak inhibition of leukotriene synthesis was observed at the maximum 800 mg dose [56]. Development of the compound has been reported to be discontinued (company communication, March 1995). More recently, Johnson & Johnson have reported a cyclooxygenase 2/5-lipoxygenase dual inhibitor, RWJ-63556 [57]. The company Merckle has reported on a series of perazolene analogues as dual cyclooxygenase/5-lipoxygenase inhibitors from which they derived the clinical candidate ML-3000. The compound has shown some oral activity in rat models of inammation [58] and is reported to work by a non-redox mechanism [59] and to show activity in a sheep model of asthma [60] by aerosol administration. The compound is reported to be in Phase II clinical trials. Another dual cyclooxygenase/5-lipoxygenase inhibitor, VUFB-16066 (obufen), has reportedly been evaluated as inammatory in clinical trials for rheumatoid arthritis [61]. Researchers from Cytomed have recently reported on the hybridizaton of N-hydroxyurea functionality to diaryltetrahydrofuran to derive a novel series of compounds showing dual 5-lipoxygenase and platelet activating factor receptor (PAF) antagonism. An optimized compound CMI-392 was reported [62]. Although many companies continue to try and develop dual inhibitors, it is clear that the difficulty to balance the inhibitory activity toward two different targets in vivo has made it difficult to optimize compounds in this area. 6. Competitive (non-redox) inhibitors of 5-lipoxygenase (gure 6) The multiple toxicities and difficulties encountered in developing redox inhibitors of 5-lipoxygenase has led many research groups to strive to nd competitive nonredox inhibitors of the enzyme. Only as a better understanding of the mechanism of the enzyme became available, has it become possible to develop reliable criteria for the evaluation and identication of non-redox inhibitors [52]. The observation by ICI (Zeneca) that a series of methoxyalkylthiazoles [63] and methoxytetrahydropyrans [64] were potent inhibitors of 5-lipoxygenase which, in some cases, showed enantioselective activity [63, 65] suggested they might act by stereospecic binding at the active site of 5-LO. Structural features suggested they were unlikely to enter into redox reactions. Subsequent studies showed that compounds in this series, such as ZM-211965, did not act as reducing substrates for 5-lipoxygenase, while they did inhibit turnoverdependent inactivation of the enzyme and therefore could be considered as true non-redox inhibitors [52]. Optimization of these compounds led to the discovery of ZD-2138 [65, 66]. This compound was evaluated extensively by Zeneca in clinical trials for arthritis and asthma, and in normal volunteers. For example, a 350 mg p.o. dose was shown to maximally inhibit leukotriene synthesis for at least 24 hours [67]. Unfortunately, in spite of this potent activity, Zeneca has recently reported to have discontinued development of this compound due to mixed and unconvincing results as an anti-arthritis agent [68]. It may still be in development for asthma. An analogue, ZM-230487, has also been evaluated preclinically by Zeneca [69, 70] and more recently structureactivity relationships on a series of thiene analogues have been published [71]. Recent studies have suggested that compounds such as ZM-230487 are potent inhibitors of 5-lipoxygenase under conditions of low peroxide tone and are less efficient under conditions of oxidative stress [72]. This could explain the disappointing activity observed for these compounds as anti-inammatory agents. A related series of non-redox inhibitors of 5-LO has been discovered by Merck Frosst scientists. Screening of the Merck sample collection identied a class of lignans derived from a natural product, Justicidin E, as a moderately potent non-redox inhibitor 5-lipoxygenase [73]. Based on similarities between these compounds and ZD-2138, hybrid molecules such as L-697,198 were prepared which were not only markedly more potent than the initial lignans, but also by virtue of the possibility of dosing as a prodrug dihydroxyacid form, showed excellent bioavailability and oral activity [74]. However, extensive metabolism of the pyran ring in these compounds [75] complicated their development. Structureactivity studies were carried out with the intention of deriving a metabolically stable analogue. These efforts resulted in the identication of L-708,780 [76], L-739,010 [77] and L-746,530 [78]. These compounds showed only traces of metabolism and had excellent bioavailability and duration in a number of animal species. Unfortunately, it was noted that recovery of drug under in vitro metabolic conditions was poor for these compounds and subsequent studies showed that both the bicyclo moiety and the furan moiety were metabolized to derive reactive metabolites which labelled plasma and hepatic proteins [79, 80]. Toxicological evaluation of L-739,010 showed a variety of toxicities [81] and, thus, development of these compounds was discontinued. Other studies in the Merck Frosst Laboratories have identied a series of thiopyranylindoles [82] structurally related to series which had been shown to indirectly inhibit lipoxygenase activity in cells (vide infra). These

680

Figure 6. Competitive (non-redox) inhibitors of 5-lipoxygenase.

681 compounds, exemplied by L-689,065 [83], showed potent in vitro activity against human 5-lipoxygenase and optimized compounds L-691,816 [84] and L-699,333 [85] were subsequently identied with excellent in vivo activity in a variety of models of asthma when dosed by the oral route. Development of this series was, however, inhibited by the modulation of ex vivo activity through binding to plasma proteins in blood and by competing development of a series of indirect inhibitors in the same laboratory (see following discussion). 7. Inhibitors of the 5-lipoxygenase activating protein (FLAP) (gure 7) Early screening studies in PMN cell assays at Merck Frosst looking for novel inhibitors of leukotriene biosynthesis identied a new class of indole inhibitors which showed potent inhibitory activity in intact PMN cells. Optimization of this series of indoles led to the discovery of L-663,536 (or MK-886) [86]. This was a specic inhibitor of leukotriene biosynthesis in a variety of intact

Figure 7. Inhibitors of the 5-lipoxygenase activating protein (FLAP).

682 cell preparations but had no effect on either cyclooxygenase or 15- or 12-lipoxygenase derived products. In broken cell preparations or on puried 5-lipoxygenase, the compound had no activity. In addition, the compound had no activity on phospholipase A2 and did not inhibit arachidonic acid release. The compound also showed excellent activity in a variety of animal models of asthma. Attempts to nd the molecular target for MK-886 led to the preparation of a photoaffinity label 125I-L-669,083 which was shown to label an 18 Kd protein in cell membrane preparations [87]. In addition, MK-886 and a variety of analogues inhibited the binding of the photo probe to this protein in a dose-dependent and potencydependent manner. An analogue of MK-886 was used to prepare an affinity column from which it was possible to isolate this 18 Kd protein and based on sequence information attained from the rat protein, the cDNA for both the rat and human protein were isolated, cloned and expressed [87, 88]. Subsequent studies indicated that this novel protein (termed 5-lipoxygenase activating protein or FLAP) is a necessary factor which must be present in cells to facilitate the transfer of arachidonic acid to 5-lipoxygenase [89, 90]. Cells which contain 5-lipoxygenase and not FLAP are unable to biosynthesize leukotriene products unless provided with a large excess of exogenous arachidonic acid substrate. It was thus proposed that MK-886 inhibited the biosynthesis of leukotrienes in PMNs and other cells by blocking the FLAPmediated transfer of the arachidonic acid substrate to the enzyme. Other photo-affinity studies showed that arachidonic acid binds to FLAP and this binding is competed by MK-886 and related analogues [91]. The discovery of this protein led to the possibility of setting up a binding assay whereby inhibitors of arachidonic acid binding to FLAP could be evaluated directly [92]. MK-886 itself entered clinical trials for asthma and while it showed some efficacy, the results were somewhat disappointing [93]. At maximally tolerated doses, however, inhibition of leukotriene biosynthesis was not complete [94] and it was felt that a more potent compound would be necessary. After the discovery of MK-886, a survey of the literature identied a weak inhibitor, REV-5901 [95], a quinoline structure, which appeared to have similar activity to MK-886 in that it was active in whole cells but not in broken cells. Evolution of structures related to REV-5901 identied a series of quinoline inhibitors characterized by L-674,636 [96]. These compounds, though relatively potent and orally absorbed, were plagued by signicant modulation of their activity through binding to plasma proteins. Observation of similarities between the quinoline series and the indole series of MK-886 led to the investigation of hybrid structures [97]. The utilization of the FLAP binding assay allowed optimization of analogues of MK-886 and the discovery of a signicantly more potent FLAP inhibitor, L-686,708 (MK-0591) [98100]. This compound inhibited FLAP binding with a potency at least 10-fold greater than MK-886, and showed potent inhibition of LT biosynthesis in stimulated whole blood and also showed good oral absorption and pharmacokinetics and was less attenuated by plasma proteins. MK-0591 has been extensively evaluated in clinical trials where it showed efficacy both for biochemical inhibition of leukotriene biosynthesis in vivo, as measured by excretion of LTE4 in urine in asthmatic patients, and in ex vivo assays as measured by inhibition of leukotriene B4 biosynthesis in stimulated whole blood [101]. Subsequent trials in chronic asthma showed efficacy by a variety of parameters including increase of FEV1, decrease in -agonist usage and symptom scores, although a small incidence of rash was observed as well [102]. The compound was administered under a b.i.d. regimen at 125 mg dose. Although MK0591 showed clinical efficacy and oral activity, its development was suspended by Merck in favour of the leukotriene D4 receptor antagonist MK-0476 (montelukast), which had superior activity and could be given by a once-a-day regimen. MK-0591 has also been evaluated in clinical trials in ulcerative colitis. Again, MK-0591 showed excellent biochemical efficacy [103, 104]. However, similar to trials previously described for Zileuton, clinical efficacy in the disease was marginal and not statistically signicantly different than placebo [104]. Other laboratories have pursued the quinoline series of leukotriene biosynthesis inhibitors and, in particular, Bayer have derived compounds which have been evaluated in clinical trials. BAY-X-1005 has been extensively evaluated in allergen and cold air challenged asthmatics in clinical trials where it has shown efficacy [105, 106]. In spite of this activity, the compound has been reported to show rather limited anti-inammatory potential [107]. The compound has been reported to be in Phase II clinical trials and with Phase III trials planned, no recent information has been available as to its status. A related compound from Bayer, BaY-Y-1015, has been investigated in animal models of IBD where it has shown efficacy better than olsalazine [108] but, to date, no reports of human clinical evaluation have been published. 8. Summary Intensive efforts to develop clinically useful drugs from inhibitors of the 5-lipoxygenase enzyme or of leukotriene biosynthesis have been carried out for almost

683 two decades now. As only one marketed drug, Zileuton, has emerged from this massive effort, this speaks strongly to the high risk and great difficulty of developing new therapies today. It is apparent from clinical results on MK-0591, BaY-X-1005 and ABT-761, that a variety of leukotriene biosynthesis inhibitors could provide useful therapy in asthma [109]. However, other issues of toxicity, pharmacokinetics or tolerability have frequently blocked bringing such compounds to the market. Of great importance has been the success of two leukotriene receptor antagonists, namely, zafraleukast and montelukast which have also shown good efficacy in the treatment of asthma [110113]. To date, the leukotriene receptor antagonists appear to have superior properties of safety, pharmacokinetics and, to some degree, efficacy. It remains to be seen whether there are disease states or situations where the more complete blockade of leukotriene biosynthesis, including the elimination of leukotriene B4, may prove to be an advantage. Acknowledgements The author wishes to acknowledge the efforts of Diane Sauv and Mary Lynn Gaal for their help in preparing this manuscript. References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] Kellaway C.H., Trethewie E.R., Q. J. Exp. Physiol. 30 (1940) 121145. Murphy R.C., Hammarstrm S., Samuelsson B., Proc. Natl. Acad. Sci. USA 76 (1979) 42754279. Samuelsson B., Borgeat P., Hammarstrm S., Murphy R.C., Prostaglandins 17 (1979) 785787. Ford-Hutchinson A.W., Bray M.A., Doig M.V., Shipley M.E., Smith M.J.H., Nature (London) 286 (1980) 264265. Ford-Hutchinson A.W., in: Rokach J. (Ed.), Leukotrienes and Lipoxygenases, Elsevier, Amsterdam, 1989, pp. 405425. Fitzsimmons B., Rokach J., in: Rokach J. (Ed.), Leukotrienes and Lipoxygenases, Elsevier, Amsterdam, 1989, pp. 427502. Ford-Hutchinson A.W., Gresser M., Young R.N., Annu. Rev. Biochem. 63 (1994) 383417. Bell R.L., Summers J.B., Harris R.R., Annu. Rep. Med. Chem. 32 (1997) 91100. Goldberg B., Stern A., Biochim. Biophys. Acta 437 (1976) 628632. Guindon Y., Girard Y., Maycock A., Ford-Hutchinson A.W., Atkinson J.G. et al., Adv. Prostaglandins Thromboxane Leuk. Res. 17A (1987) 554557. Blanger P., Maycock A., Guindon Y., Back T., Dollob A.L. et al., Can. J. Physiol. Pharmacol. 65 (1987) 24412448. Lau C.K., Blanger P.C., Scheigetz J., Dufresne C., Williams H.W.R. et al., J. Med. Chem. 32 (1989) 11901197. Rouzer C.A., Riendeau D., Falgueyret J.-P., Lau C.K., Gresser M.J., Biochem. Pharmacol. 41 (1991) 13651373. [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] Lau C.K., Blanger P.C., Dufresne C., Scheigetz J., Thrien M. et al., J. Med. Chem. 35 (1992) 12991318. Ancill R.J., Takahashi Y., Kibune Y., Campbell R., Smith J.R., J. Int. Med. Res. 18 (1990) 7588. Fujimura M., Sasaki F., Nakatsumi Y., Takahashi Y., Hifumi S. et al., Thorax 41 (1986) 955959. Scrip No. 1306 (1988) p. 28. Harris R.R., Batt D.G., Galbraith W., Ackerman N.R., Agents Actions 27 (1989) 297299. Marien K., Morren M., Degreef H., DeDoncker P., Rooman R.P., Cauwenbergh G., Arch. Dermatol. 128 (1992) 993994. Beetens J.R., Somers Y., Goossens J., Vanwauwe J., De Clerck F., Arch. Int. Pharmacodyn. Ther. 290 (1987) 151. Fort F.L., Pratt M.C., Carter G.W., Lewkowski J.P., Heyman I.A. et al., Fundam. Appl. Toxicol. 4 (1984) 216220. Bruneau P., Delvare C., Edwards M.P., McMillan R.M., J. Med. Chem. 34 (1991) 10281036. Albert D.H., Machinist J., Young P.R., Dyer R., Bouska J. et al., FASEB J. 3 (1989) A735. Duniec Z., Robak J., Gryglewski R., Biochem. Pharmacol. 32 (1983) 22832286. Bhatia P.A., Brooks C.D., Basha A., Ratajczyk J.D., Gunn B.P. et al., J. Med. Chem. 39 (1996) 39383950. Degreef H., Dockx P., DeDoncker P., De-Beule K., Cauwenbergh G., J. Am. Acad. Dermatol. 22 (1990) 751755. van de Kerkhof P.C., van Pelt H., Lucker G.P., Steijlen P.M., Heremans A., Skin Pharmacol. 9 (1996) 307311. Tateson J.E., Randall R.W., Reynolds C.H., Jackson W.P., Bhattacherjee P. et al., Br. J. Pharmacol. 94 (1988) 528539. Summers J.B., Gunn B.P., Martin J.G., Mattin M.B., Mazdiyasni H. et al., J. Med. Chem. 31 (1988) 19601964. Payne A.N., Jackson W.P., Salmon J.A., Nicholls A., Yeadon M., Garland L.G., Agents Actions Suppl. 34 (1991) 189199. Chamulitrat W., Mason R.P., Riendeau D., J. Biol. Chem. 267 (1992) 95749579. Wenzel S.E., Kamada A.K., Ann. Pharmacother. 30 (1996) 858864. Liu M.C., Dube L.M., Lancaster J., J. Allergy Clin. Immunol. 98 (1996) 859871. Rodrigues A.D., Machinist J.M., Toxicol. Appl. Pharmacol. 137 (1996) 193201. Zarif A., Eiznhamer D., Callaghan C., Doria M.I., Broutman L., Keshavarzian A., Inammation 20 (1996) 217227. Hawkey C.J., Dube L.M., Rountree L.V., Linnen P.J., Lancaster J.F., Gastroenterology 112 (1997) 718724. Bell R.L., Bouska J.B., Malo P.E., Lanni C., Harris R.R. et al., J. Pharmacol. Exp. Ther. 272 (1995) 724731. Bouska J. et al., 6th Int. Conf. Inamm. Res. Assoc., Sept. 2024, White Haven (1992) Abst. P119. Brooks D.W., Stewart A.O., Basha A., Bhatia P., Ratajczyk J.D. et al., 206th ACS Nat. Meeting, Aug. 2227, Chicago (1993) Abst. Medi 41. Annual Drug Data Report, 17 (No. 8) (1995) 711712. Marr K.A., Lees P., Page C.P., Cunningham F.M., Inamm. Res. 44 (Suppl. 1) Abst. W20/07, (1995) S289. Lehnigk B., Rabe K.F., Dent G., Herst R.S., Carpentier P.J., Magnussen H., Eur. Resp. J. 11 (1998) 617623. Wong S.L., Drajesk J., Chang M.S., Witt G., Awni W.M., Biopharm. Drug. Disp. 19 (1998) 159162. Wong S.L., ODea R., Dube L., Awni W.M., 99th Meeting Am. Soc. Clin. Pharm. Ther., New Orleans (1998) Abst. PI-106.

[40] [41] [42] [43] [44]

[11] [12] [13]

684
[45] [46] Brooks C.D., Stewart A.O., Basha A., Bhatia P., Ratajczyk J.D. et al., J. Med. Chem. 38 (1995) 47684775. Adams J.L., Garigipati R.S., Sorenson M., Ross S., Schmidt S.J. et al., 204th ACS Nat. Meeting, Aug. 2328, Washington (1992) Abst. Medi 121. Chabot-Fletcher M.C., Underwood D.C., Breton J.J., Adams J.L., Kagey-Sobotka A. et al., J. Pharmacol. Exp. Ther. 273 (1995) 11471155. Kimble E., Kowalski T., Peters P., Lee W., Koehler J. et al., Inamm. Res. 44 (Suppl. 2) (1995) S147S148. Satu Y., Stanton J.L., Hutchison A.J., Libby A.H., Kowalski T.J. et al., 205th ACS Nat. Meeting, March 28April 2; Denver (1993) Abst. Medi 143. Raychaudhuri A., Chertock H., Chovan J., Jones L.S., Kimble E.F. et al., Naunyn-Schmiedebergs Arch. Pharmacol. 355 (1997) 470474. Malamas M.S., Carlson R.P., Grimes D., Howell R., Glaser K. et al., J. Med. Chem. 39 (1996) 237245. Falgueyret J.-P., Hutchinson J.H., Riendeau D., Biochem. Pharmacol. 45 (1993) 978981. Yeadon M., Price R.C., Payne A.N., Pulmon. Pharmacol. 8 (1995) 5363. Oketani K., Murakami M., Inoue T., Yamanaka T., Yamatsu I., Jpn. J. Pharmacol. 64 (Suppl. I) Abst. (1994) P351. Hussey H.J., Bibby M.C., Tisdale M.J., Br. J. Cancer 73 (1996) 11871192. Depre M., VanHecken A., Verbesselt R., Verpooten G.A., Arnout J. et al., Int. J. Clin. Pharm. Res. 16 (1996) 18. Parnham M.J., Drug News Perspect. 9 (1996) 631639. Laufer S.A., Augustin J., Dannhardt G., Kiefer W., J. Med. Chem. 37 (1994) 18941897. Laufer S.A., Tries S., Augustin J., 13th Int. Symp. Med. Chem., Sept. 1923, Paris (1994) Abst. P181. Abraham W.M., Laufer S., Tries S., Pulmon. Pharmacol. Ther. 10 (1997) 167173. Panajotova V., Anderova E., Jandera A., Kuchar M., Arzneim. -Forsch 47 (1997) 648652. Cai X., Scannell R.T., Yaeger D., Hussoin M.S., Killian D.B. et al., J. Med. Chem. 41 (1998) 19701979. Bird T.G.C., Bruneau P., Crawley G.C., Edwards M.P., Foster S.J. et al., J. Med. Chem. 34 (1991) 21762186. Crawley G.C., Dowell R.I., Edwards P.N., Foster S.J., McMillan R.M. et al., J. Med. Chem. 35 (1992) 26002609. Lambert-van der Brempt C., Bruneau P., Lamorlette M.A., Forster S.J., J. Med. Chem. 37 (1994) 113124. Bird T.G.C., 12th Int. Symp. Med. Chem., Sept. 1317 Basel (1992) Abst. OC-05. 2. Walker E.R.H., 13th Int. Symp. Med. Chem., Paris, (1994) Abst. ML22. Scrip No. 2050 (1995) p. 7. Kusner E.J., Buckner C.K., Dea D.M., DeHaas C.J., Marks R.L. et al., Eur. J. Pharmacol. 257 (1994) 285292. Teixeira M.M., Hellewell P.G., Brit. J. Pharmacol. 111 (1994) 12051211. Crawley G.C., Briggs M.T., J. Med. Chem. 38 (1995) 39513956. Werz O., Szellas D., Henseler M., Steinhilber D., Mol. Pharmacol. 54 (1998) 445451. Thrien M., Fitzsimmons B.J., Sheigetz J., Macdonald D., Choo L.Y. et al., Bioorg. Med. Chem. Lett. 3 (1993) 20632066. [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] Ducharme Y., Brideau C., Dub D., Chan C.-C., Falgueyret J.-P. et al., J. Med. Chem. 37 (1994) 512518. Chauret N., Li C., Ducharme Y., Trimble L.A., Yergey J.A. et al., Drug Metab. Disp. 23 (1995) 6571. Delorme D., Ducharme Y., Brideau C., Chan C.C., Chauret N. et al., J. Med. Chem. 39 (1996) 39513970. Hamel P., Riendeau D., Brideau C., Chan C.C., Desmarais S. et al., J. Med. Chem. 40 (1997) 28662875. Dub D., Blouin M., Brideau C., Chan C.C., Desmarais S. et al., Bioorg. Med. Chem. Lett. 8 (1998) 12551260. Zhang K.E., Naue J.A., Arison B., Vyas K.P., Chem. Res. Toxicol. 9 (1996) 547554. Chauret N., Nicoll-Griffith D., Friesen R., Li C., Trimble L. et al., Drug Metab. Disp. 23 (1995) 13251334. Molon-Noblot S., Gillet J.P., Durand-Cavagna G., Huber A.C., Patrick D.H., Duprat P., Toxicol. Pathol. 24 (1996) 231237. Hutchinson J.H., Charleson S., Evans J.F., Falgueyret J.P. et al., J. Med. Chem. 38 (1995) 45384547. Hutchinson J.H., Prasit P., Choo L.Y., Riendeau D., Charleson S. et al., BioMed. Chem. Lett. 45 (1993) 978981. Hutchinson J.H., Riendeau D., Brideau C., Chan C., Delorme D. et al., J. Med. Chem. 36 (1993) 27712787. Hutchinson J.H., Riendeau D., Brideau C., Chan C., Falgueyret J.P. et al., J. Med. Chem. 37 (1994) 11531164. Gillard J., Ford-Hutchinson A.W., Chan C.C., Charleson S., Denis D. et al., Can. J. Physiol. Pharmacol. 67 (1989) 456464. Miller D.K., Gillard J.W., Vickers P.J., Sadowski S., Lveill C. et al., Nature 343 (1990) 278281. Dixon R.A.F., Diehl R.E., Opas E., Rands E., Vickers P.J. et al., Nature 343 (1990) 282284. Kargman S., Vickers P.J., Evans J.F., J. Cell. Biol. 119 (1992) 17011709. Kargman S., Rousseau P., Reid G.K., Rouzer C.A., Mancini J.A. et al., J. Lipid Mediators 7 (1993) 3145. Mancini J.A., Abramovitz M., Cox M.E., Wong E., Charleson S. et al., FEBS Lett. 318 (1993) 277281. Charleson S., Prasit P., Lger S., Gillard J.W., Vickers J.P. et al., Mol. Pharmacol. 41 (1992) 873879. Friedman B.S., Bel E.H., Buntinx A., Tanaka W., Han Y.-H.R. et al., Am. Rev. Respir. Dis. 147 (1993) 839844. Depre M., Friedman B., Tanaka W., VanHecken A., Buntinx A., DeSchepper P.J., Clin. Pharmacol. Ther. 53 (1993) 602607. Musser J.H., Chakraborty U.R., Sciortino S., Gordon R.J., Khandwala A. et al., J. Med. Chem. 30 (1987) 96104. Prasit P., Belley M., Evans J.F., Gauthier J.Y., Lveill C. et al., Bioorg. Med. Chem. Lett. 1 (1991) 645648. Young R.N., Gillard J.W., Hutchinson J.H., Lger S., Prasit P., J. Lipid Mediators 6 (1993) 233238. Prasit P., Belley M., Blouin M., Brideau C., Chan C., Charleson S. et al., J. Lipid Mediators 6 (1993) 239244. Mancini J.A., Prasit P., Coppolino M.G., Charleson S., Lger S. et al., Mol. Pharmacol. 41 (1992) 267272.

[47]

[48] [49]

[50]

[51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73]

[100] Brideau C., Chan C., Charleson S., Denis D., Evans J.F. et al., Can. J. Physiol. Pharmacol. 70 (1992) 799807. [101] Diamant Z., Timmers M.C., van der Veen H., Friedman B.S. et al., J. Allergy Clin. Immunol. 95 (1995) 4251. [102] Chapman K.R., Friedman B.S., Shingo S., Heyse J., Reiss T., Spector R., ALA/ATS Meeting, Boston, May 2125 (1994) A215. [103] Hillingso J., Kjeldsen J., Laursen L.S., Lauritsen K., von Spreckelsen S. et al., Clin. Pharmacol. Ther. 57 (1995) 335341.

685
[104] Roberts W.G., Simon T.J., Berlin R.G., Haggitt R.C., Snyder E.S. et al., Gastroenterology 112 (1997) 725732. [105] Hamilton A.L., Watson R.M., Wyile G., OByrne P.M., Thorax 52 (1997) 348354. [106] Fischer A.R., Rosenberg M.A., Roth M., Loper M., Jungerwirth S., Isreal E., Thorax 52 (1997) 10741077. [107] Muller-Peddinghaus R., J. Physiol. Pharmacol. 48 (1997) 529536. [108] Murthy S., Murthy N.S., Coppola D., Wood D.L., Inamm. Res. 46 (1997) 224233. [109] Negro J.M., Miralles J.C., Ortiz J.L., Funes E., Garcia A., Allergol. Immunopathol. 25 (1997) 209216. [110] OByrne P.M., Chest III (2 Suppl.) (1997) 27S34S. [111] Larson J.S., Jackson S.K., Am. J. Health Syst. Pharm. 53 (1996) 28212830. [112] Sorkness C.A., Pharmacotherapy 17 (1997) 50S54S. [113] Renzi P.M., Can. Med. Assoc. J. 160 (1999) 217223.

Eur. J. Med. Chem. 34 (1999) 687699 1999 Editions scientiques et mdicales Elsevier SAS. All rights reserved

687

Original article

The COREPA approach to lead generation: an application to ACE-inhibitors

Verginia Kamenska, Julian Ivanov, Ovanes Mekenyan*
Bourgas University As. Zlatarov, Department Of Physical Chemistry, 8010 Bourgas, Bulgaria (Received 17 August 1998; revised 19 January 1999; accepted 21 January 1999)

Abstract The recently derived algorithm for identifying the COmmon REactivity PAttern (COREPA) of structurally diverse chemicals having similar biological behaviour was employed to recognize the structural requirements for high ACE inhibition. COREPA is based on an assessment of all energetically-reasonable conformations of the compounds under study. It is not dependent upon a predetermined and specied pharmocophore or an alignment of active conformers. The approach describes the common reactivity pattern in terms of global and local stereoelectronic parameter ranges. These ranges are associated with compounds having extreme (highest and lowest) biological activity through a comparison of conformer distributions for specic descriptors. The dened parameter ranges were combined into Boolean expressions providing exible screening of chemicals by making use of a new chemical rule interpreter allowing a simultaneous search according to all available 2-D and 3-D information. The structural combinations, i.e., COREPAs, derived on a training set of ACE inhibitors were used for screening chemicals in an external test series for identifying those with high ACE inhibition activity. The obtained results showed that the structural requirements derived in the present work can be useful for screening chemicals from databases as potential lead ACE inhibitors. 1999 Editions scientiques et mdicales Elsevier SAS quantitative structure-activity relationships / active analogues / stereoelectronic structure / conformational exibility / ACE-inhibitors / lead generation / COREPA

1. Introduction Angiotensin-converting enzyme (ACE) is a Zncontaining metallopeptidase which catalyses the hydrolysis of the terminal dipeptide from the angiotensin I (decapeptide) to produce the octapeptide angiotensin II. The latter appears to be one of the most potent vasoconstrictors. Although the primary amino acid sequence of ACE is known, [13] its 3-D structure is still undetermined. Inhibitors of ACE are widely prescribed to control essential hypertension. The structural requirements for high ACE activity within congeneric series of chemicals have been derived from intensive SAR studies [46]. The latter, combined with crystallographic data from the analogous enzyme, thermolysin, and its inhibitors, dened the minimal set of the active groups necessary for a chemical to elcite ACE inhibition: a C-terminal carboxyl group for ionic binding to the enzyme; a carbonyl oxygen which hydrogen bonds to some active site residue; some Zn-binding functional groups such as a carboxylate, hydroxamate, phosphonate, or thiolate. This structural
*Correspondence and reprints

information has been used to screen databases of structurally diverse classes of ACE-inhibitors to elicite the common three-dimensional geometry of the pharmacophoric sites consistent with their activity [79]. Still, the recognition of the 3-D spatial alocation of the pharmacophore responsible for ACE inhibition potency is a subject of intensive QSAR studies employing a variety of pattern recognition techniques. A set of 28 inhibitors of angiotensin converting enzyme, selected by Mayer et al. [8], were modelled by the active analogue technique to elucidate the pharmacophore, i.e., the essential 3-D arrangement of functional groups that a molecule must possess to be recognized by the receptor under study. A strategy for systematic sampling of the conformational space has been used, thus, arriving at two plausible alternative active site hypotheses [10, 11], demonstrating that the multiple searches of the conformational space of exible ligands is an effective way to achieve increasing reliability of the modelling results. The CoMFA methodology (comparative molecular eld analysis) is one of the most applied 3D-QSAR

688 approaches to analyse ACE activity of chemicals. DePriest et al. [12] and Waller et al. [13] applied the CoMFA paradigm. They enlarged the initial set of 28 molecules with structures from additional classes of ACE inhibitors to resolve some uncertainties associated with recent hypotheses for the pharmacophoric site [8]. The CoMFA models of DePriest et al. [7], derived from potential elds, were found to be insufficient for accurately quantifying the enzyme-inhibitor interactions. The model and its predictive ability was improved by introducing a Zink indicator variable explicitly describing the Zink-ligand interaction. The derived 3D-QSAR models have been further used for predicting ACE activity of chemicals not included in the initial training set [12]. Pharmacophore search methods and receptor-site mapping, such as the active analogue technique and CoMFA, face signicant challenges, which include the selection of appropriate conformations and obtaining an alignment of these structures. There are a number of good techniques for superimposing molecules [1417], but developing a robust alignment model is not trivial. Typically, hundreds of alignments are explored to reach an optimal outcome, which, if not carefully evaluated and explained in the context of a presumed mechanism of interaction with the receptor may become susceptible to violation of the criteria of Topliss and Edwards [18] for causality in structure-activity models. Alignment errors can also lead to models that are incorrect or are poorly predictive. Further, the use of the lowest-energy conformers in methods such as CoMFA to assess similarity in pharmocophore search and receptor-mapping alogorithms seems inappropriate because, in complex systems such as biological tissues and uids, chemicals are likely to exist in a variety of conformational states. In fact the lowestenergy, gas-phase, conformations might be the least likely to interact with the solvent or macromolecules [19], and solvation and binding interactions could more than compensate for energy differences among the conformers of a chemical [2023]. In an attempt to address the issue of conformational exibility, Prendergast et al. [24] reported an augmented active analogue technique to identify specic conformers of ligands acting as antagonists to angiotensin-II. All geometrically reasonable conformers were assessed; however, conformational energies were not evaluated and an energy minimization was not performed during the search. In a sense, this methodology can be viewed as an augmented version of the active analogue approach because it accounts for conformational exibility and it eliminates the necessity of conformer alignment. Recently, we developed a new pattern recognition approach, named COREPA (COmmon REactive PAttern) employed as a pharamacophore mapping technique. COREPA is a generalization of the active analogue approach and it provides an implicit exploration of a receptors stereoelectronic shape. As opposed to existing pharmacophore search methods, the COREPA is an attempt to circumvent the problems related to the selection of appropriate conformations and obtaining appropriate template alignments [25, 26]. For each stereoelectronic parameter identied to be associated with the endpoint studied, all energetically-reasonable conformer distributions for the compounds from the training set are superimposed and the parameter ranges common for at least one conformer from all of the compounds identied. The collection of common stereoelectronic parameter ranges denes the common reactivity pattern. Due to the increasing pharmacological interest in potent vasoconstrictor agents, the present work aims to employ the COREPA method for the recognition of stereoelectronic requirements for a high ACE inhibition activity. 2. Materials and methods 2.1. The COREPA algorithm The methodology to elucidate chemical similarity is based on the assumption that chemicals that elicit similar biological behaviour through a common mechanism of action should also possess similarities in stereoelectronic descriptors. Elucidation of this common reactivity pattern within a set of biologically-similar chemicals requires examination of the conformational exibility of the compounds to evaluate molecular similarity in the context of the associated variability in specic stereoelectronic parameters. The principal steps of the algorithm are presented and discussed in detail in Mekenyan et al. [25]. However, in order to follow the forthcoming, they are summarized as follows: Step 1. Denition of the training set of chemicals. A dened subset of chemicals in the reaction series under investigation are selected as the training set. The training set can include either the most or least active chemicals, as dened by a user-imposed threshold of biological activity. This initial step establishes the extent of biological similarity among the chemicals from which stereoelectronic similarity will be discerned in the subsequent steps of the algorithm. Step 2. Evaluation of stereoelectronic parameters hypothesized to be associated with biologically similar compounds. A restricted set of parameters, hypothesized to be associated with biological activity, are evaluated based on the normalized sum of dynamic similarity

689

Figure 1. The training set of 28 AR ligands examined in this study (training set A).

690

Figure 2. The test set of 13 captopril derivatives (test set B).

indices [27] between each pair of molecules in the training set. The Tanimoto coefficient, TAN, and the related Hodgkin-Richards similarity metric, 3D-TAN, cosine index, COSINE, and an index based on information theory, IF, were used as similarity indices, whereas, the Euclidean metric, DIST, was employed as a dissimilarity measure [27]. The stereoelectronic parameters that provide the maximal measure of similarity among the chemicals in the training set are assumed to be most closely associated with the activity under consideration and are used in the subsequent step of the algorithm. Step 3. Recognition of the common reactivity pattern. For each stereoelectronic parameter identied in step 2, the conformer distributions of the chemicals from the training set are superimposed and the parameter ranges common for conformers from all of the chemicals identied. The collection of common stereoelectronic parameter ranges denes the common reactivity pattern. 2.2. ACE ligands and binding affnity A training set of 28 AR ligands examined in this study (training set A) are those included in the learning set of Mayer et al. [8]. The structures are depicted in gure 1. The pIC50 values (the negative log of the inhibition concentration providing a standard biological response)

are listed above the structures in gure 1. In the following analyses, sets of known ACE inhibitors such as captopril, pivalopril, rentiapril, zofenopril, enalapril, ramapril, quinapril, perindopril, cilazapril, delapril, lisinopril, SQ 29,852 and fosinopril [2830] were used as an external validation data set (test set B) to evaluate discrimination abilities of the derived reactivity pattern. Their structures and associated activities (pIC50) are presented in gure 2. 2.3. ACE ligand conformations A primary aspect of the COREPA approach is to evaluate the conformational space for the chemicals under study using a number of conformers that can reasonably be assumed to represent the diversity of relevant stereoelectronic character for the biological process of interest. Sampling of conformational space was performed, aiming to provide up to 50 structurally most distinct conformers for each of the structures included in the training set. The conformational search was performed by making use of the 3DGEN algorithm that initiates from molecular topology and generates all conformers consistent with steric constraints and expert rules [31]. Specic geometric constraints were imposed during structure generation. Thus, torsion resolution (TR) around saturated (SP3-SP3) acyclic bonds was chosen

691 to be 120, using an initial torsion angle of 60, with respect to the plane of the preceding three atoms. Up to 7 rotational variables were chosen for the acyclic part of each chemical. For all chemicals, 1.5 was set as the distance between non-bonded atoms, while 1.22.0 was the range imposed for ring closure. After generation of the initial set of conformers, up to 50 of the structurally most diverse conformers were screened for each chemical based on structural dissimilarity of conformers as assessed by the sum of the distances between their nonhydrogen atoms. Each of the generated conformers was submitted to force eld optimization based on a simple energy-like function, where only the electrostatic terms are omitted. Subsequently, the conformational degeneracy of the isomers was detected due to molecular symmetry and geometry convergence. Subsequent geometry optimization of the conformers was employed with MOPAC 7 [32], using the AM1 Hamiltonian with the key words PRECISE and NOMM. Finally, these conformers were screened to eliminate those structures with Hfo values that were 20 kcal/mol or higher than that calculated for the conformer associated with the absolute energy minimum (see Results and discussion for an explanation of this threshold). The COREPA analyses were performed on this resulting data set. 2.4. Molecular descriptors Stereoelectronic parameters were calculated with MOPAC 7 [32], augmented by a new computing module [33], that provides additional reactivity descriptors, using the AM1 all-valence electron semi-empirical Hamiltonian. The electronegativity (EN), dipol moment (), volume polarizability (Vol.P), energy of frontier orbitals (EHOMO and ELUMO), and electronic gap (EHOMO-LUMO) were used as global electronic descriptors, whereas the atomic charges (qi), frontier atomic charges (fiHOMO and fiLUMO), and acceptor superdelocalizability indices (SiE and SiN), as well as atomic self-polarizabilities (ii), were calculated as local electronic indices (i denotes a specic atom in a molecule). As it was already mentioned, when searching common patterns based upon local parameter distributions, the atomic reactivity indices were not restricted to specic rings in studied derivatives. The delocalizabilities were calculated by using the following equations: Si =
E occ j i a i a

where Ej are the molecular orbital (MO) energies; Cj are the corresponding eigenvectors, and a pertains to the atomic orbitals of site i; Eref is a non-zero, xed, energy reference level, that was set equal to the electronic midgap level for benzene (4.549 eV), i.e., Eref = 0.5(EHOMO + ELUMO)benzene. Superdelocalizability indices were used to assess the ability of a reactant to form bonds through charge transfer, i.e., donor superdelocalizability indices are measures of the ability of molecules to donate electron density through orbital transfers, while acceptor superdelocalizability indices are measures of the ability of molecules to accept electron density. Conformer structures were also assessed based on the steric descriptors GW (sum of geometric distances [34]), Lmax (the greatest interatomic distance), dij (steric distance between atoms i and j) and planarity (the normalized sum of torsion angles in a molecule [25]). Finally, Vol.Polar., dened as a sum of atomic selfpolarizabilities, and thus, the averaged ability of a compound to change electron density at its atoms during chemical interactions [35, 36], was selected as a physiochemical descriptor. Lower values of Vol. Polar. (Vol.P > 0) reect higher charge localizations and more polarizable (less lipophilic) molecules [36]. These descriptors were selected because hydrophobicity, steric bulk and size constraints have been reported as important criteria in predicting and interpreting ligand binding interactions. 2.5. Database screening approach The common reactivity pattern, i.e., the pharmacophore, dened in terms of 2-D structural fragments and associated 3-D stereochemical information (such as parity of stereocentres, distance between pharamacophoric sites, charge and delocalizability requirements) can be further used to screen databases for the presence of chemicals meeting these structural characteristics. A new chemical rule interpreter (CRI) has been developed for this purpose allowing simultaneous search according all available 2-D and 3-D information [37]. With that aim, substructure search techniques were combined with the range requirements for numeric descriptors as introduced in the extended SMILES language [37]. These entities, combining structural fragments with range requirement for numeric descriptors, are further called SD (structuredescriptor) screens. The CRI provides selective classication or screening of the chemicals from an OASIS database le (*.CMP) [33] according to a set of rules, given in terms of SD screens. A particular classication scheme is developed and described in a text le (*.RUL le). A separate rule of the scheme can be used to select from the *.CMP le those chemicals that satisfy it. The

Cja Cja / Eref Ej Cja Cja / Eref Ej

(1)

Si =
N

vac j

(2)

692 application of the whole classication scheme results in a case-specic assignment of descriptors to the chemicals of the database. The *.RUL le is comprised of three sections, namely, Denes, Rules and Apply sections. In the Denes section, the pertinent SD screens are described by means of extended SMILES language. It is very often practical to combine multiple SD screens in a single group that is used further as a single composite SD screen. For this purpose, the program supports a dynamic library of SD screens, partitioned in groups. Composite or simple SD screens have user dened labelsscreen identiers. Once dened, screen identiers can be further used as parts of SMILES strings in order to dene new SD screens. Screen identiers are SD screens on their own since they can serve for SD search or for denition of new screen identiers. When SD search is performed, each encountered screen identier is recursively replaced by its actual contents, and the result is positive if at least one match is found. Explicit SD screens and predened screen identiers constitute the basic logical variables of Boolean expressions (BL rules). The BL rules in use are described in the Rules section of the *.RUL le. An SD screen in the context of a BL rule takes logical values true or false if the SD screen is encountered in the current chemical or not. A BL rule combines a Boolean expression SD screen related with the logical operators and, or and not. The default priority of Boolean operations can be redened by the use of brackets in unlimited nesting. As in the case of SD screens, separate BL rules can be assigned to different rule identiers. An explicit BL rule or a rule identier can be selected, and applied to a *.CMP le to produce in an output *.CMP le the subset of chemicals that match the rule. In the Apply section, BL rules are employed in conditional if, then, else statements. The condition of a statement is a rule identier or a Boolean expression of rule identiers. Examples are given with some of the structural requirements for high ACE-activity included in Denes and Rules sections of *.RUL le (table IV). Comments are included to provide explanation of the screens.

Denes: RX: O, S, N, P Z1: O_O{5.6 < DISTANCE < 5.9} Z4: O{0.39 < Q < 0.25}_N{0.39 < Q < 0.25} {6.9 < DISTANCE < 7.4} Z5: RX{0.39 < Q < 0.25}_RX{0.39 < Q < 0.25} {8.7 < DISTANCE < 9.4} Z7: O{0.22 < DONOR_DLC < 0.30}_O{0.22 < DONOR_DLC < 0.30}{8.7 < DISTANCE < 9.4} Z9: RX{0.22 < DONOR_DLC < 0.30}_RX{0.22 < DONOR_DLC < 0.30}{8.7 < DISTANCE < 9.4} Rules: r2: Z7 and Z8 r4: Z3 or Z7 r5: (Z3 or Z4) and (Z7 or Z8) Wild-card atom RX holds for any of O, S, N and P atoms. Two O-atoms in a distance range of 5.65.9 []. O- and N-atoms having charges in the range of 0.39 to 0.25 [a.u.] to be in a distance range of 6.97.4 []. Two wild-card atoms RX (O, S, N and P) having charges in the range of 0.39 to 0.25[a.u.] to be in a distance range of 8.79.4 []. Two O-atoms having donor delocalizabilties in the range of 0.220.30 [(a.u.)2/eV] to be in a distance range of 8.79.4 []. Two wild-card atoms RX (O, S, N and P) having donor delocalizabilties in the range of 0.220.30[(a.u.)2/eV] to be in a distance range of 8.79.4 []. Simultaneous fulllment of the requirements Z7 and Z8. Fulllment of any one of requirements Z3 or Z7. Simultaneous fulllment of the structural combination in brackets. The rst of these combinations means fulllment of any one of the requirements Z3 or Z4 whereas the second means fulllment of any one of the requirements Z7 or Z8. Fulllment of any one of the requirements in brackets. The rst one means simultaneous fulllment of the requirements Z3 and Z4, whereas the second one means simultaneous fulllment of requirements Z7 and Z8. Fulllment of any one of the rst two requirements in brackets and at the same time non fulllment of the requirement dened in the third combination. The rst and third of the combinations means fulllment of any one of the requirements Z5 or Z6, and Z9 or Z10, respectively, whereas the third one means simultaneous fulllment of the requirements Z1 and Z2.

r6: (Z3 and Z4) or (Z7 and Z8) r15: (Z5 or Z6) or (Z9 or Z10) and not (Z1 and Z2)

693
Table I. ACE-inhibitors, observed binding affinities, number of conformer and parameter ranges for some signicant stereoelectronic parameters.
Structure # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 pIC50 obs. 9.638 9.222 9.000 8.959 8.921 8.921 8.854 8.796 8.585 8.553 8.523 8.523 8.495 8.432 8.398 8.222 8.155 8.046 8.000 7.921 7.700 7.638 7.420 7.398 7.301 7.000 7.155 6.000 Conformers 32 40 46 41 38 37 45 33 49 41 44 27 32 44 46 45 42 50 44 35 42 47 49 47 47 48 36 47 Vol.Polar [(a.u.)2/eV] 1.511 to 1.546 1.464 to 1.481 0.899 to 0.912 1.607 to 1.631 1.374 to 1.398 1.648 to 1.664 0.799 to 0.816 1.910 to 1.923 1.242 to 1.259 1.589 to 1.610 1.240 to 1.263 1.453 to 1.468 1.705 to 1.732 0.985 to 1.014 1.540 to 1.563 1.069 to 1.097 1.278 to 1.294 0.774 to 0.801 0.936 to 0.948 1.812 to 1.835 1.328 to 1.352 0.778 to 0.799 0.881 to 0.906 0.891 to 0.920 0.824 to 0.847 0.939 to 0.959 0.773 to 0.791 0.883 to 0.909 E (HOMO) [eV] 9.713 to 8.564 9.224 to 8.812 10.943 to 9.827 8.853 to 8.474 9.817 to 9.229 9.744 to 9.273 10.643 to 9.753 8.837 to 8.147 10.389 to 9.712 9.713 to 9.264 9.426 to 8.643 9.743 to 9.293 9.723 to 8.994 9.378 to 8.718 9.295 to 8.762 9.338 to 8.629 9.833 to 9.155 9.763 to 8.738 9.393 to 8.735 9.790 to 9.233 9.671 to 9.028 9.475 to 8.794 9.629 to 9.054 9.627 to 8.713 9.538 to 8.712 9.526 to 8.771 9.792 to 9.270 9.755 to 9.022 E (LUMO) [eV] 0.606 to 0.069 0.055 to 0.43 0.915 to 0.09 0.562 to 0.148 0.041 to 0.588 0.166 to 0.513 0.236 to 0.708 0.253 to 0.468 0.278 to 0.869 0.587 to 0.045 0.203 to 0.769 0.147 to 0.547 0.547 to 0.344 0.422 to 0.170 0.529 to 0.039 0.639 to 0.103 0.069 to 0.512 0.246 to 0.321 1.278 to 0.518 0.516 to 0.125 0.684 to 0.032 0.049 to 0.831 0.318 to 0.146 0.001 to 0.775 0.010 to 0.844 0.234 to 0.449 0.727 to 0.228 0.394 to 0.195 E(HOMO-LUMO) [D] [eV] 8.514 to 9.364 8.980 to 9.280 9.532 to 10.132 8.283 to 8.359 9.506 to 9.907 9.488 to 9.895 10.040 to 10.742 8.566 to 8.633 9.850 to 10.870 8.892 to 9.345 8.964 to 9.794 9.390 to 9.908 8.957 to 9.694 8.581 to 9.286 8.668 to 8.837 8.419 to 8.981 9.291 to 9.927 8.818 to 9.798 8.026 to 8.216 8.737 to 9.753 8.856 to 9.263 9.138 to 9.805 8.940 to 9.500 9.129 to 9.787 9.351 to 9.851 8.863 to 9.745 8.764 to 9.264 8.823 to 9.629 2.113 to 13.257 0.999 to 6.900 2.030 to 8.560 1.093 to 9.888 1.360 to 9.822 1.296 to 10.996 1.395 to 7.550 1.441 to 8.794 1.375 to 8.565 2.394 to 9.494 1.141 to 7.011 2.300 to 7.324 1.796 to 10.779 0.531 to 7.336 3.402 to 9.146 0.775 to 7.067 2.037 to 8.545 0.881 to 6.955 1.850 to 9.097 1.415 to 7.814 2.230 to 8.691 1.286 to 7.696 0.558 to 6.044 1.995 to 7.923 0.852 to 7.961 0.268 to 7.502 2.582 to 6.031 1.537 to 6.628 Hfo [kcal/mol] 213.709 to 194.315 190.112 to 170.246 162.691 to 144.868 177.810 to 158.097 201.710 to 181.716 215.523 to 195.697 359.037 to 339.913 501.299 to 481.409 245.194 to 225.583 167.064 to 151.341 184.961 to 164.971 206.092 to 188.253 152.300 to 132.998 92.215 to 73.736 247.923 to 229.786 128.738 to 109.576 271.344 to 252.291 164.372 to 148.330 237.338 to 219.398 306.403 to 287.903 140.147 to 120.234 135.315 to 118.169 125.260 to 106.604 139.030 to 119.150 113.348 to 95.449 112.093 to 97.395 59.732 to 47.805 113.843 to 93.921

3. Results and discussions 3.1. Conformational exibility and electronic structure Recently, the 20 kcal/mol threshold for Hfo was assumed to result in an energetically reasonable set of conformations, given the extent to which energy provided during ligand binding could facilitate conformational transformations [22, 23]. In this respect, the range of Hfo values for the conformers of the chemicals under study was selected to be less than 20 kcal/mol (table I). For a given compound, conformers within the specied range of Hfo, exhibited chemically-signicant variation in potentially relevant electronic descriptors, as summarized in table I. For example, the following parameter ranges are produced by the conformers of compound 7: 0.944 eV for ELUMO (from 0.2360.708), 0.89 eV for EHOMO, 0.702 eV for EHOMO-LUMO, 0.017 (a.u.)2/eV for Vol.P and 6.155 D for . Signicant parameter variations were also observed for the other compounds. The observation that relatively small energy differences between conformers can be associated with signicant variations in electronic structure (i.e., molecu-

lar electronic descriptors) highlights the necessity of including all energetically-reasonable conformers when dening common reactivity patterns. Within the employed approximation, these conformers were considered as equally probable because it is difficult to relate their gas-phase energies with the conformer preference in complex biological environments. 3.2. Application of the COREPA algorithm According to the rst step of the COREPA algorithm, two learning subsets were selected out of all 28 ACE inhibitors under investigation (training set A). The rst one includes the most active chemicals having pIC50 8.9, whereas the second consists of the least potent (passive) ligands having pIC50 7.42. Further, the dynamic similarity was calculated for each pair of chemicals belonging to the subsets of active and non-active ligands, by using molecular descriptors presented in the preceding section. The normalized (over the pairs of compared chemicals) similarity indices, associated with molecular descriptors are presented in table II.

694
Table II. Normalized similarity measurements between conformer distributions of most active and least active ACE-inhibitors. Descriptors TAN ACCEPT DLC ACCEPT MLK BOND ORDER DONOR DLC DONOR MLK POLAR POLAR MLK POPHOMO POPHOMO MLK POPLUMO POPLUMO MLK Q Q MLK SPECIAL DISTANCE E(HOMO) ELECTRONEGATIVITY E(HOMOLUMO) GEOM. WIENER VOLUME POLARIZAB. CALC. HEAT FORM. DIPOLE MOMENT E(LUMO) 0.285 0.272 0.291 0.277 0.279 0.289 0.292 0.283 0.278 0.284 0.283 0.286 0.287 0.261 0.061 0.149 0.096 0.012 0.000 0.051 0.257 0.104 Most active ACE-inhibitors DIST 0.472 0.473 0.522 0.472 0.472 0.472 0.472 0.472 0.529 0.514 0.492 0.478 0.487 0.475 0.969 0.958 0.663 0.750 0.510 0.719 0.472 1.044 IF 3.297 3.282 3.319 3.289 3.292 3.302 3.305 3.293 3.292 3.296 3.297 3.296 3.297 4.471 0.505 1.225 0.673 0.244 0.000 0.471 1.665 0.802 3D-TAN COSINE 0.878 0.865 0.817 0.655 0.884 0.333 0.600 0.700 0.924 0.579 0.652 0.750 0.207 0.125 0.154 0.000 0.000 0.000 0.000 0.070 0.056 0.077 0.935 0.928 0.931 0.792 0.958 0.756 0.866 0.841 0.961 0.937 0.894 0.949 0.455 0.223 0.272 0.000 0.000 0.000 0.000 0.134 0.113 0.147 TAN 0.269 0.279 0.289 0.277 0.279 0.289 0.292 0.283 0.278 0.284 0.283 0.286 0.287 0.261 0.061 0.149 0.096 0.012 0.000 0.051 0.257 0.104 Least active ACE-inhibitors DIST 0.473 0.472 0.517 0.472 0.472 0.472 0.472 0.472 0.529 0.514 0.492 0.478 0.487 0.475 0.969 0.958 0.663 0.750 0.510 0.719 0.472 1.044 IF 3.144 3.155 3.179 3.289 3.292 3.302 3.305 3.293 3.292 3.296 3.297 3.296 3.297 4.471 0.505 1.225 0.673 0.244 0.000 0.471 1.665 0.802 3D-TAN COSINE 0.535 0.408 0.747 0.655 0.884 0.333 0.600 0.700 0.924 0.579 0.652 0.750 0.207 0.125 0.154 0.000 0.000 0.000 0.000 0.070 0.056 0.077 0.716 0.649 0.856 0.792 0.958 0.756 0.866 0.841 0.961 0.937 0.894 0.949 0.455 0.223 0.272 0.000 0.000 0.000 0.000 0.134 0.113 0.147

The data summarised in table II suggest that the most active and inactive ACE-ligands are highly similar according to charges (qi), steric distances (dij), donor delocalizabilities (SiE, SiN) frontier charges on HOMO (fiHOMO) and LUMO (fiLUMO) and atom polarizabilities (i). In this respect, the next search of the common reactivity pattern was based on conformer distributions of chemicals across each of those local molecular descriptors. As such, we have also studied the interatomic distances dij between non-hydrogen atoms: N-X, O-X and X-X, where X stands for any nonhydrogen atom in the molecules. To establish common reactivity patterns (step 3), conformer frequency distributions of compounds from each of the two training sets were subsequently examined across all local stereoelectronic descriptors, with results based on qi and dij, illustrated in gure 3 and gure 4, respectively. To study the effect of parameter distribution partitioning (i.e., size of parametric windows) on the obtained reactivity pattern, the number of parameter partitions was set at 20 and 40 for each descriptor. The intersections between conformer distributions were identied separately for chemicals belonging to active and non-active subsets. These intersections are described in terms of parameter ranges occupied by at least one conformer from each chemical belonging to the

learning subsets. The collection of those ranges represents the common reactivity pattern necessary for eliciting similar (high or low) biological effect (ACEinhibition). Stereoelectronic parameters producing the most distinct reactivity patterns for active and passive chemicals and the respective ranges are presented in table III. Even though conformer distributions from the most active and least active sets of compounds overlap, the subset of common partitions that contain conformers from each of the active compounds does not overlap with partitions containing conformers from each of the least active compounds. Thus, the common atomic charge pattern associated with the most active ligands (gure 3a) deviates signicantly from that of the least active chemicals (gure 3b). For the most active compounds, common partitions corresponding to oxygen and nitrogen atoms having charges from 0.25 to 0.39 [a.u.] (gure 3a) and donor delocalizabilities from 0.220.30 [(a.u.)2/eV] were observed. In difference, the charge and donor delocalizability patterns corresponding to O and N atoms for non-active inhibitors were shifted to more positive charge values (0.26 to 0.36 [a.u.]; gure 3b) and lower delocalizabilities. Signicant differences have also been found in charge and donor delocalizability patterns, due to the presence of S-atoms in non-active chemicals

695

Figure 3. Common reactivity patterns for active a. and nonactive b. ACE-inhibitors across the charges of hetero-atoms (oxygen, nitrogen and sulfur). Grey bars correspond to parameter ranges populated by at least conformers of all chemicals (activities, in gure 3a, and nonactives in gure 3b), whereas the black bars correspond to parameter ranges populated by conformers of some of the chemicals from active and nonactive subsets.

(gure 3b). The maximum common distance range between oxygen atoms for active ligands was found to be from 8.79.4 (light coloured bars in gure 4a), whereas between O and N atoms was from 6.97.4 (not illustrated). As shown in gure 4b, no such common ranges have been established between the electronegative sites in non-active chemicals. As can be seen in table III, less signicant differences in reactivity patterns of active and non-active chemicals (i.e., small differences in parametric ranges for active and non-active ACE inhibitors) were established for donor delocalizabilities of carbonyl carbons and the remaining C-skeleton.

Figure 4. Common reactivity patterns for active a. and nonactive b. ACE-inhibitors across the distance between oxygen atoms. Grey bars correspond to parameter ranges populated by at least conformers of all chemicals (actives in gure 4a and non-actives in gure 4b), the black bars correspond to parameter ranges populated by conformers of some of the chemicals from active and non-active subsets. The light grey bar in gure 4a are associated with the largest common distance range for active chemicals in the training set.

696
Table III. Common reactivity patterns for most active and least active ACE-inhibitors from training set A.
Most active ACEinhibitors pIC50 8.9 Descriptor interatomic distances, [] Atoms/ fragm. O-O O-N Charge [a.u.] Donor Delocalizability [(a.u.)2/eV] X All Descriptor ranges Partitioning = 40 8.94 to 9.17 6.96 to 7.19 0.39 to 0.26 (O,N); 0.111 to 0.126 (C=(O)); 0.149 to 0.212 (C); 0.227 to 0.290 (O,N) Partitioning = 20 8.74 to 9.39 6.95 to 7.42 0.39 to 0.25 (O,N); 0.111 to 0.126 (C=(O)); 0.142 to 0.22 (C); 0.22 to 0.30 (O,N) Least active ACE-inhibitors pIC50 7.42 Descriptor ranges Partitioning = 40 5.75 to 5.88 3.71 to 3.80 0.36 to 0.26 (O,N); 0.013 to 0.010 (S); 0.113 to 0.121 (C=(O)); 0.146 to 0.163 (C(N)); 0.171 to 0.188 (C(C)); 0.230 to 0.255 (N,OH); 0.263 to 0.280 (O=(C)); 0.381 to 0.410 (S) Partitioning = 20 5.63 to 5.88 3.71 to 3.89 0.36 to 0.26 (O,N); 0.024 to 0.021 (S); 0.113 to 0.129 (C=(O)); 0.146 to 0.163 (C-(N)); 0.163 to 0.196 (C-(C)); 0.230 to 0.247 (N,OH); 0.247 to 0.280 (O=(C)); 0.381 to 0.410 (S)

X - for non hydrogens and non-carbon atoms; ALL - the partitions of the local parameter were analysed for all atoms in the molecule.

3.3. Validation of reactivity pattern The complete lists of studied structural requirements and the Boolean expressions based on the above dened common parameter ranges are presented in table IV. More or less conservative screens were imposed by using exibility of the Boolean expressions (i.e., the logical combination of the different structural rules by making use of the logical operators and, or and not).

The validation of the method was based on the applications of stereoelectronic screens on chemicals of the internal training set (A) (subset of which was used for deriving reactivity patterns) as well as to an external dataset (B) with well known ACE-inhibitors, used as pharmaceutical agents. The results are presented in table V. The quality of screens is assessed by their ability to: (i) identify conformers of the chemicals from the learning

Table IV. The structural requirements and their Boolean combinations used for selecting highly active ACE inhibitors from test series under investigation.
Expressions of the dened structural requirement used for selecting high ACE inhibition activity from test series under investigation. RX:O, S, N, P Z1:O_O{5.6 < DISTANCE < 5.9} Z2:O_N{3.7 < DISTANCE < 3.9} Z3:O{0.39 < Q < 0.25}_O{0.39 < Q < 0.25}{8.7 < DISTANCE < 9.4} Z4:O{0.39 < Q < 0.25}_N{0.39 < Q < 0.25}{6.9 < DISTANCE < 7.4} Z5:RX{0.39 < Q < 0.25}_RX{0.39 < Q < 0.25}{8.7 < DISTANCE < 9.4} Z6:RX{0.39 < Q < 0.25}_RX{0.39 < Q < 0.25}{6.9 < DISTANCE < 7.4} Z7:O{0.22 < DONOR_DLC < 0.30}_O{0.22 < DONOR_DLC < 0.30}{8.7 < DISTANCE<9.4} Z8:O{0.22 < DONOR_DLC < 0.30}_N{0.22 < DONOR_DLC < 0.30}{6.9 < DISTANCE < 7.4} Z9:RX{0.22 < DONOR_DLC < 0.30}_RX{0.22 < DONOR_DLC < 0.30}{8.7 < DISTANCE < 9.4} Z10:RX{0.22 < DONOR_DLC < 0.30}_RX{0.22 < DONOR_DLC < 0.30}{6.9 < DISTANCE < 7.4} Boolean expression of the structural requirements r1: Z3 and Z4 r2: Z7 and Z8 r3: Z3 or Z4 r4: Z3 or Z7 r5: (Z3 or Z4) and (Z7 or Z8) r6: (Z3 and Z4) or (Z7 and Z8) r7: (Z3 or Z4) or (Z7 or Z8) r8: (Z5 or Z9) r9: (Z5 and Z9) r10: (Z5 or Z6 ) r11: (Z5 and Z6 ) r12: (Z5 or Z6) or (Z9 or Z10) r13: (Z5 and Z6) or (Z9 and Z10) r14: (Z5 or Z6) and (Z9 or Z10) r15: (Z5 or Z6) or (Z9 or Z10) and not (Z1 and Z2) r16: (Z5 or Z6) or (Z9 or Z10) and not (Z1 or Z2)

697
Table V. Screened ACE-inhibitors from test series according to the employed structural requirements, as described by the Boolean expressions listed in table IV. Rules #a Number of selected chemicals 10 12 14 16 14 12 16 16 14 19 14 22 15 19 22 22 Training set A Number of selected conformers 47 75 222 186 216 65 250 222 188 428 127 495 156 416 484 350 Numbering of selected chemicalsb 1, 2, 4, 5, 6, 8, 10, 13, 20, 21 1, 2, 48, 10, 12, 1921 16, 810, 12, 13, 1921 110, 12, 13, 17, 1921 16, 810, 12, 13, 1921 1, 2, 4, 5, 7, 8, 10, 13, 1921 110, 12, 13, 17, 1921 111, 13, 17, 1921 16, 811, 13, 1921 16, 813, 16, 1921, 23, 27, 28 16, 811, 13, 1921 110, 12, 13, 1517, 1921, 23, 27, 28 111, 13, 1921 16, 811, 13, 16, 1921, 23, 27, 28 113, 1517, 1921, 23, 27, 28 113, 1517, 1921, 23, 27, 28 Number of selected chemicals 6 1 9 9 2 7 9 9 2 11 9 11 9 2 10 11 Test set B Number of selected conformers 42 9 350 248 60 51 374 317 35 596 134 624 162 79 619 265 Numbering of selected chemicalsc 57, 911 12 47, 913 47, 913 12, 13 57, 912 47, 913 47, 913 12, 13 313 47, 913 313 47, 913 12, 13 312 313

r1 r2 r3 r4 r5 r6 r7 r8 r9 r10 r11 r12 r13 r14 r15 r16

b c

The numbering of chemicals corresponds to those in gure 1 The numbering of chemicals corresponds to those in gure 2

subset; (ii) catch some chemicals with moderate activity (not used for deriving the pattern); (iii) eliminate all conformers of the chemicals with low activity; and (iv) identify as active, chemicals from external databases, experimentally shown to elicit high activity (i.e., pharmaceutical agents). Beside selection of conformers of most active chemicals from training set A (having pIC50 8.9) used for deriving the common reactivity pattern, almost all of the structural rules screened also conformers of ACE ligands with lower activity, i.e., having pIC50 < 8.9. Still, these are active and medium active inhibitors having 7.7 pIC50 8.9 (up to chemical 21, table I). Only few of the structural rules, classied as active, the inhibitors having pIC50 < 7.7. These are the least restrictive rules #10, 12, 1416 with high concentration of or logical operator, i.e., providing very low screen conservativeness. These rules, screened as active, chemicals having pIC50 6.00. It is very difficult to dene precisely the pIC50 threshold according to which a chemical could be classied as active. Usually, these thresholds are user dened. All chemicals included in training set A are ACE inhibitors and the differences in their activity are only quantitative, not qualitative. In this respect, the obtained results are reasonable: the more

restrictive screens, selected as active, the inhibitors with higher pIC50 values, whereas, the less restrictive screens select as active all inhibitors under investigation. The reactivity pattern derived by the most active ACE inhibitors from training set A was used for screening of the chemicals from the test set B. As it was already mentioned, these captopril derivatives are well known ACE inhibitors with different activity, most of them used as pharmaceutical agents. It was hypothesized that the derived reactivity pattern should, screen as active, most of the chemicals from this test set. As can be seen from table V, Rules # 3, 4, 7, 8, 11 and 13 selected 70% of the inhibitors from test set B, whereas Rules # 10, 12 and 16 selected 85% of those chemicals. The fact that the last four rules were capable of screening a large percentage of the known ACE-inhibitors from the external dataset is an indication that they could be used for screening of chemicals from databases as potential lead ACE inhibitors. A practical advise is to perform screenings with each of the four individual rule les and then to look for the intersection of predicted subsets of chemicals. The latter are assumed to be the most potent ligands because they meet simultaneously stereoelectronic requirements included in all rule les.

698 4. Summary and conclusions The present search for the structural requirements for high ACE-inhibition is based on the COREPA algorithm. This last one is a generalization of the active analogue approach circumventing the need to align conformers of active molecules, while it explicitly addresses variation in conformational exibility in the context of varying biological activity. The distributions of all energetically reasonable conformers of chemicals from the training subsets across specic molecular descriptors found to be relevant to activity under study are analysed. Thus, instead of template alignment based on single conformer representations of chemicals, their conformer distributions are naturally aligned (ordered) across specic molecular descriptors. The distribution intersections populated by conformers of each of the biologically similar chemicals (actives and non-actives) form the common reactivity pattern. Due to the population character of conformer distributions in the COREPA approach it is difficult to recognize the active conformers of chemicals in classical chemical terms as it is done in other dynamic techniques [2022, 26, 38, 39]. The common reactivity pattern, in terms of charge and distance ranges between electronegative atoms required for high ACE-inhibition, was derived. It was based on the subset of the most active chemicals from the training set A. Then, it was validated by screening all chemicals from the same training set, as well as an external set of active analogues known as pharmaceutical agents (training set B). The various logical combinations of the stereoelectronic rules provided the exibility of the screening process. The most restrictive screens selected as active the inhibitors with pIC50 values higher or equal to the activity threshold used for deriving reactivity patterns. The less restrictive screens in addition, selected as active, inhibitors with moderate activity. Six of the employed screening rules selected as active 70% of the active inhibitors from test set B, whereas four of the rules selected 85% of those chemicals. The large percentage of the screened ACE-inhibitors from the test set by these four stereoelectronic screens is indicative that the associated reactivity pattern could be used for screening of large databases of 3-D structures for the search of potentially active ACE-inhibitors. Individual screenings were performed by employing each of the stereoelectronic screens on conformers of the chemicals from the test set. Chemicals meeting multiple stereoelectronic requirements, i.e., lying on the intersection of the subsets predicted by single screens, are considered as the most likely candidates for active ligands. Such studies are presently performed by our laboratory for the needs of the Chemicals and Pharmaceutical Research Institute, Soa, aiming to design original pharmaceutical agents. In large databases, however, chemicals are represented by single conformers because the conformational search for each compound is computationally impractical. In these situations, a less restrictive screening strategy is employed initially that assesses a single conformer per chemical and which is designed to minimize the percentage of false negatives (i.e., compounds incorrectly predicted to be non-active). The Tweak technique is also used as a pre-screen alternative, in which the rotatable bonds of the structures are adjusted to produce a conformation which matches as closely as possible a given 3D requirement. Then, in a second stage, more rened screens should be employed on a smaller set of conformationally multiplied chemicals, already passed the pre-screen. Acknowledgements This research was nancially supported by the Chemicals and Pharmaceutical Research Institute, Soa (grant # RD-09-173/20.12.1995). The authors also thanks Dr Karabunarliev for developing the Chemical Rule Interpreter software as well as Drs Matey Vitev, Neno Dimov and Kiril Ninov for valuable discussions. References
[1] [2] [3] Bunning P., Kleemann S.G., Riordan J.F., Biochemistry 29 (1990) 1048810492. Chen Y.N.P., Riodan J.F., Biochemistry 29 (1990) 1049310498. Soubrier F., Alhenc-Gelas F., Hubert C., Allegrini J., John M., Tregear G., Corvol P., Proc. Natl. Acad. Sci. USA 85 (1988) 93869390. Petrillo W.W., Trippodo N.C., De Forrest J.M., in: Robertson D.W. (Ed.), Annual Reports in Medicinal Chemistry Vol 25, Academic Press, New York, 1989, pp. 5160. Hangauer D.G., in: Perun T.J., Propst C.L. (Eds.), Computer-Aided Drug Design: Methods and Applications, Marcel Dekker, New York, 1989, pp. 253295. Wyvratt M.J., Patchett A.A., Med. Res. Rev. 5 (1985) 483531. Depriest S.A., Shands R.F.B., Dammkeohler R.A., Marshall G.R., in: Silipo C., Vittoria A. (Eds.), QSAR Rational Approaches to the Design of Bioactive Compounds Vol 16, Elsevier Science, Amsterdam, 1991, pp. 405414. Mayer D., Naylor C.B., Motoc I., Marshall J.R., J. Comput. -Aided Mol. Des 1 (1987) 316. Andrews P.R., Carson J.M., Caselli A., Spark M.J., Woods R., J. Med. Chem. 28 (1985) 393399. Dammkoehler R.A., Karasek S.F., Shands E.B.F., Marshall G.R., J. Comput. -Aided Mol. Des. 3 (1989) 3. Dammkoehler R.A., Karasek S.F., Shands E.B.F., Marshall G.R., J. Comput. -Aided Mol. Des. 9 (1995) 491499.

[4]

[5]

[6] [7]

[8] [9] [10] [11]

699
[12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] Depriest S.A., Mayer D., Naylor C.B., Marshall G.R., J. Am. Chem. Soc. 13 (1993) 115 53725384. Waller C.L., Shands E.B.F., Dammkoehler R.A., Marshall G.R., J. Comput. -Aided Mol. Des. 11 (1997) 121. Kearsely S.K., Smith G.M., Tetrahedron Comput. Methodol. 3 (1990) 615633. Blaney J.M., Dixon J.S., Perspect. Drug Discov. Des. 1 (1993) 301319. Perkins T.D., Deam P.M., J. Comput. -Aided Mol. Des. 7 (1993) 155172. Martain Y.C., Bures M.G., Danahar E.A., Delazzar J., Lico I., Pavlik P.A., J. Comput. -Aided Mol. Des. 7 (1993) 83102. Topliss J.G., Edwards R.P., J. Med. Chem. 22 (1979) 12381244. Eliel E.L., in: Warr W.A. (Eds.), Chemistry in Three Dimensions in Chemical Structures, Vol. 1, Springer, Berlin, 1993, pp. 18. Mekenyan O.G., Ivanov J.M., Veith G.D., Bradbury S.P., Quant. Struct. -Act. Relat. 13 (1994) 302307. Mekenyan O.G., Veith G.D., Call D.J., Ankley G.T., Environ. Health Perspect. 104 (1996) 13021309. Bradbury S.P., Mekenyan O.G., Ankley G.T., Environ. Toxicol. Chem. 15 (1996) 19451954. Wiese T., Brooks S.C., J. Steroid. Biochem. Molec. Biol. 50 (1994) 6172. Prendergast K., Adams K., Greenlee W.J., Nachbar R.B., Patchett A.A., Underwood D.J., J. Comput. -Aided Mol. Des. 8 (1994) 491512. Mekenyan O.G., Ivanov J.M., Karabunarliev S.H., Bradbury S.P., Ankley G.T., Karcher W., Environ. Sci. Technol. 31 (1997) 37023711. [26] [27] Bradbury S.H., Mekenyan O.G., Ankley G.T., Environ. Toxicol. Chem. 17 (1998) 1525. Mekenyan O.G., Ivanov J.M., Karabunarliev S.H., Hansen B., Ankley G.T., Bradbury S.P., in: Chen F., Schrmann G. (Eds.), Quantitative Structure Activity Relationships in Environmental Sciences VII, SETAC Press, Pensacola FL, USA, 1998, pp. 39. Cambell W.B., Brooks S.N., Pettinger W.A., Science 184 (1974) 994996. Ondetti M.A., Rubin B., Cushman D.W., Science 196 (1977) 441444. Natoff I.L., Redshaw S., Drugs Future 12 (1987) 475483. Ivanov J.M., Karabunarliev S.H., Mekenyan O.G., J. Chem. Inf. Comput. Sci. 34 (1994) 234243. Stewart J.J.P., MOPAC: A general molecular orbital packages; Version 7. 0. Software. Quantum Chemistry Program Exchange no. 455, University of Indiana, Bloomington, IN, USA, (1995). Mekenyan O.G., Karabunarliev S.H., Ivanov J.M., Dimitrov D.N., Comput. Chem. 18 (1994) 173187. Mekenyan O.G., Peitchev D., Bonchev D., Trinajstic N., Bangov I., Arzneim. Forsch. 36 (1986) 176. Schrmann G., Quant. Struct. -Act. Relat. 59 (1990) 326333. Lewis D.F.V., J. Comput. Chem. 10 (1989) 145151. Karabunarliev S.H., Nikolov N., Mekenyan O.G., submitted. Mekenyan O.G., Schultz T.W., Veith G.D., Kamenska V.B., J. Appl. Toxicol. 16 (1996) 355. Veith G.D., Mekenyan O.G., Ankley G.T., Call D.J., Environ. Sci. Technol. 29 (1995) 126.

[28] [29] [30] [31] [32]

[33] [34] [35] [36] [37] [38] [39]

[25]

Eur. J. Med. Chem. 34 (1999) 701709 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

701

Original article

Synthesis of some new 1,4-benzothiazine and 1,5-benzothiazepine tricyclic derivatives with structural analogy with TIBO and their screening for anti-HIV activity#
Giuliano Grandolini*, Luana Perioli, Valeria Ambrogi
Istituto di Chimica e Tecnologia del Farmaco, Universit di Perugia, Via del Liceo 1, 06123 Perugia, Italy (Received 10 November 1998; accepted 27 January 1999)

Abstract Several new tricyclic derivatives with structural analogy to TIBO were prepared starting from properly substituted 1,4benzothiazines and 1,5-benzothiazepine. All synthesized compounds were submitted to screenings for in vitro anti-HIV-1 activity. Only two compounds showed moderate activity. 1999 ditions scientiques et mdicales Elsevier SAS 1,4-benzothiazine and 1,5-benzothiazepine derivatives / anti-HIV-1 activity / non-nucleoside reverse transcriptase inhibitor (NNRTI)

1. Introduction The dramatic increase of spread of the acquired immunodeciency syndrome (AIDS) has stimulated considerable efforts in the research in this eld. As a result of the difficulties encountered in the development of an effective vaccine, research is aimed at the discovery of new chemotherapeutic agents. Good results were obtained with the class of nonnucleoside reverse transcriptase inhibitors (NNRTIs) for both their antiviral activity and their low toxicity. These considerations prompted us to prepare some 1,4-benzothiazine and 1,5-benzothiazepine derivatives with structural analogy with TIBO [15], a nonnucleoside inhibitor of HIV-1 reverse transcriptase (gure 1). In this paper we describe the synthesis and the preliminary anti-HIV screening of some tricyclic derivatives (gure 1).

Figure 1. General structure of new tricyclic 1,4-benzothiazine and 1,5-benzothiazepine derivatives with structural analogy with TIBO.

2. Chemistry Starting materials for our work program were compounds 2 (af) in which the amino group is properly located to react with bifunctional reagents to give the desired tricyclic compounds. The 5- or 6-amino derivatives 2 (table I) were obtained by reduction of 7-methoxy-5-nitro-3,4-dihydro-2H-1,4-benzothiazin3-ones or 8-methoxy-6-nitro-2,3,4,5-tetrahydro-1,5benzothiazepin-4-one 1 already synthesized by us [6] (gure 2).

A preliminary account of this work was presented at the 4th International Conference on Heteroatom Chemistry, Seoul, Korea, July 30August 4, 1995 *Correspondence and reprints

702
Table I. Physical and chemical data of compounds 2af.

Compound R 2a H

n 0

Yield (%) 78

M.p. (C) 258260

Formula (MW) C9H10N2O2S (210.25) C10H12N2O2S (224.28) C11H14N2O2S (238.30)

H-NMR,

3.35 (s, 2H, SCH2), 3.62 (s, 3H, OCH3), 5.25 (s, 2H, NH2), 6.106.20 (m, 2H, Ar), 9.50 (s, 1H, NH) (DMSOd6) 1.30 (d, J = 9.5 Hz, 3H, SCHCH3), 3.50 (q, 1H, SCHCH3), 3.65 (s, 3H, OCH3), 5.25 (s, 2H, NH2), 6.106.20 (m, 2H, Ar), 9.48 (s, 1H, NH) (DMSO-d6) 1.05 (t, 3H, CH2CH3), 1.502.10 (m, 2H, CH2CH3), 3.25 (q, 1H, SCH), 3.75 (s, 3H, OCH3), 4.05 (brs, 2H, NH2), 6.15 (d, J = 2.4 Hz, 1H, Ar), 6.35 (d, J = 2.4 Hz, 1H, Ar), 9.40 (brs, 1H, NH) (CDCl3) 0.90 (t, 3H, CH3), 1.221.70 (m, 4H, CH2CH2CH2CH3), 1.852.03 (m, 2H, CH2CH2CH2CH3), 3.35 (q, 1H, SCH), 3.73 (s, 3H, OCH3), 3.88 (br s, 2H, NH2), 6.20 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.35 (d, Jmeta = 2.4 Hz, Ar), 8.75 (br s, 1H, NH) (CDCl3) 3.61 (s, 3H, OCH3), 4.82 (s, 1H, SCHC6H5), 5.38 (s, 2H, NH2), 6.10 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.20 (d, Jmeta = 2.40 Hz, 1H, Ar), 7.27 (s, 5H, Ar), 9.83 (s, 1H, NH) (DMSO-d6) 2.33, 3.33 (A2B2 system, 4H, CH2CH2), 3.65 (s, 3H, OCH3), 5.12 (s, 2H, NH2), 6.306.40 (m, 2H, Ar), 8.67 (s, 1H, NH) (DMSO-d6)

2b

CH3

89

178179

2c

CH2CH3

70

153154

2d

(CH2)3CH3

51

104106

C13H18N2O2S (266.36)

2e

C6H5

62

220221

C15H14N2O2S (286.35)

2f

51

198199

C10H12N2O2S (224.28)

Unfortunately, because of the problems encountered during the synthesis of 1 [3], to date only compounds with a methoxy substituent at the 7 or 8 position could be prepared. Our rst attempt to synthesize the imidazo derivatives 4 (af) was performed by heating 2 (af) with an excess of formic acid under reux [7]. In this case the N-formylderivatives 3 (ac, e and f) were obtained (table II). Tentative cyclizations were performed by treating 3 with concentrated HCl but only for 3a (R = H, n = 0) the desired tricyclic 4a was obtained. Treatment of 3b (R = CH3, n = 0) with concentrated HCl gave rise to the replacement of the N-formyl group with a chlorine atom with the formation of the 5-chloro-7-methoxy-2-methyl-3,4-dihydro-2H-1,4benzothiazin-3-one 5. Another tentative synthesis of 4 (af) was achieved by reuxing compound 2 (af) with triethyl orthoformate in

xylene (mixture of isomers) [7], in this way only compounds 4e and 4f were obtained (table III). Condensation of compounds 2 with triethyl orthoacetate in xylene was successful only for compounds 6a, b and e (table III). 5-Amino-2-ethyl-7-methoxy-3,4dihydro-2H-1,4-benzothiazin-3-one 2c and 5-amino-2butyl-7-methoxy-3,4-dihydro-2H-1,4-benzothiazin-3-one 2d afforded only a variety of decomposition products which have not yet been identied. When the same reaction was performed with the benzothiazepine 2f, the 2-ethoxy-9-methoxy-2-methyl-1H-5,6-dihydro-imidazo[3,4,5-e,f]-1,5-benzothiazepin-4-one 7 (gure 3) was obtained. Operating in the presence of pyridinium hydrochloride, compound 8 was formed (gure 3). Compounds 9ac and e (table IV, gure 2) were obtained by reuxing the corresponding amines 2ac and e with carbon disulde in anhydrous pyridine. Reaction of carbon disulde with 2d and 2f either in anhydrous

703

Figure 3. Tentative synthesis of 2-methyl-9-methoxy-4H-5,6dihydroimidazo[3,4,5-e,f]-1,5-benzothiazepin-4-one. Figure 2. Synthetic pathway for new tricyclic derivatives.

pyridine or in anhydrous xylene afforded only decomposition of the starting products. A further attempt was realized by lowering the reaction temperature, thus 2d and 2f were reacted with carbon disulde in tetrahydrofuran in the presence of triethylamine. Only compound 9f (table IV) was obtained, although in low yields. Unfortunately this attempt to synthesize the 2-butylderivative was unsuccessful. Diazotization reaction of the primary amino group gave rise to the triazoloderivatives 10a, b and df (table V). It is noteworthy that this reaction occurred immediately for 2-unsubstituted-1,4-benzothiazine and needed 240 h of stirring for 2-substituted benzothiazines. Annulation of a six-membered ring on the 1,4benzothiazine or 1,5-benzothiazepine system was very difficult. Many attempts were made with 1,2dibromoethane, phenacylbromide, ethylchlorooxalate and ethylpyruvate, but only reaction of 2 with 1,2dibromoethane was successful and gave rise to compounds 11af (table VI), operating in the presence of tetrabutylammonium bromide and powdered potassium

hydroxide in tetrahydrofuran according to phase-transfer catalysis conditions (PTC) for 11ad and reuxing in xylene for 11e and f. It is noteworthy that the building of a third nucleus was easier on the 1,5-benzothiazepine system than on the 1,4-benzothiazine one and among the 1,4-benzothiazines it was easier on the 2-unsubstituted than on the 2-substituted derivatives, perhaps because of the hindrance exerted by the substituent at the 2 position. The annulation of a six-membered ring was more difficult than that of a ve membered one.

3. Biological investigation and results All synthesized compounds were submitted to the National Cancer Institute (NCI) of Bethesda (MD) and were evaluated for in vitro anti-HIV-1 activity. All compounds were inactive with the exception of 9c which showed moderate activity (CC50 = 0.155 mM, EC50 = 0.0323 mM), the ratio between the two concentrations (therapeutic index = CC50/EC50) being 4.80.

704
Table II. Physical and chemical data of compounds 3ac, e and f.

Compound R 3a H

n 0

Yield (%) 50

M.p. (C) 208210

Formula (MW) C10H10N2O3S (238.26)

H-NMR,

3.45 (s, 2H, SCH2), 3.75 (s, 3H, OCH3), 6.83 (d, Jmeta = 2.4 Hz, 1H, Ar), 7.15 (d, Jmeta = 2.4 Hz, 1H, Ar), 8.25 (s, 1H, NHCHO), 9.60, 9.77 (2s, 2H, 2NH) (DMSO-d6) 1.29 (d, J = 7.8 Hz, 3H, CHCH3), 3.60 (q, 1H, CHCH3), 3.70 (s, 3H, OCH3), 6.80 (d, J = 2.4 Hz, 1H, Ar), 7.15 (d, J = 2.4 Hz, 1H, Ar), 8.25 (s, 1H, NHCHO), 9.61, 9.80 (2s, 2H, 2NH) (DMSO-d6) 0.75 (t, 3H, CH2CH3), 1.101.70 (m, 2H, CH2CH3), 3.103.20 (m, 1H, SCH), 3.50 (s, 3H, OCH3), 6.70 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.90 (d, Jmeta = 2.4 Hz, 1H, Ar), 8.00 (s, 1H, NHCHO), 8.10, 8.15 (2s, 2H, 2NH) (DMSO-d6) 3.72 (s, 3H, OCH3), 6.00 (s, 1H, SCHC6H5), 6.70 (br d, 1H, Ar), 6.88 (br d, 1H, Ar), 7.257.50 (m, 7H, C6H5 + NHCHO), 8.15 (s, 1H, NH) (DMSO-d6) 2.55, 3.40 (A2B2 system, 4H, CH2CH2), 3.85 (s, 3H, OCH3), 7.00 (d, J = 2.4 Hz, 1H, Ar), 7.77 (d, J = 2.4 Hz, 1H, Ar), 8.20 (s, 1H, NHCHO), 8.48, 8.56 (2s, 2H, 2NH) (CDCl3)

3b

CH3

77

203204

C11H12N2O3S (252.29)

3c

CH2CH3

23

5860

C12H14N2O3S (266.31)

3e

C6H5

27

117119

C16H14N2O3S (314.36) C11H12N2O3S (252.29)

3f

33

186188

4. Experimental protocols 4.1. Chemistry Melting points were taken on a Koer hot-stage apparatus and are uncorrected. 1H-NMR spectra were recorded in the solvent indicated, using a Bruker AC-200 (200 MHz) instrument. The chemical shift values are reported in (ppm) relative to tetramethylsilane as an internal standard. Mass spectra were recorded on a Varian MAT 311A spectrometer. Elemental microanalyses were performed for C, H and N on a Carlo Erba Elemental Analyser model 1106 and results were within 0.4% of the theoretical values. The purity of the compounds was checked by TLC (pre-coated silica-gel plates, Merck Kieselgel 60 F254). Flash chromatographies were performed on columns packed with Merck silica gel, 230400 mesh.

4.1.1. General procedure for 5-amino-7-methoxy-3,4dihydro-2H-1,4-benzothiazin-3-ones 2ae and 6-amino8-methoxy-2,3,4,5-tetrahydro-1,5-benzothiazepin-4-ones 2f (table I) Nitrobenzothiazinederivative 1ae (0.03 mol) was added portionwise under stirring to a solution of stannous chloride dihydrate (6.77 g, 0.03 mol) in concentrated HCl (10 mL). For thiazepine compound 1f (0.03 mol) a suspension of stannous chloride dihydrate (9.02 g, 0.04 mol) in tetrahydrofuran (30 mL) and concentrated HCl (2 mL) was used. The mixture was heated in a steam bath for 10 min for compound 1d, 30 min for 1a, 1 h for 1b, e and f, and 1.5 h for 1c. After cooling, in the case of 1a, an abundant precipitate was formed. It was collected by ltration, alkalinized with 5 N NaOH solution, ltered again, washed with water and nally recrystallized from EtOH to give 2a.

705
Table III. Physical and chemical data of compounds 4a, e and f and 6a, b and e.

Compound 4a

R H

R1 H

n 0

Yield (%) 71

M.p. (C) 210212

Formula (MW) C10H8N2O2S (220.25) C16H12N2O2S (296.34) C11H10N2O2S (234.27) C11H10N2O2S (234.27) C12H12N2O2S (248.30)

H-NMR,

3.78 (s, 3H, OCH3), 3.97 (s, 2H, SCH2), 6.75 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.90 (d, Jmeta = 2.4 Hz, 1H, Ar), 8.15 (s, 1H, N=CH)(DMSO-d6) 3.81 (s, 3H, OCH3), 5.90 (s, 1H, SCHC6H5), 7.04 (d, Jmeta = 2.4 Hz, 1H, Ar), 7.21 (d, Jmeta = 2.4 Hz, 1H, Ar), 7.39 (s, 5H, Ar), 8.87 (s, 1H, N=CH) (DMSO-d6) 3.17, 3.51 (A2B2 system, 4H, CH2CH2), 3.85 (s, 3H, OCH3), 6.87 (d, Jmeta = 2.4 Hz, 1H, Ar), 7.10 (d, Jmeta = 2.4 Hz, 1H, Ar), 8.71 (s, 1H, N=CH) (CDCl3) 2.85 (s, 3H, CH3), 3.78 (s, 3H, OCH3), 4.10 (s, 2H, SCH2), 6.75 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.85 (d, Jmeta = 2.4 Hz, 1H, Ar) (DMSO-d6) 1.65 (d, J = 6.3 Hz, 3H, SCHCH3), 2.85 (s, 3H, CH3), 3.84 (s, 3H, OCH3), 4.06 (q, 1H, SCHCH3), 6.78 (d, Jmeta = 2.4 Hz, 1H, Ar), 7.00 (d, Jmeta = 2.4 Hz, 1H, Ar) (CDCl3) 2.73 (s, 3H, CH3), 3.78 (s, 3H, OCH3), 5.76 (s, 1H, SCHC6H5), 6.94 (d, Jmeta = 2.4 Hz, 1H, Ar), 7.05 (d, Jmeta = 2.4 Hz, 1H, Ar), 7.36 (s, 5H, Ar) (DMSO-d6)

4e

C6H5

44

158159

4f

10

176178

6a

CH3

54

151

6b

CH3

CH3

63

166167

6e

C6H5

CH3

50

140141

C17H14N2O2S (310.37)

In the case of 1f, the mixture was evaporated in vacuo, the residue was dissolved in CHCl3 and extracted with 2 N HCl. The aqueous solution was alkalinized with 50% NaOH solution, extracted with CHCl3, dried over anhydrous Na2SO4, ltered and evaporated in vacuo. The pure solid residue was recrystallized from EtAc to give 2f. For other compounds, the reaction mixture was alkalinized with 30% ammonia solution. Compounds 2c and 2d crystallized and were collected by ltration, washed with water and recrystallized from EtOH. In the other cases the alkalinized mixture was extracted with CHCl3, washed with water, dried over Na2SO4 and dried in vacuo. The solid residue was recrystallized from EtOH. 4.1.2. General procedure for 5-formylamino-7-methoxy1,4-benzothiazine derivatives 3ac and e and 6formylamino-8-methoxy-2,3,4,5-tetrahydro-1,5-benzothiazepin-4-one 3f (table II) A solution of 2 (0.01 mol) in 30 mL formic acid (excess) was reuxed under stirring for 40 min for 2f, 2 h for 2ac and 9 h for 2e. After cooling, the solution was poured into ice-water and alkalinized with 30% ammonia

solution and the resulting precipitate was collected and recrystallized from EtOH with the exception of 3b where MeOH was used. 4.1.3. 8-Methoxy-2H-4,5-dihydro-imidazo[3,4,5-d,e]-1,4benzothiazin-4-one 4a (table III) N-(formyl)derivative 3a (2.38 g, 0.01 mol) was reuxed in concentrated HCl (10 mL) for 1 h. After cooling, the solution was poured into ice-water and alkalinized with 30% ammonia solution. The resulting precipitate was collected by ltration and recrystallized from EtOH. 4.1.4. Tentative synthesis of 5-methyl-8-methoxy-4,5dihydro-2H-imidazo[3,4,5-d,e]-1,4-benzothiazin-4-one. Formation of 5-chloro-7-methoxy-2-methyl-3,4-dihydro2H-1,4-benzothiazin-3-one 5. When the reaction described above (4.1.3) was performed using 3b, the 5-chloro-7-methoxy-2-methyl-3,4dihydro-2H-1,4-benzothiazin-3-one 5 was obtained (38% yield), m.p. 143144 C, recrystallized from EtOH [(C10H10ClNO2S)]; 1H-NMR, (DMSO-d6): 1.5 (d, J =

706
Table IV. Physical and chemical data of compounds 9ac, e and f.

Compound 9a

R H

n 0

Yield (%) 48

M.p. (C) 242243

Formula (MW) C10H8N2O2S2 (252.31) C11H10N2O2S2 (266.33)

H-NMR,

3.78 (s, 3H, OCH3), 4.05 (s, 2H, SCH2), 6.50 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.80 (d, Jmeta = 2.4 Hz, 1H, Ar), 13.18 (s, 1H, NH) (DMSO-d6) 1.48 (d, J = 6.6 Hz, 3H, CHCH3), 3.78 (s, 3H, OCH3), 4.37 (q, 1H, CHCH3), 6.50 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.78 (d, Jmeta = 2.4 Hz, 1H, Ar), 13.20 (s, 1H, NH) (DMSO-d6) 1.00 (t, 3H, CH2CH3), 1.752.05 (m, 2H, CH2CH3), 3.52 (q, 1H, SCH), 3.80 (s, 3H, OCH3), 6.75 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.90 (d, Jmeta = 2.4 Hz, 1H, Ar), 11.05 (br s, 1H, NH) (CDCl3) 3.79 (s, 3H, OCH3), 5.65 (s, 1H, SCHC6H5), 6.52 (d, J = 2.4 Hz, 1H, Ar), 6.83 (d, J = 2.4 Hz, 1H, Ar), 7.35 (m, 5H, Ar), 13.28 (s, 1H, NH) (DMSO-d6) 2.62, 3.15, 4.20 (AB2X system, 4H, CH2CH2), 3.73 (s, 3H, OCH3), 6.75 (d, J = 2.4 Hz, 1H, Ar), 6.83 (d, J = 2.4 Hz, 1H, Ar), 11.28 (s, 1H, NH) (CDCl3)

9b

CH3

89

246248

9c

CH2CH3

10

172175

C12H12N2O2S2 (280.36)

9e

C6H5

25

263265

C16H12N2O2S2 (328.41) C11H10N2O2S2 (266.34)

9f

13

117120

8.3 Hz, 3H, CH3), 3.55 (q, J = 8.3 Hz, 1H, CHCH3), 3.78 (s, 3H, OCH3), 6.75 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.83 (d, Jmeta = 2.4 Hz, 1H, Ar), 7.80 (br s, 1H, NH) ppm. 4.1.5. 8-Methoxy-2-phenyl-4,5-dihydro-2H-imidazo[3,4, 5-d,e]-1,4-benzothiazin-4-one 4e and 9-methoxy-2,4,5,6tetrahydro-imidazo[3,4,5-e,f]-1,5-benzothiazepin-5-one 4f (table III) A solution of triethyl orthoformate (1.50 g, 0.01 mol) in anhydrous xylene (5 mL) was added slowly under stirring to a suspension of 2e and f (0.005 mol) in anhydrous xylene (20 mL). The reaction mixture was reuxed with stirring for 4 h for 2e and 7 h for 2f. After cooling, the solvent was evaporated in vacuo and the solid residue was recrystallized from EtOH. 4.1.6. General procedure for 2-methyl-8-methoxy-4,5dihydro-imidazo[3,4,5-d,e]-1,4-benzothiazin-4-ones 6a, b and e (table III) A solution of triethyl orthoacetate (3.24 g, 0.02 mol) in anhydrous xylene (5 mL) was added slowly and with stirring to a suspension of 2a, b and e (0.01 mol) in

anhydrous xylene (100 mL). The mixture was reuxed for 612 h, then ltered while hot. The ltrate was concentrated under reduced pressure until ca. 40 mL. The resulting precipitate was isolated by ltration and recrystallized from EtOH. 4.1.7. Tentative synthesis of 2-methyl-9-methoxy-4H-5,6dihydroimidazo[3,4,5-e,f]-1,5-benzothiazepin-4-one. Formation of 2-ethoxy-9-methoxy-2-methyl-1H-5,6-dihydroimidazo[3,4,5-e,f]-1,5-benzothiazepin-4-one 7 and 6{[(E)-1-ethoxyethylidene]amino}-8-methoxy-2,3,4,5-tetrahydro-1,5-benzothiazepin-4-one 8 4.1.7.1. Cyclocondensation reaction in anhydrous xylene The reaction mixture, prepared as described above using 2f as starting material, was reuxed for 16 h and was evaporated under reduced pressure. The residue was puried by ash chromatography using CHCl3 as eluent and recrystallized from EtOH to give 7 (19% yield), m.p. 157159 C [(C14H18N2O3S)]; 1H-NMR, (CDCl3): 1.33 (t, J = 6.9 Hz, 3H, CH2CH3), 1.82 (s, 3H, CH3), 2.63, 3.44 (A2B2 system, 4H, CH2CH2), 3.78 (s, 3H,

707
Table V. Physical and chemical data of compounds 10a, b and df.

Compound R 10a H

n 0

Yield (%) 69

M.p. (C) 250251

Formula (MW) C9H7N3O2S (221.23) C10H9N3O2S (235.26) C13H15N3O2S

H-NMR,

3.82 (s, 3H, OCH3), 4.08 (s, 2H, SCH2), 6.80 (d, J = 1.5 Hz, 1H, Ar), 6.95 (d, J = 1.5 Hz, 1H, Ar) (DMSO-d6) 1.50 (d, J = 7.5 Hz, 3H, CHCH3), 3.80 (s, 3H, OCH3), 4.50 (q, 1H, CHCH3), 6.857.15 (m, 2H, Ar) (DMSOd6) 0.85 (t, 3H, CH3), 1.201.50 (m, 4H, CH2CH2CH2CH3), 1.701.95 (m, 2H, CH2CH2CH2CH3), 3.83 (s, 3H, OCH3), 4.254.42 (m, 1H, SCH), 6.95 (br s, 2H, Ar) (DMSO-d6) 3.80 (s, 3H, OCH3), 5.85 (s, 1H, CHC6H5), 6.807.55 (m, 7H, Ar) (DMSO-d6) 2.70, 3.40 (A2B2 system, 4H, CH2CH2), 3.88 (s, 3H, OCH3), 7.00 (d, Jmeta = 2.4 Hz, 1H, Ar), 7.05 (d, Jmeta = 2.4 Hz, 1H, Ar) (CDCl3)

10b

CH3

24

191194

10d

(CH2)3CH3

35

192195

10e 10f

C6H5 H

0 1

23 10

215217 8586

C15H11N3O2S (297.33) C10H9N3O2S (235.26)

OCH3), 4.20 (q, J = 6.9 Hz, 2H, CH2CH3), 6.32 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.85 (d, Jmeta = 2.4 Hz, 1H, Ar) ppm. 4.1.7.2. Cyclocondensation reaction in anhydrous xylene in the presence of pyridinium hydrochloride To the reaction mixture, prepared as above, was added 1.16 g (0.01 mol) of pyridinium hydrochloride. The reaction mixture was reuxed for 14 h and nally was evaporated under reduced pressure. The residue was puried by ash chromatography using CHCl3 as eluent and was recrystallized from EtOH to give 8 (72% yield), m.p. 234235 C [(C14H18N2O3S)]; 1H-NMR, (CDCl3): 1.25 (t, J = 6.9 Hz, 3H, CH2CH3), 2.60 (s, 3H, CCH3), 2.60, 3.20 (A2B2 system, 4H, CH2CH2), 3.80 (s, 3H, OCH3), 4.16 (q, J = 6.9 Hz, 2H, CH2CH3), 6.90 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.98 (d, Jmeta = 2.4 Hz, 1H, Ar), 8.90 (s, 1H, NH) ppm. 4.1.8. General procedure for 8-methoxy-4-oxo-1H,4,5dihydro-imidazo[3,4,5-d,e]-1,4-benzothiazin-2-thiones 9ac and e (table IV) Carbon disulde (20 mL) was added slowly to a solution of 2ac and e (0.01 mol) in anhydrous pyridine (50 mL). The reaction mixture was reuxed for 8 h for 2a, 14 h for 2b and e and for 30 h for 2c. After cooling the

solvent was evaporated in vacuo. The solid residue was recrystallized from EtOH to give 9a, b and e. In the case of 9c the oily residue was induced to crystallize by adding a few drops of EtOH. The compound was puried by ash chromatography using CHCl3 as eluent and nally was recrystallized from EtOH. 4.1.9. 9-Methoxy-4-oxo-1,4,5,6-tetrahydroimidazo[3,4,5e, f]-1,5-benzothiazepin-2-thione 9f (table IV) A mixture of 2f (2.24 g, 0.01 mol), carbon sulde (20 mL) and triethylamine (1.52 g, 0.015 mol) in anhydrous tetrahydrofuran (50 mL) was reuxed under stirring for 16 h. After cooling the reaction mixture was evaporated in vacuo and the residue was puried by ash chromatography using CHCl3 as eluent. The puried compound was crystallized from EtOH. 4.1.10. General procedure for 8-methoxy-4,5-dihydro1,2,3-triazolo[3,4,5-d, e]-1,4-benzothiazin-4-ones 10a, b, d and e and 9-methoxy-4H-5,6-dihydro-1,2,3-triazolo[3,4,5-e, f]-1,5-benzothiazepin-4-one 10f (table V) A solution of sodium nitrite (1.04 g, 15 mmol) in 10 mL of water was added slowly to an ice-cooled suspension of the aminoderivative 2a, b, df (0.01 mol) in 2 N HCl (10 mL).

708
Table VI. Physical and chemical data of compounds 11af.

Compound R 11a H

n 0

Yield (%) 18

M.p. (C) 136138

Formula (MW) C11H12N2O2S (236.29)

H-NMR, (CDCl3)

3.203.43 (superimposed d and t, 4H, CH2CH2 + SCH2), 3.67 (s, 3H, OCH3), 3.92 (t, J = 10.0 Hz, 2H, CH2CH2), 4.25 (s, 1H, NH), 6.00 (d, Jmeta = 2.5 Hz, 1H, Ar), 6.22 (d, Jmeta = 2.5 Hz, 1H, Ar) 1.50 (d, J = 5.1 Hz, 3H, CHCH3), 3.40 (t, 2H, CH2CH2), 3.55 (q, 1H, CHCH3), 3.73 (s, 3H, OCH3), 3.754.23 (m, 3H, CH2CH2 + NH), 6.05 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.25 (d, Jmeta = 2.4 Hz, 1H, Ar) 1.08 (t, 3H, CH2CH3), 1.502.10 (m, 2H, CH2CH3), 3.253.45 (m, 2H, CH2), 3.603.80 (m, 2H, CH2), 3.72 (s, 3H, OCH3), 4.08 (br s, 1H, NH), 4.204.35 (m, 1H, SCH), 6.05 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.25 (d, Jmeta = 2.4 Hz, 1H, Ar) 0.86 (t, 3H, CH3), 1.201.70 (m, 4H, CH2CH2CH2CH3), 1.802.00 (m, 2H, CH2CH2CH2CH3), 3.303.46 (m, 3H, NH + CH2CH2), 3.72 (s, 3H, OCH3), 4.154.28 (m, 2H, CH2CH2), 6.05 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.25 (d, Jmeta = 2.4 Hz, 1H, Ar) 3.70 (s, 3H, OCH3), 3.82, 4.25 (A2B2 system, 4H, CH2CH2), 4.07 (s, 1H, NH), 4.68 (s, 1H, SCHC6H5), 6.03 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.25 (d, Jmeta = 2.4 Hz, 1H, Ar), 7.207.40 (m, 5H, Ar) 2.653.50 (m, 8H, SCH2CH2 + NCH2CH2), 3.75 (s, 3H, OCH3), 4.20 (s, 1H, NH), 6.15 (d, Jmeta = 2.4 Hz, 1H, Ar), 6.50 (d, Jmeta = 2.4 Hz, 1H, Ar)

11b

CH3

27

118120

C12H14N2O2S (250.32)

11c

CH2CH3

oil

C13H16N2O2S (264.34)

11d

(CH2)3CH3

oil

C15H20N2O2S (292.40)

11e

C6H5

24

160162

C17H16N2O2S (312.39)

11f

45

oil

C12H14N2O2S (250.32)

After the addition was complete, compound 10a was immediately obtained as an abundant precipitate which was collected by ltration and recrystallized from MeOH. In all the other cases the reaction mixture was stirred at room temperature for 2 h for compound 2d, for 5h for compounds 2b and f and for 40 h for 2e. The resulting precipitate was ltered to give compounds 10b, d and e which were recrystallized from EtOH (10d and e) or EtAc (10c and d). In the case of the benzothiazepine derivative, as no precipitate was formed, the solution was alkalinized with dilute NaHCO3, extracted with CHCl3, washed with water, dried over anhydrous Na2SO4 and dried in vacuo. The oily residue was puried by ash chromatography using CHCl3 as eluent to give 10f.

4.1.11. General procedure for 9-methoxy-2,3,6,7tetrahydro-5H-[1,4]thiazino[4,3,2-d,e]quinoxalin-3-ones 11ad (table VI) To a stirred solution of the aminoderivatives 2ad (0.01 mol), tetrabutylammonium bromide (0.32 g, 0.001 mol) and 1,2-dibromoethane (1.88 g, 0.01 mol) in tetrahydrofuran, nely powdered potassium hydroxide (0.56 g, 0.01 mol) was added. The reaction mixture was kept at room temperature for 20 h for 2d, 48 h for 2a, 3 d for 2b and 6 d for 2c, and then ltered. The ltrate was evaporated in vacuo and the residue was taken up with chloroform, the chloroform extract washed with water, dried over anhydrous Na2SO4 and dried in vacuo. The oily residue was puried by ash chromatography using CHCl3 as eluent.

709 4.1.12. Preparation of 9-methoxy-2-phenyl-2,3,6,7-tetrahydro-5H-1,4-thiazino[4,3,2-d,e]quinoxalin-3-one 11e and 10-methoxy-2,3,6,7-tetrahydro-1H,5H-1,4-thiazepino [4,3,2-d,e]quinoxalin-5-one 11f (table VI) A suspension of aminoderivative 2e and f (0.01 mol), 1,2-dibromoethane (1.88 g, 0.01 mol), NaHCO3 (2.52 g, 0.03 mol) in anhydrous xylene was reuxed under stirring for 24 h. After cooling the reaction mixture was ltered, the ltrate was dried in vacuo and the resulting residue was puried by ash chromatography using CHCl3 as eluent. 5. Biological evaluation The procedure [8] used in the NCIs test for anti-HIV-1 screening is designed to detect agents acting at any stage of the virus reproductive cycle. Tested compounds were dissolved in dimethylsulfoxide and then diluted 1:100 in cell culture medium and then serial half-log10 dilutions were prepared. T4 lymphocytes (CEM cell line) were added and after a brief interval HIV-1 was added, resulting in a 1:200 nal dilution of the compound. Uninfected cells with the compound were used as a toxicity control and uninfected cells without the compound as basic controls. Cultures were incubated at 37 C in a 5% carbon dioxide atmosphere for 6 d. The tetrazolium salt XTT was added to all wells and cultures were incubated to allow formazan colour development by viable cells. Individual wells were analysed spectrophotometrically to quantitate formazan production and in addition were viewed microscopically for detection of viable cells and conrmation of protective activity. Drug-treated virus-infected cells were compared with drug-treated noninfected cells and with other controls (untreated-infected and noninfected cells, drugcontaining wells without cells) on the same plate. Data were reviewed in comparison with other tests done at the same time and a determination about activity was made. Anti-HIV-1 activity was expressed as 50% effective concentration (EC50) against HIV-1 cytopathic effects and drug cytotoxicity as 50% cytotoxic concentration (CC50). Acknowledgements The authors would like to express their gratitude and thanks to the staff of the anti-HIV screening division, National Cancer Institute, Bethesda, MD, USA for carrying out the in vitro anti-HIV-1 testing. Special thanks are due to the Ministero dellUniversit e della Ricerca Scientica e Tecnologica (M.U.R.S.T.) and the Consiglio Nazionale delle Ricerche (C.N.R.), Rome, for nancial support. References
[1] Pauwels R., Andries K., Desmyter J., Schols D., Kukla M., Breslin H. et al., Nature 343 (1990) 470474. [2] Parker K.A., Coburn C.A., J. Org. Chem. 56 (1991) 46004601. [3] De Clercq E., Clin. Microbiol. Rev. 10 (1997) 674693. [4] Pauwels R., in: Adams J., Merluzzi V.J., (Eds.), The Search for Antiviral Drugs, Birkhauser, Boston, 1993, pp. 71104. [5] De Clercq E., Int. J. Immunotherapy 10 (1994) 145158. [6] Grandolini G., Perioli L., Ambrogi V., Gazz. Chim. Ital. 127 (1997) 411413. [7] Liu K.C., Shih B.J., Chern J.W., J. Heterocycl. Chem. 26 (1989) 457460. [8] Weislow O.W., Kiser R., Fine D., Bader J., Shoemaker R.H., Boyd M.R., J. Natl. Cancer Inst. 81 (1989) 577586.

Eur. J. Med. Chem. 34 (1999) 711717 1999 Editions scientiques et mdicales Elsevier SAS. All rights reserved

711

Original article

Synthesis of a new series of N-hydroxy, N-alkylamides of aminoacids as ligands of NMDA glycine site
Eleonora Ghidinia*, Maurizio Delcanalea, Vittorino Servadioa, Claudio Pietraa, Marco Bergamaschia, Gino Villettia, Enrico Redentia, Paolo Venturaa, Lucio Merlinib
a

R & D Department, Chiesi Farmaceutici S.p.A., Via Palermo 26/A, 43100 Parma, Italy b DISMA, University of Milano, Via Celoria 2, I-20133 Milan, Italy (Received 16 November 1998; accepted 4 February 1999)

Abstract A new series of N-hydroxy, N-alkylamides of aminoacids structurally related to the N-hydroxy-3-amino-2 pyrrolidone [()HA-966] was synthesised and evaluated for the ability to displace [3H]Glycine, [3H]CGS19755, [3H]AMPA and [3H]Kainate binding sites. The N-heptyl glycinamide 5a was the most potent compound (IC50 = 4.5 M) in inhibiting [3H]Glycine binding. Compounds 5b, 5d, 5m, 5p, 5q and 5r showed an activity similar to ()HA-966, whereas 5h, 5i, 5n and 5s appeared less active. None of the compounds tested exhibited a signicant displacement of [3H]AMPA and [3H]Kainate binding sites. Compounds active in the [3H]Glycine binding inhibited, to a different degree, NMDA induced contractions in guinea-pig LMPP preparation. 1999 Editions scientiques et mdicales Elsevier SAS N-hydroxyamides of aminoacids / hydroxamic acids / aminoacids / glycine antagonists

1. Introduction The NMDA receptor possesses a variety of potential drug binding sites, among which the strychnineinsensitive glycine regulatory site plays an important role as up-regulator of the receptor function and as a cotransmitter site [17]. Therefore the identication of this site has stimulated intensive effort to discover selective ligands to be used as potential neuroprotective and anticonvulsant drugs. Examples of these selective compounds include the kynurenic acid analogue L-689,560 [8], the 3-substituted indole-2-carboxylate [9], the quinoxaline ACEA 1021 [10] and the 2-quinolone derivative l-701,324 [11]. Before the development of these selective glycine antagonists, manipulation of the N-hydroxy-3-amino-2-pyrrolidone [()HA-966] (gure 1) moiety has been reported to generate compounds with affinity and selectivity for the glycine site in both in vitro and in vivo functional studies [12]. To this aim, in this paper we describe the synthesis and the functional in vitro activity with regard to the antagonism of the NMDA response of a novel series of N-hydroxy, N-alkylamides
*Correspondence and reprints

Figure 1. Structure [()HA-966].

of

N-hydroxy-3-amino-2-pyrrolidone

of aminoacid derivatives 5aw obtained upon chemical manipulation of a structure corresponding to the opened ring of HA-966. In most of these compounds, the features of HA-966, i.e. the hydroxamic moiety and the amino group, have been maintained, whereas the kind, length and size of the N-linked group has been variously modied. In a few cases, also the -carbon and the NH2 group of the starting aminoacid were alkylated.

712

Figure 2. Procedures A and B used in preparing 4 and 5aw.

2. Chemistry The synthetic pathways for the synthesis of the title compounds are illustrated in gure 2 (procedure A and B). The new derivatives synthesised are numbered in table I. Usually, hydroxamic acids are prepared following classical methods of literature [13]. We have found that N-alkylhydroxylamines undergo condensation with succinic esters of aminoacids under very mild conditions [14]. The reaction proceeds through an intermediate 3 (gure 2, route A), which is soon formed, derived from the attack of the hydroxylamine oxygen on the succinic ester carboxylic group. In a few cases, the intermediate was isolated and characterised by 1H NMR. This intermediate slowly rearranges to the product 4 [15]. The route B was followed when the rearrangement of the intermediate 3, which is formed in route A, was too slow to give the product. To avoid the attack of the hydroxylamine oxygen on the active ester, the former was protected by reaction with trimethylchlorosilane and the

silylated hydroxylamine was then reacted with the active ester in pyridine. The appropriate monosubstituted N-hydroxylamines were conveniently prepared by a well-described synthesis [16] from the corresponding oxime [17, 18], while the other reactions just involve classical reactivity of aminoacids [19]. The structures of the products 5aw were conrmed by IR and 1H NMR. Although hydroxamic derivatives may exist as a mixture of Z and E isomers, only the former was present in DMSO at room temperature and at the concentration used for measuring the NMR spectra, in accordance to what is reported in the literature [20]. The conformation in solution was also conrmed by n.O.e.s experiments on the corresponding hydrochlorides in order to be able to observe the OH signal. Indeed, when the spectra are recorded on the corresponding free bases, the protons on heteroatoms are in rapid exchange with traces of water from DMSO-d6.

713
Table I. Physical properties and observed biological activity of compounds 5.

Compound R1 ()HA-966 5a 5b 5c 5d 5e 5f 5g 5h 5i 5j 5k 5l 5m 5n 5o 5p 5q 5r 5s 5t 5u 5v 5w
a

R2 H H H methyl methyl H H H H H H H H H butyl H H H H phenyl H hexyl butyl

R3 H H H H H penthyl H H H H methyl H H H H H H H H H H H H

M.W. 188.27 132.16 146.19 202.3 202.3 160.22 216.33 180.21 172.23 176.17 202.3 194.24 188.27 186.26 244.38 210.23 238.25 225.21 214.65 264.37 202.3 272.43 244.38

Salt

Conguration

Yield (%) 64 34 24 22 22 43 27 51 59 45 35 25 64 74 8 10 16 16 71 11 16 8 26

IC50a (M) % Inhibitionb 22 4.5 20 N.A. 49 N.A. N.A. N.A. 91 51 N.A. N.A. N.A 29 62 N.A. 27 28 26 42 N.A. N.A. N.A N.A. 24 100 15 N.T. 10 N.T. N.T. N.T. 19 22 N.T. N.T. N.T. 78 45 N.T. 7 20 10 13 N.T. N.T. N.T. N.T.

heptyl isopropyl butyl heptyl heptyl H nonyl benzyl cyclohexyl ethoxycarbonylmethyl heptyl 4-methylbenzyl 2,3-dimethylpenthyl cyclohexylmethyl heptyl 4-methoxybenzyl 4-carbomethoxybenzyl 4-nitrobenzyl 4-chlorobenzyl heptyl 1-methylheptyl heptyl heptyl

R S

HCl HCl HCl

RS

HCl HCl HCl HCl RS RS S

Potencies of compounds in inhibiting [3H]Glycine binding in rat cortical homogenates. N.A.: not active up to 1 mM; bEffect on the inhibition of NMDA (1mM)-induced contractions in guinea-pig LMPP. Each compound was tested at 100 M. N.T.: not tested; cAs diastereomeric mixture.

3. Pharmacology Compounds 5aw were tested for their ability to displace the [3H]Glycine binding accordingly to the procedure described in the experimental protocol. Table I shows the results in comparison with that obtained with the parent compound ()HA-966. 5a was the most potent compound in inhibiting [3H]Glycine binding, IC50 being 4.5 M. N-isopropylglycinamide 5b and N-heptyl-(D)alaninamide 5d derivatives as well as compounds 5m, 5p, 5q and 5r showed an activity similar to ()HA-966. The N-benzyl, N-cyclohexyl, N-cyclohexylmethyl, N-(4-Clbenzyl)glycinamide derivatives, respectively 5h, 5i, 5n and 5s, appeared less active than ()HA-966. The remaining compounds tested were inactive in displacing [3H]Glycine binding up to a concentration of 1 mM. As for the interaction with the competitive site of NMDA receptors, only compounds 5a, 5b, 5d, 5h and 5i

showed a weak activity in the displacement of [3H]CGS19755 binding, their IC50s being 16, 57, 72, 95 and 51 M, respectively. No signicant displacement of [3H]Kainate or [ H]AMPA binding sites was observed for all compounds (up to 1 mM) under examination. As expected, also HA-966 (up to 1 mM) was inactive on [3H]CGS19755, [3H]Kainate and [3H]AMPA binding. Compounds showing a signicant displacement of [3H]Glycine or [3H]CGS19755 binding sites exhibited a different degree of inhibition of NMDA induced contractions in the guinea-pig LMPP in vitro, when tested at the concentration of 100 M. Accordingly with the [3H]Glycine binding data, compound 5a was the most potent in inhibiting NMDA induced contractions. None of the compounds under examination including () HA-966 exhibited, per se, a contractile response.
3

714 The SAR in this class of compounds was explored, changing rst the substituents on the nitrogen atom of the hydroxamic acids. The activity seems to depend rather critically on the size and length of the alkyl group, the highest activity being reached for the N-heptyl group (5a), whereas shortening (5c) or lengthening (5g) of the linear chain was detrimental, as well as -methylation (5u). Activity, although lower, was retained with bulkier groups, such as isopropyl (5b), 2,3-dimethylpentyl (5m), cyclohexyl (5i), or cyclohexylmethyl (5n). Only a small (methyl) substituent seems to be tolerated on the aminoacid -carbon, (cf. 5d vs. 5o, 5t, 5v, 5w) but this is dependant on the conguration (5d vs. 5e). The only example of an unsubstituted hydroxamic acid 5f, was also inactive. Attempts to nd a linear correlation with one or more parameters representative of these substituents, such as logP, WDW volume, molecular connectivity or length failed (see experimental part). Similar negative results were obtained by applying a principal component analysis. Therefore, only a qualitative discussion of the results reported above seems possible. All the compounds showed no, or scarce, binding affinity towards other excitatory aminoacid receptors, such as glutamate, kainate and AMPA receptors. In conclusion, the N-alkylamides of aminoacids related to the structure of () HA-966 demonstrated a varying degree of affinity for the glycine binding site on the NMDA receptors, 5a being the most potent and compounds 5p, 5q, 5r and 5s more selective versus the other excitatory aminoacids evaluated. The antagonist effect of these substances was also conrmed in the functional studies in vitro on guinea-pig LMPP. In this respect, the level of inhibition showed by () HA-966 is consistent with that reported in other experiments [21]. Since () HA-966 is dened also as a glycine partial agonist, it is likely that these compounds behave in such a manner. However, this issue was not addressed in this study. More experimental work at the electrophysiological level or at the modulation on NMDA receptors in the presence or absence of glycine need to be performed in order to clarify this point. 4. Experimental protocols 4.1. Chemistry Melting points were determined on a Bchi capillary melting point apparatus and are uncorrected. The IR spectra were determined on a Perkin Elmer 1310 spectrophotometer. 1H NMR spectra were recorded at 200.13 MHz on a Bruker ACF 200 spectrometer; chemical shifts are in (ppm), with tetramethylsilane as internal standard. Mass spectra were measured with a Fisons-VG Trio 2000 single quadrupole spectrometer equipped with a dual EI/CI source. The []D were performed on a Perkin-Elmer 241 and 241-MC polarimeter. Sodium sulphate was employed as a drying agent for ether extracts. The petroleum ether used throughout this work had a boiling point of 4070 C. TLC on silica gel plates (Merck, 60 F 254) was used to check product purity and the spots detected with ninhydrin. The intermediate 3 can be monitored by TLC also, with potassium ferricyanideferric chloride giving a typical blue spot, whereas the nal product 4 gives a brown one with ferric chloride. Preparative chromatography was carried out on silica gel ICN (3263 m). The structures of all compounds were consistent with their analytical and spectroscopic data. 4.1.1. General synthesis of 4 (route A) To a mixture of the appropriate N-monosubstituted hydroxylamine 1 (85 mmol) in dichloromethane (360 mL) at 0 C was added the succinic ester of the corresponding N-protected aminoacid 2 (82 mmol). The solution was stirred for 1 h at 10 C and at room temperature for a few days. The composition of the crude product was determined by TLC and infrared spectroscopy. The intermediate 3 slowly rearranges to the product 4. When the conversion was complete, the solution was poured in water. The separated organic phase was washed with a 5% sodium bicarbonate solution (2 50 mL), with water (2 50 mL) and dried. The solvent was removed in vacuo at 35 C. The compound 4 was nally deprotected following literature methods. The following compounds were prepared according to this procedure from the corresponding N-protected (Cbz or Boc) aminoacid. 4.1.1.1. N-Hydroxy, N-heptylglycinamide 5a White solid; IR (Nujol) 3 355, 3 280, 1 630, 1 540 cm1; 1H NMR (DMSO-d6) 0.9 (t, 3H), 1.01.4 (s, 8H), 1.5 (m, 2H), 3.5 (t, J = 7.0 Hz, 2H), 3.9 (s, 2H), 4.04.5 (br, 3H); EI+/MS (m/z) 188 (M+). 4.1.1.2. N-Hydroxy, N-heptylalaninamide 5d and 5e Pale oil; IR (CHCl3) 3 060, 1 610, 1 580 cm1; 1H NMR (DMSO-d6) 0.8 (t, 3H), 1.0 (d, J = 6.8 Hz, 3H), 1.11.4 (m, 8H), 1.5 (m, 2H), 3.33.6 (m, 2H), 3.8 (q, J = 6.8 Hz, 1H), 3.94.5 (br, 3H); EI+/MS (m/z) 203 (M+.); 5d: [a]D = 10.5 (c = 0.6, EtOH); 5e: [a]D = + 11.2 (c = 0.6, EtOH). 4.1.1.3. N-Hydroxy, N-benzylglycinamide 5h White solid; m.p. 118120 C; IR (Nujol) 3 320, 3 290, 2 500, 1 800, 1 630, 1 550, 720, 685 cm1; 1H NMR

715 (DMSO-d6) 3.4 (s, 2H), 4.7 (s, 2H), 3.54.5 (br, 3H), 7.17.5 (m, 5H). 4.1.1.4. N-Hydroxy, N-cyclohexylglycinamide 5i White solid; m.p. 128129 C; IR (CHCl3) 3 3603 050, 1 620 cm1; 1H NMR (DMSO-d6) 0.81.6 (m, 8H), 1.7 (m, 2H), 3.3 (s, 2H), 4.1 (m, 1H), 3.54.5 (br, 3H); EI+/MS (m/z) 173 (M+.). 4.1.1.5. N-Hydroxy, N-ethoxycarbonylmethylglycinamide 5j White solid; m.p. 159162 C; IR (Nujol) 3 5002 600, 1 720, 1 640, 1 600 cm1; 1H NMR (DMSO-d6) 1.2 (t, J = 7 Hz, 3H), 3.8 (s, 2H), 4.1 (q, J = 7 Hz, 2H), 4.4 (s, 2H), 8.3 (br, 3H), 10.7 (br, 1H); EI+/MS (m/z) 176 (M+.). 4.1.1.6. N-Hydroxy, N-heptyl, N-methylglycinamide 5k White solid; IR (Voltalef) 3 430, 3 300, 2 400, 1 850, 1 630 cm1; 1H NMR (DMSO-d6) 0.9 (t, 3H), 1.11.5 (m, 8H), 1.6 (m, 2H), 2.5 (s, 2H), 3.5 (s, 2H), 3.6 (t, J = 7Hz, 2H), 3.64.4 (br, 3H); EI+/MS (m/z) 203 (MH+.). 4.1.1.7. N-Hydroxy, N-(4-methylbenzyl)glycinamide 5l White solid; m.p. 197199 C; IR (Nujol) 3 2003 000, 1 640, 880, 780 cm1; 1H NMR (DMSO-d6) 2.3 (s, 3H), 3.8 (s, 2H), 4.7 (s, 2H), 7.07.2 (AB syst., 2H), 8.2 (br, 3H), 10.5 (br, 1H); EI+/MS (m/z) 194 (M+.). 4.1.1.8. N-Hydroxy, N-(2,3-dimethylpenthyl)glycinamide 5m White solid; 1H NMR (DMSO-d6) 0.8 (d; d, J = 6.8 Hz, 3H), 0.9 (d; t, 6H), 0.91.5 (m, 3H), 1.0 (m, 1H), 3.4 (m, 2H), 3.8 (q, J = 7 Hz, 2H), 8.2 (br, 3H), 10.5 (br, 1H); EI+/MS (m/z) 188 (M+.). 4.1.1.9. N-Hydroxy, N-cyclohexylmethylglycinamide 5n White solid; IR (Nujol) 3 5002 300, 1 640 cm1; 1H NMR (DMSO-d6) 0.9 (m, 2H), 1.11.3 (m, 3H), 1.51.9 (m, 6H), 3.3 (d, J = 6.8 Hz, 2H), 3.4 (s, 2H), 3.24.2 (br, 3H); EI+/MS (m/z) 186 (M+.). 4.1.1.10. N-Hydroxy, N-heptyl, norleucinamide 5o White solid; 1H NMR (DMSO-d6) 0.9 (t; t, 6H), 1.11.5 (m, 12H), 1.6 (m, 2H), 1.8 (m, 2H), 3.4 (m, 1H), 3.7 (m, 1H), 4.2 (t, J = 7 Hz, 1H), 8.3 (br, 3H), 10.5 (br, 1H); EI+/MS (m/z) 244 (M+.). 4.1.1.11. N-Hydroxy, N-(4-methoxybenzyl)glycinamide 5p White solid; IR (KBr) 3 320, 3 290, 1 640; 1H NMR (DMSO-d6) 3.3 (s, 2H), 3.7 (s, 3H), 3.74.3 (br, 3H), 4.6 (s, 2H), 6.9 (A of AAXX syst., 2H), 7.2 (XX of AAXX syst., 2H); EI+/MS (m/z) 210 (M+.). 4.1.1.12. N-Hydroxy, N-(4-carbomethoxybenzyl)glycinamide 5q White solid; IR (KBr) 3 330, 3 290, 1 650 cm1; 1H NMR (DMSO-d6) 3.4 (s, 2H), 3.54.2 (br, 3H), 3.9 (s, 3H), 4.8 (s, 2H), 7.4 (d, J = 8.2 Hz, 2H), 8.0 (d, J = 8.2 Hz, 2H); EI+/MS (m/z) 239 (M+.). 4.1.1.13. N-Hydroxy, N-(4-nitrobenzyl)glycinamide 5r White solid; m.p. 179181 C; IR (KBr) 3 320, 3 2002 500, 1 680 cm1; 1H NMR (DMSO-d6) 3.9 (s, 2H), 4.9 (s, 2H), 7.6 (d, J = 8.2 Hz, 2H), 8.2 (d, J = 8.2 Hz, 2H), 8.3 (br, 3H), 10.8 (br, 1H); EI+/MS (m/z) 225 (M+.). 4.1.1.14. N-Hydroxy, N-(4-chlorobenzyl)glycinamide 5s White solid; m.p. 183185 C; IR (KBr) 3 5002 500, 1 660 cm1; 1H NMR (DMSO-d6) 3.9 (s, 2H), 4.7 (s, 2H), 7.27.5 (AB syst., 4H), 8.3 (br, 3H), 10.7 (br, 1H); EI+/MS (m/z) 214 (M+.). 4.1.1.15. N-Hydroxy, N-heptyl,2-phenylglycinamide 5t White solid; IR (nujol) 3 310, 3 1003 000, 1 630 cm1; 1H NMR (DMSO-d6) 0.8 (t, 3H), 1.01.4 (m, 8H), 1.5 (m, 2H), 3.33.7 (br, m, 5H), 5.0 (s, 1H); EI+/MS (m/z) 264 (M+.). 4.1.1.16. N-Hydroxy, N-(1-methylheptyl)glycinamide 5u White solid; m.p. > 95 C; 1H NMR (DMSO-d6) 0.8 (t, 3H), 1.1 (d, J = 6.8 Hz, 3H), 1.11.5 (m, 9H), 1.6 (m, 1H), 3.8 (s, 2H), 4.4 (m, 1H), 8.3 (br, 3H), 10.7 (br, 1H); EI+/MS (m/z) 202 (M+.). 4.1.1.17. N-Hydroxy, N-heptyl,2-hexylglycinamide 5v White solid; IR (KBr) 3 320, 3 280, 3 0002 800, 1 640, 1 620 cm1; 1H NMR (DMSO-d6) 0.8 (t, 6H), 1.11.5 (m, 16 H), 1.6 (m, 2H), 1.8 (m, 2H), 3.4 (m, 1H), 3.54.2 (br, 3H), 3.7 (m, 1H), 4.2 (t, 1H); EI+/MS (m/z) 272 (M+.). 4.1.1.18. N-Hydroxy, N-heptylnorleucinamide 5w White solid; 1H NMR (DMSO-d6) 0.8 (t, 6H), 1.11.5 (m, 12H), 1.8 (m, 2H), 3.5 (m, 1H), 3.7 (m, 1H), 8.3 (br, 3H), 10.7 (br, 1H); EI+/MS (m/z) 244 (M+.). 4.1.2.General synthesis of 4 (route B) The appropriate hydroxylamine hydrochloride 1 (80 mmol) was dissolved in dichloromethane (200 mL), added with potassium carbonate (80 mmol) and stirred for 10 min at room temperature. The solution was ltered and the ltrate was cooled to 10 C. Pyridine (350 mL) and trimethylchlorosilane (800 mmol) were added and the solution was stirred at room temperature for 30 min. The solution was cooled again to 10 C and the succinic

716 ester of the corresponding N-protected aminoacid 2 was added (82 mmol). The reaction was stirred at room temperature for a further 2 d. The solution was treated with hydrogen chloride until pH = 1 and the product was extracted with ethyl acetate (3 70 mL). The organic layer, that contained only one product, was dried and the solvent evaporated in vacuo. 4.1.2.1. N-Hydroxy, N-isopropylglycinamide 5b White solid; m.p. 5859 C; IR (CHCl3) 3 350, 1 640 cm1; 1H NMR (DMSO-d6) 1.1 (d, J = 6.8 Hz, 6H), 4.0 (s, 2H), 4.5 (m, 1H), 8.3 (br, 3H), 10.4 (br, 1H); EI+/MS (m/z) 132 (M+.). 4.1.2.2. N-Hydroxy, N-butylglycinamide 5c White solid; 1H NMR (DMSO-d6) 0.9 (t, J = 7 Hz, 3H), 1.3 (m, 2H), 1.6 (m, 2H), 3.6 (t, J = 7.0 Hz, 2H), 3.9 (s, 2H), 4.1 (br, 3H); EI+/MS (m/z) 176 (M+.). 4.1.2.3. N-Hydroxy, N-penthylglycinamide 5f Pale oil; m.p. 130132 C; IR (oil) 3 3002 300, 1 620 cm1; 1H NMR (DMSO-d6) 0.8 (t, 3H), 1.11.5 (m, 6H), 2.4 (t, J = 7.0 Hz, 2H), 3.0 (s, 2H), 3.13.6 (br, 3H); EI+/MS (m/z) 161 (M+.). 4.1.2.4. N-Hydroxy, N-nonylglycinamide 5g White solid; m.p. 118120 C; IR (Nujol) 3 320, 3 290, 2 500, 1 800, 1 630, 1 550, 720, 685 cm1; 1H NMR (DMSO-d6) 3.4 (s, 2H), 4.7 (s, 2H), 3.54.5 (br, 3H), 7.17.5 (m, 5H). 4.2. Pharmacology Assays of [3H]Glycine binding were performed as described by Snell et al. [22]. Rat brain cortices were removed from male Wistar rats in order to prepare a membrane fraction by standard techniques. A quantity of 12.5 mg of membrane preparation was incubated with 20 nM [3H]Glycine for 30 min at 4 C. Non-specic binding was estimated in the presence of 10 M glycine. Membranes were ltered and washed 3 times and the lters were counted to determine [3H]Glycine specically bound. Assays of [3H]CGS19755 binding were performed as described by Murphy et al. [23]. 20 mg of membrane preparation was incubated with 10 nM [3H]CGS19755 for 20 min at 4 C. Non-specic binding was estimated in the presence of 100 M L-glutamate. Membranes were ltered and washed 3 times and the lters were counted to determine [3H]CGS19755 bound. The binding to kainic receptors was measured by using [3H]Kainate [24], a selective agonist that binds to the kainate subtype of the ionotropic glutamate receptors present in rat brain. 15 mg of membrane preparation was incubated with 5 nM [3H]Kainate for 1 h at 4 C. Nonspecic binding was estimated in the presence of 1 mM native L-glutamate. Membranes were ltered and washed 3 times to separate free from bound ligand and lters were counted to determine [3H] Kainate bound. The binding to AMPA receptors was measured by using [3H]AMPA [25]. Briey, 20 mg of membrane preparation was incubated with 10 nM [3H]AMPA for 30 min at 4 C. Membranes were ltered and washed 3 times to separate free from bound ligand and lters were counted to determine [3H]AMPA bound. All conditions were tested in triplicate. For each compound, concentration response changes in [3H]Glycine, [3H]Kainate or [3H]CGS19755 binding were tted to a sigmoidal curve and IC50 values were estimated using a curve tting program to the logistic equation described by De Lean et al. [26]. Functional studies in vitro were performed in the guinea-pig ileum muscle/myenteric plexus preparation (LMPP) [27]. Two cm lenghts of intestine were stretched over a glass pipette. The LMPP was separated from the underlying circular muscle by a cotton-tipped application and mounted in a 20 mL organ bath contained Mg++-free Krebs buffer (composition in mM: NaCl 134, KCl 3.4, CaCl2 2.8, KH2PO4 1.3, NaHCO3 16 and glucose 7.7) at 37 C bubbled with 5% CO2 in oxygen. Contractions were recorded isometrically (0.5 g tension) with transducers connected to a poligraph. After 1 h of equilibration period, tissues were contracted with NMDA 1mM. Compounds that showed a signicant activity in the displacement of [3H]Glycine binding sites were evaluated in their ability to antagonise the NMDA response. The molecular models of the compounds were built on a Silicon Graphics IRIS 35, using the program Insight II. The program Discover (Molecular Simulation Inc., San Diego, CA) was used to generate low-energy conformations. The values of van der Waals volumes (V) and the length of R groups (L) were calculated on these rened conformations via Insight II. Calculation of logP was carried out using atomic parameters derived by Ghose and Crippen [28, 29]. The structure-activity relationships were explored with the program Systat 5.0 (Systat Inc.), used to nd possible simple correlations between biological activity and each of the above parameters logP, V or L and to perform a principal component analysis (PCA). In both cases no signicant relationships were found.

717 Acknowledgements The authors are grateful to Mr. Milco Lipreri for technical support.
[13] [14] [15] [16] Kehl H. (Ed.), Chemistry and Biology of Hydroxamic Acids, Karger, NewYork, 1982. Chiesi P., Ventura P., Servadio V., Italian Patent IT 01255240, (1992). Stroh R. (Ed.), Houben-Weyl Handbuch der Organischen Chemie Vol 10/1, Georg Thieme Velag, Stuttgart, 1971. Borch R.F., Bernstein M., Durst D., J. Am. Chem. Soc. 93 (1971) 28972904. Lachman A., Org. Synth. Coll. Vol. II (1943) 7071. Kahr K., Berther C., Chem. Ber. 93 (1960) 132134. Barrett G.C., Chemistry and Biology of Amino Acids, Chapman and Hall, London, 1985. Brown D.A., Glass W.K., Mageswaran R., Mohammed S.A., J. Magn. Reson. 29 (1991) 4045. Campbell B.G., Couceyro P., Keana J.F.W., Weber E., J. Pharmacol. Exp. Ther. 257 (1991) 754766. Snell L.D., Morter R.S., Johnson K.M., Eur. J. Pharmacol. 156 (1988) 105110. Murphy D.E., Adler J., Br. J. Pharmacol. 95 (1988) 932938. London E.D., Coyle J.T., Mol. Pharmacol. 15 (1979) 492505. Olsen R.W., Szamraj O., Houser C.L., Brain. Res. 402 (1987) 243254. De Lean A., Munson P.J., Rodbard D., Am. J. Physiol. 235 (1978) E97E105. Shannon H.E., Sawyer B.D., J. Pharmacol. Exp. Ther. 251 (1989) 518523. Ghose A.K., Pritchett A., Crippen G.M., J. Comp. Chem. 9 (1988) 8090. Viswanadhan V.N., Ghose A.K., Revankar G.R., Robins R.K., J. Chem. Inf. Comput. Sci. 29 (1989) 163172.

References
[1] [2] [3] [4] [5] Johnson J.W., Ascher P., Nature 325 (1987) 529531. Small D.L., Buchan A.M., Int. Rev. Neurobiol. 40 (1997) 137171. Sucher N.J., Awobuluyi M., Choi Y.B., Lipton S.A., Trends Pharmacol. Sci. 17 (1996) 348355. Leeson P.D., Iversen L.L., J. Med. Chem. 37 (1994) 40534067. Krogsgaardlarsen P., Ebert B., Lund T.M., Braunerosborne H., Slok F.A., Johansen T.N., Brehm L., Madsen U., Eur. J. Med. Chem. 31 (1996) 515537. Nankai M., Klarica M., Fage D., Carter C., Prog. NeuroPsycopharmacol. Biol. Psychiatry 22 (1998) 3564. Bigge C.F., Nikam S.S., Exp. Opin. Ther. Pat. 7 (1997) 10991114. Grimwood S., Moseley A.M., Carling R.W., Leeson P.D., Foster A.C., Mol. Pharmacol. 41 (1992) 923930. Gaviraghi G., Pietra C., Ratti E., Trist D., Cerebrovascular Dis. 6 (1996) P1220. Woodward R.M., Huettner J.E., Guastella J., Keana J.F.K., Weber E., Mol. Pharmacol. 47 (1995) 568581. Kulagowski J.J., Baker R., Curtis N.R., Leeson P.D., Mawer I.M., Moseley A.M.E.T. et al., J. Med. Chem. 37 (1994) 14021405. Fletcher E.J., Loger D., Eur. J. Pharmacol. 151 (1988) 161162.

[17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29]

[6] [7] [8] [9] [10] [11] [12]

Eur. J. Med. Chem. 34 (1999) 719727 1999 Editions scientiques et mdicales Elsevier SAS. All rights reserved

719

Original article

Synthesis and binding affinities for 5-HT1A, 5-HT2A and 5-HT2C receptors of a series of 1- and 2-(4-arylpiperazinylalkyl)-4-(benzoyl)-1,2,3-triazole derivatives
Giuseppe Caliendoa*, Ferdinando Fiorinoa, Paolo Griecoa, Elisa Perissuttia, Vincenzo Santagadaa, Stefania Albriziob, Loredana Spadolab, Giancarlo Brunic, Maria Rosaria Romeoc
a

Dipartimento di Chimica Farmaceutica e Tossicologica, Universit di Napoli Federico II, Via D. Montesano, 49-80131 Napoli, Italy b Dipartimento di Scienze Farmaceutiche, Universit di Salerno, Piazza V. Emanuele 9, 84080, Penta di Fisciano, Salerno, Italy c Istituto di Farmacologia, Universit di Siena- Via delle Scotte, 6 - 53100 Siena, Italy (Received 9 September 1998; accepted 4 February 1999)

Abstract A number of 1- and 2-(4-arylpiperazinylalkyl)-4-(benzoyl)-1,2,3-triazole derivatives (14) were prepared in order to obtain compounds with a high affinity and selectivity for 5-HT1A receptors. 5-HT1A, 5-HT2A and 5-HT2C affinities were determined by radioligand binding experiments and the most active compounds were also tested for binding affinities on dopaminergic D-1, D-2 and adrenergic 1, 2 receptors. The modication of aromatic substituents, the length of the alkyl chain and its position on the 4-benzoyl-1,2,3-triazole ring were explored. Most of the considered compounds generally showed moderate to high affinity for the 5-HT1A receptor binding site. Three derivatives 2c, 3c and 3e bind to 5-HT1A receptors in the nanomolar range (IC50 values = 2, 7.2 and 2.6 nM respectively). The most active compound, 2c, presented a high degree of selectivity versus all considered receptors. It was found that the benzoyltriazole derivatives 1h and 4c are new selective ligands for 5-HT2A (IC50 = 89 nM) and 5-HT2C receptors (IC50 = 17 nM), respectively. 1999 Editions scientiques et mdicales Elsevier SAS arylpiperazines / synthesis / 5-HT receptor / serotonin

1. Introduction The neurotransmitter serotonin is involved in various physiological (e.g. sleep and thermoregulation) and pathological processes (e.g. migraine and depression). A classication of 5-HT receptor subtypes, their role in various CNS activities and their respective ligands were recently reviewed [1, 2]. One of the serotonin receptor subtypes, 5-HT1A, plays an important role as the somatodendritic autoreceptor (presynaptic) in the dorsal raphe nucleus and as a postsynaptic receptor for 5-HT in terminal areas [3]. Buspirone and ipsapirone, which are agonists and display high affinity to the 5-HT1A receptor (IC50 = 60 and 35 nM, respectively), are presently being used as anti-anxiety agents [4, 5]. It seems useful to
*Correspondence and reprints

develop ligands that are selective for the 5-HT1A subtype to facilitate the study and characterization of this receptor, but also in view of their potential use as anxiolytics.. The most commonly used ligand, 8-OH-DPAT (8hydroxy-2-(N-N-di-n-propylamino)tetralin), is a potent 5-HT1A agonist, and in fact, the tritium-labelled compound is the ligand of choice for 5-HT1A receptor binding studies (Kd = 0.5 nM, rat hippocampal homogenates). One class of compounds with affinity for the 5-HT1A receptor is represented by the arylpiperazine derivatives [6] (general structure I) where R is a heterocyclic nucleus.

720
Table I. Physicochemical properties of 4-benzoyl-1,2,3-triazole derivatives.

1-Substituted 4-benzoyl-1,2,3-triazoles X H o-Cl m-Cl p-Cl o-OCH3 p-OCH3 o-F p-F H o-Cl m-Cl p-Cl o-OCH3 p-OCH3 o-F p-F
a

2-Substituted 4-benzoyl-1,2,3-triazoles Compoundb 2a 2b 2c 2d 2e 2f 2g 2h 4a 4b 4c 4d 4e 4f 4g 4h M.p. (C) 206 212 180 191 206 200 164 196 251 177 178 248 153 208 179 208 207 214 182 193 208 201 165 197 252 179 180 249 155 210 180 209 Yieldc % 35 32 35 37 40 38 42 38 42 45 38 40 41 34 42 39

n 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3

Formula

M.W. 397.91 432.34 432.34 432.34 428.01 428.01 415.89 415.89 411.93 446.35 446.35 446.35 441.95 441.95 429.92 429.92

Compound 1a 1b 1c 1d 1e 1f 1g 1h 3a 3b 3c 3d 3e 3f 3g 3h

M.p. (C) 201 208 204 236 210 218 142 216 183 334 231 217 207 220 188 187 202 210 206 238 212 220 143 217 184 335 232 219 209 221 189 188

Yield % 25 23 24 25 28 20 28 24 27 27 22 27 28 22 28 25

C21H23N5OHCl C21H22ClN5OHCl C21H22ClN5OHCl C21H22ClN5OHCl C22H25N5O2HCl C22H25N5O2HCl C21H22N5OFHCl C21H22N5OFHCl C22H25N5OHCl C22H24ClN5OHCl C22H24ClN5OHCl C22H24ClN5OHCl C23H27N5O2HCl C23H27N5O2HCl C22H24N5OFHCl C22H24N5OFHCl

Satisfactory microanalyses obtained: C, H, N, Cl, F values are within 0.4% of the theoretical ones; bAll compounds were crystallized by methyl alcohol and diethyl ether. cYield refers to the single structural isomer after separation by chromatography as free bases.

Affinity and specicity depend critically on the nature of the specic heterocyclic nucleus (R), on the (X) aromatic substitution and on the length (n) of the polymethylene chain. In particular, it was observed that different terminal groups (R) play an important role in the interaction with a corresponding hydrophobic region of the 5-HT1A receptor [7]. We recently reported a series of arylpiperazines of general structure I [8] having a benzotriazole group as a terminal (R) moiety with mixed 5-HT1A, 5-HT2A and 5-HT2C receptor affinity. Structure-affinity relationships in this series [8] showed that nanomolar affinity towards the 5-HT1A receptor requires a methoxy group at the ortho-position of the phenyl ring. However, the most potent benzotriazole ligands of the 5-HT1A receptor have poor selectivity, e.g., they possess affinities for the 5-HT2A receptor only 1050 times lower than those measured for 5-HT1A. Considering the important role played by the (R) terminal group, in an attempt to increase both affinity and selectivity for the 5-HT1A receptor the benzotriazole nucleus of these arylpiperazine derivatives [8], was re-

placed by a 4-benzoyl-1,2,3-triazole group. These compounds (14) represent open chain analogues of their benzotriazole counterparts with bioisosteric properties. Most of the arylpiperazine moieties used are those previously reported [8] displaying the highest affinity for the 5-HT1A receptor, typied by 1-(2- or 4-methoxyphenyl)-, 1-phenyl-, 1-(2- or 3- or 4-chlorophenyl)piperazine. In this work we considered also a new electronwithdrawing substituent such as a uoro group introduced into the 2- or 4- position of the phenylpiperazine moiety. The piperazine is connected to the terminal benzoyltriazole system via bridges of two (series 1 and 2) or three (series 3 and 4) methylene groups. Herein we report the synthesis of a series of 1- and 2-(4-arylpiperazinylalkyl)-4-(benzoyl)-1,2,3-triazole derivatives 14, (table I) and the affinities for 5-HT1A, 5-HT2A and 5-HT2C receptors obtained by radioligand binding studies. The most active compounds were also tested for their affinity at dopaminergic D-1, D-2 and adrenergic 1 and 2 receptors to substantiate their pharmacological prole in view of potential therapeutic uses.

721

Figure 1. Reagents: (a) CrO3/H2SO4; (b) NaN3; (c) N-arylpiperazine/K2CO3; (d) HCl (anhydrous).

2. Chemistry 4-benzoyl-1,2,3-triazole derivatives 1ah, 2ah, 3ah and 4ah (table I) were synthesized as described in gure 1. The parent compound 4-benzoyl-1,2,3-triazole (7) was obtained in 95% yield by a modied method described in the literature [9, 10]. Oxidation of the phenylethynylcarbinol 5 with chromium trioxide and concentrated sulphuric acid produced phenyl ethynyl ketone 6, which was successively treated with NaN3 in anhydrous dimethylacetamide providing the desired compound. Alkylation of the aromatically substituted 1-(2chloroethyl)- or 1-(3-chloropropyl)-4-phenylpiperazines, prepared as described in the literature [8], with 4-benzoyl-1,2,3-triazole in butan-2-one in the presence of potassium carbonate afforded a mixture of the expected 1-, 2- and 3- triazole isomers with an overall yield in the range of 4580%. The three isomers were separated by chromatography on a silica gel column using n-hexane/diethyl ether 95:5 as eluent. The faster moving 2-substituted isomers (2 and 4) were collected with a higher yield with respect to the slower moving 1-substituted isomers (1 and 3). The yield of intermediate moving 3-substituted isomers (8) was so low that they were not further considered. The free bases were converted into their corresponding hydrochlorides by usual methods. All the nal products were further puried by crystallization from a mixture of diethyl ether and ethyl alcohol. Synthesized compounds listed in table I were charcterized by 1H-NMR spectroscopy. 1H-NMR differentiated clearly between 1-, 2- and 3- substituted

4-benzoyl-1,2,3-triazole derivatives. In fact, it should be pointed out that there is a difference in the chemical shift values among the protons in the 5- position of the benzoyltriazole ring in the series of 1-, 2- and the proton in 4-position of the benzoyltriazole ring of the 3-substituted compounds. The triazole proton of the 1isomer appears always as a singlet at lower eld with respect to the position of the same proton of analogues 2substituted, whereas, that of the 3-isomer resonated as a singlet at higher eld with respect to the corresponding 1and 2-isomers. This evidence is in accordance with to literature data [11, 12]. All compounds gave satisfactory analyses (C, H, N, Cl, F). 3. Biological activity The compounds reported in table I (14) were tested for in vitro affinity on serotonin 5-HT1A, 5-HT2A and 5-HT2C receptors by radioligand binding assays. The more active compounds on serotonin receptors have been selected and evaluated for their affinity on dopaminergic (D-1 and D-2) and adrenergic (1 and 2) receptors. All the compounds were used as hydrochloride salts and were water-soluble. The following specic radioligands and tissue sources were used: (a) serotonin 5-HT1A receptors, [3H]-8-OH-DPAT, rat brain cortex membranes; (b) serotonin 5-HT2A receptors, [3H] ketanserin, rat brain cortex membranes; (c) serotonin 5-HT2C receptors, [3H] mesulergine, rat brain cortex membranes; (d) dopamine D-1 receptors [3H]SCH-23390, rat strial membranes; (e)

722
Table II. Binding affinities and selectivities

IC50 nM ( SEM) Compound X n 5-HT1A [3H] 8-OH-DPAT 1-substituted 1a 1b 1c 1d 1e 1f 1g 1h 2-substituted 2a 2b 2c 2d 2e 2f 2g 2h 1-substituted 3a 3b 3c 3d 3e 3f 3g 3h 2-substituted 4a 4b 4c 4d 4e 4f 4g 4h 8-OH-DPAT ketanserin cinanserin mesulergine H o-Cl m-Cl p-Cl o-OCH3 p-OCH3 o-F p-F H o-Cl m-Cl p-Cl o-OCH3 p-OCH3 o-F p-F H o-Cl m-Cl p-Cl o-OCH3 p-OCH3 o-F p-F H o-Cl m-Cl p-Cl o-OCH3 p-OCH3 o-F p-F 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 200 300 24 2 240 20 180 16 480 35 > 105 10 000 1 200 > 105 73 5.8 36 3 2 0.2 16 000 1 400 62 7 > 105 4 800 390 13 000 1 200 20 1.8 660 60 7.2 0.8 200 15 2.6 0.2 > 105 5 200 480 11 000 1 000 91 6.8 180 20 710 65 10 000 1 150 93 6.8 > 105 95 7.2 1 500 160 2.1 0.1 5-HT2A [3H] ketanserin 24 000 1 900 > 105 30 000 2 850 9 600 810 46 000 4 400 > 105 1 100 170 89 6 > 105 > 105 > 105 > 105 > 105 > 105 240 20 1 000 120 1 300 140 2 000 360 770 67 3 000 250 8 100 700 > 105 7 800 670 6 800 620 22 000 2 100 34 000 3 100 8 200 740 12 000 1 100 > 105 > 105 11 1.2 800 74 1.7 0.3 1.6 1.0 5-HT2C [3H] mesulergine 2 900 350 37 6 960 86 480 34 180 20 > 105 610 55 > 105 1 100 170 12 000 1 100 1 900 290 250 23 3 000 280 > 105 8 800 780 580 57 17 2.0 120 19 230 30 310 30 980 72 > 105 > 105 29 000 2 750 9 500 880 930 72 17 2 1 100 120 2 200 150 > 105 460 40 4 400 380 1.2 0.2

Selectivity vs. 5-HT1A receptor IC50 ratio 5-HT2A 5-HT2C

7.5 > 4 103 125 53 96 1 0.1 > 9 10-4 > 103 > 3 103 > 104 > 6.2 > 103 1 0.08 0.05 65 3.0 107 15 3115 1 1.5 0.6 242 189 11.5 1.2 > 103 1 0.1 0.5

0.9 1.5 4.0 2.7 0.4 1 0.06 1 15 333 950 0.02 48.4 1 1.8 0.04 0.8 0.2 32 1.5 377 1 > 19 3 104 5.2 0.02 0.1 24 1 5 3

dopamine D-2 receptors [3H]spiroperidol, rat strial membranes; (f) 1 adrenergic receptors [3H]prazosin, rat brain cortex membranes; (g) 2 adrenergic receptors [3H]yohimbine, rat brain cortex membranes. Concentrations

required to inhibit 50% of radioligand specic binding (IC50) were determined through 24 independent experiments with samples in triplicate using 79 different concentrations of the title compounds. Specic binding,

723
Table III. Binding affinities and selectivities for D-1, D-2, 1 and 2 receptors in compounds 1h, 2c, 3c, 3e and 4c. IC50 nM ( SEM) Compound X n D-1 [3H] SCH-23390 1h 2c 3c 3e 4c spiroperidol prazosin yohimbine p-F m-Cl m-Cl o-OCH3 m-Cl 2 2 3 3 3 > 105 > 105 11 000 1 100 7 300 620 1 900 15 3.6 0.5 D-2 [3H] spiroperidol > 105 > 105 10 000 1 500 1 500 210 1 800 270 4.6 0.7 1 [3H] prazosin 7 200 610 > 105 > 105 1 700 250 120 15 1.3 0.7 2 [3H] yohimbine 24 000 1 800 24 000 1 700 220 20 43 7 260 28 22 2

dened as described in the experimental section, represented more than 75% of total binding in all three assays. The obtained IC50 values are listed in tables II and III. 4. Results and discussion In this work we examined affinity and selectivity of 4 series of 4-benzoyl-1,2,3-triazole derivatives (14) towards 5-HT1A, 5-HT2A and 5-HT2C receptors by a preliminary binding screen. As mentioned, these compounds can be regarded as open chain analogues of previously reported benzotriazole derivatives I [8]. The IC50 nM values listed in table II reveal that several compounds exhibit nanomolar affinity for the 5-HT1A receptor, whereas generally only a few bind to 5-HT2A and 5-HT2C receptors with comparable affinity. As far as 5-HT1A receptor affinity is concerned, the results reported in table II indicate that the structures endowed with highest 5-HT1A affinity are 1b, 2ac, 2e, 3a, 3c and 3e. With respect to the length of the alkyl chain between the arylpiperazine moiety and the terminal 4-benzoyl-1,2,3-triazole nucleus, the spacer effect is different for 1- or 2-substituted compounds. For the analogues having an ethyl [(CH2)2] connecting group between the triazole and piperazine rings, several 2-substituted compounds have lower IC50 values than the corresponding 1-substituted derivatives. For example, compare 2a vs. 1a (IC50 nM = 73 and 3 200, respectively); 2c vs. 1c (IC50 nM = 2 and 240, respectively) and 2e vs. 1e (IC50 nM = 62 and 480, respectively). The only partial exceptions are compounds 2b and 2d whose affinities are in fact lower than those of 1b and 1d respectively. By contrast, for the analogues having a propyl connecting group [(CH2)3] between the triazole and the piperazine rings, several 1-substituted compounds have lower IC50 values than the corresponding 2-substituted

derivatives. For example, compare 3a vs. 4a (IC50 nM = 20 and 91, respectively); 3c vs. 4c (IC50 nM = 7.2 and 710, respectively); 3e vs. 4e (IC50 nM = 2.6 and 93, respectively). Only ortho chloro (3b) and ortho and para uoro substituted compounds (3g and 3h) exhibited IC50 values higher than 2-substituted analogues. It seems that in these series the alkyl chain is not interacting directly with the receptor on the basis of its hydrophobicity but is more likely acting as a spacer. The nature and the position of the substituent on the phenyl ring also has a signicant inuence on affinity. For all series, p-substitution (by Cl, F, or OCH3) of the piperazinyl-aromatic ring led to compounds with supra nM affinity for the 5-HT1A receptor (but note, compound 1a with no substituent did not bind with high affinity). For compounds of structure 2 and 3, m-Cl substituents afforded analogues with high 5-HT1A receptor affinity, whereas, in the benzotriazole series it had little effect on affinity for the same receptor [8]. An o-OCH3 substituent also appeared to support 5-HT1A receptor binding for these compounds. Lpez-Rodrguez et al. [13] reported that the o-methoxy group on the phenyl ring in various hydantoin-phenylpiperazine homologues exerted effects on 5-HT1A affinity strongly related to the overall ligand structure. Taken together, these ndings suggest that there are different possible orientations within the 5-HT1A binding site accessible to the X substituent on the phenyl ring. Such orientations are dictated by the global conformational arrangement of the bound ligand. As a result, contribution of X to affinity depends on the properties of X as well as on the nature of the receptor subsite into which such a substituent is docked. Among compounds evaluated in table II, 2c (m-Cl) and 3e (o-OCH3) had the highest binding affinity and show the most favourable affinity-selectivity proles for the 5-HT1A receptor (5-HT2A/5-HT1A IC50 ratios are about

724 10 000 and 3 000, respectively, while 5-HT2C/5-HT1A IC50 ratios are about 1 000 and 300, respectively). While uorine substitution generally led to compounds with low binding affinity to all 5-HT receptors, paradoxically, compound 4g (o-uoro) had IC50 values of 95 nM and 11 nM in the 5-HT1A and 5-HT2A binding assays respectively. The other two highly potent ligands of the 5-HT1A receptor, 1b and 3c, are moderately selective as they retain appreciable affinities for 5-HT2C receptors. A comparison between affinity proles of benzoyltriazole derivatives 2c and 3e versus those of the above mentioned benzotriazole counterparts [8] (IC50 ratios always less than 50) clearly show that the 5-HT1A/5-HT2A and 5-HT1A/5-HT2C selectivity ratios of the former compounds are signicantly superior. This result might be partly ascribed to poor steric and/or electronic complementarity between the benzoyltriazole residue and the 5-HT2A and 5-HT2C binding sites compared to the benzotriazole moiety in compounds I. The results of this study indicate that the affinity for the 5-HT2A receptor is usually lower than the affinity for 5-HT1A receptors. The introduction of the electronwithdrawing substituents in the ortho and para position, such as the uoro group, increased the binding affinity for the 5-HT2A receptor in all series except 3. Only 4g exhibits nanomolar affinity (IC50 = 11 nM), but this compound binds also to the 5-HT1A receptors with appreciable potency. A remarkable selectivity (1 000fold) for 5-HT2A versus 5-HT1A and 5-HT2C receptors was achieved with 1h, although the affinity of this compound for the 5-HT2A receptor fell within the submicromolar range (IC50 = 89 nM). As far as the 5-HT2C receptor is concerned, nanomolar potency was observed for compounds 1b, 3a and 4c (IC50s < 50 nM). Binding of the former two ligands to the 5-HT1A receptor was, however, slightly stronger with respect to the 5-HT2C receptor. The compound 4c was the only one showing appreciable selectivity towards the 5-HT2C receptor (5-HT1A/5-HT2C and 5-HT2A/5-HT2C IC50 ratios were about 40 and 500, respectively). As mentioned in the introduction section, in order to evaluate potential therapeutic uses, it is crucial to check whether high potency is anked by undesirable affinity for other receptors, e.g., adrenergic receptors [14, 15]. Thus the most active compounds on 5-HT1A (2c, 3c, and 3e), 5-HT2A (1h) and 5-HT2C (4c) were further evaluated for their affinity at dopaminergic and adrenergic receptors. Results are summarized in table III. As far as the dopaminergic system was concerned, the D-1 and D-2 receptor affinity consistently showed IC50 values above 106 M. As regards the affinity for 1 adrenergic receptors, the IC50 values were high for all selected compounds, with the exception of 4c (IC50 = 120 nM). The affinity for 2 receptors were in some cases quite considerable (3c, 3e and 4c). In particular, o-OCH3-Ph derivative 3e showed affinity for the 2 receptor which was only one order of magnitude lower than the affinity toward the 5-HT1A receptor. However, compound 2c, the most potent 5-HT1A ligand showed the best selectivity prole versus all considered receptors. In conclusion, some of the newly investigated benzoyltriazole derivatives were endowed with high affinity and selectivity for the 5-HT1A receptor. Particularly, the IC50 values of 2c and 3e measured on this receptor are 2 and 2.6 nM, and in terms of selectivity, these two ligands represent a substantial improvement over previously described benzotriazole analogues [8]. These observations suggest that both the arylpiperazine pharmacophore and the terminal benzoyltriazole system contribute to the 5-HT1A interaction of the compounds. Further synthesis and biological in vivo evaluation of derivatives with new substituents are currently in progress, and the results will be reported in due course. 5. Experimental protocols 5.1. Chemistry Melting points were determined using a Koer hotstage apparatus and are uncorrected. Kieselgel 60 was used for column chromatography and kieselgel 60 F254 plates from Merck were used for TLC. Where analyses are indicated only by the symbols of the elements, results obtained are within 0.4% of the theoretical values. The purity of compounds were carefully assessed using analytical TLC and the structure veried spectroscopically by proton NMR spectra recorded on a Bruker AMX-500 instrument. Chemical shifts in ppm are referenced to the DMSO signal at 2.50 ppm. Elemental analyses (C, H, N, Cl, F) were determined within 0.4% of the theoretical values. The following compounds were synthesized by published procedures: aromatically substituted 1-(2-chloroethyl)-4-phenylpiperazine and aromatically substituted 1-(3-chloropropyl)-4-phenylpiperazine derivatives [8]. 5.1.1. Phenyl ethynyl ketone (6) A solution of chromium trioxide (0.10 mol) in water (30 mL) and concentrated sulphuric acid (8.5 mL) was slowly added to a stirred and cooled solution of propylethynylcarbinol 5 (0.15 mol) in acetone (50 mL). The reaction mixture was carried out at 4 C in an atmosphere

725 of nitrogen. After stirring for 7 h, water was added to dissolve the precipitated chromium salts and the product was extracted with chloroform. Evaporation of the organic solution gave a yellow solid which was recrystallized from aqueous methanol to give 16.6 g (85%) of 6 as pale yellow needles. The physical data are in agreement with those given in reference [9]. 5.1.2. 4-benzoyl-1H-1,2,3-triazole (7) To a stirred and heated solution of NaN3 (0.10 mol) in anhydrous dimethylacetamide (80 mL) was slowly added phenyl ethynyl ketone 6 (0.10 mol) dissolved in dimethylacetamide, anhydrous (80 mL). The reaction mixture was kept at 100 C for 2 h. After stirring for a further 12 h at room temperature, evaporation of the solvent under reduced pressure gave a liquid residue which was diluted with water. The aqueous layer was acidied to pH = 5 with 10% HCl and extracted with ether (3 200 mL). The combined organic layers were dried over anhydrous Na2SO4. After evaporation of the solvent, the solid residue was puried by crystallization from ethyl alcohol to give 16.4 g (95%) of 7. The physical data are in agreement with those given in reference [10]. 5.1.3. General procedure for the preparation of 1- and 2-[2-[4-(X-Phenyl)piperazinyl]ethyl](4-benzoyl)-1,2,3-triazole (1ah and 2ah) To a stirred solution of 4-benzoyl-1,2,3-triazole (0.1 mol) and appropriate aromatically substituted 1-(2chloroethyl)-4-phenylpiperazines (0.15 mol) in 80 mL of butan-2-one was added anhydrous K2CO3 (0.3 mol). The mixture was reuxed for 20 h and the solvent was removed under reduced pressure. The residue was diluted with H2O and extracted with CHCl3 (3 200 mL). The combined organic layers were washed with water and dried over Na2SO4. Evaporation of the solvent in vacuo provided the crude residue consisting of the expected triazole isomers. The mixture of isomers was separated and puried using a silica-gel column chromatography (diethyl ether/n-hexane 95:5 as eluent) to give 1-isomers 1 and 2-isomers 2 in the relative yields provided in table I. Spectral data of the title compounds refer to the hydrochloride salts. 5.1.3.1. 1-[2-[4-(Phenyl)piperazinyl]ethyl](4-benzoyl)1,2,3-triazoleHCL (1a) M.p. 201202 C; 1H NMR (DMSO-d6) 3.34 (m, 4H, 2CH2-pip), 3.71 (t, J = 7.5 Hz, 2H, CH2N-pip), 3.94 (m, 4H, CH2-pip), 5.22 (t, J = 7.5 Hz, 2H, CH2N-triaz), 6.978.35 (mm, 10H, ArH), 9.17 (s, 1H, H-triaz). 5.1.3.2. 2-[2-[4-(Phenyl)piperazinyl]ethyl](4-benzoyl)1,2,3-triazoleHCL (2a) M.p. 206207 C; 1H NMR (DMSO-d6) 3.34 (m, 4H, 2CH2-pip), 3.71 (t, J = 7.5 Hz, 2H, CH2N-pip), 3.96 (m, 4H, 2-CH2-pip), 5.32 (t, J = 7.5 Hz, 2H, CH2N-triaz), 6.978.30 (mm, 10H, ArH), 8.60 (s, 1H, H-triaz). 5.1.4. General procedure for the preparation of 1- and 2-[3-[4-(X-Phenyl)piperazinyl]propyl](4-benzoyl)-1,2,3triazole (3ah and 4ah) These compounds were prepared in a manner similar to that mentioned above, starting from 4-benzoyl-1,2,3triazole (0.1 mol) and aromatically substituted 1-(3chloropropyl)-4-phenyl-piperazines (0.1 mol). The crude mixture of isomers was chromatographed on a silica-gel column (diethyl ether/n-hexane 95:5 as eluent) to yield pure 3 and 4 in the relative yields shown in table I. Spectral data of title compounds refer to the hydrochloride salts. 5.1.4.1. 1-[3-[4-(Phenyl)piperazinyl]propyl](4-benzoyl)1,2,3-triazoleHCL (3a) M.p. 183184 C; 1H NMR (DMSO-d6) 2.58 (m, 2H, CH2CH2CH2), 3.27 (m, 4H, 2CH2-pip), 3.32 (t, J = 7.5 Hz, 2H, CH2N-pip), 3.67 (m, 4H, 2CH2-pip), 4.76 (t, J = 7.5 Hz, 2H, CH2N-triaz), 6.958.37 (mm, 10H, ArH), 9.31 (s, 1H, H-triaz). 5.1.4.2. 2-[3-[4-(Phenyl)piperazinyl]propyl](4-benzoyl)1,2,3-triazoleHCL (4a) M.p. 251252 C; 1H NMR (DMSO-d6) 2.60 (m, 2H, CH2CH2CH2), 3.26 (m, 4H, 2CH2-pip), 3.45 (t, J = 7.5 Hz, 2H, CH2N-pip), 3.673.88 (m, 4H, 2CH2-pip), 4.81 (t, J = 7.5 Hz, 2H, CH2N-triaz), 6.848.28 (mm, 10H, ArH), 8.56 (s, 1H, H-triaz). 5.1.5. Hydrochloride salts: general procedure The hydrochloride salts were prepared by adding an HCl ethereal solution to an ethanolic solution of free bases. Recrystallization solvent, formulae and melting points are reported in table I. They were obtained as white crystals. 5.2. Biological methods 5.2.1. 5-HT1A binding assay Radioligand binding assays were performed following a published procedure [16]. Cerebral cortex from male Sprague-Dawley rats (180220 g) was homogenized in 20 volumes of ice-cold Tris-HCl buffer (50 mM, pH 7.7 at 22 C) with a Brinkmann Polytron (setting 5 for 15 s), and the homogenate was centrifuged at 50 000 g for 10 min. The resulting pellet was then resuspended in the

726 same buffer, incubated for 10 min at 37 C, and centrifuged at 50 000 g for 10 min. The nal pellet was resuspended in 80 volumes of the Tris-HCl buffer containing 10 M pargyline, 4 mM CaCl2, and 0.1% ascorbate. To each assay tube was added the following: 0.1 mL of the drug dilution (0.1 mL of distilled water if no competing drug was added), 0.1 mL of [3H]-8-hydroxy2-(di-n-propylamino)tetralin ([3H]-8-OH-DPAT) in buffer (containing Tris, CaCl2, pargyline, and ascorbate) to achieve a nal assay concentration of 0.1 nM, and 0.8 mL of resuspended membranes. The tubes were incubated for 30 min at 37 C, and the incubations were terminated by vacuum ltration through Whatman GF/B lters. The lters were washed twice with 5 mL of ice-cold Tris-HCl buffer, and the radioactivity bound to the lters was measured by liquid scintillation spectrometry. Specic [3H]-8-OH-DPAT binding was dened as the difference between binding in the absence and presence of 5-HT (10 M). 5.2.2. 5-HT2A and 5-HT2C binding assays Radioligand binding assays were performed as previously reported by Herndon et al. [17]. Briey, frontal cortical regions of male Sprague-Dawley rats (200250 g, Charles River) were dissected on ice and homogenized (1:10 w/v) in ice-cold buffer solution (50 mM Tris HCl, 0.5 mM EDTA, and 10 mM MgCl2 at pH 7.4) and centrifuged at 3 000 g for 15 min. The pellet was resuspended in buffer (1:30 w/v), incubated at 37 C for 15 min and then centrifuged twice more at 3 000 g for 10 min (with resuspension between centrifugations). The nal pellet was resuspended in buffer that also contained 0.1% ascorbate and 105 M pargyline. Assays were performed in triplicate in a 2.0 mL volume containing 5 mg wet weight of tissue and 0.4 nM [3H] ketanserin (hyphen) (76 Ci/mmol; New England Nuclear) for 5-HT2A receptor assays, and 10 mg wet weight of tissue and 1 nM [3H]mesulergine (75.8 Ci/mmol; Amersham) for 5-HT2C receptor assays. Cinanserin (1.0 M) was used to dene nonspecic binding in the 5-HT2A assay. In the 5-HT2C assays, mianserin (1.0 M) was used to dene nonspecic binding, and 100 nM spiperone was added to all tubes to block binding to 5-HT2A receptors. Tubes were incubated for 15 min at 37 C, ltered on Schliecher and Schuell (Keene, NH) glass bre lters presoaked in poly(ethylene imine), and washed with 10 mL of ice-cold buffer. Filters were counted at an efficiency of 50%. 5.2.3. D-1 dopaminergic binding assay The binding assay for D-1 dopaminergic receptors was that described by Billard et al. [18]. Corpora striata were homogenized in 30 vol. (w/v) ice cold 50 mM Tris-HCl buffer (pH 7.7 at 25 C) using a Polytron PT10 (setting 5 for 20 s). Homogenates were centrifuged twice for 10 min at 50 000 g with resuspension of the pellet in fresh buffer. The nal pellet was resuspended in 50 mM ice cold Tris-HCl containing 120 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 0.1% ascorbic acid and 10 M pargyline (pH 7.1 at 37 C). Each assay tube contained 50 L [3H]SCH-23390 to achieve a nal concentration of 0.4 nM, and 900 L resuspended membranes (3 mg fresh tissue). The tubes were incubated for 15 min at 37 C and the incubation was terminated by rapid ltration under vacuum through Whatman GF/B glass bre lters. The lters were washed three times with 5 mL ice-cold 50 mM Tris-HCl buffer (pH 7.7 at 25 C). The radioactivity bound to the lters was measured by a liquid scintillation counter. Specic [3H]SCH-23390 binding was dened as the difference between binding in the absence or in the presence of 0.1 M piutixol. 5.2.4. D-2 dopaminergic binding assay The procedure used in the radioligand binding assay was reported in detail by Creese et al. [19]. Corpora striata were homogenized in 30 vol. (w/v) ice cold 50 mM Tris-HCl buffer (pH 7.7 at 25 C) using a Polytron PT10 (setting 5 for 20 s). Homogenates were centrifuged twice for 10 min at 50 000 g with resuspension of the pellet in fresh buffer. The nal pellet was resuspended in 50 mM ice cold Tris-HCl containing 120 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 0.1% ascorbic acid and 10 M pargyline (pH 7.1 at 37 C). Each assay tube contained 50 L [3H]spiroperidol to achieve a nal concentration of 0.4 nM, and 900 L resuspended membranes (3 mg fresh tissue). The tubes were incubated for 15 min at 37 C and the incubation was terminated by rapid ltration under vacuum through Whatman GF/B glass bre lters. The lters were washed three times with 5 mL ice-cold 50 mM Tris-HCl buffer (pH 7.7 at 25 C). The radioactivity bound to the lters was measured by a liquid scintillation counter. Specic [3H]spiroperidol binding was dened as the difference between binding in the absence or in the presence of 1 M (+)-butaclamol. 5.2.5. 1 adrenergic binding assay The procedure used in the radioligand binding assay has been reported in detail by Greengrass and Bremner [20]. Brain cortex was homogenized in 30 vol. (w/v) ice-cold 50 mM Tris-HCl buffer, (pH 7.2 at 25 C) using a Polytron PT10 (setting 5 for 20 s). Homogenates were centrifuged twice for 10 min at 50 000 g with resuspension of the pellet in fresh buffer. The nal pellet was

727 resuspended in 50mM ice-cold Tris-HCl, (pH 7.4 at 25 C). Each assay tube contained 50 L drug solution, 50 L [3H]prazosin to achieve a nal concentration of 0.4 nM, and 900 L resuspended membranes (10 mg fresh tissue). The tubes were incubated for 30 min at 25 C and the incubation was terminated by rapid ltration under vacuum through Whatman GF/B glass bre lters. The lters were washed three times with 5 mL ice-cold 50 mM Tris-HCl, buffer (pH 7.2 at 25 C). The radioactivity bound to the lters was measured by a liquid scintillation counter. Specic [3H]prazosin binding was dened as the difference between binding in the absence or in the presence of 10 M phentolamine. 5.2.6. 2 adrenergic binding assay The procedure used in the radioligand binding assay was reported in detail by Perry and UPrichard [21]. Brain cortex was homogenized in 30 vol. (w/v) ice-cold 5mM tris-HCl, 5mM EDTA buffer (pH 7.3 at 25 C) using a polytron PT10 (setting 5 for 20 s). Homogenates were centrifuged three times for 10 min at 50 000 g with resuspension of the pellet in fresh buffer. The nal pellet was resuspended in 50 mM ice-cold Tris-HCl, 0.5 mM EDTA (pH 7.5 at 25 C). Each assay tube contained 50 L drug solution, 50 L [3H]yohimbine to achieve a nal concentration of 1 nM, and 900 L resuspended membranes (10 mg fresh tissue). The tubes were incubated for 30 min at 25 C and the incubation was terminated by rapid ltration under vacuum through Whatman GF/B glass bre lters. The lters were washed three times with 5 mL ice-cold 50 mM Tris-HCl, 0.5 mM EDTA buffer (pH 7.5 at 25 C). The radioactivity bound to the lters was measured by a liquid scintillation counter. Specic [3H]yohimbine binding was dened as the difference between binding in the absence or in the presence of 10 M phentolamine. Acknowledgements This work was supported by a grant from Regione Campania ai sensi della L.R. 31 December 1994. The NMR spectral data were provided by Centro di Ricerca Interdipartimentale di Analisi Strumentale, Universit degli Studi di Napoli Federico II. The assistance of the staff is gratefully acknowledged.

References
[1] [2] [3] [4] [5] [6] [7] [8] Baumgarten H.G., Gothert M. (Eds.), Serotoninergic Neuron and 5-HT Receptors in the CNS, Springer-Verlag, Berlin, 1997. van Wijngaarden I., Soudin W., Serotonin Receptors and Their Ligands, Elsevier, Amsterdam, 1997. Hamon M., Gozlan H., El Mestikawy S., Emerit M.B., Boslanos F., Schecther L., Am. N.Y. Acad. Sci. 600 (1990) 114131. Eison A.S., Eison M.S., Prog. Neuropsychopharmacol. Biol. Psychiatry 18 (1994) 4762. Eison M.S., Psychopathology 1 (1989) 1320. Glennon R.A., Drug Dev. Res. 26 (1992) 251274. van Steen B.J., van Wijngaarden I., Tulp M.T.M., Soudijn W., J. Med. Chem. 36 (1993) 27512760. Caliendo G., Greco G., Grieco P., Novellino E., Perissutti E., Santagada V., Barbarulo D., De Blasi A., Eur. J. Med. Chem. 31 (1996) 207213. Kenneth B., Heilbron I.M., Jones E.R.H., Weedon B.C.L., J. Chem. Soc. Part I (1946) 3945. Nesmeyanov A.N., Rybinskaya M.I., Dokl. Akad. Nauk. SSSR. 158 (1964) 408410. Livi O., Biagi G., Ferretti M., Lucacchini A., Barili P.L., Eur. J. Med. Chem. -Chim. Ter. 18 (1983) 471475. Biagi G., Ferretti M., Livi O., Scartoni V., Lucacchini A., Mazzoni M., Il Farmaco Ed. Sc. 41 (1986) 388400. Lpez-Rodrguez M.L., Rosado M.L., Benham B., Morcillo M.J., Ferndez E., Schaper K.J., J. Med. Chem. 40 (1997) 16481656. Lpez-Rodrguez M.L., Rosado M.L., Benham B., Morcillo M.J., Sanz A.M., Orensanz L., Beneitez M.E., Fuentes J.A., Manzanares J., J. Med. Chem. 39 (1996) 44394450. Perrone R., Berardi F., Leopoldo M., Tortorella V., J. Med. Chem. 39 (1996) 31953202. Schlegel J.R., Peroutka S.J., Biochem. Pharmacol. 35 (1986) 19431949. Herndon J.L., Ismaiel A., Ingher S.P., Teitler M., Glennon R.A., J. Med. Chem. 35 (1992) 49034910. Billard W., Ruperto V., Grosby G., Iorio L.C., Barnett A., Life Sci. 35 (1985) 18851893. Creese I., Schneider R., Snyder S.H., Eur. J. Pharmacol. 46 (1977) 377381. Greengrass P., Brenner R., Eur. J. Pharmacol. 55 (1979) 323326. Perry B.D., UPrichard D.C., Eur. J. Pharmacol. 76 (1981) 461464.

[9] [10] [11] [12] [13] [14]

[15] [16] [17] [18] [19] [20] [21]

Eur. J. Med. Chem. 34 (1999) 729744 1999 Editions scientiques et mdicales Elevier SAS. All rights reserved

729

Original article

Synthesis and QSAR of substituted 3-hydroxyanthranilic acid derivatives as inhibitors of 3-hydroxyanthranilic acid dioxygenase (3-HAO)
Mats Linderberga, Sven Hellberga, Susanna Bjrka, Birgitta Gotthammara, Thomas Hgberga, Kerstin Perssona, Robert Schwarczc, Johan Luthmanb, Rolf Johanssona*
Medicinal Chemistry, Preclinical R& D, Astra Arcus, S-15185 Sdertlje, Sweden b Cell Biology, Preclinical R& D, Astra Arcus, S-15185 Sdertlje, Sweden c Maryland Psychiatric Research Center, University of Maryland School of Medicine, Baltimore, MD 21228, USA (Received 21 January 1999; revised 12 March 1999; accepted 18 March 1999)
a

Abstract Novel 4,5-, 4,6-disubstituted and 4,5,6-trisubstituted 3-hydroxyanthranilic acid derivatives were synthesized and their ability to reduce the production of the excitotoxin quinolinic acid (QUIN) by inhibition of brain 3-hydroxyanthranilic acid dioxygenase (3-HAO) was subsequently investigated. The potency of the compounds to inhibit 3-HAO was assayed in rat brain homogenate, while chemical stability of certain compounds was studied by HPLC. The data were used to generate quantitative structure-activity relationship (QSAR) models for potency of 3-HAO inhibition and compound stability. Compounds with longer half-lives were obtained when the difference between the HOMO and LUMO was increased, while electron withdrawing groups in the 4- and 5-positions increased the potency of 3-HAO inhibition. Selected compounds that showed high potency in vitro were also found to be efficacious inhibitors in vivo after cerebral administration in rats. 1999 Editions scientiques et mdicales Elevier SAS 3-HAO / QSAR / hydroxyanthranilic acid / enzyme inhibitors / neuroprotection

1. Introduction The cytosolic, non-haem ferrous (Fe2+) enzyme 3-hydroxyanthranilic acid dioxygenase (3-HAO; EC 1.13.11.6) plays an important role in the metabolic transformation of L-tryptophan to nicotinamide in the kynurenine pathway (gure 1). The enzyme oxidizes and ring opens 3-hydroxyanthranilic acid (3-HANA) to produce -amino--carboxymuconic acid -semialdehyde, which subsequently, spontaneously cyclizes and forms quinolinic acid (QUIN) [1]. 3-HAO seems mainly to be localized to hepatic tissue, but it has also been demonstrated that the enzyme is expressed both in the brain [2, 3] and in inammatory cells [2, 4] where the production of QUIN has been shown to be stimulated by certain cytokines [59]. Biochemical and immunological analysis in the rat suggest that the brain and liver 3-HAO are identical proteins [10]. Moreover, there appears to be a
*Correspondence and reprints

high degree of homology (94% similarity) between the rat and the human 3-HAO amino acid sequence [11]. QUIN is an NMDA receptor agonist and an excitotoxin [12] that has been reported to cause convulsions [13] or neurodegeneration [14] after intracerebral administration. It has been shown that intrastriatal injections of QUIN in rats induce biochemical and morphological alterations similar to those observed in Huntingtons disease [15, 16]. Furthermore, increases in neuronal Huntingtin immunoreactivity [17] have recently been reported to occur after administration of QUIN into the striatum of mice, providing further support for a role of QUIN in the pathogenesis of Huntingtons disease. It is also possible that QUIN is involved in epilepsy. For example, the levels of QUIN are higher in epilepsy-prone mice in a transgenic model of epilepsy than in the corresponding wild type mice [18]. Moreover, increased QUIN levels, or enhanced kynurenine pathway activity, have been implicated in inammatory diseases of the central nervous system. For instance, QUIN has been

730

Figure 1. The biotransformation pathway for tryptophan yielding the excitotoxin QUIN.

associated with neurological dysfunction occurring following various viral or bacterial infections [1923]. To elucidate the possible role of 3-HAO-mediated formation of QUIN in different neurological conditions, inhibitors of the enzyme are valuable tools, and may also nd therapeutic utility. In this report, the synthesis of several analogues of 3-HANA and their structure-activity relationship as inhibitors of 3-HAO are presented. The main objective was to develop inhibitors with improved chemical and pharmacological properties compared to previously described 4-halogenated 3-HANA analogues [24]. A particular problem associated with 3-HANA analogues is the propensity of 3-HANA for auto-oxidation leading to its degradation and the formation of oxidized cinnabarinic acid [25, 26]. Hence, the stability of some of the synthesized compounds was studied by HPLC. The potency of the compounds to inhibit 3-HAO was assessed in rat brain tissue homogenates, while the ability of selected compounds to inhibit 3-HAO in vivo was also studied following intracerebral administration. 2. Chemistry A series of 4,5-, 4,6-disubstituted and 4,5,6trisubstituted 3-hydroxyanthranilic acids were prepared. The 4,5-dihalo substituted compounds were synthesized from the known 4-chloro- and 4-bromo-3-hydroxyanthranilic acids. The strong directing effect of the amino substituent gives access to derivatives 11 and 12 after reaction with bromine and chlorine, respectively. When

the 5-position is substituted, the 4-substituted halogens can be obtained by a regio-selective reaction at the 4-position. Accordingly, compounds 8 and 10 were obtained by direct bromination of the 5-substituted-2hydroxyanthranilic acids or esters (gures 2 and 3). To conrm the identity of the structures, 5-bromo-3hydroxyanthranilic acid was chlorinated and shown to give an identical product as formed by bromination of 4-chloro-3-hydroxyanthranilic acid. The identity of the 4,5-dichloro and 4,5-dibromo derivatives was conrmed by comparison with the corresponding 4,6-disubstituted compounds and by synthesis either from the 4- or 5-halogenated 3-HANA leading to identical products (gure 2). Compound 39 was synthesized by bromination of 5-hydroxy-2-methoxybenzoic acid, which after nitration and reduction of the nitro group gave the desired product

Figure 2. i) Br2/HBr/HOAc, Cl2/HCl/HOAc.

ii)

Br2/HBr/HOAc

or

731

Figure 3. i) Br2/HOAc, ii) KOH/EtOH.

Figure 6. i) EtSNa/DMF, ii) HCl/MeOH, iii) HNO3/CH3NO2/ CH2Cl2, iv) KOH/EtOH, v) SnCl2xH2O/EtOH.

(gure 4). Compounds 6, 34 and 44 (gures 5 and 6) were prepared by nitration of suitably substituted mono or dihalogenated benzoic acid. The benzoic acids 32 and 42 were obtained by a palladium catalysed carbonylation of the corresponding triates. The directing effect of the phenol gave the desired regio-selectivity for compounds 6 and 44. The more easily oxidized compounds 17, 25 and 36 (gure 7 and 8) were obtained after halogenation of an intermediate without the strongly directing and activating amino-function. To make the tri-substituted compounds 21 and 29 and the 4,6-disubstituted 50 and 54 the carboxylic acid was introduced via ring closure of the corresponding isonitrosoacetanilid followed by oxidative ring-opening of the isatin-derivative (gure 9).

Figure 7. i) Cl2/CHCl3, ii) BnBr/K2CO3, iii) Cu(I)Cl/KBH4.

Figure 8. i) Br2/HOAc/NaOAc, ii) H2/PtS2.

Figure 4. i) Br2/HOAc, ii) NaNO3/La(NO3)3/HCl(aq.)/diethyl ether, iii) H2/PtS2/EtOH.

Figure 5. i) Pd(OAc)2/CO/MeOH/dppp, ii) KOH/MeOH, iii) BBr3 or HBr, iv) HNO3/CH3NO2, v) Pd/H2.

Figure 9. i) CCl3CH(OH)2/NH2OH/DMF/H2O, ii) H2SO4/Dioxane or polyphosphoric acid, iii) NaOH/H2O2, iv) H2/Pd or BBr3.

732
Table I. Inhibition of 3-HAO and chemical stability of 3-HANA derivatives.

Compound 55 [24, 32] 56 [24] 57 [24] 6 8 10 11 12 36 39 44 50 21 29 34 54

a

R4 Cl Br F Br Br Br Cl Cl Br Br Cl Cl Cl Cl Cl Cl

R5 H H H OEt Br Me Br Cl H H H H (CH2)2 Me Me H

R6 H H H H H H H H Br MeO Cl Ph (CH2)2 Me Cl Me

IC50a nM 6 2 24 56 0.3 2.3 0.26 0.3 5.8 120 10.1 11 000 200 8.2 4.4 7.8

Stabilityb 38

21

90 0 89 79 0c 62 53

In vitro inhibition of 3-HAO in rat cortex. bPercent remaining after 24 h in a PBS buffer at pH 7.5 and 37 C. cPercent remaining after 24 h in a PBS buffer at pH 7.5 and 50 C.

3. QSAR A quantitative structure-activity relationship was developed for the 16 compounds presented in table I. The of the R1, R2 and R3 substituents, the and MR for the R4, R5 and R6 substituents were obtained from the literature [27]. The pKa values for the CO2H, NH2 and OH substituents and logP for the molecules were calculated using the Pallas software [28]. Semi-empirical calculations using AM1 (Spartan) [29] provided values for HOMO and LUMO energies, e-neg, hardness, heat of formation, polarizability, surface area, molecular volume and ovality [30] which were subsequently used as descriptors in the QSAR calculations. 4. Pharmacology The compounds were tested for their ability to inhibit 3-HAO in homogenates of rat brain tissue according to the method of Foster et al. [18]. The production of radioactive QUIN was measured after addition of [1-14C]3-HANA to determine inhibition of the enzyme.

Test compound concentrations resulting in a 50% inhibition of the enzyme (IC50) are reported. Also, the ability of selected compounds to inhibit cerebral 3-HAO in vivo was studied following intracerebroventricular (i.c.v.) administration of the compounds in rats. The de novo formation of QUIN in hippocampal tissue after concomitant i.c.v. administration of the test compounds and QUIN was measured by GC/MS. 5. Results and discussion A PLS analysis# (Codex, AP Scientic Service) of potency resulted in a four component model, accounting for 97% of the variance in pIC50 (cross-validated Q2 = 0.85) (gure 10). All the components in the model were statistically signicant according to cross-validation. The two most important components accounted for 86% of the variance. A plot of the PLS-weights for these components indicated that the most important factors inu#

Descriptors used in the PLS analysis are available on the Internet on request

733

Figure 10. Predicted vs. observed IC50 values for 16 3-HAO inhibitors. Figure 11. Plot of stability versus hardness.

encing the IC50 values are the size of the R6 substituent, the lipophilicity of the R5 substituent and the pKa of the phenol. This implies that small, electron-withdrawing groups at R6 and lipophilic, electron-withdrawing groups at R5 would increase the potency of the compounds. Nine of the compounds were used to develop a QSAR model for chemical stability. The PLS analysis resulted in a two-component model, accounting for 94% of the variance in stability. The single most important factor for the model was the difference between the HOMO and LUMO energies (hardness) of the molecules (gure 11). The very labile compound 39, having the strongly electron-donating 6-methoxy substituent could not be accounted for in the model. This may indicate a different mechanism for the degradation of this compound. The increased instability observed for compounds with a small difference in HOMO and LUMO-orbitals is in agreement with the observation that dimerization has been reported to be involved in the degradation of 4-bromo-3-hydroxyanthranilic acid [25, 26]. Several compounds that were found to be potent in vitro inhibitors of 3-HAO were also efficacious in vivo after intracerebral administration. There was a general agreement between the in vitro and in vivo ability of the 3-HANA analogues to inhibit 3-HAO (tables I and II). Hence, compounds which showed low nM IC50 values were able to almost totally inhibit de novo production of QUIN in vivo when given at doses that were equivalent to or higher than those of 3-HANA, while 21, which showed a high nM IC50 value in vitro, displayed no in vivo activity. On the other hand, certain differences in potency observed in vitro were not reected in vivo. This may, at least partially, depend on the inherent differences in chemical stability of the compounds.

For example, the low stability of the very potent compound 8 may have reduced its in vivo actions, but distribution and local metabolism in vivo could also contribute to the discrepancies seen. Importantly, the nding that potency and stability depend on separate structural features present possibilities to further improve the characteristics of 3-HANA analogues as useful inhibitors of 3-HAO.

Table II. In vivo effects of 3-HAO inhibitors on 3-HANA-induced de novo production of QUIN after i.c.v. administration in rat. Compound 55 8 36 Ratio dosea 3-HANA:compound 1:1 1:1 1:0.3 1:1 1:30 1:30 1:30 1:30 1:0.3 1:30 % Inhibition 96.9 0.1 87.4 2.0 67.2 6.6 81.9 1.4 95.2 1.4 91.9 1.9 0 13 95.1 1.2 69.6 7.9 93.9 0.6

44 21 29 34

3-HANA was given in a dose of 10 nmol i.c.v., while the different compounds were given in doses of either 3, 10 or 300 nmol i.c.v. at the same time. The cerebral levels of QUIN at 2 h after administration were determined in hippocampal tissue by GC/MS. Data are the mean SEM of 57 determinations.

734 6. Experimental protocols 6.1. Chemical methods Melting points were determined on a Bchi SMP-20 apparatus. 1H and 13C NMR spectra were recorded at ambient temperature on a Varian Unity 400 or Varian Gemini 300 instrument. Chemical shifts are given in ppm from internal standards. For 1H NMR and 13C NMR spectra the internal references were tetramethylsilane (0.0 ppm), CDCl3 ( 7.26 or 77.0 ppm), CD3OD ( 3.38 or 49.3 ppm) or DMSO-d6 ( 2.49 or 39.5 ppm), respectively. Coupling constants are given in Hertz, and the splitting patterns are designated as follows: s, singlet; d, doublet; dd, doublet of doublets; t, triplet; q, quartet and br, broad. Mass spectra were obtained on an LKB 2091 (ELI, 70 eV or CI/CH4) or on a Finnigan-MAT TSQ 70 (thermospray) spectrometer. Elemental analyses were performed by MIKRO KEMI AB, Uppsala, Sweden. The values were within 0.4% of theoretical if not otherwise indicated. 6.1.1. Ethyl 4-bromo-3,5-diethoxybenzoate 1 4-Bromo-3,5-dihydroxybenzoic acid [31] (4.68 g, 20 mmol), potassium carbonate (8.28 g) and diethyl sulfate (7.84 mL) in dry acetone (50 mL) under nitrogen atmosphere were stirred at room temperature for 20 min and then reuxed for 8 h. The salts were ltered off and the solvent was evaporated, and the remainder was dissolved in diethyl ether. The organic layer was washed with water, sodium hydroxide (aq., 5%), ammonium hydroxide (conc.), hydrochloric acid, water and nally dried (Na2SO4). The organic phase was then evaporated to give the title compound (2.63 g, 51%). 1H NMR (CDCl3) 7.22 (s, 2H), 4.39 (q, J = 7.1 Hz, 2H) 4.18 (q, J = 7.0 Hz, 4H), 1.5 (t, J = 7.0 Hz, 6H), 1.41 (t, J = 7.1 Hz, 3H), 13C NMR (CDCl3) 166.8, 157.0, 130.8, 107.9, 106.8, 65.4, 61.6, 14.7, 14.4. 6.1.2. 4-Bromo-5-ethoxy-3-hydroxybenzoic acid 2 A solution of 1 (1.02 g, 3.2 mmol) in N,N-dimethylformamide (5 mL) was added to sodium ethanethiolate (4.53 g, 51.2 mmol) in N,N-dimethylformamide (25 mL) under nitrogen atmosphere. The mixture was stirred at 100 C for 8 h. After cooling, the mixture was poured into water. The aqueous phase was acidied by the addition of hydrochloric acid (1 M) and extracted with diethyl ether. The organic phase was dried (Na2SO4) and evaporated The remainder was puried by ash chromatography (SiO2, toluene-acetic acid 5:1) to give the title compound (115 mg, 14%). 1H NMR (DMSO-d6) 7.13 (d, J = 1.8 Hz, 1H), 6.95 (d, J = 1.8 Hz, 1H) 4.08 (q, J = 7.0 Hz, 2H), 3.4 (br, 2H), 1.30 (t, J = 7.0Hz, 3H), 13C NMR (DMSO-d6) 167.5, 156.5, 155.8, 131.3, 109.7, 104.5, 104.3, 64.6, 14.6, MS (TSP) m/z: 263/261 (M + 1). 6.1.3. Methyl 4-bromo-5-ethoxy-3-hydroxybenzoate 3 A solution of 2 (115 mg, 0.44 mmol) in dry methanol (10 mL) was saturated with hydrogen chloride at 0 C and the mixture was stirred overnight at room temperature. The solution was evaporated and the remainder was dissolved in chloroform. The organic phase was washed with sodium bicarbonate (sat.), water, brine and then dried (Na2SO4) and evaporated to give the title compound (94 mg, 79%). 1H NMR (CDCl3) 7.35 (d, J = 1.8 Hz, 1H), 7.14 (d, J = 1.8 Hz, 1H) 5.85 (br, 1H), 4.17 (q, J = 7.0 Hz, 2H), 3.91 (s, 3H), 1.49 (t, J = 7.0 Hz, 3H), 13 C NMR (CDCl3) 167.2, 156.5, 154.4, 130.8, 110.1, 106.0, 105.5, 65.4, 52.7, 14.7, MS (TSP) m/z: 277/275 (M + 1). 6.1.4. Methyl 4-bromo-5-ethoxy-3-hydroxy-2-nitrobenzoate 4 A solution of 3 (89.7 mg, 0.326 mmol) in nitromethane (1 mL) and dichloromethane (1 mL) was stirred with nitric acid (90%, 15 L) at 40 C for 10 min. Ice was added, and the layers were separated. The aqueous layer was extracted with chloroform, and the combined organic layer was washed with water and evaporated. Purication of the remainder by ash chromatography (SiO2, tolueneethyl acetate 10:1) gave the title compound (37 mg, 35%). 1H NMR (CDCl3) 6.59 (s, 1H), 4.21 (q, J = 7.0 Hz, 2H) 3.90 (s, 3H), 1.47 (q, J = 7.0 Hz, 3H), 13C NMR (CDCl3) 167.1, 162.2, 154.4, 132.2, 126.9, 105.4, 103.3, 66.6, 53.9, 14.6, MS (TSP) m/z: 339/337 (M + NH4). 6.1.5. 4-Bromo-5-ethoxy-3-hydroxy-2-nitrobenzoic acid 5 Methyl 4-bromo-5-ethoxy-3-hydroxy-2-nitrobenzoate 4 (30.2 mg, 0.094 mmol) was dissolved in ethanol (1.2 mL) and potassium hydroxide (62 mg) in water (0.54 mL) was added. The mixture was stirred at 40 C for 16 h and then cooled and acidied to pH 2 by the addition of hydrochloric acid. The remainder was triturated with water, ltered and dried to give the title compound (24.6 mg, 86%). 1H NMR (CD3OD) 6.82 (s, 1H), 4.74 (br, 2H) 4.00 (q, J = 7.0 Hz, 2H), 1.26 (t, J = 7.0 Hz, 3H), 13C NMR (CD3OD) 167.3, 159.7, 151.0, 135.3, 128.2, 106.9, 106.3, 67.4, 15.4, MS (EI, 70 eV) m/z: 307/305 (M +, 99/100). 6.1.6. 4-Bromo-5-ethoxy-3-hydroxyanthranilic acid 6 A mixture of 5 (19.8 mg, 0.065 mmol) and SnCl2xH2O (73 mg) in ethanol (2.0 mL) was heated at 70 C for 5.5 h under nitrogen atmosphere. After cooling, water (2.0 mL) was added and the pH was adjusted to 7 by adding

735 sodium bicarbonate (aq., 5%). The mixture was extracted with ethyl acetate, and the organic phase was washed with brine and dried (Na2SO4). The organic phase was evaporated to give the title compound (8.4 mg, 47%). 1H NMR (CD3OD) 6.93 (s, 1H), 4.80 (br, 4H) 3.89 (q, J = 7.0 Hz, 2H), 1.30 (t, J = 7.0 Hz, 3H), 13C NMR (CD3OD) 171.7, 147.7, 145.1, 137.9, 109.5, 108.7, 107.2, 67.0, 15.7, MS (TSP) m/z: 278/276 (M + 1). 6.1.7. 5-Bromo-3-hydroxyanthranilic acid 7 To 3-hydroxyanthranilic acid (5 g, 32.6 mmol) in acetic acid (450 mL) was bromine (10.4 g, 65.2 mmol) in acetic acid (50 mL) added drop-wise. The thick slurry was stirred for 1 h at room temperature and was then evaporated. The remainder was dissolved in methanol (60 mL) and water was added. The precipitate was ltered and dried to give the title compound (7.38 g, 98%), m.p.: 210212 C. 1H NMR (DMSO-d6) 10.20 (br, 1H), 7.30 (d, J = 2.4 Hz, 1H) 6.88 (d, J = 2.4 Hz, 1H), 13C NMR (DMSO-d6) 169.2, 146.4, 141.2, 123.2, 118.9, 110.9, 104.2, MS (EI, 70 eV) m/z: 233/231 (M +, 14/13). 6.1.8. 4,5-Dibromo-3-hydroxyanthranilic acid 8 5-Bromo-3-hydroxyanthranilic acid 7 (7.38 g, 31.8 mmol) was dissolved in methanol (120 mL) and dichloromethane (250 mL) saturated with hydrogen bromide was added. More dichloromethane (300 mL) was added and the mixture was stirred for 48 h whereafter the solvent was evaporated. The residue (8.92 g) was dissolved in acetic acid (400 mL) and hydrobromic acid (48%, 100 mL) and bromine (5.5 g, 34.4 mmol) in acetic acid (50 mL) were added drop-wise for 20 min. The mixture was heated to 70 C and additional bromine (5.50 g, 34.4 mmol) in acetic acid (50 mL) was added. After 14 h at 90 C the solvent and excess reagent were evaporated. The remainder was crystallized from ethanolwater to give the title compound (8.45 g, 85%), m.p.: 223223.5 C. 1H NMR (DMSO- d6) 9.3 (br, 1H), 7.56 (s, 1H), 6.0 (br, 1H), 5.3 (br, 1H). 13C NMR (DMSO-d6) 168.9, 143.5, 142.5, 125.9, 117.9, 110.3, 107.2. MS (EI, 70 eV) m/z: 313/311/309 (M +, 51/100/50). Anal. for the hydrochloride C7H5Br2NO3xHCl (C, H, N). 6.1.9. Methyl 4-bromo-3-hydroxy-5-methylanthranilate 9 To a solution of methyl 3-hydroxy-5-methylanthranilate hydrochloride (50 mg, 0.23 mmol) dissolved in acetic acid (2.5 mL) under argon atmosphere was bromine (36 mg, 0.23 mmol) in acetic acid (2.5 mL) added drop-wise. After 3 h at room temperature, additional bromine (18 mg, 0.11 mmol) in acetic acid (1 mL) was added. The solvent and excess reagent were evaporated, and the remainder was puried by repeated preparative thin layer chromatography (SiO2, chloroform-methanolammonium hydroxide 300:10:1, chloroform-methanol 1 000:1) to give the title compound (9 mg, 15%). 1H NMR (DMSO- d6) 7.22 (s, 1H), 6.25 (br, 3H), 3.74 (s, 3H), 2.17 (s, 3H), MS (EI, 70 eV) m/z: 261/259 (M +, 85/86). 6.1.10. 4-Bromo-3-hydroxy-5-methylanthranilic acid 10 Methyl 4-bromo-3-hydroxy-5-methylanthranilate 9 (14 mg, 0.05 mmol) was dissolved in ethanol (0.7 mL) and ushed with argon. Potassium hydroxide (aq, 0.35 mL, 0.05 mmol) was added and the reaction was stirred at 40 C for 6 h. The mixture was acidied to pH 2 by the addition of hydrochloric acid (2 M). The solvent was removed and the remainder was triturated with ice-water. The crude product was puried by preparative HPLC (Lichrosorb C18) using methanol-ammonium hydroxide (0.05 M) (40:60) as the eluent to give the title compound (2 mg, 16%). 1H NMR (DMSO-d6) 7.23 (s, 1H), 3.35 (br, 2H), 2.13 (s, 2H), 2.05 (s, 3H), MS (EI, 70 eV) m/z: 247/245 (M +, 82/86). High resolution MS C8H8BrNO3. 6.1.11. 5-Bromo-4-chloro-3-hydroxyanthranilic acid 11 To 4-chloro-3-hydroxyanthranilic acid [23] (50 mg, 0.27 mmol) dissolved in acetic acid (2.0 mL) were hydrobromic acid (47%, 0.02 mL, 0.54 mmol) and then bromine (86 mg, 0.54 mmol) in acetic acid (3 mL) added, and the mixture was stirred at room temperature and under nitrogen atmosphere for 2 d. The solvent was evaporated and the remainder was crystallized from ethanol (40%), to give the title compound (37 mg, 51%), m.p.: 227228 C. 1H NMR (CD3OD) 7.72 (s, 1H), 5.0 (br, 4H), 13C NMR (CD3OD) 170.3, 143.5, 143.0, 127.2, 125.4, 111.0, 106.7. MS (TSP) m/z: 270/268/266 (M +, 28/100/80). Anal. C7H5BrClNO3 (C, H, N, O). 6.1.12. 4,5-Dichloro-3-hydroxyanthranilic acid 12 To a solution of 4-chloro-3-hydroxyanthranilic acid [32] (150 mg, 0.80 mmol) in acetic acid (12 mL) and under argon, were hydrochloric acid (12 M, 336 L, 4.8 mmol) and then chlorine (1.76 mmol) in acetic acid (1.78 mL) added drop-wise. The reaction was stirred for 20 h and then the precipitated material was ltered. The crude material was recrystallized from ethanol (50%) giving the title compound (14 mg, 8%). 1H NMR (CD3OD) 7.57 (s, 1H), 4.98 (br, 4H), 13C NMR (CD3OD) 170.7, 143.6, 143.4, 124.0, 118.7, 110.6, 109.5. MS (TSP) m/z: 224/222 (M +, 53/100). Anal. C7H5Cl2NO3 (C, H, N).

736 6.1.13. 6-Methoxy-7-nitrotetralin 13 To a solution of 6-methoxytetralin (6.80 g, 41.9 mmol) in dichloromethane (135 mL) under argon atmosphere and at 0 C was nitric acid (90%, 3.92 mL, 83.8 mmol) added drop-wise. The mixture was stirred for 25 h when iodomethane (7.85 mL, 126 mmol), tetra-n-butylammonium hydrogensulfate (28.4 g, 83.8 mmol) and sodium hydroxide (2 M, 8 mL, 16 mmol) were added and the mixture was heated at reux for 4 h. The organic layer was collected, washed with brine, sodium hydroxide (2 M), brine, dried (MgSO4) and evaporated. The remainder was puried by ash chromatography (SiO2, chloroformhexane 2:1) to give the title compound (2.62 g, 30%). 1H NMR (CDCl3) 7.60 (s, 1H), 6.74 (s, 1H), 3.90 (s, 3H), 2.79 (m, 2H) 2.71 (m, 2H) 1.811.77 (m, 4H), 13C NMR (CDCl3) 151.5, 145.5, 137.6, 130.0, 126.8, 114.2, 57.0, 30.5, 28.8, 23.3 23.1, MS (EI, 70 eV) m/z: 207 (M +, 59). 6.1.14. 6-Hydroxy-7-nitrotetralin 14 To 13 (2.92 g, 14.1 mmol) dissolved in dichloromethane (450 mL) at 65 C was boron tribromide (1 M, 28.2 mmol) in dichloromethane (28.2 mL) added for 15 min under an atmosphere of argon. The solution was stirred for 2 h when the temperature was gradually increased to 0 C. The organic phase was collected, washed with sodium bicarbonate (sat., 200 mL), brine, dried (MgSO4) and evaporated. The remainder was puried on a short column (SiO2, chloroform-hexane 2:1) to give the title compound (2.10 g, 77%). 1H NMR (DMSOd6) 10.47 (br, 1H), 7.62 (s, 1H), 6.80 (s, 1H), 2.72.6 (m, 4H), 1.651.71 (m, 4H), 13C NMR (DMSO-d6) 150.0, 146.1, 134.0, 128.3, 124.7, 118.5, 28.9, 27.5, 22.3, 22.0, MS (EI, 70 eV) m/z: 193 (M +, 100). 6.1.15. 5-Chloro-6-hydroxy-7-nitrotetralin 15 To a solution of 14 (2.44 g, 12.6 mmol) in chloroform (290 mL) under argon atmosphere was chlorine (0.99 M, 25.6 mmol) in chloroform (25.6 mL) added. The mixture was stirred at room temperature for 6 h when the solvent was evaporated. The remainder was puried by ash chromatography (SiO2, chloroform-hexane 1:1) to give the title compound (2.29 g, 80%). 1H NMR (DMSO-d6) 10.62 (br, 1H), 7.71 (s, 1H), 2.752.69 (d, 4H), 1.771.64 (d, 4H), 13C NMR (DMSO-d6) 146.3, 143.6, 134.6, 129.7, 123.4, 122.8, 28.3, 27.9, 21.7, 21.6, MS (EI, 70 eV) m/z: 229/227 (M +, 32/100). 6.1.16. 6-Benzyloxy-5-chloro-7-nitrotetralin 16 To a solution of 15 (2.28 g, 10.0 mmol) in dry N, N-dimethylformamide (40 mL) under argon atmosphere, were benzyl chloride (11.5 mL, 100 mmol), tetra-nbutylammonium hydrogensulfate (95 mg, 0.25 mmol) and potassium carbonate (41.5 g, 30.0 mmol) added. The mixture was stirred at room temperature for 24 h, the salts were then ltered and the solvent was evaporated. The remainder was dissolved in ethyl acetate, and the organic phase was washed with brine, dried (MgSO4) and evaporated. The crude material was puried by ash chromatography (SiO2, chloroform-hexane 1:1) to give the title compound (2.20 g, 69%), m.p.:8082 C. 1H NMR (DMSO-d6) 7.73 (s, 1H), 7.487.37 (m, 5H), 5.06 (s, 2H), 2.792.75 (m, 4H), 1.791.69 (m, 4H). 13C NMR (DMSO-d6) 144.8, 142.1, 135.8, 135.6, 129.6, 128.4, 123.4, 106.2, 105.5, 28.6, 27.6, 21.6, 21.4. MS (TSP) m/z: 337/335 (M + NH4, 30/100). Anal. C17H16ClNO3 (C, H, N). 6.1.17. 7-Amino-6-benzyloxy-5-chlorotetralin 17 To a suspension of 16 (2.56 g, 8.33 mmol) in methanol at 1 C was copper(I) chloride (4.95 g, 25.0 mmol) and potassium borohydride (3.25 g, 60.1 mmol) added portion-wise for 8 h. The mixture was ltered and evaporated. The remainder was dissolved in ethyl acetate and the organic phase was washed with water, brine, dried (Na2SO4) and evaporated. The crude material was puried by ash chromatography (SiO2, chloroform) to give the title compound (1.82 g, 76%). 1H NMR (DMSO-d6) 7.55 (dd, J = 1.6 Hz, J = 8.1 Hz, 2H), 7.417.34 (m, 3H), 6.42 (s, 1H), 4.84 (s, 2H), 4.79 (s, 2H), 2.582.52 (m, 4H), 1.711.60 (m, 4H), 13C NMR (DMSO-d6) 140.1, 139.3, 137.3, 134.0, 128.3, 128.2, 128.1, 127.9, 127.0, 121.8, 114.0, 72.8, 29.0, 26.2, 22.7, 22.3, MS (TSP) m/z: 290/288 (M + 1, 27/100). Anal. C17H18ClNO (C, H, N). 6.1.18. 6-Benzyloxy-5-chloro-7-isonitrosoacetaminotetralin 18 To a solution of 17 (1.84 g, 6.41 mmol) in N, N-dimethylformamide (80 mL) and water (8 mL) under argon atmosphere, were hydrochloric acid (12 M, 0.53 mL) and chloral hydrate (1.17 g, 7.05 mmol) added. The mixture was heated to 110 C when hydroxylamine hydrochloride (1.78 g, 25.6 mmol) in water (8 mL) was added. The mixture was heated to 100 C for 1 h and the solvent was evaporated. The remainder was dissolved in ethyl acetate, and the organic phase was washed with water, brine, dried (Na2SO4) and evaporated. The crude material was puried by ash chromatography (SiO2, ethyl acetate-chloroform 1:5) to give the title compound (1.35 g, 59%). 1H NMR (DMSO-d6) 12.33 and 9.72 (E/Z) (2s, 1H), 9.20 and 8.28 (E/Z) (2s, 1H), 7.84 and 7.72 (E/Z) (2s, 1H), 7.60 (s, 1H), 7.577.37 (m, 5H), 4.90 and 4.88 (E/Z) (2s, 2H), 2.712.65 (m, 4H), 1.751.68 (m, 4H), 13C NMR (DMSO-d6) 160.3, 160.0, 143.4, 143.0, 136.6, 136.1, 134.4, 134.3, 131.2, 130.6, 129.8, 129.5, 128.8, 128.3, 128.2, 128.1, 127.0, 120.7, 120.3,

737 74.8, 74.3, 29.1, 26.7, 22.2, 22.0, MS (TSP) m/z: 361/359 (M + 1, 28/100). 6.1.19. 9-Benzyloxy-8-chloro-1H-4,5,6,7-tetrahydro[e] benzindole-2,3-dione 19 To sulfuric acid (conc., 5 mL) at 60 C, 18 (500 mg, 1.39 mmol) was added portion-wise for 1 h. The mixture was poured on ice (50 mL) and extracted with ethyl acetate (200 mL). The organic phase was dried (Na2SO4) and evaporated. The remainder was dissolved in N, N-dimethylformamide (3 mL) under argon atmosphere, and benzyl bromide (0.16 mL, 1.39 mmol) and potassium carbonate (192 mg, 1.39 mmol) were added. The mixture was stirred at room temperature for 18 h. The salt was ltered off and the solvent was evaporated. The remainder was puried by repeated ash chromatography (SiO2, ethyl acetate-methanol 20:1, chloroform-methanol 50:1) to give the title compound (56 mg, 12%). 1H NMR (DMSO-d6) 11.40 (s, 1H), 7.58 (d, J = 7.3 Hz, 2H), 7.427.36 (m, 3H), 4.90 (s, 2H), 2.92 (t, J = 7.3 Hz, 2H), 2.63 (t, J = 6.0 Hz, 2H), 1.741.66 (m, 4H), 13C NMR (DMSO-d6) 183.8, 159.4, 142.3, 137.3, 137.1, 136.2, 136.2, 129.9, 128.8, 128.2, 128.1, 114.7, 74.6, 26.7, 25.4, 21.7, 20.8, MS (EI, 70 eV) m/z: 343/341 (M +, 5/15). 6.1.20. 6-Amino-7-benzyloxy-5-carboxy-8-chlorotetralin 20 To a suspension of 19 (51 mg, 0.15 mmol) in sodium hydroxide (0.68 M, 0.60 mmol) and water (0.46 mL) at 10 C, was hydrogen peroxide (30%, 86 L, 0.85 mmol) in sodium hydroxide (0.68 M, 0.90 mmol) added. The mixture was stirred for 3 h and then ltered. Water and acetic acid were added and the mixture was extracted with ethyl acetate (40 mL). The organic phase was washed with brine, dried (Na2SO4) and evaporated to give the title compound (35 mg, 70%). 1H NMR (DMSOd6) 7.56 (d, J = 7.0 Hz, 2H), 7.437.36 (m, 3H), 4.84 (s, 2H), 3.3 (br, 3H), 2.73 (m, 2H), 2.59 (m, 2H) 1.701.61 (m, 4H), 13C NMR (DMSO-d6) 169.3, 139.9, 139.5, 136.8, 133.1, 129.5, 128.3, 128.1, 122.3, 116.5, 73.1, 28.1, 26.8, 22.2, 22.0, MS (EI, 70 eV) m/z: 333/331 (M +, 7/18). 6.1.21. 6-Amino-5-carboxy-8-chloro-7-hydroxytetralin 21 6-Amino-7-benzyloxy-5-carboxy-8-chlorotetralin 20 (33 mg, 0.10 mmol) in ethanol (3 mL) was hydrogenated at room temperature and at atmospheric pressure for 2 h with Pd/C (5%, 4 mg) as the catalyst. The mixture was ltered, evaporated and dried to give the title compound (21 mg, 87%), m.p.: 147 C (dec.). 1H NMR (DMSO-d6) 7.9 (br, 4H), 2.66 (t, J = 7 Hz, 2H), 2.54 (t, J = 7 Hz, 2H), 1.701.64 (m, 2H), 1.631.57 (m, 2H), 13C NMR (DMSO-d6) 169.7, 138.1, 136.4, 128.1, 123.3, 121.4, 115.2, 27.9, 26.9, 22.4, 22.2, MS (EI, 70 eV) m/z: 333/331 (M +, 21/65). 6.1.22. 4,5-Dimethyl-2-nitrophenol 22 To a solution of 3,4-dimethylphenol (20.0 g, 164 mmol) in dichloromethane (400 mL) under argon atmosphere and at 2 C was nitric acid (90%, 7.7 mL, 165 mmol) added drop-wise. The mixture was stirred for 90 min while the temperature was slowly increased to ambient temperature. To the mixture was dichloromethane (500 mL) added and the organic phase was washed with brine, water, dried (MgSO4) and evaporated. The remainder was puried by ash chromatography (SiO2, ethyl acetate-hexane 1:1) to give the title compound (10 g, 36%), m.p.: 8283 C. 1H NMR (DMSOd6) 11.5 (br, 1H), 7.71 (s, 1H), 6.92 (s, 1H), 2.22 (s, 3H), 2.17 (s, 3H), 13C NMR (DMSO-d6) 150.8, 146.3, 133.4, 128.0, 124.9, 119.7, 19.7, 18.1, MS (EI, 70 eV) m/z: 167 (M +, 100). 6.1.23. 2-Chloro-3,4-dimethyl-6-nitrophenol 23 To a solution of 22 (7.0 g, 41.9 mmol) dissolved in chloroform (300 mL) under argon atmosphere was chlorine (0.99 M, 83.7 mmol) in chloroform (84.8 mL) added. The mixture was stirred for 26 h and then the solvent and excess reagents were evaporated. The remainder was dissolved in dichloromethane (400 mL) and the organic phase was washed with brine, dried (MgSO4) and evaporated. The crude material was puried by ash chromatography (SiO2, chloroform-hexane 1:1) to give the title compound (6.47 g, 77%), m.p.: 6263 C. 1H NMR (DMSO-d6) 11.5 (br, 1H), 7.80 (s, 1H), 2.34 (s, 3H), 2.27 (s, 3H). 13C NMR (DMSO-d6) 146.9, 143.8, 134.1, 128.8, 123.7, 122.9, 19.4, 17.4. MS (EI, 70 eV) m/z: 203/201 (M +, 37/100). 6.1.24. 2-Benzyloxy-3-chloro-4,5-dimethylnitrobenzene 24 To a solution of 23 (6.44 g, 31.9 mmol) in dry N, N-dimethylformamide (105 mL) under argon atmosphere, were benzyl bromide (4.17 mL, 35.1 mmol) and potassium carbonate (13.2 g, 95.7 mmol) added. The mixture was stirred at room temperature for 8 h and was then ltered. The solvent was evaporated, and the remainder was puried by ash chromatography (SiO2, chloroform-hexane 1:1) affording the title compound (8.44 g, 91%), m.p.: 7475 C. 1H NMR (DMSO-d6) 7.80 (s, 1H), 7.487.37 (m, 5H), 5.06 (s, 2H), 2.37 (s, 3H), 2.34 (s, 3H), 13C NMR (DMSO-d6) 145.2, 142.2, 142.1, 135.7, 134.7, 129.6, 128.4, 128.4, 123.6, 75.7, 19.7, 17.1. MS (EI, 70 eV) m/z: 293/291 (M +, 8/37).

738 6.1.25. 2-Benzyloxy-3-chloro-4,5-dimethylaniline 25 To a solution of 24 (3.00 g, 10.3 mmol) in methanol (420 mL) at 2 C, were copper(I) chloride (6.11 g, 30.8 mmol) and potassium borohydride (4.4 g, 81.6 mmol) added in portions. The mixture was ltered and evaporated. The remainder was dissolved in ethyl acetate, and the organic phase was washed with water, dried (MgSO4) and evaporated. The crude material was puried by ash chromatography (SiO2, chloroform) to give the title compound (2.14 g, 79%), m.p.: 5758 C. 1 H NMR (DMSO-d6) 7.54 (d, J = 6.5 Hz, 2H), 7.427.34 (m, 3H), 6.52 (s, 1H), 4.83 (s, 2H), 4.79 (s, 2H), 2.12 (s, 6H), 13C NMR (DMSO-d6) 140.0, 138.9, 137.3, 133.0, 128.3, 128.2, 127.9, 127.3, 121.2, 115.4, 72.7, 20.2, 15.4, MS (EI, 70 eV) m/z: 286/284 (M + 23, 42/100). 6.1.26. 2-Benzyloxy-3-chloro-4,5-dimethylisonitrosoacetanilide 26 To a solution of 25 (2.14 g, 8.19 mmol) in N, N-dimethylformamide (60 mL) and water (2 mL) were hydrochloric acid (conc., 0.68 mL, 8.19 mmol) and chloral hydrate (1.49 g, 9.00 mmol) added. The mixture was heated to 105 C and hydroxylamine hydrochloride (2.28 g, 32.8 mmol) dissolved in water (4 mL) was added. The mixture was stirred for 1 h and the solvent was evaporated. The remainder was dissolved in ethyl acetate, and the organic layer was washed with water, dried (MgSO4) and evaporated. The crude material was puried by ash chromatography (SiO2, chloroformethyl acetate 10:1) affording the title compound (1.28 g, 47%). 1H NMR (DMSO-d6) 12.33 and 9.71 (E/Z) (2s, 1H), 9.20 and 8.27 (E/Z) (2s, 1H), 7.91 and 7.83 (2s, 1H), 7.60 (s, 1H), 7.657.36 (m, 5H), 4.90 (s, 1H), 4.87 (s, 1H), 2.262.23 (m, 6H), 13C NMR (DMSO-d6) 160.2, 160.0, 143.4, 143.1, 136.6, 136.1, 133.4, 133.3, 131.0, 130.4, 129.7, 129.4, 128.7, 128.4, 128.4, 128.3, 128.2, 128.1, 128.0, 127.4, 127.2, 121.5, 121.2, 74.7, 74.2, 20.3, 16.1, MS (EI, 70 eV) m/z: 334/332 (M +, 4/12). 6.1.27. 7-Benzyloxy-6-chloro-4,5-dimethylisatin 27 To sulfuric acid (conc., 6 mL) at 80 C was 26 (700 mg, 2.10 mmol) added. The mixture was stirred at 80 C for 10 min and then poured into ice-water (200 mL). The mixture was extracted with ethyl acetate (200 mL) and the organic phase was dried (MgSO4) and evaporated. The remainder was dissolved in N, N-dimethylformamide and benzyl bromide (0.28 mL, 2.30 mmol) and potassium carbonate (318 mg, 2.30 mmol) were added. The mixture was stirred at ambient temperature for 30 h and ltered. Acetic acid (1.5 mL) was added, and the solvents were evaporated. The remainder was puried by repeated ash chromatography (SiO2, dichloromethane-methanol 50:1) to give the title compound (13 mg, 2%). 1H NMR (DMSO-d6) 11.42 (s, 1H), 7.57 (d, J = 6.2 Hz, 2H), 7.427.35 (m, 3H), 4.92 (s, 2H), 2.44 (s, 3H), 2.22 (s, 3H), 13C NMR (DMSO-d6) 184.3, 159.3, 141.9, 137.0, 137.0, 136.3, 135.4, 130.2, 128.8, 128.3, 128.2, 115.6, 74.6, 15.3, 14.2, MS (EI, 70 eV) m/z: 317/315 (M +, 3/7). 6.1.28. 3-Benzyloxy-4-chloro-5,6-dimethylanthranilic acid 28 To a suspension of 27 (13 mg, 0.04 mmol) in dioxane (0.5 mL) and sodium hydroxide (0.68 M, 0.14 mmol) at 10 C, was hydrogen peroxide (30%, 7 L, 0.22 mmol) dissolved in sodium hydroxide (0.41 mL) added in portions. The mixture was concentrated in a stream of nitrogen, when acetic acid (38 L) was added. The mixture was dissolved in ethyl acetate (3 mL) and the organic phase was washed with brine, dried (MgSO4) and evaporated to give the title compound (11 mg, 90%). 1H NMR (DMSO-d6) 7.55 (d, J = 7.0 Hz, 2H), 7.437.35 (m, 3H), 4.82 (s, 2H), 2.21 (s, 3H), 2.19 (s, 3H), MS (EI, 70 eV) m/z: 307/305 (M +, 7/19). 6.1.29. 4-Chloro-5,6-dimethyl-3-hydroxyanthranilic acid 29 3-Benzyloxy-4-chloro-5,6-dimethylanthranilic acid 28 (10 mg, 0.03 mmol) in ethanol (1.5 mL) was hydrogenated at ambient temperature and atmospheric pressure for 5 h with Pd/C (10%, 2 mg) as the catalyst. The mixture was ltered and the solvent was evaporated. The remainder was puried by preparative HPLC (Lichrosorb-C18, methanol-phosphate buffer, pH 3, 1:1). The fractions containing the product were pooled and evaporated. To the residue was ethyl acetate added and the organic phase was washed with sodium carbonate (sat), dried and evaporated to give the title compound (3 mg, 46%). 1H NMR (DMSO-d6) 3.35 (br, 1H), 3.16 (s, 2H), 2.16 (s, 3H), 2.15 (s, 3H), 13C NMR (DMSO-d6) 169.9, 137.9, 135.2, 126.4, 123.5, 121.4, 117.3, 17.7, 16.0, MS (EI, 70 eV): m/z 217/215 (M +, 21/63). 6.1.30. 2,4-Dichloro-5-methoxy-3-methylphenyl triuoromethanesulfonate 30 To a solution of 2,4-dichloro-5-methoxy-3-methylphenol [33] (7.73 g, 37.3 mmol) in dichloromethane (180 mL) were triethylamine (10.4 mL, 74.7 mmol) and N,N-dimethylaminopyridine (10 mg, 0.08 mmol) added. The solution was cooled to 78 C and triuoromethanesulfonic anhydride (9.4 mL, 74.7 mmol) was added dropwise. The mixture was then stirred for 20 min while the temperature slowly reached 0 C. To the solution was added dichloromethane (200 mL) and the organic phase

739 was washed with water (150 mL), brine, dried (MgSO4) and evaporated. The remainder was puried by ash chromatography (SiO2, ethyl acetate-hexane 1:3) affording the title compound (12.3 g, 97%), m.p.: 58.559.0 C. 1 H NMR (DMSO-d6) 7.30 (s, 1H), 3.92 (s, 3H), 2.49 (s, 3H), 13C NMR (DMSO-d6) 154.2, 143.9, 137.0, 122.8, 118.1, 118.0 (q, J = 321Hz), 105.3, 57.2, 18.0, MS (EI, 70 eV) m/z: 340/338 (M +, 47/64). 6.1.33. 4,6-Dichloro-3-hydroxy-5-methyl-2-nitrobenzoic acid 33 To a solution of 32 (90 mg, 0.41 mmol) in nitromethane (9 mL) at 40 C, was nitric acid (90%, 20 mL, 0.43 mmol) added. The mixture was stirred for 4 h, when the solvent was removed by evaporation. The remainder was puried by ash chromatography (SiO2, ethyl acetate-acetic acid 30:1) to give the title compound (79 mg, 72%), m.p.: 197199 C. 1H NMR (DMSO-d6) 2.45 (s, 3H). 13C NMR (DMSO-d6) 164.3, 147.1, 139.6, 136.3, 128.4, 126.5, 118.9, 18.6. MS (EI, 70 eV) m/z: 267/265 (M +, 66/100). 6.1.34. 4,6-Dichloro-3-hydroxy-5-methylanthranilic acid 34 4,6-Dichloro-3-hydroxy-5-methyl-2-nitrobenzoic acid 33 (69 mg, 0.26 mmol) in acetic acid (10 mL) and hydrochloric acid (conc., 33 mL) was hydrogenated at atmospheric pressure and room temperature for 2 h with Pd/C (10%, 10 mg) as the catalyst. Methanol was added and the mixture was ltered. The solvent was evaporated and the remainder was puried by ash chromatography (SiO2, ethyl acetate-acetic acid 45:1) affording the title compound (55 mg, 90%), m.p.: 192 C (dec.). 1H NMR (DMSO-d6) 2.27 (s, 3H), 13C NMR (DMSO-d6) 167.4, 139.2, 136.0, 122.9, 121.3, 120.4, 116.9, 17.0, MS (EI, 70 eV) m/z: 237/235 (M +, 45/79). 6.1.35. 4,6-Dibromo-3-hydroxy-2-nitrobenzoic acid 35 To a cooled solution of 3-hydroxy-2-nitrobenzoic acid (10.5 g, 0.057 mol) and sodium acetate (9.85 g, 0.57 mol) in acetic acid (100 mL) was bromine (6.15 mL, 0.12 mol) added drop-wise. The mixture was stirred at 60 C for 68 h and then cooled and ltered. The solution was evaporated, and the remainder was dissolved in ethyl acetate. The organic phase was washed with dilute hydrochloric acid, dried (MgSO4) and evaporated. The crude product was puried by ash chromatography (SiO2, toluene-acetic acid 5:1). The pure compound was crystallized from methanol to give the title compound (15.4 g, 79%), m.p.: 201202 C (dec.). 1H NMR (DMSO-d6) 8.20 (s, 1H), 13C NMR (DMSO-d6) 164.4, 146.9, 139.1, 130.5, 116.2, 108.1, MS (EI, 70 eV) m/z: 343/341/339 (M +, 46/98/49). Anal. C7H3Br2NO5 (C, H, N). 6.1.36. 4,6-Dibromo-3-hydroxyanthranilic acid 36 4,6-Dibromo-3-hydroxy-2-nitrobenzoic acid 35 (4.09 g, 12 mmol) in ethanol (150 mL) was hydrogenated at atmospheric pressure and room temperature for 45 h, with PtS2 (0.16 g, 0.62 mmol) as the catalyst. The mixture was ltered and the solvent was removed by evaporation. The remainder was puried by ash chromatog-

6.1.31. Methyl 2,4-dichloro-5-methoxy-3-methylbenzoate 31 To a solution of 30 (7.60 g, 22.4 mmol) in dioxane (75 mL) were 1,3-bis(diphenylphosphino)propane (0.37 g, 0.90 mmol) and palladium acetate (0.20 g, 0.90 mmol) added. The mixture was ushed with carbon monoxide, then triethylamine (6.90 mL, 49.4 mmol) and methanol (23 mL) were added. The mixture was stirred at 70 C and atmospheric pressure for 25 h and was then ltered and evaporated. The remainder was dissolved in diethyl ether and the organic phase was washed with ammonium hydroxide (2 M, 150 mL), brine (150 mL), dried (MgSO4) and evaporated. The crude material was puried by ash chromatography (SiO2, ethyl acetatehexane 1:3) to give the title compound (4.18 g, 75%), m.p.: 73.574 C. 1H NMR (DMSO-d6) 7.33 (s, 1H), 3.88 (s, 3H), 3.86 (s, 3H), 2.45 (s, 3H), 13C NMR (DMSO-d6) 165.7, 153.3, 136.1, 130.5, 125.2, 123.0, 110.8, 56.6, 52.7, 17.9, MS (EI, 70 eV) m/z: 250/248 (M +, 53/80).

6.1.32. 2,4-Dichloro-5-hydroxy-3-methylbenzoic acid 32 To a solution of 31 (0.24 g, 0.96 mmol) in methanol (30 mL) under argon atmosphere, was potassium hydroxide (0.31 g, 4.8 mmol) added and the mixture was stirred at 50 C for 19 h. The solvent was evaporated, and hydrobromic acid (48%, 30 mL) was added and the mixture was stirred at 110 C for 3 d. The acid was evaporated, and the remainder was dissolved in ethyl acetate (40 mL). The organic phase was extracted with ammonium hydroxide (dil., 11 mL) and the pH of the aqueous phase was lowered to 1. The aqueous phase was extracted with ethyl acetate (40 mL), the organic phase was washed with brine, dried (MgSO4) and evaporated to give, after crystallization from diethyl ether, the title compound (0.20 g, 94%), m.p.: 203.5204.5 C. 1H NMR (DMSO-d6) 10.68 (br, 1H), 7.15 (s, 1H), 2.42 (s, 3H), 13C NMR (DMSO-d6) 166.7, 151.8, 135.9, 131.2, 123.7, 121.2, 114.4, 17.9, MS (EI, 70 eV) m/z: 222/220 (M +, 56/100).

740 raphy (SiO2, toluene-acetic acid 5:1) affording, after crystallization from methanol-water, the title compound (2.51 g, 63%), m.p.: 162164.5 C. 1H NMR (DMSO-d6) 6.96 (s, 1H), 13C NMR (DMSO-d6) 167.7, 140.6, 140.0, 121.8, 117.2, 112.5, 110.4. MS (EI, 70 eV) m/z: 313/311/309 (M +, 36/72/34). Anal. C7H5Br2NO3 (C, H, N). 6.1.37. 4-Bromo-5-hydroxy-2-methoxybenzoic acid 37 To a solution of 5-hydroxy-2-methoxybenzoic acid [34] (1.24 g, 73.7 mmol) in acetic acid (100 mL) was bromine (0.38 mL, 7.4 mmol) added drop-wise. The mixture was stirred for 3 h and the solvent was then evaporated. The remainder was puried by ash chromatography (SiO2, toluene-acetic acid 10:1) to give the title compound (1.5 g, 83%). 1H NMR (CD3OD) 7.38 (s, 1H), 7.26 (s, 1H), 3.84 (s, 3H), 13C NMR (CD3OD) 169.0, 154.1, 149.6, 121.3, 119.6, 119.0, 116.6, 57.7. 6.1.38. 4-Bromo-3-hydroxy-6-methoxy-2-nitrobenzoic acid 38 To a solution of sodium nitrate (361 mg, 4.25 mmol), lanthanum nitrate hexahydrate (18 mg, 0.042 mmol) and hydrochloric acid (12 M, 4 mL) in water (4 mL) at 0 C, was 37 (1.05 g, 4.25 mmol) in diethyl ether (20 mL) added in portions. The mixture was stirred for 7 h, while the solution reached ambient temperature. Dichloromethane (90 mL) and water (20 mL) were added, and the organic phase was collected, dried (MgSO4), ltered and evaporated. The remainder was puried by ash chromatography (SiO2, toluene-ethyl acetate-acetic acid 8:2:1) affording the title compound (600 mg, 48%). 1H NMR (CD3OD) 7.57 (s, 1H), 3.85 (s, 3H), MS (EI, 70 eV) m/z 293/291 (M +, 21/19). 6.1.39. 4-Bromo-3-hydroxy-6-methoxyanthranilic acid 39 4-Bromo-3-hydroxy-6-methoxy-2-nitrobenzoic acid 38 (52 mg, 0.18 mmol) in ethanol (7 mL) was hydrogenated at atmospheric pressure and room temperature for 18 h with PtS2 as the catalyst. The solution was ltered and the solvent was evaporated. The remainder was puried by ash chromatography (SiO2, toluene-ethyl acetate-acetic acid 8:2:1) to give the title compound (30 mg, 64%). 1H NMR (CD3OD) 6.40 (s, 1H), 3.88 (s, 3H), 13C NMR (CD3OD) 170.8, 154.8, 145.6, 137.8, 115.5, 102.1, 101.6, 57.6, MS (EI, 70 eV): m/z 263/261 (M +, 79/80). 6.1.40. 2,4-Dichloro-5-methoxyphenyl triuoromethanesulfonate 40 To a solution of 2,4-dichloro-5-methoxyphenol [35] (4.43 g, 22.9 mmol), triethylamine (6.40 mL, 45.9 mmol) and N,N-dimethylaminopyridine (5 mg, 0.04 mmol) in dichloromethane (10 mL) at 70 C, was triuromethanesulfonic anhydride (5.79 mL, 34.4 mmol) added. The mixture was stirred for 20 min while the temperature was increased to ambient temperature. Dichloromethane (150 mL) was added and the organic phase was washed with water (100 mL), brine (100 mL), dried (MgSO4) and evaporated. The remainder was puried by ash chromatography (SiO2, dichloromethane-hexane 1:2) affording the title compound (6.37 g, 86%). 1H NMR (DMSO-d6) 7.99 (s, 1H), 7.42 (s, 1H), 3.92 (s, 3H), 13C NMR (DMSO-d6) 154.6, 144.0, 130.9, 122.3, 118.0 (q, J = 321 Hz), 117.1, 108.2, 57.3, MS (EI, 70 eV) m/z: 328/326/324 (M +, 6/37/55). 6.1.41. Methyl 2,4-dichloro-5-methoxybenzoate 41 To a solution of 40 (6.35 g, 19.5 mmol) in N, N-dimethylformamide (65 mL) ushed with carbon monoxide were 1,3-bis(diphenylphosphino)propane (314 mg, 0.76 mmol), palladium acetate (171 mg, 0.76 mmol), triethylamine (6.0 mL) and methanol (14.5 mL) added. The reaction was stirred at 70 C and atmospheric pressure for 5 h, the solvent was then evaporated. The remainder was dissolved in diethyl ether (600 mL) and the organic phase was washed with ammonium hydroxide (2 M, 300 mL), brine (200 mL), dried (MgSO4) and evaporated. The crude product was puried by ash chromatography (SiO2, dichloromethane-hexane 2:3) affording the title compound (1.65 g, 36%), m.p.: 8889 C. 1H NMR (DMSO-d6) 7.73 (s, 1H), 7.49 (s, 1H), 3.90 (s, 3H), 3.86 (s, 3H), 13C NMR (DMSO-d6) 164.8, 153.4, 131.3, 129.6, 125.3, 123.2, 114.2, 56.7, 52.7, MS (EI, 70 eV) m/z: 238/236/234 (M +, 8/55/83). 6.1.42. 2,4-Dichloro-5-hydroxybenzoic acid 42 To a solution of 41 (600 mg, 2.55 mmol) in dichloromethane (10 mL) at 70 C and under argon atmosphere was boron tribromide (0.72 mL, 7.66 mmol) added. The mixture was stirred for 4 h while the temperature slowly reached ambient temperature. Dichloromethane (20 mL) and sodium hydroxide (2 M, 15 mL) were added and stirring was continued for 30 min. The aqueous phase was washed with dichloromethane (10 mL) and was then acidied to pH 1 by the addition of hydrochloric acid (12 M). The aqueous phase was extracted with ethyl acetate (40 mL), and the organic phase washed with brine, dried (MgSO4) and evaporated to give the title compound (350 mg, 66%), m.p.: 200.5201.5 C. 1 H NMR (DMSO-d6) 10.8 (br, 1H) 7.56 (s, 1H), 7.38 (s, 1H), 13C NMR (DMSO-d6) 165.8, 152.0, 131.3, 130.3, 123.6, 121.6, 118.1, MS (EI, 70 eV) m/z: 210/208/206 (M +, 11/64/100).

741 6.1.43. 4,6-Dichloro-3-hydroxy-2-nitrobenzoic acid 43 To a solution of 42 (280 mg, 1.35 mmol) in nitromethane (35 mL) at 45 C was nitric acid (90%, 63 L, 1.35 mmol) added. The mixture was stirred for 4 h and then the solvent was evaporated. The remainder was dissolved in ethyl acetate (100 mL), and the organic phase was washed with water (5 mL), brine (5 mL), dried (MgSO4) and evaporated. The crude product was puried by ash chromatography (SiO2, ethyl acetate-acetic acid 50:1) affording the title compound (323 mg, 95%), m.p.: 186 C (dec.). 1H NMR (DMSO-d6) 7.88 (s, 1H), 13C NMR (DMSO-d6) 163.8, 147.2, 139.0, 133.1, 128.1, 126.3, 118.2, MS (EI, 70 eV) m/z: 253/251 (M +, 34/49). 6.1.44. 4,6-Dichloro-3-hydroxyanthranilic acid hydrochloride 44 4,6-Dichloro-3-hydroxy-2-nitrobenzoic acid 43 (223 mg, 0.88 mmol) dissolved in ethanol (50 mL) was hydrogenated at atmospheric pressure and room temperature for 1 h with Pd/C (10%, 30 mg) as the catalyst. The mixture was ltered and the solvent was evaporated. The remainder was puried by ash chromatography (SiO2, ethyl acetate-acetic acid 50:1) affording the free base. The base was dissolved in tetrahydrofuran (0.6 mL) and ethereal hydrogen chloride (3 M, 1 mL) was added to give the title compound (32 mg, 14%), m.p.: 231 C (dec.). 1H NMR (DMSO-d6) 6.68 (s, 1H), 13C NMR (DMSO-d6) 167.2, 140.3, 139.1, 122.5, 121.9, 116.2, 114.0, MS (EI, 70 eV) m/z: 223/221 (M +, 42/65). Anal. C7H5Cl2NO3xHCl (C, H, N). 6.1.45. 3-Chloro-2-methoxy-5-phenylnitrobenzene 45 A solution of 2-chloro-6-nitro-4-phenylphenol [36] (21.2 g, 0.08 mol), potassium carbonate (17.6 g, 0.13 mol) and iodomethane (13.5 mL, 0.22 mol) in dry N, N-dimethylformamide (150 mL) was stirred at room temperature under nitrogen atmosphere for 19 h, when the solvent was evaporated. The remainder was puried by ash chromatography (SiO2, hexane ethyl acetatehexane 10:1) affording, after crystallization from ethyl acetate-hexane, the title compound (22.2 g, 99%), m.p.: 7677.5 C. 1H NMR (CDCl3) 7.88 (d, J = 2.3 Hz, 1H), 7.81 (d, J = 2.3 Hz, 1H), 7.537.40 (m, 5H), 4.05 (s, 3H), 13 C NMR (CDCl3) 148.8, 145.5, 138.1, 137.1, 132.8, 130.8, 129.4, 129.2, 128.7, 126.8, 121.8, 62.6, MS (EI, 70 eV) m/z: 265/263 (M +, 30/98). Anal. C13H10ClNO3 (C, H, N). 6.1.46. 3-Chloro-2-methoxy-5-phenylaniline 46 3-Chloro-2-methoxy-5-phenylnitrobenzene 45 (9.1 g, 34.5 mmol) was hydrogenated in tetrahydrofuran (100 mL), methanol (100 mL) and hydrochloric acid (2 M, 34.5 mL) at atmospheric pressure and room temperature for 3 h with Pd/C (10%, 180 mg) as the catalyst. The mixture was ltered and the solvent was evaporated. The remainder was dissolved in ethyl acetate and the organic phase was washed with ammonium hydroxide (2 M), dried (MgSO4) and evaporated. The crude material was crystallized from ethyl acetate-hexane affording the title compound (5.05 g, 62%), m.p.: 6566.5 C. 1H NMR (CDCl3) 7.497.31 (m, 5H), 6.96 (d, J = 2.1 Hz, 1H), 6.82 (d, J = 2.1 Hz, 1H), 3.97 (s, 2H), 3.86 (s, 3H), 13 C NMR (CDCl3) 142.5, 141.4, 139.9, 138.4, 128.7, 127.9, 127.4, 126.8, 118.2, 112.9, 59.8, MS (EI, 70 eV) m/z: 235/233 (M +, 24/48). Anal. C13H12ClNO (C, H, N). 6.1.47. 3-Chloro-2-methoxy-5-phenylisonitrosoacetanilide 47 To a stirred solution of chloralhydrate (1.08 g, 6.53 mmol) and sodium sulfate (5.0 g, 35.2 mmol) in water (20 mL) was 46 (1.01 g, 4.33 mmol) dissolved in a mixture of water (5 mL), N,N-dimethylformamide (10 mL) and hydrochloric acid (conc., 0.43 mL) added. The mixture was stirred at 95 C for 90 min, when hydroxylamine hydrochloride (1.36 g, 19.5 mmol) was added, and the stirring was continued for another 22 h. The mixture was cooled and evaporated and then ethyl acetate was added. The organic phase was washed with water, dried (MgSO4) and evaporated. The remainder was puried by ash chromatography (SiO2, dichloromethane ethyl acetate) to give the title compound (1.04 g, 79%). 1H NMR (DMSO-d6) 9.51 (s, 1H), 8.39 (s, 1H), 7.80 (s, 1H), 7.60 (d, 2H), 7.517.34 (m, 5H), 3.83 (s, 3H), 13C NMR (DMSO-d6) 160.4, 145.3, 143.6, 138.4, 137.3, 132.8, 129.0, 128.8, 127.9, 127.0, 126.6, 126.4, 123.3, 118.6, 60.8, MS (EI, 70 eV) m/z: 306/304 (M +, 6/20). Anal. C15H13ClN2O3 (C, H, N). 6.1.48. 6-Chloro-7-methoxy-4-phenylisatin 48 3-Chloro-2-methoxy-5-phenylisonitrosoacetanilide 47 (6.5 g, 21.3 mmol) and polyphosphoric acid (66 g) were stirred at 80 C for 3 h when ice, water and ethyl acetate were added. The organic phase was washed with water, dried (Na2SO4) and evaporated. The remainder was puried by ash chromatography (SiO2, dichloromethane ethyl acetate-hexane 1:1) to give the title compound (5.06 g, 83%). 1H NMR (DMSO-d6) 11.54 (br s, 1H), 7.547.42 (m, 5H), 7.10 (s, 1H), 3.82 (s, 3H), 13C NMR (DMSO-d6) 181.8, 159.3, 145.4, 139.8, 138.0, 135.6, 134.9, 128.9, 128.8, 128.1, 124.4, 114.5, 61.2, MS (EI, 70 eV) m/z: 289/287 (M +, 27/100). Anal. C15H10ClNO3, (C, H, N). 6.1.49. 4-Chloro-3-methoxy-6-phenylanthranilic acid 49 To a solution of 48 (2.15 g, 7.47 mmol) in dioxane (20 mL) and sodium hydroxide (0.68 M, 60 mL) at 0 C

742 and under nitrogen atmosphere, was hydrogen peroxide (30%, 4.0 mL, 39.2 mmol) in sodium hydroxide (0.68 M, 50 mL) added drop-wise. The solution was stirred at ambient temperature for 6 h when acetic acid was added to pH 5. The mixture was extracted with ethyl acetate, the organic phase was dried (Na2SO4) and evaporated. The remainder was puried by ash chromatography (SiO2, ethyl acetate-hexane 1:2 toluene-ethyl acetate-acetic acid 4:1:1) to give the title compound (1.37 g, 66%). 1H NMR (DMSO-d6) 7.387.25 (m, 5H), 6.52 (s, 1H), 3.75 (s, 3H), 13C NMR (DMSO-d6) 169.3, 142.9, 141.6, 141.1, 139.0, 128.0, 127.9, 127.1, 117.5, 114.6, 59.4, MS (EI, 70 eV) m/z: 280/278 (M +, 15/47). Anal. C14H12ClNO3 (C, H, N). 6.1.50. 4-Chloro-3-hydroxy-6-phenylanthranilic acid 50 To 49 (110 mg, 0.40 mmol) dissolved in diethyl ether was ethereal hydrogen chloride added. The solvent was evaporated and dichloromethane (3 mL) was added. The solution was cooled to 70 C under nitrogen atmosphere and boron tribromide (0.08 mL, 0.84 mmol) was added. The solution was stirred for 40 min while the temperature reached ambient temperature. The mixture was left over night, and then sodium bicarbonate (sat, 15 mL) was added and the aqueous phase was stirred with dichloromethane for 1 h. Then the pH was lowered to 4 by the addition of hydrochloric acid (2 M), and the aqueous phase was extracted with dichloromethane, the organic phase was dried (MgSO4) and evaporated. The crude material was puried by ash chromatography (SiO2, toluene-ethyl acetate-acetic acid 4:1:1) affording the title compound (23 mg, 22%). 1H NMR (DMSO-d6) 7.337.22 (m, 5H), 6.52 (s, 1H), 13C NMR (DMSO-d6) 172.5, 143.6, 141.2, 141.1, 137.2, 129.6, 129.2, 128.0, 123.0, 119.8, 115.2, MS (EI, 70 eV) m/z: 265/263 (M +, 23/79). 6.1.51. 3-Chloro-2-methoxy-5-methylisonitrosoacetanilide 51 To a stirred solution of chloral hydrate (0.99 g, 6.0 mmol) and sodium sulfate (3.4 g, 24 mmol) in water (15 mL) was 3-chloro-2-methoxy-5-methylaniline [37] (0.51 g, 3.0 mmol) in N,N-dimethylformamide (7 mL), water (9 mL) and hydrochloric acid (conc., 0.37 mL) added. The mixture was stirred at 90 C for 25 min, when hydroxylamine hydrochloride (1.24 g, 18 mmol) was added and stirring was continued for another 18 h. The mixture was cooled and ethyl acetate (100 mL) was added. The organic phase was washed with water (100 mL), dried (Na2SO4) and evaporated. The remainder was puried by ash chromatography (SiO2, ethyl acetate-hexane 1:1) affording the title compound (0.62 g, 85%). 1H NMR (DMSO-d6) 9.36 (s, 1H), 7.90 (d, J = 1.6 Hz, 1H), 7.74 (s, 1H), 7.07 (d, J = 2.1 Hz, 1H), 3.75 (s, 3H), 2.25 (s, 3H), 13C NMR (DMSO-d6) 160.2, 143.6, 134.8, 132.1, 125.9, 125.5, 120.8, 60.7, 20.5, MS (TSP) m/z: 262/260 (M + 18, 30/100). Anal. C10H11ClN2O3 (C, H, N). 6.1.52. 6-Chloro-7-methoxy-4-methylisatin 52 3-Chloro-2-methoxy-5-methylisonitrosoacetanilide 51 (0.49 g, 2.0 mmol) and polyphosphoric acid (6.3 g) were stirred at 80 C for 3 h when ice was added. Ethyl acetate (150 mL) was added and the organic phase was washed with water (50 mL), dried (Na2SO4) and evaporated. The remainder was puried by ash chromatography (SiO2, ethyl acetate-hexane 1:2 ethyl acetate) to give the title compound (0.37 g, 82%). 1H NMR (DMSO-d6) 11.41 (br, 1H), 7.00 (s, 1H), 3.74 (s, 3H), 2.38 (s, 3H), 13C NMR (DMSO-d6) 183.8, 159.4, 144.5, 138.6, 136.0, 135.8, 125.1, 116.1, 61.1, 16.6, MS (TSP) m/z: 245/243 (M + 18, 37/100). Anal. C10H8ClNO3 (C, H, N). 6.1.53. 4-Chloro-3-methoxy-6-methylanthranilic acid 53 To a solution of 52 (320 mg, 1.4 mmol) in dioxane (5 mL) at 5 C and under nitrogen atmosphere was hydrogen peroxide (30%, 0.73 mL, 7.2 mmol) in water (10 mL) and sodium hydroxide (0.67 M, 26 mL) added drop-wise. The mixture was stirred for 3 h while the temperature slowly reached ambient temperature. Water (25 mL) was added and the pH was adjusted to 5 by the addition of acetic acid. The mixture was extracted with ethyl acetate (120 mL) and the organic phase was dried (Na2SO4) and evaporated. The remainder was puried by ash chromatography (SiO2, ethyl acetate ethyl acetate-acetic acid 1 000:1) affording the title compound (210 mg, 70%). 1H NMR (CD3OD) 6.45 (s, 1H), 3.76 (s, 3H), 2.37 (s, 3H), 13C NMR (CD3OD) 171.9, 146.4, 142.9, 138.3, 131.0, 120.3, 114.5, 60.2, 23.2, MS (TSP) m/z: 218/216 (M + 1, 34/100). Anal. C9H10ClNO3 (C, H, N). 6.1.54. 4-Chloro-3-hydroxy-6-methylanthranilic acid 54 To 53 (49 mg, 0.23 mmol) in diethyl ether was ethereal hydrogen chloride (3 M) added. The solvent was evaporated and dichloromethane (2 mL) was added. The solution was cooled to 65 C under nitrogen atmosphere when boron tribromide (0.1 mL, 1.12 mmol) was added. The reaction was stirred for 43 h while the temperature reached ambient temperature. Sodium bicarbonate (sat., 15 mL) was added, and the pH of the aqueous phase was adjusted to 2 by the addition of hydrochloric acid (2 M). The aqueous phase was extracted with ethyl acetate, and the organic phase was dried (Na2SO4) and evaporated. The remainder was puried by ash chromatography

743 (SiO2, toluene-acetic acid 10:1 5:1) affording the title compound (22 mg, 47%). 1H NMR (CD3OD) 6.43 (s, 1H), 2.35 (s, 3H), 13C NMR (CD3OD) 172.2, 142.9, 139.9, 133.4, 123.8, 119.8, 113.8, 22.9, MS (TSP) m/z: 204/202 (M + 1, 28/100). Anal. C8H8ClNO3 (C, H, N). 6.2. Stability assay The compounds were dissolved in PBS buffer at pH 7.5 and the solution was left in the dark for 24 h at 37 C. Aliquots were analysed by HPLC (Bondapak C18, 150 3.9, phosphate buffer pH 3-acetonitrile-acetic acid 75:25:1) with UV detection (254 nm). 6.3. Pharmacological methods 6.3.1. In vitro measurement of 3-HAO inhibition Male Sprague-Dawley rats (150200 g) were anaesthetized and perfused before decapitation. The cortex was rapidly dissected out on ice and stored at 70 C. To prepare a homogenate, the cortex was thawed, minced and sonicated in 9 volumes (w/v) of ice-cold distilled water by 3 times 5 s bursts of sonication using a cell disrupter (Branson sonier). Aliquots of the crude homogenate were stored in vials (400 L/vial) at 70 C prior to assay. To prepare cell-free homogenate, the vials were thawed in an ice-bath, and to each vial, 600 L 150 mM (N-morpholino)-2-ethanesulfonic acid (MES)/NaOH buffer (pH 6.5), was added. The vials were then vortexed and centrifuged at 15 000 g for 10 min at 4 C. 100 L of the supernatant was incubated at 37 C under gently shaking in a water bath for 1 h in a solution containing FeSO4 (0.3 mM), ascorbic acid (aq, 0.1%) and [14C]3HANA (specic activity of 5.5 mCi/mmol, 5 M, Astra Arcus AB, nal volume 200 L). Test compounds were added in a volume of 10 L prior to the substrate. The incubation was terminated by addition of HClO4 (aq., 6%, 50 L), the tubes cooled on ice and the precipitate removed by centrifugation at 10 000 g for 4 min at 4 C. An aliquot (230 L) of the supernatant was applied to a Dowex 50W (200400 mesh) cation-exchange column (0.5 2 cm), which was thereafter washed with distilled water (1 mL) to elute the [14C]-QUIN. Scintillation uid (10 mL, Beckman, Ultima Gold) was added to the eluate, and the radioactivity was determined by liquid scintillation spectrophotometry. 6.3.2. In vivo measurement of 3-HAO inhibition Male Sprague-Dawley rats (150200 g) were acclimatized to the animal quarters for at least 7 d before initiation of the experiments. The rats were housed ve animals per cage under controlled conditions of temperature (21 C), relative humidity (5565%) and light-dark cycle (12:12 h, lights on 6 a.m.). Food and tap water were available ad libitum. The rats were anaesthetized with enurane (Efrane; Abbott) in a ow mixture of O2 and N2O. They were thereafter placed in a stereotaxic frame. The anaesthesia was maintained by free breathing into a mask tted over the nose of the rat of 3.55.0% of enurane maintained at a ow of 3 L/min of O2 and 7 L/min of N2O. Body temperature was kept at 37 C using a heating pad controlled via a rectal thermometer (CMA/12). The skull was oriented with the horizontal plane passing through bregma and lambda and through the intraural line. The skin over the skull was opened and a unilateral hole on the right side of the skull bone was made by drilling with a 1 mm burr. A Hamilton microlitre syringe (gauge 22S, 25 L) was lowered 3.9 mm vertically from the surface of the brain at AP 0 mm (bregma) and L 1.0 mm. An intracerebroventricular (i.c.v.) injection of 3, 10, or 300 nmol of test compound together with 10 nmol 3-HANA dissolved in 10 l 10 mM phosphate buffer, containing 105 mM NaCl, 2.5 mM KCl, 1.18 mM MgCl2, 1.26 mM CaCl2 and 0.1% ascorbic acid was performed for 1 min. The syringe was then removed and the skin was closed by wound clips. At various timepoints after the i.c.v. injection, the rats were decapitated, and their brains rapidly removed and placed on an ice-chilled petri-dish. The right hippocampus was dissected out and stored at 70 C until QUIN determinations. For QUIN determination, the samples were homogenized in 10 volumes (w/w) 0.3 M formic acid with [18O]4-quinolinic acid added as internal standard. After ltration, the samples were applied to a pretreated anionexchanger column (0.5 mL Bio-Rad AG 1 8, 200400 mesh, formate form). The column was washed with water (4 mL) before use. QUIN and internal standard were eluted with 3 mL 6.0 M formic acid. After evaporation, the samples were esteried at 70 C with hexauoroisopropanol and triuoroacetylimidazole for 1 h. The esters were partitioned between water (0.4 mL) and n-heptane (3 mL) by whirlimixing for 1 min followed by centrifugation at 1 000 g for 5 min. The organic phase was transferred to a new tube and washed with 0.4 mL of water by whirlimixing. After a second centrifugation, the water was discarded and the samples evaporated in a stream of nitrogen at room temperature to about 5 mL. Two L of the nal solution was assayed using gas chromatography coupled to a mass spectrometer (GC/MS). The GC (Carlo Erba MFC 500) was operated in splitless injection mode, and a capillary column (NB-225, 25 m, i.d. 0.32 mm) was used. Helium was employed as carrier gas. The oven was run from 70135 C, with a ramp of 20 C/min. The MS (VG,

744 TS-250) was adjusted to record m/e 467 (QUIN) and m/e 471 (internal standard) by negative chemical ionization. Methane was used as reagent gas. The peak height ratio (QUIN/internal standard) from standards with known amounts of QUIN was calculated. The calibration curve was plotted as peak height ratio against concentration of QUIN in the standards. QUIN amount in the samples was then calculated from the standard curve. Acknowledgements The authors are grateful for the analysis of QUIN by Mr Gran Fredriksson, for surgical assistance by Ms Eva Vnerman and to Mr Gran Stening for skillful technical assistance. 4-Halogenated 3-HANA analogues were generously provided by Dr Barry Carpenter, Cornell University. References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Bokman A.H., Schweigert B.S., Arch. Biochem. Biophys. 33 (1951) 270276. Roberts R.C., McCarthy K.E., Du F., Ottersen P., Okuno E., Schwarcz R., J. Neurosci. 15 (1995) 11501161. Heyes M.P., Chen C.Y., Major E.O., Saito K., Biochem. J. 326 (1997) 351356. Alberati-Giani D., Ricciardi-Castagnoli P., Khler C., Cesura A.M., J. Neurochem. 66 (1996) 9961004. Saito K., Markey S.P., Heyes M.P., Brain Res. 546 (1991) 151154. Saito K., Chen C.Y., Masana M., Crowley J.S., Markey S.P., Heyes M.P., Biochem. J. 291 (1993) 1114. Saito K., Markey S.P., Heyes M.P., Neurosci. Lett. 178 (1994) 211215. Heyes M.P., Saito K., Major E.O., Milstien S., Markey S.P., Vickers J.H., Brain 116 (1993) 14251450. Blight R.T., Cohen T.I., Saito K., Heyes M.P., Brain 118 (1995) 735752. Okuno E., Khler C., Schwarcz R., J. Neurochem. 49 (1987) 771780. Malherbe P., Khler C., Da Prada M., Lang G., Kiefer V., Schwarcz R., Lahm H.W., Cesura A.M., J. Biol. Chem. 269 (1994) 1379213797. [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] Stone T.W., Pharmacol. Rev. 45 (1993) 309379. Lapin I.P., J. Neuronal Trans. 42 (1978) 3743. Schwarcz R., Whetsell W.O. Jr., Mangano R.M., Science 219 (1983) 316318. Beal M.F., Ferrante R.J., Swartz K.J., Kowall N.W., J. Neurosci. 11 (1991) 16491659. Huang Q., Zhou D., Sapp E., Aizawa H., Ge P., Bird E.D., Vonsattel J.P., Di Figlia M., Neuroscience 65 (1995) 397407. Tatter S.B., Galpern W.R., Hoogeveen A.T., Isacson O., Neuroreport 6 (1995) 11251129. Foster A.C., White B.J., Schwarcz R., J. Neurochem. 47 (1986) 2230. Heyes M.P., Jordan E.K., Lee K., Saito K., Frank J.A., Snoy P.J., Markey S.P., Gravell M., Brain Res. 570 (1992) 237250. Halperin J.J., Heyes M.P., Neurology 42 (1992) 4350. Eastman C.L., Urbnska E., Lve A., Kristensson K., Schwarcz R., Exper. Neurol. 125 (1994) 119124. Heyes M., Saito K., Major E.O., Milstien S., Markey S.P., Vickers J.H., Brain 116 (1993) 14251450. Heyes M.P., Saito K., Milstien S., Schiff S.J., J. Neurol. Sciences 133 (1995) 112118. Todd W.P., Carpenter B.K., Schwarcz R., Prep. Biochem. 19 (1989) 155165. Manthey M.K., Pyne S.G., Truscott R.J.W., J. Org. Chem. 53 (1988) 14861488. Manthey M.K., Pyne S.G., Truscott R.J.W., Biochim. Biophys. Acta 1034 (1990) 207212. Hansch C., Leo A., Hoekman D., Exploring QSAR: Hydrophobic, Electronic and Steric Constants, ACS, Washington DC, 1995. Pallas CompDrug Chemistry Ltd. Dewar M.J.S., Zoenisch E.G., Healy E.F., Stewart J.J.P., J. Am. Chem. Soc. 107 (1985) 39023909. Ovality is dened as ovality = A/4p(3V/4p) 2/3, where A is the molecular surface area and V is the molecular volume. Cannon J.R., Cresp T.M., Metcalf B.W., Sargent M.V., Vinciguerra G., J. Chem. Soc. C (1971) 34953504. Melikian A., Boigegrain R., Kan J.P., Soubrie P., Eur. J. Med. Chem./Chim. Ther. 25 (1990) 267270. Calam C.T., Oxford A.E., J. Chem. Soc. (1939) 280284. Schmidt U., Boekens H., Lieberknecht A., Griesser H., Tetrahedron Lett. 22 (1981) 49494952. Castelfranchi G., Perrotti E., Ann. Chim. 47 (1957) 12011214. Colbert J.C., Meigs W., Stuerke A.H., J. Amer. Chem. Soc. 56 (1934) 21282130. Watanabe Y., Nakajima K., Seki T., Ozawa H., Chem. Pharm. Bull. 18 (1970) 22082215.

Eur. J. Med. Chem. 34 (1999) 745751 1999 Editions scientiques et mdicales Elsevier SAS. All rights reserved

745

Short communication

Phenylsulfonylnitromethanes as potent irreversible inhibitors of aldose reductase

Nada H. Saaba, Isaac O. Donkora, Libaniel Rodriguezb, Peter F. Kadorb, Duane D. Millera*
b

Department of Pharmaceutical Sciences, The University of Tennessee, Memphis, TN 38163, USA Laboratory of Ocular Therapeutics, National Eye Institute, National Institutes of Health, Bethesda, MD 20892, USA (Received 27 October 1998; revised 4 March 1999; accepted 4 March 1999)

Abstract Aldose reductase (AR) inhibition provides a viable pharmacologically direct mode for the treatment of diabetic complications. We have synthesized a series of N-4 substituted analogues (1521) of the known aldose reductase inhibitor phenyl-sulfonylnitromethane. The compounds are potent inhibitors of AR with IC50s between 0.01 and 0.19 M. Some of the compounds are also potent affinity labels for AR. Compound 19 exhibits the highest and almost complete irreversible inhibition of AR known to date. 1999 Editions scientiques et mdicales Elsevier SAS aldose reductase / diabetic complication / phenylsulphonyl-nitromethane / irreversible inhibitors / aldose reductase inhibitors

1. Introduction Over 50% of diabetics develop tissue-damaging complications [1]. These complications develop in tissues capable of insulin-independent glucose uptake and result in retinopathy, nephropathy, cataract, keratopathy, neuropathy and angiopathy. Results of the recent Diabetes Control and Complications Trial suggest that tight control of blood sugar levels can reduce the incidence and severity of diabetic complications [1]; however, on a practical basis, tight control is difficult to maintain. This has spurred efforts toward the development of alternative treatments with agents acting by mechanisms independent of the control of blood glucose. Aldose reductase inhibitors (ARIs) have provided therapeutically useful agents [2], which in long-term animal studies demonstrate benecial prevention or delay in the onset and progression of diabetic complications with no signicant adverse effects [3]. The search for clinically useful ARIs has been an on-going process since the late 1960s. This effort has led to the discovery of a number of structurally diverse compounds as ARIs (gure 1). Kinetic studies suggest that despite their structural variations, ARIs exhibit either uncompetitive or noncom-

petitive inhibition and bind to a site on the enzyme which is independent of the substrate and NADPH cofactor binding sites [46]. Using affinity-labelling studies, Kador et al. [7] demonstrated that there are three distinct binding sites on the AR enzyme, namely, the substrate site, cofactor site and inhibitor site. This observation is, however, at odds with recent X-ray crystallographic studies involving the ternary complex between AR, NADPH and zopolrestat in which the inhibitor was found completely sequestered into the substrate site [8]. In an effort to study the nature of the inhibitor binding site of AR, we previously reported a series of affinity labels (710) (table I) based on the reversible inhibitor alrestatin [9]. Based on the remarkable irreversible inhibitory activity of 5-iodoacetamidoalrestatin (10) in our previous studies [7, 9] and the possibility of using affinity labels to locate the binding site(s) of ARIs on AR, we synthesized affinity labels (1517) and (1921) along with their known acetylated derivatives (14 and 18) (table I), derived from a new class (sulfonylnitromethanes) of potent AR inhibitors, and examined the compounds for AR inhibitory activity. 2. Chemistry Sulfonylnitromethanes have been synthesized by three principal synthetic routes. These include the sulfonation

*Correspondence and reprints

746

Figure 1. Aldose reductase inhibitors.

of -halo nitro compounds [10], the oxidation of -nitro sulde [11] and the alkaline nitration of sulfones [12]. Alkaline nitration, which is based on a free radical chain process rather than nucleophilic displacement of iodide by the nitro anion, was the most fruitful method for our synthesis. The synthesis of affinity labels 1517 and 1921 was achieved as depicted in gure 2. The amino moiety of dimethylaniline (22) was protected by treatment with acetic anhydride to give 23, which was reacted with chlorosulfonic acid to yield the corresponding sulfonyl chloride 25. Compound 25 as well as the commercially available N-acetylsulfonyl chloride (24) were reduced using sodium sulte in the presence of sodium bicarbonate buffer to give the corresponding sulnic acids 26 and 27 [13]. The acids were converted to their sodium salts and then reacted with two-fold excess of nitromethane in the presence of iodine and sodium methoxide at 0 C to give the corresponding N-(acetylphenyl)sulfonyl nitromethane derivatives 14 and 18 [13, 14]. Basic hydrolysis of the acetyl group afforded 12 and 13. The latter compounds were treated with thiophosgene in acetone to give the corresponding isothiocyanate analogues 15 and 19, respectively. Treatment of 12 or 13 with chloro- or iodoacetic anhydride afforded the corresponding chloro- or iodoacetamido derivatives 16 and 20 or 17 and 21, respectively.

3. Results and discussion The sulfonylnitromethane analogues were found to be potent inhibitors of recombinant rat lens AR with IC50s ranging between 0.01 to 0.19 M (table I). Subsequent gel ltration studies for irreversible binding also indicated that compounds 1517 and 1921 are potent affinity labels for AR (table I). Comparison of the activities of compounds 1517 with that of compounds 1921 indicates that ortho dimethyl substitution potentiates irreversible inhibition of AR in this class of compounds. In contrast to the affinity labels (710, table I) derived from alrestatin, the irreversible inhibitory activity for the present compounds is not directly linked to the chemical reactivity of their electrophilic groups. Irreversible inhibitory activity for both sets of compounds decreased in order: isothiocyanato analogues > iodoacetamido analogues > chloroacetamido analogues. No correlation between reversible and irreversible inhibitory activities was observed. Results from mass spectrometry and molecular modelling studies [7], carboxymethylation studies [15], site directed mutagenesis studies [16], as well as kinetic data [46] suggest the possibility of multiple inhibitor binding sites on AR which are distinct from the substrate binding site. Structurally diverse, selective affinity labels

747
Table I. Reversible (IC50) and irreversible (%) inhibition of recombinant rat lens AR.

# 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
a b c

R1

R2

Ra NCS NHCOCH2Cl NHCOCH2Br NHCOCH2I NDb 0.55 0.04 0.60 0.05 0.40 0.02 NDb 0.19 0.66c 0.13 0.04c 0.09 0.09 0.05 0.05 0.06 0.07 0.09 0.03 0.09 0.09 0.13 0.03 0.01 0.04 0.03 0.001 56.0 0.0 46.0 89.0 NDb 0.0 0.0 0.0 80.7 58.0 73.8 0.0 99.5 86.6 96.8

H NH2 NH2 NHCOCH3 NCS NHCOCH2Cl NHCOCH2l NHCOCH3 NCS NHCOCH2Cl NHCOCH2l

CH3 H CH3 H H H H CH3 CH3 CH3 CH3

The alrestatin derivatives were previously reported [9]. ND, not determined. Reported values from [13, 14].

can be useful tools for investigating the possibility of multiple inhibitor sites on AR. The present results indicate that compounds 19 and 21 are more potent affinity labels than 5-iodoacetamidoalrestatin (10) and that apparently complete irreversible inhibition of AR can be achieved with compound 19. Future determination of the exact site(s) of interaction with the enzyme by these compounds and identication of the reactive nucleophilic residue(s) that bind to these compounds will provide insight into the possibility of multiple inhibitor binding sites on AR. 4. Experimental protocols 4.1. Chemistry Melting points were determined on a Thomas-Hoover capillary melting point apparatus and are uncorrected. Infrared (IR) spectra were recorded on a Perkin Elmer System 2000 FT-IR spectrophotometer. Proton nuclear magnetic resonance (1H NMR) spectra were recorded on a Bruker AX 300 spectrometer. Chemical shift values are

reported in parts per million () relative to tetramethylsilane (TMS) as an internal standard. Spectral data are consistent with assigned structures. Elemental analyses were performed by Atlantic Microlab Inc., Norcross GA, and experimentally determined values are within 0.4% of the theoretical values. Routine thin-layer chromatography (TLC) was performed on silica gel GHIF plates (Analtech Inc., Newark DE). Flash chromatography was performed on silica gel (Merck, grade 60, 230400 mesh, 60 ). Acetonitrile (MeCN) was dried by distillation, dimethylformamide (DMF) was dried by distillation from P2O5. All solvents (except anhydrous MeCN and DMF) were stored over 3 or 4 molecular sieves. 4.1.1. N-Acetyl-3,5-dimethylaniline 23 Acetic anhydride (60 mL) was added slowly to 3,5dimethylaniline (40 mL, 0.32 mol). The hot reaction mixture was allowed to cool to room temperature to give 42 g (80%) of 23 as a white solid: m.p. 137139 C (Lit. [13] 138 C). IR (KBr, cm1) 1 663 (C=O). 1H NMR (CDCl3) 7.12 (s, 2H, aromatic), 6.75 (s, 1H, aromatic), 2.28 (s, 6H, 2CH3), 2.15 (s, 3H, COCH3), 1.63 (br, 1H, NHCO).

748

Figure 2. Synthetic route to phenylsulfonylnitromethanes.

749 4.1.2. (4-Acetamido-2,6-dimethylphenyl)sulfonyl chloride 25 In a three-necked round bottom ask tted with a mechanical stirrer was placed chlorosulfonic acid (100 g, 0.85 mol). The ask was cooled (1215 C) by means of an ice-bath and N-acetyl-3,5-dimethylaniline (28.07 g, 0.172 mol) was added slowly over a 15 min period. After the addition was complete, the mixture was allowed to warm to room temperature and then was heated to 60 C with stirring. After 2.5 h, the mixture was added slowly to ice-water mixture (330 mL) with stirring. The off-white precipitate that formed was ltered and dried in a desiccator under vacuum overnight to give 22 g of crude product. An analytical sample was recrystallized from benzene: m.p. 157159 C. IR (KBr, cm1) 3 322 (NH), 1 683 (C=O), 1 360 (SO2) and 1 172 (SO2). 1H NMR (CDCl3) 7.65 (br, 1H, NH), 7.42 (s, 2H, aromatic), 2.72 (s, 6H, 2CH3), 2.22 (s, 3H, COCH3). 4.1.3. (4-Acetamidophenyl)sulnic acid 26 N-Acetylsulfanilyl chloride (11.9 g, 0.05 mol) was added to a vigorously stirred solution of sodium bicarbonate (10 g, 0.119 mol) and sodium sulte (14.29 g, 0.113 mol) in water (60 mL) at 7080 C. The mixture was heated for 1 h and allowed to cool to room temperature. The white precipitate that separated out was redissolved in water and acidied with 60% sulfuric acid until a white precipitate appeared. The suspension was left in a refrigerator overnight. The solid was recovered by ltration and recrystallized from water to give 7.23 g (72%) of 26 as white needle-like crystals: m.p. 149151 C. IR (KBr, cm1) 3 321 (NH), 1 668 (C=O), 1 320 (SO2), 1 084 (SO2). 1HNMR (DMSO-d6) 10.20 (s, 1H, NHCO), 7.73 (d, 2H, aromatic), 7.58 (d, 2H, aromatic), 2.06 (s, 3H, COCH3). 4.1.4. (4-Acetamido-2,6-dimethylphenyl)sulnic acid 27 Sulfonylchloride 25 (22 g, 0.084 mol) was transformed to sulnic acid 27 as described above for the synthesis of 26. The crude product was recrystallized from water to give 5.5 g (29%) of 27 as white crystals: m.p. 158159 C. IR (KBr, cm1) 3 321 (NH), 1 668 (C=O), 1 320 (SO2), 1 084 (SO2). 1H NMR (CDCl3) 7.31 (s, 2H, aromatic), 2.59 (s, 6H, 2CH3), 2.10 (s, 3H, COCH3). 4.1.5. (4-Acetamidophenyl)sulfonylnitromethane 14 Sodium metal (0.62 g, 0.027 mol) was dissolved in dry MeOH and 26 (5.5 g, 0.027 mol) was added and the mixture was stirred overnight. The solvent was removed under reduced pressure and the residue was kept in a desiccator. Another sample of sodium metal (1.09 g, 0.047 mol) was dissolved in dry MeOH, the solvent was stripped off under reduced pressure, and the residue was dissolved in dry DMF (15 mL) and cooled to 0 C. Nitromethane (3.22 g, 0.052 mol) was dissolved in dry DMF (15 mL) and added slowly to the cooled NaOMe solution. The reaction mixture was stirred at 0 C for 15 min during which period a bright yellow precipitate separated out. The previously prepared sodium salt of 26 was dissolved in dry DMF, cooled to 0 C, and added to the bright yellow mixture followed immediately by the addition of iodine (6.08 g, 0.024 mol). The reaction was stirred at 0 C for 4 h and then allowed to warm to room temperature over 30 min. The dark solution was poured into ice-water and decolourized with sodium sulte. It was acidied slowly to pH 1.5 with 2 N HCl and the resulting suspension was stored in a refrigerator overnight followed by ltration to recover the product which was dried and recrystallized from MeOH to give a 63% yield of 14: m.p. 220222 C (dec.) (Lit. [14] 228229 C). IR (KBr, cm1) 3 259 (NH), 1 674 (C=O), 1 590 (NO2), 1 553 (NO2). 1H NMR (DMSO-d6) 10.54 (s, 1H, NHCO), 7.90 (s, 4H, aromatic), 6.57 (s, 2H, CH2NO2), 2.13 (s, 3H, COCH3). 4.1.6. (4-Acetamido-2,6-dimethylphenyl)sulfonylnitromethane 18 Compound 27 (2 g, 8.8 mmol) was transformed into 18 as described above for the synthesis of 14. The off-white solid obtained was puried by column chromatography over silica gel with hexane/ethyl acetate (1:2) as the eluant followed by recrystallization from ethanol to yield white crystals: m.p. 177179 C (Lit. [13] 179180 C). IR (KBr, cm1) 3 308 (NH), 1 677 (C=O), 1 534 (NO2). 1 H NMR (acetone-d6) 9.50 (s, 1H, NHCO), 7.59 (s, 2H, aromatic), 6.15 (s, 2H, CH2NO2), 2.60 (s, 6H, 2CH3), 2.05 (s, 3H, COCH3). 4.1.7. (4-Aminophenyl)sulfonylnitromethane 12 A solution of 14 (0.2 g, 0.77 mmol) in 2 N NaOH (2 mL) was heated for 1 h at 80 C. The mixture was then poured into ice-water mixture (10 mL) containing acetic acid (0.3 mL) to yield a solid which was extracted with EtOAc, dried (MgSO4), and evaporated under reduced pressure. The residue was puried by column chromatography over silica gel with hexane/EtOAc (1:1) as the eluant followed by recrystallization from EtOH to give a 78% yield of 12: m.p. 128130 C. IR (KBr, cm1) 3 396 (NH2), 1 554 (NO2). 1H NMR (DMSO-d6) 7.51 (s, 2H, aromatic), 6.67 (s, 2H, aromatic), 6.44 (s, 2H, CH2NO2), 6.30 (s, 2H, Ar-NH2). 4.1.8. (4-Amino-2,6-dimethylphenyl)sulfonylnitromethane 13 Compound 18 (0.146 g, 0.51 mmol) was transformed into 13 in 81% yield as described above for the synthesis of 12: m.p. 131133 C (Lit. [13] 132133 C). IR (KBr,

750 cm1) 3 474 (NH2), 3 376 (NH2), 1 551 (NO2). 1H NMR (DMSO-d6) 7.54 (s, 2H, aromatic), 6.40 (s, 2H, CH2NO2), 2.62 (s, 6H, 2CH3), 6.32 (s, 2H, Ar-NH2). 4.1.9. (4-Isothiocyanatophenyl)sulfonylnitromethane 15 Thiophosgene (12 drops) was added to a solution of 12 (0.06 g, 0.28 mmol) in dry acetone (4.5 mL) and the mixture was stirred at room temperature for 2 h. The solvent was removed under reduced pressure and the crude product was puried by column chromatography over silica gel with hexane/EtOAc (1:1) as the eluant to yield 0.06 g (87%) of 15 as a bright yellow solid: m.p. 159160 C. IR (KBr, cm1) 2 131 (NCS). 1H NMR (acetone-d6) 8.09 (d, 2H, aromatic), 7.72 (d, 2H, aromatic), 6.35 (s, 2H, CH2NO2). MS (EI, m/z) 258+. Anal. C8H6N2S2O4 (C, H, N, O, S). 4.1.10. (4-Isothiocyanato-2,6-dimethylphenyl)sulfonylnitromethane 19 Thiophosgene (8 drops) was added to a solution of 13 (0.06 g, 2.4 mmol) in dry acetone (3 mL) and the mixture was stirred at room temperature for 2 h. The solvent was removed under reduced pressure and the crude product was puried by column chromatography over silica gel with hexane/EtOAc (1:1) as the eluant to yield 0.065 g (93%) of 19 as a white solid: m.p. 110112 C. IR (KBr, cm1) 2 009 (NCS). 1H NMR (acetone-d6) 7.35 (s, 2H, aromatic), 6.27 (s, 2H, CH2NO2), 2.68 (s, 6H, 2CH3). MS (EI, m/z) 286+. Anal. C10H10N2S2O4 (C, H, N, O, S). 4.1.11. (4-Chloroacetamidophenyl)sulfonylnitromethane 16 Chloroacetic anhydride (73 mg, 0.43 mmol) was added to a solution of 12 (0.06 g, 0.27 mmol) in dry CH3CN (4 mL) and the mixture was stirred at room temperature for 18 h. The solvent was concentrated to precipitate pure 16 in 0.044 g (55%) yield as a white solid: m.p. 212 C. IR (KBr, cm1) 3 279 (NH), 1 684 (C=O). 1H NMR (DMSO-d6) 10.86 (s, 1H, NHCO), 7.92 (d, 4H, aromatic), 6.60 (s, 2H, CH2NO2), 4.34 (s, 2H, CH2Cl). MS (EI, m/z) 292 +. Anal. C9H9N2SO5Cl (C, H, N, O, S, Cl). 4.1.12. (4-Chloroacetamido-2,6-dimethylphenyl)sulfonylnitromethane 20 Compound 13 (0.05 g, 2.0 mmol) was treated as described for the synthesis of 16 to yield 0.06 g (87%) of 20 as a bright yellow solid: m.p. 160161 C. IR (KBr, cm1) 3 366 (NH), 1 723 (C=O). 1H NMR (DMSO-d6) 10.66 (s, 1H, NHCO), 7.53 (s, 2H, aromatic), 6.50 (s, 2H, CH2NO2), 4.31 (s, 2H, CH2Cl), 2.56 (s, 6H, 2CH3). MS (EI, m/z) 320+. Anal. C11H13N2SO5Cl (C, H, N, O, S, Cl). 4.1.13. (4-Iodoacetamidophenyl)sulfonylnitromethane 17 Iodoacetic anhydride (0.106 g, 0.29 mmol) was added to a solution of 12 (0.05 g, 0.23 mmol) in dry CH3CN (4 mL) and the mixture was stirred at room temperature for 18 h. The mixture was concentrated and CHCl3 was added and kept in a refrigerator overnight to precipitate pure 17 as pale yellow crystals in 95% yield: m.p. 165 C. IR (KBr, cm1) 3 356 (NH), 1 690 (C=O). 1H NMR (DMSO-d6) 10.88 (s, 1H, NHCO), 7.50 (d, 4H, aromatic), 6.61 (s, 2H, CH2NO2), 3.87 (s, 2H, CH2I). MS (EI, m/z) 384+. Anal. C9H9N2SO5I (C, H, N, O, S, I). 4.1.14. (4-Iodoacetamido-2,6-dimethylphenyl)sulfonylnitromethane 21 This compound was synthesized as described for the synthesis of 17 by reacting iodoacetic anhydride (0.094 g, 0.26 mmol) with 13 (0.05 g, 2.0 mmol). Compound 21 was obtained in 95% yield as white crystals. M.p. 200 C. IR (KBr, cm1) 3 366 (NH), 1 723 (C=O). 1H NMR (DMSO-d6) 10.68 (s, 1H, NHCO), 7.50 (s, 2H, aromatic), 6.49 (s, 2H, CH2NO2), 3.84 (s, 2H, CH2I), 2.51 (s, 6H, 2CH3). Anal. C11H13N2SO5I (C, H, N, O, S, I). 4.2. Biological studies 4.2.1. Enzyme purication Recombinant rat lens AR was puried by a series of chromatographic procedures as previously described [17]. Briey, AR was released from E. coli by sonication and the mixture was centrifuged at 10 000 g for 10 min. The supernatant was then subjected to gel ltration on a Sephadex G-75 column (2.5 90 cm), equilibrated with 10 mM imidazole-HCl buffer, pH 7.5 containing 7 mM 2-mercaptoethanol and eluted with the same imidazole buffer. The eluent was collected in 220-drop aliquots (ca. 10 mL) and fractions containing AR activity were applied to a Matrex Gel Orange A affinity column (2.5 15 cm). The affinity column was washed with the imidazole buffer (ca. 500 mL) and the enzyme was eluted with the same imidazole buffer containing 0.1 mM NADPH. Fractions eluted with NADPH were chromatofocused on a Mono P (HR 5/20) column developed at a ow rate of 1 mL/min with Polybuffer 74 (diluted 10-fold and containing 7 mM 2-mercaptoethanol). The protein concentration of the eluent was monitored at 280 nm and peaks containing AR activity were collected and concentrated on Centricon 10 lters. 4.2.2. Enzyme assay Reductase activity was spectrophotometrically assayed on a Shimadzu UV 2100U spectrophotometer by following the decrease in the absorption of NADPH at 340 nm

751 over a 4-min period with DL-glyceraldehyde as substrate [9]. Each 1 mL cuvette contained equal units of enzyme, 0.10 M Na, K phosphate buffer, pH 6.2, 0.3 mM NADPH with/without 10 mM substrate and inhibitor. Appropriate controls were employed to negate potential changes in the absorption of nucleotide and/or protein modication reagents or ARIs at 340 nm in the absence of substrate. 4.2.3. Affnity binding Rat lens AR was passed through NAP-5 desalting columns equilibrated with 0.1 M sodium phosphate buffer, pH 7.6, to remove potentially reactive 2-mercaptoethanol which stabilizes AR [9]. Equal aliquots of enzyme were then combined with either reversible or irreversible inhibitor dissolved in 0.1 M sodium phosphate buffer, pH 7.6, and the reaction was allowed to proceed at room temperature for 15 min. Reversible and unreacted inhibitor were then removed from the protein by gel ltration with a NAP-5 desalting column with 0.1 M phosphate buffer, pH 7.0, containing 10 mM 2-mercaptoethanol. Residual AR activity was spectrophotometrically measured. All experiments were conducted at least in triplicate. Acknowledgements This work was supported in part by NIH grant 1 R15 EY0936-02 (I.O.D). We thank Mr John Miller for running the Mass Spectra of the compounds. References
[1] [2] [3] [4] [5] [6] [7] Diabetes Control and Complications Trial Research Group, New Engl. J. Med. 329 (1993) 977986. Tomlinson D.R., Stevens E.J., Diemel L.T., Trends Pharm. Sci. (England) 15 (1994) 293297. Kador P.F., Robinson W.G. Jr., Kinoshita J.H., Ann. Rev. Pharmacol. Toxicol. 25 (1985) 691714. Kador P.F., Sharpless N.E., Mol. Pharmacol. 24 (1983) 521531. Okuda J., Miwa I., Inagaki K., Horie T., Nakayama M., Biochem. Pharmacol. 31 (1982) 38073822. Kador P.F., Nakayama T., Sato S., Smar M., Miller D.D., Enzymol. Mol. Bio. Carbonyl Metabolism 2 (1989) 237250. Kador P.F., Lee Y.S., Rodriquez L., Sato S., Bartoszko-Malik A., Abdel-Ghany Y.S., Miller D.D., Bioorg. Med. Chem. 3 (1995) 13131324. Wilson K.D., Tarle I., Petrash M.J., Quiocho F.A., Proc. Natl. Acad. Sci. USA 90 (1993) 98479851. Smar M.W., Ares J.J., Nakayama T., Itabe H., Kador P.F., Miller D.D., J. Med. Chem. 35 (1992) 11171120. Troger J., Notle E., J. Prakt. Chem. 101 (1920) 136. Kharasch N., Cameron J.L., J. Am. Chem. Soc. 75 (1953) 10771081. Truce W.E., Klingler T.C., Paar J.E., Feuer H., Wu D.K., J. Org. Chem. 34 (1969) 31043107. Brittain D.R., Brown S.P., Cooper A.L., Longridge J.L., Morris J.J., Preston J., Slater L., UK Patent Application (GB 2227745), 1990. Kelley J.L., McLean E.W., Williard K.F., J. Hetero. Chem. 14 (1977) 14151416. Liu S.Q., Bhatnagar A., Srivastavs S.K., Biochem. Biophy. Acta. 1120 (1992) 329336. Bohren K.M., Page J.L., Shankar R., Henry S.P., Gabby K.H., J. Biol. Chem. 266 (1991) 2403124037. Old S.E., Sato S., Kador P.F., Carper D.A., Proc. Natl. Acad. Sci. USA 87 (1990) 49424945.

[8] [9] [10] [11] [12] [13] [14] [15] [16] [17]

Eur. J. Med. Chem. 34 (1999) 753758 1999 Editions scientiques et mdicales Elsevier SAS. All rights reserved

753

Short communication

QSAR study on antibacterial and antifungal activities of some 3,4-disubstituted-1,2,4-oxa(thia)-diazole-5(4 H)-ones(thiones) using physicochemical, quantumchemical and structural parameters
mer Gebana, Hamide Ertepinara, Mine Yurtseverb, Sekin zdenc*, Fatma Gmsd
Middle East Technical University, Faculty of Education, Department of Science Education, 06531 Ankara, Turkey b Istanbul Technical University, Faculty of Arts and Sciences, Department of Chemistry, Istanbul, Turkey c Ankara University, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, 06100 Tandogan- Ankara, Turkey d Gazi University, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, 06100 Ankara, Turkey (Received 19 October 1998; revised 3 March 1999; accepted 11 March 1999)
a

Abstract This work demonstrated the quantitative structure-activity relationships of 3,4-disubstituted-1,2,4-oxa(thia)-diazole-5(4 H)-ones (thiones) using quantum chemical parameter R(I), hydrophobicity descriptor and structural parameters. Semiempirical molecular orbital calculations were used to determine the quantum chemical parameter R(I), which is the electron density of HOMO at the sulfur and oxygen in position 1 of the compounds investigated, divided by the orbital energy of HOMO. It was shown that the electron density of HOMO at the sulfur and oxygen of the molecules was strongly related to the biological activities of these molecules. The results obtained from the QSAR application indicated that there was a parabolic dependence between the biological activities and the R(I) index. The structural factor IY which shows the presence of a sulfur atom in position 1 was the dominant predictor for the antibacterial and antifungal activities. On the other hand, the other structural variable IX which shows the presence of a sulfur atom double bonded to the C atom in position 5 caused a decrease, but the hydrophobicity of the whole molecule () caused an increase in activity. 1999 Editions scientiques et mdicales Elsevier SAS 3,4-disubstituted-1,2,4-oxa(thia)-diazole-5(4 H)-ones(thiones) / QSAR / antibacterial activity / antifungal activity / quantum chemical calculations

1. Introduction A series of compounds having the 3,4-disubstituted1,2,4-oxa(thia)-diazole-5-(4H)-ones(thiones) structure (gure 1) was found to show antibacterial and antifungal activities [1]. These compounds were tested in vitro for their ability to inhibit the growth of representative grampositive and gram-negative bacteria and various fungal species. In order to identify the substituent effects to the chemical reactivity and the biological activity, we have examined the QSAR of a series of these compounds using the quantum chemical index R(I), physicochemical and structural parameters. Recently, much research which concerns the quantum chemical parameters in QSAR studies have been reported [26]. Theoretical calculations have been used for
*Correspondence and reprints

Figure 1. Structures of compounds 114.

the description of the mechanism of interactions at the molecular level. The electrophilic and nucleophilic superdelocalizability of nitrogen [6], sulfur and oxygen atom [5] and electron densities on the sulfur and oxygen atoms [4] in the investigated structures have been used as

754
Table I. Biological activities of the compounds 114. Compound 1 2 3 4 5 6 7 8 9 10 11 12 13 14
a

R1 Ph 2-Pyr 2-Pyr 2-Pyr 4-Pyr Ph Ph 2-Pyr 2-Pyr 2-Pyr Me 4-Pyr 4-Pyr 2-Pyr

R2 Me Me Ph Et p-tolyl H p-tolyl Et Me n-pr Ph Et Me Ph

S.a.a 3.58 3.89 4.01 3.92 3.73 3.89 4.06 2.72 2.99 3.05 2.98 3.32 2.99 3.11

B.s.b 3.58 3.89 4.01 4.22 3.73 3.89 4.06 2.72 2.99 3.05 3.28 3.32 2.99 3.11

P.a.c 3.58 3.89 4.01 4.22 3.73 3.89 4.06 3.02 2.99 3.05 2.98 3.32 2.69 3.11

C.a.d 3.89 4.19 4.31 3.92 4.03 3.89 4.06 3.02 3.29 3.35 3.28 3.02 2.99 3.41

C.p.e 3.89 4.19 4.31 4.22 4.03 4.19 4.36 3.02 3.29 3.35 3.28 3.02 3.29 3.41

C.p.f 3.89 4.19 4.31 4.22 4.03 3.89 4.06 3.02 3.29 3.35 3.28 3.02 3.29 3.41

C.n.g 4.19 4.19 4.31 4.22 4.03 4.19 4.36 3.32 3.29 3.65 3.28 3.02 3.29 3.41

S. aureus, bB. subtilis, cP. aeruginosa, dC. albicans, eC. pseudotropicalis, fC. parapsilosis, gC. neoformans.

the independent parameters in QSAR analysis. The results obtained from these researches showed that the main type of interaction at the binding site should be via interaction of the lone pairs of heteroatoms in the compounds which were investigated with any binding site(s). 2. Chemistry 2.1. Chemicals and biological activity The compounds investigated in this work were synthesized previously [79]. The antibacterial activity against Staphylococcus aureus, Bacillus substilis and Pseudomonas aeruginosa and antifungal activity against Candida albicans, Candida pseudotropicalis, Candida parapsilosis and Crytococcus neoformans were determined [1] using the tube dilution method, and given as minimum inhibitory concentration (MIC, g/mL). The potency was dened as log1/C in the QSAR analysis, where C is the molar MIC value of the compound and is used as the dependent variable in the QSAR study. The log1/C values are given in table I. 2.2. Molecular orbital calculations The AM1 [10] molecular orbital calculations were carried out using the MOPAC program [11] running on a Silicon Graphics Work station (IRIS 4D, INDIGO, POWER INDIGO 2). Initially the geometries of the molecules were fully optimized. After optimization, ISF calculations with VECTORS option were carried out to print all eigen values, eigen vectors, and HOMO energies (table II).

2.3. Quantitative structure-activity relationships (QSAR) QSAR of 3,4-disubstituted-1,2,4-oxa(thia)-diazole-5(4 H)-ones(thiones) was investigated using the quantumchemical parameter R(I), physicochemical parameter () and structural parameters. The electron density of HOMO at active sites of the molecules (position 1) are strongly related to the antibacterial and antifungal activities of the molecules. These active sites are important in such a way that they bind to some specic SH enzymes to achieve the role of inhibition. Previously Nakayama et al. [4] introduced an index R(I) which shows the electron density of HOMO at atom i which is one of the active sites (in the molecule) playing a role in the possible chemical reactions. It was dened as 2 R I = fr i / EHOMO 10 where fr(i) is the frontier electron density of HOMO at atom i and EHOMO is the energy of HOMO in eV measured from the zero level, f(i) is the measure of the delocalization of electron density of HOMO on atom i and EHOMO is the energy of the highest occupied (most energetic) molecular orbital. These two parameters are a measure of the relative reactivity of HOMO at atom i within a single molecule. Then, the ratio of fr(i) and EHOMO at atom i can be related to the biological reactivity of a series of molecules concerned. Oxygen and sulfur atoms embedded in the ve membered ring of the concerned structure were thought to be the effective part of these compounds. Then, to investigate the contribution of this index to the biological activities of the examined molecules, the R(I) values at the S and O atom in position 1 were calculated using the MOPAC program and the values are listed in table II.

755
Table II. QSAR parameters examined in this study for the compounds 114. X 1 2 3 4 5 6 7 8 9 10 11 12 13 14 O O O O O S S S S S S S S S Y S S S S S S S O O O O O O O EHOMO(eV) 9.105 9.559 9.310 9.560 9.410 8.961 8.824 9.067 9.008 9.059 9.017 9.274 9.285 8.984 R(I) 5.37 4.32 1.03 4.37 0.83 3.55 3.55 1.04 1.03 1.03 0.81 1.06 1.05 0.93 2.52 1.06 2.46 1.52 3.01 1.96 4.65 1.52 1.06 2.05 2.52 1.34 0.88 2.46 IX 0 0 0 0 0 1 1 1 1 1 1 1 1 1 IY 1 1 1 1 1 1 1 0 0 0 0 0 0 0

Concerning physicochemical parameters, the hydrophobicity of the compounds () were used in the QSAR analysis. values of the substituents were taken from the tables given by Hansch and Leo [12] and was used for the hydrophobicity of the whole molecule. The structural parameter (IY) expresses the presence of S and O atoms in position 1. IY is dened as 1 in the case of the presence of S in the position 1 and 0 for the O in the same position. The other structural variable (IX) represents the atom which is double bonded to the C atom in position 5. Similarly IX has a value of 1 for the S atom and 0 for the O atom. Table II shows the parameters used in this study. 3. Results and discussion 3.1. R(I) index and the biological activities of 3-4-disubstituted 1,2,4-oxa(thia)-diazole-5(4 H)-ones(thiones) Antibacterial and fungicidal activities of these compounds are given in table I. From table I it can be seen that molecules that contain O in position 1 (814) have lower activities than the molecules which have S in position 1 (17). Similarly the R(I) index is much lower for the O containing molecules than the others. Accordingly there is a correlation between R(I) and biological activities measured. Namely, sulfur containing molecules in position 1 have more electron density on the sulfur atom under consideration coming from HOMO, which makes that region more susceptible to chemical reactions, especially to the ones in which binding to another atom or electron transfer reactions are considered. 3.2. QSAR Correlation and regression analyses of the QSAR study were done on a microcomputer using the SPSS/PC [13].

In the equations, the gures in parantheses are the standard errors of the regression coefficients, n is the number of compounds, r is the multiple correlation coefficient, s is the standard error of estimate, and F is the ratio of regression and residual. The level of signicance accepted in all of the analyses was 0.05. The best equation(s) was selected among other equations by considering the various statistical criteria [1417]. All possible combinations of parameters were considered, except that square terms were only allowed in equations containing the corresponding linear terms. This method provided a total of 22 possible equations for each activity. The results showed that there were two best equations including the same terms for each activity. The rst best tted equations included R(I), R(I)2, , and IX for the activities against S. aureus, B. subtilis, P. aeruginosa, C. albicans, C. pseudotropicoli, C. parapsilosis, and C. neoformans. The second best tted equation for each activity included only the structural parameter IY. Table III shows the best equations for the antibacterial and antifungal activities of the examined structures. In these equations, n represents the number of compounds analysed, r, the multiple correlation coefficient, s, the standard deviation, and F, the ratio of regression and residual. The gure in parantheses is the 95% condence interval. The equations including R(I), R(I)2, , and IX together for the activities against S. aureus, B. subtilis, P. aeruginosa, C. albicans, C. pseudotropicoli, C. parapsilosis, and C. neoformans were all signicant because the overall F statistics for these equations were statistically signicant, and they have high r values (about 0.90). Also, the individual F statistics for the coefficients of R(I), R(I)2, , and IX in these equations were signicant at P < 0.05 (table IV).

756
Table III. The best equations. Number 1 2 Equation The best equations against S. aureus log1/C = 0.85 ( 0.20) IY + 3.02 log1/C = 1.05 ( 0.46) R(I) 0.18( 0.085) R(I)2 + 0.14 ( 0.11) 0.66 ( 0.28) IX + 2.59 The best equations against B. subtilis log1/C = 0.85 ( 0.24) I Y + 3.06 log1/C = 1.05 ( 0.51) R(I) 0.17 ( 0.09) R(I)2 + 0.14 ( 0.12) 0.68 ( 0.31) IX + 2.65 The best equations against P. aeruginosa log1/C = 0.89 ( 0.23) I Y + 3.02 log1/C = 1.13 ( 0.44) R(I) 0.19 ( 0.079) R(I)2+ 0.14 ( 0.11) 0.72 ( 0.27) IX + 2.57 The best equations against C. albicans log1/C = 0.85 ( 0.19) I Y + 3.19 log1/C = 0.75 ( 0.45) R(I) 0.13 ( 0.083) R(I)2 + 0.17 ( 0.11) 0.76 ( 0.28) IX + 3.06 The best equations against C. pseudotropicalis log1/C = 0.93 ( 0.18) I Y + 3.24 log1/C = 1.07 ( 0.44) R(I) 0.18 ( 0.08) R(I)2 + 0.18 ( 0.11) 0.71 ( 0.27) IX + 2.79 The best equations against C. parapsilosis log1/C = 0.85 ( 0.19) I Y + 3.24 log1/C = 0.82 ( 0.38) R(I) 0.14 ( 0.069) R(I)2 + 0.13 ( 0.09) 0.79 ( 0.23) IX + 3.14 The best equations against C. neoformans log1/C = 0.89 ( 0.18) I Y + 3.32 log1/C = 0.79 ( 0.46) R(I) 0.12 ( 0.085) R(I)2 + 0.18 ( 0.11) 0.61 ( 0.28) IX + 2.99 n 14 14 r 0.94 0.95 s 0.17 0.17 F 84.61 20.19 p 0.0000 0.0002

3 4

14 14

0.91 0.94

0.20 0.19

58.39 17.49

0.0000 0.0003

5 6

14 14

0.92 0.96

0.20 0.17

68.94 26.65

0.0000 0.0001

7 8

14 14

0.94 0.95

0.17 0.17

87.35 21.16

0.0000 0.0001

9 10

14 14

0.95 0.96

0.16 0.17

121.29 27.45

0.0000 0.0000

11 12

14 14

0.94 0.97

0.16 0.15

99.18 30.99

0.0000 0.0000

13 14

14 14

0.95 0.95

0.15 0.17

120.33 21.67

0.0000 0.0001

The predictor variables R(I), R(I)2, , and IX accounted for a signicant portion of the variance in the activities. Also, each predictor contributed a signicant proportion of the additional variance in the presence of the other variables for each activity. The structural parameter IX induced some inuences on the activity. But it showed that S, which is double bounded to the C atom in position 5, caused a decrease in the activity. R(I) and its square produced an additive contribution to the activity. There was a parabolic dependence between the biological activities and the R(I) index. The parabolic relationship of the activity against S. aureus on the reactivity of HOMO on the sulfur and oxygen of 3,4 disubstituted 1,2,4oxa(thia)-diazole-5(4 H)-one(thione) derivatives suggests that log1/C initially increases as the reactivity of HOMO increases up to an ideal R(I) value, and then it decreases after this point. is the summation of the hydrophobic parameter of the substituents R1 and R2 in the observed

structure and its positive coefficient suggested that higher hydrophobicity was favourable for the activity. An examination of the intervariable correlations involving all predictor variables in table V demonstrates the relatively large correlation between R(I) and IY (r = 0.71) (colinearity). When high colinearity exists, the regression analyses using the given set of independent variables can not be

Table V. Correlation matrix of all predictor variables used in the equations. R(I) R(I) IY IX 1.000 0.121 0.712 0.483 1.000 0.399 0.033 IY IX

1.000 0.745

1.000

757
Table IV. The t statistics (P < 0.05) for the coefficients of variables in the best equations 2, 4, 6, 8, 10, 12, 14. t Equation 2 R(I) R(I)2 IX Equation 4 R(I) R(I)2 IX Equation 6 R(I) R(I)2 IX Equation 8 R(I) R(I)2 IX Equation 10 R(I) R(I)2 IX Equation 12 R(I) R(I)2 IX Equation 14 R(I) R(I)2 IX 5.128 4.688 2.850 5.264 4.652 4.228 2.539 4.921 5.863 5.333 2.974 6.085 3.720 3.454 3.523 6.223 5.556 5.043 3.695 6.062 4.925 4.547 3.139 7.757 3.863 3.300 3.530 4.880 P 0.0006 0.0011 0.0191 0.0005 0.0012 0.0022 0.0317 0.0008 0.0002 0.0005 0.0156 0.0002 0.0048 0.0072 0.0065 0.0002 0.0004 0.0007 0.0050 0.0002 0.0008 0.0014 0.0119 0.0000 0.0038 0.0092 0.0064 0.0009 Table VI. The calculated r2 Equation Equation Equation Equation Equation Equation Equation Equation Equation Equation Equation Equation Equation Equation 1 2 3 4 5 6 7 8 9 10 11 12 13 14
PRESS

and s

PRESS

. s
PRESS

r2

PRESS

0.831 0.472 0.768 0.756 0.798 0.751 0.836 0.578 0.877 0.637 0.853 0.806 0.877 0.383

0.201 0.409 0.242 0.286 0.234 0.299 0.198 0.366 0.185 0.367 0.186 0.246 0.177 0.457

In order to judge the validity of the predictive power of the QSAR, a cross-validation method was applied to the original data set for equations 114 and the resulting PRESS (predicted residual sum of squares) for each equation was calculated according the following equation [14, 19]. 2 2 PRESS = yi yi / 1 hii
i

performed effectively [14, 18]. Also, the correlations between the biological activities and predictor variables demonstrates the highest correlation between the biological activities and structural parameter IY (r values about 0.90). For these reasons, the variable IY was entered into the correlation equation separately, and the second best equations were obtained (table III). F-test values for each activity were signicant. It showed the presence of the sulfur atom in position 1 was a dominant predictor for the activities against S. aureus, B. subtilis, P. aeruginosa, C. albicans, C. pseudotropicoli, C. parapsilosis, and C. neoformans.

here yi and yi are the response (activity) values of observation i (i = 1,2,..., n), observed and calculated by the best equation, respectively. The diagonal elements of that matrix are denoted by hii in the equation, and calculated by the SPSS computer program. The calculated overall cross-validated r2 and standard deviation values for each equation are given in table VI. The cross validation technique was used to check the validation of these equations, because this technique seems to substantially reduce the probability of chance correlation relative to multiple regression. In the present study, cross validation did not conrm some of the four-term equations. The cross-validated r2 values had a maximum at one-term equations. Cross validation results using IY as the only independent variable were much better than those including four parameters R (I), R(I)2, , and IX. One parameter equations were very stable, leading to cross-validated values in the range between 0.7680.877, whereas the four-parameter equations gave cross-validated values between 0.3830.806. This showed that one-term equations indicated higher predictive ability, as shown by cross validation. On the other hand, the greater the number of variables tested, the greater role chance will play in the observed correlation [20]. With an increasing number of observations, the probable degree of chance correlation is steadily reduced.

758 The number of variables correlating by chance increases as the number of observations decreases. In this study, for four parameter equations to be tested, the required number of observations to avoid undue risk of chance correlations can be insufficient. For these reasons, one parameter equations were sound models, and explained the data better. The other one-term, two-term, three-term, four-term, and ve-term equations were signicant, but most of them have lower r values and higher s values than the equations given in table III. However, these one-term, two-term, three-term, four-term, and ve-term equations contain at least one coefficient with a poor F statistic (P > 0.05). References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] Gms F., Drst Y., Drst N., Abbasoglu U., Pharmazie 48 (1993) 867868. Greco G., Novellino E., Silipo C., Vittoria A., Quant. Struct.-Act. Relat. 11 (1992) 461477. Mekenyan O.G., Veith G.D., Bradbury S.P., Russom C.L., Quan. Struct.-Act. Relat. 12 (1993) 132136. Nakayama A., Hagiwara K., Hashimoto S., Shimoda S., Quant. Struct.-Act. Relat. 12 (1993) 251255. Sener E., Turgut H., Yalin Y., ren Y., Trker L., elebi N., Akin A., Int. J. Pharm. 110 (1994) 109115. Veith G.D., Mekenyan O.G., Quant. Struct.-Act. Relat. 12 (1993) 349356. Agirbas H., Drst Y., Smengen D., Sulfur and Silicon 66 (1992) 321324. Drst Y., Agirbas H., Smengen D., Phosphorus Sulfur and Silicon 62 (1991) 4751. Smengen D., Agirbas H., Drst Y., Dogan N., Chim. Acta Turc. 20 (1992) 1723. Dewar M.J.S., Stewart J.J.P., Zocbisch E.G., Healy E.F., J. Am. Chem. Soc. 107 (1985) 39023909. Stewart J.J.P., MOPAC Manual, Frank J. Seiler Research Laboratory, United States Air Force Academy, USA, (1990). Hansch C., Leo A., Substituent Constants for Correlation Analysis in Chemistry and Biology, John Wiley and Sons, New York, 1979. Norusis M.J., The SPSS Guide to Data Analysis for SPSS/PC+ (2nd ed.), SPSS Inc., Chicago, 1991, pp. 849872. Rawlings J.O., Applied Regression Analysis: A Research Tool, Wadsworth Inc., California, 1988, pp. 327336. Anderson T.W., Stanley L.S., An Introduction to the Statistical Analysis of Data, Houghton Mifflin Company, Boston, 1978, pp. 619637. Darper N.R., Smith H., Applied Regression Analysis, John Wiley and Sons Inc., New York, 1981, pp. 294342. Howell D.C., Statistical Methods for Psychology, PWS-Kent Publishing Company, Boston, 1987, pp. 463507. Norman H.I., Hull C.H., Jenkins J.G., Steinbrenner K., Bent D.H., SPSS Statistical Package for the Social Sciences, McGraw Hill Inc, New York, 1975, pp. 340341. Wold S., Quant Struct.-Act. Relat. 10 (1991) 191193. Topliss J.G., Costello, R.J., J. Med. Chem. 15 (1972) 10661068.

4. Conclusion

[11] [12]

The activity contributions of 3,4 disubstituted-1,2,4oxa(thio)-diazole-5(4H)-one(thione) derivatives indicated that S containing compounds in position 1 have higher biological activities against some bacteria and fungal species. QSAR studies which concerned the quantum chemical, physical and structural parameters conrmed this result. The highest correlation between the biological activities and structural parameter IY revealed that the main interactions between the binding sites should be on the lone pair electrons of oxygen and sulfur atoms in position 1 of the ve membered ring of some 3,4 disubstituted1,2,4 oxa(thio)-diazole-5(4 H)-one(thione) derivatives.

[13] [14] [15]

[16] [17] [18]

[19] [20]

Eur. J. Med. Chem. 34 (1999) 759763 1999 Editions scientiques et mdicales Elsevier SAS. All rights reserved

759

Laboratory note

Hydroxypyridone iron chelators, L1 and analogues. An 17O NMR study

Giovanni Cerioni*a, Antonio Plumitallob, Maria G. Batzellac, Francesca Loccid
Dipartimento di Scienze Chimiche, Universit di Cagliari, Cittadella Universitaria di Monserrato, SS 554 Bivio per Sestu, I-09024 Monserrato (Ca), Italy b Dipartimento Farmaco Chimico Tecnologico, Universit di Cagliari, Via Ospedale 72, I-09124 Cagliari, Italy c Centro Trasfusionale, Ospedale di San Gavino Monreale, Via Roma 236, I-09037 San Gavino Monreale, Italy d Dipartimento di Citomorfologia, Universit di Cagliari, Cittadella Universitaria di Monserrato, SS 554 Bivio per Sestu, I-09024 Monserrato (Ca), Italy (Received 29 October 1998; revised 3 March 1999; accepted 3 March 1999)
a

Abstract Some iron chelators of the class of hydroxypyridones, including the experimental drug 1,2-dimethyl-3-hydroxy-4-pyridone (L1), have been studied by 17O NMR spectroscopy. The reciprocal inuence of the keto and hydroxy groups has been examined in order to gain a better comprehension of their bonds. A strong intramolecular hydrogen bond between 3-OH and 4-CO groups has been observed for L1 and its analogues. A comparison has also been made on the data of three different solvents. 1999 Editions scientiques et mdicales Elsevier SAS iron chelators / L1 / hydrogen bond /
17

O NMR

1. Introduction Humans have a very limited capacity for iron excretion [1], even when iron stores are markedly increased. Thus, congenital or acquired iron-loading anaemias can cause severe tissue damage and eventual death, unless the iron overload is removed. One such disease, -thalassaemia, still has a very large diffusion in our region of Sardinia. Desferrioxamine (DFO) is the only iron-chelating agent currently in widespread clinical use. Its administration, however, requires cumbersome and expensive methods of slow subcutaneous or intravenous infusion. This fact has expedited the research for oral iron chelators [2]. One of the most promising among them, 1,2-dimethyl-3-hydroxy-4-pyridone (L1) [2, 3], is currently in clinical tests by our group. Dioxobidentate ligands are generally iron (III) ligands [4], and the majority of them incorporate the carbonyl and the hydroxy groups. It would be interesting to directly investigate these groups, which actually bond to the metal atom. Moreover, one of the most useful concepts [5] to develop new chelating agents is the theory
*Correspondence and reprints

of hard and soft bases [6]. Its application is helped by a knowledge of the electronic (and charge) distribution, e.g. in the bases [5]. A very valuable method to study oxygenated groups is 17 O NMR spectroscopy [7], and both carbonyl and hydroxy groups have been extensively studied by this technique [811]. We thus decided to apply such a method to the hydroxypyridone iron chelators, L1 and analogous molecules, as a starting point for a systematic study of actual and possible new iron chelating drugs. 2. Chemistry Structures of studied compounds are shown in gure 1 and their 17O NMR spectroscopic data are shown in table I. The OH structure has also been used for the groups in positions 2 or 4, without any sign of a preference between the hydroxyl or the carbonyl structure. 17 O NMR shifts of the 3-OH group in H2O/D2O are quite similar for all the derivatives bearing this group except for 5. The observed shifts are in the range of the phenols [1215], albeit 13 are among the most shielded ones.

760 The comparatively strong deshielding noted for 5 is only partly due to the lack of an intramolecular hydrogen bond with a contiguous carbonyl group and/or to a different extent of intermolecular hydrogen bonding with the solvent. The intramolecular hydrogen bond between the 3-hydroxy and 4-keto groups of 13 should scarcely inuence the chemical shift of the 3-hydroxy group, within 1015 ppm, as discussed e.g. in two studies on salicylaldehydes and salicylanilides by Boykin and coworkers [12, 13]. Phenol itself in H2O has a = 77 ppm shift [14]. Its shift ranges, in some solvents, from 73.5 (MeCN) [12] to 83.6 ppm (DMSO-d6) [15]. All these values are thus more similar among themselves, and to those obtained for 5 in pyridine-d5 and in CDCl3, than to the shift of 5 in H2O/D2O. These data on phenol also suggest that intermolecular hydrogen bonding is not a sufficient cause for the observed deshielding. Without re-opening a somewhat old and controversial problem [1620] on the real structure of 5, we agree on an equilibrium, in H2O solution, between 5a and 5b with an estimated ca. 1:1 ratio [20] (gure 2). According to the Karplus-Pople equation [21], structure 5b with an increased electronic density at the oxygen

Figure 1. Structures of studied compounds.

Table I.

17

O NMR chemical shifts (ppm) of studied compounds in H2O/D2O, C5D5N and CDCl3.a 2 OH (H2O/D2O) 2 OH (C5D5N) 2 OH (CDCl3) 3 OH (H2O/D2O) 54 60 59 224.7 268.6 245.7 105.2 ca. 215 218 252.2 272 b b ca. 65 77.7 69.8 78.0 266.3 b 104 110 b 251.9 ca. 290 3 OH (C5D5N) 53.5 51.0 53.6 3 OH (CDCl3) 50.6 47.5 52.1 4 OH (H2O/D2O) 224 229.6 227.5 4 OH (C5D5N) 306.2 305.0 303.0 4 OH (CDCl3) 280.3 277.4 280.3

Compound 1 2 3 4 5 6 7 8

a) Linewidths range 200600 Hz, but for 5 (880 Hz) and 6 (ca. 3 000Hz) in CDCl3 and for 7 and 8 in H2O/D2O (not measurable, very faint signals). b) Insoluble, not measured.

Figure 2. Structures of compound 5.

761 atom should cause an upeld shift. The observed deshielding can be due to a considerable weight of resonance for structures 5ce, which are alike to the commonly accepted resonance structures of the phenol and phenate anion. Actually, sodium phenate in H2O solution, in a trial experiment, has given a strong deshielding (146 ppm) compared to phenol (ca. 77 ppm). An important CO double bond character has thus to be taken into account for the observed deshielding of 5. Compounds 13 and 7, in H2O/D2O, for their 3-hydroxy groups, show a narrow shift range, 5465 ppm, and are thus more shielded than most other phenols [1215]. Several causes could co-operate to give this result. As discussed, the inuence of intramolecular hydrogen bonding can partly account for this shielding, even if a possible inductive effect of the ring nitrogen should operate in an opposite sense. Whilst to the best of our knowledge data for meta-substituted phenols are lacking in the literature, an indication can be obtained from the reported data on anisoles [23]. In their case, if we assume a meta nitro group as a comparable substitute of the ring nitrogen of our compounds, a 15 ppm deshielding has been observed. Moreover, parasubstitued phenols are similarly deshielded by electronwithdrawing groups [12]. The two contributions to the chemical shifts seem opposite and comparable in magnitude. The inuence of the adjacent carbonyl group is thus apparent, irrespectively of its 2 or 4 position. Its presence causes the resonance structures with a CO double bond in the 3 position to become negligible, allowing an understanding of the shielding observed compared to other phenols [1215]. This point is also important for Fe(III) co-ordination, since the hardness as a donor of the oxygen atom is increased. Derivatives 4 and 6 are known to exist in solution mainly as 2 and 4pyridones [20]. The 17O NMR spectrum of 4 in MeCN has been reported [24] and its shift (269.5 ppm) is in good agreement with our value in pyridine-d5 (table I). The observed shifts of the 4-carbonyl group of derivatives 13 are thus in substantial agreement with the usually accepted pyridone structure of these compounds. We observe that in H2O/D2O the C=O groups of 13 are much more upeld than in the other two solvents, particularly in pyridine-d5 (table I). On the contrary, 3OH groups do not show any signicant or systematic variation with the solvent. All these observations are coherent with a strong intermolecular hydrogen bond formation, at the carbonyl level, in H2O/D2O but without a rupture of the intramolecular hydrogen bond. An alternative and/or complementary
Figure 3. Resonance formulas of compound 1.

cause of the observed shielding could be an increased importance of the resonance formulas type 1b for 13 (gure 3). Again, structures type 1b should increase the hardness of the oxygen atom. The importance of the intermolecular hydrogen bond is evident by 4, which shows the same trend of variation of the shifts with all the solvents. Its d values (table II) however are smaller than those of the other derivatives. Data for 6 are difficult to interpret. On going from H2O/D2O to less polar solvents, a retreat towards the hydroxy tautomer would be predicted [25]. On the other hand, a strong intermolecular hydrogen bonding between the CO group and the solvent is present in H2O/D2O and its effect is shielding [26]. The two effects are thus counteracting and it is not possible to know a priori which one will prevail. Moreover, in CDCl3 there is surely an increased importance of self-association, as shown by the quite large value of the line-width of the signal. Spectra of 8 clearly show that only one of the substituents is in the keto form. In accordance with literature [27], we can assign a 2-pyridone structure at this derivative.

Table II. Chemical shifts differences ()a of CO groups of selected compounds. Compound 1 2 3 4 6 P-W (ppm) +82.2 +75.4 +75.5 +43.9 14.4 C-W (ppm) +56.3 +47.8 +52.8 +21.0 +23.7 P-C (ppm) +25.9 +27.6 +22.7 +22.9 38.1

a) Subscripts: PW, chemical shifts difference between C5D5N and H2O/D2O solutions, CW, chemical shifts difference between CDCl3 and H2O/D2O solutions, PC, chemical shifts difference between C5D5N and CDCl3 solutions.

762 3. Conclusion In this report we have shown the potential of 17O NMR spectroscopy, at natural isotopic abundance, for studying the bonds and solvent interactions in small oxygencontaining molecules, like the drug L1. It has been possible to determine that the intramolecular hydrogen bond between the 3OH and the 4CO groups in this molecule is strong enough not to be broken even in aqueous solution. The mutual interactions between hydroxyl and carbonyl groups have been examined, obtaining indications of the relative importance of their resonance structures and their inuence on their ability to act as Fe(III) complexing agents. These data can thus facilitate both the understanding of the activity of L1 as well as the development of new and more efficient iron chelating drugs. The importance of these points is evident from very recent literature data [28, 29]. A possible hepatotoxicity of L1 has in fact been reported [28] from one of the main groups involved in clinical experimental use of this drug and this point has been challenged in an editorial paper [29], stressing the importance of ascertaining the safety and efficacy of this agent. 4. Experimental All studied derivatives are known compounds. 13 have been prepared according to Kontoghiorghes procedure [30]. 48 were purchased from Aldrichy and used without further purication. 17 O NMR spectra were recorded, in the Fourier transform mode, on a Varian VXR 300 spectrometer equipped with a Sun 3/60 computer and with a 10 mm broad band probe at 65 C (probe temperature = 338 K) for H2O/D2O and pyridine-d5 solutions and at 45 C (probe temperature = 318 K) for CDCl3 solutions and at natural isotopic abundance. Saturated solutions were used in all experiments. The instrumental settings were: 40.662 MHz frequency, spectral width 36 KHz, acquisition time 10 ms, pre-acquisition delay 100 s, pulse angle 90 (pulse width 28 s). Number of scans varied largely (4 1043 106) as a function of solvent and solubility. The spectra were recorded with sample spinning and without lock. The signal to noise ratio was in most cases improved by applying a 30 Hz exponential broadening factor (l.b.) to the FID prior to Fourier transformation. In some experiments (H2O/D2O solutions of 13 and 7 and CDCl3 solution of 6) an l.b. up to 120 Hz was necessary. The data point resolution was improved by zero lling to 16 K data points. Chemical shifts are expressed in ppm and referred to external tap water by the substitution method. The reproducibility of the chemical shift data is estimated to be 1 ppm ( 3 ppm when l.b. = 120 Hz has been used). Acknowledgements This work was supported by the Regione Autonoma della Sardegna, Assessorato dellIgiene e Sanit e dellAssistenza Sociale, Grant No. 3250. References
[1] [2] [3] [4] [5] Green R., Charlton R., Seftel H., Bothwell T., Mayet F., Adams B., Finch C., Layrisse M., Am. J. Med. 45 (1968) 336353. Bergeron R.J., Brittenham G.M., The Development of Iron Chelators for Clinical Use, CRC Press, Boca Raton, 1994. Kontoghiorghes G.J., Ann. NY Acad. Sci. 612 (1990) 339350. Hider R.C., Hall A.D., Prog. Med. Chem. 28 (1991) 41173. Martell A.E., Motekaitis R.J., Sun Y., Clarke E.T., in: Bergeron R.J., Brittenham G.M. (Eds.), The Development of Iron Chelators for Clinical Use, CRC Press, Boca Raton, 1994, pp. 329351. Pearson R.G., J. Am. Chem. Soc. 85 (1963) 35333539. Boykin D.W., CRC Press, Boca Raton, 17O NMR Spectroscopy in Organic Chemistry 1991. Chandrasekaran S., in: Boykin D.W. (Ed.), 17O NMR Spectroscopy in Organic Chemistry, CRC Press, Boca Raton, 1991, pp. 141204. Boykin D.W., Baumstark A.L., in: Boykin D.W. (Ed.), 17O NMR Spectroscopy in Organic Chemistry, CRC Press, Boca Raton, 1991, pp. 3968. Boykin D.W., Baumstark A.L., Boykin D.W. (Ed.), 17O NMR Spectroscopy in Organic Chemistry, CRC Press, Boca Raton, 1991, pp. 6994. Boykin D.W., Baumstark A.L., in: Boykin D. W. (Ed.), 17O NMR Spectroscopy in Organic Chemistry, CRC Press, Boca Raton, 1991, pp. 205232. Boykin D.W., Chandrasekaran S., Baumstark A.L., Magn. Reson. Chem. 31 (1993) 489494. Nowak-Widra B., Allison L.W., Kumar A., Boykin D.W., J. Chem. Res. (S) (1994) 490491. St Amour T.E., Burgar M.I., Valentine B., Fiat D., J. Am. Chem. Soc. 103 (1981) 11281136. Frey J., Eventova I., Rappoport Z., Mller T., Takai Y., Sawada M., J. Chem. Soc. Perkin Trans. II (1995) 621637. Paoloni L., Tosato M.L., Cignitti M., Theoret. Chim. Acta 14 (1969) 221231. Berthier G., Lvy B., Paoloni L., Theoret. Chim. Acta 16 (1970) 316318. Cignitti M., Paoloni L., Theoret. Chim. Acta 25 (1972) 277288. Vgeli U., von Philipsborn W., Org. Magn. Reson. 5 (1973) 551559. Johnson C.D., in: Boulton A.J., McKillop A. (Eds.), Comprehensive Heterocyclic Chemistry, The Structure, Reactions, Synthesis and Uses of Heterocyclic Compounds, 2., Part 2A, Pergamon Press, Oxford, 1984, pp. 99164. Karplus M., Pople J.A., J. Chem. Phys. 38 (1963) 28032807. Wheland G.W., in: Resonance in Organic Chemistry, Wiley, New York, 1955, pp. 341342. Katoh M., Sugawara T., Kawada Y., Iwamura H., Bull. Chem. Soc. Jpn. 52 (1979) 34753476.

[6] [7] [8] [9]

[10]

[11]

[12] [13] [14] [15] [16] [17] [18] [19] [20]

[21] [22] [23]

763
[24] [25] Boykin D.W., Sullins D.W., Pourahmady N., Eisenbraun E.J., Heterocycles 29 (1989) 307312. Boulton A.J., McKillop A., in: Boulton A.J., McKillop A., (Eds.), Comprehensive Heterocyclic Chemistry, The Structure, Reactions, Synthesis and Uses of Heterocyclic Compounds, 2., Part 2A, Pergamon Press, Oxford, 1984, p. 26. Baumstark A.L., Boykin D.W., in: Boykin D.W., (Ed.), 17O NMR Spectroscopy in Organic Chemistry, CRC Press, Boca Raton, 1991, pp. 111112. [27] [28] De Kowalewski D.G., Contreras R.H., De Los Santos C., J. Mol. Struct. 213 (1989) 201212. Olivieri N.F., Brittenham G.M., McLaren C.E., Templeton D.M., Cameron R.G., McClelland R.A., Burt A.D., Fleming K.A., N. Engl. J. Med. 339 (1998) 417423. Kowdley K.V., Kaplan M.M., N. Engl. J. Med. 339 (1998) 468469. Kontoghiorghes G.J., Sheppard L., Inorg. Chim. Acta 136 (1987) L11L12.

[26]

[29] [30]

Eur. J. Med. Chem. 34 (1999) 765771 1999 Editions scientiques et mdicales Elsevier SAS. All rights reserved

765

Laboratory note

Antimicrobial activities of new analogues of benzalkonium chloride

J. Pernaka*, I. Mirskab, R. Kmiecika
b a Poznan University of Technology, Sklodowskiej-Curie 2, 60-965 Poznan, Poland K. Marcinkowski University of Medical Sciences, Sieroca 10, 61-771 Poznan, Poland

(Received 29 December 1998; revised 15 March 1999; accepted 18 March 1999)

Abstract (Alkoxymethyl)dimethyl{2-hydroxy-5-[(4-X-phenyl)azo]benzyl}ammonium chlorides were prepared in high yield. All these chlorides, new analogues of benzalkonium chloride, showed antimicrobial activity. Activity depends on the length and kind of substituent at the quaternary nitrogen atom. 1999 Editions scientiques et mdicales Elsevier SAS analogue of benzalkonium chloride / 4-hydroxyazobenzenes / Mannich bases / chloromethylalkyl ether / antimicrobial activity

1. Introduction Benzalkonium chloride (BAC) is the product of a nucleophilic substitution reaction of alkyldimethylamine with benzyl chloride [1]. Chemically, it is monoalkyldimethylammonium chloride with one long-chain alkyl group representing a mixture of the alkyls from C8H17 to C18H37. Following Domagks publication in 1935 [2], a large number of application areas were developed for BAC. It is used as a pharmaceutical aid (preservative), cationic surface active agent, germicide, antiseptic (topical), antiseptic for skin preoperatively or for wounds, burns, etc. BAC is often present as a preservative or stabilising agent in nebulizer solutions used to treat asthma and chronic obstructive pulmonary disease [3]. Also it is widely used as an antimicrobial agent in the treatment of common infections of the mouth and throat. We now report the synthesis and antimicrobial activities of new quaternary ammonium compounds, the analogues of benzalkonium chloride. We plan to nd compounds with antimicrobial activity which are diluted in water giving a coloured solution. Commercial products with these compounds will not have to contain any dye.

2. Chemistry Mannich reaction (or aminomethylation) of variously substituted phenols is a well known process and comprehensive reviews have been published [47]. A few Mannich bases of phenolic azobenzenes have demonstrated cytotoxicity towards murine and human cancers [8]. Mannich bases methiodides have a promising cytotoxic activity in a wide variety of tumours [9]. The new analogues of benzalkonium chloride (310) were prepared by the reaction of 2-[(dimethylamino)methyl]-4-[(4-chlorophenyl)azo]phenol (2a) or 2-[(dimethylamino)methyl]-4-[(4-methylphenyl)azo]phenol (2b) with chloromethylalkyl or chloromethylcycloalkyl ethers giving yields between 8098%. In this case, Mannich base is a compound which is easily transformed into a quaternary ammonium salt. Chloromethylalkyl and chloromethylcycloalkyl ethers were synthesised from the corresponding alcohols. Mannich bases (2ab) were prepared by treatment of the 4-hydroxyazobenzenes with equimolar quantities of formaldehyde and dimethylamine in 75% yield. The 4-hydroxyazobenzenes (azo dyes) were prepared by several authors in the 19th century [10]. At the present time, only one compound, 4-hydroxyazobenzene (Solvent Yellow 7), is commercially available.

*Correspondence and reprints

766

3. Antimicrobial activity All synthesised quaternary ammonium chlorides (310) were tested for antimicrobial activity against cocci, rods and fungi. 4. Results and discussion 4-Hydroxyazobenzenes react readily with formaldehyde and secondary amines to give Mannich bases. The structures of two prepared mono Mannich bases (2ab) were characterized by their microanalysis CHN and by their 1H and 13C NMR spectroscopy. Quaternization of 2a and 2b with chloromethylalkyl or chloromethylcycloalkyl ethers produced red crystalline quaternary ammonium chlorides 310 (table I) diluted in water. The water solution of these chlorides is stable and orange in colour. 1H and 13C NMR-spectral analysis of prepared chlorides allowed easy elucidation of their structure. The minimal inhibitory concentration (MIC) and the minimal bactericidal concentration (MBC) values determined for all forty chlorides are given in tables II and III. The chlorides studied were divided into four groups with respect to the kind of substituent: group 1, chlorides with an alkoxymethyl substituent with an even number of carbon atoms (3 and 7); group 2, chlorides with the same substituent but with an odd number of carbon atoms (4 and 8); group 3, chlorides with a cycloalkoxymethyl

substituent (5 and 8); and group 4, chlorides with CH2O(CH2)nC6H11 substituent (6 and 10). The calculated average MIC values for cocci, rods and fungi are shown in gures 1 and 2. As shown by the results in these tables and gures, all the chlorides studied are very active against cocci and active against rods and fungi. The microbial activity depends on the length and kind of substituent at the quaternary nitrogen atom. Figures 1 and 2 reveal a decrease in the MIC value to the optimum value and these values increase for chlorides from groups 1, 2 and 3. The same correlation is observed for the MBC values. Generally the MBC values are slightly higher than MIC values. To the most active compounds in group 1 belong chlorides which have butoxymethyl, hexyloxymethyl and octyloxymethyl substituent, in group 2 chloride with heptyloxymethyl chain, and in group 3 chlorides which have cyclopentyloxymethyl, cyclohexyloxymethyl and cycloheptyloxymethyl substituent. To the worst compounds belong chlorides with dodecyloxymethyl and cyclododecyloxymethyl chain. Chlorides from group 4 have comparable values of MIC. In this group the microbial activity does not depend on the substituent CH2O(CH2)nC6H11 where n = 1 or 2. The most active chlorides against microorganisms were: {5-[(4-chlorophenyl)azo]-2-hydroxybenzyl}dimethyl(cyclohexyloxymethyl)ammonium (5b), {5-[(4-chlorophenyl)azo]-2-hydroxybenzyl}dimethyl(cyclohexylmethyloxymethyl)ammonium (6a), {2-hydroxy-5-[(4methylphenyl)azo]benzyl}dimethyl(hexyloxymethyl)-

767
Table I. (Alkoxymethyl)dimethyl{5-[(4-chlorophenyl)azo]-2-hydroxybenzyl}ammonium chlorides (36) and (alkoxymethyl)dimethyl{2hydroxy-5-[(4-methylphenyl)azo]benzyl}ammonium (710) chlorides. Chloride 3a 3b 3c 3d 3e 3f 4a 4b 4c 4d 4e 5a 5b 5c 5d 5e 5f 6a 6b 6c
a

R1 Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl Cl
b

R2 C2H5 C4H9a C6H13a C8H17a C10H21a C12H25a C3H7a C5H11a C7H15a C9H19a C11H23a C5H9b C6H11b C7H13b C8H15b C12H23b C6H10CH3b CH2C6H11b CH2CH2C6H11b CH2CH2CH2C6H11b alicyclic.

Yield (%) 80 96 98 90 85 86 90 80 80 90 96 80 80 90 80 80 80 90 86 85

Chloride 7a 7b 7c 7d 7e 7f 8a 8b 8c 8d 8e 9a 9b 9c 9d 9e 9f 10a 10b 10c

R1 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3 CH3

R2 C2H5 C4H9a C6H13a C8H17a C10H21a C12H25a C3H7a C5H11a C7H15a C9H19a C11H23a C5H9b C6H11b C7H13b C8H15b C12H23b C6H10CH3b CH2C6H11b CH2CH2C6H11b CH2CH2CH2C6H11b

Yield (%) 80 86 90 80 80 96 80 87 80 80 90 88 84 92 80 86 84 80 80 80

linear alkyl,

ammonium (7c), {2-hydroxy-5-[(4-methylphenyl)azo]benzyl}dimethyl(cyclopentyloxymethyl)ammonium (9a). The results presented demonstrate that the new BAC analogues are very active against cocci. Their activities are similar to the activity of BAC. The antimicrobial activities of BAC (Aldrich product, in which R represents a mixture of alkyls from C8H17 to C18H37) against cocci, Micrococcus luteus, Staphylococcus epidermidis and Staphylococcus aureus as measured in the same MIC test are 1.5, 3.0 and 1.5 mmol/L, respectively. Tomlinson and coworkers [11] reported the antibacterial activities of homologues series (C8C18) of alkylbenzyldimethylammonium chlorides against Pseudomonas aeruginosa. BAC resistance is a potential problem for application, for example in the food processing industry [12]. We found compounds with large molecular weights, crystalline, diluted in water and the orange water solution is stable. 5. Experimental protocols 5.1. Chemistry NMR spectra were recorded on a Varian Model XL 300 spectrometer at 300 MHz for 1H and 75 MHz for 13C at 20 C with tetramethylsilane as internal reference. Satisfactory elemental analyses were obtained: C 0.32, H 0.29 and N 0.24.

Chloromethylalkyl ethers and chloromethylcycloalkyl ethers were prepared via the procedures which were reported earlier [13]. The percentage of ether in a crude product was determined by an alkalimetric method [14]. 2-[(Dimethylamino)methyl]-4-(phenylazo)phenols (2ab); general procedure: Pathway A. To a solution of dimethylamine (30 mmol) in 95% EtOH (10 mL) paraformaldehyde powder (0.9 g, 30 mmol) was added. The mixture was stirred and heated when the paraformaldehyde had dissolved, then the corresponding 4-hydroxyazobenzene (30 mmol) in 100 mL EtOH was added. The reaction mixture was stirred for 1 h at 60 C. Pathway B. To 2-[(dimethylamino)methyl]phenol (20 mmol) in 50 mL MeOH was added diazonium salt prepared from dimethylamine (20 mmol). The mixture was stirred at room temperature for 2 h. The solid substrate was ltered and then recrystallized from EtOH. 2-[(Dimethylamino)methyl]-4-[(4-chlorophenyl)azo]phenol (2a): m.p. 120122 C, 1H NMR (CDCl3) ppm = 11.6 (s, OH), 7.83 (dd, J = 6 Hz, 1H), 7.82 (d, J = 9 Hz, 2H), 7.61 (d, J = 2 Hz, 1H), 7.45 (d, J = 8 Hz, 2H), 6.95 (d, J = 8 Hz, 1H), 3.72 (s, 2H), 2.36 (s, 6H); 13 C NMR ppm = 161.8, 151.0, 145.5, 135.7, 129.1, 125.2, 123.6, 122.6, 122.1, 116.6, 62.6, 44.4. 2-[(Dimethylamino)methyl]-4-[(4-methylphenyl)azo]phenol (2b): m.p. 99100 C, lit. 103 C [10].

768
Table II. The MICa and MBCa values of examined chlorides (36).
Strainsb 3a Cocci M. luteus S. epidermidis S. aureus Rods P. aeuruginosa P. vulgaris K. pneumoniae E. coli S. marescens Fungi C. albicans T. mentagrophytes Rh. rubra MIC MBC MIC MBC MIC MBC
b

3b 10 10 10 18 38 76 304 304 76 152 304 304 38 76

3c 18 18 9 18 217 217 284 568 18 36 141 284 141 141

3d 17 34 8 17 267 534 267 534 132 267 132 267 132 267 534 534 267 534 66 132 132 267

3e 32 32 8 16 252 252 504 504 252 504 504 504 125 252

3f 59 118 30 30 238 477

4a 78 78 156 156

4b 37 37 73 146

4c 9 18 9 18 9 36

Chlorides 4d 4e 5a 33 518 16 128 128 518 61 61 31 31 245 490 19 19 19 19 9 19 73 146 19 73 38 146 38 73

5b 36 71 18 18 9 18 71 142 18 71 36 71 71 142 142 285 71 142 36 36 36 71

5c 18 18 9 9 9 9 553 553 9 137 36 36 18 18 256 553 68 256 36 36 36 36

5d 34 34 8 17 214 214 429 429 268 536 133 133 66 133

5e 479 479 8 118 8 8

5f 9 9 9 9 18 36

6a 9 9 18 18 18 36 137 553 18 137 36 36 18 36

6b 8 8 8 8 8 8 133 268 8 66 66 133 34 66 133 268 34 66 34 34 17 34

6c 16 16 16 64 16 129 520 520 16 129 64 129 32 64 1041 1041 129 260 64 129 64 64

MIC MBC MIC MBC MIC MBC MIC MBC MIC MBC MIC MBC MIC MBC MIC MBC

21 21 21 21 260 260 260 520 81 161 260 520 81 326

156 146 1256 219

1908 2512 1750 275 1908 2515 4695 550 477 954 954 954 477 954 628 219 2512 438 9 36

518 980 2074 980 259 518 245 490

1915 137 1915 256 16 9 1915 68 239 958 36 68

1256 875 9 2512 1750 36 628 628 438 875 136 275

64 490 1037 980 259 490 1037 980

958 36 1915 36

1302 608 568 1302 1216 568 325 651 161 161 161 161 304 608 76 152 152 304 70 141 18 18 36 70

1008 3816 5025 1750 275 4032 3816 5025 4695 550

1037 1960 73 2074 1960 146 146 146 73 73 38 73

1073 3831 1106 68 2146 3831 256 137 268 536 268 286 268 268 3831 256 68 3831 1106 137 1915 137 1915 256 1915 256 3831 256 68 68 36 68

504 3816 1256 875 136 259 490 1008 3816 2512 1750 1101 4149 980 125 252 252 504 477 477 1256 438 1256 875 36 36 36 68 64 122 2074 245 259 490 2074 490

3816 628 219 3816 1256 219

in mmol/L,

the number of microorganisms in mL ranged from 104105.

The quaternary ammonium chlorides 310 were prepared by dissolving 2-[(dimethylamino)methyl]-4-(Xphenylazo)phenol in CH2Cl2 and adding an equimolar amount of the appropriate chloromethylalkyl or chloromethylcycloalkyl ether. The mixture was stirred at room temperature for 24 h. The solvent was evaporated and the crude product was extracted three times with hexane. Finally, the products were crystallized from CH3COOC2H5/MeOH and dried in vacuum oven. {5-[(4-chlorophenyl)azo]-2-hydroxybenzyl}dimethyl(dodecyloxymethyl)ammonium chloride (3f): m.p. 124126 C, 1H NMR (DMSO-d6) ppm = 11.9 (s, OH), 8.06 (d, J = 2 Hz, 1H), 7.96 (dd, J = 7 Hz, 1H), 7.86 (d, J = 9 Hz, 2H), 7.66 (d, J = 9 Hz, 2H), 7.46 (d, J = 9 Hz, 1H), 4.82 (s, 2H, CH2N), 4.58 (s, 2H, NCH2O), 3.86 (t, J = 7 Hz, 2H), 3.02 (s, 6H), 2.00 (m, 2H), 1.63 (m, 18H), 0.86 (t, J = 7 Hz, 3H); 13C NMR ppm = 161.4, 150.5, 144.6, 135.2, 130.5, 129.5, 125.9, 123.8, 117.1, 114.9, 90.0, 73.0 (NCH2O), 57.9 (CH2N), 46.3 [N(CH3)2], 33.7, 31.3, 29.2, 29.0, 28.8, 28.7, 25.3, 22.1, 13.9.

{5-[(4-chlorophenyl)azo]-2-hydroxybenzyl}dimethyl(cyclohexyloxymethyl)ammonium chloride (5b): m.p. 120123 C, 1H NMR (DMSO-d6) ppm = 11.5 (s, OH), 8.07 (d, J = 2 Hz, 1H), 7.96 (dd, J = 7 Hz, 1H), 7.85 (d, J = 9 Hz, 2H), 7.58 (d, J = 9 Hz, 2H), 7.38 (d, J = 9 Hz, 1H), 4.86 (s, 2H, CH2N), 4.60 (s, 2H, NCH2O), 3.84 (m, 1H), 3.05 (s, 6H), 1.92 (m, 2H), 1.74 (m, 2H), 1.49 (m, 6H); 13C NMR ppm = 159.3, 148.7, 143.0, 133.6, 128.3, 127.4, 124.4, 121.8, 115.2, 113.2, 82.5, 78.0 (NCH2O), 56.1 (CH2N), 44.4 [N(CH3)2], 33.4, 27.7, 21.9. (Cycloheptyloxymethyl)dimethyl{2-hydroxy-5-[(4-methylphenyl)azo]benzyl}ammonium chloride (9c): m.p. 152154 C, 1H NMR (DMSO-d6/CDCl3) ppm = 11.3 (OH), 8.03 (d, J = 2 Hz, 1H), 7.92 (dd, J = 6 Hz, 1H), 7.76 (d, J = 9 Hz, 2H), 7.34 (d, J = 9 Hz, 3H), 4.82 (s, 2H, CH2N), 4.59 (s, 2H, NCH2O), 4.02 (m, 1H), 3.05 (s, 6H), 2.42 (s, 3H), 1.99 (m, 2H), 1.81 (m, 4H), 1.58 (m, 4H), 1.54 (m, 2H); 13C NMR ppm = 158.7, 148.4,

769
Table III. The MICa and MBCa values of examined chlorides (710).
Strainsb Cocci M. luteus S. epidermidis S. aureus Rods P. aeuruginosa P. vulgaris K. pneumoniae E. coli S. marescens Fungi C. albicans T. mentagrophytes Rh. rubra MIC MBC MIC MBC MIC MBC
b

Chlorides 7a MIC MBC MIC MBC MIC MBC MIC MBC MIC MBC MIC MBC MIC MBC MIC MBC 22 22 22 22

7b 20 40 10 79

7c 9 9 9 9 19 19

7d 9 9 9 18 9 9

7e 65 65 33 33

7f 61 61

8a 42 82

8b 20 20 10 10 10 20

8c 18 18 9 9 9 36

8d 9 34 34 67

8e

9a

9b 10 19 19 38

9c 9 37 9 18 9 18

9d 9 9 9 18 9 9

9e

9f

10a 10b 10c 9 9 9 9 37 72 9 9 9 9 9 9 17 17 17 17 17 17

BAC 1.5 3 3 3 1.5 6

63 20 126 40 8 8 20 40

249 9 249 9 8 16 9 18

248 42 248 42 42 82

11 20 170 40

16 16 262 16

134 126 77 19 270 126 154 38

249 9 997 9

344 158 147 138 2103 3972 331 76 143 1083 2043 1239 148 143 139 1994 143 143 280 1088 96 688 319 298 279 4206 3972 662 308 288 4333 4086 1239 299 289 1122 1994 289 143 561 1088 192 85 79 19 69 688 319 298 69 344 79 19 344 158 38 85 85 40 79 19 19 262 496 662 308 9 270 255 20 19 9 262 993 662 154 143 1083 1021 619 299 18 18 36 67 63 40 38 1083 255 154 74 18 71 18 37 280 16 9 18 9 17 12 280 249 289 143 139 561 12 36 997 18 139 997 37 36 70 498 9 997 18 72 72 37 37 18 67 12 139 134 24 36 69 134 3 134 6

138 262 993 331 39 138 529 1986 331 76 69 529 1986 82 20 138 1051 1986 164 76

36 270 255 77 19 143 1083 510 154 38

688 319 147 279 4206 3972 1324 152 143 2166 4086 1239 148 289 1122 1994 289 143 561 2176 12 1376 638 596 279 4206 3972 1324 308 143 2166 2043 2478 299 289 1122 3988 289 143 561 2176 24 688 319 38 1376 319 38 170 40 170 79 38 74 36 69 9 18 9 36 1051 1986 331 152 288 541 1021 309 148 37 36 2103 1986 662 308 576 2166 1021 619 299 143 70 262 248 164 39 288 1083 510 154 38 18 529 248 331 152 288 1083 255 308 148 71 1051 1986 164 20 1051 1986 331 39 143 270 255 154 74 143 1083 255 154 74 37 37 9 9 18 36 1994 72 143 139 272 3 1994 143 289 280 272 3 489 37 489 72 997 72 1994 72 72 69 143 69 72 72 272 3 282 6

344 158 19 344 158 38

69 272 12 131 272 12

in mmol/L,

the number of microorganisms in mL ranged from 104105.

143.3, 139.0, 127.9, 127.8, 124.4, 120.5, 115.2, 112.5, 86.6, 81.1, 56.4, 44.5, 31.9, 26.1, 20.3, 19.4 (CH3). 5.2. Antimicrobial activity Microorganisms used: eleven standard strains representative of cocci; Micrococcus luteus ATCC 9341, Staphylococcus epidermidis ATCC 12228, Staphylococcus aureus ATCC 6538, rods; Pseudomonas aeruginosa ATCC 15442, Proteus vulgaris NCTC 4635, Klebsiella pneumoniae ATCC 4352, Escherichia coli NCTC 8196, Serratia marcescens ATCC 8100, yeast-like fungi; Candida albicans ATCC 10231, Rhodotorula rubra PhB, and dermatophytes Trichophyton mentagrophytes var. gypseum ATCC 9533. Standard strains were supplied by National Collection of Type Cultures (NCTC), London and American Type Culture Collection (ATCC). Rhodotorula rubra (PHB) strain was taken from the Department of Pharmaceutical

Microbiology, K. Marcinkowski University of Medical Sciences, Poznan. Antimicrobial activity was determined by the tube dilution method. The method shows, the lowest concentration of a chloride inhibiting cell multiplication (MIC) or killing them (MBC). Two-fold dilutions of the chlorides were prepared in the Mueller-Hinton broth medium (bacteria) or in the Sabouraud broth medium (fungi). A suspension of the standard microorganisms prepared from 24 h cultures of bacteria in the Mueller-Hinton broth medium and from 5 and 10 day cultures in the Sabouraud agar medium for fungi at a concentration of 105 cfu/mL were added to each dilution in a 1:1 ratio. Growth (or its lack) of the microorganisms was determined visually after incubation for 24 h at 37 C (bacteria) or 510 days at 2830 C (fungi). The lowest concentration at which there was no visible growth (turbidity) was taken as the MIC.

770

Figure 1. The MIC mean values for (alkoxymethyl)dimethyl{5-[(4-chlorophenyl)azo]-2-hydroxybenzyl}ammonium chlorides (35).

Figure 2. The MIC mean values for (alkoxymethyl)dimethyl{2-hydroxy-5-[(4-methylphenyl)azo]benzyl}ammonium chlorides (79).

Then from each tube, one loopful was cultured on an agar medium with inactivates [14] (0.3% lecithin, 3% polysorbate 80 and 0.1% cysteine L) and incubated for 48 h at 37 C (bacteria) or for 510 days at 2830 C (fungi). The lowest concentration of the chloride supporting no colony formation was dened as the MBC.

Acknowledgements This investigation received nancial support from the Polish Committee of Scientic Research, Grant KBN 3 T09B 010 15.

771 References
[1] [2] [3] [4] [5] [6] [7] [8] Dunn C.C., Proc. Soc. Exp. Biol. Med. 35 (1936) 427429. Domagk G., Dtsch. Med. Wochenschr. 61 (1935) 829832. Beasley R., Fishwick D., Miles J.F., Hendeles L., Pharmacotherapy 18 (1998) 130139. Tramontini M., Synthesis (1973) 703775. Tramontini M., Angiolini L., Tetrahedron 46 (1990) 17911837. Tramontini M., Angiolini L., Mannich-Bases, Chemistry and Uses, CRC, Boca Raton USA, 1994. Arend M., Westermann B., Risch N., Angew. Chem. Int. Ed. 37 (1998) 10451070. Dimmock J.R., Erciyas E., Kumar P. et al., Eur. J. Med. Chem. 32 (1997) 583594. [9] [10] Dimmock J.R., Pandeya S.N., Allen T.M., Koa G.Y., Pharmazie 53 (1998) 201202. a. Kimich C., Ber. 8 (1875) 10261032; b. Nolting E., Kohn O., Ber. 17 (1884) 351369; c. Heumann K., Oeconomides L., Ber. 20 (1887) 904909; d. Krause M., Ber. 32 (1899) 124127; e. Farmer R., Hantzsch A., Ber. 32 (1899) 30893101. Tomlinson E., Brown M.R.W., Davis S.S., J. Med. Chem. 20 (1977) 12771282. Heir E., Sundheim G., Holck A.L., J. Appl. Bacteriol. 79 (1995) 149156. Bedford C.D., Harris R.N., Howd R.A. et al., J. Med. Chem. 32 (1989) 493516. Skrzypczak A., Brycki B., Mirska I., Pernak J., Eur. J. Med. Chem. 32 (1997) 661668.

[11] [12] [13] [14]

Eur. J. Med. Chem. 34 (1999) 773779 1999 Editions scientiques et mdicales Elsevier SAS. All rights reserved

773

New products

New 2-sulfonamidothiazoles substituted at C-4: synthesis of polyoxygenated aryl derivatives and in vitro evaluation of antifungal activity
Pierre Beucheta, Martine Varache-Lembgea, Arlette Neveua, Jean-Michel Lgerb, Joseph Vercauterena, Stphane Larrouturea, Grard Deffieuxa, Alain Nuhricha*
a

Groupe dtude des Substances Naturelles dIntrt Thrapeutique (GESNIT), Facult de Pharmacie, Universit Victor-Segalen (Bordeaux 2), 146, rue Lo-Saignat, 33076 - Bordeaux cedex, France b Laboratoire de Chimie Analytique, Facult de Pharmacie, Universit Victor-Segalen (Bordeaux 2), 3, Place de la Victoire. 33076 -Bordeaux cedex, France (Received 29 September 1998; revised 4 February 1999; accepted 10 February 1999)

Abstract Polymethoxylated and polyhydroxylated derivatives of 2-amino-4-arylthiazoles bearing a halogenobenzenesulfonamide moiety at position 2 were synthesized as azole antifungal analogues. X-ray crystallography studies revealed the predominance of the 2-imino-2,3dihydrothiazole form in the amino/imino tautomerism. In vitro assays against various pathogenic fungal strains (Candida and Trichophyton species) showed no activity in comparison to econazole as reference. These results are discussed on the basis of the estimated global lipophilicity of the molecules (Rekkers method) and the -electron distribution (Mulliken population analysis, AM1 method) within the ve-membered heterocycle. 1999 Editions scientiques et mdicales Elsevier SAS 2,4-disubstituted thiazoles / sulfonamides / antifungal testing

1. Introduction The most commonly used imidazole antifungal agents, such as miconazole 1, econazole 2 and ketoconazole 3 (gure 1), exhibit side effects including interference with steroidogenesis or hepatic toxicity [1]. Over the past few years, the discovery of triazole compounds including itraconazole 4 and uconazole 5 with better tolerance gave rise to the synthesis of numerous related analogues [2]. Since the sulfonamide moiety seems to be a possible pharmacophore in fungicidal agents [3], we decided to introduce this functionality in position 2 of a thiazole nucleus. On the other hand, the lack of hydrosolubility which limits systemic distribution of azoles led us to include hydrophilic fragments and thus to synthesize new thiazole derivatives with a polyoxygenated phenyl component (69), as represented in gure 1.
Figure 1. Structure of azole antifungals (15) and target compounds (69).

*Correspondence and reprints

774
Table I. Chemical and physical properties of compounds 69.

Compound 6a 6b 6c 6d 7a 7b 7c 7d 8a 8b 8c 8d 9a 9b 9c 9d
a b

(RO)n 2,5-OMe

Hal 4-Cl 4-F 2,4-Cl 2,4-F 4-Cl 4-F 2,4-Cl 2,4-F 4-Cl 4-F 2,4-Cl 2,4-F 4-Cl 4-F 2,4-Cl 2,4-F

M.p.(C) 130 136 148150 140 260 218 240 232 b b 165166 178180 220 218220 b 219

Yield (%) 75 36 51 31 45 47 56 78 58 67 56 20 67 53 65 84

Formula (MW)

log Pcalc 2.44 1.92 3.18 2.13 1.25 0.73 1.99 0.94 2.51 1.99 3.25 2.20 0.73 0.20 1.47 0.42

C17H15ClN2O4S2 (410.90) C17H15FN2O4S2 (394.43) C17H14Cl2N2O4S2 (445.35) C17H14F2N2O4S2 (412.44) C15H11ClN2O4S2 (382.84) C15H11FN2O4S2 (366.39) C15H10Cl2N2O4S2 (417.29) C15H10F2N2O4S2 (384.38) C18H17ClN2O5S2 (440.92) C18H17FN2O5S2 (424.47) C18H16Cl2N2O5S2 (475.37) C18H16F2N2O5S2 (442.46) C15H11ClN2O5S2 (398.85) C15H11FN2O5S2 (382.39) C15H10Cl2N2O5S2 (433.29) C15H10F2N2O5S2 (400.38)

2,5-OH

3,4,5-OMe

3,4,5-OH

All compounds were analysed for C, H, N, S and when present Cl or F; analytical results are within 0.4% of the theoretical values. Decomposition over 80 C. cAccording to Rekker fragmental system [10].

2. Chemistry The target compounds (table I) were synthesized following gure 2. The key step in the preparation of derivatives 6 and 8 resulted in the reaction of bromomethylketones 13 with the appropriate arylsulfonylthioureas 12 in reuxing THF, according to a modied Hantzsch condensation [4]. The thioureas 12 were obtained by reaction of N-cyanoarylsulfonamide sodium salts 11 with thiosulfuric acid (generated in situ upon treatment of sodium thiosulfate with aqueousH2SO4 [5]). 3. Structural data As an example, 1H and 13C NMR data for 6a are shown in table II. The chemical shifts for the carbon atoms at positions 7, 8 and 9 are consistent with data of substituted thiazoles [6] and therefore conrm the heterocyclisation reaction. The main information is provided by the long-range 1H -13C correlations (HMBC spectrum). The quaternary carbon 7 exhibits correlation with proton H6, belonging to the oxygenated ring The 7,9-

functionalization on the heterocyclic ring was therefore established. X-ray diffraction analysis of 6b (gure 3) conrmed a structure close to that of a thiazoline derivative [7]. Noteworthy, the exocyclic C(9)-N(12) bond shows signicant shortening from the value of the endocyclic N(8)-C(9) (table III). Consequently, this phenomenon is indicative of an exocyclic C=N double bond and conrms that the imino form is the predominant tautomer, at least in the crystalline state. This conclusion agrees with previous reports on 2-substituted thiazoles [8, 9]. 4. Antifungal evaluation and discussion The antifungal activity of compounds 69 was evaluated in vitro, against yeasts (Candida albicans ATCC 10231 and Candida krusei CBS 573) and lamentous fungi (Trichophyton mentagrophytes IP 1468-83 and T. rubrum IP 2073-92), according to described methods [11, 12]. By comparison with econazole, used as reference substance, all the tested thiazoles were found inactive (data not shown). These results conrm that lipophilicity plays a fundamental role for antifungal

775 antifungals, might be a determining factor in the biological response. 5. Experimental protocols Melting points were determined with a Ker apparatus and are uncorrected. NMR spectra were obtained in CDCl3, at 29 C, using a Bruker AMX 500 spectrometer (1H: 500 MHz, 13C: 125 MHz). Chemical shifts are given in ppm from TMS as an internal standard (coupling constants J, in Hz). Multiplicities in 13C NMR spectra were derived from JMOD experiments. When necessary, HMQC, HMBC and 1H-1H COSY experiments were used for structural assignments. 5.1. Chemistry 5.1.1. General procedure for N-aminothioxomethylarylsulfonamides 12ad To a stirred suspension of sodium cyanamide [16] (0.08 mol) in anhydrous Et2O (20 mL) was added appropriate arylsulfonyl chloride 10ad (0.04 mol). Stirring was continued for 6 h under reux, and the insoluble Na salt of N-cyanoarylsulfonamide 11ad was washed (dry ether, then acetone) and dried (yields: 6575%). A cooled solution (0 C) of the convenient sodium salt 11 (20 mmol) in H2O (10 mL) was neutralized with 2 N H2SO4 (10 mL). Solid Na2S2O3, 5H2O (60 mmol) was added in one portion under stirring and the mixture was treated with more 2 N H2SO4 (15 mL). After further stirring overnight at 20 C, the resulting precipitate was washed with cold H2O (2 10 mL), and recrystallized from methanol/H2O (50:50) to afford the expected product 12 as white needles (4561%). 5.1.1.1. N-Aminothioxomethyl-4-chlorobenzenesulfonamide 12a M.p. = 149151 C. 1H-NMR 7.61 (d, J = 8.8, 2H); 7.94 (d, J = 8.8, 2H); 13C- NMR 130.0 (C2, C6); 130.6 (C3, C5); 139.3 (C4); 141.3 (C1); 182.3 (C7). 5.1.1.2. N-Aminothioxomethyl-4-uorobenzenesulfonamide 12b M.p. = 120 C. 1H- NMR 7.33 (dd, J = 8.8, 8.6, 2H); 8.02 (dd, J = 9.0, 5.0, 2H); 13C- NMR 117.5 (C3, C5); 131.4 (C2, C6); 136.8 (C1); 167.1 (C4); 182.4 (C7). 5.1.1.3. N-Aminothioxomethyl-2,4-dichlorobenzenesulfonamide 12c M.p. = 205 C. 1H- NMR 7.81 (dd, J = 2.0, 8.6, 1H); 7.94 (d, J = 2.0, 1H); 8.34 (d, J = 8.6, 1H); 13C- NMR 128.5 (C2); 132.5 (C4); 133.8 (C6); 134.3 (C1); 136.6 (C5); 140.9 (C3); 181.1 (C7).

Figure 2. General synthetic pathway for 2,4-functionalized thiazoles 69.

properties [13]: the estimated log P values for compounds 69, ranging from 0.23.25 (table I) are signicantly lower than that calculated for econazole (log P = 5.17). We have also investigated the electron density distribution in econazole and compound 6b, respectively. The AM1 method [14] with Mulliken population analysis was selected for its usefulness in predicting the aromaticity of 5-membered heterocycles [15]. From the data presented in gure 4, it is shown that the imidazole ring of econazole displays a high level of aromaticity ( overlap populations ranging from 2340%), while in the thiazole moiety, the -electrons are unequally spread over the intracyclic bonds (furthest values: 1144%), thus indicative of a weak aromatic character for 6b. In conclusion, it appears that the association of a polyoxygenated phenyl ring and an arylsulfonamido moiety on a thiazolic framework does not lead to antifungal activity. The electronic properties of the ve-membered ring, which differ strongly from those of the imidazole

776
Table II. NMR dataa and main HMQC, HMBC, and 1H-1H COSY correlations for 6a (CDCl3).

Positions 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
a d

13

HMQC C8, H8 C3, H3 C4, H4 C6, H6 C8, H8 C11, H11 C12, H12 C14, H14 C15, H15

HMBC C1, H8 C2, H6; C2, H4 C4, H6 C5, H6; C5, H4 C6, H4 C7, H6; C7, H8 C9, H8 C10, H12; C10, H14 C11, H15 C12, H14 C13, H11; C13, H15 C14, H12 C15, H11

6.94 6.92 7.02 6.62 7.91 7.42 7.42 7.91

(d, 8.8)b (dd, 9.1, 2.6)b (d, 2.6)b (s) (d, 8.6)b (d, 8.6) (d, 8.6) (d, 8.6)b

117.0 149.6 116.0 113.0 154.1 113.2 134.4 102.2 168.4 140.8 127.8 128.9 138.4 128.9 127.8

Chemical shifts, ppm (multiplicity, J in Hz). b1H-1H COSY correlations. cThe methoxy protons appeared at 3.79 and 3.91 ppm, as singlets. The NH-proton was not obviously identiable.

5.1.1.4. N-Aminothioxomethyl-2,4-diuorobenzenesulfonamide 12d M.p. = 102 C. 1H- NMR 7.037.15 (m, 2H); 7.95 (ddd, J = 8.0, 6.0, 6.0, 1H); 13C- NMR 106.4 (C3); 112.4 (C5); 122.7 (C1); 131.8 (C6); 160.1 (C2); 166.9 (C4); 179.6 (C7).

Table III. Selected bond lengths () for 6ba C(1)C(7) C(7)C(11) C(7)N(8) N(8)C(9) C(9)N(12) C(9)S(10)
a

1.477 (4) 1.322 1.395 1.346 1.325 1.737 (4) (4) (3) (4) (3)

S(10)C(11) N(12)S(13) S(13)O(15) S(13)O(14) S(13)C(16)

1.740 (3) 1.588 1.441 1.444 1.762 (2) (2) (2) (3)

Estimated standard deviations are given in brackets

Figure 3. Molecular conformation of compound 6b with special atom-numbering used in the crystallographic analysis. (displacement ellipsoids are shown for non-hydrogen atoms at 50% probability level).

Figure 4. -Interatomic overlap populations (%) calculated using the Mulliken population analysis for the heterocycles of econazole 2 and compound 6b (AM1 results from X-ray structure data). Atomic charges are indicated in italic numbers.

777 5.1.2. General procedure for 1-aryl-2-bromoethanones 13 These products were obtained by bromination of commercial acetophenones in CCl4 [17], and chromatographed on silica gel column (eluent: hexane/CH2Cl2, 70: 30) prior to use. 5.1.2.1. 1-(2,5-Dimethoxyphenyl)-2-bromoethanone 13a M.p. = 80 C. 1H- NMR 3.80 (s, 3H); 3.93 (s, 3H); 4.65 (s, 2H); 7.05 (m, 2H); 7.40 (d, J = 8.0, 1H). 5.1.2.2. 1-(3,4,5-Trimethoxyphenyl)-2-bromoethanone 13b M.p. = 71 C. 1H- NMR 4.00 (s, 9H); 4.46 (s, 2H); 7.30 (s, 2H). 5.1.3. General procedure for N-(4-aryl-2,3dihydrothiazol-2-ylidene) arylsulfonamides 6ad and 8ad A suspension of N-aminothioxomethylbenzenesulfonamide 12ad (4 mmol) and 1-aryl-2-bromoethanone 13a or 13b (4 mmol) in anhydrous THF (30 mL) was heated at reux for 5 h under constant stirring. The mixture was then cooled to 20 C and after removal of the volatile materials under vacuo, the crude residue was chromatographed on a silica-gel column (eluent: CH2Cl2/EtOAc, 80:20), then recrystallized from heptane (yields: 4051%). 5.1.3.1. N-[2,3-Dihydro-4-(2,5-dimethoxyphenyl) thiazol-2-ylidene]-4-chlorobenzenesulfonamide 6a (see table II for spectral data) 5.1.3.2. N-[2,3-Dihydro-4-(2,5-dimethoxyphenyl) thiazol-2-ylidene]-4-uorobenzenesulfonamide 6b 1 H- NMR 3.80 (s, 3H); 3.92 (s, 3H); 6.62 (s, 1H); 6.92 (dd, J = 9.0, 2.6, 1H); 6.95 (d, J = 9.0, 1H); 7.03 (d, J = 2.6, 1H); 7.13 (dd, J = 8.8, 8.6, 2H); 7.99 (dd, J = 8.8, 5.2, 2H). 13C- NMR 55.8; 56.4; 102.2; 113.0; 113.3; 115.6; 115.8; 116.0; 117.0; 128.9; 129.0; 134.4; 138.5; 149.6; 154.1; 164.8 (d, 1JC, F = 250); 168.4. 5.1.3.3. N-[2,3-Dihydro-4-(2,5-dimethoxyphenyl) thiazol-2-ylidene]-2,4-dichlorobenzenesulfonamide 6c 1 H-NMR 3.80 (s, 3H); 3.92 (s, 3H); 6.64 (s, 1H); 6.92 (dd, J = 9.1, 2.6, 1H); 6.95 (d, J = 8.8, 1H); 7.04 (d, J = 2.4, 1H); 7.34 (dd, J = 8.0, 2.0, 1H); 7.47 (d, J = 2.0, 1 H); 8.16 (d, J = 8.5, 1H). 13C-NMR 55.8; 56.3; 102.2; 113.0; 113.3; 116.0; 117.0; 126.8; 131.1; 131.2; 133.5; 134.5; 138.4; 149.7; 154.1; 169.0. 5.1.3.4. N-[2,3-Dihydro-4-(2,5-dimethoxyphenyl) thiazol-2-ylidene]-2,4-diuorobenzenesulfonamide 6d 1 H-NMR 3.80 (s, 3H); 3.92 (s, 3H); 6.65 (s, 1H); 6.89 (m, 2H); 6.95 (m, 1H); 7.05 (d, J = 2.7, 1H); 8.03 (m, 1H); 13C-NMR 55.8; 56.3; 102.3; 105.2 (dd, 2JC, F = 25, 2JC, F = 25); 111.1; 113.0; 113.3; 116.0; 117.0; 126.7; 131.1; 134.5; 149.7; 154.1; 160.1 (dd, 1JC, F = 250, 3 JC, F = 10); 165.1 (dd, 1JC, F = 250, 3JC, F = 10); 169.0. 5.1.3.5. N-[2,3-Dihydro-4-(3,4,5-trimethoxyphenyl) thiazol-2-ylidene]-4-chlorobenzenesulfonamide 8a 1 H-NMR 3.88 (s, 9H); 6.46 (s, 1H); 6.71 (s, 2H); 7.38 (d, J = 8.6, 2H); 7.89 (d, J = 8.6, 2H). 13C-NMR 56.3; 60.8; 101.2; 103.2; 123.9; 127.9; 128.9; 137.5; 138.6; 139.5; 140.2; 153.7; 168.7. 5.1.3.6. N-[2,3-Dihydro-4-(3,4,5-trimethoxyphenyl) thiazol-2-ylidene]-4-uorobenzenesulfonamide 8b 1 H-NMR 3.85 (s, 9H); 6.46 (s, 1H); 6.73 (s, 2H); 7.07 (dd, J = 8.6, 8.6, 2H); 7.97 (dd, J = 8.8, 5.1, 2H). 13 C-NMR 56.3; 60.8; 101.3; 103.2; 115.7; 115.9; 124.5; 129.1; 137.6; 137.8; 139.2; 153.7; 164.8 (d, 1JC, F = 250); 168.7. 5.1.3.7. N-[2,3-Dihydro-4-(3,4,5-trimethoxyphenyl) thiazol-2-ylidene]-2,4-dichlorobenzenesulfonamide 8c 1 H-NMR 3.88 (s, 9H); 6.49 (s, 1H); 6.71 (s, 2H); 7.32 (d, J = 4.2, 1H); 7.45 (s, 1H); 8.10 (d, J = 4.2, 1H). 13 C-NMR 56.3; 60.9; 101.4; 103.2; 123.9; 126.8; 131.1; 131.2; 133.4; 137.1; 137.8; 138.6; 139.5; 153.8; 169.3. 5.1.3.8. N-[2,3-Dihydro-4-(3,4,5-trimethoxyphenyl) thiazol-2-ylidene]-2,4-diuorobenzenesulfonamide 8d 1 H-NMR 3.85 (s, 9H); 6.50 (s, 1H); 6.67 (s, 2H); 6.82 (m, 1H); 6.91 (m, 1H); 7.98 (m, 1H); 13C-NMR 56.3; 60.8; 101.3; 103.2; 105.4 (dd, 2JC, F = 25, 2JC, F = 23); 111.2 (d, 2JC, F = 22); 123.9; 126.7 (d, 2JC, F = 23); 131.2 (d, 3JC, F = 10); 137.3; 139.3; 153.7; 159.9 (dd, 1JC, 3 1 3 F = 250, JC, F = 10); 165.3 (dd, JC, F = 250, JC, F = 10); 169.4. 5.1.4. General procedure for phenolic compounds 7ad and 9ad To a cooled (0 C) solution of 6ad or 8ad (1 mmol) in anhydrous CH2Cl2, boron tribromide was added dropwise (2.2 mmol for dimethoxy compounds 6ad, or 3.3 mmol for trimethoxy compounds 8ad) under nitrogen. The mixture was stirred at 0 C for 1 h, then quenched cautiously with H2O (50 mL). Following 15 min hydrolysis, the resulting precipitate was collected, washed with cold H2O and recrystallized from 50% ethanol, affording the expected polyhydroxylated compound in 5378% yield.

778 5.1.4.1. N-[2,3-Dihydro-4-(2,5-dihydroxyphenyl) thiazol-2-ylidene]-4-chlorobenzenesulfonamide 7a 1 H-NMR 4.77 (s, 2H); 6.69 (dd, J = 8.7, 2.8, 1H); 6.74 (d, J = 8.7, 1H); 6.88 (d, J = 2.8, 1H); 6.93 (s, 1H); 7.50 (d, J = 8.6, 2H); 7.88 (d, J = 8.6, 2H). 13C-NMR 105.3; 114.7; 117.1; 118.3; 118.7; 129.1; 130.0; 137.1; 139.4; 142.2; 148.8; 151.6; 170. 5.1.4.2. N-[2,3-Dihydro-4-(2,5-dihydroxyphenyl) thiazol-2-ylidene]-4-uorobenzenesulfonamide 7b 1 H-NMR 4.76 (s, 2H); 6.69 (dd, J = 8.7, 2.8, 1H); 6.74 (d, J = 8.7, 1H); 6.89 (d, J = 2.8, 1H); 6.92 (s, 1H); 7.22 (dd, J = 8.9, 8.8, 2H); 7.99 (dd, J = 8.9, 5.1, 2H). 13 C-NMR 105.2; 114.7; 116.7; 116.9; 117.1; 118.3; 118.7; 130.2; 130.3; 139.7; 148.7; 151.6; 166.2 (d, 1JC, F = 250); 170.0. 5.1.4.3. N-[2,3-Dihydro-4-(2,5-dihydroxyphenyl) thiazol-2-ylidene]-2,4-dichlorobenzenesulfonamide 7c 1 H-NMR 6.93 (dd, J = 8.7, 2.9, 1H); 7.03 (d, J = 8.7, 1H); 7.24 (s, 1H); 7.25 (d, J = 2.9, 1H); 7.69 (dd, J = 8.5, 2.0, 1H); 7.77 (d, J = 2, 1H); 8.3 (d, J = 8.5, 1H). 13 C-NMR 104.8; 114.4; 116.8; 118.5; 118.6; 127.9; 131.9; 132.2; 134.0; 138.7; 140.1; 147.9; 151.7; 171.4. 5.1.4.4. N-[2,3-Dihydro-4-(2,5-dihydroxyphenyl) thiazol-2-ylidene]-2,4-diuorobenzenesulfonamide 7d 1 H-NMR 6.69 (dd, J = 8.8, 2.8, 1H); 6.75 (d, J = 8.8, 1H; 6.90 (d, J = 2.8, 1H); 6.95 (s, 1H); 7.057.13 (m, 2H); 6.69 (dd, J = 8.4, 6.3, 1H); 13C-NMR 105.5; 106.5 (dd, 2JC, F = 25; 2JC, F = 23); 112.2; 114.7; 117.0; 118.3; 118.8; 127.7; 132.3; 137; 148.8; 151.7; 161.2 (dd, 1JC, F = 250, 3JC, F = 10); 166.9 (dd, 1JC, F = 250, 3JC, F = 10); 170.4. 5.1.4.5. N-[2,3-Dihydro-4-(3,4,5-trihydroxyphenyl) thiazol-2-ylidene]-4-chlorobenzenesulfonamide 9a 1 H-NMR 6.55 (s, 1H); 6.57 (s, 2H); 7.49 (d, J = 8.8, 2H); 7.86 (d, J = 8.8, 2H). 13C-NMR 101.7; 106.4; 121.4; 129.0; 130.0; 136.1; 139.4; 139.7; 142.3; 147.4; 171.1. 5.1.4.6. N-[2,3-Dihydro-4-(3,4,5-trihydroxyphenyl) thiazol-2-ylidene]-4-uorobenzenesulfonamide 9b 1 H-NMR 6.55 (s, 1H); 6.57 (s, 2H); 7.21 (dd, J = 8.8, 8.7, 2H); 7.93 (dd, J = 8.8, 5.2, 2H); 13C-NMR 101.6; 106.4; 116.7; 116.8; 121.4; 130.1; 130.2; 136.0; 139.7; 139.8; 147.4; 166.2 (d, 1JC, F = 250); 171.0. 5.1.4.7. N-[2,3-Dihydro-4-(3,4,5-trihydroxyphenyl) thiazol-2-ylidene]-2,4-dichlorobenzenesulfonamide 9c 1 H-NMR 6.58 (s, 2H); 6.59 (s, 1H); 7.45 (dd, J = 8.5; 2.0, 1H); 7.59 (d, J = 2.0, 1H); 8.10 (d, J = 8.5, 1H);
13

C-NMR 101.9; 106.4; 121.3; 129.0; 132.4; 134.5; 136.1; 139.5; 139.6; 139.7; 147.5; 171.4.

5.1.4.8. N-[2,3-Dihydro-4-(3,4,5-trihydroxyphenyl) thiazol-2-ylidene]-2,4-diuorobenzenesulfonamide 9d 1 H-NMR 6.57 (s, 2H); 6.60 (s, 1H); 7.067.11 (m, 2H); 7.98 (ddd, J = 8.0, 6.0, 6.0, 1H); 13C-NMR 102.0; 106.3 (dd, 2JC, F = 25, 2JC, F = 22); 106.4; 112.2 (d, 2 JC, F = 22); 121.3; 126.9; 132.3; 136.0; 139.7; 147.5; 161.2 (dd, 1JC, F = 250, 3JC, F = 10); 166.8 (dd, 1JC, F = 250, 3JC, F = 10); 171.4. 5.2. X-ray structure determination A suitable crystal of 6b (size 0.30 0.15 0.01 mm) was obtained from a CHCl3 solution. C17H15FN2O4S2, M = 394.43, triclinic, space group P-1, a = 7.026(4), b = 9.068, c = 14.442(7) , = 88.74(4), = 77.70(5), = 80.63(4), V = 886.9(8) 3, Z = 2, D = 1.477 Mg m3. The data collection was performed on a Nonius CAD4 diffractometer using graphite monochromated Cu K radiation ( = 1.54178 ). 2 977 independent reexions were measured, of which 2 461 were used in the renements. Renement was carried out by the full-matrix leastsquares method based on F2 with the SHELXL-93 program [18]. (Final R indices: R1 = 0.0466, WR2 = 0.1546). Full experimental details are available as supplementary material. 5.3. Computational studies Calculations were performed using the MOPAC [19] program implemented in Chem3D, V 3.5 [20]. The atomic coordinates of the heavy atoms considered in the calculations were those obtained by crystallographic data (compound 6b: see above; econazole: see reference [21]). The atomic charges were analysed using the AM1 potential function with the Mulliken charges option and the PI keyword. Acknowledgements The authors thank E. Bezanon for his able technical assistance in providing high resolution NMR spectra. The secretarial help of A. Carrre is gratefully acknowledged. References
[1] [2] [3] Hoeprich P.D., Prog. Drug Res. 44 (1995) 87127. Lartey P.A., Moehle C.M., Ann. Rep. Med. Chem. 32 (1997) 151160. Habib O.M.O., Kandel E.M., Askou S.A., Moawad E.B., Hung. J. Ind. Chem. 14 (1986) 477483.

779
[4] [5] [6] [7] [8] [9] [10] [11] Das Gupta P.K., Gupta P., J. Indian Chem. Soc. 23 (1946) 1315. Fldi Z., Fldi T., Fldi A., Acta Chim. Acad. Sci. Hung. 13 (1957) 111115. Faure R., Galy J.P., Vincent E.J., Elguero J., Can. J. Chem. 56 (1978) 4655. Kalcheva V., Tosheva M., Hadjieva P., Liebigs Ann. Chem. (1993) 13191322. Form G.R., Paper E.S., Downie T.C., Acta Cryst. B 30 (1974) 342348. Argay G., Kalman A., Lazar D., Ribar B., Toth G., Acta Cryst. B 33 (1977) 99105. Rekker R.F., Mannhold R., Calculation of Drug Lipophilicity. The Hydrophobic Fragmental Approach, VCH, Weinheim, 1992. Scheven M., Senf L., Mycoses 37 (1994) 205207. [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] Drouhet E., Barale T., Bastide J.E. et al., Bull. Soc. Fr. Mycol. Med. 10 (1981) 131134. Wahbi Y., Tournaire C., Caujolle R., Payard M., Linas M.D., Seguela J.P., Eur. J. Med. Chem. 29 (1994) 701706. Dewar M.J.S., Zoebisch E.G., Healy E.F., Stewart J.P., J. Am. Chem. Soc. 107 (1985) 39023909. Bean G.P., J. Org. Chem. 58 (1993) 73367340. Drechsel E., J. Prakt. Chem. 11 (1875) 284353. Menass R., Klein G., Erlenmeyer H., Helv. Chim. Acta 38 (1955) 12891291. Sheldrick G.M., SHELXL-93, University of Gttingen, (1993). Stewart J.J.P., QCPE Bull. 9 (1989) 1015. Chem3D: Cambridge Soft Corporation, Cambridge, (1996). Freer A.A., Pearson A., Salole E.G., Acta Cryst. C42 (1986) 13501352.

Eur. J. Med. Chem. 34 (1999) 783790 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

783

Original article

Demonstration of the strength of focused combinatorial libraries in SAR optimisation of growth hormone secretagogues
Michael Ankersena*, Birgit Sehested Hansenb, Thomas Kruse Hansena, Jesper Laua, Bernd Peschkea, Kjeld Madsena, Nils Langeland Johansena
b a MedChem Research, Health Care Discovery, Novo Nordisk A/S, Novo Nordisk Park, DK-2760 Mlv, Denmark Assay and Cell Technology, Health Care Discovery, Novo Nordisk A/S, Novo Nordisk Park, DK-2760 Mlv, Denmark

(Received 9 December 1998; accepted 10 February 1999)

Abstract A series of 96 growth hormone secretagogues, derived from ipamorelin are described. The compounds are prepared as a 6 4 4 member library on solid support using a PAL resin. The compounds are all acylated dipeptides, based on two aromatic amino acids and a free amino N-terminal. All compounds are characterised by HPLC, LC-MS and their ability to release GH in a pituitary cell based assay. The most potent compounds show EC50 values at 1 nM and are full agonists. We demonstrate the strength of focused combinatorial libraries and conrm the pitfall in broad SAR exploration by giving examples where selected fragments obviously show poor receptor interaction except in very dened structural arrangements. 1999 ditions scientiques et mdicales Elsevier SAS GHRP-2 / GHRP-6 / MK677 / ipamorelin / growth hormone secretagogue / combinatorial libraries

1. Introduction Since Bowers [1] in 1977 disclosed the discovery of a new class of compounds with the ability to release growth hormone (GH) from the pituitary in a manner distinct from GH releasing hormone (GHRH), an increasing interest in this type of compound has emerged. These compounds, called GH releasing peptides (GHRP) or GH secretagogues (GHS) [2], may offer a new potential treatment for a number of conditions due to their ability to maintain the physiological pulsatile pattern of GH secretion. Such conditions could be GH deciency, osteoporosis or obesity [35]. The most prominent members of this class of GHSs are GHRP-2, GHRP-6 and MK677 [6]. Recently we discovered a new series of acylated dipeptides derived from ipamorelin [7] via a peptidomimetic strategy [810]. Representatives of this series are the acylated dipeptides 1, 2 and 3, which are all based on a free aminogroup at the N-terminal, two N-methylated aromatic D-amino acids and methylamide in the C-terminal (gure 1).
*Correspondence and reprints

To optimize this series, a compound library of single isolated compounds was constructed in a 6 4 4 matrix, based on a total combination of six N-terminals and two sets of N-methylated aromatic D-amino acids (4 building blocks in each), to give 96 compounds with the generic structure as shown in gure 2. Of the 96 compounds in the library, 10 compounds were prepared in solution and have been described previously [810], and 86 compounds were prepared on solid support. Herein we report the synthesis of the 86 compounds prepared on solid support and discuss the structure-activity relationship of the entire library (ie. 96 compounds). 2. Chemistry The 86 compounds, prepared on solid phase, were prepared in parallel on N-Me-PAL [11, 12] (gure 3) as solid support using an Fmoc/Boc peptide synthesis strategy with traditional amide coupling conditions [O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexauorophosphate, 1-hydroxy-7-azabenzotriazole, diisopropylethylamine in dimethylformamide, deprotection

784

Figure 1. Peptidomimetic GH secretagogues.

with piperidine in dimethylformamide for the Fmocprotected building blocks B (14), i.e. B1, B2, B3 and

B4, and C (14) and deprotection/cleavage with TFA for the Boc-protected A (16) (gures 4 and 5)]. Six Bocprotected N-terminals (A (16)), two sets of four Fmocprotected aromatic D-amino acids [B (14) and C (14) (gure 1)] and N-methylamine as a xed C-terminal were used for the 3 step synthesis. The compounds were prepared in parallel as single compounds on an automated shaker with overnight runs for each coupling. Each well was furnished with 0.045 mmol of N-methyl-PAL-Resin (75 mg, load: 0.6) and each coupling was carried out with 0.09 mmol of the respective building block, 0.09 mmol of O-(7azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexauorophosphate, 0.09 mmol of 1-hydroxy-7-azabenzo-

Figure 2. General scaffold holding three building blocks (Nterminal and two aromatic groups) which constitute the 96 member GH secretagogue library.

785 triazole and 0.18 mmol of diisopropylethylamine in a total of 6 mL of dimethylformamide. After deprotection, the resin was treated with TFA at 0 C and the obtained crude product was applied to a C-18 Sep-Pak Classic Cartridge (Waters) and a gradient on water, acetonitrile and TFA. The yield was measured by weight (table I) after lyophilisation and all compounds were characterised by two HPLC systems (table II). The molecular weight of all compounds sent for testing was in accordance with the mass obtained by LC-MS.

Figure 3. The N-Me PAL resin used in the preparation of the 96 member GH secretagogue library.

Figure 4. Building blocks A (16), B (14) and C (14). The building blocks A (16) were all Boc-protected and the building blocks B (14) and C (14) were Fmoc protected for synthesis.

786 sequence B (14)-C (14)-NHCH3 usually runs without problems, while the coupling of the last amino acid A (16) and in particular the nal deprotection/cleavage with TFA causes several problems and should be carried out with great care (temp. 100 C). Thus, it is interesting, that three of 12 compounds containing A6 failed, and that only two of 12 compounds gave yields over 5 mg. In contrast 13 of 14 compounds containing A4 gave yields over 5 mg, and if we consider those N-terminals which contain ether linkages, i.e. A3, A4 and A5, and compare them with those containing conjugation, i.e. A1, A2 and A6, we clearly see a better nal yield in the ether linkage group, indicating that the nal yield depended on the N-terminal. All compounds were characterised by two HPLC systems and LC-MS. The retention-times for the prepared compounds are given in table II in two different systems. It should be noted that although the building block A1 is a racemic mixture, only one peak was observed in both HPLC systems. The molecular weight of all compounds was determined by LC-MS (electrospray) and were in accordance with the obtained mass, except in those cases where the synthesis is denoted as failed. In such cases no further characterisation was carried out. HPLC data for the compounds prepared in solution have been described elsewhere, and are not included in table II. The GH releasing properties of the prepared compounds, and those in the series previously prepared (marked with a reference) in solution are presented in table III by EC50 values as is the in vitro potency in a rat pituitary cell assay [7, 13]. Figures 68 are graphic

Figure 5. Reaction scheme for preparation of the 96 member GH secretagogue library.

3. Results and discussion A total of 86 compounds were prepared on solid support. The synthesis of 5 of the 86 compounds (noted SF in table I) failed completely. The remaining 81 compounds were puried by a SepPac Classic Cartridge (Waters) and isolated in amounts ranging from 125 mg and purities from 70100% as shown in table I. Although no characterisation of intermediates was carried out in this study, it is our experience that the preparation of the

Table I. Isolated amount (mg) and purity (%) after SepPac purication of the 86 compounds prepared on solid support. Entry B1C1 B2C1 B3C1 B4C1 B1C2 B2C2 B3C2 B4C2 B1C3 B2C3 B3C3 B4C3 B1C4 B2C4 B3C4 B4C4 A1 4 mg, 100 % 3 mg, 100 % 10 mg, 87 % 2 mg, 100 % 5 mg, 100 % 4 mg, 81 % 5 mg, 100 % 10 mg, 81 % NS 4 mg, 100 % 16 mg, 100 % 8 mg, 92 % 8 mg, 100 % 3 mg, 100 % 6 mg, 100 % 8 mg, 88 % A2 5 mg, 100 % 2 mg, 100 % 8 mg, 70 % 12 mg, 85 % 3 mg, 100 % 3 mg, 95 % 5 mg, 100 % 7 mg, 85 % NS 3 mg, 100 % 13 mg, 90 % 10 mg, 90 % 4 mg, 90 % 6 mg, 95 % 7 mg, 100 % 7 mg, 92 % A3 10 mg, 88 % SF 10 mg, 100 % 14 mg, 90 % 10 mg, 90 % 11 mg, 100 % 24 mg, 93 % 23 mg, 78 % NS 1 mg, 100 % 15 mg, 94 % 13 mg, 95 % 18 mg, 86 % 8 mg, 100 % 2 mg, 100 % 6 mg, 100 % A4 9 mg, 89 % 8 mg, 100 % 11 mg, 93 % 11 mg, 85 % 12 mg, 93 % 7 mg, 90 % 26 mg, 100 % 21 mg, 71 % NS 3 mg, 100 % 18 mg, 100 % 21 mg, 90 % 24 mg, 91 % 11 mg, 90 % 19 mg, 100 % 10 mg, 100 % A5 11 mg, 90 % 6 mg, 100 % 12 mg, 87 % 13 mg, 71 % 12 mg, 95 % 8 mg, 100 % 20 mg, 79 % 12 mg, 89 % NS 3 mg, 100 % 17 mg, 100 % NS 16 mg, 88 % 10 mg, 100 % 21 mg, 84 % SF A6 NS 2 mg, 100 % 2 mg, 100 % 13 mg, 70 % NS SF SF 5 mg, 95 % NS 1 mg, 100 % NS 25 mg, 89 % 3 mg, 75 % 4 mg, 94 % 4 mg, 100 % SF

NS = not synthesised; SF = synthesis failed; purities based on HPLC system A (table II) at 215 nm.

787
Table II. HPLC retention-time (system A, system B; min) of the 86 compounds prepared on solid support. Entry B1C1 B2C1 B3C1 B4C1 B1C2 B2C2 B3C2 B4C2 B1C3 B2C3 B3C3 B4C3 B1C4 B2C4 B3C4 B4C4 A1 31.05, 30.62, 28.85, 34.58, 32.00, 31.60, 31.07, 35.40, NS 31.15, 30.87, 35.25, 31.32, 30.90, 30.82, 34.23, 33.00 32.57 30.67 36.67 33.98 33.57 32.97 37.58 33.10 32.50 36.93 32.92 32.52 31.82 N/A A2 30.32, 29.82, 30.45, 33.92, 31.30, 30.82, 30.25, 34.70, NS 30.45, 30.02, 34.63, 30.62, 30.15, 36.23, 34.21, 32.22 31.73 30.43 35.20 33.23 32.77 32.10 36.08 32.28 31.65 36.38 32.15 31.67 N/A 35.90 A3 30.13, SF 28.23, 33.57, 31.25, 30.62, 29.32, 34.13, NS 30.48, 29.38, 34.30, 30.78, 30.18, 36.15, 33.63, 31.98 29.65 34.47 33.02 32.45 30.97 36.08 32.16 30.85 36.10 32.27 31.72 38.22 35.65 A4 30,07, 29.42, 27.77, 33.30, 31.90, 30.47, 29.45, 27.65, NS 30.47, 29.00, 34.47, 30.85, 30.20, 28.65, 33.73, 31.88 31.23 29.40 35.24 32.73 32.30 30.87 30.18 31.97 30.45 36.17 32.40 31.70 30.02 35.71 A5 29.77, 29.22, 28.50, 27.68, 30.83, 30.33, 30.17, 34.32, NS 30.18, 29.15, NS 30.37, 29.87, 28.80, SF 31.62 31.00 29.23 30.22 32.63 32.10 30.77 36.13 31.68 30.57 31.90 31.35 30.17 A6 NS 29.62, 28.03, 28.48, NS SF SF 27.68, NS 30.57, NS 34.05, 30.42, 30.42, 28.37, SF 31.50 29.77 30.03

30.20 32.18 35.78 32.10 31.62 N/A

NS = not synthesised; SF = synthesis failed; N/A = data not available.

illustrations of all 96 compounds in three different skyline views to clarify the signicance of each group of building blocks. Each bar depicts the potency of the compound expressed as how many times the compound is more potent than 10 nM, i.e., a bar height of 5 equals a potency of 2 nM. This highlights the most potent compounds and neglects the less potent compounds.
Table III. Potency (EC50, nM) of the growth hormone secretagogue library. Mean of at least two separate experiments. Entry B1C1 B2C1 B3C1 B4C1 B1C2 B2C2 B3C2 B4C2 B1C3 B2C3 B3C3 B4C3 B1C4 B2C4 B3C4 B4C4 A1 14 2 600 2 11 75 40 7 13b 125 600 9 285 700 360 8 A2 155 80 inactive 18 2 145 22 9 40b 105 975 40 775 inactive inactive 30 A3 20 SF 40 10 5 70 inactive 10 22b 195 1 850 1 525 840 inactive 31 A4 10 25 1 000 2 13 155 1 370 1 16b 725 150 4 150 inactive inactive 80 A5 220 435 inactive 40 15 28 500 2 13b 620 650 55* 590 2 850 inactive SF A6 2a 70 405 10 75* SF SF 18 18a 75 5a 3 19 120 650 SF

As illustrated in gures 68 and table III, a number of combinations are preferable. The combination of the biphenyl derivative B4 with C1, C2 or C3 gives compounds with potencies in the range of 155 nM independent of the six building blocks A (16), while the combination of the naphthyl-derivative B1 only with C2 or C3 gives compounds with potencies in the range of 275 nM. In the same manner gures 68 illustrate a number of unfavourable combinations. The combination

NS = not synthesised; SF = synthesis failed; N/A = data not available; aprepared in solution and described in ref. [9]; bprepared in solution and described in ref. [10]; *prepared in solution as in refs. [9, 10] but not previously described; inactive dened as EC50 > 2 000 nM.

Figure 6. 3-Dimensional graphic illustration of the potencies of the growth hormone secretagogue library. The y-axis expresses how many times the compound is more potent than 10 nM.

788 six building block A (16) gave compounds with poor potencies, except when combined as A6-B1-C4 (19 nM) or A1-B4-C4 (8 nM). A similar pattern is observed, when the phenyl derivative C3 is combined with B2 or B3, where the combinations with all six building blocks A (16) show compounds with poor potencies, except when combined as A6-B3-C3 (5 nM). The building blocks B2 and B3 give, in general, compounds with poor potencies, except when combined as A1-B2-C1 (2 nM) or A6B3-C3 (5 nM) (gures 6 and 7). By observing each column A1A6 in table III and gure 6, it is interesting to see that the most potent compound in each column never contains the same combination of building blocks, i.e. A1 gives the most potent compound combined with B4-C1, A2 gives the most potent compound combined with B1-C2, and A3 gives the most potent compound combined with B4-C3 etc. Likewise, it is interesting to note that the two building blocks A3 and A4, which give the two most potent compounds in combination with B4-C3 (1 nM) and B4-C2 (1 nM), respectively, give inactive compounds when combined with B3-C2 or B3-C4 (for A3) and B2-C4 or B3-C4 (for A4). Altogether, these observations indicate the danger in conserving one part of the molecule while modifying the other part, without simultaneously modifying the remaining part of the molecule. For instance, it seems obvious that the p-methoxyphenylderivative C4 as a building block has poor receptor interaction since in 23 combinations the compounds show poor potencies, but nevertheless the building block in combination with A1-B4 shows good potency. Also, the combination of B4-C3 with A3 gives a very potent compound (1 nM), while combined with A2 the compound shows poor potency, indicating that A2 as an N-terminal is a bad choice. But when A2 is combined with B1-C2 it gives a compound more potent than when B1-C2 is combined with A3. Other discrepancies may be highlighted studying gures 68 and table III. The use of combinatorial chemistry and highthroughput screening has increased tremendously over the last few years. Combinatorial chemistry began with the concept of huge libraries of mixtures and the deconvolution of biologically active mixtures to detect new leads. Nowadays, the automated parallel synthesis of specially designed and focused small libraries, made up from single compounds, is at the forefront of research [14]. Most combinatorial libraries are prepared for lead-nding purposes, and the diversity of the library is of utmost importance since the chance of nding active compounds in a library increases with the diversity and dissimilarity of the compounds in the library. However, if a known lead structure is to be optimised, the library may

Figure 7. 3-Dimensional graphic illustration of the potencies of the growth hormone secretagogue library. The y-axis expresses how many times the compound is more potent than 10 nM.

of the benzyloxy-derivative B3 with C1 or C4, and the benzothiophene-derivative B2 with C3 or C4, with any of the six building blocks A (16) gave compounds which only showed potencies in the range of 40 nMinactive. It is noteworthy that some building blocks generally give inactive or low potent compounds, except for a few particular combinations, where an active compound can be obtained. This is, for instance, illustrated by the p-methoxyphenyl-derivative C4, which in combination with any of the four building block B (14) and any of the

Figure 8. 3-Dimensional graphic illustration of the potencies of the growth hormone secretagogue library. The y-axis expresses how many times the compound is more potent than 10 nM.

789 have comparatively little diversity [15]. Such small focused combinatorial libraries for optimisation of leadcompounds have been described in the literature earlier ( [15] and references therein), but the strength of such a strategy has never been manifested so clearly as we describe here. 4. Conclusion We have described a fast and reliable method to synthesize GH secretagogues derived from ipamorelin on solid phase. All prepared compounds were tested for their ability to release GH from rat pituitary cells. The compounds show in vitro potencies with EC50 values from 1 nMinactive. Most striking is the unpredictability in the SAR, demonstrated by the fact that some structural elements (i.e. building blocks) in one combination show high potency but totally lack potency in combinations which previously had seemed favourable. This fact emphasizes that even small structural modications may cause dramatic change in the ligand-receptor interaction, and it demonstrates the necessity for a very dened structural arrangement to get optimal potency. Illustrated by compounds with potencies at 1 nM and by a number of discrepancies in the SAR, we have clearly demonstrated the strength of focused combinatorial libraries in SAR optimisation. 5. Experimental protocols All compounds were prepared in analogy to the synthesis of 3-(1-aminoethyl)-N-methyl-N-((1R)-1-(Nmethyl-N-((1R)-1-(methylcarbamoyl)-2-(2-thienyl)ethyl)carbamoyl)-2-(2-naphthyl)ethyl)benzamide described below using the following buildingblocks: Boc-A1, 3-(1(tert-butyloxycarbonylamino)ethyl)benzoic acid (racemic mixture); Boc-A2, 3-(tert-butyloxycarbonylaminomethyl) benzoic acid; Boc-A3, (2-(2S)-(tert-butoxycarbonylamino)butoxy)acetic acid; Boc-A4, (2S)-2-(((carboxy) methoxy)methyl)pyrrolidin-1-carboxylic acid tert-butyl ester; Boc-A5, (2-tert-butoxycarbonylamino-2-methylpropoxy) acetic acid; Boc-A6, (2E)-5-(tert-butyloxycarbonylamino)-5-methylhex-2-enoic acid; Fmoc-B1, 2-[(9H-uoren-9-ylmethoxycarbonyl)-methyl-amino]-3naphthalen-2-yl-propionic acid; Fmoc-B2, (2R)-3(benzo[b]thiophen-2-yl)-2-(N-(9H-uoren-9-ylmethoxycarbonyl)-N-methylamino)propionic acid; Fmoc-B3, (2R)-2-(N-((9H-uoren-9-ylmethoxy)carbonyl)-N-methylamino)-3-benzyloxypropionic acid; Fmoc-B4, (2R)-3(biphenyl-4-yl)-2-(N-((9H-uoren-9-yl)methoxycarbonyl)N-methylamino)propionic acid; Fmoc-C1, (2R)-2-(N(9H-uoren-9-ylmethoxycarbonyl)-N-methylamino)-3(2-thienyl)propionic acid; Fmoc-C2, (2R)-2-(N-((9Huoren-9-yl)methoxycarbonyl)-N-methylamino)-3-(2uorophenyl)propionic acid; Fmoc-C3, (2R)-2-(N-(((9Huoren-9-yl)methoxy)carbonyl)-N-methylamino)-3-phenylpropionic acid; Fmoc-C4, (2R)-2-(N-((9H-uoren9-yl)methoxycarbonyl)-N-methylamino)-3-(4-methoxyphenyl)propionic acid. All building blocks have been purchased from Synthetech Inc. (Albany, Oregon 97321) or prepared as previously described [810]. The RP-HPLC analysis was performed using UV detection at 214, 254, 276 and 301 nm on a Vydac 218TP54 4.6 250 mm 5 C-18 silica column, which was eluted at 1 mL/min at 42 C. Two different elution conditions were used: System A: the column was equilibrated with 5% acetonitrile in a buffer consisting of 0.1 M ammonium sulphate, which was adjusted to pH 2.5 with 4 M sulphuric acid. After injection the sample was eluted by a gradient of 560% acetonitrile in the same buffer for 50 min. System B: the column was equilibrated with 5% acetonitrile/0.1% TFA/water and eluted by a gradient of 5% acetonitrile/0.1% TFA/water to 60% acetonitrile/ 0.1% TFA/water for 50 min. The LC-MS analyses were performed on a PE Sciex API 100 LC/MS system using a Waterst 3 150 mm 3.5 C-18 symmetry column and positive ionspray with a ow rate of 20 L/min. The column was eluted with a linear gradient of 590% acetonitrile, 850% water and 10% triuoroacetic acid (0.1%)/water in 15 min at a ow rate of 1 mL/min. 5.1. 3-(1-Aminoethyl)-N-methyl-N-((1R)-1-(N-methyl-N((1R)-1-(methylcarbamoyl)-2-(2-thienyl)ethyl)-carbamoyl)2-(2-naphthyl)ethyl)benzamide The N-Methyl-PAL-Resin (75 mg, 0.045 mmol, load: 0.60) was washed with 5% diisopropylethylamine in dichloromethane (2 2 mL), dichloromethane (3 2 mL) and dimethylformamide (3 2 mL) and then swelled in dimethylformamide (2 mL). Then 2-(N-((9H-uoren-9yl)methoxycarbonyl)-N-methylamino)-3-(2-thienyl)propionic acid (46 mg, 0.09 mmol) in dimethylformamide (1 mL), O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexauorophosphate (34 mg, 0.09 mmol) in dimethylformamide (1 mL), 1-hydroxy-7-azabenzotriazole (15 mg, 0.09 mmol) in dimethylformamide (1 mL) and diisopropylethylamine (31 L, 0.18 mmol) in dimethylformamide (1 mL) were added and the mixture

790 was shaken overnight. The resin was ltered and washed with dimethylformamide (3 2 mL), dichloromethane (3 2 mL) and dimethylformamide (2 mL). Then 20% piperidine in dimethylformamide (5 mL) was added and the mixture was shaken for 20 min, ltered and washed with dimethylformamide (3 2 mL), dichloromethane (3 2 mL) and dimethylformamide (2 mL). Then 2-(N-((9Huoren-9-yl)methoxycarbonyl)-N-methylamino)-3-(2naphthyl)propionic acid (41 mg, 0.09 mmol) in dimethylformamide (1 mL), O-(7-azabenzotriazol-1-yl)-1,1,3,3tetramethyluronium hexauorophosphate (34 mg, 0.09 mmol) in dimethylformamide (1 mL), 1-hydroxy-7azabenzotriazole (15 mg, 0.09 mmol) in dimethylformamide (1 mL) and diisopropylethylamine (31 L, 0.18 mmol) in dimethylformamide (1 mL) were added and the mixture was shaken overnight. The resin was ltered and washed with dimethylformamide (3 2 mL), dichloromethane (3 2 mL) and dimethylformamide (2 mL). Then 20% piperidine in dimethylformamide (5 mL) was added and the mixture was shaken for 20 min, ltered and washed with dimethylformamide (3 2 mL), dichloromethane (3 2 mL) and dimethylformamide (2 mL). Then 3-(1-(tert-butoxycarbonylamino)ethyl)benzoic acid (22 mg, 0.09 mmol) in dimethylformamide (1 mL), O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexauorophosphate (34 mg, 0.09 mmol) in dimethylformamide (1 mL), 1-hydroxy-7-azabenzotriazole (15 mg, 0.09 mmol) in dimethylformamide (1 mL) and diisopropylethylamine (31 L, 0.18 mmol) in dimethylformamide (1 mL) were added and the mixture was shaken overnight. The resin was ltered and washed with dimethylformamide (3 2 mL), dichloromethane (3 2 mL) and dimethylformamide (2 mL). The resin was cooled to 0 C and 50% triuoroacetic acid in dichloromethane (4 mL) was added and the mixture was shaken for 10 min at 0 C. The resin was ltered and washed with 50% triuoroacetic acid in dichloromethane (2 0.5 mL) and the combined ltrates were concentrated under a stream of nitrogen. The obtained product was dissolved in acetonitrile/water 1:20 (10 mL) and applied to a C-18 Sep-Pak Classic cartridge (0.25 g, purchased from Watersy), which had been prewashed with acetonitrile (10 mL) and water (10 mL). Then water/triuoroacetic acid 99.9:0.1 (5 mL), followed by water/acetonitrile/ triuoroacetic acid 89.9:20:0.1 (4 mL) was run through the Sep-Pak and the ltrate was discarded. Then the Sep-Pak was washed with water/acetonitrile/ triuoroacetic acid 64.9:35:0.1 (4 mL) and the ltrate was diluted with water (11 mL) and lyophilised to 4 mg of the title product. 5.2. In vitro charaterisation in rat pituitary cell assay The Sprague-Dawley male albino rats (250 25 g) were purchased from Mllegaard, Lille Skensved, Denmark. The rats were housed in group cages (four to eight animals/cage) and placed in rooms with 12 h light cycle. The room temperature varied from 1924 C and the humidity from 3060%. The cells were isolated from rat pituitaries and dispersed into single cells using trypsin and cultured for 3 d. The cells were then washed and stimulated for 15 min with different GH secretagogues. The supernatant was decanted and assayed for GH content in a rat GH SPA-assay. Acknowledgements We are thankful to Peter Andersen, Annette Heewagen, Lotte G. Srensen, Anne G. Christensen, Sanne Kold, Annette Nielsen and Nille Birkebk Larsen for technical assistance. References
[1] [2] [3] [4] [5] [6] Bowers C.Y., Chang J., Momany F.A., Folkers F., in: MacIntyre I. (Ed.), Molecular Endocrinology, Elsevier, Amsterdam, 1977, p. 287. Smith R., Van Der Ploeg L.H.T, Howard A.D., Feighner S.D., Cheng K, Hickey G.J. et al., Endocr. Rev. 5 (1997) 621645. Clemmesen B., Overgaard K., Riis B., Christiansen C., Osteoporosis Int. 3 (1993) 330336. Brixen K., Nielsen H.K., Mosekilde L., Flyvbjerg A., Bone Miner. (1990) 609618. Rudman D., Feller A.G., Nagraj H.S., Gergans G., Lalitha P., Goldberg A.F. et al., N. Engl. J. Med. 323 (1990) 19. Patchett A.A., Nargund R.P., Tata J.R., Chen M.H., Baraket K.J., Johnston D.B.R. et al., Proc. Natl. Acad. Sci. USA 92 (1995) 70017005. Raun K., Hansen B.S., Johansen N.L., Thgersen H., Madsen K., Ankersen M., Andersen P.H., Eur. J. Endocrinol. 139 (1998) 552561. Ankersen M., Johansen N.L., Madsen K., Hansen B.S., Raun K., Nielsen K.K. et al., J. Med. Chem. 41 (1998) 36993704. Hansen T.K., Ankersen M., Hansen B.S., Raun K., Nielsen K.K., Lau J. et al., J. Med. Chem. 41 (1998) 37053714. Peschke B., Ankersen M., Hansen B.S., Hansen T.K., Johansen N.L., Madsen K. et al., Eur. J. Med. Chem. 34 (1998) 361380. Albericio F., Kneib-Cordonier N., Biancalana S., Gera L., Masada R.I., Hudson D., Barany G., J. Org. Chem. 55 (1990) 37303743. PALaldehyde is commercially available from the BioSearch division of PerSeptive Biosystems(Framingham, MA). Heiman M.L., Nekola M.V., Murphy W.A., Lance V.A., Coy D.H., Endocrinology 116 (1985) 410415. Kubinyi H., Curr. Opin. Drug Discov. Develop. 1 (1998) 1627. Balkenhohl F., von dem Bussche-Hnnefeld C., Lansky A., Zechel C., Angew. Chem. Int. Ed. Engl. 35 (1996) 22882337.

[7]

[8] [9] [10] [11] [12] [13] [14] [15]

Eur. J. Med. Chem. 34 (1999) 791798 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

791

Original article

Heterocyclic congeners of PD 128,907 with a partially hydrogenated benzomorpholine moiety as potential dopamine D3-receptor ligands
Norbert Matzankea, Werner Lwea, Sylvie Perachonb, Pierre Sokoloffb, Jean-Charles Schwartzb, Holger Starka*
b a Freie Universitt Berlin, Institut fr Pharmazie, Knigin-Luise-Strasse 2+4, 14195 Berlin, Germany Unit de Neurobiologie et Pharmacologie Molculaire (U. 109), Centre Paul Broca de lINSERM, 2ter, rue dAlsia, 75014 Paris, France

(Received 19 November 1998; revised 14 April 1999; accepted 26 April 1999)

Abstract With a straightforward seven-step synthesis, racemic perhydro[1,4]benzoxazin-6-on was synthesized in overall good yields via regioselective epoxid ring-opening to the corresponding -aminoalcohol. The oxazine derivative was the key intermediate for the preparation of heteroaromatic analogues of the dopamine D3-receptor preferring agonist PD 128,907. The morpholine moiety of PD 128,907 was incorporated in diazole and diazine compounds obtained by different ring closure reactions. The target compounds obtained were structurally related to non-ergot heteroaromatic dopamine agonists which display preferential activity at the D3 receptor, e.g., quinpirole, quinerolane, or pramipexole. The ve membered aminothiazole, aminoselenazole, and pyrazole derivatives showed at least one order of magnitude higher binding at the human D3 receptor than that at the D2L receptor. Although the novel compounds displayed Ki values only in the micromolar concentration range, the most active ones showed full agonist activity in a functional assay on mitogenesis. 1999 ditions scientiques et mdicales Elsevier SAS dopamine / D3-receptor / D2-receptor / PD 128,907 / agonist / mitogenesis / pramipexole / quinpirole / quinerolane

1. Introduction Dopamine is among the most widely studied neurotransmitters of the mammalian central nervous system (CNS). Recent advances in the molecular biology of dopamine receptors have resulted in the classication into D1-like (D1 and D5) and D2-like (D2D4) receptor subtypes [1, 2]. Each of the ve characterized dopamine receptor subtypes belongs to the superfamiliy of G protein-coupled receptors and contains approximately 400 amino acids arranged in similar tertiary structures of seven putative transmembrane domains. Although the receptors were well characterized by molecular biology and localized in different brain areas, their specic functions are poorly understood at present. The D3 receptor has restricted expression in limbic brain areas, associated with cognitive functions and motivated behaviour [3, 4]. Due to the distinct neuroanatomical distribution of different D2-like subreceptors, as well as pharma*Correspondence and reprints

cological and behavioural studies, it is suggested that D3 receptors seem to be important targets for the development of drugs for the treatment of Parkinsons disease [5], schizophrenia [6], and drug abuse, e.g., ligands with agonist activity at D3 receptors reduced the selfadministered cocaine intake in rats [7]. The development of more structurally diverse and selective D3-receptor agonists is required to provide a better understanding of pharmacology and pathophysiology of dopamine receptor subtypes. Dopamine itself had 20 times higher affinity at the D3 receptor than that at the D2 receptor [3]. Different non-ergot agonists were described possessing varying degrees of preference for the D3 receptor. Tetralin derivatives like 7-OH-DPAT (7-hydroxy-N,N-dipropyl-2aminotetralin) [4] or PD 128,907 ((+)-4aR,10bR3,4,4a,10b-tetrahydro-4-propyl-2H,5H-[1]benzopyrano [4,3-b]1,4-oxazin-9-ol) [8, 9] showed higher affinity at the D3 receptor than that at other D2-like receptors. The same is true for some heterocyclic analogues like pramipexole, quinerolane and quinpirole [10, 11] containing an aminothiazole, an aminopyrimidine and a pyrazole ring,

792 cyclic moiety of other D3-receptor preferring agonists or related structures, e.g., pramipexole, quinerolane, and quinpirole. 2. Chemistry
Figure 1. Structure of PD 128,907.

respectively. The reference agonist showing high activity in combination, with one of the highest degrees of D3 receptor preferring selectivity, is PD 128,907 (gure 1). This compound contains a morpholine moiety as one structural element which may be responsible for the increased selectivity compared to the structurally related 7-OH-DPAT. The afore mentioned heterocyclic agonists also displayed a preferential higher affinity at the D3 receptor compared to that at other D2-like receptors. The aim of this study was the synthesis and pharmacological investigation of compounds which have unied the structural morpholine element of PD 128,907 and the hetero-

The racemic N-substituted perhydro[1,4]benzoxazin-6on (()-7) as key intermediate was synthesized by the straightforward seven-step route outlined in gure 2. The synthesis of trans-congurated ()-3 relied on a modication of chemistry originally described by Cheng and co-workers [12] for the preparation of ligands for the j opioid receptor. Cheng et al. obtained the epoxid 2 via a four-step route starting from 1,4-cyclohexanedione monoethylene ketal, but this can be done in a simple two-step procedure previously described [13, 14]. Thus, the enol ether of 1-methoxy-1,4-cyclohexadiene [15] was cleaved by a catalytic amount of p-toluyl sulfonic acid (p-TsOH) and the resulting carbonyl group was subsequently protected as ethylene ketal. Epoxidation of 1 with m-chloroperoxybenzoic acid (m-CPBA) provided 2 in

Figure 2. Synthesis of precursor ()-7. (a) ethylene glycol, p-TsOH, toluene, T; (b) m-chloroperoxybenzoic acid, CH2Cl2; (c) propylamine, water, ambient temperature; chloroacetyl chloride, Et3N, CH2Cl2; (e) n-tetrabutylammonium hydrogensulphate (cat.), NaOH; (f) LiAlH4, THF, T; (g) 2 N HCl, T.

793

Figure 3. Synthesis of heterocyclic dopamine receptor ligands. (a) i: Br2, HBr, 0 C, ii: H2N-CX-NH2, T; (b) HC(NMe2)3, DMF, 65 C; (c) hydrazine, MeOH, ambient temperature; (d) guanidine x HCl, K2CO3, EtOH, T.

good overall yield. Epoxid ring-opening of 2 with propyl amine resulted exclusively in the formation of the desired trans-congurated -aminoalcohol ()-3. The isomeric aminoalcohol could not be isolated, as depicted in gure 2. The selectivity and rate of conversion of this reaction is remarkable, since such nucleophilic epoxide ring openings with amines normally require heterogeneous catalysis [16] or amine activation with aluminiaorganic reagents [17]. This observation of selectivity is congruent to the experiments of Chen et al. [12], who isolated, in the reaction of this epoxid with pyrrolidine, the two regioisomeric aminoalcohols in a ratio of 13:1 (89% yield). Reaction of 3 with chloroacetyl chloride [18] led to the amide 4, which is cyclized under phase-transfer conditions to the 1,4-benzoxazine 5. In the 1H NMR spectra of 4 both rotameres are present in equal amounts. This amide was reduced with LiAlH4 in THF to give the amine 6. Initial attemps to deprotect 6 with PPTS/wet acetone [19] failed, but this could be achieved by 2 N HCl [20] and afforded the desired racemic aminoketone 7. The heterocyclic substituted 1,4-benzoxazines were synthesized as shown in gure 3. The thiazolo- and selenazolo substituted compounds 8 and 9 were prepared in a well known manner [21] by treatment of 7 with bromine in acidic medium, and subsequent reaction of the

-haloketone intermediate either with thiourea or selenourea. Bromination of 7 could have taken place on either side of the carbonyl group to produce either the desired linear compounds 8 and 9 or the undesired angular products. However, bromination of position 5 is sterically unfavoured and the evidence of substitution could be done by an 1H NMR off resonance decoupling experiment of compound 8: decoupling of the 9a-proton resulted in a double doublet pattern of the protons in the 9-position, which would be impossible if the molecule was angularly annelated. Analogous observations were described by Kornfeld et al. [22] in the course of the preparation of rigid 3-(2-aminoethyl)pyrroles as dopamine agonists. The precursor 10 as a masked 1,3-dicarbonyl compound was synthesized by the treatment of the free base of 7 with tris(dimethylamino)methane [23] in good yield. Treatment of this enaminone 10 either with guanidine hydrochloride/K2CO3 [24] or hydrazine [20, 22] resulted in the formation of the 2-aminopyrimidine derivative 12 and the pyrrole derivative 11 in its corresponding tautomeric forms, respectively. All compounds which were selected for pharmacological testing were converted with anhydrous HCl in their water soluble salts (see Experimental section). Due to the moderate pharmacological affinity the compounds were not separated into their enantiomers.

794
Table I. Binding and functional studies at Dopamine D2 and D3 receptors. Inhibition of binding at Compound ()-3 ()-6 ()-8 ()-9 ()-11 ()-12 Quinpirolec,d Quinerolanec,d Pramipexolec PD 128,907c D2 receptor Ki [M] 600 > 600 99 2 49 3 99 5 > 600 1.4 0.34 0.79 0.389 D3 receptor Ki [M] 68 3 172 28 6.2 0.5 2.9 0.4 2.3 0.3 46 5 0.039 0.004 0.004 0.002 Ki (D2)/ Ki (D3) 13 > 3.4 16 17 43 > 12 36 95 193 216 Stimulation of D3 receptor EC50 [M] ndb nd 7.1 3.0 1.3 0.08 0.54 0.09 nd 0.00086 0.00015 0.00023 0.00064 i.a.a nd nd 106 101 100 nd 100 96 98 96

(a) i.a. = intrinsic activity (i.a. (quinpirole) = 100); (b) nd = not determined; (c) ref. [32]; (d) ref. [33].

3. Pharmacology Displacement assays were performed with chinese hamster ovary (CHO) cell lines stably transfected with human D2L or D3 receptor cDNA using [125I]iodosulpiride at a concentration of 0.1 nM [25]. The nonspecic binding was determined in the presence of enomapride. Binding data were analysed by computerized nonlinear regression for a one-site model. The Ki values were derived from IC50 values according to the ChengPrussoff equation [26]. Additional functional experiments were performed to clarify the mode of action for the most active compounds of this series. NG 108-15 cells expressing the human dopamine D3 receptor were incubated with forskoline and the drug in increasing concentrations [25]. Then, [3H]thymidine was added, and increase in mitogenesis, as a measurement of agonist activity, could be measured by counting the radioactivity incorporated into the cells after 2 h. A decrease in, or inhibition of, mitogenesis in relation to the effect of the reference compound quinpirole would show a partial agonist or antagonist effect, respectively. 4. Pharmacological results and discussion The target compounds ()-812 displayed at least 10 times higher binding affinities at the D3 receptor than that at the D2 receptor (table I). Their affinity constants at the D3 receptor were found to be in the micromolar concentration range. Although the compounds ()-811 were found to possess higher D3-receptor affinity than the intermediate products of their chemical synthesis, the difference between these heteroaromatic products and their chemical educts were not important due to their

overall moderate affinity constants. The aminopyrimidine derivative ()-12 showed equipotency to the secondary amine intermediate ()-3 and even lower D3-receptor affinity than the non-aromatic morpholine precursor molecule ()-6. Compounds possessing a ve-membered heteroaromatic moiety (()-811) were found in this series to be active in a lower concentration range than the other tested compounds. Interestingly, the seleno analogue ()-9 of the pramipexole related derivative ()-8 showed about two times higher affinity than that of compound ()-8. For dopamine agonists used in therapy for Parkinsons disease, an anti-oxidative or radical scavenger effect seemed to be benecial. These properties may be enhanced in seleno heterocycles compared to sulfur heterocycles. Therefore, further investigations to study the anti-oxidative effects of aminothiazole and aminoselenazole derivatives are currently in progress. Compared to PD 128,907, pramipexole, quinerolane, and quinpirole, the novel compounds possess the basic amino functionality on another side of the molecule (reversed orientation), i.e., the positions of the oxygen and the nitrogen in the morpholine moiety have to be changed to be able to superimpose with the reference agonists. Concerning the pharmacological activity, this seemed to be the major drawback for the compounds prepared. The position of the basic amino functionality in relation to the aromatic hydrogen-bond area were described in different molecular modelling and structureactivity relationship studies to be essential for high affinity D2-like receptor binding [27]. These ndings could be emphasized by the results of the present study for binding at D3-receptors. Compounds with an opposite orientation of oxygen and nitrogen atom in the morpholine moiety possessing a stronger relationship to PD

795 128,907 compared to the ones described in this paper have already been reported as potent dopamine receptor agonists [28, 29]. Nevertheless, when the three most active compounds ()-811 were tested for their mode of action in the mitogenesis test it was found that all compounds were able to induce a full mitogenesis rate indicating a full agonist behaviour. Although the compounds displayed only moderate affinity they were still able to induce a change in the receptor structure that is required for signal transduction indicated by complete receptor activation. 5. Experimental protocols Melting points were uncorrected. H NMR spectra were determined with a Bruker AC 300 (300 MHz) or a Bruker Avance DPX 400 (400 MHz) instrument as indicated by spectrometer frequency. Chemical shifts were reported in values (ppm) relative to an internal standard of tetramethylsilane. Abbreviations used in NMR analysis were as follows: s, singlet; d, doublet; t, triplet; m, multiplet; br s, broad singlet; and dd, doublet of doublets. IR spectra were determined with a PerkinElmer 297 spectrophotometer and Mass spectra were obtained on a CH-7A-Varian MAT (70 eV) instrument. Microanalyses were performed with a Perkin-Elmer 240 B and C analyzer. Analyses indicated by the symbols of the elements or functions were within 0.4% of theoretical values. Thin layer chromatography was performed on Merck precoated TLC plates with silica gel 60-F254 and visualized with UV light (254 nm), after treatment with iodine or after treatment with Fast Blue B salt. Column (ash) chromatography [30] was performed with Merck silica gel 60 (230400 mesh). Solvents and reagents were used as purchased, except as noted. Ether, THF and toluene were distilled from sodium metal/ benzophenone ketyl. [125I]Iodosulpiride (2 000 Ci/mmol) and [3H]thymidine (110 Ci/mmol) were obtained from Amersham International, Ltd. (Buckinghamshire, UK) and emonapride (YM 09152) from the Yamanouchi Pharmaceutical Co. (Tokyo, Japan). 5.1. Chemistry 5.1.1. 1,4-Dioxaspiro[4,5]dec-7-ene (1) Compound 1 was prepared according to the procedure of Lambert et al. [13] from 1-Methoxy-1,4-cyclohexadiene [15] in 80% yield: b.p. 7071 C (13 mbar) (lit. [13] b.p. 6264 C, 7 Torr); IR (CHCl3 solution) 3 027, 2 934, 2 879, 1 687 cm1; 1H NMR (CDCl3, 300 MHz) 1.741.81 (m, 2H), 2.272.28 (m, 4H), 3.99
1

(s, 4H), 5.605.74 (m, 2H); MS m/e 140 (M, 35), 125 (8), 112 (8), 99 (62), 86 (100). 5.1.2. ()-Spiro[1,3-dioxolane-2,3-[7]oxabicyclo[4,1,0] heptane] (2) Compound 2 was prepared by the procedure of Cheng et al. [12] from 1 in 67% yield: b.p. 99 C (9 mbar) (lit. [14] b.p. 100 C, 16 Torr). IR (CHCl3 solution) 2 942, 2 882 cm1; 1H NMR (CDCl3, 300 MHz) 1.411.70 (m, 2H), 2.002.20 (m, 4H), 3.153.18 (m, 2H), 3.873.98 (m, 4H); MS m/e 156 (M, 2), 140 (22), 112 (12), 99 (85), 86 (100). 5.1.3. ()-trans-7-Propylamino-1,4-dioxaspiro[4,5]decan8-ol (()-3) To a stirred mixture of 2 (6.25 g, 40.0 mmol) and propylamine (6.58 mL, 80.0 mmol) was added 4 mL H2O dropwise at 0 C. The reaction mixture was stirred at ambient temperature under N2. After 40 h the mixture was evaporated in vacuo to remove the water and the excess of propylamine to give a red oil. This oil was puried by ash chromatography (10% MeOH in CH2Cl2 with 1% concentrated NH4OH) on silica gel to give ()-3 (5.148 g, 60%) as a light red oil which slowly crystallized on standing: m.p. 6566 C; IR (KBr) 3 390, 3 300, 3 166, 2 956, 2 876, 2 824, 1 635 cm1; 1H NMR (CDCl3, 400 MHz) 1.00 (m, 3H), 1.392.55 (m, 10H), 2.79 (m, 1H), 3.29 (m, 1H), 4.02 (m, 4H) (no OH/NH signals were detected); MS m/e 215 (M, 5), 186 (37), 156 (64), 129 (6), 101 (10), 86 (100). Anal. C11H21NO3 (C, H, N). 5.1.4. ()-trans-2-Chloro-N-(8-hydroxy-1,4-dioxaspiro [4,5]decan-7-yl)-N-propyl-acetamid (()-4) To a stirred solution of ()-3 (5.08 g, 23.6 mmol) and triethylamine (3.29 mL, 23.6 mmol) in 100 mL CH2Cl2 was added chloroacetyl chloride (1.88 mL, 23.6 mmol) dropwise over a period of 10 min at 0 C. After 30 min the mixture was allowed to warm to ambient temperature and stirred for a further 2 h. The organic layer was separated, washed with 20 mL 2 N HCl, 20 mL H2O, 20 mL brine, dried (MgSO4), and concentrated in vacuo whereupon red crystals formed. Recrystallization from CH2Cl2/n-hexane gave pure ()-4 (4.50 g, 65%) as colourless crystals: m.p. 172 C; IR (KBr) 3 408, 2 961, 2 880, 1 647 cm1; 1H NMR (CDCl3, 400 MHz) 0.900.94 (m, 6H), 1.492.14 (m, 21H), 2.22 (s, 1H, collapse after D2O-exchange), 2.56 (1H, collapse after D2O-exchange), 2.94 (m, 1H), 3.233.34 (m, 2H), 3.653.76 (m, 2H), 3.974.13 (m, 9H), 4.334.35 (m, 1H); MS m/e 291 (M, 1), 256 (14), 232 (23), 156 (49), 86 (100). Anal. C13H22ClNO4 (C, H, N).

796 5.1.5. ()-trans-4-Propyl-2,3,4,4a,5,7,8,8a-octahydrospiro[6H-1,4-benzoxazine-6,2-[1,3]dioxolane]-3-on (()-5) Compound ()-4 (4.50 g, 15.4 mmol) and tetra-nbutylammonium hydrogen sulphate (525 mg, 1.54 mmol) were dissolved in 300 mL CH2Cl2 under an atmosphere of nitrogen. 61.7 mL of 0.5 N NaOH was added via syringe, and the biphasic system was stirred vigerously at ambient temperature for 20 h. The organic layer was separated, washed with water (2 40 mL), brine (40 mL), dried (MgSO4), and concentrated in vacuo to give a yellow oil. This oil was puried by ash chromatography (10% MeOH in CH2Cl2) on silica gel to give ()-5 (3.804 g, 97%) as a light yellow oil which slowly crystallized on standing: m.p. 67 C; IR (KBr) 3 448, 2 964, 2 880, 1 655 cm1; 1H NMR (CDCl3, 400 MHz) 0.91 (m, 3H), 1.432.13 (m, 8H), 3.043.74 (m, 4H), 3.98 (s, 4H), 4.26 (dd, J = 16.0/21.2 Hz, 2H); MS m/e 255 (M, 92), 226 (31), 198 (53), 155 (68), 141 (47), 99 (100). Anal. C13H21NO4 (C, H, N). 5.1.6. ()-trans-4-Propyl-2,3,4,4a,5,7,8,8a-octahydrospiro[6H-1,4-benzoxazine-6,2-[1,3]dioxolane] hydrochloride (()-6) LiAlH4 (1.61 g, 42.3 mmol) was suspended in 20 mL THF under an atmosphere of nitrogen. The solution of ()-5 (3.60 g, 14.1 mmol) in 30 mL THF was added dropwise via syringe at 0 C over a period of 20 min. The reaction mixture was allowed to warm to ambient temperature and then heated at reux for 2 h. After cooling, excess hydride was destroyed by the careful addition of 10 mL water. The mixture was ltered and the solid was washed with ether (5 20 mL). The ltrate was washed with brine (2 30 mL), dried (K2CO3) and concentrated in vacuo. The resulting oil was puried by ash chromatography (10% MeOH in CH2Cl2 with 1% concentrated NH4OH) on silica gel to give the free base of ()-6 as a light yellow oil. This oil was dissolved in dry ether and a stream of dry gasous HCl was bubbled through the solution at 0 C to produce the hydrochloride ()-6 (3.793 g, 97%) as colourless crystals: m.p. 180181 C; IR (KBr) 3 431, 2 965, 2 881, 2 585, 2 475, 2 325 cm1; 1 H NMR (DMSO-d6, 400 MHz) 0.90 (m, 3H), 1.441.86 (m, 7H), 2.29 (m, 1H), 2.853.38 (m, 5H), 3.73 (m, 1H), 3.92 (m, 6H), 11.41 (br s, 1H); MS m/e 241 (M, 12), 212 (100), 139 (29). Anal. C13H23NO3HCl (C, H, N). 5.1.7. ()-trans-4-Propyl-3,4,4a,5,6,7,8,8a-octahydro2H-1,4-benzoxazin-6-on hydrochloride (()-7) Compound ()-6 (3.793 g, 13.7 mmol) was dissolved in 50 mL 2 N HCl and heated to reux for 2 h, cooled, neutralized with 10% NaOH, saturated with NaCl and extracted with CH2Cl2 (5 20 mL). The organic extracts were dried over K2CO3 and concentrated in vacuo to provide a yellow oil. This oil was puried by ash chromatography (10% MeOH in CH2Cl2 with 1% concentrated NH4OH) on silica gel to give the free base of ()-7 as a light yellow oil, which was converted in the hydrochloric salt as described before to give ()-7 (2.954 g, 93%) as colourless crystals: m.p. 197 C; IR (KBr) 3 419, 2 969, 2 880, 2 581, 2 479, 2 419, 1 719 cm1; 1H NMR (DMSO-d6, 400 MHz) 0.91 (t, J = 7.3 Hz, 3H), 1.631.69 (m, 3H), 2.063.47 (m, 10H), 4.024.03 (m, 2H), 4.134.14 (m, 1H), 11.83 (br s, 1H); MS m/e 197 (M, 34), 168 (100), 140 (42). Anal. C11H19NO2HCl (C, H, N). 5.1.8. ()-trans-8-Propyl-4a,6,7,8,8a,9-hexahydro-4Hthiazolo[4,5-g][1,4]benzoxazin-2-amine dihydrochloride (()-8) To a stirred solution of ()-7 (233.7 mg, 1 mmol) in 10 mL 47% HBr was added Br2 (1.76 mL of a 10% solution in 47% HBr, 1.1 mmol) dropwise at 0 C. After 2 h, thiourea (83.7 mg, 1.1 mmol) was added as a solid, and the mixture was heated to reux for further 2 h. After evaporation of the HBr in vacuo, the residue was treated with 10 mL 10% NaOH with cooling and extracted with CH2Cl2 (5 10 mL). The combined organic layers were dried (MgSO4) and concentrated in vacuo to provide a crude base of ()-8 as a brown solid. This residue was puried by ash chromatography (10% MeOH in CH2Cl2 with 1% concentrated NH4OH) on silica gel to give the free base of ()-8 as a foam. This foam was dissolved in dry ethanol (15 mL), treated with saturated HCl in ethanol (2 mL) and evaporated in vacuo again. Recrystallization from dry 2-propanol/ether provided the title compound (185.7 mg, 57%) as a grey solid: m.p. 240241 C; IR (KBr) 3 375, 3 247, 3 064, 2 936, 2 720, 2 621, 2 510, 1 627, 1 577 cm1; 1H NMR (DMSO-d6, 400 MHz) 0.97 (t, J = 7.1 Hz, 3H), 1.731.87 (m, 2H), 2.513.47 (m, 9H), 4.304.44 (m, 3H), 9.30 (br s, 3H), 12.67 (br s, 1H); MS m/e 253 (M, 28), 224 (29), 151 (8), 127 (100). Anal. C12H19N3OS2 HClH2O (C, H, N). 5.1.9. ()-trans-8-Propyl-4a,6,7,8,8a,9-hexahydro-4Hselenazolo[4,5-g][1,4]benzoxazin-2-amine dihydrochloride (()-9) This compound was prepared and puried in the same manner as described in the synthesis of ()-8 in a 1.5 mmol scale to provide ()-9 (137.8 mg, 25%) as a light red solid: m.p. 220221 C; IR (KBr) 3 397, 3 260, 3 073, 2 933, 2 722, 2 624, 1 623, 1 579 cm1; 1H NMR (DMSO-d6, 400 MHz) 0.94 (t, J = 7.3 Hz, 3H),

797 1.731.77 (m, 2H), 2.893.52 (m, 9H), 4.034.19 (m, 3H), 9.58 (br s, 3H), 12.12 (br s, 1H); MS m/e 301 (M, Se main isotope, 12), 272 (13), 127 (100). Anal. C12H19N3OSe2 HClH2O (C, H, N). 5.1.10. ()-trans-(E)-7-Dimethylaminomethylidene-4propyl-2,3,4,4a,5,6,8,8a-octahydro-1,4-benzoxazin-6-on (()-10) To a solution of ()-7 (986.4 mg, 5 mmol, free base) in 5 mL of dry DMF was added tris(dimethylamino)methane (2.33 mL, 15 mmol). The solution was stirred for 16 h at 65 C under a nitrogen atmosphere and then concentrated in vacuo to provide a brown oil. This oil was puried by ash chromatography (10% MeOH in CH2Cl2 with 1% concentrated NH4OH) on silica gel to give ()-10 (1.066 g, 85%) as a yellow oil which very slowly crystallized on standing. A portion of this material was recrystallized from cyclohexane/n-hexane to give ()-10 as yellow crystals: m.p. 65 C; IR (KBr) 3 451, 2 959, 2 859, 2 803, 1 648, 1 559 cm1; 1H NMR (DMSO-d6, 400 MHz) 0.83 (t, J = 7.2 Hz, 3H), 1.291.46 (m, 2H), 1.89 (dd, J = 5.8/11.6 Hz, 1H), 2.032.62 (m, 6H), 2.73 (d, J = 11.4 Hz, 1H), 2.96 (dd, J = 4.2/6.7 Hz, 1H), 3.05 (s, 6H), 3.233.27 (m, 1H), 3.56 (t, J = 11.1 Hz, 1H), 3.77 (d, J = 11.1 Hz, 1H), 7.31 (s, 1H); MS m/e 252 (M, 29), 235 (7), 193 (25), 84 (84), 58 (100). Anal. C14H24N2O2 (C, H, N). 5.1.11. ()-trans-8-Propyl-1,4,4a,6,7,8,8a,9-octahydropyrazolo[3,4-g][1,4]benzoxazine dihydrochloride (()-11) (cf. [28, 29]) To a solution of ()-10 (309.5 mg, 1.23 mmol) in 10 mL dry methanol was added hydrazine (0.39 mL, 12.3 mmol) and the mixture was stirred for 19 h. The solvent was evaporated and the crude product was puried by ash chromatography (10% MeOH in CH2Cl2 with 1% concentrated NH4OH) on silica gel to give the base of ()-11 (206.8 mg, 76%) as a yellow oil. The HCl salt was formed in ethanol and crystallized from dry 2-propanol/ether to provide ()-11 as a hygroscopic foam. IR (KBr) 3 395, 2 932, 2 877, 2 688, 2 615, 2 487, 2 360, 1 676, 1 645, 1 570 cm1; 1H NMR (DMSO-d6, 400 MHz) 0.96 (m, 3H), 1.75 (m, 2H), 2.513.48 (m, 9H), 4.034.11 (m, 3H), 7.59 (s, 1H), 7.75 (br s, 2H), 11.97 (br s, 1H); MS m/e 221 (M, 48), 192 (100), 127 (39), 98 (22). Anal. C12H19N3O2 HCl (C, H, N). 5.1.12. ()-trans-9-Propyl-5a,7,8,9,9a,10-hexahydro-5Hpyrimidino[4,5-g][1,4]benzoxazin-2-amine dihydrochloride (()-12) To a solution of ()-10 (293.1 mg, 1.16 mmol) in 15 mL dry ethanol was added dry guanidineHCl (1.11 g, 11.6 mmol) and K2CO3 (1.61 g, 11.6 mmol). The suspension was reuxed for 2 h and then evaporated to dryness in vacuo. The residue was treated with water (30 mL) and extracted with CH2Cl2 (5 10 mL). The combined organic layers were dried (K2CO3) and concentrated in vacuo. The solid residue was converted into its hydrochloric salt and recrystallized from dry ethanol/ether to provide ()-12 (252.8 mg, 68%) as a yellow solid: m.p. 196 C. IR (KBr) 3 229, 3 137, 2 969, 2 935, 2 880, 2 689, 2 542, 2 484, 2 349, 1 664 cm1; 1H NMR (DMSO-d6, 400 MHz) 0.96 (t, J = 7.3 Hz, 3H), 1.721.77 (m, 2H), 2.543.53 (m, 9H), 4.054.16 (m, 3H), 7.99 (br s, 3H), 8.33 (s, 1H), 12.24 (br s, 1H); MS m/e 248 (M, 21), 219 (100). Anal. C13H20N4O2 HCl (C, H, N). 5.2. Pharmacological testing 5.2.1. Displacement studies [25] Transfected cell lines consisting of CHO cell lines stably transfected with human dopamine D2L or D3 receptor DNA [11] were cultured in Dulbeccos Modied Eagle Medium supplemented with 10% foetal calf serum in a 5% CO2 humidied atmosphere. They were harvested from culture dishes in the presence of 0.2% trypsin, centrifuged at 2 000 g for 5 min and homogenized in 10 mM Tris-HCl, pH 7.4 containing 5 mM MgCl2 using a Polytron. The homogenate was centrifuged at 20 000 g for 15 min at 4 C, and the pellet was resuspended by sonication in 50 mM Tris-HCl, pH 7.4, containing 120 mM: NaCl, 5 mM KCl, 2 mM CaCl2, and 8 mM MgCl2 (incubation buffer). Membranes were either used immediately or after storage at 70 C. Binding assays were started by the addition of 200 L of membranes (120 g of protein) from transfected cells diluted in incubation buffer 1 supplemented with 0.2% bovine serum albumin to polystyrene tubes containing, in 100 L, 0.1 nM [125I]iodosulpiride and drug diluted in 100 L of incubation buffer. Nonspecic binding (120% of total binding) was determined in the presence of 1 M enomapride. Incubations were run at 30 C for 30 min. All reactions were stopped by vacuum ltration through Whatman GF/B glass-bre lters coated in 0.3% polyethylenimine with an automated cell harvester (BrandelBeckmann, Gauthersburg, MD) and were rinsed 3 times with 5 mL of ice-cold incubation buffer. Filters were counted by gamma spectrometry in 5 mL of ACS II (Amersham). EC50 values ( SE) and maximal responses were calculated from concentration-response curves. Intrinsic activity was calculated relative to quinpirole, a full agonist [31].

798 5.2.2. Mitogenesis [25] NG 108-15 cells expressing the human dopamine D3-receptor [31] were cultured in Dulbeccos Modied Eagle Medium supplemented with 10% foetal calf serum in a 5% CO2 humidied atmosphere and plated in collagen-coated 96-well plates. After a 24 h culture, cells were washed twice with culture medium without foetal calf serum and incubated for 16 h with 1 M forskolin and the drug in increasing concentrations or quinpirole as control. Then, [3H]thymidine (1 Ci/well) was added for 2 h and cells were harvested by vacuum ltration through Whatman GF/C glass-bre lters by using an automated cell harvester and were rinsed 15 times with 200 L of phosphate-buffered saline. Radioactivity was counted by liquid scintigraphy in 5 mL of ACS (Amersham). Ki values were derived from IC50 values according to the Cheng-Prussoff equation [26], taking into account the Kd of [125I]iodosulpiride for respective receptors. Data were means of Ki values from data obtained in at least three separate experiments, and the statistical error was expressed as SEM.
[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] Caine S.B., Koob G.F., Parsons L.H., Everitt B.J., Schwartz J.-C., Sokoloff, P., Neuroreport 8 (1997) 23732377. Pugsley A., Davis M.D., Akunne H.C., Mackenzie R.G., Shih Y.H., Damsma G. et al., J. Pharmacol. Exp. Ther. 275 (1995) 13551366. Bancroft G.N., Morgan K.A., Flietstra R.J., Levant B., Neuropsychopharmacology 18 (1998) 305316. Mierau J., Schneider F.J., Ensinger H.A., Chio C.L., Lajiness M.E., Huff R.M., Eur. J. Pharmacol. 290 (1995) 2936. Sokoloff P., Andrieux M., Besanon R., Pilon C., Martres M.P., Giros B., Schwartz J.-C., Eur. J. Pharmacol. 225 (1992) 331337. Cheng C.Y., Wu S.C., Hsin L.W., Tam S.W., J. Med. Chem. 35 (1992) 22432247. Lambert J.B., Xue L., Bosch R.J., Taba K.M., Marko D.E., Urano S., Lebreton P.R., J. Am. Chem. Soc. 108 (1986) 75757579. Van der Louw J., Van der Baan J.L. Out G.J.J., De Kanter F.J.J., Bickelhaupt F., Klumpp G.W., Tetrahedron (1992) 99019905. Wilds A.L., Nelson N.A., J. Am. Chem. Soc. 75 (1953) 53605365. Posner G.H., Rogers D.Z., J. Am. Chem. Soc. 99 (1977) 82088214. Gorzynski Smith J., Synthesis (1984) 629656. Dijkstra D., Mulder T.B., Rollema H., Tepper P.G., Van der Weide J., Horn A.S., J. Med. Chem. 31 (1988) 21782182. Sterzycki R., Synthesis (1979) 724725. McQuaid L.A., Latz J.E., Clemens J.A., Fuller R.W., Wong D.T., Mason N.R., J. Med. Chem. 32 (1989) 23882396. Berney D., Schuh K., Helv. Chim. Acta 65 (1982) 13041309. Bach N.J., Kornfeld E.C., Jones N.D., Chaney M.O., Dorman D.E., Paschal J.W., Clemens J.A., Smalstig E.B., J. Med. Chem. 23 (1980) 481491. White J.D., Badger R.A., Kezar III H.S., Pallenberg A.J., Schiehser G.A., Tetrahedron 45 (1989) 66316644. De D., Mague J.T., Byers L.D., Krogstad D.J., Tetrahedron Lett. 36 (1995) 205208. Sautel F., Griffon N., Sokoloff P., Schwartz J.C., Launay C., Simon P. et al., J. Pharmacol. Exp. Ther. 275 (1995) 12391246. Cheng Y.C., Prussoff W.H., Biochem. Pharmacol. 22 (1973) 30993108. Casagrande C., Bertolini G., in: Giardin D., Piergentili A., Pigini M. (Eds.) Ser. Pharmacochemistry Library, Vol. 24, Elsevier, Amsterdam, 1996, pp. 6784. Booher R.N., Kornfeld E.C., Smalstig E.B., Clemens J.A., J. Med. Chem. 30 (1987) 580583. Booher R.N., Cornfeld E.C., (Lilly, Eli and Co.) U. S. US 4,552,956. (1986); Chem. Abstr. 104 (1986) P186431k. Still W.C., Kahn M., Mitra A., J. Org. Chem. 43 (1978) 29232925. Pilon C., Lvesque D., Dimitriadou V., Griffon N., Martres M.P., Schwartz J.-C., Sokoloff P., Eur. J. Pharmacol. 268 (1994) 129139. Sautel F., Griffon N., Lvesque D., Pilon C., Schwartz J.-C., Sokoloff P., Neuroreport 6 (1995) 329332. Sokoloff P., Andrieux M., Besancon R., Pilon C., Martres M.P., Giros B., Schwartz J.-C., Eur. J. Pharmacol. 225 (1992) 331337.

Acknowledgements
[23]

This work was supported by the Biomedical & Health Research programme (BIOMED) of the European Union and by the National Institute on Drug Abuse (NIDA) (DAII534-01), USA.

[24] [25] [26]

References
[1] [2] [3] [4] Schwartz J.-C., Giros B., Martres M.P., Sokoloff P., Semin. Neurosci. 4 (1992) 99108. Gingrich J.A., Caron M.G., Annu. Rev. Neurosci. 16 (1993) 731745. Sokoloff P., Giros B., Martres M.P., Bouthenet M.L., Schwartz J.-C., Nature (London) 347 (1990) 146151. Lvesque D., Diaz J., Pilon C., Martres M.P., Giros B., Souil E., Schott D., Morgat J.L., Schwartz J.-C., Sokoloff P., Proc. Natl. Acad. Sci. USA 89 (1992) 81558159. Hagan J.J., Midlemiss D.N., Sharpe P.C., Poste G.H., Trends Pharmacol. Sci. 18 (1997) 156163. Shafer R.A., Levant B., Psychopharmacology 135 (1998) 116.

[27]

[28] [29] [30] [31] [32] [33]

[5] [6]

Eur. J. Med. Chem. 34 (1999) 799808 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

799

Original article

Carbonic anhydrase inhibitors part 70#. Synthesis and ocular pharmacology of a new class of water-soluble, topically effective intraocular pressure lowering agents derived from nicotinic acid and aromatic/heterocyclic sulfonamides
Claudiu T. Supurana*, Andrea Scozzafavaa, Luca Menabuonib, Francesco Mincionec, Fabrizio Brigantia, Giovanna Mincionea
a

Universit degli Studi, Laboratorio di Chimica Inorganica e Bioinorganica, Via Gino Capponi 7, I-50121, Firenze, Italy b Ospedale San Giovanni di Dio, S.O. Oculistica, Via Torregalli 3, I-50123, Firenze, Italy c Universit degli Studi, Institute of Ophthalmology, Viale Morgagni 85, I-50134, Firenze, Italy (Received 14 January 1999; revised 30 April 1999; accepted 6 May 1999)

Abstract Reaction of twenty aromatic/heterocyclic sulfonamides containing a free amino, imino, hydrazino or hydroxyl group, with nicotinoyl chloride afforded a series of water-soluble (as hydrochloride or triate salts) compounds. The new derivatives were assayed as inhibitors of three carbonic anhydrase (CA) isozymes, hCA I, hCA II (cytosolic forms) and bCA IV (membrane-bound form); h = human, b = bovine isozyme. Efficient inhibition was observed against all three isozymes, but especially against hCA II and bCA IV (in nanomolar range), two isozymes known to play a critical role in aqueous humour secretion within the ciliary processes of the eye. Some of the best inhibitors synthesized were applied as 2% water solutions directly into the eye of normotensive or glaucomatous albino rabbits. Very strong intraocular pressure (IOP) lowering was observed for many of them, and the active drug was detected in eye tissues and uids. This result prompted us to re-analyse the synthetic work done by other groups for the design of water soluble, topically effective antiglaucoma sulfonamides. According to these researchers, the IOP lowering effect is due to the intrinsic nature of the specic heterocyclic sulfonamide considered, among which the thienothiopyran-2-sulfonamide derivatives represent the best studied case. Indeed, the rst agents developed for such applications, such as dorzolamide, are derivatives of this ring system. In order to prove that the tail (in this case the nicotinoyl moiety) conferring water solubility to a sulfonamide CA inhibitor is critically important, similarly to the ring to which the sulfonamido group is grafted, we also prepared a dorzolamide derivative to which the nicotinoyl moiety was attached. This new compound is more water soluble than dorzolamide (as hydrochloride salt), behaves as a strong CA II inhibitor, and acts similarly to the parent derivative in lowering IOP in experimental animals. Thus, it seems that the tail conferring water solubility is more important for topical activity as an antiglaucoma drug than the heterocyclic/aromatic ring to which the sulfonamido moiety is grafted. 1999 ditions scientiques et mdicales Elsevier SAS carbonic anhydrase / aromatic, heterocyclic sulfonamides / nicotinoyl chloride / antiglaucoma drugs / hydrochloride salts / dorzolamide

1. Introduction The sulfonamides represent an important class of biologically active compounds, with at least ve different classes of pharmacological agents that have been obtained from the sulfanilamide structure as lead, the derivative initially studied by Domagk [2] as the rst modern chemotherapeutic drug. Indeed, the antibacterial
#

See [1]. *Correspondence and reprints

sulfonamides [3] continue to play an important role in chemotherapy, alone or in combination with other drugs [4], the sulfonamides that inhibit the zinc enzyme carbonic anhydrase (CA, EC 4.2.1.1) possess many applications as diuretic, antiglaucoma or antiepileptic drugs among others [57], the hypoglycaemic sulfonamides are extensively used in the treatment of some forms of diabetes [8], whereas the thiazides and highceiling diuretics might be considered as a fortunate development of the CA inhibitors [9], but these compounds possess a different pharmacological prole, independent of CA inhibition [10, 11]. Finally, some antithy-

800

Figure 1. Structures of dorzolamide 1 and brinzolamide 2.

roid drugs have also been developed starting from the sulfonamide structure as lead molecule [12]. The second class of the above mentioned pharmacological agents, ie., the sulfonamides with CA inhibitory action, have been thoroughly investigated in the last 10 years, mainly in the search for a topically effective antiglaucoma drug [1319]. The possibility of administering a sulfonamide via the topical route directly into the eye, although investigated in the 1950s [20, 21], has been totally unsuccessful, whereas the systemic administration, quite useful in lowering intraocular pressure (IOP), was generally accompanied by undesired side effects, due to CA inhibition in other tissues than the eye [21]. In 1983, Marens group [13] postulated that a water-soluble sulfonamide, also possessing a relatively balanced lipid solubility, would be an effective IOP lowering drug via the topical route, but at that moment no inhibitors possessing such physico-chemical properties existed. They started to be developed in several laboratories soon thereafter [1319], and in 1995 the rst such pharmacological agent, dorzolamide 1 entered into clinical use in the USA and Europe [22]. A second compound, brinzolamide 2, quite similar structurally to dorzolamide has also recently been approved for the topical treatment of glaucoma in the USA (gure 1) [23]. Thus, in a series of interesting papers [15, 2429], the Merck, Sharp and Dohme group has developed the synthesis of a large series of generally bicyclic heterocyclic sulfonamides (derivatives of benzothiazole- [24]; benzofuran- [25]; indole- [26]; benzo[b]-thiophene- [27, 28]; thieno-thiopyran [15, 29], etc), which were then tested as IOP lowering agents, and which led to the above mentioned drug (dorzolamide). Still, the greatest majority of the synthesized compounds proved to be potent allergens in vivo since their sulfonamido group was nucleophilically displaced by reduced glutathione. More than that, the only compounds with acceptable water solubility proved to be hydrochlorides of amino-derivatives of the thienothiopyran-sulfonamides of the dorzolamide type [15, 29]. Obviously, the approach followed by this group

was to explore as many heterocyclic rings as possible on which the sulfonamido moiety should be grafted, and this approach was extremely benecial to the chemistry of heterocyclic sulfonamides. Still, this approach seemed to us not a very fortunate one for the design of topically active IOP lowering agents, and we decided to explore the opposite one, i.e., to graft moieties that would ensure water solubility (as salts of a strong acid) on the classical ring systems of the aromatic/heterocyclic sulfonamides possessing CA inhibitory properties. In this paper we report the reaction of twenty aromatic/ heterocyclic sulfonamides containing a free amino, imino, hydrazino or hydroxyl group, with nicotinoyl chloride, which afforded a series of water-soluble (as hydrochloride or triate salts) sulfonamides with strong CA inhibitory properties. Moreover, dorzolamide has been derivatized similarly, at its secondary amino group, and the obtained compound also possessed a good water solubility as the hydrochloride salt. The new compounds reported here were tested for the inhibition of three CA isozymes, hCA I, hCA II and bCA IV (h = human, b = bovine isozyme). Affinities in the nanomolar range were detected for some compounds for isozymes II and IV. The most active derivatives were assayed in vivo in normotensive and glaucomatous rabbits for their IOP lowering properties. Very strong intraocular pressure (IOP) lowering was observed for many of them, and the active drug was detected in eye tissues and uids. Our conclusion is that the water-solubilizing tail seems to be more important than the ring on which the sulfonamido moiety is grafted, and that topically active antiglaucoma drugs might be obtained from many other classes of sulfonamides than the thienothiopyran-sulfonamides and their derivatives. 2. Experimental protocols Melting points were determined with a heating plate microscope and are not corrected; IR spectra were obtained in KBr pellets with a Perkin-Elmer 16PC FTIR

801 spectrometer, whereas 1H-NMR spectra were obtained with a Varian 300CXP apparatus in solvents specied in each case. Chemical shifts are expressed as values relative to Me4Si as standard. Elemental analyses were done by combustion for C, H, N with an automated Carlo Erba analyser, and were 0.4% of the theoretical values. Sulfonamides 322 used in synthesis were either commercially available compounds (from Sigma, Acros or Aldrich) or were prepared as described previously: 4-hydrazino-benzenesulfonamide 6 by diazotization of sulfanilamide followed by reduction of the diazonium salt with tin(II) chloride [30]; halogenosulfanilamides 912 by halogenation of sulfanilamide as reported in the literature [31]; compound 17 from 5-amino-1,3,4-thiadiazole-2-sulfonamide (obtained from acetazolamide) [32] by acylation with the phthalimido-derivative of -alanine, followed by hydrazinolysis [33], whereas imine 16 by deprotection of methazolamide with concentrated hydrochloric acid [34]. The benzothiazole-2sulfonamide derivatives 1820 were prepared as described in ref. [35], whereas the alcohols 21 and 22 from the corresponding amines by diazotization followed by hydrolysis of the diazonium salts [31]. Dorzolamide 1 was prepared as described in the literature [36]. Nicotinoyl chloride hydrochloride, triic acid and triethylamine were from Acros. Acetonitrile, acetone (Merck) or other solvents used in the synthesis were doubly distilled and kept on molecular sieves in order to maintain them in anhydrous conditions. 2.1. Chemistry 2.1.1. General procedure for the preparation of nicotinoyl derivatives of the aromatic/heterocyclic sulfonamides 2343 An amount of 10 mM sulfonamide 322 or 1 was dissolved/suspended in 50 mL of anhydrous acetonitrile or acetone and then treated with 0.178 g (10 mM) of nicotinoyl chloride hydrochloride. The stoichiometric amount (200 L) of triethylamine was then added and the reaction mixture was magnetically stirred at 4 C for 410 h. By means of TLC, the conversion of all the sulfonamides to the corresponding nicotinoyl derivatives has been monitored. When the reaction was completed, the solvent was evaporated until a small volume of the reaction mixture was obtained. Generally the new compounds crystallized spontaneously by leaving the above mixture at 4 C overnight. In some cases, the concentrated liquor obtained after the evaporation of the solvent was poured into 50 mL of cold water, then the reaction products precipitated and ltered. The prepared compounds were recrystallized from ethanol or ethanol-water (1:1, v/v). Yields were in the range of 7090%. Hydrochlorides of derivatives 2343 were obtained from the free bases and a methanolic HCl solution, in methanol as solvent. The hydrochlorides precipitated by leaving the above mixtures at 4 C overnight. The hydrochlorides were analysed for the presence of Cl by potentiometric titrations. The obtained data were 0.5% of the theoretical values calculated for the proposed formulas (data not shown). Triate salts were similarly obtained from the free bases 2343 and the stoichiometric amount of triic acid, in water as solvent. 2.1.1.1. 2-(Nicotinoylamido)-benzenesulfonamide 23 White crystals, m.p. 250252 C; IR (KBr), cm1: 1 140 (SO2sym), 1 290 (amide III), 1 370 (SO2as), 1 550 (amide II), 1 680 (amide I), 3 090 (NH); 3 360 (NH2); 1 H-NMR (DMSO-d6), , ppm: 7.057.98 (m, 4H, ArH from nicotinoyl); 7.157.66 (m, 4H, ArH, 1,2phenylene); 7.50 (br s, 2H, SO2NH2); 8.14 (br s, 1H, CONH); Anal., found: C, 51.90; H, 4.10; N, 14.96%; C12H11N3O3S requires: C, 51.98; H, 4.00; N, 15.15%. 2.1.1.2. 3-(Nicotinoylamido)-benzenesulfonamide 24 White crystals, m.p. 265266 C (dec.); IR (KBr), cm1: 1 135 (SO2sym), 1 290 (amide III), 1 370 (SO2as), 1 570 (amide II), 1 690 (amide I), 3 080 (NH); 3 360 (NH2); 1H-NMR (DMSO-d6), , ppm: 7.057.98 (m, 4H, ArH from nicotinoyl); 7.107.50 (m, 4H, ArH, 1,3phenylene); 7.56 (br s, 2H, SO2NH2); 8.11 (br s, 1H, CONH); Anal., found: C, 51.78; H, 3.85; N, 15.07%; C12H11N3O3S requires: C, 51.98; H, 4.00; N, 15.15%. 2.1.1.3. 4-(Nicotinoylamido)-benzenesulfonamide 25 White crystals, m.p. 280281 C (dec.); IR (KBr), cm1: 1 150 (SO2sym), 1 290 (amide III), 1 345 (SO2as), 1 560 (amide II), 1 690 (amide I), 3 060 (NH); 3 360 (NH2); 1H-NMR (DMSO-d6), , ppm: 7.057.98 (m, 4H, ArH from nicotinoyl); A 7.18, B 7.75 (AABBsystem, 4H, JAB = 7.9 Hz, ArH from 4-sulfamoylphenyl); 7.56 (br s, 2H, SO2NH2); 8.19 (br s, 1H, CONH); Anal., found: C, 51.67; H, 4.05; N, 14.88%; C12H11N3O3S requires: C, 51.98; H, 4.00; N, 15.15%. 2.1.1.4. 4-(Nicotinoylhydrazido)-benzenesulfonamide 26 White crystals, m.p. 265267 C; IR (KBr), cm1: 980 (NN), 1 150 (SO2sym), 1 290 (amide III), 1 365 (SO2as), 1 555 (amide II), 1 690 (amide I), 3 090 (NH); 3 360 (NH2); 1H-NMR (DMSO-d6), , ppm: 7.057.98 (m, 4H, ArH from nicotinoyl); A 7.18, B 7.71 (AABBsystem, 4H, JAB = 7.8 Hz, ArH from 4-sulfamoylphenyl); 7.59 (br s, 2H, SO2NH2); 8.06 (br s, 2H, CONHNH); Anal., found: C, 49.40; H, 4.16; N, 19.03%; C12H12N4O3S requires: C, 49.31; H, 4.14; N, 19.17%.

802 2.1.1.5. 4-(Nicotinoylamidomethyl)-benzenesulfonamide 27 White crystals, m.p. 271273 C (dec.); IR (KBr), cm1: 1 170 (SO2sym), 1 290 (amide III), 1 372 (SO2as), 1 545 (amide II), 1 690 (amide I), 3 090 (NH); 3 360 (NH2); 1H-NMR (DMSO-d6), , ppm: 4.90 (s, 2H, CH2); 7.057.98 (m, 4H, ArH from nicotinoyl); A 7.22, B 7.79 (AABBsystem, 4H, JAB = 7.9 Hz, ArH from 4-sulfamoylphenyl); 7.67 (br s, 2H, SO2NH2); 8.16 (br s, 1H, CONH); Anal., found: C, 53.81; H, 4.78; N, 14.21%; C13H13N3O3S requires: C, 53.60; H, 4.50; N, 14.42%. 2.1.1.6. 4-(Nicotinoylamidoethyl)-benzenesulfonamide 28 White crystals, m.p. 278280 C (dec.); IR (KBr), cm1: 1 150 (SO2sym), 1 290 (amide III), 1 359 (SO2as), 1 540 (amide II), 1 690 (amide I), 3 080 (NH); 3 360 (NH2); 1H-NMR (DMSO-d6), , ppm: 3.10 (t, 2H, CH2 from the CH2CH2 bridge); 3.70 (t, 2H, CH2 from the CH2CH2 bridge); 7.057.98 (m, 4H, ArH from nicotinoyl); A 7.15, B 7.62 (AABBsystem, 4H, JAB = 7.9 Hz, ArH from 4-sulfamoylphenyl); 7.67 (br s, 2H, SO2NH2); 8.17 (br s, 1H, CONH); Anal., found: C, 55.40; H, 5.03; N, 13.45%; C14H15N3O3S requires: C, 55.07; H, 4.95; N, 13.76%. 2.1.1.7. 3-Fluoro-4-(nicotinoylamido)-benzenesulfonamide 29 White crystals, m.p. 234235 C. IR (KBr), cm1: 1 150 (SO2sym), 1 290 (amide III), 1 348 (SO2as), 1 550 (amide II), 1 680 (amide I), 3 060 (NH); 3 360 (NH2); 1 H-NMR (DMSO-d6), , ppm: 6.60 (br s, 2H, SO2NH2); 7.057.98 (m, 4H, ArH from nicotinoyl); 7.057.89 (m, 3H, Ar H from the F-substituted ring); 8.15 (br s, 1H, CONH); Analysis, found: C, 48.54; H, 3.61; N, 14.07%; C12H10FN3O3S requires: C, 48.81; H, 3.41; N, 14.23%. 2.1.1.8. 3-Chloro-4-(nicotinoylamido)-benzenesulfonamide 30 White crystals, m.p. 238239 C. IR (KBr), cm1: 1 155 (SO2sym), 1 290 (amide III), 1 339 (SO2as), 1 550 (amide II), 1 690 (amide I), 3 090 (NH); 3 360 (NH2); 1 H-NMR (DMSO-d6), , ppm: 6.70 (br s, 2H, SO2NH2); 7.057.98 (m, 4H, ArH from nicotinoyl); 7.057.76 (m, 3H, Ar H the 2-Cl-substituted ring); 8.15 (br s, 1H, CONH); Analysis, found: C, 46.27; H, 3.37; N, 13.39%; C12H10ClN3O3S requires: C, 46.23; H, 3.23; N, 13.48%. 2.1.1.9. 3-Bromo-4-(nicotinoylamido)-benzenesulfonamide 31 White crystals, m.p. 230232 C. IR (KBr), cm1: 1 160 (SO2sym), 1 290 (amide III), 1 356 (SO2as), 1 540 (amide II), 1 690 (amide I), 3 060 (NH); 3 360 (NH2); 1 H-NMR (DMSO-d6), , ppm: 6.65 (br s, 2H, SO2NH2); 7.057.98 (m, 4H, ArH from nicotinoyl); 7.057.86 (m, 3H, Ar H the 2-Br-substituted ring); 8.14 (br s, 1H, CONH); Analysis, found: C, 40.55; H, 3.00; N, 11.50%; C12H10BrN3O3S requires: C, 40.46; H, 2.83; N, 11.80%. 2.1.1.10. 3-Iodo-4-(nicotinoylamido)-benzenesulfonamide 32 White crystals, m.p. 210212 C. IR (KBr), cm1: 1 145 (SO2sym), 1 290 (amide III), 1 360 (SO2as), 1 545 (amide II), 1 690 (amide I), 3 070 (NH); 3 360 (NH2); 1 H-NMR (DMSO-d6), , ppm: 6.60 (br s, 2H, SO2NH2); 7.057.98 (m, 4H, ArH from nicotinoyl); 7.087.79 (m, 3H, Ar H the 2-I-substituted ring); 8.14 (br s, 1H, CONH); Analysis, found: C, 35.87; H, 2.43; N, 10.36%; C12H10IN3O3S requires: C, 35.75; H, 2.50; N, 10.42%. 2.1.1.11. 4,5-Dichloro-6-nicotinoylamido-benzene-1,3disulfonamide 33 White crystals, m.p. 267268 C. IR (KBr), cm1: 1 140 (SO2sym), 1 290 (amide III), 1 370 (SO2as), 1 550 (amide II), 1 690 (amide I), 3 080 (NH); 3 360 (NH2); 1 H-NMR (DMSO-d6), , ppm: 7.057.98 (m, 4H, ArH from nicotinoyl); 7.54 (s, 1H, ArH from the pentasubstituted benzene ring); 7.68 (br s, 4H, 2 SO2NH2); 8.10 (br s, 1H, CONH); Analysis, found: C, 33.69; H, 2.40; N, 13.08%; C12H10Cl2N4O5S2 requires: C, 33.89; H, 2.37; N, 13.17%. 2.1.1.12. 6-Chloro-4-nicotinoylamido-benzene-1,3-disulfonamide 34 White crystals, m.p. 290294 C (dec.). IR (KBr), cm1: 1 150 (SO2sym), 1 290 (amide III), 1 330 (SO2as), 1 540 (amide II), 1 680 (amide I), 3 060 (NH); 3 360 (NH2); 1H-NMR (DMSO-d6), , ppm: A 7.057.98 (m, 4H, ArH from nicotinoyl); 7.35 (s, 1H, ArH from disulfamoylphenyl); 7.59 (s, 1H, ArH from disulfamoylphenyl); 7.75 (br s, 4H, 2 SO2NH2); 8.14 (br s, 1H, CONH); Analysis, found: C, 36.59; H, 2.90; N, 14.21%; C12H11ClN4O5S2 requires: C, 36.88; H, 2.84; N, 14.34%. 2.1.1.13. 5-Nicotinoylamido-1,3,4-thiadiazol-2-sulfonamide 35 White crystals, m.p. > 310 C; IR (KBr), cm1: 1 180 (SO2sym), 1 295 (amide III), 1 340 (SO2as), 1 545 (amide II), 1 690 (amide I), 3 060 (NH), 3 375; 1H-NMR (DMSO-d6), , ppm: 6.94 (br s, 2H, SO2NH2); 7.057.98 (m, 4H, ArH from nicotinoyl); 8.26 (br s, 1H, CONH); Anal., found, C, 33.60; H, 2.60; N, 24.45%; C8H7N5O3S2 requires: C, 33.68; H, 2.47; N, 24.55%. 2.1.1.14. 5-Nicotinoylimido-4-methyl-2-sulfonamido-21,3,4-thiadiazoline 36 White crystals, m.p. > 310 C; IR (KBr), cm1: 1 182 (SO2sym), 1 290 (amide III), 1 366 (SO2as), 1 540 (amide

803 II), 1 690 (amide I), 3 080 (NH), 3 380 (NH2); 1H-NMR (DMSO-d6), , ppm: 3.90 (s, 3H, Me); 6.96 (br s, 2H, SO2NH2); 7.057.98 (m, 4H, ArH from nicotinoyl); Anal., found, C, 35.99; H, 3.12; N, 23.25%; C9H9N5O3S2 requires: C, 36.11; H, 3.03; N, 23.40%. 2.1.1.15. 5-(Nicotinoylamidoethylcarboxamido)-1,3,4thiadiazol-2-sulfonamide 37 White crystals, m.p. 287289 C (dec.), IR (KBr), cm1: 1 150 (SO2sym), 1 270 and 1 290 (amide III), 1 330 (SO2as), 1 450, 1 570 (amide II), 1 690 and 1 710 (amide I), 3 090 (NH); 3 360 (NH2); 1H-NMR (DMSO-d6), , ppm: 2.252.60 (m, 4H, CH2CH2); 6.88 (br s, 3H, CONH + SO2NH2); 7.057.98 (m, 4H, ArH from nicotinoyl); 8.24 (br s, 1H, CONH from nicotinoylamido moiety); Analysis, found: C, 37.15; H, 3.19; N, 23.46%; C11H12N6O4S2 requires: C, 37.07; H, 3.39; N, 23.58%. 2.1.1.16. 6-Nicotinoylamido-benzothiazol-2-sulfonamide 38 White crystals, m.p. 290294 C (dec.), IR (KBr), cm1: 1 165 (SO2sym), 1 290 (amide III), 1 344 (SO2as), 1 540 (amide II), 1 680 (amide I), 3 060 (NH); 3 360 (NH2); 1H-NMR (DMSO-d6), , ppm: 7.057.98 (m, 4H, ArH from nicotinoyl); 6.94 (dd, 1H, J = 9 Hz; J = 3 Hz, H-5 of benzothiazole); 7.10 (d, 1H, J = 3 Hz, H-7 of benzothiazole); 7.78 (d, 1H, J = 9 Hz, H-4 of benzothiazole); 8.10 (br s, 2H, SO2NH2); 8.18 (br s, 1H, CONH); Analysis, found: C, 46.85; H, 2.94; N, 16.47%; C13H10N4O3S2 requires: C, 46.70; H, 3.01; N, 16.76%. 2.1.1.17. 6-Nicotinoyloxy-benzothiazol-2-sulfonamide 39 White crystals, m.p. 281283 C (dec.), IR (KBr), cm1: 1 030 (COO), 1 160 (SO2sym), 1 350 (SO2as), 1 450, 1 775 (COO), 3 360 (NH2); 1H-NMR (DMSO-d6), , ppm: 7.057.98 (m, 4H, ArH from nicotinoyl); 6.90 (dd, 1H, J = 9 Hz; J = 3 Hz, H-5 of benzothiazole); 7.11 (d, 1H, J = 3 Hz, H-7 of benzothiazole); 7.79 (d, 1H, J = 9 Hz, H-4 of benzothiazole); 8.10 (br s, 2H, SO2NH2); Analysis, found: C, 46.40; H, 2.90; N, 12.38%; C13H9N3O4S2 requires: C, 46.56; H, 2.71; N, 12.53%. 2.1.1.18. 6-Nicotinoyloxyethyloxy-benzothiazol-2-sulfonamide 40 White crystals, m.p. 245246 C, IR (KBr), cm1: 1 030 (COO), 1 175 (SO2sym), 1 341 (SO2as), 1 450, 1 770 (COO), 3 360 (NH2); 1H-NMR (DMSO-d6), , ppm: 2.89 (t, 3H, CH2); 3.14 (t, 3H, CH2); 6.95 (dd, 1H, J = 9 Hz; J = 3 Hz, H-5 of benzothiazole); 7.057.98 (m, 4H, ArH from nicotinoyl); 7.10 (d, 1H, J = 3 Hz, H-7 of benzothiazole); 7.79 (d, 1H, J = 9 Hz, H-4 of benzothiazole); 8.15 (br s, 2H, SO2NH2); Analysis, found: C, 47.58; H, 3.66; N, 10.89%; C15H13N3O5S2 requires: C, 47.49; H, 3.45; N, 11.07%. 2.1.1.19. 4-(Nicotinoyloxymethyl)-benzenesulfonamide 41 White crystals, m.p. 244246 C; IR (KBr), cm1: 1 040 (COO), 1 155 (SO2sym), 1 325 (SO2as), 1 780 (COO), 3 310 (NH2); 1H-NMR (DMSO-d6), , ppm: 4.90 (s, 2H, CONHCH2); 7.057.98 (m, 4H, ArH from nicotinoyl); 7.087.41 (m, AABB, J = 7.2 Hz; 4H, ArH, phenylene); 7.49 (s, 2H, SO2NH2); Anal., found, C, 53.19; H, 4.21; N, 9.37%; C13H12N2O4S requires: C, 53.42; H, 4.14; N, 9.58%. 2.1.1.20. 4-(Nicotinoyloxyethyl)-benzenesulfonamide 42 White crystals, m.p. 240243 C. IR (KBr), cm1: 1 040 (COO), 1 157 (SO2sym), 1 332 (SO2as), 1 760 (COO), 3 300 (NH2); 1H-NMR (DMSO-d6), , ppm: 3.10 (t, 2H, CH2 from the CH2CH2 bridge); 3.70 (t, 2H, CH2 from the CH2CH2 bridge); 6.95 (br s, 2H, SO2NH2); 7.057.98 (m, 4H, ArH from nicotinoyl); 7.057.52 (m, AABB, J = 7.3 Hz, 4H, ArH, phenylene); Anal., found, C, 54.95; H, 4.67; N, 8.97%; C14H14N2O4S requires: C, 54.89; H, 4.61; N, 9.14%. 2.1.1.21. 5,6-Dihydro-4-[N-nicotinoyl-(ethylamido)]-6methyl-4H-thieno-[2,3-b]thiopyran-2-sulfonamide 7,7dioxide 43 White crystals, m.p. 270272 C; IR (KBr), cm1: 1 135 (SO2sym), 1 290 (amide III), 1 345 (SO2as), 1 545 (amide II), 1 680 (amide I), 3 366 (NH2); 1H-NMR (DMSO-d6), , ppm: 1.29 (d, 3H, Me); 1.39 (t, 3H, Me from ethyl); 2.55 (m, 1H, CH); 2.80 (m, 1H, CH); 3.053.20 (m, 2H, CH2 from ethyl); 4.37 (m, 2H, CH2); 7.057.98 (m, 4H, ArH from nicotinoyl); 8.03 (s, 1H, CH, ArH from thienyl); 8.25 (br s, 2H, SO2NH2); Anal., found, C, 44.57; H, 4.39; N, 9.66%; C16H19N3O5S3 requires: C, 44.74; H, 4.46; N, 9.78%. 2.2. Pharmacology 2.2.1. Enzyme assay Human CA I and CA II cDNAs were expressed in Escherichia coli strain BL21 (DE3) from the plasmids pACA/hCA I and pACA/hCA II described by Forsman et al. [37] (the two plasmids were a gift from Prof. Sven Lindskog, Umea University, Sweden). Cell growth conditions were those described by Lindskogs group [38] and enzymes were puried by affinity chromatography according to the method of Khalifah et al. [39]. Enzyme concentrations were determined spectrophotometrically at 280 nm, utilizing a molar absorptivity of 49 mM1.cm1 for CA I and 54 mM1.cm1 for CA II, respectively, based on Mr = 28.85 kDa for CA I, and 29.30 kDa for CA II, respectively [40, 41]. CA IV was isolated from bovine lung microsomes as described by

804 Maren et al., and its concentration has been determined by titration with ethoxzolamide [42]. Initial rates of 4-nitrophenyl acetate hydrolysis catalysed by different CA isozymes were monitored spectrophotometrically, at 400 nm, with a Cary 3 instrument interfaced with an IBM compatible PC [43]. Solutions of substrate were prepared in anhydrous acetonitrile; the substrate concentrations varied between 2 102 and 1 106 M, working at 25 C. A molar absorption coefcient e of 18 400 M1.cm1 was used for the 4-nitrophenolate formed by hydrolysis, in the conditions of the experiments (pH 7.40), as reported in the literature [43]. Non-enzymatic hydrolysis rates were always subtracted from the observed rates. Duplicate experiments were done for each inhibitor concentration, and the values reported throughout the paper are the mean of such results. Stock solutions of inhibitor (1 mM) were prepared in distilled-deionized water with 1020% (v/v) DMSO (which is not inhibitory at these concentrations) and dilutions up to 0.01 nM were done thereafter with distilled-deionized water. Inhibitor and enzyme solutions were preincubated together for 10 min at room temperature prior to assay, in order to allow for the formation of the E-I complex. The inhibition constant KI was determined as described by Pocker and Stone [43]. Enzyme concentrations were 3.5 nM for hCA II, 12 nM for hCA I and 36 nM for bCA IV (this isozyme has a decreased esterase activity [44] and higher concentrations had to be used for the measurements). 2.2.2. Measurement of tonometric IOP Adult male New Zealand albino rabbits weighing 33.5 kg were used in the experiments (three animals were used for each inhibitor studied). The experimental procedures conform to the Association for Research in Vision and Ophthalmology Resolution on the use of animals. The rabbits were kept in individual cages with food and water provided ad libitum. The animals were maintained on a 12 h:12 h light/dark cycle in a temperature controlled room, at 2226 C. Solutions of inhibitors (2%, as hydrochlorides, by weight) were obtained in distilled deionized water. The pH of these solutions was around 5.506.40. IOP was measured using a Digilab 30R pneumatonometer (BioRad, Cambridge, MA, USA) as described by Marens group [45, 46]. The pressure readings were matched with two-point standard pressure measurements at least twice each day using a Digilab calibration verier. All IOP measurements were done by the same investigator with the same tonometer. One drop of 0.2% oxybuprocaine hydrochloride (novesine, Sandoz) diluted 1:1 with saline was instilled in each eye immediately before each set of pressure measurements. IOP was measured three times at each time interval, and the means reported. IOP was measured rst, immediately before drug administration, then at 30 min after the instillation of the pharmacological agent, and then every 30 min for a period of several hours. For all IOP experiments, drug was administered to only one eye, leaving the contralateral eye as an untreated control. The ocular hypotensive activity is expressed as the average difference in IOP between the treated and control eye, in this way minimizing the diurnal, seasonal and interindividual variations commonly observed in the rabbit [45, 46]. All data are expressed as mean SE, using a one-tailed t test. Ocular hypertension was elicited in the right eye of albino rabbits by the injection of -chymotrypsin (from Sigma) as described by Melena et al. [48]. The IOP of operated animals was checked after approximately four weeks, and animals with an elevated pressure of 3036 mm Hg were used at least one month after the injection of -chymotrypsin. 2.2.3. Drug distribution in ocular uids and tissues The general procedure of Marens group has been followed [45, 46]. The animals were killed with an intracardiac air injection. Aqueous humour (both posterior and anterior chamber uids) were withdrawn. Then, the cornea and anterior uvea (iris plus attached ciliary body) were dissected, rinsed well with water, blotted, weighed and put into 12 mL of water. For isolation of the ciliary processes, intact anterior uvea rings were placed on a paralm covered piece of polystyrene foam in a Petri dish. The tissue was wetted with normal saline and dissected under a microscope, then ciliary processes were liberated from their attachment to the iris, cut, weighed and put in 0.5 mL of distilled water. The tissue from 4 eyes (average weight of 8 mg/eye) was pooled for drug analysis. Samples were boiled for 5 min (in order to denature CA, and free drug from the E-I complex), diluted and then incubated with a known amount of enzyme. The activity of the free enzyme and in the presence of the inhibitor were determined as described above. A calibration curve has been used in order to determine the fractional inhibition in the different tissues, as described in [45, 46]. 3. Results Compounds prepared by reaction of nicotinoyl chloride with aromatic/heterocyclic sulfonamides, of type 2343, are shown below (gure 2). Inhibition data against three CA isozymes, hCA I, hCA II and bCA IV with compounds 143 are presented in table I. In vivo IOP

805
Table I. CA inhibition data with standard inhibitors 12, the parent sulfonamides 322 and the new derivatives 2343 reported in the present study, against isozymes I, II and IV. Inhibitor Dorzolamide 1 Brinzolamidec 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43
*

KI* (nM) hCA Ia 50 000 45 400 25 000 28 000 78 500 25 000 21 000 8 300 9 800 6 500 6 000 6 100 8 400 8 600 9 300 455 70 55 50 24 000 18 000 21 800 20 500 16 000 23 200 1 200 1 100 547 630 650 650 540 628 36 28 21 13 9 9 2 160 2 100 2 000 hCA IIa 9 3.2 295 240 300 320 170 160 60 110 40 70 28 75 60 19 3 9 8 7 125 110 310 285 142 324 79 66 38 49 45 42 36 39 2 3 5 5 3 2 79 64 5 bCA IVb 45 45.3 1 310 2 200 3 000 3 215 2 800 2 450 180 320 66 125 175 160 540 355 125 19 17 15 560 450 570 310 165 400 115 102 70 78 80 69 77 65 9 11 12 19 8 7 140 116 11

Figure 2. Structure of derivatives 343.

lowering data with some of the most active CA inhibitors reported here, in normotensive and glaucomatous rabbits, after topical administration of the drug, are shown in tables II and III, respectively. Ex vivo distribution data of compound 35 in ocular tissues and uids after the topical administration in normotensive rabbits, are shown in table IV.

Standard error for the determination of KI-s was of 510% (from two different assays). aHuman (cloned) isozyme; bIsolated from bovine lung microsomes; cFrom [47].

4. Discussion Reaction of sulfonamides 322 or 1 with nicotinoyl chloride afforded a series of new compounds of type

806
Table II. Fall of IOP of normotensive rabbits (20.1 2.0 mm Hg), after treatment with one drop (50 L) of a solution 2 % of CA inhibitor (as hydrochloride salt, with the pH value shown below) directly into the eye, at 30, 60 and 90 min after administration. Inhibitor 1 (dorzolamide) 35 36 37 39 40 43 * IOP = IOPcontrol pH 5.5 5.5 5.8 5.5 5.7 5.5 5.5
eye

IOP (mm Hg)* t=0 0 0 0 0 0 0 0

eye

t = 30 min 2.2 0.10 5.4 0.12 6.2 0.10 5.1 0.12 2.1 0.05 2.4 0.05 2.5 0.06

t = 60 min 4.1 0.15 10.9 0.11 12.4 0.14 8.1 0.11 4.0 0.10 4.3 0.11 4.5 019

t = 90 min 2.7 0.08 12.5 0.17 14.1 0.12 8.6 0.12 4.5 0.10 3.5 0.08 7.0 0.14

IOPtreated

; Mean average spread (n = 3).

2343. The reaction was generally performed in acetone or acetonitrile as solvent, in the presence of triethylamine as base. In the case of compounds 15 and 16, the above procedure led to very low yields of nicotinoylamido derivatives, and Schotten-Baumann conditions were applied for obtaining 35 and 36 in good yields. Hydrochlorides of the new derivatives were then prepared by reacting the free bases 2343 with a methanolic HCl solution. Similarly were obtained the triate salts, by reaction of bases 2343 with triic acid in water as solvent. These salts possess a very good water solubility, generally in the range of 35% (data not shown). The pH of such solutions were generally around 5.56.0, making them appropriate for topical application directly into the eye. Compounds 343 were characterized by standard chemical and physical methods that conrmed their structure (see Experimental protocols for details) and were assayed for the inhibition of isozymes hCA I, hCA II and bCA IV (table I). Inhibition data against the three CA isozymes, hCA I, hCA II and bCA IV with the new derivatives (table I) prove that the nicotinoylamido-sulfonamides 2343 reported here generally behave as strong inhibitors, with greatly increased efficiencies as compared to the parent compounds from which they were prepared (the sulfon-

amides 322). The efficiency of the obtained inhibitor generally varied in the following way, based on the parent sulfonamide from which it was prepared: the derivative of p-hydrazino-benzenesulfonamide 26 < the orthanilamide 23 the metanilamide 24 < the sulfanilamide 25 < the homosulfanilamides 27 < the p-aminoethyl-benzenesulfonamides 28 < the 1,3-benzene-disulfonamides 33 and 34 the halogeno-substituted sulfanilamides 2932 < the 1,3,4-thiadiazole-2-sulfonamides 35 and 37 4-methyl-2-1,3,4-thiadiazoline-2-sulfonamide 36 the dorzolamide derivative 43 < the benzothiazole-2sulfonamides 3840. All three CA isozymes investigated here were susceptible to inhibition with this type of sulfonamide, with hCA II and bCA IV the most inhibitable, followed by hCA I, generally less susceptible to inhibition as compared to the rst two isozymes. The promising in vitro CA inhibitory activity of some of the newly prepared compounds prompted us to investigate their effect in vivo, on the intraocular pressure (IOP), after topical application directly into the eye, in normotensive and glaucomatous rabbits, frequently used as an animal model of glaucoma [1315, 22, 23, 45]. Some of these results are shown in tables II and III. The inhibitors selected for in vivo studies were among the most active against hCA II and IV, in the prepared series, such as compounds 3540, and 43. The following

Table III. Fall of IOP of glaucomatous rabbits (34.1 2.0 mm Hg), after treatment with one drop (50 L) of a solution 2 % of CA inhibitor (as hydrochloride salt, with the pH value shown below) directly into the eye, at 30, 60 and 90 min after administration. Inhibitor 1 35 36 43 * IOP = IOPcontrol pH 5.5 5.5 5.8 5.5
eye

IOP (mm Hg)* t=0 0 0 0 0

eye

t = 30 min 4.3 0.25 10.3 0.20 8.5 0.10 4.8 0.10

t = 60 min 7.1 0.30 15.2 0.20 13.2 0.20 6.6 0.10

t = 90 min 5.0 0.25 19.1 0.15 20.4 0.20 11.6 0.15

IOPtreated

; Mean average spread (n = 3).

807
Table IV. Ocular tissue concentrations (M) after 1 and 2 h, following corneal application of one drop (50 L) of a 2 % solution of the compound 35 in normotensive albino rabbits. Time (h) 1h 2h Drug concentration (M)* Cornea 150 5 47 4 Aqueous humour 283 10 39 3 Ciliary process 51 3 10 1
[2] [3] [4]

References
[1] Menabuoni L., Scozzafava A., Mincione F., Briganti F., Mincione G., Supuran C.T., J. Enzyme Inhib. 15 in press. (1999) Domagk G., Dt. Med. Wocheschr. 61 (1935) 250254. Northey E.H., in: The Sulfonamides and Allied Compounds, Reinhold, New York, 1948, pp. 1267. Mandell G.L., Sande M.A., in: Gilman A.G., Rall T.W., Nies A.S., Taylor P. (Eds.), The Pharmacological Basis of Therapeutics, 8th Edition, Pergamon Press, New York, 1990, pp. 10471064. Supuran C.T., in: Puscas I. (Ed.), Carbonic Anhydrase and Modulation of Physiologic and Pathologic Processes in the Organism, Helicon Timisoara, 1994, pp. 29111. Weiner I.M., in: Gilman A.G., Rall T.W., Nies A.S., Taylor P. (Eds.), The Pharmacological Basis of Therapeutics, 8th Edition, Pergamon Press, New York, 1990, pp. 713732. Supuran C.T., Scozzafava A., Ilies M.A., Iorga B., Cristea T., Briganti F., Chiraleu F., Banciu M.D., Eur. J. Med. Chem. 33 (1998) 577595. Boyd A.E., Diabetes 37 (1988) 847850. Beyer K.H., Baer J.E., Pharmacol. Rev. 13 (1961) 517562. Maren T.H., Drug Dev. Res. 10 (1987) 255276. Supuran C.T., Conroy C.W., Maren T.H., Eur. J. Med. Chem. 31 (1996) 843846. Maren T.H., Annu. Rev. Pharmacol. Toxicol. 16 (1976) 309327. Maren T.H., Jankowska L., Edelhauser G.F., Sanyal G., Exp. Eye Res. 36 (1983) 457480. Katritzky A.R., Caster K.C., Maren T.H., Conroy C.W., Bar-Ilan A., J. Med. Chem. 30 (1987) 20582062. Ponticello G.S., Freedman M.B., Habecker C.N., Lyle P.A., Schwam H., Varga S.L. et al., J. Med. Chem. 30 (1987) 591597. Boriack P.A., Christianson D.W., Kingery-Wood J., Whitesides G.M., J. Med. Chem. 38 (1995) 22862291. Supuran C.T., Popescu A., Ilisiu M., Costandache A., Banciu M.D., Eur. J. Med. Chem. 31 (1996) 439448. Supuran C.T., Briganti F., Scozzafava A., J. Enzyme Inhib. 12 (1997) 175190. Supuran C.T., Scozzafava A., Popescu A., Bobes-Tureac R., Banciu A., Creanga A., Bobes-Tureac G., Banciu M.D., Eur. J. Med. Chem. 32 (1997) 445452. Becker B., Am. J. Ophthalmol. 39 (1955) 177183. Maren T.H., Physiol. Rev. 47 (1967) 595782. Ponticello G.S., Sugrue M.F., Plazonnet B., Durand-Cavagna G., Pharm. Biotechnol. 11 (1998) 555574. Silver L.H., Am. J. Ophthalmol. 126 (1998) 400408. Graham S.L., Shepard K.L., Anderson P.S., Baldwin J.J., Best D.B., Christy M.E. et al., J. Med. Chem. 32 (1989) 25482554. Graham S.L., Hoffman J.M., Gautheron P., Michelson S.R., Scholz T.H., Schwam H. et al., J. Med. Chem. 33 (1990) 749754. Graham S.L., Scholz T.H., Synthesis (1986) 10311033. Prugh J.D., Hartmann G.D., Mallorga P.J., McKeever B.M., Michelson S.R., Murcko M.A. et al., J. Med. Chem. 34 (1991) 18051818. Hartmann G.D., Halczenko W., Prugh J.D., Smith R.L., Sugrue M.F., Mallorga P.J. et al., J. Med. Chem. 35 (1992) 30273033. Baldwin J.J., Ponticello G.S., Anderson G.S., Christy M.E., Murcko M.A., Randall W.C. et al., J. Med. Chem. 32 (1989) 25102513. Crippa G.B., Maffei S., Gazz. Chim. Ital. 71 (1941) 9799. Cingolani E., Gazz. Chim. Ital. 78 (1948) 275282.

* Mean standard deviation (n = 3).

[5]

facts should be noted regarding the data of tables II and III. Some of the new compounds assayed in vivo, such as 35, 36, 37 and 43, showed much more effective IOP lowering effects as compared to dorzolamide 1, both after 30 min from the administration of the inhibitor within the rabbit eye, as well as at other times (1, 1.5 and 2 h, respectively), in normotensive as well as glaucomatous animals. A second group of inhibitors, such as 39 and 40, showed IOP lowering effects of the same order of magnitude as those of dorzolamide, both after half an hour or longer periods after the administration. Mention should be made that the pH of the solutions administered in these experiments was in the range of 5.05.9 for all inhibitors used. In table IV, the drug distribution in ocular uids and tissues of normotensive rabbits is shown, after the topical administration of one of the most active topical inhibitors in the prepared series, i.e., compound 35. It is seen from the above data that at 1 and 2 h after topical administration of the drug, high levels of 35 were found in the cornea, aqueous humour and ciliary processes. Based on the inhibition constant of this compound (2 nM for CA II, and 9 nM for CA IV, respectively), the fractional inhibition estimated in these tissues/uids is of 99.599.9%, proving the fact that the IOP decrease is indeed due to CA inhibition [45, 46]. In conclusion, we report here a general approach for the preparation of water-soluble, topically effective antiglaucoma sulfonamides, by attaching water-solubilizing moieties (such as isonicotinoyl) to well-known aromatic/ heterocyclic sulfonamides. The new compounds reported here might lead to the development of more efficient antiglaucoma drugs. Acknowledgements This research was nanced by the EU grant ERB CIPDCT 940051. Thanks are addressed to Drs M.A. Ilies and M. Barboiu for expert technical assistance with the preparation of several sulfonamide intermediates used in this work.

[6]

[7]

[8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19]

[20] [21] [22] [23] [24] [25] [26] [27]

[28] [29] [30] [31]

808
[32] [33] [34] [35] [36] [37] [38] [39] [40] Jitianu A., Ilies M.A., Scozzafava A., Supuran C.T., Main Group Met. Chem. 20 (1997) 147153. Barboiu M., Scozzafava A., Supuran C.T., J. Enzyme Inhib. 15 in press. (1999) Scozzafava A., Supuran C.T., J. Enzyme Inhib. 13 (1998) 419442. Schoenwald R.D., Eller M.G., Dixson J.A., Barfknecht C.F., J. Med. Chem. 27 (1984) 810812. Blacklock T.J., Sohar P., Butcher J.W., Lamanec T., Grabowski E.J.J., J. Org. Chem. 58 (1993) 16721679. Forsman C., Behravan G., Osterman A., Jonsson B.H., Acta Chem. Scand. B42 (1988) 314318. Behravan G., Jonasson P., Jonsson B.H., Lindskog S., Eur. J. Biochem. 198 (1991) 589592. Khalifah R.G., Strader D.J., Bryant S.H., Gibson S.M., Biochemistry 16 (1977) 22412247. Nyman P.O., Lindskog S., Biochim. Biophys. Acta 85 (1964) 141151. [41] [42] [43] [44] [45] [46] [47] [48] Henderson L.E., Henriksson D., Nyman P.O., J. Biol. Chem. 251 (1976) 54575463. Maren T.H., Wynns G.C., Wistrand P.J., Mol. Pharmacol. 44 (1993) 901906. Pocker Y., Stone J.T., Biochemistry 6 (1967) 668678. Baird T.T., Waheed A., Okuyama T., Sly W.S., Fierke C.A., Biochemistry 36 (1997) 26692678. Maren T.H., Brechue W.F., Bar-Ilan A., Exp. Eye Res. 55 (1992) 7379. Brechue W.F., Maren T.H., Invest. Ophthalmol. Vis. Sci. 34 (1993) 25812587. Stams T., Chen Y., Boriack-Sjodin P.A., Hurt J.D., Liao J., May J.A. et al., Protein Sci. 7 (1998) 556563. Melena J., Santafe J., Segarra-Domenech J., Purras G., J. Ocul. Pharmacol. Ther. 15 (1999) 1927.

Eur. J. Med. Chem. 34 (1999) 809824 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

809

Original article

5-Substituted UTP derivatives as P2Y2 receptor agonists#

Bernd H.A. Knoblaucha,d, Christa E. Mllera,c*, Leif Jrlebarkd, Grace Lawokod, Thomas Kottkeb, Martin A. Wikstrme, Edith Heilbronnd
Institut fr Pharmazie und Lebensmittelchemie, Pharmazeutische Chemie, Universitt Wrzburg, Am Hubland, D-97074 Wrzburg, Germany b Institut fr Anorganische Chemie, Universitt Wrzburg, Am Hubland, D-97074 Wrzburg, Germany c Universitt Bonn, Pharmazeutisches Institut, Pharmazeutische Chemie Poppelsdorf, Germany d Stockholm University, Department of Neurochemistry and Neurotoxicology, Arrhenius Laboratories of Natural Sciences, Stockholm, Sweden e Karolinska Institute, Nobel Institute for Neurophysiology, Department of Neuroscience, Stockholm, Sweden (Received 8 February 1999; revised 30 April 1999; accepted 6 May 1999)
a

Abstract A series of 5-alkyl-substituted UTP derivatives, which had been synthesized previously with a moderate degree of purity, was resynthesized, puried, and characterized. Synthetic and purication procedures were optimized. New spectroscopic data, including 13C- and 31 P NMR data, are presented. Phosphorylation reactions yielded a number of side products, such as the 2-, 3-, and 5-monophosphates, the 2,3-cyclic monophosphates, and the 2,3-cyclic phosphates of the 5-triphosphates. Furthermore, raw products were contaminated with inorganic phosphates, including cyclometatriphosphate, phosphate, and pyrophosphate. The uracil nucleotides were investigated for their potency to increase intracellular calcium concentrations by stimulation of P2Y2 receptors (P2Y2R) on NG10815 cells, a mouse neuroblastoma glioma cell line, and in human basal epithelial airway cells, including a cystic brosis (CF/T43) cell line. UTP exhibited EC50 values of ca. 1 M (in NG10815 cells) and of 0.1 M (in CF/T43 cells), respectively. 5-Substituted UTP derivatives were agonists at the P2Y2R, but were less potent than UTP. 5-Ethyl-UTP, for example, exhibited an EC50 value of 99 M at P2Y2R of NG10815 cells and proved to be a full agonist. With increasing volume of the 5-substituent of UTP derivatives, P2Y2 activity decreased. 1999 ditions scientiques et mdicales Elsevier SAS P2Y2 receptor agonists / P2U receptor agonists / UTP analogues / pyrimidine nucleotides / cystic brosis

1. Introduction Membrane receptors for endogenous nucleotides, such as ATP and UTP (gure 1) belong to the purine/ pyrimidine receptor family. Purine receptors are subdivided into two separate families, the adenosine recepPreliminary results were presented at the Joint Meeting of the German and Swiss Pharmaceutical Societies in Zrich, Switzerland, 1997, and at the International Symposium on Adenosine and Adenine Nucleotides in Ferrara, Italy (abstract in Drug Dev. Res. 43 (1998) 34). * Correspondence and reprints Dr Christa Mller, Pharmazeutisches Institut der Universitt Bonn, Pharmazeutische Chemie Poppelsdorf, Kreuzbergweg 26, D-53115 Bonn.
#

tors, or P1-receptors (P1R), and the purine and pyrimidine nucleotide receptors or P2-receptors (P2R) [1, 2]. The growing family of known P2 receptors comprises subfamilies of ionotropic (P2X) and metabotropic (P2Y) receptors. While P2XR subtypes appear to be activated exclusively by adenine nucleotides, subtypes of P2YR exist, including P2Y2, P2Y4, P2Y6, at which uracil nucleotides, namely UTP or UDP, show activity and may act as physiological agonists. The P2Y2R is a ubiquitously expressed UTP-sensitive P2YR subtype, also known as P2UR [3]. ATP and UTP are equipotent as agonists at the P2Y2R. It has recently been found that the impaired chloride transport in the bronchi of patients suffering from cystic brosis can be bypassed by stimulation of a P2Y2R to activate an alternative chloride channel [4]. UTP is currently under development as a

810

Figure 1. Structures of standard nucleotides.

811

Figure 2. Preparation of 5-substituted UTP derivatives.

novel therapeutic agent for the symptomatic treatment of cystic brosis. Only few P2Y2R agonists are known so far and structure-activity relationships for P2Y2R agonists are virtually unknown [5, 6]. In the present study we synthesized a series of 5-substituted UTP derivatives. We selected UTP as a lead structure to develop P2Y2R ligands, rather than ATP, because of the higher selectivity of UTP versus other P2R subtypes, e.g., P2XR. In addition, degradation products of ATP and analogues (adenosine and analogues) may interact with P1R (adenosine receptors), but not those of UTP and analogues [7]. The synthesized nucleotides were investigated in vitro for their potency to increase intracellular calcium concentration ([Ca2+]i) caused by a P2Y2R-mediated stimulation of phospholipase C [8]. 2. Chemistry 5-Alkyl-substituted UTP derivatives 10ae were synthesized according to published procedures with modications (gure 2) [912]. Uracil derivatives 6ae were silylated using hexamethyldisilazane (HMDS) in the presence of trimethylsilylchloride, or ammonium sulfate,

respectively. Derivatives with longer side chains dissolved faster in HMDS due to their higher lipophilicity as compared to those with smaller substituents. After complete dissolution, which indicated completion of the reaction, silylated uracil derivatives 7ae were characterized by their 1H and 13C NMR spectra (table I). Condensation with 1-O-acetyl-2,3,5-tri-O-benzoyl--D-ribofuranose in the presence of tin(IV) chloride yielded nucleosides 8ae, essentially as described [9]. For 5-ethyl derivative 7b, as an example, different reaction conditions were investigated. As solvents, 1,2-dichloroethane or acetonitrile were used. As Lewis acid catalysts tin(IV) chloride, or trimethylsilyltriate, were applied. All different combinations gave very good yields of 8b (8090%). In 1,2-dichloroethane as a solvent, traces of N3nucleoside were formed, which could be removed by silica gel column chromatography. Trimethylsilyltriate was found not to be superior to tin(IV) chloride. Hydrolysis was performed using a saturated aqueous sodium hydrogencarbonate solution. In our hands, potassium hydrogencarbonate showed no advantage over the sodium salt, in contrast to results reported by Szemo et al. [9]. Filtration over silica gel efficiently removed tin(IV) oxide hydrate formed during the reaction. The

812
Table I. 7a 7b 7c 7d 7e
13

C-NMR data of silylated uracil derivatives (in CDCl3) (ppm). O-Si(CH3)3 0.32 0.29 0.26 0.29 0.41 C2 168.15 167.84 167.87 167.39 167.26 C4 161.78 161.69 161.66 167.39 161.08 C5 112.57 118.33 116.75 122.30 116.08 C6 158.87 157.90 158.63 158.63 157.83 R5 12.20 13.42, 13.67, 21.63, 13.08, 20.30 22.27, 28.91 26.24 21.54, 25.91, 30.76

Compound

benzoylated nucleosides 8ae were deprotected via transesterication using methanolic sodium methylate solution. The formed benzoic acid methyl ester was conveniently removed by extraction with diethyl ether and subsequent lyophilization to yield the nucleosides 9ae. A recently described direct method for nucleoside synthesis [13], which does not require protection of the sugar, was investigated for the preparation of the 5-propyl derivative 10c, but was not successful in our hands. At low temperatures no reaction occurred and the starting material was recovered. At temperatures over 80 C only degradation was observed. Nucleotides 10ae were prepared from nucleosides 9ae according to published procedures with slight modications [912]. The lyophilized nucleosides were dissolved in dry trimethylphosphate and reacted with phosphorus oxychloride, with or without the addition of 1,8-bis(dimethylamino)naphthalene (proton sponge). The reactive intermediate phosphorodichloridate was coupled with a mixture of one equivalent of tri-nbutylamine and a ve-fold excess of tri-n-butylammonium diphosphate in dimethylformamide to yield 10ae (gure 2). Nucleosides 8ae and 9ae were characterized by their 1 H NMR spectra, which were in accordance with published data [9]. In addition, 13C NMR spectra were recorded (tables II and III). Crystals suitable for X-ray crystallography could be obtained from nucleosides 8b and 8d. The structure of nucleoside 8d is shown in gure 3 (see also table VI).
Table II.
Compound 8a 8b 8c 8d 8e
13

Compound 8d showed a nucleosidic torsion angle (O3C5-N1-C1) of 133.22. Thus, the nucleoside crystallized in the anti-conformation despite bulky 5- and 5substituents. The torsion angle about the exocyclic C8C9 bond (gure 3) may be described as +sc (gauche/ gauche). A vibrational disorder of the isopropyl side chain could be seen, probably facilitated by the low data collection temperature of 173 K. As an interesting feature, the sugar puckering in 8d deviated from the conformation observed in unprotected 5-methyluridine [14], which adopted a C3-endo conformation, while 8d showed a sugar puckering that can be described as C1-exo. The synthesized nucleotides were puried by anion exchange chromatography using diethylaminoethyl(DEAE-) Sephadex A25. The triethylammonium salts were converted to the corresponding sodium salts by dissolution in absolute methanol and subsequent precipitation by the addition of sodium iodide in acetone. The sodium salts were obtained as amorphous powders and therefore better to handle for biological tests than the triethylammonium salts, which were obtained as oily, transparent liquids. Nucleotides were characterized by 13C and 31P NMR spectroscopy and plasma desorption mass spectrometry. Purity was determined by HPLC/UV and 31P NMR spectroscopy (tables III and IV). The combination of both methods proved to be important for purity control (table IV).

C-NMR data of protected nucleosides (in DMSO-d6), (ppm).

C3 70.7 70.6 70.7 70.6 70.7 C2 73.1 73.0 72.9 73.0 73.0 C4 78.8 78.8 78.8 78.8 78.9 C1 88.5 88.8 88.4 89.4 88.6 Caromat. 128.4, 129.3, 128.4, 129.3, 128.4, 133.5, 128.4, 129.3, 128.4, 129.3, 128.5, 133.5, 128.6, 133.5, 128.5, 133.6, 128.6, 133.5, 128.5, 133.5, 128.7, 133.7 128.7, 133.8 128.6, 133.7 128.7, 133.8 128.7, 133.7 129.2, 129.2, 129.3, 129.2, 129.2, benzoyl 164.4, 164.6, 165.5 164.5, 164.6, 165.5 164.5, 164.6, 165.5 164.5, 164.6, 165.5 164.6, 164.7, 165.5 C5 C6 C2 C4 R5 110.1 136.9 150.9 162.9 11.9 115.7 136.7 150.2 163.2 12.9, 19.5 114.1 137.0 150.2 163.2 13.3, 21.1, 28.1 119.9 136.1 150.0 162.9 21.0, 21.1, 25.5 114.4 137.1 150.3 163.3 13.6, 21.7, 25.9, 30.2

C5 63.7 63.6 63.6 63.7 63.8

813
Table III.
Compound 9a 9b 9c 9d 9e 10a 10b 10c 10d 10e
13

C-NMR data of nucleosides and nucleotides (in DMSO-d6), (ppm).

C5 60.9 60.8 60.8 60.7 60.8 66.6 66.7 66.8 67.0 66.9 C3 69.8 69.9 69.8 69.9 69.8 74.9 74.6 64.8 75.0 74.8 C2 73.4 73.5 73.4 73.8 73.6 71.5 71.6 71.6 71.8 71.6 C4 84.8 84.8 84.7 84.7 84.8 85.0 85.1 85.1 85.3 85.1 C1 87.6 87.7 87.8 88.1 87.8 89.2 89.0 89.2 89.4 89.3 C5 109.2 115.1 113.2 119.4 113.5 113.2 118.9 117.2 123.5 117.7 C6 136.4 135.8 136.4 134.9 136.4 138.7 138.2 139.0 138.7 138.9 C2 150.8 150.7 150.6 150.4 150.7 153.4 153.2 153.3 153.2 153.8 C4 163.8 163.4 163.3 163.0 163.5 167.8 167.1 167.3 166.9 167.5 R5 12.2 12.9, 12.9, 21.2, 12.9, 14.2 12.1, 14.4, 22.2, 14.9, 19.7 21.0, 28.4 21.4, 25.4 21.7, 25.9, 30.1 21.2 22.8, 29.6 22.5, 21.2 23.1, 27.5, 31.9 9.2 9.2 9.2 9.2 9.2 6.1 4.9 6.1 6.1 6.1
3

JC4-P

JC5-P

Figure 3. X-ray structure of nucleoside 8d.

814
Table IV.
Compound 10a 10b 10c 10d 10e
31

P-NMR (in D2O), MS data and HPLC retention times of nucleotides 10ae, (ppm).
31

pH 5.9 6.5 6.7 7.7 5.8

P 10.69 d J = 19.6 Hz 10.26 d J = 19.6 Hz 10.38 d J = 19.6 Hz 10.58 d J = 20.0 10.98 d J = 20.2 Hz

P-NMR P 22.56 t J = 19.5 21.90 t J = 19.6 22.30 t J = 19.5 22.03 t J = 20.2 22.56 t J = 20.0

P 10.04 d J = 19.5 Hz 9.26 d J = 19.5 Hz 9.05 d J = 19.5 Hz 7.14 d J = 19.8 Hz 10.04 d J = 19.8 Hz

Mole Peak 496.9 511.1 524.9 524.8 539.1

MSa Ion [M H]

System Ab Retention time 13.51 21.88 25.97 17.36 31.15

HPLC (UV; =266 nm) System Bc (Purity) Retention time (94 %) (95 %) (93 %) (94 %) (92 %) 7.56 7.82 8.01 9.40 8.07

(Purity) (95 %) (96 %) (94 %) (92 %) (93 %)

Hz Hz Hz Hz Hz

[M H] [M H] [M H] [M H]

Negative ion plasma desorption MS. b System A: Nucleosil RP-18 column with eluent A, 0.1 M triethylammonium acetate buffer and eluent B, acetonitrile (gradient: 015%, 30 min). c System B: Macherey-Nagel ET125/4 Nucleosil 4000-7 PEI column with eluent A, 0.01 M Tris/HCl (pH = 8.4) and eluent B, Tris/HCl (pH = 8.4), 1.0 M NaCl.

Phosphorylation reactions were investigated in some detail. It had been reported that reaction of nucleosides with phosphorus oxychloride was somewhat faster in trimethyl phosphate as compared to triethyl phosphate [15]. We investigated the reaction of ethyl derivative 9b with phosphorus oxychloride under various conditions. Reaction time was always virtually identical in both solvents. Proton sponge accelerated reaction time from 7 to 2 h, while pre-heating of reagents before the addition of phosphorus oxychloride had no effect. Phosphorylation of 9ae to obtain triphosphates 10ae yielded a number of side products (gure 4). For butyl derivative 10e, structure elucidation of the major side products was performed as preparation of 10e yielded the largest quantity of side products. A fraction isolated at a concentration of 0.130.16 M triethyl ammonium bicarbonate (TEAB) buffer showed ve spots in thin layer chromatography on cellulose-coated plates, all of which contained phosphorus as shown by spraying with a phosphate-specic reagent. Four of the spots showed UV absorption at 266 nm (table V). The mass spectra (anionic mode) conrmed triphosphate 10e as the major product. Several pH-dependent signals were detected in the 31P NMR spectrum. According to Cozzone and Jardetzky [16], the signals could be assigned to the 2monophosphate (11), the 3-monophosphate (12), the 5-monophosphate (13), and the 2,3-cyclomonophosphate (14, gure 5). The identication of the side products is shown in table V. A fraction isolated at 0.43 M TEAB buffer concentration contained two more side products and a small amount of 10e. The major side product showed a mass of 600.7. The 31P NMR spectrum in dimethylsulfoxide showed four signals, a triplet at 23.87 ppm (P, JPP =

21.8 Hz, JPP = 22.9 Hz), a doublet at 11.93 ppm (P, JPP = 21.8 Hz), a doublet at 11.39 ppm (P, JPP = 22.9 Hz), and a singlet at 16.41 ppm. These data indicate that the compound (15) contains a 2,3-cyclophosphate group in addition to the 5-triphosphate residue of 10e. The second side product with a molecular mass of 619.1 is a ring-open analogue of 15, since the 31P NMR spectrum clearly shows a signal in the range of a monophosphate in addition to the 5-triphosphate structure. Bisphosphorylation at the 2- and 3-hydroxyl groups was not observed. The ratio of formed 10e:15 was ca. 2:1. The particularly large amount of side products observed in the preparation of the 5-butyluridine triphosphate 10e may be due to steric interaction between the bulky butyl residue and the 5-hydroxyl function in polar aprotic solvents, which leads to an increased phosphorylation of the secondary hydroxyl groups at C2 and C3. The preparation of other nucleotides yielded the analogous side products, but in smaller quantities. Use of proton sponge as reaction accelerator was not the cause for the formation of 2- and 3-phosphoric acid esters and 2,3-cyclophosphates as side products, as we observed the formation of the same compounds in the absence of proton sponge. In addition to nucleotide side products, large amounts of inorganic phosphates, including cyclometatriphosphate (P3O93), phosphate (PO43), pyrophosphate (P2O74) and linear triphosphate (P3O105) were obtained (table V). The stable cyclometatriphosphate was particularly difficult to remove by standard chromatographic methods due to partial coelution with the target UTP derivatives (gure 4 and table V).

Table V. Analysis of isolated fractions obtained from ion exchange chromatography of crude 10e (compare gure 3).
Fraction TEAB concentration [%]a Yield [mg] (of 1 820 mg crude product)b UV maximum at [nm] 31 P-NMR (in D2O) (pH value) I 0 810 218c 7.3 (s); 2.4 (s); 3.0 (s); 3.1 (s) (pH = 7.2) II 1316 241 III 3136 250 IV 4144 67 V 7080 114 266 4.8 (s); 4.2 (s); 2.2(s); 7.7 10.3 (m); 20.0 (s); 21.7 (s) d (pH = 7.4) [540, 619], 700f 4 spots 5 spots VI 80 33 266 21.8 (s); 4.5. (s); 8.1 (s)d; 10.3 (s)d; 21.7 (s)d (pH=8.0) 760f 3 spots 5 spots

MS (neg. Mode) UV detection at 266 nm Phosphate detection (FeCl3/5-sulfosalicylic acid reagent) Inorganic phosphate as determined by 31PNMR Nucleotidesg and other organic compounds

n.d.e 3 spots 3 spots

Spectroscopic methods 266 266 266 (s); 2.8 (s); 2.3 (s); 9.8 (d); 11.0 (d); 21.8 (s); 5.2 (d); 0.4 (s); 0.0 (s) 20.8 (s); 22.4 (t) 9.8 (d); (pH = 4.5) (pH = 6.2) 20.8 (t) 19.2 (s); 4.0 (s); 3.7 (pH = 8.0) (s); 3.5 (s); 2.7 (s) (pH = 9.5) 379f 538f 601f, [619] Cellulose-TLC (solvent: 2-propanol:NH3(25 %):H2O = 6:3:1) 4 spots 1 spots 2 spots 5 spots 2 spots Deduced structure P3O93 10e 15 2 spots

PO43 P2O7
4

PO43 11 12 13 14

P3O93, P3O105

PO43,

PO43, P3O105 structure could not be determinedi

phosphates of 1,8bis-(dimethyl)ammoniumnaphthalene

structure could not be determinedh

100% corresponds to 1 M TEAB buffer. b Unidentied: 305 mg. c Typical spectrum for aromatic compounds. d Signal not resolved. e n.d. = not determined. f Main peak. g For structures see gures 2 and 5. h Structure contains a total of 5 phosphate groups but no 23-cyclic phosphate groups. i Structure contains a total of 6 phosphate groups (one of them is a 23-cyclic phosphate group).

815

816

Figure 4. Separation of crude 10e on DEAE-Sephadex A25 by FPLC.

3. Biological evaluation P2Y2 receptor activity was determined in a mouse neuroblastoma glioma hybrid cell line (NG 108-15) by uorimetric measurement of the increase in intracellular calcium concentration caused by stimulation of P2Y2 receptors via activation of phospholipase C. Selected compounds were additionally tested in a human cystic brosis epithelial airway cell line (CF/T43) and a human basal epithelial airway cell line (BEA). 4. Results and discussion 4.1. Neuroblastoma glioma cells (NG108-15 cell line) The putative physiological P2Y2 receptor agonists UTP (1) and ATP (4), the ATP analogue ZTP (5), uridine diphosphate UDP (2) and the corresponding monophosphate UMP (3) were investigated at NG108-15 cells for comparison. Initially, single concentrations (50 and/or 500 M) of UTP derivatives and standard nucleotides were used (table VII). Before testing it was shown that not even at high concentrations (500 M) did the inor-

ganic phosphates pentasodium triphosphate (Na5P3O10), trisodium tricyclometaphosphate (Na3P3O9), and sodium pyrophosphate (Na4P2O7), which could be present as side products in some of the nucleotides, show any effects. As previously shown, UTP acted as a potent P2Y2 receptor agonist, exhibiting an EC50 value of 1.25 M and a maximal increase in intracellular calcium of 69% (table VIII). The observed potency for UTP is in accordance with published data [5, 17]. ATP was somewhat less potent, exhibiting an EC50 value of 17.7 M, but showed about the same maximal stimulation as UTP. The lower activity of ATP may be due to faster enzymatic degradation of ATP as compared to UTP under test conditions [18, 19]. The ring-open, base-modied ATP analogue ZTP (5) was found to be nearly equipotent to ATP in our assay system. In the system described, addition of UDP (2) also resulted in an increase in intracellular calcium concentration, with an EC50 value of 3.2 M (table VIII). Efficacy, however, was only 61% of that of UTP. UDP has previously been shown to be inactive at P2Y2R [19, 20], but it might have been contaminated with UTP, or could have been phosphorylated enzymatically, and the result-

817
Table VI. Crystal data of compound 8d. Empirical formula Formula weight Temperature Wavelength Crystal system Space group Unit cell dimensions C33H30N2O9 598.59 173 (2) K MoK, 71.073 pm orthorhombic C2221 a = 1 482.37 (7) pm = 90 deg. b = 2 322.16 (14) pm = 90 deg. c = 1 780.29 (13) pm = 90 deg. 6.1283 (6) nm3 8 25 1.298 Mg/m3 0.095 mm-1 2 512 0.3 0.2 0.2 mm 2.2925.05 deg. 17 h 17, 27 k 27, 21 l 21 20 046 5 425 [R(int) = 0.0561] 3 / 0% 99.6 % N/a Full-matrix least-squares on F2 mixeda 5 425 / 203 / 436 0.912 R1 = 0.0392, wR2 = 0.0835 R1 = 0.0638, wR2 = 0.0913 0.5(9) 0.0017(2) 157 and 201 e.nm3 ENRAF-NONIUS CAD4 SHELXS-93 [28], SHELXL-97 [29]
2 1/2

Volume Z Reections used for cell renement Density (calculated) Absorption coefficient F(000) Crystal size Theta range for data collection Index ranges Reections collected Independent reections No. of standard reections / decay Completeness to 2Theta = 25.05 Absorption correction Renement method Treatment of hydrogen atoms Data / restraints / parameters Goodness-of-t on F^2 Final R indices [I > 2sigma(I)]b R indices (all data)b Absolute structure parameter Extinction coefficient Largest diff. peak and hole Diffractometer: Program:
a

See Experimental. b Denition of R values: 2 2 2 2 R1 = R Fo||Fc /R| and wR2 = Rw Fo Fc /Rw Fo

; w = 1/ r Fo + 0.0494 P }; P = Max 0, Fo + 2 Fc /3.

2 2 2 2 2

ing UTP may have been responsible for the effects measured. Uridine monophosphate (UMP, 3) exhibited only very weak activity at high concentrations of 500 M, which might, again, be due to contamination by UTP and/or enzymatic conversion to UTP [18, 19]. 5-Alkyl-substituted UTP derivatives showed lower activity at P2Y2 receptors than UTP. Activity decreased with increasing size of the 5-substituent exhibiting the following rank order of potency: UTP > 5-methyl-UTP > 5-ethyl-UTP > 5-isopropyl-UTP > 5-propyl-UTP > 5-butyl-UTP (table VII). A recorded dose-response curve for 5-ethyl-UTP (10b) showed this compound to be a full agonist at P2Y2R of NG108-15 cells exhibiting an EC50 value of 99 M (table VIII).

4.2. CF/T43 and basal epithelial airway (BEA) cells Selected compounds were further tested in CF/T43 cells [4], a human epithelial airway cell line containing the defective chloride channel (CFTR), that causes cystic brosis, and compared with a basal epithelial airway (BEA) cell line. Results from measurements of P2Y2Rmediated increase [Ca2+]i were virtually identical in CF/T43 and BEA cells (tables IX and X). Activity of nucleotides and maximal increase in [Ca2+]i was higher in CF/T43 and BEA cells as compared to NG108-15 cells. ATP was found to be nearly equipotent to UTP in CFT/43 and BEA cells (table IX), while ZTP was somewhat less potent (table X). 5-Methyl-UTP (10a) showed a dose-

818

Figure 5. Structures of identied side products in the phosphorylation of 9e (preparation of 10e).

dependent increase in [Ca2+]i and appeared to be nearly equipotent to ATP and UTP (table X), while 5-ethyl-UTP (10b, 10100 nM) only slightly increased [Ca2+]i in BEA cells (table X). The higher potency of nucleotides in epithelial airway cells as compared to NG108-15 cells may reect species

differences (human, mouse). Differences in enzyme pattern and enzymatic activity of nucleotidases and phosphatases may also contribute to the differences measured [18, 19, 21]. A further factor, which has to be taken into account, may be differences in techniques used for the measurements. Thus, NG108-15 cells were used in

Table VII. Effects of nucleotides on P2Y2 receptor-mediated increase in intracellular calcium concentration in NG108-15 cells, results from testing of single concentrations of compounds. Compounda 1 2 3 4 5 10a 10b 10c 10d 10e
a

Percent increase in [Ca2+]i SEM (number (n) of independent experiments) 50 M concentration of test compound 500 M concentration of test compound UTP UDP UMP ATP ZTP 5-Methyl-UTP 5-Ethyl-UTP 5-Propyl-UTP 5-Isopropyl-UTP 5-Butyl-UTP 69.3 8.2 (n = 10) 42.4 4.5 (n = 4) n.d.c 44.6 5.9 (n = 3) 54.3 13.5 (n = 2) n.d. 52.0 29.7 (n = 2) at 100 M 28.9 7.8 (n = 3) 0 (n = 2) n.d. n.d.
b

(179.0 26.9)b (n = 5) 31.3 (n = 1) 10.6 1.3 (n = 2) 67.2 17.9 (n = 2) n.d. n.d. 71.7 38.6 46.2 25.0 12.0 (n = 3) 5.3 (n = 3) 25.0 (n = 2) 0.7 (n = 3)
c

For structures see gures 1 and 2. determined.

High values are due to lysis of cells observed at a concentration of 500 M of UTP.

n.d. = not

819
Table VIII. Effects of nucleotides on P2Y2 receptor-mediated increase in intracellular calcium concentration in NG108-15 cells, results from dose-response curves. Compound 1 2 4 5 10b
a

EC50 [M]a (95 % condence intervals) 1.25 (0.1114) 3.21 (0.2836) 17.7 (5.656) 17.1b 99 (20480)

Maximal effect in % of max. UTP effect (= 100 %) (number (n) of independent experiments) 100 61 95 78 104 12 (n = 10) 7 (n = 4) 12 (n = 3) 19 (n = 2) 17 (n = 3)

UTP UDP ATP ZTP 5-Ethyl-UTP

Results from three independent experiments unless otherwise noted. b Single dose-effect curve.

suspension while CF/T43 and BEA cells were used as attached cell layers in culture dishes (see Experimental protocol). It has been shown that stressing of cells, e.g,. by agitation, may result in a massive release of endogenous nucleotides [22], a fact which clearly affects the test results obtained with exogenously applied nucleotides. Such variations in EC50 values for nucleotides, e.g., UTP and ATP, at P2Y2R in different test systems have been described [5]. In the present study, degradation
Table IX. Effects of standard nucleotides on P2Y2 receptormediated increase in intracellular calcium concentration in CF/T43 cells and basal epithelial airway (BEA) cells, results from doseresponse curves. Compound 1 UTP 4 ATP Cell line CF/T43cells BEA cells CF/T43cells BEA cells EC50 [M] (95 % condence intervals)a 0.10 (0.010.93) 0.16 (0.0055.6) 0.17 (0.012.9) 0.20b

of compounds was not investigated, though enzymatic and chemical hydrolysis undoubtedly play a role [21]. Calcium measurements were, however, performed very fast, within seconds up to a few minutes, and therefore, only moderate degradation will have occurred. It had been shown that in airway epithelial cells, where strong UTP degradation is observed, after 5 min, about 70% of intact UTP is still present [21]. 4.3. Structure-activity relationships The P2Y2R is unique in that it is stimulated by purine (ATP) as well as pyrimidine nucleotides (UTP). We conrmed previous ndings [5, 20, 22, 23] that UTP (1) and ATP (4) activate P2Y2R in low micromolar to submicromolar concentrations (EC50 of UTP: 1.25 M at NG108-15 cells, 0.10 M at CF/T43 cells, and 0.16 M at BEA cells; EC50 of ATP: ca. 18 M at NG108-15 cells, and ca. 0.2 M at CF/T43 and BEA cells, see tables VIII and IX). Efficacy was similar for both physiological agonists. A ring-open analogue of ATP, ZTP (5) exhibited similar potency as ATP indicating that purine base modication was possible without loss of activity. Truncation

a b

Determined in 34 separate experiments unless otherwise noted. Two separate experiments.

Table X. Effects of nucleotides on P2Y2 receptor-mediated increase in intracellular calcium concentration in CF/T43 cells and BEA cells, results from testing of single concentrations of compounds Compound Cell line Percent increase in [Ca2+]i SEM (number of independent experiments) 10 nM concentration of test compound 108 80 (n = 3) 110 49 (n = 3) 103 20 (n = 3) 199 94 (n = 3) 36 29 (n = 2) 80 35 (n = 2) 155 46 (n = 2) 113 32 (n = 2) 39 11 (n = 2) 17 24 (n = 2)

100 nM concentration of test compound 197 58 (n = 3) 173 113 (n = 3) 266 94 (n = 3) 208 42 (n = 3) 117 55 (n = 2) 138 22 (n = 2) 269 24 (n = 2) 312 15 (n = 2) 43 26 (n = 2) 41 7 (n = 2)

1 UTP 4 ATP 5 ZTP 10a 5-Methyl-UTP 10b 5-Ethyl-UTP

CF/T43cells BEA cells CF/T43cells BEA cells CF/T43cells BEA cells CF/T43cells BEA cells CF/T43cells BEA cells

820 of the triphosphate chain in UTP to UDP (2) and further to UMP (3), led to a decrease or loss in activity as reported [5, 20, 22], suggesting that the four negative charges are important for electrostatic interactions with the receptor protein. Only few synthetic UTP analogues have previously been investigated as P2Y2R ligands [5, 6, 21, 22]. Exchange of the bridging --oxygen atom in the triphosphate chain by NH, CH2, or CF2, respectively, in order to enhance stability towards nuceotidases, have led to a decrease in activity [23]. The only substitution tolerated in the phosphate chain was the replacement of a -oxygen atom by sulfur, as in UTPS [21]. The compound was nearly as potent as UTP itself. A series of UTP derivatives in which the oxygen in the 4-position of the uracil moiety was replaced by various substituents, including (alkyl)amino, (alkyl)thio or alkoxy, had been investigated, but all of the derivatives were less potent than UTP [6]. A 5-substituted UTP derivative that had been investigated was 5-bromo-UTP. It was found to be less potent than UTP or ATP at P2Y2R [22]. The present study shows that alkyl substituents in the 5-position of UTP are not tolerated by the receptors either. Introduction of an ethyl group in the 5-position of UTP, for example, decreased activity at P2Y2R of NG108-15 cells ca. 80-fold (from an EC50 for UTP of 1.25 M to 99 M for 5-ethyl-UTP 10b). With increasing volume of the 5-substitutent P2Y2 activity was decreased (table VII). Since both polar 5-substituents, such as bromo, and non-polar, especially bulky alkyl substituents, led to decreased activity, it can be concluded, that steric reasons, i.e., lacking bulk tolerance of the receptor in that area, are responsible for this effect. 5. Conclusion Methods and conditions for the preparation and purication of 5-substituted UTP derivatives were investigated and optimized. New spectroscopic data on UTP derivatives are presented, including 13C and 31P NMR data. 5-Substituted UTP derivatives were found to be full agonists at P2Y2 receptors of NG108-15 cells, and basal epithelial airway cells without or with defective CFTR channel (CF/T43 cell line). Potency of 5-substituted UTP derivatives decreased with increasing volume of the 5-substitutent. It is planned to investigate the potency of 5-substituted UTP derivatives at other P2Y receptor subtypes, such as P2Y4 or P2Y6. The presented SAR of UTP derivatives may contribute to the design of selective ligands for subtypes of the uracil nucleotide-sensitive P2Y receptors. 6. Experimental protocols 6.1. Chemistry Melting points were determined with a Bchi 530 apparatus and are uncorrected. Nuclear magnetic resonance spectra were determined using a Bruker AC-200 spectrometer (1H: 200 MHz, 13C: 50.3 MHz). Chemical shifts are given in ppm downeld from tetramethylsilane as internal standard. 31P-NMR spectra were recorded on a Bruker AMX 400 spectrometer at 161 MHz with H3PO4 as an internal standard. UV spectra were recorded with a Perkin Elmer Lambda 12 spectrometer. HPLC was performed using the following equipment: Pharmacia 2249 gradient pump, 2141 variable wavelength monitor set to 266 nm, and a 2221 integrator. Purity controls were performed on two systems: system A: Nucleosil RP-18 column with eluent (A) 0.1 M triethylammonium acetate buffer, and eluent (B) acetonitrile/H2O (gradient: 015%, 30 min). System B: Macherey-Nagel ET125/4 Nucleosil 4000-7 PEI column with eluent (A) 0.01 M Tris/HCl (pH 8.4) and eluent (B) 0.01 M Tris/HCl (pH 8.4), 1.0 M NaCl. Negative ion plasma desorption mass spectra were obtained on Applied Biosystem BIO-ION 20 plasma desorption mass spectrometer, using 252Cf as source of ssion fragments. Nucleotides were puried on a Pharmacia High Load system using a triethyl ammonium hydrogen carbonate buffer gradient (080%). Thin layer chromatography was performed on silica gel F254 (Merck), and in some cases on cellulose TLC plates (Merck). The following solvent systems were used for development of the TLC: (S1) dichloromethane:methanol = 9:1 for compounds 8ae; (S2) dichloromethane:methanol = 3:1 for compounds 9ae; (S3) 2-propanol:NH4OH:water = 6:3:1 for compounds 10ae, 1115. Phosphate spraying agent (0.1% FeCl3 x 6 H2O; 7% 5-sulfosalicylic acid in 75% ethanol) according to Wade and Morgan [24] was used for identication of organic and inorganic phosphates. RF-values of phosphates were compared to data published by Ebel [25]. 6.1.1. General procedure for silylation A suspension of 5 mmol of uracil derivatives 6ae in 20 mL of 1,1,1,3,3,3-hexamethyldisilazane (HMDS) and 1 mL of trimethylsilyl chloride, or a few crystals of (NH4)2SO4 respectively, was reuxed until a clear solution was obtained. The solution was allowed to cool, and excess HMDS was removed under reduced pressure. The silylated uracil derivatives 7ae were kept under nitrogen. A small sample of the oily sirup was dissolved in

821 CDCl3 and checked for complete silylation by 1H NMR spectroscopy. Yields ranged from 9698%. Selected 1H-NMR data in CDCl3, (ppm): 7d: 0.30 (s, 9H, C-O-Si(CH3)3); 0.32 (s, 9H, C-OSi(CH3)3); 1.14 (d, 6H, C5-CH-(CH3)2, J = 6.9Hz); 2.85 (sept., 1H, C5-CH-(CH3)2, J = 6.9Hz); 7.99 (s, 1H, C6-H). 6.1.2. General procedure for the synthesis of nucleosides Silylated uracil derivatives 7ae (5 mmol) under nitrogen were dissolved in 5 mL of 1,2-dichloroethane, and 1.30 g (5.5 mmol) of SnCl4 (10% excess) was added with vigorous stirring. 1-O-Acetyl-2,3,5-tri-O-benzoyl--Dribofuranose (2.2 g, 0.4 mmol) in 40 mL of 1,2dichloroethane was added dropwise to the slightly yellowish solution. The mixture was stirred as indicated below, completion of reaction was controlled by TLC (dichloromethane:methanol = 9:1). Reaction times were between 45 h. The reaction solution was poured into a saturated aqueous NaHCO3 solution (ca. 50 mL), or on humidied solid NaHCO3 with successive adding of saturated NaHCO3 solution, respectively, under vigorous stirring, and then allowed to stand overnight. The suspension was ltered over silica gel and the gel was washed twice with 50 mL of dichloromethane and twice with 100 mL of ethyl acetate. The organic phase was separated, dried (Na2SO4), ltered over silica gel and evaporated to dryness. The glassy residue was crystallized from methanol, or ethanol respectively, to yield 8ae. If necessary, the removal of unreacted sugar and side products was achieved by column chromatography on silica gel using dichloromethane:methanol = 9:1 as eluent. Yields ranged from 8187%. Selected 1H-NMR data in DMSO-d6, (ppm): 8b: 0.93 (t, 3H, C5-CH2-CH3, J = 7.3 Hz); 2.08 (q, 2H, CH3, J = 7.3 Hz); 4.644.77 (m, 3H, C4H, C5H2); 5.96 (m, 2 H, C2H, C3H); 6.21 (d, 1H, C1H, J12 = 3.8 Hz); 7.398.04 (m, 16H, C6-H, Haromat.); 11.48 (br s, 1H, N1-H, exchangeable). 8c: 0.76 (t, 3H, propyl CH3, J = 7.1 Hz); 1.33 (sext., 2H, propyl CH3CH2-, J = 7.3Hz); 2.05 (t, 2H, propyl CH3CH2CH2-, J = 7.2 Hz); 4.614.78 (m, 3H, C4H, C5H2); 5.90 (m, 2H, C2-H, C3-H); 6.23 (d, 1H, C1-H, J = 3.7 Hz); 7.388.06 (m, 16H, C6H, Haromat); 11.48 (br s, 1H, N3-H, exchangeable). 8d: 1.00 (d, 6H, CH(CH3)2, J = 6.9Hz); 2.85 (sept., 1H, CH(CH3)2, J = 6.6Hz); 4.654.75 (m, 3H, C4H, C5H2); 6.00 (d, 2H, C3H, C2H); 6.25 (d, 1H, C1H, J = 5.1Hz); 7.01 (d, 1H, C6H, J = 0.9Hz); 7.388.05 (m, 16H, C6H, Haromat); 11.47 (br s, 1H, N3-H). 6.1.3. Deprotection of nucleosides Tribenzoyl nucleoside (8ae, 0.92 mmol) was dissolved in a mixture of 20 mL of absolute methanol and 2.76 mL of 5% methanolic sodium methylate solution. The solution was stirred for 58 h and completion of the reaction was determined by TLC (dichloromethane: methanol = 3:1). Neutralisation of the solution was achieved by adding DOWEX-50 WX-8 ion exchange resin washed previously with methanol. After ltering the resin off, 20 mL of water was added and methanol was evaporated. The benzoic acid methyl ester was extracted with diethyl ether and the water fraction was lyophilized to yield the nucleosides 9ae in 9498% yield. Selected 1H-NMR data in DMSO-d6, (ppm): 5-n-Propyluridine (9c): 0.84 (t, 3H, C5-CH2-CH2-CH3, J = 7.3Hz); 1.42 (sext., 2H, C5-CH2-CH2-CH3, J = 7.5 Hz); 2.14 (dt, C5-CH2-CH2-CH3, J12 = 7.5 Hz); 3.493.68 (m, 2H, C5H2); 3.82 (q, 1H, C4H, J = 3.46/2.94Hz); 3.964.03 (m, 2H, C2H, C3H); 5.06, 5.33 (br s, 3H, OH (exchangeable)); 5.77 (d, 1H, C1H, J = 5.3 Hz); 7.73 (s, 1H, C6H); 11.24 (br s, 1H, N3-H (exchangeable)). 6.1.4. Phosphorylation Synthesis of nucleoside 5-triphosphate synthesis was adapted from literature procedures [10, 26, 27]. 6.1.5. Preparation of tri-n-butylammonium diphosphate Sodium diphosphate decahydrate (6.69 g; 15 mmol) was dissolved in 150 mL of water (twice distilled). Excess of Dowex ion exchange resin 50 8, 2050 mesh, proton form, prewashed several times with water, was added to the solution of sodium diphosphate, and the mixture was gently stirred for 60 min. A mixture of 60 mL of ethanol and 7.14 mL of tributylamine in a ask was placed in an ice-water bath, and the diphosphate solution was ltered directly into the ask. The resin was repeatedly washed with water until the ltrate was no longer acidic. The solvent was then evaporated under vacuum at 40 C, yielding a thick, nearly colourless syrup. This residue was treated twice with 100 mL of ethanol and then evaporated. The residue was taken up in 40 mL of dimethylformamide (DMF, anhydrous grade, from Fluka) and evaporated again. This residue was taken up in 30 mL of anhydrous DMF, yielding 30 mL of 0.5 M tri-n-butylammonium diphosphate in DMF. The solution was stored sealed and cooled at 4 C until used. 6.1.6. Preparation of triethylammoniumhydrogencarbonate (TEAB) buffer A 1 M solution of TEAB was prepared by bubbling CO2 through a 1 M triethylamine solution in water at 04 C for several hours (pH approx. 7.47.6).

822 6.1.7. Preparation of nucleotides 10ae Lyophilized nucleoside (9ae, 1 mmol) was dissolved in 5 mL of trimethyl phosphate (dried over 10 molecular sieve). The mixture was stirred at room temperature under nitrogen and then cooled to 4 C. Dry 1,8bis(dimethylamino)naphtalene (proton sponge, 0.32 g, 1.5 mmol) was added, followed by 1.3 mmol (0.20 g) of POCl3, 5 min later. After several hours of stirring at 04 C, tri-n-butylamine (0.1 mL, 0.72 mmol) was added to the solution followed by 10 mL (5 mmol) of 0.5 M tri-n-butylammonium diphosphate solution in DMF. After 25 min the mixture was poured into 0.5 M cold aqueous TEAB solution (30 mL, pH 7.5) and stirred at 04 C for several minutes. The solutions were allowed to reach room temperature with stirring and were then left standing for 1 h. Trimethylphosphate was extracted with t-butylmethyl ether and the aqueous solutions were evaporated and lyophilized to yield glassy colourless oils. The reactions were controlled by TLC using freshly prepared solvent system S3. TLC plates were dried before UV absorption was detected and the plates were subsequently sprayed with phosphate reagent. 6.1.8. Purication of nucleotides The crude nucleoside 5-triphosphates were puried by ion exchange column chromatography using DEAESephadex A25 (Pharmacia) HCO3-form swelled in 1 M NH4HCO3 solution at 4 C. After equilibrating the column with deionized water, the crude product was dissolved in 5 mM aqueous triethylammonium carbonate buffer. The column was washed with deionized water, followed by a solvent gradient of 0800 mM TEAB in approximatly 3 000 mL of solvent to elute the triphosphates. Fractions were collected and appropriate fractions pooled, evaporated, diluted in water and lyophilized to yield a product with a clear glass appearance. Purity was assessed by HPLC/UV and 31P NMR spectroscopy. Isolated yields: 33% (10a), 24% (10b), 37% (10c), 40% (10dd), 26% (10e). 6.1.9. Preparation of sodium salts Nucleotides (50 mmol) were dissolved in absolute methanol (dried over Mg) and evaporated. This procedure was repeated twice. The solid was then dissolved in 5 mL of absolute methanol with stirring. A 1.5-fold excess of a 1 M sodium iodide solution in acetone was added dropwise. The solution was then diluted with absolute acetone. The formed precipitate was ltered off and washed several times with absolute acetone. The white solid was dried under high vacuum. 6.2. Determination of X-ray structure Single pure crystals of 8d were obtained by slow crystallization from ethanol at room temperature over several weeks. Low temperature X-ray intensity data were collected on an Enraf-Nonius CAD4 diffractometer using MoK radiation. The structure was resolved with direct methods using SHELXS-97 [28] and rened by full-matrix least squares iteration against F2, using SHELXL-97 [29]. All non-hydrogen atoms were rened anisotropically. The coordinates of hydrogen atoms bonded to C5, C6, C7, and C8 were freely rened using isotropic displacement parameters. For all other hydrogen atoms a riding model was employed in the renement with their Uiso constrained to equal 1.2 times the Ueq of the parent atoms (1.5 times Ueq in the case of the hydrogen atoms of the methyl groups). Crystallographic data were deposited at the Cambridge Crystallographic Data Centre as supplementary publication (deposition number 120084). Free copies of data are available from: CCDC; 12 Union Road, Cambridge CB21EZ (fax: + 44 1223 336 033; e-mail: deposit@ccdc.cam.ac.uk). 6.3. Biological studies ATP, UTP, UDP, UMP, and ZTP were purchased from Sigma Chemical Co. in the highest available purity grade. Nucleotides were checked for autouorescence prior to testing. 6.3.1. NG108-15 cell culture NG108-15 cells were kindly provided by Berit Frjh, Dept. of Medical Neurochemistry, Lund University Hospital, Sweden. NG108-15 cell line (hybrid cell line generated by fusion of mouse neuroblastoma and C6 glioma cells) was cultured in tissue culture asks (75 or 150 cm2). Cells were grown adherent in tissue culture asks in a 5% CO2 atmosphere at 37 C until conuent. The growth medium consisted of a mixture of 500 mL of Dulbeccos modied Eagle medium (DMEM) supplemented with 4.5 g/L of glucose and pyruvate, 55 mL of foetal calf serum, 10 mL of HAT (hypoxanthine, aminopterin, thymidine) supplement (Gibco article No. 15140-114) and 5.5 mL of penicillin/ streptomycin (Gibco article no. 21060-017). Medium was changed nearly every day depending on cell density as determined by visual inspection. 6.3.2. Preparation of fura2-loaded cells for measurements NG10815 cells were examined under a microscope and cells were used when conuent. The medium was removed from the ask, growth medium (10 mL at 37 C) was added and the cells were dispersed into a single cell

823 suspension by gentle, brief, repeated pipetting. The cells were centrifuged and the pellet was resuspended in Krebs-Ringer-Hepes (KRH) buffer (conc. in mM: NaCl 125; KCl 5.0; MgSO4 1.2; KH2PO4 1.2; CaCl2 2.0; glucose 6.0 Hepes 25; pH adjusted to 7.4, with NaOH) in a concentration of 106 cells per mL. The cells were incubated with 2 M of fura-2/AM for 45 min at 37 C in the dark, spun down (at 1 000 g) and washed twice with the same amount of KRH buffer as used before. 6.3.3. Measurement of [Ca2+]i using NG108-15 cells in suspension [Ca2+]i was determined as described by Lin et al. [30] measuring uorescence of fura-2 loaded cells (2 mL of cell suspension containing 106 cells per experiment, 37 C) using a Hitachi Model F-2000 dual wavelength uorescence spectrophotometer. Test compound (10 L) was added. The additon of P2Y2R agonists resulted in a transient increase in intracellular calcium concentration. After baseline was reached again, 10 L of Triton X 100 (10% aqueous solution) was added, followed by the addition of 10 L of a saturated solution of EGTA and 10 L of a saturated solution of Tris base. [Ca2+]i was calculated as described [31]. 6.3.4. CF/T43 and BEA cell cultures Airway epithelial cell lines CF/T43 and BEA (kindly provided by Dr J.R. Yankaskas), which express an endogenous P2Y2R [4] were routinely grown in 75 cm2 tissue culture asks (Costar) using keratinocyte growth medium (KGM; Clonetics, catalog No. CC-3115) at 37 C in a humidied 4% CO2 atmosphere. Medium was changed 23 times a week depending on cell density, and cells were passaged every 710 d. For passaging, cells were washed twice with phosphate-buffered saline (PBS) and then treated with 0.1% trypsin/EDTA for 510 min to detach the cells. Soybean trypsin inhibitor (three-fold excess) was used to neutralize trypsin prior to passaging. 6.3.5. Measurement of intracellular calcium in CF/T43 and BEA cell monolayers Cell culture dishes (20 cm2) with a conuent monolayer of CF/T43 or BEA cells were washed twice with 2 mL of Krebs-Ringer-Hepes (see above). After the addition of fura-2/AM (2 M), cells were incubated at 37 C in the dark for 1 h. The medium was changed to 4 mL of fresh KRH buffer and cells were washed 3 times before being incubated for another 15 min to remove non-hydrolysed dye. Immediately before measurement, the experimental incubation medium was replaced again (2 mL) and the culture dishes were placed in the spectrouorometer (photon counter; Curt Lindmark Innovation AB, Sweden). After base line stabilization, the test compound was rapidly added with careful shaking, culture dishes were re-inserted within 10 s, and changes in uorescence were registered until base line levels were reached again (310 min). Then, Ca2+ -ionophore ionomycin (10 M) and subsequently MnCl2 (20 mM) were added in order to determine maximum and minimum values of fura-2 uorescence. Calculations of [Ca2+]i were made according to Grynkiewicz et al. [31].

Acknowledgements We thank Gaby Huhurez for skillful technical assistance. C.E. Mller is grateful for support given by the Fonds der Chemischen Industrie and the Industrie- und Handelskammer Wrzburg-Schweinfurt. This project was supported by grants from the Deutscher Akademischer Austauschdienst (DAAD) to B.K. and C.E.M. and from the Svenska Institutet to E.H. References
[1] [2] [3] [4] Fredholm B., Abbracchio M.P., Burnstock G. et al., Pharm. Rev. 46 (1994) 143156. Bhagwat S.S., Williams M., Eur. J. Med. Chem. 32 (1997) 183193. Abbracchio M.P., Burnstock G., Pharmac. Ther. 64 (1994) 445475. Parr C.E., Sullivan D.M., Paradiso A.M., Lazarowski E.R., Burch L., Olsen J.C., Erb L., Weisman G.A., Boucher R.C., Turner J.T., Proc. Natl. Acad Sci. USA 91 (1994) 32753279. Heilbronn E., Knoblauch B.H.A., Mller C.E., Neurochem. Res. 22 (1997) 10411050. Shaver S.R., Pendergast W., Siddiqi S.H., Yerxa B.R., Croom D.K., Dougherty R.W., James M.D., Jones A.N., Rideout J.L., Nucleosides Nucleotides 16 (1997) 10991102. Mller C.E., Stein B., Curr. Pharm. Design. 2 (1996) 501530. Jrlebark L., Erlandsson M., Uri A., King B.F., Ziganshin A.U., Johansson C., Heilbronn E., Biochem. Biophys. Res. Comm. 229 (1996) 363369 . Szem A., Szabocs A., Sgi J., tvs L.J., Carbohydr. Nucleosides Nucleotides 7 (1980) 365379. Zimmet J., Jrlebark L., Hammarberg T. et al., Nucleosides. Nucleotides 12 (1993) 120. Ludwig L., Acta. Biochim. Biophys. Acad. Sci. Hung. 16 (1981) 131133. Mishra N.C., Broom A.D., J. Chem. Soc. Chem. Commun. (1991) 12761277. Bennua-Skalmowski B., Krolikiewicz K., Vorbrggen H., Tetrahedron. Lett. 36 (1995) 78457848. Hunt D.J., Subramanian E., Acta. Crystallogr. B25 (1969) 21442152. Ikemoto T., Haze A., Hatano H., Kitamoto Y., Ishida M., Nara K., Chem. Pharm. Bull. 43 (1995) 210215. Cozzone P.J., Jardetzky O., Biochemistry 15 (1976) 48534859. Conigrave A.D., Jiang L., Cell. Calcium 17 (1995) 111119. Harden T.K., Lazarowski E.R., Boucher R.C., Trends. Pharmacol. Sci. 18 (1997) 4346.

[5] [6]

[7] [8]

[9] [10] [11] [12] [13] [14] [15] [16] [17] [18]

824
[19] [20] [21] [22] [23] [24] Nicholas R.A., Watt W.C., Lazarowski E.R., Qing L., Harden T.K., Mol. Pharmacol. 50 (1996) 224229. Erb L., Lustig K.D., Sullivan D.M., Turner J.T., Weisman G.A., Proc. Natl. Acad. Sci. USA 90 (1993) 1044910453. Lazarowski E.R., Watt W.C., Stutts M.J., Stutts M.J., Brown H.A., Boucher R.C., Harden T.K., Br. J. Pharmacol. 117 (1996) 203209. Lazarowski E.R., Watt W.C., Stutts M.J., Boucher R.C., Harden T.K., Br. J. Pharmacol. 116 (1995) 16191627. Pendergast W., Siddiqi S.H., Rideout J.L., James M.K., Dougherty R.W., Drug. Dev. Res. 37 (1996) 133 (abstract). Wade H.E., Morgan D.M., Biochem. J. 60 (1955) 264270. [25] [26] [27] [28] [29] [30] [31] Ebel J.P., Bull. Soc. Franc. (1953) 10891095. Moffat J.G., Can. J. Chem. 42 (1964) 599604. Kovacs T., tvs L., Tetrahedron. Lett. 29 (1988) 45254528. Sheldrick G.M., Acta. Crystallogr. A46 (1990) 467. Sheldrick G.M., Program for crystal structure renement, University of Gttingen, 1997. Lin T.A., Lustig K., Sportiello M.G., Weisman G., Sun G.Y., J. Neurochemistry 60 (1993) 11151125. Grynkiewicz G., Poenie M., Tsien R.Y., J. Biol. Chem. 260 (1985) 34403450.

Eur. J. Med. Chem. 34 (1999) 825835 1999 Editions scientiques et mdicales Elsevier SAS. All rights reserved

825

Original article

Investigation of SAR requirements of SR 142801 through an indexed combinatorial library in solution

Luca F. Ravegliaa*, Mauro Vitalic, Marco Articoa, Davide Graziania, Douglas W.P. Hayb, Mark A. Luttmannb, Renzo Menaa, Giorgio Pifferic, Giuseppe A.M. Giardinaa
b a Department of Medicinal Chemistry, SmithKline Beecham S.p.A., Via Zambeletti, 20021 Baranzate, Milano, Italy Department of Pulmonary Pharmacology, SmithKline Beecham Pharmaceuticals, 709 Swedeland Road, King Of Prussia, PA, 194106, USA c Istituto di Chimica Farmaceutica, V.le Abruzzi 42, 20133 Milano, Italy

(Received 3 May 1999; accepted 25 May 1999)

Abstract To rapidly gain information on structure-activity relationship (SAR) requirements of the human neurokinin 3 (hNK-3) receptor antagonist SR 142801, an indexed combinatorial library was synthesised in solution and screened on the hNK-3 receptor. SAR considerations drawn from binding affinity of combinatorial mixtures were conrmed through the synthesis and biological evaluation of some individual compounds. 1999 Editions scientiques et mdicales Elsevier SAS tachykinins / neurokinin 3 receptor antagonists / combinatorial chemistry / indexed libraries / structure-activity relationships

1. Introduction The neurokinin 3 (NK-3) is one of the three receptors for the family of peptides named tachykinins or neurokinins and belongs to the G protein-coupled receptor superfamily [1]. In 1996, our group described the identication of potent and selective non-peptide NK-3 receptor antagonists featuring the 2-phenyl-4-quinolinecarboxamide framework, including SB 223412 [2, 3]; this is one of the three main chemical classes of nonpeptide NK-3 receptor antagonists reported to date [4, 5]. The other two are the peptide-derived structures reported by Parke-Davis (PD 161182) [6] and the 3,4-dichlorophenylpiperidines which were rst described by Sano (exemplied by SR 142801, gure 1) [7] and subsequently by Merck Sharp and Dohme [8]. Due to our involvement in the NK-3 area, we were interested in investigating the SAR requirements of the 3,4-dichlorophenylpiperidine NK-3 receptor antagonists, only marginally described in the literature [8]. For this reason, the
*Correspondence and reprints

Figure 1. Structure of SR 142801.

synthesis of a small library of analogues of SR 142801 was planned via a combinatorial chemistry approach. Since the 3-(3,4-dichlorophenyl)-3-propylpiperidine framework A (gure 1) of SR 142801 appears to be a key feature of this class of NK-3 receptor antagonists, able to drive their selectivity for the NK-3 receptor with respect to the NK-1 and NK-2 [4, 8], we decided to keep this

826

Figure 2. Library design.

moiety xed and focus our attention on variations in the basic head X and the acylating group Y. In designing the library, 7 amines (X1X7) and 7 acylating groups (Y1Y7) were selected, on the basis of commercial availability of corresponding reagents, to give a total of 49 compounds (X17AY17), as all the possible combinations of X and Y substituents (gure 2). To simplify the synthetic pathway, it was planned to work on racemic mixtures. Solution phase combinatorial chemistry was the technique of choice for the synthesis of the library because a suitable point of attachment to the resin could not be found, ruling out a solid phase approach. Moreover, since in the synthetic scheme the combinatorial reaction was a well consolidated amide coupling at the last step (gures 4 and 5), and since the number of compounds to synthesise was small (49), we reasoned that pools of compounds could easily be obtained by using mixtures of reactants, without risk of getting too many by-products. In particular, to help the structural determination of active components of the library, the strategy of indexed (or orthogonal) combinatorial libraries was utilised [911]. 2. Chemistry 2.1. Synthesis of intermediate compounds In gure 3, the synthesis of intermediate compounds is described. Racemic aminol 1 was obtained according to

literature procedure [12], protected at the nitrogen and activated to nucleophilic substitution by converting the hydroxy group into the mesylate and this, in turn, into the more reactive iodo derivative 2 [13]. Displacement of iodine with the 7 amines X1X7 (CH3CN, K2CO3, 90 C, 18 h) afforded the desired intermediate compounds X1AX7A, which were puried by ash column chromatography. 2.2. Synthesis of the library: set 1 As depicted in gure 4, each intermediate compound was subsequently deprotected (20% TFA/CH2Cl2, r.t., 18 h) and reacted with a stoichiometric and equimolar mixture of the 7 acylating agents Y1Y7, (CH2Cl2, K2CO3, r.t., 18 h). Simple washing of the reaction mixtures with H2O and evaporation to dryness afforded the rst set (set 1) of desired mixtures (or sublibraries) X1AY17X7AY17 in satisfactory to high yields (6080%) [14a] and high purity (9598%, LC/UV) [14b]. Peak identity was assessed by LC/MS and all expected molecular weights were detected in each mixture. It is worth noting that the seven sublibraries of set 1 contain all the desired 49 compounds; each sublibrary has the X substitution xed and different from one another, while containing all possible Y substituents. Since the Y portion is common to all the sublibraries, binding affinity data on this set will clarify the SAR related to the X group.

827

Figure 3. Synthesis of intermediates.

2.3. Synthesis of the library: set 2 In gure 5, the synthesis of the second set (set 2) of sublibraries is described. An equimolar amount of each of the intermediate compounds X1AX7A (gure 3) was mixed into the same reaction vessel and Boc-deprotected; the mixture was then split into seven equal portions (X17A), and each portion was reacted with a stoichiometric amount of acyl chloride/phenylisocyanate/ phenylsulfonyl chloride Y1Y7 (CH2Cl2, K2CO3, r.t., 18 h). Again, simple washing of the reaction mixtures

with H2O and evaporation to dryness produced the desired sublibraries X17AY1X17AY7 in high yields (7495%) [14a] and high purity (9098%, LC/UV, except for X17AY4: 83%) [14b]. All expected molecular weights were detected in each mixture by LC/MS. The seven sublibraries of set 2 contain the same 49 compounds of set 1, but in a different combination: in this case, the Y substituent is xed and peculiar to any sublibrary, while all possible X substitutions are common to all. Binding affinity data on this second set will clarify

828

Figure 4. Synthesis of the library: set 1.

the SAR related to the Y group. By combining information coming from the two sets of sublibraries, the best X (set 1) and the best Y (set 2) substituents should be identied, thus allowing structural determination of the most active compounds of the whole library. 2.4. Synthesis of individual compounds To prove the validity of the technique and the predictions made from binding affinity data of mixtures, the six individual compounds reported in table II were synthesised, deprotecting the appropriate intermediate com-

pound (20% TFA/CH2Cl2, r.t., 18 h) and reacting it with the appropriate acylating agent (CH2Cl2, K2CO3, r.t., 18 h). 3. Pharmacology Receptor binding assays were performed with crude membranes from CHO cells expressing the hNK-3 receptor as detailed previously [3, 15]. For NK-3 receptor competition binding studies, [125I]-[MePhe7]-NKB binding to hNK-3-CHO membranes was performed using the

829

Figure 5. Synthesis of the library: set 2.

procedure of Sadowski and co-workers [16]. Specic binding was determined by subtracting total binding from non-specic binding, which was assessed as the binding in the presence of 0.5 M cold [MePhe7]-NKB. Percent inhibition of specic binding was determined for each concentration of the compounds and the IC50, dened as the concentration required to inhibit 50% of the specic binding, obtained from concentration-response curves. Values reported in tables I and II are the apparent inhibition constant (Ki), which was calculated from the IC50 as described by Cheng and Prusoff [17].

4. Results and Discussion Human NK-3 receptor binding affinity data for the 14 sublibraries synthesised are reported in table I. Results of sublibraries of set 1 revealed that pyrrolidinocarbonylmethylpiperazine (X7, Ki = 113 nM) should be slightly more active than phenylpiperidine (X1, Ki = 160 nM) which, in turn, should be more active than phenylpiperazine (X2, Ki = 232 nM). A strongly reduced affinity is forseen for isopropylpiperazine (X3, Ki = 743 nM) and morpholine (X4, Ki = 1235 nM).

830
Table I. Binding affinities of library compounds.

Set 1 X Y Y1-7 hNK-3 Ki (nM) a,b 160 X X1-7 Y Y1

Set 2 hNK-3 Ki (nM)a,b CH3CO 9163

Y1-7

232

X1-7

Y2

PhCO

242

Y1-7

743

X1-7

Y3

PhCH2CO

394

Y1-7

1235

X1-7

Y4

(CH3)2C=CHCO

503

Y1-7

525

X1-7

Y5

2,4-(MeO)2PhCO

1835

Y1-7

440

X1-7

Y6

PhNHCO

1497

Y1-7

113

X1-7

Y7

PhSO2

570

Inhibition of [125I]MePhe7-NKB binding in hNK-3-CHO cell membranes. bsingle determination (n = 1).

As far as the Y substitution is concerned (set 2), benzoyl (Y2, Ki = 242 nM), also present in SR 142801, appears to be the best substituent. Increasing the distance of the phenyl ring from the carbonyl with a methylene spacer (Y3, Ki = 394 nM) slightly decreases the binding affinity, while substitution at the phenyl ring with two methoxy groups (Y5, Ki = 1 835 nM) or replacement of the phenyl with a methyl (Y1, Ki = 9 163 nM) resulted in a marked loss of activity. Dimethylacryloyl (Y4, Ki = 503 nM) and phenylsulfonyl (Y7, Ki = 570 nM), although less potent than the benzoyl (Y2, Ki = 242 nM), maintain a certain activity, while the urea (Y6, Ki = 1497 nM) does

not appear to be a good replacement for the benzamide. To test the reliability of the predictions made from mixtures, six individual compounds (some predicted to be the most active and some predicted to have medium activity) were synthesised and screened (table II). The rank order of potency of individual compounds was found to be in good agreement with the prediction, X7AY2 being the most active (hNK-3 binding affinity, Ki = 35.4 10.4 nM) and X1AY5 the least active (hNK-3 binding affinity, Ki = 513 99.5 nM) amongst individual compounds prepared. It is worth noting that the best individual compound synthesised, X7AY2, appears to

831
Table II. Binding affinities of individual compounds.

piperidine NK-3 receptor antagonists with an improved biological prole compared to SR 142801 [19]. 5. Conclusions In conclusion, an indexed combinatorial library of analogues of SR 142801 was synthesised in solution and screened for its hNK-3 receptor binding affinity. Synthesis and biological evaluation of some individual compounds proved the reliability of this very simple, accessible and rather under-utilised technique and allowed us to gain information on SAR requirements of 3,4dichlorophenylpiperidine NK-3 receptor antagonists, signicantly decreasing the number of reactions and samples to test. 6. Experimental protocols 6.1. Chemistry

Individual compounds X Y PhCO hNK-3 Ki (nM)a,b 46.2 9.2

2,4(MeO)2PhCO PhCO

513 99.5

50.9 9.2

PhCO

35.4 10.4

PhCH2CO

289 68.0

PhSO2

175 50.2

PhCO

SR 142801c,d

1.2 0.3

Inhibition of [125I]MePhe7-NKB binding in hNK-3-CHO cell membranes. bmean SEM for three determinations (n = 3). cin house data. dSR 142801 is a single enantiomer [18].

Melting points were determined with a Bchi 530 hot stage apparatus and are uncorrected. Proton NMR spectra were recorded on a Bruker ARX 300 spectrometer at 303 K unless otherwise indicated. Chemical shifts were recorded in parts per million () downeld from tetramethylsilane (TMS); NMR spectral data are reported as a list. IR spectra were recorded in Nujol mull or neat on sodium chloride disks or in KBr with a Perkin-Elmer 1420 spectrophotometer; mass spectra were obtained on a Finnigan MAT TSQ-700 spectrometer. Silica gel used for ash column chromatography was Kiesegel 60 (230400 mesh) (E. Merck AG, Darmstadt, Germany). All evaporations were performed at reduced pressure. Combustion elemental analyses were performed by Redox s.n.c., Milan, Italy and analyses indicated by the symbols of the elements were within 0.4% of the theoretical values. All reagents utilised in gures 35 are commercially available compounds and were used without further purication. Racemic aminol 1 and SR 142801 were synthesised according to Giardina et al. [12] and to Chen et al. [13]. 6.2. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3(3-iodopropyl)piperidine 2 Racemic aminol 1 [12] (4.7 g, 16.3 mmol) and di-tertbutyl dicarbonate (4.3 g, 19.6 mmol) were dissolved, under nitrogen atmosphere, in dry CH2Cl2 (75 mL). The reaction mixture was stirred at room temperature for 8 h and left standing overnight. The solvent was evaporated to dryness and the crude material was puried by ash column chromatography, eluting with a mixture of

have an hNK-3 binding affinity about 30 times lower than SR 142801 (hNK-3 binding affinity, Ki = 1.2 0.3 nM) [18]. Since X7AY2 differs from SR 142801 only for the basic head X, this SAR study outlines the importance of the X substitution in conferring high binding affinity to the 3,4-dichlorophenylpiperidine NK-3 receptor antagonists. Additional work is needed to understand the requirements of this portion of the molecule and X7 could be a good starting point for further chemical modications in order to obtain novel 3,4-dichlorophenyl-

832 CH2Cl2/MeOH 98:2, to obtain 5.4 g (85%) of N-tertbutoxycarbonyl-3-(3,4-dichlorophenyl)-3-(3-hydroxypropyl)piperidine as a pale yellow oil. This compound (5.4 g, 13.9 mmol) was dissolved in CH2Cl2 (50 mL); triethylamine (TEA) (2.2 mL, 15.8 mmol) was added and the solution was cooled to 0 C. Methanesulfonyl chloride (1.2 mL, 15.4 mmol), dissolved in CH2Cl2 (7 mL), was added dropwise and the reaction mixture was allowed to reach room temperature and stirred for additional 2 h. The reaction was quenched with water and the extracted organic layer was washed with 10% citric acid, 5% NaHCO3 and brine, dried over MgSO4, ltered and evaporated to dryness to yield 6.0 g (93%) of N-tertbutoxycarbonyl-3-(3,4-dichlorophenyl)-3-(3-methanesulfonyloxypropyl)piperidine as a pale yellow oil. This compound (6.0 g, 12.9 mmol) was dissolved in acetone (120 mL) and KI (3.5 g, 20.9 mmol) was added to the solution which was reuxed for 16 h. Insoluble material was ltered off and the ltrate was evaporated to dryness to yield 6.4 g (100%) of the title compound as a red oil, which was used in the following reactions without further purication. IR (neat) 2 940, 2 885, 1 692, 1 470, 1 428 cm1. 1H-NMR (CDCl3) 7.44 (d, 1H), 7.38 (d, 1H), 7.17 (dd, 1H), 3.91 (d, 1H), 3.583.47 (m, 1H), 3.31 (d, 1H), 3.313.20 (m, 1H), 3.01 (t, 2H), 2.071.98 (m, 1H), 1.801.40 (m, 7H), 1.49 (s, 9H). 6.3. General procedure for the synthesis of intermediate compounds X1AX7A K2CO3 (0.303 g, 2.2 mmol) and the appropriate amine (1.1 eq.) were added to a solution of N-tertbutoxycarbonyl-3-(3,4-dichlorophenyl)-3-(3-iodopropyl) piperidine 2 (0.99 g, 1.98 mmol) in CH3CN (7 mL). The reaction mixture was heated to 80 C under magnetic stirring for 18 h; then it was ltered and the ltrate was evaporated to dryness. The crude solid obtained was dissolved in EtOAc and washed with H2O. The organic layer was dried over MgSO4, ltered and evaporated to dryness to yield the title compounds which were puried by ash column chromatography. Yields after chromatography were in the range: 7596%. Spectroscopic data for compounds X1AX7A are reported below. 6.3.1. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3(4-phenylpiperidin-1-yl)propyl]piperidine X1A IR (Nujol) 2 926, 1 690, 1 674, 1 604, 1 554 cm1. 1 H-NMR (CDCl3343 K) 7.47 (d, 1H), 7.37 (d, 1H), 7.307.14 (m, 6H), 4.00 (d, 1H), 3.56 (dt, 1H), 3.26 (d, 1H), 3.263.18 (m, 1H), 2.902.83 (m, 2H), 2.502.40 (m, 1H), 2.21 (t, 2H), 2.091.91 (m, 3H), 1.801.13 (m, 11H), 1.49 (s, 9H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 531 (MH+). 6.3.2. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3(4-phenylpiperazin-1-yl)propyl]piperidine X2A IR (KBr) 2 942, 2 816, 1 692, 1 602, 1 554, 1 502 cm1. 1H-NMR (CDCl3) 7.44 (d, 1H), 7.38 (d, 1H), 7.24 (dd, 2H), 7.18 (d br, 1H), 6.90 (d, 2H), 6.84 (dd, 1H), 4.05 (m br, 1H), 3.60 (m, br, 1H), 3.203.10 (m, 6H), 2.43 (m, 4H), 2.21 (t, 2H), 2.05 (m br, 1H), 1.751.10 (m, 7H), 1.46 (s, 9H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 532 (MH+), 554 (MNa+). 6.3.3. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3(4-isopropylpiperazin-1-yl)propyl]piperidine X3A IR (neat) 2 942, 2 812, 1 690, 1 554, 1 470 cm1. 1 H-NMR (CDCl3) 7.42 (d, 1H), 7.31 (d, 1H), 7.15 (d br, 1H), 4.05 (m br, 1H), 3.60 (m br, 1H), 3.13 (d, 1H), 3.12 (m, 1H), 2.64 (m, 1H), 2.50 (m, 4H), 2.35 (m, 4H), 2.20 (t, 2H), 2.05 (m, 1H), 1.701.00 (m, 7H), 1.45 (s, 9H), 1.00 (d, 6H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 498 (MH+). 6.3.4. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3(morpholin-4-yl)propyl]piperidine X4A IR (neat) 2 944, 2 858, 2 810, 1 690, 1 590, 1 554 cm1. 1H-NMR (CDCl3) 7.44 (d, 1H), 7.35 (d, 1H), 7.14 (d br, 1H), 4.02 (m br, 1H), 3.63 (m br, 4H), 3.59 (m br, 1H), 3.13 (d, 1H), 3.12 (m, 1H), 2.30 (m, 4H), 2.18 (t, 2H), 2.05 (m br, 1H), 1.711.00 (m, 7H), 1.46 (s, 9H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 457 (MH+), 479 (MNa+). 6.3.5. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3(1,2,3,4-tetrahydroisoquinolin-2-yl)propyl]piperidine X5A IR (neat) 2 932, 1 692, 1 554 cm1. 1H-NMR (CDCl3343 K) 7.48 (d, 1H), 7.36 (d, 1H), 7.20 (dd, 1H), 7.107.03 (m, 3H), 6.94 (m, 1H), 4.01 (d, 1H), 3.58 (dt, 1H), 3.47 (s, 2H), 3.25 (d, 1H), 3.20 (ddd, 1H), 2.83 (t, 2H), 2.59 (t, 2H), 2.34 (t, 2H), 2.092.00 (m, 1H), 1.781.13 (m, 7H), 1.49 (s, 9H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 503 (MH+).

833 6.3.6. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3(1-phenyl-1,3,8-triazaspiro[4,5]decan-4-on-8-yl)propyl] piperidine X6A IR (neat) 2 938, 2 856, 1 708, 1 690, 1 602, 1 502 cm1. 1H-NMR (CDCl3343 K) 7.46 (d, 1H), 7.36 (d, 1H), 7.29 (dd, 2H), 7.19 (dd, 1H), 6.996.89 (m, 3H), 5.61 (s br, 1H), 4.70 (s, 2H), 3.90 (d br, 1H), 3.56 (m, 1H), 3.293.19 (m, 2H), 2.752.42 (m, 6H), 2.23 (m, 2H), 2.05 (m, 1H), 1.781.05 (m, 9H), 1.47 (s, 9H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 601 (MH+), 545. 6.3.7. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-{3[4-(pyrrolidinocarbonylmethyl)piperazin-1-yl]propyl} piperidine X7A IR (neat) 2 944, 2 874, 2 812, 1 690, 1 642, 1 554 cm1. 1H-NMR (CDCl3343 K) 7.44 (d, 1H), 7.36 (d, 1H), 7.18 (dd, 1H), 3.90 (d, 1H), 3.55 (dt, 1H), 3.543.43 (m, 4H), 3.23 (d, 1H), 3.20 (ddd, 1H), 3.09 (s, 2H), 2.55 (m, 4H), 2.36 (m, 4H), 2.19 (t, 2H), 2.081.99 (m, 1H), 1.981.80 (m, 4H), 1.781.00 (m, 7H), 1.49 (s, 9H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 567 (MH+). 6.4. General procedure for the synthesis of library (set 1) compounds X1AY1-7X7AY17 6.4.1 Synthesis of sublibrary X1AY17 The synthesis of sublibrary X1AY17 is reported as an example of the synthesis of all sublibraries of the rst set of the library. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl]piperidine X1A (0.301 g, 0.55 mmol) was dissolved in dry CH2Cl2 (8 mL); TFA (2 mL) was added and the reaction mixture was stirred at room temperature overnight. After evaporation to dryness, the residue was taken up with CH2Cl2, washed with 5% NaHCO3, dried over MgSO4, ltered and evaporated to dryness, to obtain 3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl]piperidine. This compound was dissolved in dry CH2Cl2 (10 mL), and K2CO3 (165.3 mg, 1.2 mmol) was added; then a solution of acetyl chloride (564 L, 7.9 mmol), benzoyl chloride (918 L, 7.9 mmol), phenylacetylchloride (1 047 L, 7.9 mmol), dimethylacryloyl chloride (880 L, 7.9 mmol), 2,4-dimethoxybenzoyl chloride (1 585 mg, 7.9 mmol), phenylisocianate (859 L, 7.9 mmol) and phenylsulfonyl chloride (1 013 L, 7.9 mmol), brought to 100 mL with dry CH2Cl2, was prepared and 1 mL of this solution was added to the reaction mixture, which was stirred at room temperature overnight. The reaction was quenched with H2O (5 mL) and the extracted organic layer was washed with H2O, dried over MgSO4, ltered and evaporated to dryness to obtain 0.233 g (78%) of the title sublibrary. All seven compounds forming the sublibrary were characterised by LC/MS. Purity of the seven sublibraries of set 1, calculated as the sum of the areas of the seven peaks of the LC/UV chromatogram attributed by LC/MS to the seven compounds forming each sublibrary, is reported below: X1AY17 98%; X2AY17 98%; X3AY17 95%; X4AY17 98%; X5AY17 97%; X6AY17 98%; X7AY17 97%. 6.5. General procedure for the synthesis of library (set 2) compounds X17AY1X17AY7 6.5.1. Synthesis of sublibrary X17AY1 The synthesis of sublibrary X17AY1 is reported as an example of the synthesis of all sublibraries of the second set of the library. N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl] piperidine X1A (116 mg, 0.217 mmol), N-tertbutoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperazin-1-yl)propyl]piperidine X2A (116 mg, 0.217 mmol), N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(4-isopropylpiperazin-1-yl)propyl]piperidine X3A (108 mg, 0.217 mmol), N-tert-butoxycarbonyl3-(3,4-dichlorophenyl)-3-[3-(morpholin-4-yl)propyl] piperidine X4A (100 mg, 0.217 mmol), N-tertbutoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(1,2,3,4tetrahydroisoquinolin-2-yl)propyl]piperidine X5A (110 mg, 0.217 mmol), N-tert-butoxycarbonyl-3-(3,4dichlorophenyl)-3-[3-(1-phenyl-1,3,8-triazaspiro[4,5] decan-4-on-8-yl)propyl]piperidine X6A (131 mg, 0.217 mmol) and N-tert-butoxycarbonyl-3-(3,4-dichlorophenyl)-3-{3-[4-(pyrrolidinocarbonylmethyl)piperazin1-yl]propyl}piperidine X7A (124 mg, 0.217 mmol) were dissolved, under magnetic stirring, in CH2Cl2 (36 mL); TFA (9 mL) was added and the reaction mixture was stirred at room temperature overnight. After evaporation to dryness, the crude material was taken up with CH2Cl2 and washed with 5% NaHCO3, dried over MgSO4, ltered and evaporated to dryness to yield 0.826 g of a mixture of Boc-deprotected X1AX7A. This mixture was split into seven equal portions (one for each sublibrary) and K2CO3 (65 mg, 0.47 mmol) and acetyl chloride (15.49 L, 0.217 mmol) were added to one of these portions, dissolved in CH2Cl2 (15 mL). After stirring overnight, the reaction was quenched with H2O (5 mL), the organic layer was separated, washed with H2O, dried over MgSO4, ltered and evaporated to dryness to yield 75 mg (74%) of the title sublibrary. All seven compounds forming the sublibrary were characterised by LC/MS.

834 Purity of the seven sublibraries of set 2, calculated as the sum of the areas of the seven peaks of the LC/UV chromatogram attributed by LC/MS to the seven compounds forming each sublibrary, is reported below: X17AY1 98%; X17AY2 94%; X17AY3 98%; X17AY4 83%; X17AY5 98%; X17AY6 94%; X17AY7 98%. 6.6. Synthesis of individual compounds reported in table II 6.6.1. Synthesis of N-benzoyl-3-(3,4-dichlorophenyl)-3[3-(4-phenylpiperidin-1-yl)propyl]piperidine hydrochloride X1AY2 The synthesis of this compound is reported as a general procedure for the synthesis of all six individual compounds prepared. N-tert-Butoxycarbonyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl] piperidine X1A (0.23 g, 0.43 mmol) was dissolved in CH2Cl2 (8 mL); TFA (2.2 mL), dissolved in CH2Cl2 (2 mL), was added, and the reaction was stirred at room temperature overnight. After evaporation to dryness, the crude material was taken up with CH2Cl2 and washed with 5% NaHCO3, dried over MgSO4, ltered and evaporated to dryness to yield crude 3-(3,4dichlorophenyl)-3-[3-(4-phenylpiperidin-1-yl)propyl] piperidine. Half of this crude material ( 0.21 mmol) was dissolved in CH2Cl2 (5 mL), K2CO3 (80 mg, 0.58 mmol) and benzoyl chloride (24.4 L, 0.21 mmol) were added and the reaction was stirred at room temperature overnight and then quenched with H2O (2 mL). The organic layer was separated, washed with H2O, dried over MgSO4, ltered and evaporated to dryness to yield 140 mg (61%) of N-benzoyl-3-(3,4-dichlorophenyl)-3[3-(4-phenylpiperidin-1-yl)propyl]piperidine as a yellow oil. This compound was transformed into its hydrochloride by dissolving in MeOH and treating with HCl/Et2O. Evaporation to dryness and trituration with Et2O afforded the title compound as a solid. IR (KBr) 3 425, 2 937, 2 532, 1 622, 1 444 cm1. 1H-NMR (CDCl3333 K) 7.407.10 (m, 13H), 4.25 (m, 2H), 3.583.30 (m, 4H), 2.902.81 (m, 2H), 2.45 (m, 1H), 2.22 (t, 2H), 2.202.09 (m, 1H), 2.001.55 (m, 9H), 1.351.10 (m, 2H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 535 (MH+). Anal. C32H37Cl3N2O (C, H, N, Cl). 6.6.2. N-(2,4-dimethoxy)benzoyl-3-(3,4-dichlorophenyl)3-[3-(4-phenylpiperidin-1-yl)propyl]piperidine hydrochloride X1AY5 IR (KBr) 3 447, 2 939, 2 668, 1 607, 1 467 cm1. 1 H-NMR (CDCl3333 K) 7.507.20 (m, 9H), 6.506.40 (m, 2H), 4.25 (m, 2H), 3.80 (s, 6H), 3.703.10 (m, 4H), 2.902.81 (m, 2H), 2.45 (m, 1H), 2.20 (m, 2H), 2.152.00 (m, 1H), 2.001.60 (m, 9H), 1.301.10 (m, 2H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 595 (MH+). Anal. C34H41Cl3N2O3 (C, H, N, Cl). 6.6.3. N-benzoyl-3-(3,4-dichlorophenyl)-3-[3-(4-phenylpiperazin-1-yl)propyl]piperidine dihydrochloride X2AY2 IR (KBr) 3 447, 2 947, 2 403, 1 614, 1 440 cm1. 1 H-NMR (CDCl3333 K) 7.427.20 (m, 10H), 6.90 (d, 2H), 6.80 (dd, 1H), 4.25 (m, 1H), 3.53 (d, 1H), 3.533.30 (m, 2H), 3.15 (m, 4H), 2.45 (m, 4H), 2.24 (t, 2H), 2.182.08 (m, 1H), 1.88 (ddd, 1H), 1.791.42 (m, 4H), 1.401.10 (m, 2H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 536 (MH+). Anal. C31H37Cl4N3O (C, H, N, Cl). 6.6.4. N-benzoyl-3-(3,4-dichlorophenyl)-3-{3-[4-(pyrrolidinocarbonylmethyl)piperazin-1-yl]propyl}piperidine dihydrochloride X7AY2 IR (KBr) 3 436, 2 954, 2 437, 1 662, 1 618, 1 438 cm1. 1H-NMR (CDCl3333K) 7.407.10 (m, 8H), 4.40 (m, 1H), 3.523.30 (m, 7H), 3.09 (s, 2H), 2.51 (m, 4H), 2.31 (m, 4H), 2.18 (t, 2H), 2.152.05 (m, 1H), 1.951.80 (m, 5H), 1.701.40 (m, 4H), 1.351.05 (m, 2H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 571 (MH+). Anal. C31H42Cl4N4O2 (C, H, N, Cl). 6.6.5. N-phenylacetyl-3-(3,4-dichlorophenyl)-3-{3-[4(pyrrolidinocarbonylmethyl)piperazin-1-yl]propyl}piperidine dihydrochloride X7AY3 IR (KBr) 3 435, 2 954, 2 546, 1 653, 1 630, 1 451 cm1. 1H-NMR (CDCl3333 K) 7.407.05 (m, 8H), 4.40 (m, 1H), 3.70 (s, 2H), 3.49 (m, 5H), 3.23 (m, 2H), 3.09 (s, 2H), 2.52 (m, 4H), 2.31 (m, 4H), 2.11 (t, 2H), 2.102.00 (m, 1H), 1.981.75 (m, 4H), 1.751.00 (m, 7H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 585 (MH+). Anal. C32H44Cl4N4O2 (C, H, N, Cl). 6.6.6. N-benzenesulfonyl-3-(3,4-dichlorophenyl)-3-{3[4-(pyrrolidinocarbonylmethyl)piperazin-1-yl]propyl} piperidine dihydrochloride X7AY7 IR (KBr) 3 427, 2 943, 2 582, 1 658, 1 469, 1 340, 1 162 cm1. 1H-NMR (CDCl3333 K) 7.78 (m, 2H), 7.55 (m, 3H), 7.40 (m, 2H), 7.21 (m, 1H), 3.47 (m, 5H), 3.10 (m, 1H), 3.10 (s, 2H), 2.79 (m, 2H), 2.51 (m, 4H), 2.32 (m, 4H), 2.17 (t, 2H), 2.001.50 (m, 10H), 1.301.10 (m, 2H). ESI-MS (positive, solvent: methanol, spray 4.5 keV, skimmer: 60 eV, capillary 220 C) m/z 607 (MH+). Anal. C30H42Cl4N4O3S (C, H, N, Cl).

835 Acknowledgements
[8] [9] Harrison T., Korsgaard M.P.G., Swain C.J., Cascieri M.A., Sadowski S., Seabrook G.R., Bioorg. Med. Chem. Lett. 8 (1998) 13431348. Smith P.W., Lai J.Y.Q., Whittington A.R., Cox B., Houston J.G., Stylli C.H., Banks M.N., Tiller P., Bioorg. Med. Chem. Lett. 4 (1994) 28212824. Pirrung M.C., Chen J., J. Am. Chem. Soc. 117 (1995) 12401245. Deprez B., Williard X., Bourel L., Coste H., Hyal F., Tartar A., J. Am. Chem. Soc. 117 (1995) 54055406. Giardina G.A.M., Grugni M., Rigolio R., Vassallo M., Erhard K., Farina C., Bioorg. Med. Chem. Lett. 6 (1996) 23072310. Chen H.G., Chung F.Z., Goel O.P., Johnson D., Kesten S., Knobelsdorf J., Lee H.T., Rubin J.R., Bioorg. Med. Chem. Lett. 7 (1997) 555560. (a) Calculated utilising the average molecular weight. (b) Calculated as the sum of the areas of the seven peaks of the LC/UV chromatogram, attributed by LC/MS to the seven compounds forming each mixture. Sarau H.M., Griswold D.E., Potts W., Foley J.J., Schmidt D.B., Webb E.F. et al., J. Pharmacol. Exp. Ther. 281 (1997) 13031311. Sadowski S., Huang R.R.C., Fong T.M., Marko O., Cascieri M.A., Neuropeptides 24 (1993) 317319. Cheng Y.C., Prusoff W.H., Biochem. Pharmacol. 22 (1973) 30993108. Being SR 142801 a single enantiomer about 2-fold more potent than the corresponding racemate (in-house data), the real difference in binding affinity for the hNK-3 receptor between X7AY2 and () SR 142801 is about 15-fold. In house data demonstrate that SR 142801 has high systemic plasma clearance (36.9 10.6 mL/min/kg) and low oral bioavailability (11 2%) in rats; in addition, it showed signicant interactions with some isoenzymes of P450 family (specically, P2D6 and P3A4).

The authors wish to thank Dr Alberto Cerri for NMR spectroscopic determinations.

[10] [11] [12] [13]

References
[1] [2] Pritchard M.C., Boden P., Drugs Fut. 20 (1995) 11631173. Giardina G.A.M., Sarau H.M., Farina C., Medhurst A.D., Grugni M., Foley J.J., Raveglia L.F. et al., J. Med. Chem. 39 (1996) 22812284. Giardina G.A.M., Sarau H.M., Farina C., Medhurst A.D., Grugni M., Raveglia L.F. et al., J. Med. Chem. 40 (1997) 17941807. Giardina G.A.M., Raveglia L.F., Exp. Opin. Ther. Patents 7 (1997) 307323. von Sprecher A., Gerspacher M., Anderson G., Idrugs 1 (1998) 7391. Boden P., Eden J.M., Hodgson J., Horwell D.C., Pritchard M.C., Raphy J., SumanChauhan N., Bioorg. Med. Chem. Lett. 5 (1995) 17731778. Edmonds-Alt X., Bichon D., Ducoux J.P., Heaulme M., Miloux B., Poncelet M. et al., Life Sci. 56 (1995) PL27PL32.

[14]

[15] [16] [17] [18]

[3] [4] [5] [6]

[19]

[7]

Eur. J. Med. Chem. 34 (1999) 837842 1999 Editions scientiques et mdicales Elsevier SAS. All rights reserved

837

Preliminary communication

Synthesis and antimicrobial activity of heterocyclic ionone-like derivatives

Maria Anzaldia, Enzo Sottofattoria, Rolando Rizzettob, Barbara Granello di Casaletob, Alessandro Balbia*
a

Dipartimento di Scienze Farmaceutiche, Universit degli Studi di Genova, Viale Benedetto XV, 3, 16132 Genova, Italy b Istituto di Igiene e Prevenzione, Universit degli Studi di Genova, Viale Pastore, 1, 16132 Genova, Italy (Received 2 November 1998; revised 2 March 1999; accepted 3 March 1999)

Abstract A number of heterocyclic ionone-like derivatives 5 were prepared with appropriate bifunctional reagents by one-pot cyclisation of 3-dimethylamino-5-(2,6,6-trimethyl-2-cyclohexen-1-yl)-2,4-pentadienal 3a, which was, in turn, obtained from -ionone with N, N-dimethylformamide/phosphorus oxychloride. All compounds 5 possess remarkable activity against the selected Gram-positive, Gramnegative bacterial strains and against Candida albicans. Derivatives 5a2 and 5a6, acting at a high level especially against both Propionibacterium acnes and Staphylococcus aureus, could play a potential role in the treatment of acne and related skin disorders. 1999 Editions scientiques et mdicales Elsevier SAS ionone-like derivatives / short retinoids / antimicrobial activity

1. Introduction Since their introduction nearly two decades ago [1], the use of retinoids for topical and systemic treatment of psoriasis and other hyperkeratotic and parakeratotic skin disorders has increased. They are also used for the treatment of severe acne and acne-related dermatoses. Nonetheless, various side effects, the most serious of which being the teratogenicity, coupled with their long elimination half life, have limited their use [2]. Furthermore, while the recovery of Propionibacterium acnes and other anaerobic bacteria in the skin is markedly reduced, there is an increase in the colonisation of Staphylococcus aureus on the skin and, in general, a signicant rise in the incidence of cutaneous staphylococcal infection in patients treated with oral retinoids [3]. While searching for more biologically active and less toxic compounds, the structure of retinoic acid 1 (gure 1) has been modied both in the cyclohexenyl ring and in the polyene side chain. Major transformations have led to retinoids which barely resemble the original retinoic acid [4, 5]. All these modications, which lead to the three generations of
*Correspondence and reprints

Figure 1. The structure of retinoic acid 1.

retinoids known today (non-aromatic, mono-aromatic and poly-aromatic), are often associated with reduced toxicity while biological activity is maintained or even enhanced. Among these, Etretinate and Acitretin are currently prescribed in systemic use, while Adapalene [4, 5] and Tazarotene [4, 6] have recently been introduced on the market for topical use. Having a method to synthesise new heterocyclic ionone-like derivatives on hand [7] and because of the similarity of these compounds to the so-called short heteroretinoids [811], we focused our biological attention both on the reduction of the P. acnes and the decrease of staphylococcal infections. In fact, retinoids would be of interest for topical use if they were found to possess

838 yield and without side products, conrming the particular behaviour of the enamines 3. 3. Results and discussion Table I summarises the germicidal (GE log values) and killing percentage effects of compounds 5a1a6, 5b and the synthon 3a against two Gram-positive, two Gramnegative bacterial strains and Candida albicans (see Experimental for biological methods). Pyrazole derivatives 5a1a3 exhibited a good to excellent activity against both C. albicans and P. acnes. Introduction of a phenyl ring in the position1 of the pyrazole led to considerable enhancement of the antibacterial activity versus S. aureus, while the presence of the 2,4dinitrophenyl group in the same position decreased the activity against Pseudomonas aeruginosa, Escherichia coli and S. aureus compared to the other two, although the activity against P. acnes and C. albicans remained almost unchanged. The substitution of one nitrogen with oxygen, which led to the isoxazole derivative 5a4, decreased almost all the activity against all the microbes but for P. acnes. Moreover, in this case the activity was only second to phenyl-pyrazole derivative 5a2. Introduction of a pyrimidine in place of a pyrazole or an isoxazole ring did not improve the activity against C. albicans, P. aeruginosa and E. coli. On the contrary, they showed a good to excellent level of potency against S. aureus and P. acnes. From a general point of view it could be stated that the already notable activity of the key intermediate 3a versus C. albicans, S. aureus and P. acnes, is enhanced and specically directed towards one of the three above bacterial strains when a particular heterocyclic moiety is introduced in synthon 3a: namely pyrazole versus C. albicans, phenylpyrazole versus P. acnes, methyltiopyrimidine versus S. aureus. Moreover, lipophilicity, while not increasing selectivity, enhances activity towards all the species. In fact, the more lipophilic 5a2 and 5a6 present the highest activity towards P. acnes and S. aureus, respectively, and also possess a good to excellent activity versus the other species. 4. Conclusion From a structural-activity point of view, certain features can be noted. Compounds 5 could be classied as short-heteroretinoids [811], related to the natural ones through the presence of the cyclohexenyl ring which is a feature of - and -ionones. On the other hand, the potential therapeutic use of these compounds (in partic-

Figure 2. The formation of the classical -chlorovinylaldehydes 2 and the unexpected products 3.

the following two properties: counteracting both the proliferation of P. acnes and of other bacteria (including S. aureus) which are responsible for the majority of infections observed. 2. Chemistry Our laboratory recently began to investigate the synthesis of new short retinoid-like derivatives [7]. In fact, while studying the Vilsmeier reaction (VR) on - and -ionones, we outlined both the formation of the classical -chlorovinylaldehydes 2 and the unexpected products 3 (gure 2), which is unprecedented in a Vilsmeier reaction. Surprisingly, the chlorine of compounds 2 had difficulty in reacting, while the dialkylamino group of 3 acted as an outstanding leaving group, allowing for the formation of new heterocyclic retinoids. In fact, all attempts to cyclise compounds 2 with bifunctional reagents failed. We were only able to isolate the classical formylderivatives as the semicarbazones 4a and 4b. On the other hand, compound 3a reacted rapidly with bifunctional reagents such as hydrazine and guanidine. The reaction, which involved both the formyl and the dimethylamino groups gave a new series of heterocyclic ionone-like derivatives. To complete both the previously reported study and for biological purposes, we have re-synthesised compounds 5a1, 5a3a5, thus improving the yield, and have extended the ring closure on key intermediates 3a and 3b by other bifunctional reagents, therefore obtaining new short heteroretinoids (gure 3). All reagents used reacted quickly, giving the expected compounds in high

839

Figure 3. Synthesis of compounds 4a, 4b, 5a and 5b.

ular 5a2 and 5a6) is closer to the new Adapalene and Tazarotene [46] than classical retinoids. The latter, in fact, may not only cause S. aureus infections when administered orally, but at the same time do not have a signicant effect on aerobic and anaerobic bacteria in vitro [3]. Compounds 5a2 and 5a6, acting both on S. aureus and on P. acnes, could be considered as candidates for topical treatment of acne and/or a coadjuvant in the oral administration of classical retinoids. 5. Experimental protocols 5.1. Chemistry Melting points were determined with Fisher-Johns apparatus and are uncorrected. The IR spectra were recorded in chloroform or in potassium bromide disks on

a Perkin-Elmer 398 spectrometer. The 1H and 13C-NMR spectra were recorded on a Bruker AC 300 (300 MHz, 1 H; 75 MHz, 13C) or a Varian Gemini 200 (200 MHz, 1H; 50 MHz, 13C) spectrometers in deuteriochloroform solutions with tetramethylsilane as the internal standard ( = 0). The purity of all compounds was checked by thinlayer chromatography on silica gel 60-F-254 pre-coated plates and the spots were located in UV light or by vanillin in sulphuric acid. Elemental analyses were performed on a Carlo Erba 1106 Elemental Analyser in the Microanalysis Laboratory in our Institute and the results were within 0.4% of theoretical values. For elemental analysis and spectral data of 5a1, 5a3a5 see [7]. 5.1.1. 5-(2, 6, 6-Trimethyl-2-cyclohexen-1-yl)ethenyl-1Hpyrazole 5a1 A solution of 0.25 g (1 mmol) of 3a and 1 mL of hydrazine hydrate (20 mmol) in 10 mL of ethanol was

840
Table I. Germicidal effect (GE) and killing percentage effect (KE %) of short-retinoids 5 and 3a against the ve species mentioned. Compound 3a 5a1 5a2 5a3 5a4 5a5 5a6 5b GE KE % GE KE % GE KE % GE KE % GE KE % GE KE % GE KE % GE KE % mean % DS C. albicans 2.447 99.64 3.56 99.97 2.05 99.11 2.263 99.46 1.21 93.90 0.75 82.31 1.83 98.52 0.75 82.50 94.43 7.67 P. aeruginosa 0.286 48.18 0.26 45.46 1.16 93.13 0.176 33.33 0.36 53.67 0.30 50.00 1.15 92.94 0.27 45.45 57.77 22.54 E. coli 0.456 65.00 0.56 72.50 1.50 96.87 0.301 50.00 0.60 75.00 0.58 73.91 1.29 94.82 0.58 74.40 75.31 15.13 S. aureus 2.058 99.13 0.98 89.41 1.94 98.85 0.503 68.57 1.38 95.79 1.93 98.82 3.02 99.91 1.90 97.42 93.49 10.61 P. acnes 2.041 99.09 2.35 99.55 4.91 99.99 2.981 99.90 3.01 99.90 2.01 99.03 2.17 99.33 2.06 99.13 99.49 0.4 mean GE DS 1.46 1.01 1.54 1.38 2.31 1.19 1.25 1.28 1.31 1.04 1.12 0.80 1.89 0.75 1.11 0.81

allowed to stand at 10 C for 30 min and 1 h at room temperature. The resulting oil, puried by chromatography on silica gel (toluene/ethyl acetate 1:1), gave 5a1 in 88% yield as a thick oil. 5.1.2. 1-Phenyl-5-[(2,6,6-trimethyl-2-cyclohexen-1-yl) ethenyl]-1H-pyrazole 5a2 A solution of 0.4 g (1.6 mmol) of 3a and phenylhydrazine hydrochloride [0.5 g (3.2 mmol) in 5 mL of water and 0.8 g of sodium acetate trihydrate] in 10 mL of ethanol was reuxed for 15 min and allowed to stand at room temperature for 1 d. The resulting oil was extracted with chloroform. The combined extracts were dried (sodium sulphate) and evaporated to afford a red oil which was puried by chromatography on silica gel (toluene/ethyl acetate 1:1), giving 5a2 in 85% yield as a thick oil; IR (lm): 3 100, 2 920, 1 670, 1 600, 1 500, 1 450, 1 390 cm1; 1H-NMR (200 MHz): 0.88 (3H, s, CH3), 0.92 (3H, s, CH3), 1.20 (1H, m, H-5), 1.45 (1H, m, H-5), 1.60 (3H, s, CH3), 2.03 (2H, m, H-4), 2.25 (1H, d, H-1 J 8.98 Hz), 5.45 (1H, m, H-3), 6.07 (1H, dd, ethene, J 8.98; 15.76 Hz), 6.20 (1H, d, ethene, J 15.76 Hz), 6.47 (1H, d, H-3, J 1.98 Hz), 7.47 (5H, m, arom-H) 7.61 (1H, d, H-4, J 1.98 Hz); 13C-NMR (50 MHz, CDCl3) 23.44 (CH3), 23.53 (CH2), 27.39 (CH3), 28.23 (CH3), 32.00 (CH2), 33.11 (C), 55.35 (CH), 104.37 (CH), 119.44 (CH), 122.17 (CH), 125.72 (CH), 125.72 (CH), 128.17 (CH), 129.50 (CH), 129.50 (CH), 133.69 (C), 136.62 (CH), 140.17 (C), 140.56 (CH), 141.71 (C). Anal. C20H24N2 (C, H, N).

5.1.3. 1-(2,4-Dinitrophenyl)-5-[(2,6,6-trimethyl-2-cyclohexen-1-yl)ethenyl]-1H-pyrazole 5a3 A solution of 0.25 g (1 mmol) of 3a and 2,4-dinitrophenylhydrazine [0.2 g (1 mmol) in ethanol/H2SO4] in 10 mL of ethanol was allowed to stand at 10 C for 30 min and 2 h at room temperature. The precipitated solid was collected by ltration and washed with ethanol, yielding 5a3 as already pure yellow crystals (90% yield, m.p. 116117 C from ethanol). 5.1.4. 5-(2,6,6-Trimethyl-2-cyclohexen-1-yl)ethenyl-isoxazole 5a4 A solution of 0.4 g (1.6 mmol) of 3a and hydroxylamine hydrochloride [0.5 g (7.0 mmol) in 3 mL of water] in 30 mL of ethanol and 2 mL of a 10% water solution of sodium hydroxide was allowed to stand at room temperature for 3 d. The resulting oil, puried by chromatography on silica gel (toluene/ethyl acetate 1:1), gave 5a4 in 88% yield as a thick oil. 5.1.5. 6-(2,6,6-Trimethyl-2-cyclohexen-1-yl)ethenyl-2-aminopyrimidine 5a5 A solution of 0.5 g (2.0 mmol) of 3a and guanidine carbonate [0.36 g (2.0 mmol) in 3 mL of water] in 30 mL of ethanol was reuxed for 48 h. The resulting oil was extracted with chloroform. The combined extracts were dried (sodium sulphate) and evaporated to afford an oil which was puried by chromatography on silica gel (toluene/ethyl acetate 1:1), giving 5a5 in 85% yield as a thick oil.

841 5.1.6. 2-Methylthio-4-[(2,6,6-trimethyl-2-cyclohexen-1yl)ethenyl]-pyrimidine 5a6 To the solution of 0.5 g (2.0 mmol) of 3a in 10 mL of ethanol, 0.55 g (2.0 mmol) of S-methylisourea hydrogen sulphate dissolved in 2 mL of sodium hydroxide 2 N and 10 mL of water was added and the mixture reuxed for 48 h. The resulting oil was extracted with chloroform. The combined extracts were dried (sodium sulphate) and evaporated to afford an oil which was puried by chromatography on silica gel (toluene/ethyl acetate 1:1), giving 5a6 in 85% yield as a thick yellow oil; IR (lm): 3 100, 2 900, 2 860, 1 680, 1 560, 1 450, 1 375, 1 200 cm1; 1H-NMR (200 MHz): 0.85 (3H, s, CH3), 0.92 (3H, s, CH3), 1.24 (1H, m, H-5), 1.44 (1H, m, H-5), 1.59 (3H, s, CH3), 2.04 (2H, m, H-4), 2.27 (1H, d, H-1 J 8.98 Hz), 2.29 (3H, s, SCH3), 5.47 (1H, m, H-3), 6.19 (1H, d, ethene, J 8.65 Hz), 6.59 (1H, d, H-4, J 5.30 Hz), 6.70 (1H, dd, ethene, J 15.15, 8.65 Hz), 8.15 (1H, d, H-5, J 5.30 Hz); 13C-NMR (50 MHz) 18.95 (CH3S), 23.41 (CH3), 23.56 (CH2), 27.81 (CH3), 28.35 (CH3), 31.78 (CH2), 33.11 (C), 55.18 (CH), 108.82 (CH), 122.53 (CH), 130.17 (CH), 133.27 (C), 141.89 (CH), 158.34 (CH), 164.408 (C), 164.55 (C). Anal. C16H22N2S (C, H, N, S). 5.1.7. 6-(2,6,6-Trimethyl-1-cyclohexen-1-yl)ethenyl-2aminopyrimidine 5b By use of the procedure in 5.1.5, compound 5b was prepared from 3b. The product was nally separated by chromatography on silica gel in 70% yield as a thick oil; IR (lm): 3 300, 3 180, 2 900, 1 680, 1 610, 1 550, 1 440 cm1; 1H-NMR (200 MHz): 1.02 (6H, s, CH3), 1.48 (2H, m, H-5), 1.60 (2H, m, H-4), 1.74 (3H, s, CH3), 2.02 (2H, t, H-3), 5.19 (2H, s, NH2), 5.92 (1H, d, ethene, J 16.28 Hz), 6.48 (1H, dd, ethene, J 16.28 Hz), 6.58 (1H, d, H-4, J 5,26 Hz), 8.19 (1H, d, H-5, J 5.26). Anal. C15H21N3 (C, H, N). 5.1.8. Semicarbazone 4a An ethanolic solution of 0.5 g of 3-chloro-5-(2,6,6trimethyl-2-cyclohexen-1-yl)-2,4-pentadienal 2a, 0.7 g of sodium acetate trihydrate and 0.5 g of semicarbazide hydrochloride was heated at 90 C with stirring. After cooling, the precipitate obtained was ltered off and crystallised from ethanol. The yellow crystals formed were pure 4a, (35% yield; m.p. 209210 C from ethanol). IR (KBr): 3 480, 3 140, 2 900, 1 700, 1 580, 1 420, 1 160 cm1; 1H-NMR (200 MHz, DMSO-d6): 0.80 (3H, s, CH3), 0.90 (3H, s, CH3), 1.20 (1H, m, H-5), 1.40 (1H, m, H-5), 1.55 (3H, s, CH3), 2.02 (2H, m, H-4), 2.30 (1H, d, H-1, J 9.06 Hz), 5.48 (1H, m, H-3), 5.95 (1H, dd, ethene, J 9.06 and J 9.30 Hz), 6.35 (2H, s, NH2), 6.38 (1H, d, ethene, J 9.30 Hz), 6.50 (1H, d, CHCH=N, J 10.66 Hz), 7.92 (1H, d, CHCH=N, J 10.66 Hz), 10.45 (1H, s, NH); Anal. C15H22N3OCl (C, H, N, Cl). 5.1.9. Semicarbazone 4b By use of the above procedure, semicarbazone 4b was obtained in 38% yield as yellow powder from 2b, m.p. 196198 C from ethanol. IR (KBr): 3 480, 3 120, 2 850, 1 690, 1 580, 1 420, 1 160 cm1; 1H-NMR (200 MHz, DMSO-d6): 1.05 (6H, s, CH3), 1.45 (2H, m, H-5), 1.55 (2H, m, H-4), 1.70 (3H, s, CH3), 2.05 (2H, m, H-3), 6.35 (2H, s, NH2), 6.45 (1H, d, ethene, J 15.78 Hz), 6.60 (1H, d, CHCH=N, J 10.52 Hz), 6.65 (1H, d, ethene, J 15.78 Hz), 7.98 (1H, d, CHCH=N, J 10.52 Hz), 10.48 (1H, s, NH); Anal. C15H22N3OCl (C, H, N, Cl). 5.2. Biological methods The micro-organisms used in this study were Pseudomonas aeruginosa ATCC 9027 (16 107 CFU/ mL), Escherichia coli ATCC 8739 (11 107 CFU/mL), Staphylococcus aureus ATCC 6538 (40 107 CFU/mL), Candida albicans ATCC 10231 (77 107 CFU/mL), Propionibacterium acnes ATCC 11827 (1.54 107 CFU/ mL). The microbicidal activity has been determined by a Reybrouck in vitro test [12, 13] which was carried out using 5 min of medication time at 25 1 C and 30 min of contact time with LPHT inactivator (0.3% lecithin, 3% polysorbate 80, 0.1% histidine, 0.5% sodium thiosulfate). The bacterial strains were obtained by the 4th subculture at 37 1 C/48h on TSA slants from freeze-dried stock cultures. Bacterial suspension was obtained in TSB after an incubation time of 24 h at 37 1 C followed by centrifugation at 2 000 g for 15 min and resuspension in tryptone dilution water (TDW). The number of viable organisms in the inoculum were determined by the plating technique, mixing 1.0 mL samples from the dilutions 106, 107 and 108 with 20 mL of TSA, melted and tempered to 45 C. All inoculations were carried out in triplicate. The test was carried out in a water bath at 25 1 C. At time zero, 0.1 mL volumes of the different bacterial suspensions were added to 10 mL of the retinoid solution to be tested. After 5 min, 1 mL volumes were transferred from the medication mixtures into 9 mL of inactivator solution (LPHT). After a contact time of 30 min, 0.1 mL portions of the undiluted inactivated mixture and of the 101, 102, 103 and 104 dilutions in TDW were spread in triplicate on plates containing 20 mL of TSA. The colony forming units were counted after incubation at 37 1 C for 24 h.

842 For the negative controls, which were submitted to the same medication procedure, distilled water was used instead of the retinoid derivatives. The germicidal activity is numerically expressed using the Germicidal Effect (GE, the decimal reduction time), meaning the ratio, expressed as logarithm, between the number of colony forming units (CFU) per mL in the control mixture without the tested retinoid (Nc) to the number of CFU after the exposure to the retinoids (Ns) or GE = log Nc log Ns. The corresponding killing effect percentage is obtained by the following mathematical expression: KE % = Nc Ns 100 Nc References
[1] [2] [3] [4] [5] [6] [7] [8] [9] Orfanos C.E., Schuppli R., Dermatologica Suppl. (1) (1978) 164. Orme M., Back D.J., Shaw M.A., Allen W.L., Tjia J., Cunliffe W.J., Jones D.H., Lancet 2 (8405) (1984) 752753. Flemetakis A.C., Tsambaos D.G., J. Chemother. 1 (6) (1989) 374376. Gollnick H., Schramm M., Dermatology (Basel) 196 (1998) 119125. LOreal S.A., Fr. Demande 2761600, 1998. Sefton G., Lew-Kaya D.A., Allergan Sales Inc., USA, PCT Int. Appl., WO 9856375, 1998. Sottofattori E., Anzaldi M., Balbi A., J. Heterocyclic Chem. 35 (1998) 13771380. Coquelet C., Roussillon S., Sincholle D., Bonne C., Alzatet A., PCT Int. Appl. (1985) WO 8504652; Drugs of the future (1986) 557558. Nastruzzi C., Simoni D., Manfredini S., Barbieri R., Feriotto G., Baraldi P.G., Spandidos D., Guarneri M., Gambari R., Anticancer Res. 9 (1989) 13771384. Simoni D., Manfredini S., AghazadehTabrizi M., Bazzanini R., Baraldi P.G., Ferroni R., Traniello F., Nastruzzi C., Feriotto G., Gambari R., Drug Design and Discovery 8 (1992) 165177. Cortesi R., Esposito E., Gambari R., Menegatti E., Nastruzzi C., Eur. J. Pharmacol. Sci. 2 (1994) 281291. Reybrouck G., Werner H.P., Zbl. Bakteriol. Parasitenkunde Infektions-krankh. Hyg. Orig. B 165 (1977) 126137. Reybrouck G., Borneff J., Van De Voorde H., Werner H.P., Zbl. Bakteriol. Parasitenkunde Infektionskrankh. Hyg. Orig. B 168 (1979) 463479.

[10]

Acknowledgements

[11] [12]

We wish to thank CNR and MURST (Rome) for nancial support.

[13]

Eur. J. Med. Chem. 34 (1999) 843852 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

843

Preliminary communication

4-(4-Fluorobenzoyl)-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidine and its derivatives: synthesis and affinity at 5-HT2A, 5-HT2B and 5-HT2C serotonin receptors
Francesco Claudia*, Loredana Scocciaa, Gianfabio Giorgionia, Beatrice Accorronia, Gabriella Maruccia, Stefania Gessib, Anna Siniscalchib, Pier Andrea Boreab
b a Department of Chemical Sciences, University of Camerino, Via S. Agostino 1, 62032 Camerino, Italy Department of Experimental and Clinical Medicine, Section of Pharmacology, University of Ferrara, Via Fossato di Mortara 1719, 44100 Ferrara, Italy

(Received 24 July 1998; accepted 19 May 1999)

Abstract 4-(4-Fluorobenzoyl)-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidine (7) and its derivatives modied at the carbonyl group of the uorobenzoyl moiety were prepared and evaluated for affinity at 5-HT2A, 5-HT2C (rat cortex) and 5-HT2B (rat stomach fundus) serotonin receptors. Compound 7 bound the 5-HT2A sites with higher affinity (Ki = 8.2 nM) than the 5-HT2B (Kb = 1 290 nM) and 5-HT2C ones (Ki = 54.2 nM). Modication of the benzoyl carbonyl group decreased the 5-HT2A and 5-HT2C affinities but did not signicantly inuence 5-HT2B affinity. This suggests that the carbonyl group is the determinant for the interaction with 5-HT2A and 5-HT2C receptor subtypes. Compound 7 was found to be a 5-HT2A receptor antagonist. 1999 ditions scientiques et mdicales Elsevier SAS 4-(4-Fluorobenzoyl)-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidine derivatives / synthesis / 5-HT2A, 5-HT2B, and 5-HT2C serotonin receptor affinity / 5-HT2A antagonistic activity

1. Introduction Serotonin (5-hydroxytryptamine, 5-HT) mediates a number of neuronal processes both in the central nervous system and peripheral tissues. During the last decade multiple 5-HT receptor subtypes have been characterized and grouped in seven classes (5-HT15-HT7) [1]. The 5-HT2 class includes the subtypes 5-HT2A, 5-HT2B, and 5-HT2C which are grouped together considering their high degree of transmembrane sequence homology and second messenger coupling system. The 5-HT2A subtype is present in the brain (cortical regions) [2] and periphery (gastrointestinal tract, cardiovascular system) [3], and is involved in various cardiovascular and mental disorders, such as depression and schizophrenia [4]. The 5-HT2B receptor is expressed in rat stomach fundus, where it mediates a contractile response to 5-HT. Its mRNA transcript is present in the human brain [5], and it has been suggested to be involved in the pathophysiology of
*Correspondence and reprints

migraine [6]. The 5-HT2C receptor was initially characterized in the choroid plexus, is widely distributed in the brain [2], and has been suggested as playing a role in migraine, obsessive compulsive disorders, and anxiety [7]. To date, few antagonists discriminating between the 5-HT2 subtypes are available [814]. With the lack of selective agents has come the realization that many pharmacological and physiological effects once attributed to 5-HT2A receptors (previously 5-HT2) may, in fact, involve more than one receptor subtype. Hence, there is a need to discover new agents able to bind with high selectivity at one of the three subtypes. The 3-{2-[4-(4-uorobenzoyl)-1-piperidinyl]ethyl}2,4(1H,3H)-quinazolinedione (ketanserin, 1, gure 1) is a prototypical 5-HT2A antagonist reported to bind with as little as 15-fold [9] to as much as 140-fold selectivity for 5-HT2A versus 5-HT2C receptors [10]. It was used to explain the molecular structural prerequisites for the 5-HT2A antagonist binding [9, 10], and to build the

844 Some time ago, Ariens suggested that an agonist can be turned into a competitive antagonist by appending to the agonist structure a hydrophobic bulky group able to bind the accessory binding sites of the receptor [22]. On the basis of the Ariens strategy and of the above mentioned observations, we decided to connect the 4-(4uorobenzoyl)piperidine with the 2-(4-iodo-2,5-dimethoxyphenyl)ethyl fragment of 4 to obtain 4-(4-uorobenzoyl)-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidine 7 (gure 2). This compound was chosen in order to obtain a new 5-HT2 antagonist and to investigate how a modied tail tied to 4-(4-uorobenzoyl)piperidine could inuence selectivity for the 5-HT2A, 5-HT2B, and 5-HT2C subtypes. Studies with ketanserin and related 5-HT2 antagonists have suggested that the carbonyl group of 4-(4uorobenzoyl)piperidine is involved in a hydrogenbonding interaction with the 5-HT2A/2C receptors [23] and may have a prominent role in anchoring ketanserin to 5-HT2A receptors [10, 15]. Thus, in order to obtain further information about the role of the carbonyl group in the affinity and selectivity for the three 5-HT2 subtypes, two series of compounds were synthesized. The rst series keeps an sp2 carbon atom placed between the piperidine and the uorophenyl ring (811, gure 3), while in the other (1216, gure 4) the same carbon has an sp3 hybridization. 2. Chemistry The desired 4-(4-uorobenzoyl)-1-[2-(4-iodo-2,5dimethoxyphenyl)ethyl]piperidine 7 was synthesized by condensation of 4-(4-uorobenzoyl)piperidine 2 with the tosyl ester of 2-(4-iodo-2,5-dimethoxyphenyl)ethanol 6 (gure 2). Iodination of 2-(2,5-dimethoxyphenyl)ethanol with iodine in the presence of silver triuoroacetate gave the 2-(4-iodo-2,5-dimethoxyphenyl)ethanol 5 which was esteried with tosyl chloride. Oxime 8 and the O-alkyloximes 10 and 11 were obtained by reaction of the ketone 7 in pyridine and absolute ethanol with hydroxylamine or O-ethylhydroxylamine or O-benzylhydroxylamine (gure 3). The oxime acetate 9 was obtained from 8 by reaction with acetic anhydride. Oxime isomers were not separated. Ketone 7 was also reduced by sodium borohydride to alcohol 12 or transformed into the dioxolane 13. From 12, by reaction with acetic anhydride or benzoyl chloride, the esters 14 and 15 were obtained (gure 4). In order to obtain the derivative 16 with a methylene replacing the carbonyl group, the tosylate ester 6 was reacted with 4-(4-uorobenzyl)piperidine.

Figure 1. Structures of ketanserin, 4-(4-uorobenzoyl) piperidine (2), DOI, and 1-(4-iodo-2,5-dimethoxyphenyl)-2aminoethane (4).

three-dimensional models of 5-HT2A and 5-HT2C receptors [1517]. Several studies on ketanserin analogues showed that the 4-(4-uorobenzoyl)piperidine (2) fragment is endowed with 5-HT2A antagonistic activity and seems to be essential for binding at the 5-HT2A receptor [9, 18]. Among the 5-HT2 agonists, the most extensively studied is the 1-(4-iodo-2,5-dimethoxyphenyl)-2-aminopropane (3 R(-)-DOI), currently reported as a 5-HT2C/2A agonist [19]. Investigations on 1-(2,5-dimethoxyphenyl)2-aminopropane derivatives have established that the methyl group to the amine in 3 does not inuence the in vitro receptor affinity. Thus, the 1-(4-iodo-2,5dimethoxyphenyl)-2-aminoethane 4 binds with high afnity (Ki = 1.53 nM) the 5-HT2 receptors labelled by R-[125I]-DOI [20], and has been reported to act as agonist at 5-HT2 receptors [21]. Site-directed mutagenesis and molecular modelling studies indicate that the piperidine nitrogen atom of 1 and the amino group of 3 have a strong elecrostatic interaction with the carboxylate anion of the aspartate residue Asp 155 in transmembrane helix 3 (TM 3) of the 5-HT2A receptor [15, 17]. Therefore, it would appear that Asp155 is a common binding site for agonists 3 and 4, and for antagonist 1.

845

Figure 2. Synthesis of 7.

3. Pharmacology The affinity of compounds for 5-HT2A and 5-HT2C receptors was assessed in vitro in the cerebral cortex preparations. [3H]Ketanserin as radiolabelled ligand for the 5-HT2A receptors, and [3H]mesulergine for the 5-HT2C receptors were used. The antagonistic affinity at the 5-HT2B receptors was determined by the inhibition of 5-HT-induced contractions of rat stomach fundus. The results are reported in table I. The 5-HT2B receptor affinities were obtained from functional data; however, caution should be exercised when comparing them with data from binding assays. Derivatives 7, 9, 10, and 13 were evaluated for their antagonist activity at central 5-HT2A receptors by testing their ability to antagonize the facilitatory effect of 5-HT on basal acetylcholine release from guinea-pig striatal slices [24]. Acetylcholine release induced by 5-HT could be attributed to 5-HT2A receptor activation. In fact, it was concentration-dependently antagonized by 5-HT2A antagonists added to the superfusion medium from the

beginning of the experiment, while the 5-HT2C antagonist mesulergine was unable to counteract it, except at 10 mol, a high concentration able to block 5-HT2A receptors as well. 4. Results and discussion The data in table I indicate that the 4-(4-uorobenzoyl)piperidine 2 binds with the same affinity at 5-HT2A and 5-HT2C sites. Moreover, 2-(4-iodo-2,5dimethoxyphenyl)ethylamine 4 binds with lower affinity to the 5-HT2A than to the 5-HT2C sites (Ki = 300 33 nM vs. 2.5 0.14 nM). Accordingly, in the assays of acetylcholine release from guinea-pig striatal slices, its 5-HT2A agonistic action is negligible (net [3H]choline efflux increase at 30 M = 0.11 0.03%, n = 5), as is the antagonistic one (5 HT 30 M-induced net [3H]choline efflux increase in the presence of 4 = 0.55 0.06%, not signicantly different from 5-HT 30 M alone (0.89 0.11%). Functional assays on rat stomach fundus indicate that 4 acts as a full agonist (pD2 = 6.54 0.12, less potent

846

Figure 3. Synthesis of compounds 911.

than 5-HT: pD2 = 8.20 0.09). Introduction on the piperidine nitrogen of the 2-(4-iodo-2,5-dimethoxyphenyl)ethyl fragment increases the affinity for all 5-HT2 receptor subtypes, and this effect is more pronounced for 5-HT2A sites (80-fold) than for 5-HT2B and 5-HT2C

(13-fold) ones. On the other hand, when the amino group of 4 is substituted by the 4-(4-uorobenzoyl)piperidine moiety, the 5-HT2A and 5-HT2B affinities increase and the 5-HT2C affinity decreases. Compound 7 binds the 5-HT2A sites with lower affinity than does the reference com-

847

Figure 4. Synthesis of compounds 1216.

pound 1 (8.2 nM vs. 0.24 nM) and, under our assay conditions, displays the same selectivity for 5-HT2A versus 5-HT2C receptors (about 7-fold). Compound 7 is also 157-fold more selective for 5-HT2A versus 5-HT2B

receptors. These results suggest that the 2-(4-iodo-2,5dimethoxyphenyl)ethyl fragment enhances to a greater extent the affinity for 5-HT2A than that for 5-HT2B or 5-HT2C receptors.

848
Table I. 5-HT2A, 5-HT2C (rat cortex) receptor binding affinities, and 5-HT2B (rat stomach fundus) affinity of compounds 1, 2, 4, 716a. Compound 1 2 4 7 8 9 10 11 12 13 14 15 16
a c

Ki (nM)b 5-HT2A 0.24 0.01 667 44 300 33 8.2 0.4 666 44 100 6 150 6 275 14 616 60 118 14 325 14 275 14 238 7 5-HT2C 1.76 0.05 710 59 2.5 0.14 54.2 2.1 2 270 176 1 450 87 887 59 598 37 1 220 78 432 19 687 59 1 660 137 497 20

Kb (nM)c 5-HT2B 8 660 810 > 10 000 > 10 000d 1 290 60 1 300 150 1 350 170 1 000 130 NDe 807 115 1 080 170 ND 7 140 980 ND

All values represent means SEM; n 3 determinations. bBinding affinity (rat cortex; 5-HT2A [3H]keranserin, 5-HT2C [3H]mesulergine). Apparent antagonist dissociation constant, rat stomach fundus. The compounds were tested at 105 M and behave as competitive antagonists. d Affinity constant (Kb) determined as reported in [28]. eNot determined.

The modication of the carbonyl group in 7 always decreases the 5-HT2A and 5-HT2C affinities and does not affect signicantly the 5-HT2B affinity. In the compounds having an sp2 carbon atom (811), the carbonyl substitution with the oxime group decreases the 5-HT2A and 5-HT2C affinities. Both acetylation and etherication of the oxime hydroxyl group increases the affinities. These are inuenced in an opposite way: the 5-HT2A affinity increases and 5-HT2C affinity decreases following the order 9, 10, 11. This could suggest a different interaction of 911 with the 5-HT2A and 5-HT2C sites. It can also be seen from table I that the replacement of the carbonyl group with an sp3 carbon atom (compounds 1216) decreases the 5-HT2A and 5-HT2C affinities. In the series with an sp3 carbon atom between the piperidine and the uorophenyl ring, only the dioxolane 13 presents moderate 5-HT2A affinity. The highest decrease in affinity is observed when the carbonyl group is reduced to alcohol 12. Of importance is that the 5-HT2A and 5-HT2C affinities decrease to a greater extent in derivatives 8 and 12 bearing a hydroxyl group. The results indicate that the carbonyl oxygen participates in a key binding interaction and that a hydrogen-bond acceptor seems to be required by the 5-HT2A and 5-HT2C receptors. Moreover, it should be observed that also the carbonyl group reduction to methylene (compound 16) decreases the affinities. With regard to the 5-HT2B receptor, it is evident that the carbonyl group modication does not signicantly inuence the affinity. Derivatives 7, 9, 10, and 13 show antagonist activity at central 5-HT2A receptors counteracting the release of

acetylcholine induced by 5-HT (gure 5). They behave as competitive antagonists and are less active than the standard 1. The potency order was 1 > 7 > 9 > 10 > 13, in good agreement with the binding data. All the antagonists, when tested at the highest concentration, proved to

Figure 5. Basal tritium efflux from guinea-pig striatal slices prelabelled with [3H]choline. Relationship between 5-HT2A antagonist concentrations (mol, abscisses, log scale) and inhibition of the facilitatory effect of 5-HT 30 mol (%, ordinate). Points represent the means SEM of 49 experiments. s Ketonserin, m 7, . 9, x 10, 5 13.

849 be devoid of intrinsic activity, except 13, which at 30 M evoked a net increase in [3H]choline efflux of 0.67 0.11% (n = 3). In conclusion, compound 7 represents a successful application of the Ariens strategy in the design of a 5-HT2A antagonist by substituting the amino group of 2-(4-iodo-2,5-dimethoxyphenyl)ethylamine with the bulky lipophilic moiety of 4-(4-uorobenzoyl)piperidine. Comparing 1 with 7, it is evident that the 3-ethyl-2,4quinazolinedione moiety gives higher 5-HT2A and 5-HT2C affinities than does the 2-(4-iodo-2,5-dimethoxyphenyl)ethyl fragment. 5. Experimental protocols 5.1. Chemistry Melting points were determined on a Buchi 510 apparatus and are uncorrected. Microanalyses were performed on a 1106 Carlo Erba CHN Analyzer, and the results were within 0.4% of the theoretical values. 1H NMR spectra were recorded on a Varian VXR 200 MHz spectrometer. Chemical shifts are reported in parts per million () downeld from the internal standard tetramethylsilane (Me4Si). The identity of all new compounds was conrmed both by elemental analysis and NMR data; homogeneity was conrmed by TLC on silica gel Merck 60 F254. Chromatographic purications were performed by Merck-60 silica gel columns 70230 mesh ASTM from Merck with a reported solvent. 5.1.1. 2-(4-Iodo-2,5-dimethoxyphenyl)ethanol 5 A solution of iodine (0.92 g, 3.6 mmol) in chloroform (20 mL) was added dropwise, under stirring at room temperature, to a slurry of silver triuoroacetate (0.89 g, 3.6 mmol) and 2-(2,5-dimethoxyphenyl)ethanol (0.66 g, 3.6 mmol) in chloroform (5 mL). The mixture was stirred for 30 min. The insoluble material was ltered. The ltrate was washed with aqueous 10% Na2SO3, brine and dried (Na2SO4). The solvent was evaporated and the residue was crystallized from isopropyl ether, m.p.: 9394 C; yield 76%. 1H-NMR (CDCl3) 1.59 (1H, bs, OH), 2.88 (2H, t, J = 6.3 Hz, ArCH2), 3.80 (8H, m, OCH3, CH2O), 6.68 (1H, s, ArH), 7.21 (1H, s, ArH). Anal. C10H13IO3 (C, H). 5.1.2. 2-(4-Iodo-2,5-dimethoxyphenyl)ethyl-p-toluensulfonate 6 p-Toluenesulfonyl chloride (0.63 g, 3.3 mmol) was added portionwise to a solution of 5 (0.92 g, 3 mmol) in dry pyridine (1 mL). After stirring for 5 h at room temperature, 2 N HCl (9 mL) was added to yield a precipitate. This was ltered and crystallized from cyclohexane, m.p.: 110112 C; yield 82%. 1H-NMR (CDCl3) 2.46 (3H, s, CH3), 2.89 (2H, t, J = 6.3 Hz, ArCH2), 3.63 (3H, s, OCH3), 3.88 (3H, s, OCH3), 4.22 (2H, t, J = 6.3 Hz, CH2O), 6.57 (1H, s, ArH), 7.08 (1H, s, ArH), 7.23 (2H, d, J = 8.5 Hz, ArH), 7.58 (2H, d, J = 8.5 Hz, ArH). Anal. C17H19IO5S (C, H). 5.1.3. 4-(4-Fluorobenzoyl)-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidine hydrochloride 7 A mixture of 6 (3 g, 6.5 mmol), 4-uorobenzoylpiperidine (1.35 g, 6.5 mmol) and K2CO3 (1.28 g, 9.3 mmol) in acetone (30 mL) was warmed to reux for 20 h. The solvent was removed by evaporation, and the residue was partitioned between H2O and CHCl3. The organic extracts were dried (Na2SO4) and evaporated. The residue was puried by chromatography eluting with CHCl3:acetone:MeOH (6.7:3:0.3, v/v) and crystallized from isopropyl ether; m.p.: 98100 C; yield 84%. 1HNMR (CDCl3) 1.88 (4H, m, H3,5pip), 2.21 (2H, m, H2,6ax-pip), 2.58 (2H, m, ArCH2), 2.80 (2H, m, NCH2), 3.08 (2H, m, H2,6eq-pip), 3.21 (1H, m, H4pip), 3.78 (3H, s, OCH3), 3.83 (3H, s, OCH3), 6.70 (1H, s, ArH), 7.14 (2H, m, ArH), 7.20 (1H, s, ArH), 7.98 (2H, m, ArH). Hydrochloride: crystallized from absolute EtOH, m.p.: 232234 C. Anal. C22H25FINO3.HCl (C, H, N). 5.1.4. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidin-4-yl}methanone oxime hydrochloride 8 A mixture of 7 (1 g, 2 mmol), hydroxylamine hydrochloride (0.35 g, 5 mmol), and pyridine (2 mL) in absolute EtOH (5 mL) was warmed to reux for 3 h. After evaporation of the solvent, water was added. The precipitate was ltered and crystallized from absolute EtOH; m.p.: 239241 C; yield 92%. 1H-NMR (DMSO-d6) 1.88 (4H, m, H3,5pip), 2.23 (1H, m, H4pip), 2.96 (4H, m, ArCH2, H2,6ax-pip), 3.11 (2H, m, NCH2), 3.55 (2H, m, H2,6eq-pip), 3.75 (6H, s, OCH3), 6.93 (1H, s, ArH), 7.38 (5H, m, ArH), 10.24 (1H, bs, NH+), 10.90 e 11.45 (1H, 2s, OH). Anal. C22H26FIN2O3.HCl (C, H, N). 5.1.5. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidin-4-yl}methanone O-acetyloxime hydrochloride 9 A mixture of 8 (0.8 g, 1.6 mmol) and acetic anhydride (2 mL) was stirred at 60 C for 1.5 h. The mixture was cooled and partitioned between 5% aqueous Na2CO3 and Et2O. The organic extracts were washed with H2O, dried (Na2SO4) and evaporated. The residue was crystallized from isopropyl ether; m.p.: 7880 C, yield 77%. The residue was dissolved in Et2O and treated with ethereal HCl. The solid obtained was crystallized from acetone;

850 m.p.: 168170 C. NMR (DMSO-d6) 2.0 (7H, m, H3,5pip, CH3), 3.11 (7H, m, ArCH2, H4pip, H2,6ax-pip), 3.61 (2H, m, H2,6eq-pip), 3.78 (6H, s, OCH3), 6.96 (1H, m, ArH), 7.42 (5H, m, ArH), 10.35 (1H, bs, NH+). Anal. C24H28FIN2O4.HCl (C, H, N). 5.1.6. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidin-4-yl}methanone-O-ethyloxime hydrochloride 10 This was made in the same way as 8 from 7 and O-ethylhydroxylamine hydrochloride. The precipitate obtained after addition of water was ltered and treated with a saturated solution of Na2CO3. The mixture was extracted with Et2O. The organic extracts were dried (Na2SO4) and evaporated. The residue was dissolved in absolute EtOH, and EtOH saturated with HCl gas was added. The solvent was evaporated and the residue was crystallized from 2-propanol; m.p.: 203205 C, yield 75%. 1H-NMR (CDCl3) 1.1 (3H, t, J = 6.7 Hz, CH3), 2.0 (2H, m, H3,5ax-pip), 2.39 (2H, m, H3,5eq-pip), 2.68 (3H, m, H2,6ax-pip, H4pip), 3.19 (4H, m, ArCH2, NCH2), 3.66 (2H, m, H2,6eq-pip), 3.78 (3H, s, OCH3), 3.80 (3H, s, OCH3), 4.08 (2H, q, J = 6.7 Hz, OCH2), 6.82 (1H, s, ArH), 7.10 (2H, m, ArH), 7.25 (3H, m, ArH), 12.42 (1H, bs, NH+). Anal. C24H30FIN2O3.HCl (C, H, N). 5.1.7. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidin-4-yl}methanone O-benzyloxime hydrochloride 11 This was made in the same way as 8 from 7 and O-benzylhydroxylamine hydrochloride. The precipitate obtained after addition of water was ltered and treated with a saturated solution of Na2CO3. The mixture was extracted with Et2O. The organic extracts were dried (Na2SO4) and evaporated. The residue was dissolved in absolute EtOH, and EtOH saturated with HCl gas was added. The solvent was evaporated and the residue was crystallized from 2-propanol; m.p.: 145147 C, yield 71%. 1H-NMR (DMSO-d6) 1.82 (3H, m, H3,5ax-pip, H4pip), 2.94 (6H, m, ArCH2, H2,6ax-pip, H3,5eq-pip), 3.12 (2H, m, NCH2), 3.45 (2H, m, H2,6eq-pip), 3.76 (6H, s, OCH3), 5.02 (2H, s, OCH2), 6,92 (1H, s, ArH), 7.30 (10H, m, ArH), 9.97 (1H, bs, NH+). Anal. C29H32FIN2O3.HCl (C, H, N). 5.1.8. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidin-4-yl}methanol fumarate 12 To a solution of the ketone 7 (0.5 g, 1 mmol) in CH3OH (12 mL), stirred at room temperature, NaBH4 (0.038 g, 1 mmol) was added. The mixture was stirred for 12 h. The solvent was evaporated and the residue was partitioned between H2O and CH2Cl2. The organic extracts were dried (Na2SO4) and evaporated. The residue was puried by chromatography eluting with CHCl3:acetone:MeOH (6.7:3:0.3, v/v) to give an uncrystallizable oil; yield 84%. 1H-NMR (CDCl3) 1.43 (4H, m, H3,5pip), 2.0 (4H, m, OH, H4pip, H2,6ax-pip), 2.49 (2H, m, ArCH2), 2.74 (2H, m, NCH2), 3.02 (2H, m, H2,6eq-pip), 3.70 (3H, s, OCH3), 3.78 (3H, s, OCH3), 4.35 (1H, m, HCO), 6.65 (1H, s, ArH), 7.01 (2H, m, ArH), 7.18 (1H, s, ArH), 7.25 (2H, m, ArH). The oil was dissolved in absolute EtOH and treated with a solution of fumaric acid in absolute EtOH. The fumarate was ltered and crystallized from absolute EtOH, m.p.: 189191 C. Anal. C22H27FINO3.C4H4O4 (C, H, N). 5.1.9. 4-[2-(4-Fluorophenyl)-[1,3]dioxolan-2-yl]-1-[2(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidine hydrochloride 13 A mixture of ketone 7 (0.5 g, 1 mmol) and ethylene glycol (3 mL) was saturated with HCl gas and heated at 90 C for 1 h. The resulting solid was collected and crystallized from absolute EtOH; m.p.: 263265 C, yield 93%. 1H-NMR (DMSO-d6) 1.64 (4H, m, H3,5pip), 2.09 (1H, m, H4pip), 2.88 (4H, m, ArCH2, H2,6ax-pip), 3.11 (2H, m, NCH2), 3.50 (2H, m, H2,6eq-pip), 3.70 (8H, m, OCH3, OCH2), 3.98 (2H, m, OCH2), 6.92 (1H, s, ArH), 7.21 (2H, m, ArH), 7.38 (3H, m, ArH), 9.93 (1H, bs, NH+). Anal. C24H29FINO4.HCl (C, H, N). 5.1.10. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidin-4-yl}methylacetate fumarate 14 A mixture of 12 (0.5 g, 1 mmol), dry pyridine (5 mL) and acetic anhydride (0.47 mL, 5 mmol) was stirred at room temperature for 12 h, and then poured into ice. EtOH (2 30 mL) was added and the mixture was partially evaporated. The aqueous concentrate was made basic with 2 N NaOH, and the mixture was extracted with Et2O. The organic extracts were dried (Na2SO4) and evaporated to give an uncrystallizable oil; yield 76%. 1 H-NMR (CDCl3) 1.40 (3H, m, H3,5ax-pip, H4pip), 1.85 (4H, m, H2,6ax-pip, H3,5eq-pip), 2.07 (3H, s, CH3), 2.52 (2H, m, ArCH2), 2.77 (2H, m, NCH2), 3.01 (2H, m, H2,6eq-pip), 3.75 (3H, s, OCH3), 3.81 (3H, s, OCH3), 5.48 (1H, d, J = 8.8 Hz, OCH), 6.68 (1H, s, ArH), 7.0 (2H, m, ArH), 7.18 (1H, s, ArH), 7.25 (2H, m, ArH). The oil was dissolved in absolute EtOH and treated with a solution of fumaric acid in absolute EtOH. The fumarate was ltered and crystallized from absolute EtOH, m.p.: 188190 C. Anal. C24H29FINO4.C4H4O4 (C, H, N). 5.1.11. (4-Fluorophenyl)-{1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidin-4-yl}methylbenzoate fumarate 15 A mixture of 12 (0.5 g, 1 mmol), dry pyridine (5 mL) and acetic anhydride (0.52 mL, 4.5 mmol) was stirred at room temperature for 12 h, and then poured into ice.

851 EtOH (2 30 mL) was added and the mixture was partially evaporated. The aqueous concentrate was made basic with 2 N NaOH and the mixture was extracted with Et2O. The organic extracts were dried (Na2SO4) and evaporated to give an uncrystallizable oil; yield 74%. 1 H-NMR (CDCl3) 1.51 (3H, m, H3,5ax-pip, H4pip), 2.0 (4H, m, H2,6ax-pip, H3,5eq-pip), 2.57 (2H, m, ArCH2), 2.79 (2H, m, NCH2), 3.07 (2H, m, H2,6eq-pip), 3.75 (3H, s, OCH3), 3.81 (3H, s, OCH3), 5.78 (1H, d, J = 8.1 Hz, OCH), 6.70 (1H, s, ArH), 7.04 (2H, m, ArH), 7.19 (1H, s, ArH), 7.42 (5H, m, ArH), 8.08 (2H, m, ArH). The oil was dissolved in absolute EtOH and treated with a solution of fumaric acid in absolute EtOH. The fumarate was ltered and crystallized from absolute EtOH, m.p.: 181183 C. Anal. C29H31FINO4.C4H4O4 (C, H, N). 5.1.12. 4-(4-Fluorobenzyl)-1-[2-(4-iodo-2,5-dimethoxyphenyl)ethyl]piperidine hydrochloride 16 A mixture of 6 (0.46 g, 1 mmol), 4-uorobenzoylpiperidine (0.19 g, 1 mmol) and K2CO3 (0.2 g, 1.5 mmol) in acetone (10 mL) was warmed to reux for 20 h. The precipitate was ltered and the solution was evaporated. The residue was partitioned between H2O and CHCl3. The organic extracts were dried (Na2SO4) and evaporated. The residue was puried by chromatography eluting with CHCl3:acetone:MeOH (6.7:3:0.3, v/v) and crystallized from acetone; m.p.: 9698 C; yield 66%. 1H-NMR (CDCl3) 1.47 (5H, m, H3,5pip, H4pip), 1.97 (2H, m, H2,6ax-pip), 2.50 (4H, m, H2,6eq-pip, ArCH2), 2.75 (2H, m, ArCH2), 2.98 (2H, m, NCH2), 3.77 (3H, s, OCH3), 3.82 (3H, s, OCH3), 6.70 (1H, s, ArH), 6.98 (2H, m, ArH), 7.10 (2H, m, ArH), 7.20 (1H, s, ArH). The puried residue was dissolved in Et2O and treated with ethereal HCl. The solid obtained was crystallized from absolute EtOH; m.p.: 204205 C. Anal. C22H27FINO2.HCl (C, H, N). 5.2. Pharmacology 5.2.1. Materials Ketanserin tartrate was purchased from Research Biochemicals International (Natick, MA, USA). [3H]Ketanserin (64.1 Ci/mmol) was purchased from New England Nuclear, Boston, Mass., USA. [3H]Mesulergine (76 Ci/mmol) and [3H]choline (81 Ci/mmol) were purchased from Amersham Radiochemical Centre (Buckinghamshire, UK). All substances employed in the binding assays were dissolved in distilled water. 5.2.2. Animals In the radioligand-binding studies rat cortex was obtained from male Wistar rats (250300 g body weight) purchased from Nossan (Varese, Italy). Sections of stomach fundus were obtained from male CD Outbred rats (Charles River, Calco, Italy) weighing 125150 g. 5.2.3. Binding assays Cerebral cortices of male Wistar rats (150200 g) were dissected on ice. The tissue was homogenized in 50 mmol Tris-HCl buffer (pH = 7.7 at 25 C). The homogenate was centrifuged at 40 000 g for 10 min. The supernatant was discarded and the pellet was resuspended in the same volume of Tris-HCl buffer and incubated at 37 C for 10 min prior to a second centrifugation. Binding experiments [23] with [3H]ketanserin (64.1 Ci/mmol) and [3H]mesulergine (76 Ci/mmol) were performed in 250 L of buffer, which contained 1 nmol [3H]ketanserin or [3H]mesulergine, membranes from 10 mg (wet weight) of tissue and the compounds to be tested. After 30 min of incubation at 25 C, separation of bound from free radioligand was performed by rapid ltration through Whatman GF/B glass bre lters, which were washed three times with ice-cold buffer, dried and counted in 5 mL of Aquassure (Packard, Downers Grove, USA). Non-specic binding was measured in the presence of 10 mol 5-HT for 5-HT2A sites and 10 mol cinanserin for 5-HT2C sites with specic binding dened as the total binding minus the non-specic binding. Ki values were calculated from the Cheng-Prusoff equation [25] Ki = IC50/1 + (ligand/Kd), where Kd = 0.8 nmol/L for [3H]ketanserin and Kd = 1.9 nmol/L for [3H]mesulergine [26]. 5.2.4. Determination of apparent 5-HT2B receptor antagonist dissociation constant Experiments were performed as described by Nozulak et al. [27]. Male CD Outbred rats were sacriced by CO2, and longitudinal sections of the stomach fundus were prepared for in vitro examination. Strips were set up in organ baths of 10 mL containing Krebs solution (composition in mmol: NaCl, 118; KCl, 4.7; CaCl2, 1.25; KH2PO4, 1.2; MgSO4, 1.2; glucose, 11; NaHCO3, 25) constantly bubbled with 5% CO2 in oxygen. Contractions were measured isotonically under a resting tension of 1 g. Prior to testing, the strips were allowed to equilibrate for 1 h, during which time the bath was replaced every 15 min. After control cumulative contractile responses to serotonin were obtained in the stomach fundus, the tissues were incubated with an appropriate concentration of antagonist for 1 h. Contractile responses to serotonin were then repeated in the presence of the antagonist. Only one antagonist concentration was examined in each tissue. Apparent dissociation constants (Kb) were deter-

852 mined for each concentration of antagonist according to the following equation: Kb = [B]/(dose ratio 1) where [B] is the concentration of the antagonist, and the dose ratio is the ED50 of the agonist in the presence of the antagonist divided by the control ED50. 5.2.5. Inhibition of acetylcholine release Inhibition of the facilitatory effect of serotonin on basal acetylcholine release from guinea-pig striatal slices was determined as previously described [24]. Caudate nucleus slices (400 m thick) were incubated with 0.1 mol [3H]choline (81 Ci/mmol) for 30 min and superfused at 0.25 mL/min with Krebs solution (compositon in mmol: NaCl, 118.5; KCl, 4.8; CaCl2, 2.5; KH2PO4, 1.2; MgSO4, 1.2; NaHCO3 25, glucose, 11; hemicholinium-3, 0.01), bubbled with 95% O2 and 5% CO2. The radioactivity of the 5 min superfusate samples was determined by liquid scintillation. The effect of 5-HT, both in absence and in presence of antagonists, was quantied as the net increase of tritium efflux over the basal one, calculated as fractional rate (FR), i.e. as percent of tissue tritium content. The net increase of [3H]choline efflux, induced by 5-HT 30 M, added to the Krebs solution from the 45th min of superfusion, was 0.89 0.11% (n = 11). Acknowledgements This work was supported by Consiglio Nazionale delle Ricerche (CNR, Rome) grant 97.02771.CT03, and by the University of Camerino. References
[1] [2] [3] [4] Hoyer D., Martin G., Neuropharmacology 36 (1997) 419428 . Pazos A., Probst A., Palacios J.M., Neuroscience 21 (1987) 123139. Peroutka S.J., Snyder S.H., Mol. Pharmacol. 16 (1979) 687699. Glennon R.A., Neurosci. Biobehav. Rev. 14 (1990) 3547. [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] Kursar J.D., Nelson D.L., Wainscott D.B., Baez M., Mol. Pharmacol. 46 (1994) 227234. Fozard J.R., Kalkman H.O., Naunyn-Schmiedebergs Arch. Pharmacol. 350 (1994) 225229. Curzon G., Kenneth G.A., Trends Pharmacol. Sci. 11 (1990) 181182. Kennett G.A., Curr. Opin. Invest. Drugs 2 (1993) 317362. Herndon J.L., Ismaiel A., Ingher S.P., Teitler M., Glennon R.A., J. Med. Chem. 35 (1992) 49034910. Ismaiel A.M., Arruda K., Teitler M., Glennon R.A., J. Med. Chem. 38 (1995) 11961202. Forbes I.T., Jones G.E., Murphy O.E., Holland V., Baxter G.T., J. Med. Chem. 38 (1995) 855857. Audia J.E., Evrard D.A., Murdoch G.R., Droste J.J., Nissen J.S., Schenck K.W. et al., J. Med. Chem. 39 (1996) 27732780. Bromidge S.M., Duckworth M., Forbes I.T., Ham P., King F.D., Thewlis K.M. et al., J. Med. Chem. 40 (1997) 34943496. Weinhardt K.K., Bonhaus D.W., De Souza A., Bioorg. Med. Chem. Lett. 6 (1996) 26872692. Kristiansen K., Edvardsen ., Dahl S.G., Med. Chem. Res. 3 (1993) 370385. Kristiansen K., Dahl S.G., Eur. J. Pharmacol. 306 (1996) 195210. Westkaemper R.B., Glennon R.A., Med. Chem. Res. 3 (1993) 317334. Holtje H.D., Jendretzki U.K., Arch. Pharm. 328 (1995) 577584. Glennon R.A., Raghupathi R., Bartyzel P., Teitler M., Leonhardt S., J. Med. Chem. 35 (1992) 734740. Johnson M.P., Mathis C.A., Shulgin A.T., Hoffmann A.J., Nichols D.E., Pharmacol. Biochem. Behav. 35 (1990) 211217. Nichols D.E., Frescas S., Marona-Lewicka D., Huang X., Roth B.L., Gudelsky G.A., Nash J.F., J. Med. Chem. 37 (1994) 43464351. Ariens E.J., Beld A.J., Rodrigues de Miranda J.F., Simonis A.M., in: OBrien R.D. (Ed.), The Receptors: A Comprehensive Treatise, Vol. 1, Plenum, New York, 1979, pp. 3391. Pierce P.A., Kim J.Y., Peroutka J., Naunyn-Schmiedebergs Arch. Pharmacol. 346 (1992) 411. Siniscalchi A., Beani L., Bianchi C., Neuropharmacology 29 (1992) 10911093. Cheng Y.C., Prusoff W.H., Biochem. Pharmacol. 22 (1973) 30993108. Pazos S., Hoyer D., Palacios J.M., Eur. J. Pharmacol. 106 (1985) 531538. Nozulak J., Kalkman H.O., Floersheim P., Hoyer D., Schoeffter P., Buerki H.R., J. Med. Chem. 38 (1995) 2833. Furchgott R.F., Bursztyn P., Ann. NY Acad. Sci. 144 (1967) 882899.

[23] [24] [25] [26] [27] [28]

Eur. J. Med. Chem. 34 (1999) 853858 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

853

Short communication

Synthesis, physicochemical characterization and cytotoxic screening of new complexes of cerium, lanthanum and neodymium with Niffcoumar sodium salt
Ilia Manolova*, Irena Kostovab, Spiro Konstantinovc, Margarita Karaivanovac
c

Department of Industrial Pharmacy, Faculty of Pharmacy, Medical University, 2 Dunav St., 1000 Soa, Bulgaria b Department of Chemistry, Faculty of Pharmacy, Medical University, 2 Dunav St., 1000 Soa, Bulgaria Department of Pharmacology and Toxicology, Faculty of Pharmacy, Medical University, 2 Dunav St., 1000 Soa, Bulgaria (Received 20 August 1998; accepted 20 January 1999)

Abstract The complexes of the types Ln(HN)3.nH2O [where Ln = Ce, La, Nd; HN = Niffcoumar] have been synthesized by reaction of Niffcoumar sodium salt and the appropriate lanthanide nitrates. The solid complexes have been characterized and identied on the basis of elemental analysis, conductivities, IR- and 1H-NMR-spectroscopy. We suppose that the lacton- and keto-carbonyl oxygen atoms of Niffcoumar are bonded to the metal ion probably as a bidentate ligand. The experimental data obtained by MTT-assay show that all tested compounds manifest a similar low cytotoxic activity on lymphoma derived P3HR1 cells (20% maximum cell growth inhibition). The complexes of La and Nd have been shown to be more cytotoxic than the Ce complex against K-562 cells (about 75% maximal growth inhibition). Low to moderate cytotoxic activity has been established for the complexes of La and Ce on TPH-1 cells (25% maximum inhibition of cell growth). Our data indicate that the lanthanide complexes may have some antitumour potential which depends on the metal and the cell line used. 1999 ditions scientiques et mdicales Elsevier SAS metal complexes of Niffcoumar / lanthanides / cytotoxic screening

1. Introduction 4-Hydroxycoumarin derivatives are of interest because of their physiological, photodynamic and bacteriostatic activity [1]. 4-Hydroxycoumarin and its derivatives are known to exhibit an inclination for complexation [25]. Recently, some interesting lanthanide complexes of coumarin derivatives like bis (4-hydroxy-3-coumarinyl)acetic acid [6], N,N-bis(8-aceto-7-hydroxy-4methylcoumarin)ethylenediamine [7, 8], coumarin-3-carboxylic acid [9, 10] have been reported. As a part of our investigations on the synthesis and characterization of solid lanthanide complexes, we report herein the synthesis and characterization of lanthanide complexes with Niffcoumar sodium salt. The ligand is of special interest as it contains several potential donor sites. Niffcoumar sodium salt has not been used as a ligand up to now. Cytotoxic screening was carried out for the newly synthesized complexes with lanthanides (III). The present paper is a continuation of the authors earlier work on the
*Correspondence and reprints

complexes of lanthanides (III) with Warfarin sodium salt and Coumachlor sodium salt and their cytotoxic screening [11]. Literature data show that coumarins have an antitumour activity [12, 13]. The polycyclic coumarins hold promise as agents for the chemoprevention of cancer [14]. Lanthanides are a subject of increasing interest in bioinorganic and coordination chemistry. Furthermore, the literature data show that lanthanides manifest an antitumour activity [15, 16], which is in accordance with our previous investigations. 2. Chemistry 2.1. Preparation of the complexes The compounds used for preparing the solutions were Merck products, p.a. grade: Ce(NO3)3.6H2O, La(NO3)3.6H2O and Nd(NO3)3.6H2O. The sodium salt of 4-hydroxy-3[1-(4-nitrophenyl)-3-oxobutyl]-2H-1-benzo-

854
Table I. Elemental analyses of the complexes of Niffcoumar. Complex Ce(HN)3.4H2O La(HN)3.4H2O Nd(HN)3.6H2O M.p. (C) 270 > 300 > 300 H2O 5.93 5.68 6.05 5.68 7.95 8.26

Figure 1. Structure of Niffcoumar sodium salt.

HN = C19H15NO6

pyran-2-one (Niffcoumar sodium salt (gure 1) was prepared as follows: 3.53 g (10 mmol) 4-hydroxy-3[1-(4nitrophenyl)-3-oxobutyl]-2H-1-benzopyran-2-one (Niffcoumar) were added to 0.38 g (9.5 mmol) sodium hydroxide in 30 mL water. The mixture was stirred vigorously at room temperature for 1 h until it was clear. The solution was ltered and the ltrate was evaporated to dryness. The viscous residue was recrystallized from ethylacetate. Yield 3.2 g (86%). TLC: Al-foils Kieselgel 60 F254, Merck (FRG), developed by cyclohexanechloroform-acetic acid (10:10:4, vol. parts), detection: UV 254 nm. M.p. 132134 C, Bchi 510 apparatus (Switzerland), uncorrected. Niffcoumar sodium salt was used for the preparation of metal complexes as a ligand. General method of synthesis: the complexes were synthesized by mixing a water solution of the ligand with a water solution of the corresponding metal (III) salts in amounts equal to a metal/ligand molar ratio of 1:3. At the moment of mixing of the solutions a precipitate was obtained. The reaction mixture was stirred with an electromagnetic stirrer at 25 C for 1 h. The products thus obtained were separated from the solutions at pH 45, ltered off, washed three times with water and dried in a desiccator to constant weight. The complexes were insoluble in water, methanol and ethanol but had good solubility in DMSO. The complexes were characterized by elemental analysis. They were analysed for their metal content using standard procedures after destroying the organic matter. The ligand contained sodium ions. The complexes obtained were analysed for sodium ions by means of ame photometry and the analyses revealed a lack of these ions. The probable reason for this phenomenon was the hydrolysis of the salt in an aqueous solution. The water content in the complexes was determined by Karl Fisher analysis. The formation of the complexes was conrmed by IRand 1H-NMR-spectroscopy. Table I shows the results of the Karl Fisher analysis. All complexes behaved as non-electrolytes (m < 14 Ohm1 cm2 mol1).

3. Pharmacology 3.1. Colorimetric MTT (tetrazolium) assay This assay is based on the cellular reduction of MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] by the mitochondrial dehydrogenase of viable cells to a blue formazan product. The production of formazan can be measured spectrophotometrically following solubilization. We use the method described by Mossmann [17] with some modications. P3HR1 Burkitt lymphoma, CML derived K-562 and AML derived THP-1 cells were seeded in 96-well plates (100 L/well at a density of 1 105 cells/mL) and exposed to various concentrations of the tested compounds. After incubation for 48 or 72 h, the 5 mg/mL MTT solution in PBS was added to each well and was further incubated for 4 h at 37 C. The formazan crystals formed were dissolved by adding 100 L/well of 5% formic acid in 2-propanol. After a few minutes at room temperature to ensure that all crystals were dissolved, absorption was measured by an ELISA reader, using a test wavelength of 580 nm. For each concentration at least 8 wells were used (tables IIIV). Cell growth inhibition was calculated according to [OD of drug treatment/OD of control] 100. Data processing was executed with Microsoft Excel, Microsoft Word and Sigma Plot Windows.

4. Results and discussion The analytical data of the complexes (table I) conrm the composition Ln (HN)3.nH2O. The mode of bonding of the ligands to Ce(III), La(III) and Nd(III) ions was elucidated by recording the IR-spectra of the complexes as compared to those of the free ligands. IR-spectra of the compounds were recorded on solid state in Nujol in the range 3 600400 cm1 (table V).

855
Table II. Spectrophotometric data from MTT assays concerning the cytotoxic activity of complexes on P3HR1 cells in comparison with the inorganic salts. Compound MTT-formazan absorption at 580 nm untreated control Ce(NO3)3.6H2O La(NO3)3.6H2O Nd(NO3)3.6H2O Ce(HN)3.4H2O La(HN)3.4H2O Nd(HN)3.6H2O 0.3936 0.0672 0.3936 0.0672 0.3936 0.0672 0.8167 0.0565 0.8167 0.0565 0.8167 0.0565 25 M 0.4835 0.0419 0.4429 0.0579 0.4875 0.0885 0.7270 0.0270 0.7454 0.0363 0.8485 0.0732 100 M 0.4216 0.0511 0.4258 0.0801 0.4843 0.0245 0.7346 0.1086 0.7154 0.0939 0.7710 0.0712 400 M 0.3649 0.0823 0.4124 0.0775 0.3613 0.0955 0.6410 0.0163 0.6494 0.1234 0.6531 0.1327

Table III. Spectrophotometric data from MTT assays concerning the cytotoxic activity of complexes on K-562 cells in comparison with the inorganic salts. Compound MTT-formazan absorption at 580 nm untreated control Ce(NO3)3.6H2O La(NO3)3.6H2O Nd(NO3)3.6H2O Ce(HN)3.4H2O La(HN)3.4H2O Nd(HN)3.6H2O 0.4694 0.030 0.4694 0.030 0.4694 0.030 0.3168 0.020 0.3168 0.020 0.3168 0.020 25 M 0.3846 0.040 0.3906 0.020 0.3895 0.030 0.3791 0.030 0.4971 0.030 0.3720 0.040 100 M 0.3502 0.010 0.3342 0.050 0.3485 0.030 0.2897 0.030 0.3975 0.070 0.4192 0.030 400 M 0.3283 0.020 0.2068 0.010 0.2987 0.020 0.2390 0.020 0.1382 0.020 0.1538 0.020

Table IV. Spectrophotometric data from MTT assays concerning the cytotoxic activity of complexes on THP-1 cells in comparison with the inorganic salts. Compound MTT-formazan absorption at 580 nm untreated control Ce(NO3)3.6H2O La(NO3)3.6H2O Ce(HN)3.4H2O La(HN)3.4H2O 0.3745 0.080 0.3745 0.080 0.3745 0.080 0.3745 0.080 25 M 0.4416 0.010 0.4078 0.010 0.3855 0.010 0.3968 0.020 100 M 0.4312 0.030 0.4217 0.030 0.3828 0.030 0.3763 0.060 400 M 0.4230 0.020 0.4026 0.040 0.2746 0.020 0.2871 0.008

Table V. IR spectral data of Niffcoumar, Niffcoumar sodium salt and La(HN)3.4H2O. Compound C19H15NO6 C19H14Na NO6 La(HN)3.4H2O OH 3 295 3 582 3 432 C=O (lacton) 1 686 1 715 1 696 1 619 1 653 1 651 1 572 1 595 1 597 1 512, 1 456 1 516, 1 456 1 514, 1 446 1 173 1 163 1 181 C=O C=C arom. COH

Table VI. 1H-NMR spectral data (100 MHz; DMSO-d6) of Niffcoumar sodium and its complexes. Substance NaN Ce(HN)3.4H2O La(HN)3.4H2O Nd(HN)3.6H2O H5H8 6.957.40 6.907.40 6.907.35 7.207.60 m m m m H9 5.05 4.40 4.35 6.80 t s t s H10 3.40 3.35 3.30 3.50 d bs bd bs H12 3.05 1.90 2.05 1.90 s s s s H14, 15 7.608.10 7.508.00 7.508.10 7.507.90 d d d d

NaN = C19H14NaNO6 (Niffcoumar sodium salt); HN = C19H15NO6

856 4.1. IR-spectra of Niffcoumar sodium salt complexes The bands appearing in the IR spectrum of the free ligand at 3 582, 1 715, 1 653, 1 595, 1 516, 1 456, 1 163 were shifted in the complexes (table V). The weak band observed at 3 582 cm1 in the spectra of the free ligand shifted to a lower wave number in the complexes. A broad band characteristic of OH of crystalline or coordinated water was observed at 3 432 cm1 in the spectra of all the complexes and had higher intensity than the one of the free ligand. This was attributed to the presence of crystalline water. A band at 1 715 cm1 in the spectra of the ligand could be attributed to the stretching vibrations of the carbonyl group in the lacton ring. The band was stretched from 2030 cm1 to lower wave number values. This might be taken as evidence for the participation of the >C=O group in the coordination process. The band at 1 653 cm1 corresponded to the keto-group. The coordination with oxygen was likely, but we couldnt say which oxygens were coordinated to Ln3+ ions. We supposed that the lacton- and keto-carbonyl oxygen atoms of Niffcoumar were bonded to the metal ion probably as a bidentate ligand. The band at 1 595 cm1 could be related to the stretching vibrations of the conjugated olenic system. The vibrations in the region 1 5161 456 cm1 corresponded to the aromatic system. The band at 1 163 cm1 assigned as C-OH was observed at more or less the same position in the complexes and it might be due to the existence of an OH group in all cases. The data of the IR-analyses conrmed the compositions presented in table I. 4.2. 1H-NMR analysis Proton spectra of the compounds, recorded at 100 MHz in DMSO-d6 as a solvent, conrmed the formation of the complexes. The chemical shifts are given in -scale. The typical chemical shifts of the 1H-NMR spectra in DMSO-d6 solvent are shown in table VI. The typical values of coupling constants in Hz were: H9, 7.0; H10, 7.0; H14, 8.1 and H15, 8.1. The chemical shifts of the H9 and H10 protons varied in the lanthanide complexes because of the shift properties of these metals. It was evident that concerning the La(III) and Ce(III) complexes there was an observable weak negative shift effect on the H9 and H10. The effect in the Nd(III) complex was more complicated. A strong positive shift effect for H9 and a weaker one for H10 existed. 5. Conclusion The experimental data presented in gures24 show that the lanthanide complexes possess certain antitumour
Figure 2. A. Cytotoxic effect of the cerium complex of Niffcoumar on P3HR1 cells after 48 h incubation. B. Cytotoxic effect of the lanthanum complex of Niffcoumar on P3HR1 cells after 48 h incubation. C. Cytotoxic effects of the neodymium complex of Niffcoumar on P3HR1 cells after 48 h incubation.

857 A A

Figure 4. A. Cytotoxic effects of the cerium complex of Niffcoumar on AML derived THP-1 cells after 48 h incubation. B. Cytotoxic effect of the lanthanum complex of Niffcoumar on AML derived THP-1 cells after 48 h incubation.

Figure 3. A. Cytotoxic effect of the cerium complex of Niffcoumar on K-562 cells after 72 h incubation. B. Cytotoxic effect of the lanthanum complex of Niffcoumar on K-562 cells after 72 h incubation. C. Cytotoxic effect of the neodymium complex of Niffcoumar on K-562 cells after 72 h incubation.

potential against the three leukaemic cell lines used. In gure 2 the cytotoxic prole of cerium, lanthanum and neodimium complexes is presented. All three compounds cause similar low cytotoxic effects, even at the highest concentration applied (400 M, about 20% growth inhibition) after 48 h incubation. In gure 3 the effects of the complexes against the relatively resistant cell line K-562 after 72 h incubation are summarized. It is noteworthy that the lanthanum and neodimium compounds exert more pronounced cytotoxic effects in comparison to cerium (gure 3B and 3C). Furthermore the cerium and lanthanum complexes were tested on AML derived THP-1 cells. Both compounds induce similar low cytotoxic effects, but only at the highest applied concentration (gure 4A and 4B). The corresponding lanthanide salts are found to have very low or no activity as shown in

858 tables IIIV. So far we can conclude that the metal determines the antitumour spectrum of the newly synthesised and characterised complexes. The LD50 has been estimated for Niffcoumar and its cerium complex. When administered i.p. (single dose) the LD50 of Niffcoumar is 600.0800.0 mg/kg b.w. and the LD50 of its complex is 312.5625.0 mg/kg b.w. In both of the cases the highest dose induces 100% mortality. It means that the molar toxicity of the complex is approximately 4.5-fold higher than Niffcoumar toxicity. 6.2. Biological evaluation A laminar ow cabinet (Haereus HV 2436) and a CO2 gassed incubator (Haereus BB16) were used for cell culture. An ELISA-reader (Labsystems Uniskan I) was used for MTT-assay measurements. References
[1] [2] [3] Patonay T., Litkei G., Bognar R., Erdei J., Miszi C., Pharmazie 39 (2) (1984) 8491. Sharada L.N., Ganorkar M.C., Indian J. Chem. 27A (6) (1988) 542544. Bharat K.B., Jaya T.R.V., Ranabaore V., J. Archaeol. Chem. 4 (1986) 3537. Sharada L.N., Ganorkar M.C., Indian J. Chem. 27A (2) (1988) 163165. Kumar B.B., Raju V.J., Ranabaore V., Ganorkar M.C., Orient J. Chem. 3 (1) (1987) 3439. Deng R., Wu J., Long L., Bull. Soc. Chim. Belg. 101 (6) (1992) 439443. Pokhariyal G.P., Proc. Natl. Acad. Sci. India 58A (3) (1988) 369373. Pokhariyal G.P., Indian J. Chem. 28A (10) (1989) 922923. Castellani C.B., Carugo D., Inorg. Chim. Acta 159 (2) (1989) 157161. Dutt N.K., Sarma U.U.M., Indian J. Chem. 15A (4) (1977) 361362. Kostova I., Manolov I., Konstantinov S., Karaivanova M.H., Eur. J. Med. Chem. 34 (1999) 6368. Meyer B.N., Wall M.E., Wani M.C., Teylor H.L., J. Nat. Prod. 48 (6) (1985) 952956. Baraldi P.G., Barco A., Benetti S., Guarner M., Manfredini S., Pollini G.P., Simoni D., Tetrahedron Lett. 26 (43) (1985) 53195322. Harvey R.G., Corter C., Ananthanarayan T.P., Schmolka S., J. Org. Chem. 53 (17) (1988) 39363943. Haiduc I., Silvestru C., Coord. Chem. Rev. 99 (1990) 253296. Anghileri L.J., Crone-Escanye M.C., Anticancer Res. 7 (1987) 12051208. Mossmann T., J. Immunol. Methods 65 (1983) 5563.

6. Experimental protocols 6.1. Chemistry The carbon, hydrogen and nitrogen content of the compounds were determined by elemental analysis. TLC: Al-foils Kieselgel 60 F254, Merck (FRG), developed by cyclohexane-chloroform-acetic acid (10:10:4, vol. parts), detection: UV 254 nm. M.p., Bchi 510 apparatus (Switzerland), uncorrected. The water content was determined using a Metrohn Herizall E55 Karl Fisher Titrator and the presence of sodium ions was checked by means of ame photometry. Conductometric measurement was carried out at 25 C on 103 M solutions in DMSO, by using a Metrohm 660 AG CH-9101 Herisau conductometer with a platinum electrode and a cell having a constant of 0.79 cm1. IR spectra (Nujol) were recorded on an IR-specrometer FTIR-8101M Shimadzu. 1H-NMR spectra were recorded at room temperature on a Brucker WP 100 (100 MHz) spectrometer in DMSO-d6. Chemical shifts are given in ppm.

[4] [5] [6] [7] [8] [9] [10] [11] [12] [13]

[14] [15] [16] [17]

Eur. J. Med. Chem. 34 (1999) 859865 1999 Editions scientiques et mdicales Elsevier SAS. All rights reserved

859

Short communication

Vasodilating and antiarrhythmic activity of heteryl lactones

Ludmila Leitea*, Daina Jansonea, Maris Veverisa, Helena Cirulea, Yuris Popelisa, Gagik Melikyanb, Anna Avetisyanb, Edmunds Lukevicsa
b a Latvian Institute of Organic Synthesis, 21 Aizkraukles Str., LV-1006 Riga, Latvia Department of Organic Chemistry, Yerevan State University, Yerevan 375049, Armenia

(Received 27 October 1998; revised 22 March 1999; accepted 23 March 1999)

Abstract A new series of unsaturated - and -lactones with pyridyl, quinolyl and nitrophenyl substituents (9, 10) have been synthesized by the condensation of unsaturated methyl lactones with heteryl aldehyde or nitrobenzaldehyde in the base-catalysed aldol reaction. The antiarrhythmic, vasodilating and cardiotonic activities of the synthesized compounds have been studied in vivo and ex vivo. 3-Cyano-5,5dimethyl-4-[4'-(4-pyridyl)-1',3'-butadienyl)]-2(5H)-furanone (9e) displayed a signicant vasodilating activity. The antiarrhythmic activity of this compound was higher, but its toxicity lower than that of the procainamide reference drug. Five-membered lactones, particularly 3-cyano-4-(4-pyridylvinyl)-5,5-dimethyl-2(5H)-furanone (9c), exhibited a remarkable cardiotonic activity. The replacement of a pyridyl substituent by a nitrophenyl group in the pyranone derivative did not change the cardiovascular activity and toxicity. 1999 Editions scientiques et mdicales Elsevier SAS heteryl lactones / antiarrythmic activity / vasodilating activity / cardiotonic activity

1. Introduction Some of the synthetic and natural compounds containing an unsaturated ve-membered lactone moiety exhibit a cardiotonic activity [1]. For example, the heart glycosides contain a -lactone unit. On the other hand, it should be noted that the majority of cardiotonic substances of a novel generation is derived from N-heterocyclic compounds [2]. To determine the role of lactone units and heteryl substituents as pharmacophores responsible for antiarrhythmic and cardiotonic activity, pyridyl, quinolyl and nitrophenyl derivatives of unsaturated and -lactones have been prepared and tested in vivo and ex vivo. 2. Chemistry Pyridyl, quinolyl and nitrophenyl derivatives of unsaturated - and -lactones were synthesized according to the procedure elaborated for the pyridyl derivatives of furanone and pyranone [3]. Methyl lactones 7 and 8 were
*Correspondence and reprints

condensed with aldehydes (16) in the presence of NaOH as a catalyst (gure 1). The unusual reactivity of 2-, 3and 4-pyridinecarboxaldehyde with - and -lactones in a base-catalysed aldol reaction was shown earlier [3]. It was found that the synthesis of heteryl lactones 9ac and 10ac was accompanied by the formation of the [bis(2oxo-3-cyano-5,5-dimethyl-2(5H)-furanyl-4-methyl)-methyl]pyridines and [bis(2-oxo-3-cyano-6,6-dimethyl-5,6-dihydropyranyl-4-methyl)-methyl]pyridines. To avoid this addition we used a 2-fold excess of pyridine carboxaldehyde. The aldol condensation reaction of aldehydes 46 with lactones occurred traditionally and yielded the corresponding unsaturated derivatives of lactones only. Lactone 9d was synthesized by the reaction of the ethoxycarbonyl derivative of methyl furanone 7b with aldehyde 2. The ester was converted into the sodium salt by the additional amount of sodium hydroxide. The reaction of aldehyde 4 with furanone 7a was carried out in the presence of a catalytic amount of NaOH at room temperature to give furanone 9e in 21% yield. This was signicantly lower than in the case of aldehyde 3 condensation reaction (yield 54%). Raising the temperature did not increase the yield. In all cases the values of

860

Figure 1. Synthesis of heteroaryl lactones.

spin-spin coupling constants of the double bond CH=CH protons (1617 Hz) indicated that only E- or E,E-isomers were isolated. The coupling constant of ,-protons in chain CH=CHCH=CH (compound 9e) equalled 10.7 Hz. A similar value of coupling constant (10.4 Hz) is in agreement with the E-isomer of butadiene [4]. To obtain an additional proof for the conguration of the unsaturated synthesized compounds, E-Z photoisomerization caused by UV irradiation of the condensation products 9a and 10b, d and e was studied. The isomerization process was controlled by electron absorption spectroscopy. When the solution in ethanol was irradiated with UV light the intensity of the E-isomer absorption band at 332344 nm decreased and a band at 220285 nm appeared. According to NMR data the formation of the E- and Z-isomer mixture caused such changes in spectra.

The spin-spin coupling constants of the double bond CH=CH protons in isomers appeared during irradiation and were 1113 Hz. This means that the compounds described in the present article really have E-conguration. The characteristics of the compounds are listed in tables I and II. 3. Results and discussion The proposed cardiovascular activity of newly synthesized compounds was tested in ex vivo experiments on the isolated rabbit ear artery and in vitro on anaesthetized laboratory animals. These models were chosen in order not to miss the compounds with antiarrhythmic, vasodilating and/or cardiotonic activities. Table III summarizes the data of the antiarrhythmic screening test and acute

861
Table I. Reaction conditions and characteristics of compounds 9 df and 10 d and e. Compound 9d 9e 9f 10d 10e
a

Reaction temperature (C) 80 80 20 20 80

Reaction time (h) 4 2 1.5 3 1.2

Yield (%) 52 21 41 44 61

M.p. (C) 275-280 161-163 229-230 222-224 (dec) 224-226

Formulaa C14H12NO4Na C16H14N2O2 C18H14N2O2 C19H16N2O2 C16H14N2O4

all compounds were analysed for C, H and N.

toxicity obtained for heteryl lactones. The acute toxicity of the studied compounds was low (LD50 was over 400 mg/kg). Only the acute toxicity of compound 9c (LD50 180 mg/kg) and 9d (200 mg/kg) was comparable with that of lidocaine (LD50 238 mg/kg). The vemembered pyridyl lactones 9 ae had a toxicity similar to its six-membered analogues except 3- and 4pyridylfuranones, the toxicity of which was about 2-fold higher than that of the substituted pyranones. The replacement of pyridyl substituents for a quinolyl group appeared slightly less toxic (see 9c and f, 10c and d). Heteryl derivatives 9c, e and 10d injected i.p. in doses of 15 and 30 mg/kg caused the statistically signicant protection against CaCl2-induced arrhythmia. Their activity was higher but toxicity lower than that of the reference drug procainamide. The antiarrhythmic activity was more marked for heteryl lactones with 4-pyridyl substituent within the furanone series. Compounds 9b and c were also investigated on anaesthetized rats. It was shown (table IV) that these pyridyl furanones decreased the number of animals with lethal arrhythmia induced by calcium chloride, but to a lower extent than lidocaine. Compound 9c (dose = 1 mg/kg,

i.v.) also protected against ventricular ectopic beats. So, we established that the 4-isomer of pyridyl furanone, 9c, revealed the most pronounced antiarrhythmic activity and the highest toxicity among all compounds studied (table III). Almost all of the studied heteryl derivatives of unsaturated -and -lactones caused a more or less marked relaxation of the previously contracted vessels of the rabbit ear (table V). Compound 9e exhibited a pronounced vasodilating activity. At the same time, pyridyl lactone 9a demonstrated two phases of vasodilation/ vasoconstriction but 9b showed a trend towards vasoconstriction. In this test, six-membered lactones (10a, b and c) possessed more pronounced vasodilation than vemembered ones (9a, b and c) and 10a (2-pyridyl) was more active than 10b and c (3- and 4-isomers). In the series of pyridyl lactones, we found a tendency of 2-pyridyl substituents to induce a more selective vasodilation in comparison with 3- and 4-pyridyl substituents, as already stated with calcium channel modulators [5]. Compounds 9e, 9f and 10d induced a signicant vasodilation at a concentration of 50 M (9e being the most active). This means that these compounds may show a

Table II. 1H NMR chemical shifts (, ppm) and spin-spin coupling constants of compounds 9df and 10d and e. Compound 9d* 9e** 9f** 10d* CH=CH (JHH Hz) 7.52, d (16.6); 7.31, d (16.6) 6.54, d (15.7); 7.01, d (15.7); 7.14, dd (10.7, 15.7); 7.55, dd (10.7, 15.7) 7.06, d (16.4); 8.60, d (16.4) 7.54, d (15.7); 8.42, d (15.7) Pyrone CH2 3.35, s Pyrone or furanone CH3 1.54, s 1.65, s 1.75, s 1.50, s Phenyl, pyridyl or quinolyl CH (JHH Hz) 8.74, d, H-2, (1.4); 8.49, dd (4.2, 1.4); 8.00, dt, H-4, (8.2, 1.4); 7.40, dd, H-5 (8.2, 4.2) 7.34, m, H-3, 5 (6.5); 8.66, m, H-2, 6 (6.1) 7.65, d, H-3, (4.4); 7.70, m, H-6 (1.4; 6.9; 8.5); 7.82, m, H-7 (1.4; 6.9; 8.5); 8.10, d, H-5 (8.5); 8.20, d, H-8 (8.5); 9.01, d, H-2 (4.4) 7.74, m, H-6, (1.2, 7.8, 7.9); 7.86, m, H-7 (1.2, 7.8, 7.9); 7.93, d, H-3 (4.6); 8.11, d, H-5 (7.8); 8.50, d, H-8 (8.0); 9.00, d, H-8 (8.0); 9.00, d, H-2 (4.6) 7.70-8.35, m, H-4

10e** DMSO-d6,
**

7.24, d (16.0); 7.59, d (16.0) CDCl3

2.86, s

1.55, s

862
Table III. Antiarrhythmic activity and acute toxicity of pyridyl lactones in mice. Compound Dose Antiarrhythmic (mg/kg, i.p.) activity (Y/N)a 30 90 30 90 15 30 15 30 15 30 90 30 90 30 90 30 90 30 90 15 30 90 30 90 1/5 2/5 1/5 1/5 3/5* 3/5* 1/5 2/5 3/5* 2/5 2/5 1/5 2/5 1/5 2/5 1/5 1/5 1/5 1/5 1/5 3/5* 2/5 1/5 1/5 1/5 3/5* 2/5 4/5* 0/10 Acute toxicity LD50 (mg/kg, i.p.) > 400 > 400 180 (138.5-234) 200 (153.8-260) 480 (347.8-662.4) > 600 > 400 > 400 > 400 > 600 600 360 (257-504) 238 (180.3-314) Table V. Vasodilating activity of the investigated compounds in rabbit ear artery. Compound 9a 9b 9c 9e 9f 10a Concentration (M) 10 50 10 50 10 50 10 50 10 50 10 50 10 50 10 50 10 50 10 50 1 10 50 Relaxationa (%) 5(-3) 10(-5) -2 -8 -3 -5 30* 50* 20 30* 12 27* 15 20 2 15 19 32* 5 15 8 35* 76* 0 8
*

Table IV. Effect of compounds 9b and c on CaCl2-induced arrhythmia and lethality in rats. Compound 9b Dose (mg/kg, i.v.) 0.3 1.0 3.0 0.3 1.0 3.0 0.3 1.0 3.0 Arrhythmia-scores (M SD) 3.3 0.4 3.0 0.5 2.8 0.5 2.9 0.5 2.3 0.3* 2.6 0.5 3.0 0.4 2.1 0.3* 1.8 0.4* 3.8 0.5 Survival (%) 20 40 60* 40 60* 60* 20 60* 80* 0

9a 9b 9c 9d 9e

9c

Lidocaine

9f 10a 10b 10c 10d

P < 0.05 vs. control.

arterial blood pressure by 18%. At the same time, the heart rate remained unchanged. The positive inotropic effect was observed with 9b and 9c at doses of 0.12.0 mg/kg (duration of action was slightly shorter for

10e

Procainamide 30 90 Lidocaine 15 30 Control

a

Y, the number of mice protected against CaCl2-induced arrhythmia; N, total number of experimental animals; *P < 0.05 vs. control group.

strong TXA2 receptor antagonist activity in the rabbit blood vessels. The replacement of a 4-pyridyl substituent of furanone and pyranone (9c, 10c) for a 4-quinolyl group (9f, 10d) patently increased vasodilating properties in both cases. Unlike pyridyl derivatives of furanone and pyranone a nitrophenyl pyranone 10e did not possess the vasodilating activity. Cardiotonic and vasodilating properties of heteryl lactones were also studied on anaesthetized cats. Maximum variations in the haemodynamic parameters were registered about 15 min after i.v. administration. The results obtained are shown in table VI. The compounds 9b and c possess a signicant cardiotonic activity. Compound 9c (dose = 0.5 mg/kg) increased the systolic pressure of the left ventricle by 30%, but dP/dt by 32%, and the mean

10b 10c 10d 10e Papaverine

Control Solvent

a positive value means relaxation; negative one, vasoconstriction; *P < 0.05 vs. control.

863
Table VI. Haemodynamic effects. Results are in % (in reference to the control initial value). Compound 9a 9b Dose (mg/kg, i.v.) 1.0 3.0 0.1 0.5 2.0 0.1 0.5 2.0 0.1 0.5 2.0 0.1 0.5 2.0 0.02 0.1 MABPa 5 -205 0 12 29 3 18 30 0 0 6 -3 5 10(-5) -8 -35 HRa 0 -5 -5 3(-10) 5 -2 0 6 0 0 -3 0 0 5(-5) -5 -12 LVSP 6 23 45 10 30 50 0 0 5 0 5 7 -3 -9 dP/dt 8 26 33 15 32 38 0 5 8 0 7 10 -3 -16 FBFa 0 10 12 -5 12 15 0 0 5 0 0 5(-5) 18 45 CBFa 12 -512 0 12 22 3 16 20 0 3 5 0 6 3 15 80

9c

10b

10c

Verapamil (hydrohloride)

MABP, mean arterial blood pressure; HR, heart rate; LVSP, left ventricular systolic pressure; dP/dt, left ventricular contractility; FBF, femoral artery blood ow; CBF, carotid artery blood ow. a Haemodynamic parameters are expressed as % change: predrug/postdrug; increase/decrease; positive value means increase; negative one, decrease.

9b). However, 9b caused a more marked change in heart rate (within +5 to 10%), compared with 9c. Both pyridyl furanones 9b and c similarly increased the blood ow in the femoral and carotid arteries, but this increase was less than the change in the blood pressure. The vasodilating effect of a nitrophenyl substituent was similar to that of a pyridyl one. The vasodilating activity of pyranone with a nitrophenyl substituent is similar to that of pyridyl derivatives of furanone and pyranone. 4. Conclusions New derivatives of - and -lactones with pyridyl, quinolyl and phenyl substituents with E-conguration were synthesized, their antiarrhythmic, vasodilating and cardiotonic activities investigated. The acute toxicity of the compounds studied was low. On the whole, the ve-membered pyridyl lactones had a toxicity similar to their six-membered analogues (except 3- and 4-pyridylfuranones). The pyridyl substituents of pyridyl lactones may be responsible for a higher toxicity in comparison with quinolyl lactones. Pyridyl and quinolyl lactones caused a signicant protection against CaCl2-induced arrhythmia. The activity depended on the position of a substituent in the

heterocycle. The most active were 4-isomers 3-cyano4-(4-pyridylvinyl)-5,5-dimethyl-2(5H)-furanone 9c and 3-cyano-4-(4-quinolylvinyl)-6,6-dimethyl-(5,6-dihydro)2-pyranone 10d. The replacement of a 4-pyridyl substituent of furanone for a 4-quinolyl group decreased the antiarrhythmic activity. The similar replacement in the case of 4-pyridylpyranone caused the activity increase. The vasodilating activity depended on the ring size of the lactone, on the type of heteryl substituent, and on the position of the substituent in the ring. Six-membered lactones revealed more pronounced vasodilation than ve-membered ones. 2-Pyridylfuranone and 2-pyridylpyranone demonstrated more marked vasodilation than 3and 4-isomers. Besides, this 2-pyridylfuranone also possessed a vasoconstricting action. The insertion of the second vinyl group in the aliphatic chain between the pyridine and lactone rings increased the activity of pyridyl furanone. Thus 3-cyano-5,5-dimethyl-4-[4'-(4pyridyl)-1',3'-butadienyl)]-2(5H)-furanone 9e was the most active among the compounds studied. The vasodilating effect of a nitrophenyl substituent is similar to that of a pyridyl substituent. Cardiotonic activity is more pronounced for 3- and 4-isomers of pyridyl furanones. Cardiotonic activity of the corresponding 6-membered lactones is lower.

864 5. Experimental protocols 5.1. Chemistry 5.1.1. Methods The 1H NMR spectra were recorded on a Bruker WH-90/DS and on an AM-360 (360 MHz) spectrometers in CDCl3 or DMSO-d6, TMS internal standard. The mass spectra were obtained on a Kratos MS-25 chromatograph-mass-spectrometer with an ionizing energy of 70 eV. Elemental analysis was performed on an Elemental Analyzer Carlo Erba 1108. Silufol UV-254 plates were used for TLC analysis, eluents benzeneacetone 3:1. The melting points were determined on a Boetius stage and reported without corrections. E-Z photoisomerization has been studied in ethanol solution of compounds 9a, 10b, d and e by irradiation with UV light at 336, 332, 340 and 344 nm respectively for 45 min. Electronic absorption spectra were recorded on a Specord UV-VIS spectrometer. Starting furanones 7 and pyrone 8 were synthesized according to [6] and [7], respectively. Pyridine aldehydes 13 were synthesized by vapour phase oxidation of methyl pyridines over V2O5MoO3 catalyst [8]. Starting 1'-(4-pyridyl)acroleine 4 and 4-quinolinealdehyde 5 were synthesized according to the methods described in [9] and [10]. All solvents were of an analytical grade and used without further purication. 5.1.2. General procedure for the synthesis of compounds 9df, 10d and e A mixture of methyl lactone 7 or 8 (20 mmol), the aldehyde 46 (20 mmol) and NaOH (1.25 mmol) in MeOH (20 mL) was stirred at room temperature or reuxed for 1.24 h. The condensation products precipitated, and were ltered off after cooling to room temperature. The precipitates of compounds 9e, f, 10d and e were recrystallyzed from EtOH. Compound 9d was synthesized as described above but when the condensation was completed NaOH (21 mmol) was added to the cooled mixture. The resulting precipitate was ltered off, washed with EtOH and air-dried. The reaction conditions, yields and NMR data are shown in tables I and II. Procedures for the synthesis of compounds 9ac and 10ac and their chemical and physical data are given in [3]. 5.2. Pharmacological methods 5.2.1. Experiments ex vivo The modied method for the experiments on the isolated perfused rabbit ear blood vessels was used [11, 12]. Male and female rabbits (2.63.3 kg) were euthanazed by i.v. injection of Na pentabarbital (80 mg/kg). The central ear artery was dissected free at the ear base and cannulated with polyethylene tube and perfused at a constant ow from a 4-channel peristaltic pump, Geminy (Italy). The perfusion uid was (mmol): NaCl 136.9; KCl 2.68; CaCl2 1.8; MgCl2 1.05; NaHCO3 11.9; NaH2PO4 0.42; glucose 5.6 (pH 7.35). Intraluminal inow perfusion pressure was measured with a Statham P23I transducer and recorded on a physiograph DMP-4B (Narco Bio-System, Huston, USA). As ow remained constant, the alterations in perfusion pressure reected changes in the blood vessel resistance, i.e., the degree of vasoconstriction or relaxation. Vasoconstriction was caused by intraluminal infusion of U-46619 (thromboxane A2 receptor agonist). The relaxant responses to the investigated compounds used in different concentrations were tested. The responses were expressed as a per cent relaxation (% changes in perfusion pressure) without and with the investigated compounds or solvent. 5.2.2. Experiments in vivo 5.2.2.1. Antiarrhythmic activity 5.2.2.1.1. Antiarrhythmic screening test Antiarrhythmic activity was tested in the experimental antiarrhythmic screening model on male ICR:JCL mice (1923 g) as was described earlier [13]. The tested compounds or solvents were administered i.p. 15 min before i.v. infusion of 2% CaCl2 solution at a constant rate (0.02 mL/sec) in a dose of 180 mg/kg. The number of animals protected from CaCl2-induced lethal arrhythmia was dened. The investigated compounds were dissolved in NaCl 0.9% solution or in dimethyl acetamide and then diluted with the NaCl 0.9% solution. 5.2.2.1.2. Calcium chloride-induced arrhythmia. The experiments were performed according to the classical method [14]. In brief, Wistar male rats (180220 g) were anaesthesized with urethan (1.20 mg/ kg, i.p.). ECG was registered in II standard lead on a physiograph DMP-4B (Narco Bio-Systems, Huston, USA). The 5% CaCl2 solution was i.v. injected at a dose of 180 mg/kg. Heart rhythm disturbances in scores [15] and lethality of animals were estimated. Solutions of the compounds to be tested were administrated into the femoral vein 3 min prior to CaCl2. Every dose was tested on ve rats. 5.2.2.2. Cardiotonic action and hypotensive activity The method used was described already in [16]. Adult mongrel male and female cats (3.04.2 kg) were anaesthesized with -glucochloralose and urethan (80 and 200 mg/kg, i.p.). The trachea was catheterized and con-

865 nected to an intermittent positive-pressure respirator (DP-8, SU). Arterial blood pressure from the right femoral artery, left ventricular pressure (electromanometer P 23 ID) and the rst derivative (dP/dt) were registered on a physiograph DMP-4B (Narco Bio-Systems, Houston, USA). The blood ow from the left femoral and left common carotid arteries was measured with an electromagnetic owmeter MFV-1200 (Nihon Kohden, Tokyo, Japan). The heart rate was calculated using the blood pressure wave. The solutions of the compounds investigated were injected through a catheter placed in the femoral vein. 5.2.2.3. Toxicological examination Acute toxicity of the unsaturated ve- and sixmembered lactones was studied in albino male and female ICR:JCL mice (1822 g). Solutions or suspensions of the compounds (prepared with 0.6% Tween-80) were injected i.p. The experimental animals were observed for 10 days. Common state, mobility of animals, toxic symptoms and survival were estimated. To reduce the number of used animals, the maximal dose was 400 mg/kg. If possible, LD50 was calculated when 50% of the animals died. All references drugs (Procainamide, Lidocaine, Verapamil, Papaverine) were from commercial sources. acknowledge with gratitude and affection the nancial help they have also received from the Taiho Latvian Foundation. References
[1] [2] [3] Avetisyan A.A., Tokmadzhyan G.G., Chem. Heterocycl. Compds. 23 (1987) 595609. Yuzhakov S.D., Mastafanova L.I., Mashkovski M.D., Yahontov L.N., Khim-Farm. Zh. 26 (1992) 417; CA 117 (1992) 123832q. Jansone D., Leite L., Fleisher M., Popelis J.U., Mazheika I., Lukevics E., Melikyan G., Avetisyan A., Chem. Heterocycl. Compds. 34 (1998) 267270. Pretsch E., Clerc T.H., Simon W., Tables of Spectral Data for Structure Determination of Organic Compounds, Springer, 1989, p. H210. Ramesh M., Matowe W.C., Akula M.R., Vo D., Dagnio L., MoyCheong L.K.K., Wolowyk M.W., Knaus E.E., J. Med. Chem. 41 (1998) 509514. Pavlova L.A., Belgorodski V.V., Venus-Danilova E.D., Zh. Obsh. Khim. 36 (1966) 13861391; CA 66 (1967) 10797f. Avetisyan A.A., Kasparyan B.K., Dzandzapanyan A.N., Dangyan M.T., Arm. Khim. Zh. 36 (1983) 341343; CA 99 (1983) 139700r. Shymanska M.V., Leitis L.J., Iovele I.G., Goldberg Ju.Sh., Skolmeistere R.A., Golender L.O., in: Krilov O. (Ed.), Problems of Kinetics and Catalysis. XIX. Partial Oxidation of Organic Compounds, Nauka, Moscow, 1985, pp. 175187. Leitis L.J., Rubina K.I., Goldberg Yu.Sh., Jansone D.P., Shymanska M.V., Izv. AN LatvSSR. Ser. khim. 4 (1980) 469472; CA 94 (1981) 14734v. Achremowicz L., Synthetic Commun. 26 (1996) 16811684. Blattner R., Classen H.G., Dehnert H., Doring H.G., Experiments on Isolated Smooth Muscle Preparations, Hugo Sachs Elektronik KG (BRD), 1980. Wilson S.K., Steinsland O.S., Nelson S.H., J. Cardiovasc. Pharmacol. 23 (1994) 127135. Lukevics E., Veveris M., Dirnens V., Appl. Organomet. Chem. 11 (1997) 805811. Yoshidomi M., Sukamoto T., Morita T., Ito K., Nose T., Arzneim.-Forsch./Drug Res. 32 (1982) 10561059. Rees S.A., Tsuchihashi K., Hearne D.J., Curtis M.J., J. Cardiovasc. Pharmacol. 22 (1993) 343349. Veveris M., Syomin R., Proc. Latv. Acad. Sci. B 9 (1992) 9096; CA 118 (1993) 139512s.

[4]

[5]

[6] [7] [8]

[9]

[10] [11]

[12]

Acknowledgements We are grateful to the Latvian Science Council for the nancial support of the programme The Development of Modern Directions in the Organic Chemistry for the Production of New Medical Agents. The authors

[13] [14] [15] [16]

Eur. J. Med. Chem. 34 (1999) 867875 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

867

Short communication

New amino derivatives of 1,2,3-triazolo[4,5-d]pyrimidines and their affinity towards A1 and A2A adenosine receptors
Laura Bettib, Giuliana Biagia, Gino Giannaccinib, Irene Giorgia, Oreste Livia*, Antonio Lucacchinib, Clementina Maneraa, Valerio Scartonia
b a Dipartimento di Scienze Farmaceutiche, Facolt di Farmacia, Universit di Pisa, via Bonanno 6, 56126 Pisa, Italy Dipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotecnologie, Universit di Pisa, via Bonanno 6, 56126 Pisa, Italy

(Received 19 December 1998; revised 7 April 1999; accepted 21 April 1999)

Abstract Starting from the appropriate azides (4-chlorobenzyl-, 2-thiophenemethyl-, 2-uorobenzyl-, and 4-uorobenzylazides) in which the variation of the substituent is at the basis of the four series of derivatives (ad), the 7-aminosubstituted 1,2,3-triazolo[4,5-d]pyrimidines 4 were prepared by a well known synthetic route. The biological activity of compounds 4 was expected on the basis of the presence of particular substituents on N(7), and these substituents were introduced by the reaction of the 7 lactamic carbonyl function, present on precursors 3, with cycloalkyl-, aralkyl- and arylamines. Radioligand binding assays at bovine brain adenosine A1 and A2A receptors showed that some compounds possessed a high affinity and selectivity for the A1 receptor subtype. Furthermore, biological results indicated that the p-chlorobenzyl substituent lowered receptor binding, compared with the previously prepared benzyl and 2-chlorobenzyl derivatives, suggesting certain particular steric requirements of the lipophilic region which interacts with the benzyl substituent. The thiophenemethyl substituent conferred more activity than the benzyl one. The presence of a uorine atom on the benzyl group determined a high affinity, especially when it was in the ortho position. Compounds 4c.1 (R = 2-uorobenzyl, R = cyclopentyl, Ki = 10.5 nM), 4c.2 (R = 2-uorobenzyl, R = cyclohexyl, Ki = 19.5 nM) and 4d.1 (R = 4-uorobenzyl, R = cyclopentyl, Ki = 26 nM) were the most active for A1 receptors. 1999 ditions scientiques et mdicales Elsevier SAS 1,2,3-triazoles / 1,2,3-triazolo[4,5-d]pyrimidines / A1-adenosine and A2A-adenosine receptor antagonists

1. Introduction Our previous studies on 7-aminosubstituted 1,2,3triazolopyrimidines [1] indicated that certain compounds, bearing a benzyl or 2-chlorobenzyl as the lipophilic substituent in the 3 position, showed a high affinity and selectivity towards the A1 receptor subtype. These and other studies of ours on 1,2,3-triazolopyrimidines [26], in accordance with the receptor binding site models proposed for adenosine agonists and antagonists [7], revealed three lipophilic binding regions (corresponding to the N-6, C-2 and N-9 positions of the azapurine ring) in the A1 adenosine receptors, arranged as

a fan-shape with respect to the NH function, which was engaged as a hydrogen bond donor. On the basis of our previous results, the investigations into these structures were continued in order to study the mode of binding of these compounds with A1-receptors by comparative SAR analysis. We report here the synthesis and biological evaluation of a series of 1,2,3triazolo[4,5-d]pyrimidines (4) characterized by new benzyl-type substituents (R = 4-chlorobenzyl, 2-thenyl, 2-uorobenzyl and 4-uorobenzyl). In the 7 position, amino substituents which conferred a high biological activity to previously prepared molecules (cyclopentyland cyclohexylamino, meta- and para-toluidino, -methylbenzylamino, amphetamino) as well as other amino substituents (isopropylamino, 2-butylamino, 2-pentylamino, 3-pentylamino, p-bromoanilino, furfurylamino, 2-pyridylamino and 2-hydroxy-cyclohexylamino) were introduced.

*Correspondence and reprints

868 lipophilic substituent in the 3 position: a (4chlorobenzyl), b (2-thiophenemethyl), c (2-uorobenzyl) and d (4-uorobenzyl) (table I). The structures of all the new compounds were assigned on the basis of the well-known reaction mechanisms already regularly carried out in our laboratory: 1,3dipolar cycloaddition of azides to activated methylenic compounds, formation of the pyrimidine ring, chlorination and nucleophilic displacement of the halogen by amines. The structures were also conrmed by analytical and spectroscopic data. 3. Biochemistry The 7-(aminosubstituted)-1,2,3-triazolo[4,5-d]pyrimidines were tested in radioligand binding assays for affinity at A1 and A2A adenosine receptors in bovine brain cortical membranes and in bovine brain striatal membranes respectively. [3H]R-(-)-N6-cyclohexyl-adenosine (CHA) was used as the A1 radio-ligand and [3H]-2-{[[p(2-carboxyethyl)phenyl]ethyl]amino}-5-(N-ethylcarbamoyl)adenosine (CGS 21680) as the A2A radioligand. The experimental details of the receptor binding assays were reported in a previous paper [1]. 4. Results and discussion 2. Chemistry The scheme for the preparation of compounds 4 started from the appropriate azide, which determined the R substituent (series ad) and followed a wellexperimented synthetic route (gure 1). All the azides have been reported in the literature (4chlorobenzylazide, [8]; 2-thiophenemethylazide, [9]; 2-uorobenzylazide [10]; and 4-uorobenzylazide [11]) but preparation of 2-thiophenemethylazide has not been described. The appropriate azide was reacted with cyanoacetamide in the presence of sodium ethoxide in reuxing ethanol to give the 1-substituted-4-carboxamido-5amino-1H-1,2,3-triazoles 1ad in good yields. These compounds, by heating with formamide, easily cyclized to the corresponding 7-hydroxy-triazolopyrimidines 2ad which, by reaction with thionyl chloride in chloroform, provided the expected chloroderivatives 3ad. Compounds 1a, 2a and 3a have been described in the literature [8]. The chlorine atom in the 7 position was reactive enough to undergo a nucleophilic displacement by primary aliphatic or aromatic amines (consequently, several triazolopyrimidine derivatives 4 were prepared, corresponding to the four series characterized by the The results of the A1 and A2A adenosine receptor binding assays, expressed as inhibition constants (Ki, nM) in table II show that introduction of the p-chlorobenzyl substituent on the triazole ring in the 3 position (compounds 4a.17) reduces A1 and A2A adenosine receptor binding compared with the 2-chlorobenzyl substituent [1]. The azapurine compound 4a.1, a cyclopentylamino derivative, is an exception, with an A1 affinity constant of Ki = 128 nM, followed by the cyclohexylamino derivative 4a.2 with a Ki = 1 360 nM. Both compounds showed a high receptor selectivity (table II); the inhibition percentages of the A2A adenosine receptor binding assays were so low that the corresponding Ki values were not calculated. The shift of the chlorine from the ortho to the para position suggested a moderate and well-dened extent of the lipophilic region which receives the benzyl substituent. Whereas the orthochlorine atom on the benzyl group increased the affinity towards the A1 receptor [1], the results of the 4a series indicated that the chlorine atom in the para position decreased the capacity to bind with the receptor site. This fact could mean that the para-chloro substitution generally caused a steric repulsion within the receptor, owing to the limited depth of the lipophilic site.

Figure 1. Synthetic route for compounds 4.

869
Table I. Physicochemical data of triazolopyrimidine derivatives and intermediates. Compound Yield% 1b 1c 1d 2b 2c 2d 3b 3c 3d 4a.1 4a.2 4a.3 4a.4 4a.5 4a.6 4a.7 4b.1 4b.2 4b.3 4b.4 4b.5 4b.6 4c.1 4c.2 4c.3 4c.4 b 4c.5 4c.6 4c.7 4c.8 4c.9 4c.10 4c.11 4c.12 4c.13 4c.14 4d.1 4d.2 4d.3 4d.4 4d.5 4d.6 4d.7 4d.8 4d.9 4d.10 4d.11 4d.12
a

crystall. solvent EtOH EtOH EtOH EtOH EtOH EtOH 6080 C petr. ether 6080 C petr. ether 6080 C petr. ether MeOH 6080 C petr. ether EtOH EtOH EtOH EtOH MeOH MeOH MeOH EtOH MeOH EtOH EtOH MeOH MeOH EtOH EtOH EtOH MeOH EtOH EtOH EtOH EtOH MeOH EtOH EtOH MeOH MeOH 6080 C petr. ether EtOH EtOH EtOH 100140 C petr. ether EtOH EtOH Toluene EtOH/H2O 6080 C petr. ether 6080 C petr. ether

M.p. C 197198 186188 221223 225230 221224 225228 7980 8588 7476 115116 109111 110112 158161 175176 163164 115117 120122 132135 148152 139142 149151 170172 109112 107109 159162 181185 139142 139141 153156 170173 196198 182185 127129 160165 167170 107109 121124 115118 154157 184186 139141 142144 195198 205208 214217 99101 8284 9092

Molecular formula (molecular weight) C8H9N5OS (223.25) C10H10N5OF (235.22) C10H10N5OF (235.22) C9H7N5OS (233.25) C11H8N5OF (245.22) C11H8N5OF (245.22) C9H6N5SCl (251.69) C11H7N5FCl (263.66) C11H7N5FCl (263.66) C16H17N6Cl (328.80) C17H19N6Cl (342.83) C19H17N6Cl (364.84) C20H19N6Cl-HCl (415.33) C18H15N6Cl (350.81) C18H15N6Cl (350.81) C17H12N7O2Cl (381.78) C14H16N6S (300.38) C15H18N6S (314.41) C18H18N6S-HCl (386.90) C17H16N6S-HCl (372.88) C16H14N6S (322.39) C16H14N6S (322.39) C16H17N6F (312.35) C17H19N6F (326.38) C20H19N6F-HCl (398.87) C20H19N6F-HCl (398.87) C18H15N6F (334.36) C18H15N6F (334.36) C17H19N6OF (342.38) C16H13N6OF (324.32) C17H12N6BrF (399.22) C16H12N7F-HCl (357.78) C14H15N6F (286.31) C15H17N6F-HCl (336.80) C16H19N6F-HCl (350.83) C16H19N6F (314.37) C16H17N6F (312.35) C17H19N6F (326.38) C20H19N6F-HCl (398.87) C18H15N6F (334.36) C18H15N6F (334.36) C17H19N6OF (342.38) C16H13N6OF (324.32) C17H12N6BrF (399.22) C16H12N7F (321.32) C15H17N6F (300.34) C16H19N6F (314.37) C16H19N6F (314.37)

MS M+ 223 235 235 233 216 245 251 263 263 328 342 364 379 350 350 316 300 314 259 336 322 322 312 326 362 271 334 334 342 324 399 321 286 300 314 314 312 326 362 334 334 342 324 399 321 300 314 314

m/z base peak 97 109 109 204 109 109 97 109 109 125 125 125 125 125 125 125 203 97 97 97 97 97 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109 109

56 73 87 59 56 52 27 63 61 18 36 65 50a 83 84 16 49 42 10a 18a 56 73 46 71 27a 42a 54 71 44 84 52 45a 68 41a 23a 42 59 31 67a 78 54 45 88 84 71 66 73 65

(M+-29)

(M+-65)

(M+-91)

(M+-91)

Isolated as a hydrochloride; bSpecic rotation of the free base []D28 = 42.3 (c = 1.30, CHCl3).

870
Table II. A1 and A2A adenosine receptor binding of triazolopyrimidines 4ad.

Compound 4a.1

R1

A1 Ki nM 128 12

A2A Ki nM > 10 000

Ki A2A/Ki A1 > 78

4a.2

1 360 82 > 10 000

> 10 000 > 10 000

> 7.3

4a.3

4a.4

> 10 000

> 10 000

4a.5

> 10 000 > 10 000

> 10 000 > 10 000

4a.6

4a.7

> 10 000

> 10 000

----------------------------------------------------------------------------------------------------------------------------------------4b.1 37 3 1 460 131 39.4

4b.2

75 5 39 3

1 740 156 > 10 000

23.2 > 256

4b.3

4b.4

68 4

798 56

11.7

4b.5

42 3

616 43

14.6

4b.6

172 7

1 305 104

7.5

871
Table II. continued. Compound R R1 A1 Ki nM 4c.1 10.5 0.7 A2A Ki nM 3 422 205 325.9 Ki A2A/Ki A1

4c.2

19.5 1.2 73.4 6.6

3 973 198 2 181 153

203.7

4c.3

29.7

4c.4

91 6

932 83

10.2

4c.5

56 4

1 906 172

34.0

4c.6

402 32

> 10 000

> 24.8

4c.7

1 188 95

> 10 000

> 8.4

4c.8

1 650 149

9 800 686

5.9

4c.9

112 7

> 10 000

> 89

4c.10

45 3

2 613 183

58.0

4c.11

204 18 81 6 262 23 55 4

> 10 000 > 10 000 > 10 000 4 000 240

> 49 > 123 > 38

4c.12

4c.13

4c.14

72.7

872
Table II. continued. Compound 4d.1 R R1 A1 Ki nM 26 2 72 6 287 26 A2A Ki nM > 10 000 > 10 000 > 10 000 Ki A2A/Ki A1 > 384 > 138 > 34

4d.2

4d.3

4d.4

254 18 1 190 83

> 10 000 > 10 000

> 39 >8

4d.5

4d.6

1 137 102

> 10 000

>8

4d.7

> 10 000

> 10 000

4d.8

377 26 429 30 119.5 9.5 268 24 84 6

> 10 000 > 10 000 > 10 000 > 10 000 7 586 683

> 26 > 23 > 83 > 37 90.3

4d.9

4d.10 4d.11

4d.12

The 2-thiophenemethyl (2-thenyl) substituent (series b) was chosen because it could partially imitate the 2-chlorobenzyl substituent in view of its steric and electronic characteristics. This substituent appeared to be actively involved in A1 adenosine receptor binding: the triazolopyrimidines 4b.15 (table II) showed a high A1 adenosine affinity (Ki < 100 nM) except for the

m-toluidino derivative 4b.6 (Ki = 172 nM). The most active compound of this series was again the cyclopentylamino derivative 4b.1 (Ki = 37 nM) but its affinity was lower than those of the 2-chlorobenzyl (Ki = 21 nM) [1] and 2-uorobenzyl 4c.1 (Ki = 10.5 nM) derivatives. These results further conrmed the ability of the A1 receptor to interact with lipophilic substituents bearing a

873 sulphur or a chlorine atom in the 2 position. The amphetamino 4b.3 and the p-toluidino 4b.5 derivatives possessed equivalent affinities (Ki = 39 and 42 nM respectively), conrming the effectiveness of the amphetamino substituent and of the para-methyl position, compared with the meta-methyl one (4b.6, Ki = 173 nM). This effect was also recorded in the para-uorobenzyl series (4d.4, Ki = 254 nM; 4d.5, Ki = 1 190 nM). Among compounds of the b series, 4b.3 showed the best selectivity (Ki A2A/Ki A1 > 256) and the most active compound 4b1 followed with a ratio Ki A2A / Ki A1 = 39.5. The selection of the uorobenzyl substituents, to compare with the corresponding chlorobenzyl substituents, was based upon the consideration that the strong electronegativity of the uorine, which is capable of accepting a hydrogen bond, allowed us to evaluate the effect of possible electronic, as well as steric factors on receptor binding. As regards the 2-uorobenzyl substituent, used for the preparation of the triazolopyrimidines 4c.114 (table II), a generalized increase in affinity was observed towards the A1 adenosine receptors, compared with the 2-chlorobenzyl derivatives [1], excluding the amphetamino derivatives 4c.3 and 4c.4. The most active compounds were once again the cyclopentylamino 4c.1 (Ki = 10.5 nM) and the cyclohexylamino 4c.2 (Ki = 19.5 nM) derivatives, which also showed the best selectivity towards the A1 adenosine receptors (Ki A2A/Ki A1 = 326 and 203, respectively); the introduction of an alcoholic OH function on the cyclohexyl ring (compound 4c.7) (see GR-79326, GlaxoWellcome A1 adenosine agonist) markedly decreased binding affinity. The amphetamino substituent maintained a good affinity (Ki < 100 nM), but it is worth noting that the racemic derivative 4c.3 was found to be slightly more effective than the levorotatory isomer 4c.4. The new furfurylamino substituent (4c.8) markedly lowered receptor binding. As regards the known 7-arylamino substituents, the binding affinity was conrmed to be high with the para-toluidino derivative 4c.5 (Ki = 56 nM), while it decreased with the meta-toluidino derivative 4c.6. Also the new pyridylamino substituent of 4c.10 induced a high receptor affinity (Ki = 45 nM) and a good selectivity. As for the four derivatives bearing branched aliphatic amines not previously tested by us, coming from the simplication or opening of the cyclopentyl ring, it is interesting to point out the high affinity of 4c.14 (3-pentylamino derivative, Ki = 55 nM) and 4c.12 (2-butylamino derivative, Ki = 81 nM). Finally, as regards the 4-uorobenzyl substituent used for the preparation of the triazolopyrimidine derivatives 4d.112, the results reported in table II show that inhibition constants Ki < 100 nM were obtained for 4d.1 (cyclopentyl derivative, Ki = 26 nM), 4d.2 (cyclohexyl derivative, Ki = 72 nM) and 4d.12 (3-pentyl derivative, Ki = 84 nM). Thus in this series, a high affinity towards A1 adenosine receptors was conferred by aliphatic aminosubstituents, which also showed a high receptor selectivity. The other aminosubstituents induced a noticeable decrease in binding affinity compared with the corresponding derivatives of the 2-chlorobenzyl [1] or 2-uorobenzyl series. 5. Experimental protocols 5.1. Chemistry Melting points were determined on a Koer hot-stage and are uncorrected. IR spectra in nujol mulls were recorded on a Perkin-Elmer Mod. 1310 spectrometer. Mass spectra were performed with a Hewlett Packard MS/System 5988. Elemental analyses (C, H, N) were within 0.4% of theoretical values and were performed on a Carlo Erba Elemental Analyzer Mod. 1106 apparatus. Optical rotations were measured with a Violet AA-5 polarimeter. 5.2. 2-Azidomethyl-thiophene NaN3 (7.68 g, 118 mmol) was added to a solution of 2-chloromethyl-thiophene (5.24 g, 39.5 mmol) in 25 mL of MeOH and 4 mL of H2O, and the suspension was stirred at room temperature for 24 h. The inorganic precipitate was removed by ltration and the ltrate was concentrated, treated with H2O and extracted with CHCl3. The chloroform extract was dried and evaporated in vacuo to give the title compound as a yellow oil which was used without purication: 5.19 g, 94.5% yield. The azide was short distilled in a tubular oven at 4045 C, 1.8 mm Hg; IR (cm1): 2 105 (N3). 5.2.1. 1-Substituted-4-carboxamido-5-amino-1H-1,2,3triazoles 1bd 0.934 g (11 mmol) of cyanacetamide was added to a stirred solution of sodium ethoxide (0.253 g, 0.011 g atom of Na) in 10 mL of absolute EtOH, and stirring was continued for 30 min. A solution of 10.0 mmol of the suitable azide (2-thenyl-, 2-uorobenzylor 4-uorobenzylazide) in 10 mL of absolute EtOH was slowly added to the suspension, and then the mixture was heated under reux for 1.52 h. The reaction mixture was concentrated in vacuo and treated with H2O, and the insoluble material, consisting of the title compounds, was collected and puried by crystallization (table I).

874 5.2.2. 3-Substituted-7-hydroxy-1,2,3-triazolo[4,5-d]pyrimidines 2bd A solution of 10.0 mmol of the suitable triazole derivative (1b, 1c or 1d) in 8 mL of formamide was reuxed for 2 h. After cooling the reaction mixture was diluted with H2O and the solid precipitate was collected by ltration and recrystallized (table I). 5.2.3. 3-Substituted-7-chloro-1,2,3-triazolo[4,5-d]pyrimidines 3bd 0.8 ml of DMF and 4.5 mL of SOCl2 were added to a suspension of 5.0 mmol of the suitable triazolopyrimidine (2b, 2c or 2d) in 22 mL of boiling anhydrous CHCl3. The reaction mixture was reuxed for 2 h, the solvent was evaporated in vacuo (temperature 35 C), and the residue, after cooling at 0 C, was triturated with crushed ice. The solid formed was collected by ltration, dried and extracted repeatedly with boiling 6080 C petroleum ether. The combined extracts were evaporated in vacuo to give the title compounds as white solids (table I). 5.2.4. 3-(4-Chlorobenzyl)-7-(substituted amino)-1,2,3triazolo[4,5-d]pyrimidines 4a.17 A mixture of 3-(4-chlorobenzyl)-7-chloro-1,2,3triazolo[4,5-d]pyrimidine 3a (0.40 g, 1.43 mmol), triethylamine (0.24 mL, 1.70 mmol) and the suitable amine (1.70 mmol of cyclopentyl-, cyclohexyl-, ()--methylphenethylamine, 3.70 mmol of ()--methylbenzylamine, 5.0 mmol of para- and meta-toluidine or 1.0 mmol of para-nitroaniline) in 10 mL of absolute EtOH was reuxed for 2.5 h. For the isolation of compounds 4a.1 and 4a.2, the reaction mixture was evaporated in vacuo, the residue was treated with H2O and 5% HCl (pH 3) and the insoluble material was collected and puried by crystallization (4a.1) or by extraction with 6080 C petroleum ether (4a.2) respectively. For the isolation of compounds 4a.3, 4a.5, 4a.6, and 4a.7, the reaction mixture was allowed to cool and the crystallized precipitate was collected by ltration. Isolation of 4a.7 required further treatment of the precipitate with 10% HCl to remove unreacted p-nitroaniline. For the isolation of 4a.4, the reaction mixture was evaporated in vacuo, the residue was dissolved in MeOH and the derivative was precipitated as a hydrochloride by addition of Et2O-HCl (table I). 5.2.5. 3-(2-Thenyl)-7-(substituted amino)-1,2,3-triazolo[4,5-d]pyrimidines 4b.16 A mixture of 3b (0.25 g, 1.0 mmol), triethylamine (0.17 mL, 1.2 mmol) and the suitable amine (1.2 mmol of cyclopentyl-, cyclohexyl-, ()--methylphenethyl-, ()-methylbenzylamine, para- or meta-toluidine) in 10 mL of absolute EtOH was heated under reux for 2.5 h. The reaction mixture was evaporated in vacuo and the residue was treated with H2O and puried by crystallization to give compounds 4b.1, 4b.2, 4b.5, and 4b.6. For the isolation of 4b.3 and 4b.4, the residue was extracted with Et2O, and then Et2O-HCl was added to the ether solution to precipitate the derivatives as hydrochlorides which were collected and crystallized (table I). 5.2.6. 3-(2-Fluorobenzyl)-7-(substituted amino)-1,2,3triazolo[4,5-d]pyrimidines 4c.114 A mixture of 3c (0.30 g, 1.14 mmol), triethylamine (0.16 mL, 1.15 mmol) and the suitable amine (1.40 mmol of cyclopentyl-, cyclohexyl-, ()--methylphenethyl- or ()--methylphenethylamine; 4.0 mmol of para or metatoluidine; 2.2 mmol of trans-2-cyclohexanol-, furfuryl-, 4-bromophenyl-, 2-pyridyl-, 2-butyl-, 2-pentyl- or 3-pentylamine) in 10 mL of absolute EtOH (10 mL of toluene for the pyridylamino derivative 4c.10) was heated under reux for 2.5 h. For the isolation of 4c.5, 4c.6, 4c.8, 4c.9 and 4c.14, the reaction mixture was allowed to cool and the crystallized precipitate was collected by ltration (table I). For the isolation of 4c.1, 4c.2, 4c.7 and 4c.10, the reaction mixture was evaporated in vacuo, the residue was treated with H2O and 10% HCl and the insoluble material was collected by ltration and crystallized (table I). For the isolation of 4c.3, 4c.4, 4c.12 and 4c.13, the reaction mixture was evaporated in vacuo, the residue was washed with H2O and 5% HCl and dissolved in Et2O or absolute EtOH or MeOH. Addition of Et2O-HCl to the solution precipitated the compounds as hydrochlorides which were puried by crystallization (table I). The isopropylamino derivative 4c.11 was obtained by heating the 7-chloro-triazolopyrimidine 3c (0.39 g, 1.48 mmol) in 2 mL of isopropylamine at 100110 C in a closed tube for 2.5 h and treating the reaction mixture with H2O to precipitate the expected derivative (table I). 5.2.7. 3-(4-Fluorobenzyl)-7-(substituted amino)-1,2,3triazolo[4,5-d]pyrimidines 4d.112 A mixture of 3d (0.30 g, 1.14 mmol), triethylamine (0.16 mL, 1.15 mmol) and the suitable amine (1.40 mmol of cyclopentyl-, cyclohexyl-, ()--methylphenethyl- or furfurylamine; 4.0 mmol of para or meta-toluidine; 2.3 mmol of trans-2-cyclohexanol-, 4-bromophenyl-, 2-pyridyl-, 2-butyl-, 2-pentyl- or 3-pentylamine) in 10 mL of absolute EtOH (10 mL of toluene for the pyridylamino derivative 4d.9) was heated under reux for 2.5 h. The reaction mixture was evaporated in vacuo,

875 the residue was washed with H2O and 5% HCl (pH 3) and the insoluble material was collected and puried by crystallization (4d.1, 4d.6, 4d.10, 4d.11 and 4d.12) or by extraction with 6080 C petroleum ether (4d.2) (table I). For the isolation of 4d.3, the liquid residue obtained after the rinses was dissolved in Et2O or MeOH, and Et2O-HCl was added to the solution to precipitate the derivative as a hydrochloride (table I). For the isolation of 4d.4, 4d.5, 4d.7, 4d.8 and 4d.9 the reaction mixture was allowed to cool and the crystallized precipitate was collected by ltration (table I). References
[1] [2] [3] [4] [5] [6] [7] Betti L., Biagi G., Giannaccini G., Giorgi I., Livi O., Lucacchini A., Manera C., Scartoni V., J. Med. Chem. 41 (1998) 668773. Biagi G., Giorgi I., Livi O., Lucacchini A., Martini C., Scartoni V., Tacchi P., Il Farmaco 49 (1994) 183186. Biagi G., Giorgi I., Livi O., Lucacchini A., Martini C., Scartoni V., Tacchi P., Il Farmaco 49 (1994) 187191. Biagi G., Giorgi I., Livi O., Scartoni V., Lucacchini A., Martini C., Tacchi P., Il Farmaco 50 (1995) 1319. Biagi G., Breschi C., Giorgi I., Livi O., Martini C., Scartoni V., Scatizzi R., Il Farmaco 50 (1995) 659667. Biagi G., Giorgi I., Livi O., Lucacchini A., Scartoni V., Il Farmaco 51 (1996) 395399. Dudley M.W., Peet N., Demeter D., Weintraub H.J.R., Ijzerman A.D.P., Nordvall G., van Galen P.J.M., Jacobson K.A., Drug Dev. Res. 28 (1993) 237243. Ried W., Laoutidis J., Chemiker-Ztg. 114 (1990) 246248. Aubert T., Farnier M., Guilard R., Can. J. Chem. 68 (1990) 842851. Kreher R., Bergman U., Tetrahedron Lett. 47 (1976) 42594262. Baraldi P.G., Manfredini S., Simoni D., Zappaterra L., Zocchi C., Dionisotti S., Ongini E., Med. Chem. Lett. 4 (1994) 25392544.

Acknowledgements We wish to thank the Consiglio Nazionale delle Ricerche (C.N.R.) for nancial support.

[8] [9] [10] [11]

Eur. J. Med. Chem. 34 (1999) 877882 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

877

Short communication

Synthesis and cardiovascular evaluation of N-substituted 1-aminomethyl-2-naphthols

A.Y. Shena*, C.T. Tsaib, C.L. Chena
a

Department of Pharmaceutical Science, Foo Yin Institute of Technology, Ta-Liao, Kaohsiung County 831, Taiwan b Department of Biology, National Changhua University of Education, Changhua, Taiwan (Received 9 December 1998; revised 15 April 1999; accepted 6 May 1999)

Abstract A series of 1-alkylaminomethylnaphthols have been prepared. These compounds were readily prepared in good yields by addition of 2-naphthol to formalin and alkylamines. The hypotensive and bradycardiac effects of these compounds in normotensive rats as well as their in vitro inotropic and aortic contraction effects in isolated rat left atria and aorta have been evaluated. A higher depressor and bradycardiac activity was found for compounds substituted on nitrogen by naphthol with primary amines, i.e., ethylamine, propylamine, isopropylamine, or butylamine and with a cyclic secondary amine, i.e., pyrrolidinyl. These compounds produced biphasic changes in contractile force in isolated rat atria which was correlated to blood pressure and heart rate activity. 1-Isopropylaminomethyl-2-naphthol hydrochloride relaxed isolated rat aortic rings precontracted with high extracellular K+ (80 mM) and Ca2+ (1.9 mM). The suppressive effects of the compounds may involve a direct inhibition of Ca2+ channels. The biological activity of these compounds can be explained in terms of substitution on nitrogen. 1999 ditions scientiques et mdicales Elsevier SAS 1-alkylaminomethyl-2-naphthol / blood pressure / heart rate / left atrial / aorta

1. Introduction 1-Pyrrolidinylmethyl-2-naphthol hydrochloride (TPY-) has been shown to produce a reduction in blood pressure (BP) and heart rate (HR) in anaesthetized rats [1]. The ionic mechanism of the cardiovascular activity of TPY- has also been examined. The results indicated that suppressive effects of TPY- involve a direct depressant action on heart cells and vascular smooth cells [2]. The direct inhibition of voltagedependent L-type Ca2+ channels is involved in the TPY- mediated vasodilatory action. In addition, the inhibitory effect of TPY- on cardiac contractibility through the blockade of L-type Ca2+ channels can be prevented by TPY- mediated inhibition of the transient outward potassium current. Bril et al. [3] reported that a compound

*Correspondence and reprints

with a combination of potassium and calcium channel antagonistic properties might consititute a novel antiarrhythmic agent with reduced proarrhythmic risk. Some of the drugs used to treat ventricular arrhythmias have been shown to also act on the transient outward current at therapeutic concentrations [46]. The actions of TPY- on the cardiovascular system encouraged us to search for and synthesize novel derivatives of aminomethylnaphthol. This brought forth the modication of the pyrrolidinyl group in TPY- with primary amines via the condensation of 2-naphthol with primary amines. The method exists for attaching carbon substituents to the 1-position of 2-naphthol, and many 2-naphthols bearing such substituents are known in the literature [7]. A selection of compounds possessing a variety of 1-substituents was desired in order to broadly dene the structure activity relationship for this portion of the molecule. These compounds, shown in table I, either were known in the literature or were prepared according to well-established synthetic methods [7, 8]. The purpose of the present study was to evaluate the importance of the side substitution at the 1-position of the aminomethylnaphthols.

878
Table I. Hypotensive and bradycardic response following intravenous injection of aminomethylnaphthol derivatives (2.2 mol/kg) in rats.

1 Compound N.S. 1 2 3 4 5 6 7 8 9 10 11 12 13 R1 R2 X HR (beats/min)a maximum change % maximum 42 -8 4 -38 15 -240 15* -276 31* -268 54* -328 43* -26 3* -15 4* -13 5* 0 -258 33* -137 37* -121 17* 11 21 11 5 62 11* 67 13* 65 7* 75 21* 7 2* 42 32 0 64 15* 32 8* 28 8*

213 BP (mm Hg)a maximum change % maximum 43 -5 4 -10 5 -75 19* -94 21* -82 11* -103 16* -16 5* -6 4 0 0 -81 2* -59 13* -38 8* 32 65 10 5 69 16* 81 18* 70 20* 85 15* 15 4* 54 0 0 77 17* 53 11* 23 9*

CH3 H C2H5 H n-C3H7 H i-C3H7 H C4H9 H CH3 CH3 C2H5 C2H5 C3H7 C3H7 C4H9 C4H9 TPY-(CH2)5-(CH2)2-O-(CH2)2-

HCl HCl HCl HCl HCl

a Percent decrease in blood pressure and heart rate were calculated from the decrease in blood pressure and heart rate of the treatment group and the blood pressure (normally 105120 mm Hg) and heart rate (normally 385405 beats/min) of the control group; 2.2 mol/kg usually amounts to 0.40.5 mg/kg; values are expressed as the mean SEM. Asterisks indicate signicant difference (t-test, calculated on the changes) from the N.S. (normal saline). *P < 0.05.

2. Chemistry In the present work, reaction of 2-naphthol with formaldehyde and methylamine in a molar ratio of 1:2:1, respectively, in a methanol solution at 60 C was found to produce 2,3-dihydro-2-methyl-1H-naphth[1,2-e]-moxazine (1). Upon treating 1 with hot aqueous hydrochloric acid, formaldehyde was liberated and the hydrochloride of 1-methylaminomethyl-2-naphthol (2) was formed (gure 1) [7]. The condensation, run in the same manner except with ethylamine, propylamine, isopropylamine or butylamine, resulted in the formation of 1-alkylaminomethyl-2-naphthols directly instead of naphthoxazine products. Compounds 1, 2, 5, and 6 were previously synthesized and reported [7]. The temperature and the particular amine seem to be the important factors in determining the course of the reaction of 2-naphthol with formaldehyde and primary amine as reported by Burke et al. [7]. The condensation of naphthols with formaldehyde and secondary amines has been studied previously [8]. Reaction of piperidine and formaldehyde with 1- and 2-naphthol has been shown to result in the introduction of

a piperidinomethyl group into the 1-position of 2-naphthol and into the 2-position of 1-naphthol. This indicates that the substituent ortho to the naphthylene

Figure 1. Synthetic routes for the 2,3-dihydro-2-methyl-1Hnaphth[1,2-e]-m-oxazine (1) and aminomethylnaphthol derivatives (213).

879

Figure 2. Effects of intravenous injection of 1-isopropylaminomethyl-2-naphthol (5) and dipropylaminomethyl-2-naphthol (9) on blood pressure and heart rate in one anaestherized normotensive rat. Compounds were administered at the time indicated by an arrow.

hydroxyl group plays a role in determining the course of the reaction [7, 8]. 3. Results and discussion Aminonaphthols have been reported to exhibit antihypertensive, adrenoceptor blocking, and Ca2+ channel blocking activities [911]. As described in the experimental section, the hypotensive and bradycardiac activity of the analogue series was tested in normotensive rats at a dosage of 2.2 mol/kg. Figure 2 shows the chart recordings of blood pressure (top traces) and heart rate (bottom traces) of an anaesthetized rat. Intravenous injection of compound 5 (2.2 mol/kg) induced a maximum reduction of mean arterial pressure from 120 mm Hg to 25 mm Hg within 30 s and the BP returned to the control value in 20 min. Along with the sudden decrease in BP, the heart rate was also reduced from 400 beats min1 to a minimum value of 130 beats min1, and it took 40 min for full recovery to the control level. The acute hypotensive and bradycardiac responses induced by 2.2 mol/kg of these compounds are summarized in table I. Substitution of the amine hydrogen atom by alkyl groups of increasing chain length altered the activity. In the dialkylaminomethyl-2naphthol series, the straight chain derivatives (710) had less hypotensive and bradycardiac activity than the cyclic analogues (1113). An increase in the size of the N-dialkyl group appeared to reduce hypotensive and bradycardiac activity. Despite the limited series from which to evaluate the structure-activity relationship, primary amine substitution (36) showed better potency than secondary amine derivatives except methyl side

chain derivatives (12), which were not active even at a higher dosage (data not shown). The present results (table II) in vitro show that the 1-alkylaminomethyl-2-naphthols (36) and the secondary amines with cyclic side chains (1113), at 30 M, produced a biphasic effect on the contractile force in isolated rat left atria, i.e., an initial decrease and a sustained increase. A representative inotropic response of an electrically driven left atrium to 1-isopropylaminomethyl-2-naphthol hydrochloride (5) is depicted in gure 3. However, the straight chain derivatives (710) show the negative inotropic effect. The 4-Aminopyridinesensitive transient outward current (Ito) has been shown to be present in rabbit, rat, cat, dog and human cardiac myocytes, but not in guinea-pig cardiac myocytes [12 15]. It has been suggested that the compound 11 (TPY-) mediated inhibition of Ito would reverse the rat myocardial depressant effect caused by its initial inhibition of L-type Ca2+ channels [3]. The similarity of the molecular structure suggests that 1-isopropylaminomethyl-2naphthol hydrochloride (5) and other analogues work in the same manner as compound 11 (TPY-). In addition, Ca2+ (1.9 mM) elicited a 100% contraction in rat aorta in the presence of high K+ (80 mM), 5 was more potent in suppressing the tone induced by high K+ and Ca2+ than 7 as shown in table III. High K+ has been shown to evoke smooth muscle contraction by promoting Ca2+ inux through voltage-sensitive Ca2+ channels which are readily activated by membrane depolarization [16]. Therefore, extracellular Ca2+ entry is thought to be the main cause of the high K+-induced contraction. The present studies indicate that the suppressive effects of

880
Table II. Characteristics of aminomethylnaphthols (30 M) on the contractile force in rat left atria stimulated at 1.0 Hz. Compound Maximal force of contraction (%) phase 1 1 2 3 4 5 6 7 8 9 10 11 12 13 100 100 90 85 82 83 78 80 88 97 77 90 89 18 (3) 20 (3) 16* (5) 17* (5) 13* (6) 17* (5) 19* (4) 21* (4) 15* (4) 19* (4) 12* (5) 21* (5) 16* (6) phase 2 100 100 220 235 238 219 78 80 88 97 225 160 120 16 (3) 19 (3) 36* (4) 42* (5) 29* (5) 23* (5) 19* (4) 21* (4) 15* (4) 19* (4) 49* (4) 23* (4) 25* (5)

Values are means SEM. Number of preparations (n) is given in parentheses.*P < 0.05 compared with the respective control by student t test.

Figure 3. Effect of 1-isopropylaminomethyl-2-naphthol (5) and dimethylaminomethyl-2-naphthol (7) on the contractile force of isolated rat left atria. Stimuli were driven at 1 Hz with 2 times the threshold. The presence of compound 5 (30 M) produced a biphasic response, while compound 7 (30 M) produced only a negative intropic effect.

these compounds could interfere with Ca2+ inux through the depolarized cell membrane to induce relaxation of rat aorta. The results observed are consistent with previous ndings that 11 (TPY-) or aminonaphthols effectively suppressed the voltage-gated Ca2+ current [2, 17, 18]. In conclusion, the biological activity of these compounds can be explained in terms of substitution on nitrogen. The development of N-substituted 1 aminomethyl-2-naphthols with dual effects would be of potential therapeutic advantage. 4. Experimental protocols 4.1. Chemistry Melting points were determined on a Yanagimoto MP-3 micromelting apparatus and are uncorrected.
Table III. Effects of dimethylaminomethyl-2-naphthol (7) and 1-isopropylaminomethyl-2-naphthol hydrochloride (5) on the rat aortic contraction induced by KCl and Ca. Compound Control 7 (45 M) 5 (45 M) 5 (150 M) Contraction (% of Control) KCl (80 mM) + Ca (1.9 mM) 100 79 50.4 15.3 3.9 4.8 6.4 1.0

Analyses indicated by the symbols of the elements were within 0.4% of the theoretical values. Infrared spectra were obtained on a Shimadzu IR-408 spectrophotometer. Nuclear magnetic resonance spectra were recorded on a Varian Gemini T-300 spectrometer at the National Sun Yat-Sen University, Kaohsiung, and are expressed in parts per million () with tetramethylsilane used as an internal standard. Mass spectra recorded for the purposes of structure conrmation were obtained on a Jeol JMS-HX 110 mass spectrometer at the National Sun Yat-Sen University, Kaohsiung. Elemental analysis was performed on a CHN-O-Rapid Heraeus Elemental Analyzer at the National Cheng-Kung University, Tainan. Thin layer chromatography was carried out on precoated silica gel F254 chromatographic plates (20 20 cm; 0.2 mm). Methylamine, ethylamine, propylamine, isopropylamine and butylamine were all obtained from the Tokyo Chemical Industry Co. (TCI). 2-Naphthol was the product of the Sigma Co. All other reagents used in this study were EP grade products of E Merck. Dialkylaminomethyl-2-naphthol derivatives were synthesized previously [8]. The synthesis and characterization of 1-alkylaminomethyl-2-naphthol will be reported here. 4.1.1. 2,3-Dihydro-2-methyl-1H-naphth[1,2-e]-m-oxazine (1) Following the procedure of Burke et al. [7], to a cooled solution of 12.4 g of 25% aqueous methylamine (0.1 mol)

Percentages of the control contraction were calculated and presented as mean SEM (n = 4). *P < 0.05 as compared with the respective control.

881 in 60 mL of methanol was added 18.5 mL of 37% (0.2 mol) aqueous formaldehyde in 40 mL of methanol and 14.4 g of 2-naphthol (0.1 mol) in 50 mL of methanol. After 1.5 h of gentle reuxing at 60 C, the reaction mixture was poured into 400 mL of cold water. The product (95% yield) was collected and recrystallized from methanol. M.p. 6870 C, IR (nuzol) cm1: 2 994 (aromatic CH), 2 909 (CH3), 1 427 (CO), 1 214 (CN). 1 H NMR (300 MHz, DMSO): 7.82 (d, J = 8.4 Hz, 1H), 7.65 (m, 2H), 7.48 (m, 1H), 7.38 (m, 1H), 7.03 (d, J = 9.0 Hz, 1H), 4.82 (s, 2H), 4.22 (s, 2H), 2.53 (s, 3H). MS m/z (relative intensity): 128 M+ (100), 144 (72), 156 (42). Anal. C13H13NO (C, H, N). 4.1.2. 1-Methylaminomethyl-2-naphthol hydrochloride (2) Following the procedure of Burke et al. [7], a solution of 0.013 mol of 2,3-dihydro-2-methyl-1H-naphth[1,2-e]m-oxazine and 1.0 mL of concentrated aqueous hydrochloric acid in 60 mL of 85% aqueous propanol-1 was distilled for about 30 min. After the distillation was interrupted, 60 mL of acetone was added to the reaction mixture. Upon cooling and ltration the product was obtained (91% yield). M.p. 187189 C (dec.). IR (nuzol) cm1: 2 987 (aromatic CH), 2 904 (CH3), 1 443 (CO). 1 H NMR (300 MHz, D2O): 7.93 (m, 3H), 7.62 (dd, J = 8.7, 7.2 Hz, 1H), 7.45 (dd, J = 7.5 Hz, 1H), 7.24 (d, J = 9.6 Hz, 1H), 4.62 (s, 2H), 2.73 (s, 3H). MS m/z (relative intensity): 128 M+ (100), 156 (42), 187 (26). Anal. C12H13NO HCl (C, H, N). 4.1.3. 1-Ethylaminomethyl-2-naphthol hydrochloride (3) To a solution of 3.6 g of 2-naphthol (0.025 mol) and 4.7 mL of 37% aqueous formaldehyde (0.06 mol) in 30 mL of methanol 1.6 g of 70% ethylamine (0.025 mol) in 10 mL methanol was added dropwise. After 1.5 h of gentle reuxing at 60 C the oily residue was dissolved in absolute alcohol. The hydrochloride was obtained by treating a cold alcohol solution of the product with concentrated hydrochloric acid (yield 60%). M.p. 175177 C (dec.). IR (nuzol) cm1: 3 454 (broad, NH, OH), 3 010 (aromatic CH), 2 909 (CH3). 1H NMR (300 MHz, D2O): 7.88 (m, 3H), 7.59 (dd, J = 8.7, 7.2 Hz, 1H), 7.42 (dd, J = 7.5 Hz, 1H), 7.20 (d, J = 8.7 Hz, 1H), 4.54 (s, 2H), 3.16 (q, J = 7.5 Hz, 2H). 1.29 (t, J = 7.5 Hz, 3H). MS m/z (relative intensity): 128 M+ (100), 156 (57), 201 (47). Anal. C12H13NO HCl (C, H, N). 4.1.4. 1-Propylaminomethyl-2-naphthol hydrochloride (4) The product was prepared by a method similar to that described in the procedure for compound 3. It was recrystallized from absolute alcohol (yield 65%). M.p. 188190 C (dec). IR (nuzol) cm1: 3 486 (NH), 3 442 (OH), 2 990 (aromatic CH), 2 913 (CH2, CH3), 1 426 (CO). 1H NMR (300 MHz, D2O): 7.89 (m, 3H), 7.5 (dd, J = 8.7, 7.2 Hz, 1H), 7.42 (dd, J = 7.5, 7.5 Hz, 1H), 7.20 (d, J = 9.0 Hz, 1H), 4.54 (s, 2H), 3.04 (t, J = 7.2 Hz, 2H), 1.70 (m, 2H). 0.92 (t, J = 7.5 Hz, 3H). MS m/z (relative intensity): 128 M+ (42), 215(30). Anal. C14H17NO HCl (C, H, N). 4.1.5. 1-Isopropylaminomethyl-2-naphthol hydrochloride (5) The product was prepared by a method similar to that described in the procedure for compound 3. It was recrystallized from absolute alcohol (yield 70%). M.p. 183185 C (dec.). IR (nuzol) cm1: 2 990 (aromatic CH), 2 989 (CCC), 1 440 (CO). 1H NMR (300 MHz, DMSO): 8.07 (d, J = 8.7 Hz, 1H), 7.88 (d, J = 8.7 Hz, 1H), 7.85 (d, J = 6.6 Hz, 1H), 7.55 (dd, J = 7.2, 7.5 Hz, 1H), 7.37 (m, 2H), 4.09 (bs, 2H), 3.45 (m, 1H), 1.37 (s, 3H). 1.34 (s, 3H). MS m/z (relative intensity): 128 M+ (42), 215(12). Anal. C14H17NO HCl (C, H, N, O). 4.1.6. 1-butylaminomethyl-2-naphthol hydrochloride (6) The product was prepared by a method similar to that described in the procedure for compound 3. It was recrystallized from absolute alcohol. M.p. 150153 C (dec.). IR (nuzol) cm1: 3 473 (broad, OH, NH), 3 006 (aromatic CH), 2 916 (CH3), 1 436 (CO), 1 299 (CN). 1 H NMR (300 MHz, D2O): 7.86 (m, 3H), 7.59 (dd, J = 8.1, 6.9 Hz, 1H), 7.42 (dd, J = 7.5, 7.5 Hz, 1H), 7.15 (d, J = 8.7 Hz, 1H), 4.43 (s, 2H), 3.31 (m, 1H), 1.84 (m, 1H), 1.62 (m, 1H), 1.36 (d, J = 7.2 Hz, 3H), 0.96 (t, J = 7.5 Hz, 3H). MS m/z (relative intensity): 128 M+ (100), 156 (52), 229 (17). Anal. C15H19NO HCl (C, H, N). 4.2. Measurement of BP and HR by intravenous injection The in vivo experiments have been described previously [2]. In brief, Wistar rats (250300 g) of either sex were used in these studies. The animals were housed in an animal room with a light/dark cycle of 12 h/12 h and were fed rat chow and tap water ad libitum. The animals were anaesthetized with urethane (Aldrich, 1.0 g/kg IP, supplemented with 300 mg/kg i.v. if necessary). The trachea was intubated to keep the airway patent. Femoral arterial blood pressure was measured through a PE 50 tubing lled with heparin solution (25 units/mL) (Sigma) connected to a polygraph (Lectromed, U.K.) via a transducer (PDCR 75, Lectromed, UK). The femoral vein was cannulated for drug administration. The HR was derived by means of a cardiotachometer that was triggered by the arterial pressure pulse. The BP and HR were monitored

882 continuously. Rectal temperature was monitored and maintained between 37 and 38 C. 4.3. Measurement of contractile force in isolated rat left atria Experiments were performed following the methods described previously [2]. In brief, the hearts were quickly removed and rinsed in ice-cold Tyrodes solution. Then, left atria were dissected and mounted at 0.5 g resting tension on stainless steel hooks in a 50 mL organ bath, and bathed at 37 C in physiological saline solution containing (mM): NaCl 118, KCl 4.8, CaCl2 2.5, MgSO4 1.2, KH2PO4 1.2, NaHCO3 24, glucose 11. The bath was aerated with 95% O2 and 5% CO2 mixture. One end of the preparation was xed to the bottom of the bath and the other end was connected by a hook to the level of a force-displacement transducer (Ugo Basile, Comero, Italy). Stimuli were delivered as rectangular pulses of 5 ms duration at 2 times the threshold via two platinum electrodes placed on either side of the muscle. The tissues were always allowed to equilibrate for 90 min before the experiments were begun [2, 19]. 4.4. Measurement of contractile force in isolated rat aorta Experiments were performed following the methods described previously [20]. In brief, the thoracic aorta was isolated and excess fat and connective tissue were removed. Aortic rings (5 mm) were mounted in organ baths containing 5 mL and bathed at 37 C in physiological saline solution containing (mM): NaCl 118.4, KCl 4.7, CaCl2 1.9, MgSO4 1.2, KH2PO4 1.2, NaHCO3 25, glucose 11.7. The bath was aerated with a 95% O2 and 5% CO2 mixture. Two stainless steel hooks were inserted into the aortic lumen, one was xed while the other was connected to the transducer. The contractile effects of calcium were studied in rings stabilized in high-K+ solution without Ca2+. Calcium was then added to obtain the desired concentrations. The high K+ solution was prepared by substituting NaCl with KCl (80 mM) in an equimolar amount. Contractions were recorded isometrically via a force-displacement transducer connected to a Gould polygraph (Model 2400). 4.5. Statistical analysis Values are expressed as means standard error of the mean (SEM). Statistical analysis was performed with a paired t test, and a P value smaller than 0.05 was regarded as statistically signicant. Acknowledgements This work was supported by a grant from the National Science Council (NSC 87-2314-B-242-001). The authors thank Drs Sheng-Nan Wu, Chih-Tsao Chiu and Che-Ming Teng for their helpful suggestions. The technical assistance of Hui-Ya Chang is also greatly appreciated. References
[1] [2] [3] Shen A.Y., Chen C.L., Lin C.I., Chin. J. Physiol. 35 (1992) 4554. Shen A.Y., Wu S.N., Drug. Dev. Res. 44 (1998) 8796. Bril A., Gout B., Bonhomme M., Landais L., Faivre J.F., Linee P., Poyser R.H., Ruffolo Jr. R.R., J. Pharmacol. Exp. Ther. 276 (1996) 637646. Nabauer M., Beuckelmann D.J., Erdmann E., Circ. Res. 73 (1993) 386394. Imaizumi Y., Giles W.R., Am. J. Physiol. 253 (1987) H704H708. Nabauer M., Beuckelmann D.J., Circulation 86 (suppl 1) (1992) I-697. Burke W.J., Kolbezen M.J., Stephens C.W., J. Am. Chem. Soc. 74 (1952) 36013605. Shen A.Y., Tsai M.I., Lien E.J., Acta. Pharm. 44 (1994) 117126. Atwal K.S., OReilly B.C., Ruby E.P., Turk C.F., Aberg G., Asaad M.M., Bergey J.L., Moreland S., Powell J.R., J. Med. Chem. 30 (1987) 627635. Grundke M., Himmel H.M., Wettwer E., Borbe H.O., Ravens U., J. Cardiovasc. Pharmacol. 18 (1991) 918925. Jim K.F., Matthews W.D., J. Pharmacol. Exp. Ther. 234 (1985) 161165. Josephson I.R., Sanchez-Chapula J., Brown A.M., Circ. Res. 54 (1984) 157162. Dukes I.D., Morad M., J. Physiol. (London) 435 (1991) 395420. Furukawa T., Myerburg R.J., Furukawa N., Bassett A.L., Kimura S., Circ. Res. 67 (1990) 12871291. Simurda J., Simurdova M., Christe G., Pugers. Arch. 415 (1989) 244246. Van Breemen C., McNaughton E., Biochem. Biophys. Res. Commun. 39 (1970) 567574. Atwal K.S., OReilly B.C., Ruby E.P., Turck C.F., Aberg G., Asaad M.M., Bergery J.L., Moreland S., Powell J.R., J. Med. Chem. 30 (1987) 627635. Grundke M., Himmel H.M., Wettwer E., Borbe H.O., Ravens U., J. Cardiovasc. Pharmacol. 18 (1991) 918925. Wu S.N., Shen A.Y., Hwang T.L., Chin. J. Physiol. 39 (1996) 2329. Ko F.N., Guh J.H., Yu S.M., Hou Y.S., Wu Y.C., Teng C.M., Br. J. Pharmacol. 112 (1994) 11741180.

[4] [5] [6] [7] [8] [9]

[10] [11] [12] [13] [14] [15] [16] [17]

[18] [19] [20]

Eur. J. Med. Chem. 34 (1999) 883889 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

883

Laboratory note

6-Thienyl and 6-phenylimidazo[2,1-b]thiazoles as inhibitors of mitochondrial NADH dehydrogenase

Aldo Andreania*, Mirella Rambaldia, Alberto Leonia, Rita Morigia, Alessandra Locatellia, Gianluca Giorgib, Giorgio Lenazc, Anna Ghellid, Mauro Degli Espostie
b a Dipartimento di Scienze Farmaceutiche, Via Belmeloro 6, 40126 Bologna, Italy Centro Interdipartimentale di Analisi e Determinazioni Strutturali, Via A. Moro, 53100 Siena, Italy c Dipartimento di Biochimica G. Moruzzi, Via Irnerio 48, 40126 Bologna, Italy d Dipartimento di Biologia E.S., Via Irnerio 42, 40126 Bologna, Italy e Department of Biochemistry and Molecular Biology, Monash University, Clayton, VIC, Australia

(Received 16 November 1998; accepted 10 February 1999)

Abstract Starting from the potent inhibitory effect of the previously described 2-methyl-6-(2-thienyl)imidazo[2,1-b]thiazole on mitochondrial complex I, we prepared a series of derivatives in order to study the effect of a different substitution at the positions 2, 5 and 6. The replacement of the thienyl group at position 6 with a phenyl group does not modify the biological behaviour of the lead compound, whereas the shift of the methyl group from position 2 to position 5 yields a compound devoid of inhibitory effects. In both the 6-thienyl and 6-phenyl series, the lengthening of the chain at position 2 has provided useful information to outline the structural determinants of the ubiquinone antagonist action in imidazothiazole derivatives. 1999 ditions scientiques et mdicales Elsevier SAS imidazo[2,1-b]thiazole / mitochondrial complex I / ubiquinone / rotenone

1. Introduction NADH-ubiquinone reductase, commonly known as complex I, is the most complicated and least understood of the respiratory complexes of mitochondria and bacteria [13]. The interest in this intricate enzyme complex is increasing due to its possible involvement in the pathogenesis of human neurodegenerative diseases such as Alzheimers [4], Parkinsons [5], diabetes [6] and the fact that complex I is becoming a preferred target of commercial pesticides, especially acaricides [7]. There is a plethora of complex I inhibitors that act as antagonists of the hydrophobic substrate ubiquinone, and generally differ in their action depending upon which quinone intermediate they preferentially antagonize (quinone, semiquinone or quinol) [8]. Because the chemical determinants of the
Presented in part at the First Italian-Swiss Meeting on Medicinal Chemistry (Torino, Italy, September 2326, 1997). *Correspondence and reprints

different antagonistic actions are unclear [1, 8, 9], the study of a series of derivatives in which the chemical structures have been systematically modied is useful to elucidate how complex I inhibitors interact with the enzyme [10]. Recently, we have undertaken a systematic work to study the structure-activity relationships in complex I inhibitors bearing an indole [11, 12] and imidazothiazole moiety [13]. 2-Methyl-6-(2-thienyl)imidazo[2,1-b]thiazole 1 (gure 1) was the most potent compound and was found to have a mode of action overlapping that of the classical inhibitor rotenone as well as that of the productlike inhibitor myxothiazol [8, 13]. Using 1 as the lead compound, we have evaluated rst the effect of the substitution of the thienyl group in position 6 with a phenyl group (2) and the shift of the methyl group from position 2 to 5 (3). The results indicated that the thienyl group may be replaced by a phenyl group without loss of potency, whereas the position of the methyl group is critical (compound 3 is inactive). We next planned the synthesis of 6-thienyl and 6-phenyl derivatives with

884

Figure 1. Synthesis of 6-thienyl and 6-phenylimidazo[2,1-b]thiazoles.

various substituents at position 2, to identify whether this position in the imidazothiazole ring corresponded to the attachment point for mimics of the polyisoprenoid side chain of ubiquinone, whose length is known to be critical for Complex I activity [14, 15]. If this were the case, 2-substituents with increasing hydrophobicity (1317 and 1924) were expected to have an enhanced inhibitory potency, proportional to their hydrophobicity as previously reported for 2-substituted acridones [16]. However, the structure-activity results indicated that, in imidazothiazoles, position 2 is sterically critical for complex I inhibition, but is unlikely to be the junction for mimics of the ubiquinone side chain. During the synthesis, some unexpected compounds were isolated (18 and 2527) and two of them were tested (26 and 27).

2. Chemistry The 5-alkyl-2-aminothiazoles 6 (gure 1) were prepared from the appropriate aldehydes 4 which were brominated at the -position (5) and treated with thiourea. When compounds 6 (R = ethyl, propyl) were treated with the bromoacetylthiophens 78, white solids were isolated, corresponding to the intermediate salts 1011 which were characterized only on the basis of their IR spectrum (C=O 1 670 cm1). These compounds were reuxed with dilute hydrochloric acid in order to obtain the imidazothiazoles 1317, whose spectroscopic data (table I) are in agreement with the assigned structures. Under the same experimental conditions, starting from 6 (R = butyl) and 3-(bromoacetyl)thiophene 8, the resulting precipitate had a structure different from 1011 since the

885
Table I. Imidazothiazoles 1327. Compound Formula(mw) 13 14
b

M.p., C 111113 105106

max, cm-1 1 265, 850, 725, 685 1 270, 850, 720, 680

(ppm); J (Hz) in DMSO-d6

C11H10N2S2 (234.3) C12H12N2S2 (248.4) C12H12N2S2 (248.4) C11H10N2S2 (234.3) C12H12N2S2 (248.4)

15 16 17

108112 103105 126130

1 660, 1 620, 1 560, 1 060 1 270, 1 070, 780, 715 1 270, 1 260, 780, 725

18

C13H16N2O2S2 176180 (296.4)

d

3 5002 200, 1 620, 1 590, 1 070, 780 1 600, 1 195, 1 065, 720

19

C13H12N2S (228.3) C14H14N2S (242.3) C14H14N2S (242.3) C15H16N2S (256.4) C16H18N2S (270.4) C17H20N2S (284.4) C15H18N2O2S (290.4) C16H20N2O2S (304.4) C17H22N2O2S (318.4)

136140

20

142145

1 600, 1 260, 1 065, 720

21

136140

1 595, 1 180, 770, 710

22

128130

1 595, 1 535, 1 260, 720

23

130133

1 590, 1 250, 1 055, 710

24

109111

1 590, 1 255, 760, 715

25

176179

3 5002 200, 1 615, 1 595, 1 060, 750 3 5002 200, 1 610, 1 590, 1 055, 750 3 5002 200, 1 610, 1 595, 1 060, 750

26

185187

27

185188

1.25 (3H, t, CH3, J = 7.5) 2.77 (2H, q, CH2, J = 7.5) 7.06 (1H, m, T) 7.35 (1H, m, T) 7.41 (1H, m, T) 7.70 (1H, s, th) 8.01 (1H, s, im) 0.95 (3H, t, CH3, J = 7) 1.64 (2H, sex, CH2, J = 7) 2.73 (2H, t, CH2, J = 7) 7.07 (1H, m, T) 7.36 (1H, m, T) 7.41 (1H, m, T) 7.71 (1H, s, th) 8.02 (1H, s, im) 1.29 (6H, d, CH3, J = 7) 3.12 (1H, sep, CH, J = 7) 7.07 (1H, m, T) 7.36 (1H, m, T) 7.41 (1H, m, T) 7.71 (1H, s, th) 8.00 (1H, s, im) 1.25 (3H, t, CH3, J = 7.5) 2.76 (2H, q, CH2, J = 7.5) 7.46 (1H, m, T) 7.55 (1H, m, T) 7.69 (1H, m, T) 7.71 (1H, s, th) 7.98 (1H, s, im) 0.95 (3H, t, CH3, J = 7.5) 1.63 (2H, sex, CH2, J = 7.5) 2.72 (2H, t, CH2, J = 7.5) 7.46 (1H, m, T) 7.56 (1H, m, T) 7.69 (1H, m, T) 7.71 (1H, s, th) 7.98 (1H, s, im) 0.85 (3H, t, CH3, J = 7) 1.32 (4H, m, CH2) 1.76 (2H, m, CH2) 3.94 (1H, t, CH, J = 7) 7.44 (1H, m, T) 7.56 (1H, s, im) 7.58 (1H, m, T) 7.67 (1H, m, T) 1.26 (3H, t, CH3, J = 7.5) 2.78 (2H, q, CH2, J = 7.5) 7.24 (1H, t, ar, J = 8) 7.38 (2H, t, ar, J = 8) 7.72 (1H, s, th) 7.81 (2H, d, ar, J = 8) 8.13 (1H, s, im) 0.94 (3H, t, CH3, J = 7.5) 1.63 (2H, sex, CH2, J = 7.5) 2.72 (2H, t, CH2, J = 7.5) 7.23 (1H, t, ar, J = 8) 7.37 (2H, t, ar, J = 8) 7.71 (1H, s, th) 7.80 (2H, d, ar, J = 8) 8.12 (1H, s, im) 1.29 (6H, d, CH3, J = 7) 3.12 (1H, sep, CH, J = 7) 7.24 (1H, t, ar, J = 8) 7.38 (2H, t, ar, J = 8) 7.71 (1H, s, th) 7.81 (2H, d, ar, J = 8) 8.11 (1H, s, im) 0.89 (3H, t, CH3, J = 7.5) 1.34 (2H, sex, CH2, J = 7.5) 1.58 (2H, qui, CH2, J = 7.5) 2.73 (2H, t, CH2, J = 7.5) 7.22 (1H, t, ar, J = 8) 7.36 (2H, t, ar, J = 8) 7.71 (1H, s, th) 7.79 (2H, d, ar, J = 8) 8.12 (1H, s, im) 0.88 (3H, t, CH3, J = 7) 1.32 (4H, m, CH2) 1.62 (2H, qui, CH2, J = 7) 2.75 (2H, t, CH2, J = 7) 7.24 (1H, t, ar, J = 8) 7.38 (2H, t, ar, J = 8) 7.73 (1H, s, th) 7.81 (2H, d, ar, J = 8) 8.13 (1H, s, im) 0.86 (3H, t, CH3, J = 7) 1.29 (6H, m, CH2) 1.60 (2H, qui, CH2, J = 7) 2.74 (2H, t, CH2, J = 7) 7.23 (1H, t, ar, J = 8) 7.38 (2H, t, ar, J = 8) 7.71 (1H, s, th) 7.80 (2H, d, ar, J = 8) 8.13 (1H, s, im) 0.85 (3H, t, CH3, J = 7) 1.33 (4H, m, CH2) 1.77 (2H, m, CH2) 3.93 (1H, t, CH, J = 7) 7.20 (1H, t, ar, J = 8) 7.35 (2H, t, ar, J = 8) 7.65 (1H, s, im) 7.72 (2H, d, ar, J = 8) 0.84 (3H, t, CH3, J = 7) 1.26 (4H, m, CH2) 1.38 (2H, m, CH2) 1.77 (2H, m, CH2) 3.94 (1H, t, CH, J = 7) 7.20 (1H, t, ar, J = 8) 7.36 (2H, t, ar, J = 8) 7.64 (1H, s, im) 7.75 (2H, d, ar, J = 8) 0.83 (3H, t, CH3, J = 7) 1.23 (6H, s, CH2) 1.38 (2H, t, CH2, J = 7) 1.79 (2H, m, CH2) 3.94 (1H, t, CH, J = 7) 7.20 (1H, t, ar, J = 8) 7.36 (2H, t, ar, J = 8) 7.64 (1H, s, im) 7.73 (2H, d, ar, J = 8)

Abbreviations: T = thiophene, th = thiazole, im = imidazole, ar = aromatic. bRef. [24], m.p. 116118. cRef. [23], m.p. not reported. dRef. [24, 25], m.p. 125127.

IR showed an additional C=O stretching absorption. This compound was considered as the iminothiazolone 12 which, after reuxing with dilute hydrochloric acid, gave the imidazole 18. The structure of this new derivative was conrmed by IR, 1H-NMR (table I) and 13C-NMR, MS (see Experimental section).

The same reaction was repeated with 2-bromoacetophenone 9 and similar behaviour was observed: when R was ethyl or propyl, the intermediate salt was analogous to 11, leading to compounds 1921, whereas, when R was a longer chain, the precipitate was analogous to 12 (it led to compounds 2527) and the ltrate

886 contained the intermediate analogous to 11 which, after the usual treatment with dilute hydrochloric acid, gave the imidazothiazoles 2224. For a deeper insight into the structure of the carboxylic acids 18 and 2527, one of these, 2-(4-phenyl-2imidazolylsulfanyl)heptanoic acid 26, was subjected to DEPT, HETCOR, MS (see Experimental section) and to crystallographic studies. Its crystal structure shows that the imidazole and the phenyl ring are almost planar as observed in analogous compounds [17, 18], whereas the orientation of the carboxylic group is almost perpendicular to them. Further crystallographic details will be given elsewhere. 3. Biological results Table II reports the biological activity of the reference compounds (13) and of the newly synthesized imidazothiazole derivatives (1317 and 1924). Among the unexpected imidazole by-products, compounds 2627 were quite inactive. Consequently, the analogues 18 and 25 were not tested. The increase of hydrophobicity at position 2 in both 6-(2-thienyl) (1315) and 6-(3-thienyl) derivatives (16 and 17) decreased the inhibitory potency on complex I activity in comparison with the 2-methyl derivative 1. Therefore, position 2 seems to be very critical for the potency of complex I inhibition.
Table II. Biological activity of compounds 13, 1317, 1924 and 2627. Compound 1 2 3 13 14 15 16 17 19 20 21 22 23 24 26 27
a

Residual activity of NADH:UBQ reductase (%) a 12 24 80 63 91 63 63 100 82 17 100 73 46 0 114 107

Contrary to 6-thienyl derivatives, 6-phenyl derivatives with 2-substituents of increasing hydrophobicity either maintained or reduced the inhibitory capacity of the lead compound. There was no direct correlation between the increase in length and hydrophobicity of the substituent and inhibition, since the 2-ethyl (19) and the 2-butyl (22) derivatives had a markedly reduced potency with respect to the 2-propyl derivative (20). Nevertheless, 2-pentyl (23) and mostly 2-esyl (24) derivatives showed an increased inhibitory effect, thus suggesting that the length of the chain is important for the interaction with the complex I binding site. The different behaviour of the thienyl or phenyl derivatives indicated a slightly different interaction as inhibitors of complex I. In previous work we studied in detail the inhibitory properties of compound 1, which appeared to act as a quinol antagonist [13]. In order to clarify the relationship between the new phenyl derivatives and quinone interaction in complex I, compounds 2 and 20 were tested with two different substrates, namely the hydrophilic ubiquinone-1 (Q1) and the hydrophobic undecyl-benzoquinone (UBQ). The potency of 2 was higher with Q1 (I50 = 75 M) than with UBQ (I50 = 130 M), similarly to compound 1 [13]. Conversely, compound 20 showed the same I50 for both Q1 and UBQ (120 M). Thus, the inhibitory effect of the more hydrophobic 20 could depend on some internal reaction steps that are not extensively inuenced by the nature of the quinone substrate for complex I activity. Indeed, compound 20 acted as a classical non-competitive inhibitor with respect to both Q1 and UBQ, similarly to rotenone (gure 2) [8], while 2 showed mixed competition with UBQ, as previously reported for 1 [13]. Taken together, the biological data suggest that the substitution of a thienyl group with a bulkier phenyl group at position 6 could shift the inhibitory action of imidazothiazole compounds from antagonists of the quinol product to antagonists of the semiquinone intermediate. The latter would correspond to the action of rotenone in complex I [8]. 4. Discussion Structurally, most complex I inhibitors have a head-tail module that is critical for the antagonist action versus the ubiquinone substrate [8, 10]. The results of the present study clarify that complex I inhibition is strictly dependent on the substitutions at the extreme positions 2 and 6 of the imidazothiazole ring. The different potency of the substituents at the 2 position, when position 6 is occupied by either a thienyl or a phenyl group, indicates that steric constraints in the binding site limit the length and bulkiness of imidazothiazoles for their optimal interac-

The numbers are the average of at least three separate experiments. The assay conditions are described in the experimental section. UBQ (30 M) was used as electron acceptor. The nal concentration of the compounds was 0.25 mM.

887 A 5. Experimental 5.1. Chemistry Compounds 23 have been previously described [19]. The aldehydes 4 (butyraldehyde, valeraldehyde, hexanal, heptanal, octanal) and 2-bromoacetophenone 9 are commercially available. 2-(Bromoacetyl)thiophene 7 [20] and 3-(bromoacetyl)thiophene 8 [21] were prepared according to the literature. The bromoaldehydes 5 were prepared under experimental conditions analogous to those reported for the synthesis of -bromopropionaldehyde [22]. The imidazo[2,1-b]thiazoles 13 [23, 24], 14 [23] and 19 [25] were already reported in the literature. The melting points are uncorrected. Analyses (C, H, N) were within 0.4% of the theoretical values. TLC was performed on Bakerex plates (Silica gel IB2-F): the eluent was a mixture of petroleum ether/acetone in various proportions. The IR spectra (table I) were recorded in nujol on a Perkin-Elmer 683. The 1H-NMR (table I) and the 13C-NMR spectra were recorded on a Varian Gemini (300 MHz), and were referenced to solvent signals (DMSO). The EI mass spectra were recorded on a VG 7070E. X-ray crystallography: the data were collected on a Siemens P4 four-circle diffractometer. 5.1.1. 5-Alkyl-2-aminothiazoles 6 The appropriate bromoaldehyde 5 (100 mmol) was treated with 80 mmol of thiourea and stirred at 90100 C for 2 h. After cooling, the reaction mixture was treated with 2 N NaOH until basic and extracted with chloroform. Compounds 6 thus obtained were used, without further purication, in the following step. 5.1.2. Reaction of the 5-alkyl-2-aminothiazoles 6 with the bromoacetylthiophenes 78 (synthesis of 1318) The appropriate compound 6 (45 mmol) was dissolved in acetone (100 mL) and treated with 2-(bromoacetyl)thiophene 7 or 3-(bromoacetyl)thiophene 8 (45 mmol). The reaction mixture was reuxed for 36 h (according to a TLC test) and the resulting salt 1012 was treated, without further purication, with 200 mL of 2 N HCl. After 1 h reux, the solution was cautiously basied by adding dropwise 15% NH4OH. The resulting base (1318) was collected by ltration and crystallized from ethanol with a yield of 3540% (table I). The following spectra were also recorded: 1) as an example of compound 10, R = isopropyl was chosen, 1H-NMR: 1.21 (6H, d, CH3, J = 7); 3.01 (1H, sep, CH, J = 7); 5.69 (2H, s, CH2); 7.20 (1H, s, th); 7.38 (1H, m, T); 8.11 (1H, m, T); 8.18 (1H, m, T); 9.59 (1H, s, NH).

Figure 2. Effect of compound 2 and 20 on the Q1 (A) and UBQ (B) titration of the NADH:Q reductase activity. (q) Control. () In the presence of 0.18 mM compound 2. ([) In the presence of 0.18 mM compound 20.

tion with complex I. In phenyl derivatives, the aliphatic chain substituents in position 2 might mimic the hydrophobic tail of ubiquinone. A further lengthening of the chain which resembles the polyprenil tail of ubiquinone should better clarify if position 2 corresponds to the attachment point for hydrophobic substituents. Future studies with systematic substitutions at other positions such as nitrogen 7 will dene the determinants to optimize the ubiquinone antagonist action of imidazothiazoles and produce extremely potent inhibitors of complex I.

888 2) compound 12, IR: 3 5002 200, 1 760, 1 685, 1 640, 1 560 cm1. 1H-NMR: 0.87 (3H, t, CH3, J = 7); 1.40 (4H, m, CH2); 2.05 (2H, m, CH2); 4.90 (1H, t, CH, J = 7); 5.37 (2H, s, CH2); 7.58 (1H, m, T); 7.74 (1H, m, T); 8.78 (1H, m, T). MS: 296 (M+, 11), 278 (M-H2O, 4), 268 (M-CO, 9), 124 (C6H4SO, 30), 111 (C5H3SO, 100). 3) Compound 18, MS: 296 (M+, 13), 278 (M-H2O, 100), 235 (278-C3H7, 85), 207 (235-CO, 40), 182 (49), 123 (26), 109 (27), 55 (34). 5.1.3. Reaction of the 5-alkyl-2-aminothiazoles 6 with 2-bromoacetophenone 9 (synthesis of 1927) The appropriate compound 6 (10 mmol) was dissolved in acetone (50 mL) and treated with 10 mmol of 2-bromoacetophenone 9. The mixture was reuxed for 25 h (according to a TLC test) and worked up with two procedures. For compounds 1921, acetone was evaporated under reduced pressure and the resulting residue was reuxed for 1 h with 100 mL of 2 N HCl. The solution thus obtained was then basied with 15% NH4OH and the resulting base 1921 was crystallized from ethanol with a yield of 3035% (table I). For Compounds 2227 the precipitate was collected by ltration and worked up as above. Compounds 2527 were crystallized from ethanol with a yield of 15%. The ltrate, evaporated and worked up in the same manner, gave compounds 2224 which were crystallized from ethanol with a yield of 3%. The following data were also recorded for compound 26: 13 C-NMR (with DEPT and HETCOR): 13.80 (CH3), 21.90 (CH2), 26.13 (CH2), 30.81 (CH2), 31.83 (CH2), 50.29 (S-CH), 118.04 (im-5), 124.34 (ar-2 + 6), 126.56 (ar-4), 128.59 (ar-3 + 5), 133.13 (im-4), 138.09 (ar-1), 139.67 (im-2), 172.43 (COOH). MS: 304 (M+, 45), 286 (M-H2O, 8), 260 (M-CO2, 16), 229 (286-C4H9, 15), 203 (48), 190 (70), 176 (100), 117 (56). Crystal data: C16H20N2O2S, M = 304.4, Monoclinic, a = 17.305 (2), b = 8.5690 (10), c = 10.8620 (10) , = 99.67 (1), V = 1587.8 (3) 3 (by least-squares renement on diffractometer angles for 43 randomly selected and automatically centred reections), space group P21/c (n. 14), Z = 4, F (000) = 648, Dc = 1.27 g cm3, (Mo-K) = 0.21 mm1. 5.2. Biology 5.2.1. Mitochondrial preparation Mitochondria from beef heart were prepared according to standard procedures. They were used after at least two cycles of freezing and thawing in order to brake down membranes. 5.2.2. Biochemical assay Submitochondrial particles from beef heart were prepared essentially as described by Hansen and Smith [26].
[14] [15]

The redox activity of NADH:ubiquinone oxidoreductase was measured by using Q1 or UBQ as electron acceptor. The test compound, dissolved in DMSO, was added to 0.4 mg/mL of mitochondrial protein in 2 mL of 50 mM K2HPO4 buffer containing 1 mM EDTA, 2 mM KCN (pH 7.4) and 0.1 mM NADH at room temperature. The reaction was started by quinone and followed by decrease in absorbance of NADH in the dual wavelength mode at 350 minus 410 nm with an extinction of 5.5 mM1 cm1 as previously described [27].

Acknowledgements This work was supported by grants from the Italian National Research Council (CNR, Rome) and from the University of Bologna (Funds for Selected Research Topics).

References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] Walker J.E., Q. Rev. Biophys. 25 (1992) 253324. Videira A., Biochim. Biophys. Acta 1364 (1998) 89100. Yagi T., Yano T., Di Bernardo S., Matsuno-Yagi A., Biochim. Biophys. Acta 1364 (1998) 125133. Weyer G., Babej-Dolle R.M., Hadler D., Hofmann S., Hermann W.M., Neuropsychobiology 36 (1997) 7382. Schapira A.H.V., Biochim. Biophys. Acta 1364 (1998) 261270. Degli Esposti M., Ngo A., Mayers M.A., Diabetes 45 (1996) 15311534. Lummen P., Biochim. Biophys. Acta 1364 (1998) 287296. Degli Esposti M., Biochim. Biophys. Acta 1364 (1998) 222235. Degli Esposti M., Ghelli A., Ratta M., Cortes D., Estornell E., Biochem. J. 301 (1994) 161167. Miyoshi H., Biochim. Biophys. Acta 1364 (1998) 236244. Andreani A., Rambaldi M., Locatelli A., Andreani F., Poglayen G., Degli Esposti M., Eur. J. Med. Chem. 27 (1992) 729734. Andreani A., Rambaldi M., Locatelli A., Leoni A., Ghelli A., Degli Esposti M., Pharm. Acta Helv. 69 (1994) 1520. Andreani A., Rambaldi M., Leoni A., Locatelli A., Ghelli A., Ratta M., Benelli B., Degli Esposti M., J. Med. Chem. 38 (1995) 10901097. Lenaz G., Biochim. Biophys. Acta 1364 (1998) 207221. Degli Esposti M., Ngo A., McMullen G.L., Ghelli A., Sparla F., Benelli B., Ratta M., Linnane A.W., Biochem. J. 313 (1996) 327334. Oettmeier W., Masson K., Soll M., Reil E., Biochem. Soc. Trans. 22 (1994) 213216. Shilcrat S.C., Hill D.T., Bender P.E., Griswold D.E., Baures P.W., Eggleston D.S., Lantos I., Pridgen L.N., J. Heterocycl. Chem. 28 (1991) 11811187. Jain P.C., Acta Crystallogr. Sect. C: Cryst. Struct. Commun. 43 (1987) 24152418.

[16] [17]

[18]

889
[19] [20] [21] [22] [23] Pyl T., Giebelmann R., Beyer H., Liebigs Ann. Chem. 643 (1961) 145153. Kipnis F., Soloway H., Ornfelt J., J. Amer. Chem. Soc. 71 (1949) 1011. MacDowell D.W.H., Greenwood T.D., J. Heterocycl. Chem. 2 (1965) 4448. Reichel L., Jahns H.J., Liebigs Ann. Chem. 751 (1971) 6976. Tanaka M., Otsuka K., Obata M., Tanabe T., Saida T., Nomura K., Amemija K., Saga K., Kano S., Japan Kokai 74 70, 986; Chem. Abstr. 81 (1974) 136142c. Meakins G.D., Musk S.R.R., Robertson C.A., Woodhouse L.S., J. Chem. Soc. Perkin Trans. I (1989) 643648. ODaly M.A., Hopkinson C.P., Meakins G.D., Raybould A.J., J. Chem. Soc. Perkin Trans. I (1991) 855860. Hansen M., Smith A.L., Biochim. Biophys. Acta 81 (1964) 214222. Degli Esposti M., Ghelli A., Crimi M., Estornell E., Fato R., Lenaz G., Biochem. Biophys. Res. Commun. 190 (1993) 10901096.

[24] [25] [26] [27]

Eur. J. Med. Chem. 34 (1999) 895901 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

895

Original article

NO-Donors, part 3: nitrooxyacylated thiosalicylates and salicylates synthesis and biological activities#
Stefan Endresa, Andreas Hackerb, Eike Noackb, Georg Kojdab, Jochen Lehmanna*
a

Pharmazeutisches Institut, Rheinische Friedrich-Wilhelms-Universitt Bonn, An der Immenburg 4, D-53121 Bonn, Germany b Institut fr Pharmakologie, Heinrich-Heine-Universitt Dsseldorf, Moorenstr. 5, D-40225 Dsseldorf, Germany (Received 8 December 1998; revised 29 March 1999; accepted 31 March 1999)

Abstract Organic nitrates release nitric oxide when incubated with thiosalicylic acid. S-Nitrooxyacylated esters and amides of thiosalicylic acid, together with the corresponding salicylates, were synthesized in order to perform a rst in vitro evaluation of these new nitrate-thiol-hybrid prodrugs. These prodrugs might release NO in vivo after biotransformation without the use of endogenous reductives. None of these prodrugs released NO spontaneously when dissolved in buffer solution, but they did activate soluble guanylyl cyclase and induced vasodilatation of phenylephrine-pretreated male Wistar rat aorta in a potency range between that of isosorbiddinitrate and glycerole trinitrate. 1999 ditions scientiques et mdicales Elsevier SAS organic nitrates / nitric oxide / thiosalicylates / salicylates / vasodilatation

1. Introduction For more than a century organic nitrates have been used in the treatment of angina pectoris and congestive heart failure. The activity of these nitrates is mainly attributed to the release of nitric oxide (NO), an endogenous mediator with a rapidly growing list of physiological and pathophysiological functions [2]. The liberation of NO from organic nitrates in vivo is most likely an enzymatic metabolic reduction process [2, 3]. This does not however rule out the possibility of non-enzymatic reduction of nitrates by specic thiols. In general, all thiols reduce organic nitrates to inorganic nitrite, but only a very few of them are able to produce NO as well. The basic structural feature of these special thiols is a carbonyl function located two carbons away from a thiol group in a coplanar orientation as it is realized in cysteine, N-acetyl-cysteine and thiosalicylic acid [46]. Combining organic nitrate and a NO-liberating thiol in one

molecule leads to prodrugs which might be endowed with typical properties of organic nitrates and may show a more facilitated mechanism of NO-release, which also might reduce the development of nitrate tolerance. Some promising approaches have already been made for nitrate-cysteine combinations, primarily N-nitrooxyacylated cysteines [7] but also cysteinamides, nitrooxyacylated at the amide nitrogen [8], and isosorbiddinitrate connected with a cyclized L-cysteinamide [9]. We established S-nitrooxyacylation both for various cysteines and thiosalicylic acid derivatives. Here we give a rst report on the synthesis, stability and biological activities of S-nitrooxyacylthiosalicylates. To evaluate the inuence of the thiosalicylate substructure the analogue salicylates were synthesized as well. 2. Chemistry The target compounds (SE) could not be obtained from S-bromoacyl-thiosalicylates and silver nitrate. They were obtained instead by acylation [10] of various thiosalicylates and salicylates with nitratoacids, catalysed by carbonyldiimidazole (CDI) (gure 1).

For part 2 see [1] *Correspondence and reprints

896

Figure 1. General route to the target compounds.

The nitratoacids 3 were obtained by following two different methods. Treatment of the hydroxyesters 1 with a mixture of fuming nitric acid/acetanhydride followed by acid or base catalysed hydrolysis of the esters 2 (route A, gure 2), or treatment of the halogenated acids 4 with silver nitrate in dry acetonitrile (route B).

These nitratoacids were linked with thiosalicylates and salicylates via S- resp. O-acylation. Not all of the planned combinations of nitratoacids with thioles or phenoles were possible. Two nitratoacids (3a and d) underwent decomposition when activated with carbonyldiimidazole. 3b could be activated but decomposed as soon as a thiole

Figure 2. Synthesis of the nitratoacids.

897

Figure 3. Synthesis of the target compounds (part 1).

was added and compound 3e showed neither decomposition nor acylation. Thiosalicylic acid itself, as well as the N-unsubstituted thiosalicylamide, suffered rapid decomposition with all nitratoacids. Obviously there is a competition between reduction of the nitrate by the thiole [24] and formation of the thioesters. Figure 3 gives these target compounds which we have obtained successfully as stable solid compounds. O-Nitratoacylation of salicylamid did not yield the expected compound 8 but rather a mixture of the N-acylated 10 and the diacylated SE 161. A rearrangement producing N-acetyl- from O-acetyl-salicylamid is

described in the literature [11]. In order to obtain a completely O-acylated target compound we treated 9 with two equivalents of 3c and obtained pure SE 161. Finally, we nitrooxylated thiophenol (a thiole which is unable to release NO from organic nitrates) and obtained SE 135 (gure 4). 2.1. Stability Stability of the target compounds under simulated physiological conditions was investigated. Substances were dissolved in 50 mM phophate-buffer pH =

898

Figure 4. Synthesis of the target compounds (part 2).

7/acetonitrile (5:1) and stored at 37 C for 24 h. Before and after that period, HPLC analysis using an RP-8 column with UV-detector was performed to detect possible instabilities due to the combination of organic nitrates and NO releasing thioles resp. analogues. The diminishment of the starting material at 37 C/24 h ranged between 03.7% with the exception of SE 145 (16%) and SE 161 (18%) By means of its lipophilic character, SE 85 could not be chromatographed on this column. A sufficient thermal stability of this compound was certied by NMR-spectroscopy after 24 h at 37 C in DMSO. 3. Biological results and discussion 3.1. Spontaneous liberation of NO Liberation of NO was determined electrochemically by a Clark-type NO-sensitive electrode (Iso-NO, World Precision Instruments Inc., Berlin, Germany). Measurements were performed with 108 up to 104 M solutions of the compounds in Krebs-Henseleit buffer (pH 7.4) with constant stirring at 37 C. No release of NO was found with the SE-compounds. This conrms that an in vivo biotransformation of these prodrugs is indispensable. 3.2. Guanylyl cyclase activation The most active compound, SE 175, was chosen to verify that the vasorelaxation observed with the SE

compounds was mediated by sGC-stimulation. After treatment with 100 mM SE 175 the cGMP level in a rat aorta segment increased signicantly from 2.8 0.6 (n = 10) to 8.8 2.8 (n = 5) pmol cGMP/mg protein (P = 0.012). 3.3. Vasorelaxation of isolated vessel segments Vasorelaxation after precontraction with phenylephrine was measured using aorta segments of rats. All of the nine investigated nitrates proved to be potent vasodilators. Figure 1 gives the relaxation curves for the three most potent, and the activities (EC50 values) for all compounds in decreasing order. The benzylnitrate derivative showed to be the most active compound. The carrier of the nitrato group (thiosalicylic acid ester, thiosalicylic amide, thiophenol, salicylic acid ester or salicylic amide) does not seem to have any inuence on the biological activity in this in vitro assay. In vivo investigations, like decreasing blood pressure under repeated treatment, will be necessary to evaluate this inuence (gure 5). 4. Experimental protocols 4.1. Vasorelaxation of isolated vessel segments Method: after cervical translocation, the aorta of male Wistar rats (250300 g) was dissected free and rapidly immersed in cold oxygenated Krebs-Henseleit solution (pH 7.4). Four ring segments (5 mm) of the thoracic aorta were suspended in individual organ chambers (10 mL)

899 toluene was reuxed under argon for 48 h, washed with water, saturated NaHCO3 solution, and water again (100 mL each), then dried (Na2SO4) and evaporated. The residue was distilled. 15.9 g (73%) colourless oil; b.p. = 6769 C, 0.03 mm Hg ( [13]: 8793 C, 0.40 mm Hg); IR (KBr) cm1: 1 700 (C=O); 1H-NMR 3.83 (s, 3H, CH3), 5.26 (s, 1H, SH), 7.17.5 (m, 3H, aromat. H-3,4,5), 7.92 (dd, J = 7.1/2.4 Hz, 1H, aromat. H-6). 4.3.2. Ethyl thiosalicylate (6b) As described for 6a, using 15 mL ethanol. 16.7 g (70%) colourless oil; b.p. = 6972 C, 0.04 mm Hg ( [14]: b.p. not given); IR (KBr) cm1: 1 700 (C=O); 1 H-NMR 1.28 (t, J = 8.5 Hz, 3H, CH3), 4.41 (q, J = 8.5 Hz, 2H, CH2), 7.27.7 (m, 3H, aromat. H-3,4,5), 8.05 (dd, J = 7.2/2.2 Hz, 1H, aromat. H-6). The nitrooxycarboxylic acids were prepared as reported (3a [7, 15], 3c [7], 3d [15], 3e [15]), with the exception of the following: 4.3.3. 2-Nitrooxyisobutyric acid (3b) 6.28 mL of fuming HNO3, followed by 14.3 mL acetic acid anhydride were dropped into a stirred solution of 5.0 g (76 mmol) ethyl 2-hydroxyisobutyrate and 0.2 g urea in 500 mL CH2Cl2, maintaining a temperature < 10 C. After stirring at room temperature for 24 h, the mixture was poured into 800 mL of icewater. The organic layer was separated, washed with water, saturated NaHCO3 solution, and water again, then evaporated at < 40 C. Distillation of the oily residue produced 8.9 g (66%) of colourless ethyl 2-nitrooxyisobutyrate (b.p. = 3334 C, 0.04 mm Hg). For hydrolysis, 5.0 g (28 mmol) of this ester were dissolved in 250 mL of CH3OH and added at 10 C to a solution of 3.17 g (56 mmol) KOH in 20 mL of water. After stirring at room temperature for 2 h (TLC control) the mixture was acidied with concentrated hydrochloric acid and the solvent was evaporated. 100 mL water were added and the mixture was extracted with CH2Cl2 (2 200 mL). The organic layer was dried (Na2SO4), evaporated, and the residue crystallized from n-hexane. 2.80 g (67%) colourless crystals; m.p. = 76 C ( [16]: 7879 C); IR (KBr) cm1: 1 715 (C=O), 1 630 and 1 290 (N=O); 1H-NMR 1.59 (s, 6H, 2 CH3); 13.50 (s, 1H, COOH). Anal. C4H7NO5 (C, H, N). 4.3.4. 11-Nitrooxyundecanoic acid (3g) 5.0 g (18.85 mmol) 1l-bromoundecanoic acid in 30 mL of dry acetonitrile were added to 3.52 g (20.72 mmol) of AgNO3 dissolved in 30 mL acetonitrile and stirred for 3 h at 60 C. The ltrate of the reaction mixture was poured into 250 mL of icewater and the precipitated product was separated, washed with water and dried. 3.38 g (73%) white powder; m.p. = 4041 C; IR (KBr) cm1: 1 700

Figure 5. Relaxation-dose curves (n = 910) and halfmaximal vasorelaxing activity (EC50 values, given in mol/L, n = 210) of nitrooxyacylated thiosalicylates and salicylates in thoracic aorta vessels of rats.

lled with Krebs buffer. The solution was aerated continuously with 95% O2 and 5% CO2 maintained at 37 C. A linear force transducer (Statham force displacement transducer) recorded the actual tension of the aortic rings. After equilibration at a resting tension of 4 g (1 h) contractile function of segments was tested by application of KCl (60 mM). The vasorelaxing activity was studied by cumulative application of the compounds after precontraction with a single dose of phenylephrine (1 M). 4.2. Activity of soluble guanylyl cyclase Activity of soluble guanylyl cyclase was measured as reported earlier [12] by formation of [32P]-cGMP from [-32P]-GTP. 4.3. Chemistry Melting points were determined in open capillary tubes on a digital Gallenkamp melting point apparatus and are not corrected. The IR spectra were recorded on a PerkinElmer 1420 using KBr pellets for solids and NaCl plates for liquid substances. The 1H-NMR spectra were obtained on a Bruker WM 200 (200 MHz) spectrometer using DMSO-d6 as the solvent. Chemical shifts are reported in = ppm relative to tetramethylsilane as the internal standard. Elemental analyses were carried out with a Herus CHN-O-Rapid or an Elementar Vario EL and were within 0.4% of the calculated values. 4.3.1. Methyl thiosalicylate (6a) A mixture of 20.0 g (0.13 mol) thiosalicylic acid, 5.0 g 4-toluenesulfonic acid, 30 mL methanol and 500 mL

900 (C=O), 1 620 and 1 280 (N=O); 1H-NMR 1.201.80 (m, 16H, 8 CH2); 2.20 (t, J = 7.8 Hz, 2H, CO-CH2); 4.50 (t, J = 7.8 Hz, 2H, CH2-ONO2); 12.00 (s, 1H, COOH). Anal. C11H21NO5 (C, H, N). 4.3.5. 4-Nitrooxymethylbenzoic acid (3f) 5.38 g 4-bromomethylbenzoic acid in 30 mL of dry acetonitrile were added to 4.95 g (29.0 mmol) silver nitrate in acetonitrile and stirred overnight. The ltrate of the mixture was poured into 500 mL of icewater. The precipitated crude product was separated, dried in vacuo and recrystallized from diisopropylether. 4.12 g (84%) white powder; m.p. = 165 C; IR (KBr) cm1: 1 690 (C=O), 1 670 and 1 270 (N=O); 1H-NMR 5.65 (s, 2H, CH2-ONO2); 7.58 (d, J = 8.1 Hz, 2H, aromat. H-3,5); 7.98 (d, J = 8.1 Hz, 2H, aromat. H-2,4). Anal. C8H7NO5 (C, H, N). 4.3.6. Ethyl S-(11-nitrooxyundecanoyl)-thiosalicylate (SE 85) 3.0 g (12.1 mmol) of 11-nitrooxyundecanoic acid (3g) and 2.21 g (12.1 mmol) ethyl thiosalicylate (6b) were dissolved in 30 mL of dry CH2Cl2. 2.73 g (13.23 mmol) DCC, dissolved in 20 mL CH2Cl2, were added with protection by argon and stirred for 24 h. The solid was separated, the ltrate washed with 0.1 N hydrochloric acid (3 30 mL) and dried (Na2SO4). After evaporation, the crude product was chromatographed on a silica gel column with ethylacetate. 0.86 g (17%) yellow paste; IR (NaCl) cm1: 1 710 (C=O), 1 630 and 1 725 (N=O); 1 H-NMR 1.251.75 (m, 19H, 8 CH2, CH2-CH3), 2.65 (t, J = 7.2 Hz, 2H, CO-CH2-(CH2)8), 4.25 (q, J = 7.1 Hz, 2H, CH2-CH3), 4.50 (t, J = 6.6 Hz, 2H, -CH2-ONO2), 7.507.70 (m, 3H, aromat. H-3,4,5); 8.82 (dd, J = 6.8/2.2 Hz, 1H, aromat. H-6). Anal. C20H29NO6S (C, H, N). 4.3.7. S-(3-Nitrooxypivaloyl)-thiophenole (SE 135) 1.63 g (10.0 mmol) of 3c in 50 mL of dry DMF were cooled to 10 C and 1.78 g (11.0 mmol) of CDI were added. After stirring (argon) for 2 h, 1.10 g (10 mmol) of thiophenol were added and the reaction mixture was stirred for another 2 h at 10 C. 50 mL ethylacetate were added, the mixture washed with saturated NaCl solution (3 30 mL), dried over Na2SO4 and evaporated. The residue was chromatographed on a silica gel column with petrolether/acetone (7:1). 0.7 g (28%) colourless oil; IR (NaCl) cm1: 1 690 (C=O), 1 640 and 1 275 (N=O); 1 H-NMR 1.35 (s, 6H, C-CH3), 4.68 (s, 2H, CH2ONO2), 7.347.52 (m, 5H, aromat. H). Anal. C11H13NO4S(C, H, N). 4.3.8. Methyl O-(3-nitrooxypivaloyl)-salicylate (SE 136) 1.52 g (10.0 mmol) methyl salicylate were treated with 1.63 g (10.0 mmol) 3c and 1.78 g (11.0 mmol) of CDI as described for SE 135. The crude product was chromatographed on a silica gel column with petroether/ ethylacetate (5:1). 2.37 g (80%) white powder; m.p. = 4750 C; IR (KBr) cm1: 1 750 (C=O), 1 730 (C=O); 1 625 and 1 280 (N=O); 1H-NMR 1.39 (s, 6H, C-CH3), 3.80 (s, 3H, O-CH3), 4.75 (s, 2H, CH2-ONO2), 7.20 (dd, J = 8.1/1.0 Hz, 1H, aromat. H-3), 7.42 (ddd, J = 8.1/1.0 Hz, 1H, aromat. H-5), 7.68 (ddd, J = 8.1/1.5 Hz, 1H, aromat. H-4), 7.93 (dd, J = 8.1/1.5 Hz, 1H, aromat. H-6). Anal. C13H15NO7 (C, H, N). 4.3.9. Methyl S-(3-nitrooxypivaloyl)-thiosalicylate (SE 145) 2.58 g (15.3 mmol) 5a were treated with 2.50 g (15.3 mmol) 3c and 2.73 g (16.86 mmol) CDI as described for SE 135. The crude product was chromatographed on a silica gel column with petrolether/ ethylacetate (5:1). 3.19 g (67%) white crystals; m.p. = 3839 C (crystallized from petrolether in the cold); IR (KBr) cm1: 1 730 (C=O), 1 690 (C=O), 1 630 and 1 275 (N=O); 1H-NMR 1.35 (s, 6H, C-CH3), 3.79 (s, 3H, O-CH3), 4.66 (s, 2H, CH2-ONO2), 7.507.70 (m, 3H, aromat. H-3,4,5), 7.88 (dd, J = 7.6/2.0 Hz, 1H, aromat. H-6). Anal. C13H15NO6S (C, H, N). 4.3.10. Ethyl S-(3-nitrooxypivaloyl)-thiosalicylate (SE 152) Synthesized from 2.0 g (10.81 mmol) ethyl thiosalicylate (6b), 1.76 g (10.81 mmol) 3c and 1.93 g (11.89 mmol) CDI as described for SE 135. The crude product was chromatographed on a silica gel column with petroleum ether/ethylacetate (4:1). 2.42 g (68%) yellowish oil; IR (NaCl) cm1: 1 720 (C=O), 1 700 (C=O), 1 635 and 1 280 (N=O); 1H-NMR 1.28 (t, J = 7.1 Hz, 2H, CH2-CH3), 1.35 (s, 6H, C-CH3), 4.26 (q, J = 7.1 Hz, 2H, CH2-CH3), 4.65 (s, 2H, CH2-ONO2), 7.507.65 (m, 3H, aromat. H-3,4,5), 7.87 (dd, J = 7.1/2.0 Hz, 1H, aromat. H-6). Anal. C14H17NO6S (C, H, N). 4.3.11. O-(3-Nitrooxypivaloyl)-salicylic acid dimethylamide (SE 157) Synthesized from 1.65 g (10.0 mmol) salicylic acid dimethylamide (7b) [17], 1.63 g (10.0 mmol) 3c and 1.78 g (11.0 mmol) CDI as described for SE 135. The crude product was chromatographed on a silica gel column with ethylacetate. 3.0 g (96%) colourless oil; IR (NaCl) cm1: 1 750 (C=O), 1 650 (C=O), 1 640 and 1 280 (N=O); 1H-NMR 1.32 (s, 6H, C-CH3); 2.76 (s, 3H, N-CH3); 2.94 (s, 3H, N-CH3); 4.68 (s, 2H, CH2-ONO2); 7.19 (d, J = 8.1 Hz, 1H, aromat. H-3); 7.337.36 (m, 2H,

901 aromat. H-4,5); 7.48 (m, 1H, aromat. H-6). Anal. C14H18N2O6 (C, H, N). 4.3.12. S-(3-Nitrooxypivaloyl)-thiosalicylic acid dimethylamide (SE 158) Synthesized from 1.81 g (10.0 mmol) 7a [17], 1.63 g (10.0 mmol) 3c and 1.78 g (11.0 mmol) CDI as described for SE 135. The crude product was chromatographed on a silica gel column with ethylacetate. 1.79 g (55%) thick brown oil; IR (NaCl) cm1: 1 685 (2 C=O), 1 630 and 1 280 (N=O); 1H-NMR 1.32 (s, 6H, C-CH3); 2.64 (s, 3H, N-CH3); 2.95 (s, 3H, N-CH3); 4.65 (s, 2H, CH2ONO2); 7.347.60 (m, 4H, aromat. H). Anal. C14H18N2O5S (C, H, N). 4.3.13. N, O-Di-(3-nitrooxypivaloyl)-salicylamide (SE 161) Synthesized from 1.0 g (7.29 mmol) salicylamide, 2.38 g (14.58 mmol) 3c and 2.48 g (15.31 mmol) CDI as described for SE 135. The crude product was chromatographed on a silica gel column with petroleum ether/ acetone (1:1). 1.12 g (36%) white powder; m.p. = 7375 C; IR (KBr) cm1: 1 760 (C=O), 1 675 (2 C=O), 1 630 and 1 275 (N=O); 1H-NMR 1.27 (s, 6H, C-CH3); 1.33 (s, 6H, C-CH3); 4.67 (s, 2H, CH2-ONO2); 4.68 (s, 2H, CH2-ONO2); 7.21 (d, J = 7.6 Hz, 1H, aromat. H-3); 7.36 (dd, J = 7.6/7.1 Hz, 1H, aromat. H-5); 7.48 (dd, J = 7.1/1.5 Hz, 1H, aromat. H-6); 7.56 (ddd, J = 7.6/7.6/1.5 Hz, 1H, aromat. H-4); 10.93 (s, 1H, NH). Anal. C17H21N3O10 (C, H, N). 4.3.14. Methyl S-(4-nitrooxymethylbenzoyl)-thiosalicylate (SE 175) Synthesized from 0.50 g (2.97 mmol) 6a, 0.55 g (2.97 mmol) 3f and 0.53 g (3.27 mmol) CDI as described for SE 135. The crude product was chromatographed on a silica gel column with petroleum ether/acetone (1:1). 0.53 g (51%) white powder; m.p. = 6164 C; IR (KBr) cm1: 1 715 (C=O), 1 660 (C=O), 1 620 and 1 270 (N=O); 1H-NMR 3.76 (s, 3H, O-CH3); 5.70 (s, 2H, CH2-ONO2); 7.607.75 (m, 5H, aromat. H-3,4,5,3,5); 7.908.08 (m, 3H, aromat. H-6,2,6). Anal. C16H13NO6S (C, H, N). 4.3.15. Stability The HPLC equipment used was a Consta Metric 3200 (LCD-Analytical) pump with a Spherisorb 5 ODS 2 pre(20 4 mm) and main-column (250 4 mm), a Spektro Monitor 3200 (LCD-Analytical) UV detector and an HP 4496 Series II (Hewlett Packard) integrator. Eluent: acetonitrile/phosphoric acid 0.1% (50:50); samples: 3 mg substance in 10 mL 50 mM phosphate buffer pH 7.4/ acetonitrile (1:5; ow: 1 mL/min; column temperature: 2025 C; detection wavelength: 190 mm. Samples were heated at 37 C for 24 h in a common dry oven.

References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] Lehmann J., Kahlich R., Meyer Zum Gottesberge C., Fricke U., Arch. Pharm. Pharm. Med. Chem. 330 (1997) 247252. Vallance P., Br. J. Clin. Pharmacol. 45 (1998) 433439. Bennett B.M., McDonald B.J., James S.T., J. Pharmacol. Exp. Ther. 261 (1992) 716723. Ahlner J., Andersson R.G.G., Torfgrd K., Axelsson K.L., Pharmacol. Rev. 43 (1991) 351423. Feelisch M., Noack E.A., Eur. J. Pharmacol. 139 (1987) 1930. Chung S., Fung H.L., Biochem. Pharmacol. 42 (1991) 14331439. Htter J., Noack E., (Schwarz Pharma AG) (1990) E P 89116700 9; ref C A 113, 212672x. Ishihara S., Saito F., Yoshioka T., Koike H., Miyake S., Mizuno H., (1993) PCT Int. Appl. WO 9303, 163. Nallet J.P., Dreux J., Berdeaux A., Richard V., Martorana P., Bohn H., (1991) PCT Int. Appl. WO 9303, 037. Gais H.J., Angew. Chem. 89 (1977) 251252. Gordon A.J., Tetrahedron 23 (1967) 863870. Kojda G., Kottenberg K., Hacker A., Noack E., Pharm. Acta Helv. 73 (1998) 2735. Hellwinkel D., Bohnet S., Chem. Ber. 120 (1987) 11511174. Baldwin D., Duckworth P.A., Erickson G.R., Lindoy L.F., McPartlin M., Aust. J. Chem. 40 (1987) 18611872. McCallum K.S., Emmons W.D., J. Org. Chem. 21 (1956) 367368. Ustarshchikov B.F., Podgornova V.A., Dormidontova N.V., Farberov M.J., Dokl. Akad. Nauk. SSSR 157 (1964) 143144. Schindlbauer H., Monatsh. Chem. 99 (1968) 17991807.

Eur. J. Med. Chem. 34 (1999) 903917 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

903

Original article

Synthesis of oxypropanolamine derivatives of 3,4-dihydro-2H-1,4-benzoxazine, -adrenergic affinity, inotropic, chronotropic and coronary vasodilating activities
Kriton Iakovoua*, Michalis Kazanisa, Andreas Vavayannisa, Giancarlo Brunib, Maria Raffaella Romeob, Paola Massarellib, Shuji Teramotoc, Hiroyuki Fujikic, Toyoki Moric
a Department of Pharmacy, Division of Pharmaceutical Chemistry, University of Athens, Panepistimiopolis Zografou, GR-157 71 Athens, Greece b Istituto di Farmacologia, Universita di Siena, 53100 Siena, Italy c 2nd Tokushima Institute of New Drug Research, Otsuka Pharmaceutical Co. Ltd., 463-10 Kagasuno, Kawauchi-cho, Tokushima 771-01, Japan

(Received 18 January 1999; accepted 26 April 1999)

Abstract A series of oxypropanolamine derivatives of 3,4-dihydro-2H-1,4-benzoxazine were synthesized and evaluated for inotropic, chronotropic and coronary vasodilating activities in the canine heart, affinity to 1-adrenergic receptor in turkey erythrocytes and affinity to the 2-adrenergic receptor in the rat lung. Among these compounds, 4-acetyl-6-(3-tert-butylamino-2-hydroxy)propoxy-3,4-dihydro-2H-1,4benzoxazine showed 2.1-fold more potent affinity to the 1 receptor than propranolol and 7-(3-tert-butylamino-2-hydroxy)propoxy-N-butyryl3,4-dihydro-2H-1,4-benzoxazine showed 2.5-fold more potent affinity to the 2 receptor and furthermore 4 386-fold more potent selectivity to the 2 receptor than propranolol. In addition, 4-acetyl-6-[3-(3,4-dimethoxybenzyl)amino-2-hydroxy]propoxy-3,4-dihydro-2H-1,4benzoxazine showed 1.1-fold more potent affinity to the 1 receptor than propranolol and also 1 147-fold more potent selectivity to the 1 receptor. With a few exceptions, negative inotropic and chronotropic actions of these compounds were dependent on the size of the 4-substituent obeying the order: unsubstituted < acetyl < propanoyl < butanoyl, while the benzoyl substituent conferred even stronger negative actions in the 6-oxypropanolamine derivatives. Neither negative inotropic and chronotropic actions related with affinity to 1-adrenoceptor nor coronary vasodilator action related with affinity to 2-adrenoceptor were observed. 4-acetyl-7-[3-(3,4-dimethoxybenzyl)amino-2hydroxy]propoxy-3,4-dihydro-2H-1,4-benzoxazine exerted potent positive inotropic, chronotropic and coronary vasodilating actions. The inotropic and chronotropic actions of the latter compound may be attributed to the release of intrinsic catecholamines, as concluded by the absence of 1-adrenoceptor affinity and disappearance of activity in the presence of a -blocker. 1999 ditions scientiques et mdicales Elsevier SAS 1,4-benzoxazine / oxypropanolamines / -adrenoceptor affinity / cardiovascular effects

1. Introduction A number of 1,4-benzoxazine derivatives [14] have been synthesized so far and various pharmacological activities have been reported with this class of molecules. Ethanolamine and oxypropanolamine derivatives of 1,4benzothiazine active on the adrenergic system are already known [5, 6], as well as of the isoster nuclei 1,4benzoxazine [710] and 3,4-dihydro-2(1H)-quinolinone
*Correspondence and reprints

[11, 12]. Among the carbostyril derivatives, 5-(3-tertbutylamino-2-hydroxypropoxy)-3,4-dihydro-2(1H)-quinolinone (carteolol) [13, 14] is a -blocker more potent than propranolol and is apparently devoid of the side effects which are usually associated with -blocking therapy as it retains an intrinsic sympathomimetic activity. Oxypropanolamine derivatives of 1,4-benzodioxin [15, 16], which are analogues of 1,4-benzoxazine, have been recently synthesized. The 4-acyl derivatives of 7-oxypropanolamine-3,4-dihydro-2H-1,4-benzoxazines which are included in the present work may be consid-

904

Figure 1. Structure of the synthesized 4-acyl derivatives of 7-oxypropanolamine-3,4-dihydro-2H-1,4-benzoxazines compared to practolol.

ered as dicyclic analogues of practolol, a prototype cardioselective blocking agent, as shown in gure 1 [9, 17, 18]. It is well known that the catecholamine -stimulants (arylethanolamines and aryloxypropanolamines) have been used as cardiotonic agents in the management of heart failure. Unfortunately these compounds have some disadvantages, such as powerful vasodilating effects, which cause a reex rise in heart rate, lack of oral absorption and short half-life [19]. On the other hand, some non-oxypropanolamine [20] or oxypropanolamine derivatives of 2(1H)-quinolinone [21, 22], have demonstrated remarkable positive inotropic activities, the latter in some cases without having -agonistic action. Furthermore, a series of non-oxypropanolamine derivatives of 3,4-dihydro-1,4-benzoxazine [23] have exhibited potent, long-acting positive inotropic and peripheral vasodilating activities. In order to extend the investigation on inotropic, chronotropic and vasodilating activities of non-catechol derivatives with possible affinity to the -adrenergic receptors, we prepared molecules structurally similar to the above. The synthesized compounds are novel 4-acyl substituted 3,4-dihydro-2H-1,4-benzoxazines bearing the typical oxypropanolamine chain at positions 6- and 7- of the aromatic ring. 2. Chemistry The target compounds 2449 (tables I and II) were prepared in a standard three-step procedure, shown in gure 2, which involves: a) synthesis of the 6and 7-hydroxy-4-acyl-3,4-dihydro-2H-1,4-benzoxazines [24], b) reaction thereof with epichlorhydrin in the presence of potassium carbonate [25] and c) treatment of the resulting 1-aryloxy-2,3-epoxypropanes with the appropriate amines [26]. The derivatives 27 (table III) were synthesized through the reaction of 6- or 7-hydroxy-3,4-dihydro-2H-

1,4-benzoxazine in aqueous medium with the appropriate anhydrides. The reaction of 6- or 7-hydroxy-3,4-dihydro2H-1,4-benzoxazine with benzoic or chloroacetic anhydride was not successful in water and therefore ethyl acetate was used instead. The intermediates 12 and 13 (table III) were synthesized via the corresponding 4-chloroacetyl compounds 10 and 11 (table III) which were condensed with diethylamine in ethanol. The general synthetic procedure for the hydroxy derivatives 213 is shown in gure 2. The 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazine was synthesized by a method which is presented in gure 3. Initially 7-methoxy-3,4-dihydro-2H-1,4-benzoxazin-3one [2729] was converted into 1 through reduction with lithium aluminium hydride in anhydrous tetrahydrofuran [30]. The intermediate 1 was demethylated by means of concentrated hydrobromic acid to give 7-hydroxy-3,4dihydro-2H-1,4-benzoxazine hydrobromide, which was alkalized by concentrated ammonium hydroxide to afford the free base [31]. The synthetic route of 6-hydroxy-3,4-dihydro-2H-1,4benzoxazine, as shown in gure 4, was achieved as follows: 2,5-dimethoxyaniline was reuxed with 2-bromoethanol in the presence of calcium carbonate in water to yield 2,5-dimethoxy-N-(2-hydroxyethyl)aniline [32] which was isolated in pure form by distillation in vacuo. The latter was reuxed with concentrated hydrobromic acid leading to 6-hydroxy-3,4-dihydro-2H-1,4-benzoxazine hydrobromide [32], which in turn, was alkalized by concentrated ammonium hydroxide to provide the free base. The 4-unsubstituted products 4447 (table II) were prepared, as shown in gure 2, by hydrolysis of the corresponding 4-acetyl derivatives 2427 with potassium hydroxide in aqueous methanol [33]. IR absorption and NMR spectra were in conformity with the structures expected. However, we observed that the resonance of the aromatic proton 5 either appears as a broad peak or disappears completely, depending on the temperature. This is possibly due to stereochemical

905
Table I. 4-Acyl-6- and 7-(3-isopropylamino- and 3-tert-butylamino-2-hydroxypropoxy)-3,4-dihydro-2H-1,4-benzoxazines and 4-acetyl-6and 7-[3-(3,4-dimethoxybenzyl)amino-2-hydroxy-propoxy]-3,4-dihydro-2H-1,4-benzoxazines.

Compound 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 48 49
a

Position 7 7 6 6 7 7 6 6 7 7 6 6 7 7 6 6 7 7 6 6 7 6

R1 CH3 CH3 CH3 CH3 CH2CH3 CH2CH3 CH2CH3 CH2CH3 CH2CH2CH3 CH2CH2CH3 CH2CH2CH3 CH2CH2CH3 Ph Ph Ph Ph CH2N(C2H5)2 CH2N(C2H5)2 CH2N(C2H5)2 CH2N(C2H5)2 CH3 CH3

R2 iPr tBu iPr tBu iPr tBu iPr tBu iPr tBu iPr tBu iPr tBu iPr tBu iPr tBu iPr tBu 3,4-dimethoxybenzyl 3,4-dimethoxybenzyl

M.p. (C) 176179a 161164a 134a 169171a 120a 203a 129131a 140142a 175178a 171172a 117120a 142145a 9698 8385 176180a 112115b 5456 8590 6870 7073 140142b 138140b

Yield (%) 24 25 20 25 26 25 27 23 25 24 26 23 57 54 17 21 15 13 16 14 31 33

Formula C16H24N2O4. 0.5 C4H4O4 C17H26N2O4. 0.5 C4H4O4 C16H24N2O4. 0.5 C4H4O4 C17H26N2O4. 0.5 C4H4O4 C17H26N2O4. 0.5 C4H4O4 C18H28N2O4. 0.5 C4H4O4 C17H26N2O4. 0.5 C4H4O4 C18H28N2O4. C4H4O4 C18H28N2O4. 0.5 C4H4O4 C19H30N2O4. 0.5 C4H4O4. 0.5 H2O C18H28N2O4. 0.5 C4H4O4 C19H30N2O4. 0.5 C4H4O4 C21H26N2O4 C22H28N2O4 C21H26N2O4. 0.5 C4H4O4 C22H28N2O4. C4H4O4. 0.5 H2O C20H33N3O4. 0.5 H2O C21H35N3O4. H2O C20H33N3O4. 0.5 H2O C21H35N3O4. 0.5 H2O C22H28N2O6. C4H4O4. 2H2O C22H28N2O6. C4H4O4. 2H2O

: fumarate, b: maleate.

effects caused by the equilibrium between two extreme conformations of the morpholine ring. A further study of this effect is still under investigation and will be published elsewhere.

3. Pharmacology The biological proles of the compounds listed in table IV on 1 and 2 adrenoceptors were respectively

Table II. 6- and 7-(3-Isopropylamino- and 3-tert-butylamino-2-hydroxypropoxy)-3,4-dihydro-2H-1,4-benzoxazines.

Compound 44 45 46 47
a

Position 7 7 6 6

R2 iPr tBu iPr tBu

M.p. (C ) 130134 7680b 128130c 138140a

a

Yield (%) 19 17 20 22

Formula C14H22N2O3. C15H24N2O3. C14H22N2O3. C15H24N2O3. C4H4O4. 0.5 H2O 2 C4H4O4 2 C2H2O4 C4H4O4

: fumarate, b: maleate, c: oxalate.

906

Figure 2. General synthetic procedure of the target 6- and 7-(alkylamino-2-hydroxypropoxy)-3,4-dihydro-2H-1,4-benzoxazines.

907
Table III. 4-Acyl-6- and 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazines.

Compound 2 3 4 5 6 7 8 9 10 11 12 13

Position 7 6 7 6 7 6 7 6 7 6 7 6

R1 CH3 CH3 CH2CH3 CH2CH3 CH2CH2CH3 CH2CH2CH3 Ph Ph CH2CI CH2CI CH2N(C2H5)2 CH2N(C2H5)2

M.p (C ) 152154 185188 104106 146148 123125 102103 190193 155 120122 128130 110113 124125

Yield (%) 72 65 79 84 80 79 65 61 82 57 57 63

Solvents of crystallization a a a a a a a a a a b b

a: benzene-pentane, b: cyclohexane.

assessed on turkey erythrocytes and on rat lung. The inotropic (changes in contractile force, CF), chronotropic (changes in sinus rate, SR) and coronary vasodilator (changes in coronary blood ow, CBF) effects of the synthesized products were evaluated in the isolated blood-perfused preparations of canine heart.

4. Results and discussion 4.1. 1-Adrenenoceptor binding The affinities of the 4-acetyl substituted 6-oxypropanolamines obeyed the order: 27 > 49 > 26. The

Figure 3. Preparation of 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazine.

Figure 4. Preparation of 6-hydroxy-3,4-dihydro-2H-1,4-benzoxazine.

908
Table IV. Inhibition of [3H]DHA binding on 1 and 2 adrenoreceptors. Compound 1 Ki ( s.e.) (M) Propranolol 24a 25a 26a 27a 28a 29a 30a 31a 32a 33a 34a 35a 36 37 39b 40 41 42 43 44a 45b 46c 47a 48b 49b
a

2 Ki ( s.e.) (M) 0.132 0.116 0.327 0.684 0.782 0.296 0.131 0.214 0.121 0.188 0.282 0.849 0.121 0.628 2.5 109 6.4 107 3.6 105 2.8 109 1.0 109 3.3 108 1.6 108 inactive 1.6 107 1.3 106 1.0 109 5.6 106 5.5 106 8.2 106 2.3 107 4.5 106 1.2 107 inactive inactive 2.5 105 5.3 106 1.2 106 3.4 107 4.5 108 inactive 2.5 106 0.171 0.201 0.319 0.267 0.326 0.715 0.220 0.359 0.264 0.136 0.454 0.481 2.720 0.234 0.459 0.169

1/2 selectivity ratio 1.520 0.106 12.290 0.350 1.280 0.012 0.013 0.014 555 0.0003 681 61 4.13

1.6 109 6.0 106 2.9 106 7.4 108 8.01 1010 2.7 106 1.3 106 1.2 106 1.1 105 2.0 109 2.9 106 inactive 8.1 109 1.3 107 5.6 107 inactive inactive 6.1 107 7.1 105 3.1 106 1.9 108 inactive inactive 2.7 108 inactive 1.4 109

0.606 0.684 0.306 0.186

0.296 0.106

0.651 0.386 0.172 0.238 0.185

7.95 280

1.70 1 744

0.306

: fumarate, b: maleate, c: oxalate.

observed affinity order of the side-chain amino substituents was tert-Bu > 3,4-dimethoxybenzyl > iPr. The highest affinities observed with derivatives 27 and 49, were respectively 2.1- and 1.1-fold higher than that of propranolol. For the 6-iPr derivatives the elongation from the acetyl to the propanoyl chain led to a signicant drop in affinity (30). Additional lengthening of the chain by a methylene group led to a loss in affinity (34). In the case of the corresponding 6-tert-Bu-derivatives, the presence of the propanoyl group (31) reduced the affinity, whereas the replacement thereof with butanoyl (35) led to increased activity. In the group of the 4-unsubstituted derivatives, 44 and 47 showed affinities ca. 10-fold lower than that of propranolol, while 45 and 46 had no affinity to the 1 adrenoceptor. The 4-acetyl substituted 7oxypropanolamine compounds (24 and 25) showed a low degree of affinity. For the 7-oxypropanolamine derivatives, the elongation from the acetyl to the propanoyl chain led to little increase in affinity (28 and 29). The observed affinity order for the amino substituents follows

the order tert-Bu > iPr. Additional lengthening of the chain by a methylene group increased the affinity (32). The replacement of the iPr group of 32 with tert-Bu led to a drastic decrease in affinity (33). Among the 4-benzoyl and 4-diethylaminoacetyl derivatives (36, 37, 39, and 4043) only compound 36 showed an EC50 in the 107 range. 4.2. 2-Adrenenoceptor binding The high affinities for the 2-adrenenoceptor were observed with the 4-acetyl substituted 6-oxypropanolamine derivatives. The affinities of 26 and 27 were 1.1-fold lower and 2.5-fold higher than that of propranolol, respectively. On the contrary, the 7-isomers 24 and 25 had low affinities to the 2-adrenenoceptor. As to the 4-acetyl substituted 6-oxypropanolamine derivatives, the change of the amino substituent from iPr or tBu into 3,4-dimethoxybenzyl dramatically decreased the afnity (49), whereas it led to a loss of affinity of the correspond-

909 ing 7-isomer (48). For the 6-derivatives the elongation from the acetyl to the propanoyl or butanoyl chain led to a substantial reduction, up to deletion, of affinity (30, 31, 34, and 35). On the contrary, the lengthening of the acyl chain of the 7-isomers, led to considerable increase in affinity, except for compound 32. The 4-unsubstituted derivatives 4447 showed low affinity to the 2adrenoceptor. The 4-benzoyl (36, 37, and 39) and 4-diethylaminoacetyl derivatives (4043) showed practically no affinity. 4.3. 1/2 selectivity Compound 32 showed considerable 1-adrenergic afnity with a high 1/2 selectivity ratio (365 times more than that of propranolol). Compound 35 was 5-fold less 1-affinitive than propranolol, but 448-fold more 1selective. Compound 44 was 11-fold less 1-affinitive than propranolol, but 185-fold more 1-selective. Compound 49 was 1.1-fold more 1-affinitive than propranolol and also 1 147-fold more 1-selective. Compounds 28 and 29 exhibited a moderate 2-affinity, but considerable 2-selectivity (about 120-fold more than that of propranolol). It is noteworthy that compound 33 was 2.5-fold more 2-affinitive and furthermore 4 386-fold more 2-selective than propranolol. The results discussed in the preceding paragraphs 4.1 to 4.3 and presented in table IV, have shown that it is impossible to draw any straightforward conclusions regarding structure-activity relationships. For example, although compound 24 has the closest structure to practolol (gure 1), it did not show the expected activity, which in fact was much lower than its 6-substituted isomer 26. It is possible that other parameters, such as the difference in steric conguration or the absence of amidic hydrogen, affect the biological activity of these cyclic analogues of practolol. Another example lies in the group of N-acyl 7-substituted compounds, where each additional methylene in the acyl chain may have a drastic effect on receptor affinity and selectivity. Although this fact is obvious in the cases of compounds 25 vs. 29 and 28 vs. 32 it cannot serve as a general rule. 4.4. Inotropic, chronotropic and coronary vasodilating activities As shown in gures 57, negative inotropic and chronotropic actions of the tested compounds were dependent on the size of the 4-substituent. In the series of the 6-oxypropanolamine derivatives the actions followed the order: unsubstituted (46 and 47) < acetyl (26 and 27) < propanoyl (30 and 31) < butanoyl (34 and 35). The benzoyl compounds (38 and 39) exerted stronger nega-

Figure 5. Structure-inotropic activity relationships of 6- and 7-oxypropanolamine-3,4-dihydro-2H-1,4-benzoxazine derivatives in canine isolated blood-perfused heart preparations. Effects of test compounds on right ventricular papillary muscle contractile force (CF) were expressed by the % changes from basal CF. Test compounds were dissolved in DMSO and the action of the solvent itself was subtracted for compensation.

tive actions. Similar structure-activity relationships were observed in the series of 7-oxypropanolamine derivatives, i.e. unsubstituted (44 and 45) < acetyl (24 and 25) < propanoyl (28 and 29) < butanoyl (32 and 33), but no further negative action was observed in the benzoyl subsutituted compounds 36 and 37. The negative inotropic and chronotropic actions of these compounds were difficult to explain by the 1-adrenoceptor antagonistic actions because of inconsistency with the 1adrenoceptor affinity as described earlier. In general, coronary vasodilator actions of the 6-oxypropanolamine derivatives were more potent than those of their 7-substituted counterparts (gure 7). It was also difficult to explain the coronary vasodilator action of these compounds in relation to their 2-adrenoceptor agonistic actions. The 4-diethylaminoacetyl substituted 6-oxypropanolamine derivatives with iPr (42) and tert-Bu (43) functions, as well as the 4-acetyl substituted compound bearing a 3,4-dimethoxybenzyl group (49), exerted weak positive inotropic and chronotropic actions. The same was true for the isomer of 42 bearing the side chain at position 7 (40). The 4-unsubstituted compounds 44 and 45 showed moderate positive actions (gure 8). The 3,4-dimethoxybenzyl derivative 48 exerted potent positive inotropic, chronotropic and coronary vasodilatory actions (gure 8). It should be noted that the inotro-

910

Figure 6. Structure-chrononotropic activity relationships of 6and 7-oxypropanolamine-3,4-dihydro-2H-1,4-benzoxazine derivatives in canine isolated blood-perfused heart preparations. Effects of test compounds on right atrial sinus rate (SR) were expressed by changes from basal SR. Test compounds were dissolved in DMSO and the action of the solvent itself was subtracted for compensation. Figure 8. Effects of 6- and 7-oxypropanolamine-3,4-dihydro2H-1,4-benzoxazine derivatives on inotropic, chronotropic, and coronary vasodilator actions in canine isolated blood-perfused heart preparations. Effects of test compounds on the right ventricular papillary muscle contractile force (CF) were expressed by the % changes from basal CF, on the right atrial sinus rate (SR) were expressed by changes from basal SR and coronary blood ow (CBF) through anterior septal arteries were expressed by the mL/min changes from basal CBF. Test compounds were dissolved in DMSO and the action of the solvent itself was subtracted for compensation.

pic and chronotropic actions of this compound were inhibited by pretreatment with the -blocker carteolol (gure 9).

5. Experimental protocols 5.1. Chemistry Melting points were determined on a Bchi micro melting point apparatus without correction. 1H-NMR and 13 C-NMR spectra were recorded on a 200 MHz Bruker AC 200 spectrometer in CDCl3 or DMSO-d6 using tetramethylsilane as internal standard. All target compounds were analysed for C, H and some additionally analysed for N. Elemental analyses indicated by the symbols of the elements or functions were within 0.4% of the theoretical values. 5.1.1. 7-Methoxy-3,4-dihydro-2H-1,4-benzoxazine 1 A solution of 7-methoxy-3,4-dihydro-2H-1,4benzoxazin-3-one (0.028 mol) in dry tetrahydrofuran (95 mL) was added dropwise to a suspension of lithium aluminum hydride (0.066 mol) in dry tetrahydrofuran

Figure 7. Structure-coronary vasodilator activity relationships of 6- and 7-oxypropanolamine-3,4-dihydro-2H-1,4-benzoxazine derivatives in canine isolated blood-perfused heart preparations. Effects of test compounds on coronary blood ow (CBF) through anterior septal arteries were expressed by the mL/min changes from basal CBF. Test compounds were dissolved in DMSO and the action of the solvent itself was subtracted for compensation.

911

Figure 9. Effects of -blocker carteolol on the positive inotropic and chronotropic response induced by compound 48 (4-acetyl-7[3-(3,4-dimethoxybenzyl)amino-2-hydroxy]propoxy-3,4-dihydro-2H-1,4-benzoxa-zine) in canine isolated blood-perfused preparations. The atrial and papillary muscle preparations were preloaded with 2 g and 1 g respectively.

(75 mL). The mixture was stirred and reuxed for 3 h and cooled. Aq. sodium hydroxide 5% (40 mL) was added dropwise under cooling and continuous stirring. The mixture was stirred at room temperature for 1 h and the liquid part separated through a lter. The solution was dried over anhydrous sodium sulfate and evaporated in vacuo to give an oil. Yield 85%. 1H-NMR (CDCl3, 200 MHz) 3.00 (brs, 1H, NH), 3.35 (brs, 2H, CH2NH), 3.7 (s, 3H, CH3O), 4.22 (t, 2H, OCH2, J = 4.3Hz), 6.386.58 (m, 3H, arom). 5.1.2. 7-Hydroxy-3,4-dihydro-2H-1,4-benzoxazine A solution of 1 (0.04 mol) in 25 mL concentrated hydrobromic acid 62% (0.528 mol) was stirred and reuxed for 2 h. The reaction mixture was then basied with concentrated ammonium hydroxide 28% (32 mL) and evaporated in vacuo. Ethyl acetate (250 mL) was added to the residue and the resulting mixture was stirred for 1 h and ltered. The ltrate was dried over anhydrous sodium sulfate and evaporated in vacuo to give 0.031 mol of 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazine. Yield: 85%, m.p.: 9597 C (benzene-hexane). 1H-NMR (DMSO-d6, 200 MHz) 2.07 (s, 1H, NH), 3.17 (t, 2H, CH2NH, J = 4.3 Hz), 4.06 (t, 2H, OCH2, J = 4.3 Hz), 6.116.16 (m, 2H, arom), 6.41 (d, 1H, arom, J = 8.8 Hz), 8.62 (s, 1H, OH).

5.1.3. 2,5-Dimethoxy-N-(2-hydroxyethyl)aniline A mixture of 2,5-dimethoxy-aniline (0.19 mol), calcium carbonate (0.0135 mol), water (150 mL) and 2-bromoethanol (0.24 mol) was reuxed for 4 h. The reaction mixture was cooled, then extracted with ethyl acetate (300 mL) and evaporated in vacuo. The oily residue was fractionated in vacuo (160 C, 2 mm Hg). Yield: 41%, m.p.: 4347 C. 1H-NMR (CDCl3, 200 MHz) 2.15 (s, 1H, NH), 3.25 (t, 2H, NHCH2, J = 5.5 Hz), 3.713.84 (m, 9H, 2CH3O, CH2OH and OH), 6.15 (dd, 1H, arom, J = 8.5 Hz, 2.9 Hz), 6.24 (d, 1H, arom, J = 2.8 Hz), 6.65 (d, 1H, arom, J = 8.6 Hz). 5.1.4. 6-Hydroxy-3,4-dihydro-2H-1,4-benzoxazine A solution of 2,5-dimethoxy-N-(2-hydroxyethyl)aniline (0.02 mol) in 20 mL of concentrated hydrobromic acid 62% (0.44 mol) was stirred and reuxed for 2 h. The reaction mixture was basied with concentrated ammonium hydroxide 28% (22 mL) and evaporated in vacuo. Ethyl acetate (120 mL) was added to the residue and the resultant mixture was stirred for 1 h and ltered. The ltrate was dried over anhydrous sodium sulfate and evaporated in vacuo to give 0.015 mol of 6-hydroxy-3,4dihydro-2H-1,4-benzoxazine. Yield: 76%, m.p.: 108109 C (benzene). 1H-NMR (DMSO-d6, 200 MHz) 3.19 (t, 2H, CH2NH, J = 3.9 Hz), 3.97 (t, 2H, OCH2, J

912 = 4.2 Hz), 5.62 (brs, 1H, NH), 5.83 (dd, 1H, arom, J = 8.4 Hz, 2.8 Hz), 5.97 (d, 1H, arom, J = 2.2 Hz), 6.38 (d, 1H, arom, J = 8.4 Hz), 8.52 (s, 1H, OH). 5.1.5. 4-Acyl-6- and 7-hydroxy-3,4-dihydro-2H-1,4benzoxazines 27 To a suspension of 6- or 7-hydroxy-3,4-dihydro-2H1,4-benzoxazine (0.02 mol) in water (12 mL), acetic or propanoic or butanoic anhydride (0.025 mol) was added. The reaction mixture was heated in a water bath for 15 min. After cooling, ethyl acetate (130 mL) was added and the resulting mixture was stirred, ltered and extracted with ammonium hydroxide 14% (24 mL). The alkaline solution was discarded and the organic layer was washed with water (100 mL). The organic layer was separated, shaken with hydrochloric acid 5% (35 mL), water (35 mL), dried over anhydrous sodium sulfate and evaporated to dryness under reduced pressure. 1H-NMR (DMSO-d6, 200 MHz) 2: 2.15 (s, 3H, NCOCH3), 3.75 (t, 2H, CH2N, J = 4.4 Hz), 4.16 (brs, 2H, OCH2), 6.246.30 (m, 2H, arom), 7.03 (brs, 1H, arom), 7.34 (s, 1H, OH). Anal. (C10H11NO3) C, H, N. 3: 2.21 (s, 3H, NCOCH3), 3.77 (t, 2H, CH2N, J = 4.4 Hz), 4.14 (t, 2H, OCH2, J = 4.5 Hz), 6.43 (dd, 1H, arom, J = 8.6 Hz, 2.2 Hz), 6.66 (d, 1H, arom, J = 8.7 Hz), 8.97 (brs, 1H, OH). Anal. (C10H11NO3) C, H, N. 1H-NMR (CDCl3, 200 MHz) 4: 1.13 (t, 3H, CH3, J = 6.2 Hz), 2.56 (q, 2H, NCOCH2, J = 6.8 Hz), 3.89 (s, 2H, CH2N), 4.24 (t, 2H, OCH2, J = 4.8 Hz), 6.356.40 (m, 2H, arom), 6.93 (brs, 1H, arom), 7.34 (s, 1H, OH). Anal. (C11H13NO3) C, H, N. 1 H-NMR (DMSO-d6, 200 MHz) 5: 1.03 (t, 3H, CH3, J = 6.2 Hz), 2.502.57 (q, 2H, NCOCH2, J = 7.0 Hz), 3.79 (t, 2H, CH2N, J = 4.1 Hz), 4.13 (t, 2H, OCH2, J = 4.0 Hz), 6.46 (dd, 1H, arom, J = 8.6 Hz, 2.1 Hz), 6.67 (d, 1H, arom, J = 8.6 Hz), 7.32 (brs, 1H, arom), 8.95 (brs, 1H, OH). Anal. (C11H13NO3) C, H, N. 1H-NMR (CDCl3, 200 MHz) 6: 0.91 (t, 3H, CH3, J = 7.2 Hz), 1.61.8 (m, 2H, COCH2CH2), 2.51 (t, 2H, COCH2CH2, J = 7.6 Hz), 3.9 (brs, 2H, CH2N), 4.24 (t, 2H, OCH2, J = 4.7 Hz), 5.3 (brs, 1H, OH), 6.356.45 (m, 2H, arom), 6.9 (brs, 1H, arom). Anal. (C12H15NO3) C, H, N. 7: 0.94 (t, 3H, CH3, J = 7.3 Hz), 1.601.76 (m, 2H, COCH2CH2), 2.54 (t, 2H, COCH2CH2, J = 7.4 Hz), 3.68 (brs, 1H, OH), 3.87 (brs, 2H, CH2N), 4.21 (t, 2H, OCH2, J = 4.7 Hz), 6.57 (dd, 1H, arom, J = 8.9 Hz, 2.7 Hz), 6.75 (d, 1H, arom, J = 8.9 Hz), 7.34 (s, 1H, arom). Anal. (C12H15NO3) C, H, N. 5.1.6. 4-Benzoyl-6- and 7-hydroxy-3,4-dihydro-2H-1,4benzoxazines 8 and 9 A mixture of 6- or 7-hydroxy-3,4-dihydro-2H-1,4benzoxazine (0.02 mol) and benzoic anhydride (0.02 mol) in ethyl acetate (50 mL) was reuxed for 1 h. The mixture was then cooled, ltered, extracted with ammonium hydroxide 14% (60 mL). The alkaline solution was discarded and the organic phase was washed with water, shaken with hydrochloric acid 10% (25 mL) and nally dried over anhydrous sodium sulfate and concentrated in vacuo. 1H-NMR (DMSO-d6, 200 MHz) 8: 3.80 (brs, 2H, CH2N), 4.24 (brs, 2H, OCH2), 6.106.30 (m, 2H, arom), 7.377.52 (m, 6H, arom), 9.37 (s, 1H, OH), IR (Nujol): OH 3 389 cm1, C=O 1 710 cm1. Anal. (C15H13NO3) C, H, N. 1H-NMR (DMSO-d6, 200 MHz) 9: 3.77 (t, 2H, CH2N, J = 4.4 Hz), 4.18 (t, 2H, OCH2, J = 4.3 Hz), 6.43 (dd, 1H, arom, J = 8.4 Hz, 2.0 Hz), 6.68 (d, 1H, arom, J = 8.4 Hz), 6.83 (brs, 1H, arom), 7.407.55 (m, 5H, arom), 8.9 (s, 1H, OH), IR (Nujol): OH: 3 406 cm1, C=O: 1 720 cm1. Anal. (C15H13NO3) C, H, N. 5.1.7. 4-Chloroacetyl-6- and 7-hydroxy-3,4-dihydro-2H1,4-benzoxazines 10 and 11 Compounds 10 and 11 were synthesized through reaction of 6- or 7-hydroxy-3,4-dihydro-2H-1,4-benzoxazine with chloroacetic anhydride in ethyl acetate according to the previous method. 1H-NMR (DMSO-d6, 200 MHz) 10: 3.79 (t, 2H, CH2N, J = 4.5 Hz), 4.22 (t, OCH2, J = 4.1 Hz), 4.55 (s, 2H, CH2CI), 6.256.33 (m, 2H, arom), 7.75 (brs, 1H, arom), 9.45 (brs, 1H, OH). IR (Nujol): OH: 3 327 cm1, C=O: 1 688 cm1. Anal. 1 (C10H10ClNO3) C, H, N. H-NMR (DMSO-d6, 200 MHz) 11: 3.82 (t, 2H, CH2N, J = 4.4 Hz), 4.19 (t, 2H, OCH2), 4.60 (s, 2H, CH2CI), 6.47 (dd, 1H, arom), 6.68 (d, 1H, arom), 7.35 (s, 1H, arom), 9.03 (s, 1H, OH). IR (Nujol): OH: 3 315 cm1, C=O: 1 679 cm1. Anal. (C10H10ClNO3) C, H, N. 5.1.8. 4-Diethylaminoacetyl-6- and 7-hydroxy-3,4dihydro-2H-1,4-benzoxazines 12 and 13 A mixture of 10 or 11 (0.012 mol) and diethylamine (0.034 mol) in absolute ethanol (100 mL) was reuxed for 3 h. The mixture was then cooled, ltered and evaporated in vacuo. After addition of ethyl acetate (70 mL) the mixture was stirred, ltered and the ltrate was concentrated in vacuo. Water (20 mL) was added to the obtained viscous oil and the resulting precipitate was collected by ltration. 1H-NMR (DMSO-d6, 200 MHz) 12: 0.94 (t, 6H, N(CH2CH3)2, J = 6.7 Hz), 2.482.58 (m, 4H, N(CH2CH3)2), 3.44 (s, 2H, COCH2N), 3.84 (brs, 2H, CH2N), 4.19 (brs, 2H, OCH2), 6.226.30 (m, 2H, arom), 7.75 (brs, 1H, arom), 9.34 (brs, 1H, OH). Anal. (C14H20N2O3) C, H, N. 13: 0.95 (t, 6H, N(CH2CH3)2, J = 7.0 Hz), 2.482.58 (m, 4H, N(CH2CH3)2), 3.41 (s, 2H, COCH2N), 3.86 (t, 2H, CH2N, J = 4.2 Hz), 4.15 (t,

913 2H, OCH2, J = 4.2 Hz), 6.42 (dd, 1H, arom, J = 8.7 Hz, 2.4 Hz), 6.65 (d, 1H, arom J = 8.7 Hz), 7.46 (brs, 1H, arom), 8.92 (s, 1H, OH). Anal. (C14H20N2O3) C, H, N. 5.1.9. 4-Acyl-6- and 7-(2,3-epoxypropoxy)-3,4-dihydro2H-1,4-benzoxazines 1423 A mixture of 29, 12 or 13 (0.013 mol), epichlorohydrin (0.11 mol) and potassium carbonate (0.026 mol) was stirred at 9095 C for 3 h. After cooling, ethyl acetate (50 mL) or chloroform was added and the mixture was stirred, ltered and extracted with sodium hydroxide 5% (40 mL). The organic phase was dried over anhydrous sodium sulfate and evaporated in vacuo to give a viscous oil. All epoxides were used in the next step without purication, except for compounds 15 and 20 which were puried as follows: the crude oily residue was worked up with hot cyclohexane and ltered. After cooling the ltrate, compound 15 separated as an oily layer which was isolated after decantation and evaporation of the remaining solvent (yield 59%), whereas compound 20 was isolated as a white crystalline solid (m.p. 7577 C, yield 65%). 1H-NMR (CDCl3, 200 MHz) 15: 2.32 (s, 3H, COCH3), 2.72 (dd, 1H, CH2-oxiranic, J = 4.7 Hz, 2.6 Hz), 2.88 (t, 1H, CH2-oxiranic, J = 4.5 Hz), 3.293.33 (m, 1H, CH-oxiranic), 3.803.89, 4.154.25 (2m, 6H, CH2O, CH2CH2N, OCH2CH2), 6.66 (dd, 1H, arom, J = 8.9 Hz, 2.5 Hz), 6.8 (d, 1H, arom, J = 9.0 Hz), 6.95 (brs, 1H, arom). Anal. (C13H15NO4) C, H, N. 20: 2.7 (dd, 1H, CH2-oxiranic, J = 4.8 Hz, 2.6 Hz), 2.87 (t, 1H, CH2-oxiranic, J = 4.5 Hz), 3.263.35 (m, 1H, CHoxiranic), 3.84 (dd, 1H, CH2O, J = 10.9 Hz, 5.6 Hz), 3.953.97 (m, 2H, CH2CH2N), 4.13 (dd, 1H, CH2O, J = 11.0 Hz, 3.0 Hz), 4.34.32 (m, 1H, OCH2CH2), 6.26 (d, 1H, arom, J = 6.9 Hz), 6.44 (d, 1H, arom, J = 2.8 Hz), 6.91 (brs, 1H, arom), 7.347.48 (m, 5H, arom). Anal. (C18H17NO4) C, H, N. 5.1.10. 4-Acyl-6- and 7-(3-isopropylamino- and 3-tertbutylamino-2-hydroxypropoxy)-3,4-dihydro-2H-1,4-benzoxazines 2443 A mixture of 1423 (0.008 mol) and isopropylamine or tert-butylamine (0.08 mol) in isopropyl alcohol (90 mL) was reuxed for 3 h. The resulting solution was thoroughly evaporated in vacuo, the viscous oily residue was worked up with hot cyclohexane, ltered immediately and the solvent was evaporated. Compounds 4043 were insoluble in hot cyclohexane and were converted directly into salts. With the exception of 36 and 37 which were isolated as crystalline solids, the oily amines were converted into the fumarates (dry acetone-diethyl ether), except for 39 which was converted into the maleate (ethyl acetate). Because the salts of the amines 4043 could not be crystallized, the bases were puried as follows: the oily salts were diluted in water, made alkaline with sodium hydroxide, extracted with chloroform and the organic layer was dried over anhydrous sodium sulfate and evaporated in vacuo. 1H-NMR (CDCl3, 200 MHz) 24: 1.06 (d, 6H, CH(CH3)2, J = 6.3 Hz), 2.25 (s, 3H, COCH3), 2.39 (brs, 2H, OH and NH), 2.622.89 (m, 3H, CH(CH3)2 and OCH2CH(OH)CH2), 3.883.99 (m, 5H, OCH2CH(OH)CH2, CH2CH2N), 4.24 (t, 3H, OCH2CH2, J = 4.7 Hz), 6.436.49 (m, 2H, arom), 6.93 (brs, 1H, arom). 1H-NMR (DMSO-d6, 200 MHz) 24 (fumarate): 1.15 (dd, 6H, CH(CH3)2, J = 6.2 Hz, 1.9 Hz), 2.18 (s, 3H, COCH3), 2.753.19 (m, 3H, CH(CH3)2 and OCH2CH (OH)CH2), 3.354.50 (m, OCH2CH(OH)CH2, CH2CH2N, OCH2CH2, NH, OH, HOOCCH=CHCOOH), 6.37 (s, 1H, HOOCCH=CHCOOH), 6.426.55 (m, 2H, arom), 7.20, 7.90 (2brs, 1H, arom). Anal. C18H26N2O6 (C, H, N). 1H-NMR (CDCl3, 200 MHz). 25: 1.08 (s, 9H, C(CH3)3), 2.252.83 (m, 7H, COCH3, OH, NH and OCH2CH(OH)CH2), 3.90 (s, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.24 (t, 3H, CH(CH3)2 and OCH2CH (OH)CH2) 3.883.99 (m, 5H, OCH2CH(OH)CH2, CH2CH2N), 4.24 (t, 2H, OCH2CH2, J = 4.5 Hz), 6.446.49 (m, 2H, arom), 6.97 (brs, 1H, arom). 1H-NMR (DMSO-d6, 200 MHz) 25 (fumarate): 1.2 (s, 9H, C(CH3)3), 2.2 (s, 3H, COCH3), 2.73.1 (m, 2H, OCH2CH(OH)CH2), 3.204.40 (m, OCH2CH(OH)CH2, CH2CH2N, OCH2CH2, NH, OH, HOOCCH=CHCOOH), 6.37 (s, 1H, HOOCCH=CHCOOH), 6.426.55 (m, 2H, arom), 7.20, 7.90 (2brs, 1H, arom). 13C-NMR (DMSOd6) 22.26, 25.47, 30.27, 43.81, 52.84, 53.63, 59.3, 65.62, 69.87, 101.76, 106.21, 121.01, 124.27, 135.29, 145, 167.94, 169. Anal. C19H28N2O6 (C, H, N). 1H-NMR (CDCl3, 200 MHz) 26: 1.08 (d, 6H, CH(CH3)2, J = 6.3 Hz), 2.29 (s, 3H, COCH3), 2.642.91 (m, 3H, CH(CH3)2 and OCH2CH(OH)CH2), 3.14 (brs, 2H, NH and OH), 3.864.23 (m, 7H, OCH2CH(OH)CH2, CH2CH2N and OCH2CH2), 6.66.8 (m, 2H, arom). 1H-NMR (DMSOd6, 200 MHz) 26 (fumarate): 1.1 (d, 6H, CH(CH3)2, J = 6.3 Hz), 2.24 (s, 3H, COCH3), 2.613.09 (m, 3H, CH(CH3)2 and OCH2CH(OH)CH2), 3.204.20 (m, OCH2CH (OH)CH2, CH2CH2N, OCH2CH2, NH, OH, HOOCCH=CHCOOH), 6.37 (s, 1H, HOOCCH= CHCOOH), 6.64 (dd, 1H, arom, J = 6.9 Hz), 6.79 (d, 1H, arom, J = 7.1 Hz). Anal. C20H28N2O8 (C, H, N). 1 H-NMR (CDCl3, 200 MHz) 27: 1.09 (s, 9H, C(CH3)3), 2.31 (s, 3H, COCH3), 2.572.84 (m, 2H, OCH2CH(OH)CH2), 3.89 (brs, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.22 (t, 2H, OCH2CH2, J = 4.7 Hz), 6.65 (dd, 1H, arom, J = 8.9 Hz, 2.6 Hz), 6.79 (d, 1H, arom, J = 9.0 Hz). Anal. C19H28N2O6 (C, H). 28: 1.091.26 (m, 9H, CH(CH3)2 and COCH2CH3), 2.502.92 (m, 5H,

914 COCH2CH3, CH(CH3)2 and OCH2CH(OH)CH2), 3.654.21 (m, 9H, OH, NH, OCH2CH(OH)CH2, CH2CH2N and OCH2CH2), 6.356.44 (m, 2H, arom), 6.95 (brs, 1H, arom). Anal. C19H28N2O6 (C, H, N). 29: 11.25 (m, 12H, C(CH3)3 and COCH2CH3), 2.472.87 (m, 6H, COCH2CH3, OCH2CH(OH)CH2, OH and NH), 3.653.98 (m, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.174.27 (m, 2H, OCH2CH2), 6.46.5 (m, 2H, arom), 6.95 (brs, 1H, arom). Anal. C20H28N2O6 (C, H, N). 30: 1.07 (d, 6H, CH(CH3)2, J = 6.3 Hz), 1.17 (t, 3H, COCH2CH3, J = 6.2), 2.522.88 (m, 7H, COCH2CH3, CH(CH3)2, OCH2CH(OH)CH2, OH and NH), 3.844.05 (m, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.22 (t, 2H, OCH2CH2, J = 4.0 Hz), 6.64 (dd, 1H, arom, J = 7.5 Hz, 2.0 Hz), 6.77 (d, 1H, arom, J = 8.7 Hz). Anal. C19H28N2O6 (C, H). 31: 1.06 (s, 9H, C(CH3)3), 1.15 (t, 3H, COCH2CH3, J = 6.2), 2.42.85 (m, 6H, COCH2CH3, OCH2CH(OH)CH2, OH and NH), 3.86 (brs, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.19 (t, 2H, OCH2CH2, J = 4.0 Hz), 6.64 (dd, 1H, arom, J = 7.5 Hz, 2 Hz), 6.77 (d, 1H, arom, J = 8.7 Hz). Anal. C22H30N2O8 (C, H). 32: 0.91 (t, 3H, COCH2CH2CH3, J = 7.2 Hz), 1.07 (d, 6H, CH(CH3)2, J = 6.2 Hz), 1.61.8 (m, 2H, COCH2CH2CH3), 1.83 (brs, 2H, OH and NH), 2.51 (t, 2H, COCH2CH2CH3, J = 7.6 Hz), 2.632.9 (m, 3H, CH(CH3)2 and OCH2CH(OH)CH2), 3.894.07 (m, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.24 (t, 2H, OCH2CH2, J = 4.7 Hz), 6.456.5 (m, 2H, arom), 6.97 (brs, 1H, arom). Anal. C20H30N2O6 (C, H). 33: 0.871.13 (m, 12H, COCH2CH2CH3 and C(CH3)3), 1.61.8 (m, 2H, COCH2CH2CH3), 1.86 (brs, 2H, OH and NH), 2.452.89 (m, 4H, COCH2CH2CH3 and OCH2CH(OH)CH2), 3.9 (brs, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.25 (brs, 2H, OCH2CH2), 6.46.5 (m, 2H, arom), 6.97 (brs, 1H, arom). Anal. C21H32N2O6 (C, H). 34: 0.95 (t, 3H, COCH2CH2CH3, J = 7.2 Hz), 1.08 (d, 6H, CH(CH3)2, J = 6.2 Hz), 1.62 (m, 4H, COCH2CH2CH3, OH and NH), 2.52.9 (m, 5H, COCH2CH2CH3, CH(CH3)2, and OCH2CH(OH)CH2), 3.854.04 (m, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.22 (t, 2H, OCH2CH2, J = 4.7 Hz), 6.64 (dd, 1H, arom, J = 7.5 Hz, 2.0 Hz), 6.8 (d, 1H, arom, J = 8.7 Hz). Anal. C20H30N2O6 (C, H). 35: 0.95 (t, 3H, COCH2CH2CH3, J = 7.2 Hz), 1.09 (s, 9H, C(CH3)3), 1.61.8 (m, 2H, COCH2CH2CH3), 2 (brs, 2H, OH and NH), 2.52.88 (m, 4H, COCH2CH2CH3 and OCH2CH(OH)CH2), 3.87 (s, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.22 (t, 2H, OCH2CH2, J = 4.7 Hz), 6.67 (dd, 1H, arom, J = 7.5 Hz, 2.0 Hz), 6.78 (d, 1H, arom, J = 8.7 Hz). Anal. C21H32N2O6 (C, H). 36: 1.05 (d, 6H, CH(CH3)2, J = 6.2 Hz), 2.1 (brs, 2H, NH and OH), 2.62.86 (m, 3H, CH(CH3)2 and OCH2CH(OH)CH2), 3.853.99 (m, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.32 (t, 2H, OCH2CH2, J = 4.4 Hz), 6.25 (d, 1H, arom, J = 7.5 Hz), 6.44 (d, 1H, arom, J = 2.7 Hz), 7.297.48 (m, 5H, arom). Anal. C21H26N2O4 (C, H, N). 37: 1.20 (s, 9H, C(CH3)3), 2.602.95 (m, 4H, OCH2CH(OH)CH2, OH and NH), 3.874.00 (m, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.30 (d, 2H, OCH2CH2, J = 3.7 Hz), 6.25 (d, 1H, arom, J = 6.9 Hz), 6.44 (d, 1H, arom, J = 2.8 Hz), 7.347.48 (m, 5H, arom). Anal. C22H28N2O4 (C, H, N). 38: 1.10 (d, 6H, CH(CH3)2, J = 6.2 Hz), 2.553.00 (m, 3H, CH(CH3)2 and OCH2CH(OH)CH2), 3.554.47 (m, 7H, OCH2CH(OH)CH2, CH2CH2N and OCH2CH2), 6.556.84 (m, 1H, arom), 7.307.57 (m, 6H, arom), 8.03 (d, 1H, arom, J = 8.2 Hz). Anal. C23H28N2O6 (C, H). 1 H-NMR (DMSO-d6, 200 MHz) 39 (maleate): 1.27 (d, 9H, C(CH3)3), J = 4.5 Hz), 2.713.00 (m, 2H, OCH2CH (OH)CH2), 3.654.29 (m, 7H, OCH2CH(OH)CH2, CH2CH2N and OCH2CH2), 5.836.00 (m, 2H, CH=CHmaleic), 6.646.90 (m, 1H, arom), 7.497.68 (m, 6H, arom), 8.02 (d, 1H, arom, J = 8.2 Hz), 8.28 (brs, 1H, COOH). Anal. C26H32N2O8 (C, H). 1H-NMR (CDCl3, 200 MHz) 40: 0.91.25 (m, 12H, CH(CH3)2, and N(CH2CH3)2), 1.702.10 (m, 6H, N(CH2CH3)2, OH and NH), 2.502.85 (m, 3H, CH(CH3)2 and OCH2CH (OH)CH2), 3.15 (d, 2H, NCOCH2N, J = 5.7 Hz), 3.40 (brs, H2O), 3.713.91 (m, 5H, OCH2CH(OH)CH2 and CH2CH2N), 4.19 (brs, 2H, OCH2CH2), 6.40 (d, 2H, arom, J = 2.5 Hz), 6.63 (d, 1H, arom, J = 9.5 Hz). Anal. C20H33N3O4 (C, H). 41: 0.851.30 (m, 15H, C(CH3)3 and N(CH2CH3)2), 1.702.10 (m, 6H, N(CH2CH3)2, OH and NH), 2.502.85 (m, 2H, OCH2CH(OH)CH2), 3.13 (brs, 2H, NCOCH2N), 3.40 (brs, H2O), 3.74.2 (m, 7H, OCH2CH(OH)CH2, CH2CH2N and OCH2CH2), 6.346.69 (m, 3H, arom). Anal. C21H35N3O4 (C, H). 42: 0.91.27 (m, 12H, CH(CH3)2, and N(CH2CH3)2), 1.72.1 (m, 6H, N(CH2CH3)2), OH and NH), 2.52.85 (m, 3H, CH(CH3)2 and OCH2CH(OH)CH2), 3.20 (d, 2H, NCOCH2N) 3.40 (brs, H2O), 3.84.2 (m, 7H, OCH2CH(OH)CH2, CH2CH2N and OCH2CH2), 6.13 (dd, 1H, arom, J = 8.0 Hz, 2.8 Hz), 6.3 (d, 1H, arom, J = 2.8 Hz), 6.65 (d, 1H, arom, J = 8.5 Hz). 13C-NMR (CDCl3) 22.58, 29.66, 49.08, 50.04, 55.88, 64.19, 65.83, 67.71, 68.27, 70.98, 100.01, 102.08, 116.25, 135.96, 138.44, 153.63, 163.58. Anal. C20H33N3O4 (C, H). 1H-NMR (CDCl3, 200 MHz) 43: 0.851.27 (m, 15H, C(CH3)3 and N(CH2CH3)2), 1.652.06 (m, 6H, N(CH2CH3)2), OH and NH), 2.452.8 (m, 2H, OCH2CH(OH)CH2), 3.20 (brs, 2H, NCOCH2N) 3.40 (brs, H2O), 3.74.2 (m, 7H, OCH2CH(OH)CH2, CH2CH2N and OCH2CH2),), 6.13 (dd, 1H, arom, J = 8.0 Hz, 2.8 Hz), 6.3 (d, 1H, arom, J = 2.8 Hz), 6.65 (d, 1H, arom, J = 8.5 Hz). Anal. C21H35N3O4 (C, H).

915 5.1.11. 6- and 7-(3-Isopropylamino- and 3-tert-butylamino-2-hydroxypropoxy)-3,4-dihydro-2H-1,4-benzoxazines 4447 A mixture of 2427 (0.0031 mol), potassium hydroxide (0.0372 mol), water (2.5 mL) and methanol (6 mL) was stirred at 6570 C for 3 h. The methanol was removed in vacuo and the residue was extracted successively with ethyl acetate (50 mL) and water (20 mL). The aqueous layer was discarded, the organic phase was dried over anhydrous sodium sulfate and evaporated in vacuo to give a viscous dark-coloured oil. The amines were taken up from the crude product with warm cyclohexane and the light-coloured oils which remained after evaporation, were converted into fumarates or maleates. Compound 46 was converted into the oxalate from ethyl acetate. 1H-NMR (CDCl3, 200 MHz) 44: 1.05 (d, 6H, CH(CH3)2, J = 6.3 Hz), 2.34 (brs, 3H, OH and 2NH), 2.602.86 (m, 3H, CH(CH3)2 and OCH2CH(OH)CH2), 3.34 (t, 2H, CH2CH2N, J = 4.4 Hz), 3.833.98 (m, 3H, OCH2CH(OH)CH2), 4.21 (t, 2H, OCH2CH2, J = 4.3 Hz), 6.326.40 (m, 2H, arom), 6.51 (d, 1H, arom, J = 8.2 Hz). Anal. C18H26N2O7 (C, H). 45: 1.09 (s, 9H, C(CH3)3), 2.2 (brs, 3H, OH and 2NH), 2.602.85 (m, 2H, OCH2CH(OH)CH2), 3.35 (t, 2H, CH2CH2N, J = 4.4 Hz), 3.803.95 (m, 3H, OCH2CH(OH)CH2), 4.20 (t, 2H, OCH2CH2, J = 4.3 Hz), 6.306.40 (m, 2H, arom), 6.52 (d, 1H, arom, J = 8.2 Hz). Anal. C23H32N2O11 (C, H, N). 46: 1.12 (s, 6H, CH(CH3)2), 2.552.87 (m, 6H, OH, 2NH, CH(CH3)2 and OCH2CH(OH)CH2), 3.704.04 (m, 5H, CH2CH2N and OCH2CH(OH)CH2), 4.22 (t, 2H, OCH2CH2, J = 4.3 Hz), 6.65 (dd, 1H, arom, J = 8.4 Hz, 2.8 Hz), 6.80 (d, 1H, arom, J = 8.5 Hz). Anal. C18H26N2O11 (C, H). 47: 1.10 (s, 9H, C(CH3)3), 2.35 (brs, 3H, OH and 2NH), 2.572.85 (m, 2H, OCH2CH(OH)CH2), 3.38 (t, 2H, CH2CH2N, J = 4.3 Hz), 3.703.87 (m, 3H, OCH2CH(OH)CH2), 4.17 (t, 2H, OCH2CH2, J = 4.3 Hz), 6.156.23 (m, 2H, arom), 6.66 (dd, 1H, arom, J = 8.4 Hz, 2.8 Hz). Anal. C19H28N2O7 (C, H). 5.1.12. 4-Acetyl-6- and 7-[3-(3,4-dimethoxybenzyl) amino-2-hydroxypropoxy]-3,4-dihydro-2H-1,4-benzoxazines 48 and 49 A mixture of 14 or 15 (0.0025 mol) and 3,4dimethoxybenzylamine (0.025 mol) in isopropyl alcohol (30 mL) was reuxed for 3 h. The solvent was evaporated in vacuo and the residue was puried by crystallization from diethyl ether. 1H-NMR (CDCl3, 200 MHz) 48: 1.60 (brs, 2H, OH and NH), 2.25 (s, 3H, NCOCH3), 2.62.85 (m, 2H, OCH2CH(OH)CH2), 3.703.92 (m, 13H, CH2CH2N, CH2NHCH2CH(OH), 2CH3O and OCH2CH(OH)CH2), 4.204.30 (m, 2H, OCH2CH2), 6.406.50 (m, 2H, arom), 6.726.87 (m, 3H, arom), 7.75 (s, 1H, arom). Anal. C26H32N2O10 (C, H). 49: 2 (brs, 2H, OH and NH), 2.30 (s, 3H, NCOCH3), 2.672.92 (m, 2H, OCH2CH(OH)CH2), 3.703.95 (m, 13H, CH2CH2N, CH2NHCH2CH(OH), 2CH3O and OCH2CH(OH)CH2), 4.23 (t, 2H, OCH2CH2, J = 4.3 Hz), 6.606.88 (m, 5H, arom), 7.35 (s, 1H, arom). Anal. C26H32N2O10 (C, H). 5.2. Pharmacological methods 5.2.1. 1-Adrenoceptor binding assay Pellets containing 1 type adrenergic receptors were obtained from turkey erythrocyte membranes as described in the literature [34]. [3H]Dihydroalprenolol ([3H]DHA) (NEN), having a specic activity of 99.9 Ci/mmol and radiochemical purity > 98.5%, was used as ligand. 1-adrenergic receptor binding assays were determined as follows: 100 L of membrane ( 431 g/mL of protein diluted 1:8 v/v) were incubated for 15 min at 37 C with 100 L of 6 nM [3H]DHA and 100 L of various concentrations of the test compounds (dissolved in saline with DMSO 2.5%) and 12 mM Tris-HCI, pH = 7.5 (total vol. 1 mL). The incubations were stopped by adding 4 mL of cold buffer (12 mM Tris-HCI) followed by rapid ltration through glass bre lter disks (Whatman GF/B). The samples were subsequently washed 3 times with 4.5 mL of the same buffer and placed into scintillation vials, 10 mL of Filter-Count (Packard) liquid scintillation cocktail was then added to each vial and counting was carried out by scintillation spectrometer (Packard TRICARB 300C). Non-specic binding was dened as nondisplaceable binding in the presence of 100 L of 10 M propranolol. Blank experiments were carried out to determine the effect of the solvent (2.5%) on the binding. The concentrations of the test compounds that inhibited [3H]DHA binding by 50% (IC50) were determined by log-probit analysis with six concentrations of the displacers, each performed in triplicate. The IC50 values obtained were used to calculate apparent inhibition constants (Ki) by the method of Cheng and Prussoff [35], from the following equation: Ki = IC50/(1 + S/KD) where S represents the concentration of the ligand used and KD its receptor dissociation constant (KD values, obtained by Scatchard analysis [36], for [3H]DHA is 3.6 109 M). 5.2.2. 2-Adrenoceptor binding assay Preparation of lung homogenate: male SpragueDawley rats were sacriced by decapitation. The right lung was removed free of the major bronchi. Lungs were homogenized with a Brinkman Polytron (setting 5 for 15 s) in 50 volumes of buffer, 75 mM Tris-HCI (pH 7.65), 25 mM MgCl2 and then centrifuged at 30 000 g for

916 10 min twice. The resulting pellets were resuspended in 100 volumes of buffer, 75 mM Tris-HCI (pH 7.65), 25 mM MgCl2, then were frozen at 80 C before being assayed [37, 38]. [3H]Dihydroalprenolol was used as ligand. 300 L of lung membrane were incubated for 30 min at 37 C with 50 L of 6 nM [3H]DHA, 50 L of ketanserine 107 M 5HT antagonist, 50 L of practolol 106 M as 2 antagonist and 50 L of various concentrations of the test compounds (dissolved in saline with DMSO 5%, or H2O) and 75 mM Tris-HCI (pH 7.65), 25 mM MgCl2 (total volume 0.5 mL). The samples were subsequently washed with 4.5 mL of the same buffer and placed into scintillation vials. 10 mL of Filter-Count (Packard) liquid scintillation cocktail was then added to each vial and counting was carried out by scintillation spectrometer (Packard Tricarb 300C). Non-specic binding was dened as non-displaceable binding in the presence of 50 L of 10 M ICI 118551. The concentration of the test compounds that inhibited [3H]DHA binding by 50% (IC50) were determined by log-probit analysis with four concentrations of the displacers, each performed in triplicate. Non-specic binding was measured in the presence of 50 L of 10 M unlabelled ICI 118551, and specic binding as the difference between total and non-specic binding. Blank experiments were carried out to determine the effect of the solvent (5%) on the binding. 5.2.3. Inotropic, chronotropic and coronary vasodilating activity assays The inotropic and chronotropic effects of the test compounds were examined with the use of isolated blood-perfused dog heart preparations. The hearts were excised from mongrel dogs of either sex weighing 814 kg. The isolated blood-perfused papillary muscle and sinoatrial node preparations were prepared according to the methods of Endoh and Hashimoto [39] and Kubota and Hashimoto [40] respectively. The preparations were cross-circulated through the cannulated arteries with blood from a donor anaesthetized with sodium pentobarbital and receiving heparin. The perfusion pressure was kept constant at 100 mm Hg. The papillary muscle was stimulated at a frequency of 2 Hz and the tension developed by the muscle was measured with a force displacement transducer (Shinkoh, UL-20-240). Sinus rate was measured with the use of a cardiotachometer (Data Graph, T-149) triggered by the developed tension of the right atrium. Blood ow through the cannulated arteries was measured with an electromagnetic ow meter (Nihon Kohden, MF-27). Signals of these parameters were recorded on a thermal pen recorder (NECSanei, Recti-Horiz 8K). The compounds were injected intra-arterially with microsyringes.

Acknowledgements The authors thank Sig. Michele Guerrini for technical assistance. Financial support from MURST (60%) and LAVIFARM S.A. is gratefully acknowledged.

References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] Kajino M., Shibouta Y., Nishikawa K., Meguro K., Chem. Pharm. Bull. 39 (1991) 28962905. Matsumoto Y., Tsuzuki R., Matsuhisa A., Takayama K., Yoden T., Uchida W. et al., Chem. Pharm. Bull. 44 (1996) 103114. Kuroita T., Sakamori M., Kawakita T., Chem. Pharm. Bull. 44 (1996) 756764. Matsuoka H., Ohi N., Mihara M., Suzuki H., Miyamoto K., Maruyama N. et al., J. Med. Chem. 40 (1997) 105111. Yoshitomi Pharmaceutical Industries, Ltd., Japan Kokai Tokyo 81 51, 465; Chem. Abstr. 95 (1981) 115577. Cecchetti V., Fravolini A., Fringuelli R., Schiaffella F., Mascellani G., Pagella P., Rugarli P., Il Farmaco Ed. Sci 42 (1987) 6175. Schromm K., Mentrup A., Renth E., Fuegner A., U.S. Patent, US 4, 460, 581; Chem. Abstr. 101 (1984) 191939. Jaeggi K., Ostermayer F., Schroeter H., Ger. Offen. 2, 700, 193; Chem. Abstr. 87 (1977) 152206. Erez M., Shtacher G., Weinstock M., J. Med. Chem. 21 (1978) 982984. Angelopoulos C., (1997) Ph. D., University of Athens, Department of Pharmacy. Nakagawa K., Nishi T., Oshiro Y., Japan Kokai 76 68, 575; Chem. Abstr. 86 (1977) 89632. Nakagawa K., Murakami N., Yoshizaki S., Tominaga M., Mori H., Yabuuchi Y., Shihtani S., J. Med. Chem. 17 (1974) 529533. Winkler W., Arzn. Forsch. 33 (1983) 279280. Yabuuchi Y., Kinoshita D., Jpn. J. Pharmacol. 24 (1974) 853886. Khouili M., Guillaumet G., Coudert G., Pujol M.D., Il Farmaco. 51 (1996) 175184. Khouili M., Guillaumet G., Coudert G., Pujol M.D., Il Farmaco. 51 (1996) 185188. Crowther A.F., Howe R., Smith L.H., J. Med. Chem. 14 (1971) 511513. Hoee M.L., Hastings S.G., Meyer R.F., Corey R.M., Holmes A., Stratton C.D., J. Med. Chem. 18 (1975) 148152. Keh G.S., Raper C., Dowd H., Clin. Exp. Pharmacol. Physiol 5 (1978) 393398. Tominaga M., Yo E., Ogawa H., Yamashita S., Yabuuchi Y., Nakagawa K., Chem. Pharm. Bull. 32 (1984) 21002110. Tominaga M., Ogawa H., Yo E., Yamashita S., Yabuuchi Y., Nakagawa K., Chem. Pharm. Bull. 35 (1988) 36993704. Fujioka T., Teramoto S., Mori T., Hosokawa T., Sumida T., Tominaga M., Yabuuchi Y., J. Med. Chem. 35 (1992) 36073612. Combs D.W., Rampulla M.S., Bell S.C., Klaubert D.H., Tobia A.J., Falotico R., Haertlein B., Lakas-Weiss C., Moore J.B., J. Med. Chem. 33 (1990) 380386.

917
[24] Furniss B.S., Hannaford A.J., Rogers V., Smith P.W.G., Tatchell A.R. (Eds.), Vogels textbook of practical Organic Chemistry fourth edition, Longman, Harlow, 1981, p. 753. Liu L.L., Chen H.C., Cao S.L., Li R.T., Synth. Commun. 24 (1994) 833838. Shtacher G., Erez M., Cohen S., J. Med. Chem. 16 (1973) 516519. Loudon J.D., Ogg J., J. Chem. Soc. (1955) 739744. Booher R.N., Kornfeld E.C., Smalstig E.B., Clemens J.A., J. Med. Chem. 30 (1987) 580583. Ramage G.R., Hill J., British Patent, 1, 024, 626 (1966); Chem. Abstr. 64 (1966) 17614. Hill J., Ramage G.R., J. Chem. Soc. (1964) 37093713. Grollier J.F., Ger. Offen. DE 3, 545, 371 (1986); Chem. Abstr, 105 (1986) 158608a. Bugant A., Estradier F., Ger. offen 1, 940, 085 (1970); Chem. Abstr., 73 (1970) 36576p. [33] [34] [35] [36] [37] [38] [39] [40] Coudert G., Guillaumet G., Loubinoux B., Synthesis (1979) 543. Minneman K.P., Weilland G.A., Molinoff P.B., Mol. Pharmacol. 16 (1979) 3446. Cheng Y.C., Prusoff W.H., Biochem. Pharmacol. 22 (1973) 30993108. Scatchard G., Ann. NY Acad. Sci. 51 (1949) 660672. Minneman K.P., Hegstrand L.R., Molinoff P.B., Mol. Pharmacol. 16 (1979) 2133. Aarons R.D., Molinoff P.B., J. Pharmacol. Exp. Ther. 221 (1982) 439443. Endoh M., Hashimoto K., Am. J. Physiol. 218 (1970) 14591463. Kubota K., Hashimoto K., Naunyn-Schmiedergs Arch. Pharmacol. 278 (1973) 135150.

[25] [26] [27] [28] [29] [30] [31] [32]

Eur. J. Med. Chem. 34 (1999) 919937 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

919

Original article

1- and 2-Adrenoceptor antagonist activity of a series of para-substituted N-isopropylphenoxypropanolamines

Simon N. Louis*, Tracy L. Nero, Dimitri Iakovidis, Felicia M. Colagrande, Graham P. Jackman, William J. Louis
Clinical Pharmacology and Therapeutics Unit, The University of Melbourne, Department of Medicine, Austin and Repatriation Medical Centre Heidelberg 3084, Victoria, Australia (Received 8 July 1998; revised 14 April 1999; accepted 20 May 1999)

Abstract To further explore the structure-activity relationships of -adrenoceptor (-AR) antagonists, a series of 25 para-substituted N-isopropylphenoxy-propanolamines were synthesised, nine of which are new compounds. All have been examined for their ability to antagonise 1-ARs in rat atria and 2-ARs in rat trachea. Substitution in the para-position of the phenyl ring is thought to confer 3-specicity and the selectivity of these compounds for the 1-AR ranges from 1.5234. None of the compounds tested were selective for the 2-AR. Of the 25 compounds studied, 22 had reasonable (pA2 > 7) potencies for the rat 1-AR. Only compound 1 displayed reasonable (pA2 > 7) potency for the rat 2-AR. Twenty two compounds were used as the training set for comparative molecular eld analysis (CoMFA) of antagonist potency (pA2) at the rat 1- and 2-ARs. The inclusion of a number of additional physical characteristics improved the QSAR analysis over models derived solely using the CoMFA electrostatic and steric elds. The nal models predicted the 1- and 2-AR potency of the compounds in the training set with high accuracy (r2 = 0.93 and 0.86 respectively). The nal 1-AR model predicted the 1-potencies of two out of the three test compounds, not included in the training set, with residual pA2 values < 0.14, whereas the test compounds were not as well predicted by our nal 2-AR model (residual pA2 values < 0.38). 1999 ditions scientiques et mdicales Elsevier SAS comparative molecular eld analysis / rat -adrenoceptors / -adrenoceptor antagonists / N-isopropylphenoxypropanolamines / quantitative structure activity relationships

1. Introduction -Adrenoceptors (-ARs) are members of the large family of G-protein coupled receptors [14]. It has been established so far that there are at least three -AR subtypes, designated the 1-, 2- and 3-ARs [57]. N-Isopropylphenoxypropanolamine (gure 1, R = H) is considered to be a non-selective -AR antagonist. Many structure-activity studies in the past have focused upon the N-isopropylphenoxypropanolamine core structure with substitutions in the ortho and/or meta positions [8, 9]. However, the effects of para-substituents upon the -blocking activity of this core structure have received
*Correspondence and reprints
Abbreviations: Aryloxypropanolamine, AOPA; -adrenoceptors, -ARs; comparative molecular eld analysis, CoMFA; ESP phenyl ring charge, ESP; exibility of the para-substituent, exibility; length of the para-substituent, length; optimal number of components, ONC; partial least squares, PLS; standard error, SE.

Figure 1. General structure of para-substituted N-isopropylphenoxypropanolamines. The torsion angles 1 and 2 dene the conformation of the para-substituent and the oxypropanolamine side-chains respectively in relation to the phenyl ring, for example when 1 = 0 the para-substituent is planar to the phenyl ring.

less attention [1012]. Previous work by us [10, 13] and others [11, 12, 14, 15] has established that high 1-AR selectivity and potency can be achieved by parasubstitution of the phenyl ring and/or appropriate substitution of the phenoxypropanolamine amino group (e.g. -ethyl-3-(4-hydroxyphenyl)urea [13]).

920

Figure 2. General procedure for preparing N-isopropylphenoxypropanolamines. Reagents: (i) aqueous KOH, ETOH; (ii) isopropylamine, ETOH.

To further examine the effect of para-substitution upon both 1- and 2-AR selectivity and potency, we have synthesised a series of para-substituted N-isopropylphenoxypropanolamines. Nine of these compounds are new (6, 7, 9, 10, 15, 16, 18, 20 and 22), that is to say that a chemical abstracts structure database search indicated that these compounds have not been previously reported. Eleven previously reported compounds (1, 2, 3, 4, 5, 8, 12, 14, 19, 23 and 25) and ve commercially available compounds (11 (metoprolol), 13 (H 87/07), 17 (betaxolol), 21 (RO 31-1118) and 24 (cicloprolol)) have also been synthesised and included in the study in order to obtain comparative pharmacological data in the rat. All of these compounds have been examined in our laboratory for their ability to antagonise 1-ARs in rat atria and 2-ARs in rat trachea, which are well established as sources of the 1- and 2-AR subtypes respectively [16]. The synthesis and structure-activity relationships, using comparative molecular eld analysis (CoMFA), of these para-substituted N-isopropylphenoxypropanolamines are presented here. 2. Chemistry All compounds were prepared as their racemic mixtures by the general procedure shown in gure 2. It has

been established for simple phenoxypropanolamines that the S-isomer is the active isomer, with little -AR activity residing with the R-isomer [1719]. Resolution of the racemate into the individual isomers or stereospecic synthesis was therefore not carried out. The corresponding phenols A reacted under aqueous alkaline conditions with epichlorohydrin to produce the epoxides B. After isolation, the crude epoxides B were allowed to react with isopropylamine overnight to furnish the desired compounds C. Compounds 1, 2, 3, 4, 5, 8, 11 (metoprolol), 12, 13 (H 87/07), 14, 17 (betaxolol), 22, 21 (RO 31-1118), 23, 24 (cicloprolol) and 25, have been previously reported and were prepared following literature procedures [2033]. All intermediates and nal compounds had virtually identical physical and chemical data with those reported. The precursor phenols of compounds 6, 10, 15 and 16 were prepared by reacting bromoethane, n-bromopropane, 4-cyclohexylethyl bromide and 4-uorophenylethyl bromide with 4-benzyloxyphenol, thus producing the benzyl ethers D which were subsequently hydrogenated giving the desired phenols E (4-ethoxyphenol), F (4-n-propoxyphenol), G (4-cyclohexylethoxyphenol) and H (4-uorophenethoxyphenol) as shown in gure 3. The precursor phenols of compounds 18, 20 and 22 were prepared by the sequence of reactions shown in

Figure 3. Reagents: (i) K2CO3/NaI, anhydrous acetone; (ii) H2, 10% Pd/C.

921

Figure 4. Reagents: (i) ClCH2COOH, NaH, DMF; (ii) LiAlH4, THF; (iii) pTsCl, pyridine; (iv) NaH, DMF; (v) H2, 10% Pd/C.

gure 4. The corresponding alcohols 2-methylpropane-1ol, 2-hydroxymethyltetrahydrofuran and n-pentane-1-ol reacted with chloroacetic acid to furnish the corresponding alkoxy acetic acids I. These were reduced to their corresponding alkoxy-ethanols J whose tosylates K reacted with 4-benzyloxyphenol, to produce the benzyl ethers L which were subsequently hydrogenated giving the desired phenols M, N and O. The precursor phenol of compound 7 was prepared by the reactions shown in gure 5 [34, 35]. Phenyl-nbutanoate reacted with aluminium chloride to furnish a mixture of regioisomers P (2- and 4-hydroxypropiophenone), of which 4-hydroxypropiophenone, after iso-

lation, was reduced with sodium amalgam to produce the desired phenol Q. The precursor phenol of compound 9 was prepared according to the reactions shown in gure 6. Ethyl-4hydroxyphenyl propionate was benzylated to produce the ether R, which was subsequently reduced to give the protected alcohol S. Hydrogenation of S produced the desired phenol T. The above mentioned phenols were then used to produce compounds 6, 7, 9, 10, 15, 16, 18, 20 and 22 following the general procedure shown in gure 2. The general physical data for these compounds are given in table I.

Figure 5. Reagents: (i) anhydrous AlCl3; (ii) Zn-Hg, MeOH, HCl.

922

Figure 6. Reagents: (i) BzBr, K2CO3, NaI, anhydrous acetone; (ii) LiALH4, THF; (iii) H2, 10% Pd/C.

3. Pharmacology All compounds were evaluated for their ability to antagonise 1- and 2-ARs in rat atria and tracheal rings respectively. Cumulative concentration-response curves were obtained in each preparation as described by Van Rossum [36] and curves were tted by computer analysis according to the method of Zabrowsky et al. [37] using the sigmoidal t function of the Origin graphics package [38]. The antagonist potencies, or pA2 values, were calculated using equation 1 [39] and represent the mean

SE from 49 individual experiments. The -AR antagonist potencies and the subtype selectivity are given in table II. pA2 = log where the dose ratio =
(-)-isoprenaline at EC50 in the presence of antagonist (-)-isoprenaline at EC50 in the absence of antagonist

antagonist dose ratio 1

(1)

Table I. Physical data for compounds 6, 7, 9, 10, 15, 16, 18, 20 and 22. Compound 6 7 9 10 15 16 18 20 22 Yield (from phenol) 3.66 g (63.4 %) 4.35 g (72.1 %) 2.41 g (39.7 %) 3.56 g (58.6 %) 4.91 g (66.3 %) 11.71 g (61.4 %) 4.00 g (55.3 %) 3.59 g (48.4 %) 3.90g (52.1 %) M.p. (C) HCl salta 87.588.5 70.573.0 68.571.5 116.5119.0 131.0133.0 136.0138.0 89.591.3 165.5167.0 77.079.0 Formulab C14H23NO3.HCl C16H27NO2.HCl C15H25NO3.HCl C15H25NO3.HCl C20H33NO3.HCl C20H26FNO3.HCl C18H31NO4.HCl C11H31NO5.HCl C19H33NO4.HCl

Ethanol-ether was the solvent system used for recrystallisation. bEmpirical formula; elemental analyses were carried out for the nine compounds shown and were within 0.4% per element.

923
Table II. Antagonist activity, subtype selectivity, exibility of para-substituent, , ESP phenyl ring charge at 90 and 0 and length of the para-substituent for the compounds examined in this study.

Compound R Group

1-ARa pA2

2-ARb pA2

1-AR selectivityc

Flexibilityd

ESP phenyl ring charge at 90o (e) 0.83 0.67 0.71 0.68 0.76 0.71 0.72 0.75 0.72 0.68 0.72 0.74 0.68 0.68

ESP phenyl ring charge at 0o (e) 0.83 0.67 0.71 0.72 0.72 0.74 0.68 0.69 0.69 0.72 0.69 0.70 0.72 0.69

Lengthf ( )

Training Compounds 1 2 3 4 5 6 7 8 9 10 11 12 13 14

8.09 0.14 6.60 0.18 7.13 0.04 6.58 0.19 7.41 0.13 7.08 0.27 7.49 0.26 7.38 0.19 7.13 0.10 7.16 0.14 7.60 0.17 7.98 0.21 7.37 0.09 6.67 0.09

7.47 0.14 6.43 0.17 6.47 0.11 6.34 0.09 6.56 0.18 5.94 0.16 6.23 0.23 5.39 0.16 6.58 0.34 5.72 0.38 6.43 0.13 6.77 0.28 5.97 0.19 5.36 0.18

4.2 1.5 4.6 1.7 7.1 14 18 98 3.5 28 15 16 25 20

0 0 0 1 2 2 3 3 3 3 3 4 4 2

0.00 0.86 0.47 0.04 1.37 0.36 1.82 0.09 0.06 0.84 0.11 0.57 0.13 2.06

0.00 1.41 1.53 2.30 3.84 3.61 5.00 5.06 4.86 4.79 4.86 6.14 5.91 5.08

15

7.65 0.12

5.93 0.23

52

2.57

0.66

0.70

6.45

16

8.01 0.12

5.64 0.12

234

2.13

0.70

0.73

8.72

17

8.06 0.15

6.38 0.13

48

0.96

0.73

0.70

7.39

924
Table II. Continued
Compound R Group 1-ARa pA2 2-ARb pA2 1-AR selectivityc Flexibilityd e ESP phenyl ring charge at 90o (e) 0.73 ESP phenyl ring charge at 0o (e) 0.70 Lengthf ( )

Training Compounds 18

7.75 0.22

5.69 0.20

115

1.13

8.54

19

7.61 0.15

5.40 0.18

162

1.76

0.72

0.69

9.51

20

7.74 0.09

5.49 0.27

178

0.32

0.69

0.73

9.36

21

7.98 0.21

5.68 0.10

200

2.05

0.72

0.73

12.4

22 Test Compounds 23 24

8.04 0.23

5.80 0.24

174

1.66

0.72

0.72

10.8

7.22 0.08 7.73 0.07

6.47 0.17 5.44 0.10

5.6 195

1 6

0.92 0.72

0.77 0.73

0.75 0.70

2.50 8.59

25

8.04 0.30

5.81 0.25

170

1.31

0.74

0.70

8.36

1-AR antagonist pA2 value SEM determined in isolated spontaneously beating rat atria. b2-AR antagonist pA2 value SEM determined in rat tracheal chain preparation. cSelectivity = antilog (pA2 1-AR pA2 2-AR). dFlexibility was dened as the number of fully rotatable non-H bonds in the para-substituent. e was logP of the compound in question minus logP of compound 1, logP was calculated using PrologP. f Length was dened as the distance between the para-carbon atom of the phenyl ring and the most distant non-H atom.
a

4. Results and discussion 4.1. Isolated tissue preparations Functional potencies of compounds 125 for inhibiting (-)-isoprenaline-induced: (i) 1-AR mediated chronotropic effects in spontaneously beating rat atria; and (ii) 2-AR mediated relaxation of rat tracheal chain previously contracted with 1 M carbachol are listed in table II. The unsubstituted reference compound 1 had a high potency at both the 1-AR (pA2 = 8.09) and 2-AR (pA2 = 7.47). Para-substitution reduced the potency of the compounds for both 1- and 2-ARs (compounds 211, 1315, 18, 19, 20, 23 and 24; table II), except for compounds 12, 16, 17, 21, 22 and 25 (1-AR pA2s = 7.988.06) which had similar 1-AR potencies to the unsubstituted compound 1. Compound 12 (2-AR pA2 =

6.77) was the most potent of the para-substituted compounds at inhibiting rat 2-ARs (c.f. reference compound 1, pA2 = 7.47). Overall, the compounds had higher potencies for the 1-AR than the 2-AR, with the 1/2 selectivities ranging from 1.5234. Compounds 1, 2, 3, 4, 9 and 23, however, were at the lower end of the selectivity scale and are considered to be non-selective in the rat. The animal species used in this study to determine both the 1- and 2-AR functional potency of the compounds was the rat. Previously published 1-AR antagonist functional potency, or activity data has been obtained from both the rat and guinea-pig [30, 40], whereas the guinea-pig is the commonly used animal species for published 2-AR antagonist activity [30, 40]. By using the rat as the source for both 1- and 2-ARs, any

925

Figure 7. (a) Structural alignment of the training set of compounds (compounds 122) when 1 = 90. (b) Structural alignment of the training set of compounds when 1 = 0.

interspecies differences between the receptor subtypes was removed. 4.2. Alignment of molecules The alignment of the compounds is the most important feature of CoMFA analysis [41]. It has been proposed that AOPA type compounds interact with -ARs via a three point pharmacophore consisting of the -hydroxyl group, the amino group and an electron rich moiety which is usually a phenyl ring [42, 43] and more recent sitedirected mutagenesis studies support this hypothesis [4447]. If we assume the compounds all act via the same mechanisms, then these common points of interaction must superimpose and hence we have assumed that the AOPA core structure remains xed for all compounds in our conformational analysis. Previous studies within our laboratory [10] examined two conformations of each molecule ie. 1 = 90 or 0. In the present study we have examined the effect of these values of 1 on our CoMFA analysis. Figure 7a displays

the superimposition of the compounds when 1 = 90 and gure 7b when 1 = 0. 4.3. CoMFA analysis for antagonist potency at the rat 1-AR The CoMFA (SYBYL version 6.40) results for antagonist potency at rat 1-ARs for the twenty-two compounds in the training set are given in table III. A range of column ltering values were used (1.016.0 kcal/mol). Column ltering omits from the analysis columns, lattice points whose variance is less than the specied value (SYBYL). Partial least squares analysis (PLS) of the data identied that when s1 was set to 90 (gure 1) and the default CoMFA settings were used, which included both steric and electrostatic elds, the highest cross validated r2 (q2) value was obtained when column ltering was set to 4.0 kcal/mol (q2 = 0.24; optimum number of components (ONC) = 4; table III). When s1 = 0, q2 was maximised when column ltering was set at 5.0 kcal/mol (q2 = 0.38 and ONC = 2; table III).

926
Table III. CoMFA results for rat 1-ARs (n = 22). Analysis variables CoMFA CoMFA, exibilitya CoMFA, ESPb,c CoMFA, lengthd CoMFA, CoMFA, exibility, ESP CoMFA, exibility, length CoMFA, exibility, CoMFA, ESP, length CoMFA, ESP, CoMFA, length, CoMFA, exibility, ESP, length CoMFA, exibility, ESP, CoMFA, exibility, length, CoMFA, ESP, length, CoMFA, exibility, ESP, length, CoMFA CoMFA, exibilitya CoMFA, ESPb CoMFA, lengthc CoMFA, CoMFA, exibility, ESP CoMFA, exibility, length CoMFA, exibility, CoMFA, ESP, length CoMFA, ESP, CoMFA, length, CoMFA, exibility, ESP, length CoMFA, exibility, ESP, CoMFA, exibility, length, CoMFA, ESP, length, CoMFA, exibility, ESP, length,
a

1 90 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o
o

Column ltering 4 4 13 4 4 9 4 4 13 13 4 9 10 4 5 9 5 5 5 5 5 5 5 5 5 5 5 5 2 5 5 5

q2 0.244 0.527 0.747 0.322 0.186 0.686 0.375 0.398 0.672 0.638 0.247 0.655 0.632 0.259 0.636 0.610 0.384 0.592 0.604 0.524 0.553 0.657 0.571 0.515 0.618 0.601 0.407 0.635 0.599 0.482 0.563 0.564

ONC 4 3 10 6 5 4 4 4 10 10 5 4 5 4 7 5 2 2 4 2 2 2 2 3 2 4 3 2 6 3 4 6

SE 0.461 0.354 0.332 0.465 0.493 0.331 0.419 0.411 0.358 0.388 0.474 0.311 0.332 0.457 0.352 0.341 0.394 0.320 0.334 0.346 0.335 0.294 0.328 0.359 0.310 0.335 0.397 0.303 0.357 0.371 0.351 0.373

Flexibility of the para-substituent. bBold indicates the options used for the nal model. cESP phenyl ring charge. dLength of the para-substituent.

In addition to the PLS analysis of the CoMFA data, several other physical parameters of the compounds were determined and included in further CoMFA analysis. These parameters included the exibility, length and of the para-substituent and the ESP phenyl ring charge (table II). The exibility of the para-substituent (exibility) was determined by assigning an integer value for the number of torsion angles which affected the conformation of the substituent. The term ESP phenyl ring charge (ESP) was dened in this study as the sum of the six phenyl ring carbon atom esp charges. The phenyl ring charges of all the compounds studied (table II) were calculated using the AM1 Hamiltonian within MOPAC (version 6.0), specifying the key words ESP, PRECISE and NOMM. The length of the para-substituent (length) was dened as the distance between the para-

carbon of the phenyl ring and the most distant non-H atom, hence the para-substituent length of compound 1 is zero (table II). The value of the para-substituent was dened as the logP of the compound minus the logP of compound 1, logP having been calculated using Pallas PrologP (version 1.1). When these additional parameters were included with CoMFA they generally improved q2. For instance, when 1 = 90 a combination of the CoMFA elds and ESP gave the highest q2 (q2 = 0.769; ONC = 11; SE = 0.332; table III), and when 1 = 0 a combination of the CoMFA elds, ESP and exibility maximised q2 (q2 = 0.657; ONC = 2; SE = 0.29; table III). The former model was selected for further analysis as it possessed the highest q2 of any of the 1-AR models examined (table III).

927

Figure 8. (a) Plot of the actual 1-AR pA2 verses the predicted 1-AR pA2 compounds in table II (n = 22) using the nal model, non cross validated, and an optimum number of components of 4. (b) Plot of the actual 2-AR pA2 verses the predicted 2-AR pA2 compounds in table II (n = 22) using the nal model, non cross validated, and an optimum number of components of 4.

The nal model, without cross validation, was obtained using the default CoMFA options, 10 ONC and 13 kcal/ mol column ltering, 1 = 90o, including ESP. This model had an r2 value of 0.95, a standard error (SE) of 0.15 and an F (11, 10) of 21.47. The regression equation for ESP, taken from the PLS output le, was: 1-pA2 = 0.9958.49 ESP The relative contributions of the components were: steric, 0.88; electrostatic, 0.00; and ESP, 0.12. Figure 8a displays the relationship between calculated and mea-

sured pA2 values for the non-cross validated analysis and residual values are given in table IV. 4.4. CoMFA analysis for antagonist potency at the rat 2-AR Similarly, the default CoMFA settings and a range of column ltering values (1.016.0 kcal/mol) were used when analysing the 2-AR data. When 1 = 90o the highest q2 value was obtained when column ltering was

928
Table IV. CoMFA Experimental, calculated and residual activities of the training set of compounds for 1-ARs. Compound 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Experimental 8.09 6.60 7.13 6.58 7.41 7.08 7.49 7.38 7.13 7.16 7.60 7.98 7.37 6.67 7.65 8.01 8.06 7.75 7.61 7.74 7.98 8.04 Calculated 8.04 6.68 7.02 6.73 7.48 6.96 7.31 7.58 7.12 7.19 7.56 7.95 7.43 6.68 7.61 8.03 8.06 7.75 7.80 7.52 8.03 8.01 Residual 0.05 0.08 0.11 0.15 0.07 0.12 0.18 0.20 0.01 0.03 0.04 0.03 0.06 0.01 0.04 0.02 0.00 0.00 0.19 0.22 0.05 0.03

column ltering values of 5 kcal/mol when 1 = 90 and 6 kcal/mol when 1 = 0. 4.5. Comparison of contour maps Figures 9A and 9B show compound 16 binding within the steric contour maps for the nal 1- and 2AR models. The experimental data suggests that parasubstitution decreases antagonist potency at both 1- and 2-ARs. However, the contour maps, of the 1- and 2-AR pharmacophores display important differences, the most striking of which appears to be a bulk preferring pocket in the 1-AR which can be accessed by long exible para-substituents. Compounds that access this pocket have their 1-AR potency restored and the longer the substituent the greater the restoration of their activity, compare the activity of compounds 4, 13 and 22. Compounds with shorter bulky substituents, like compound 14 cannot access this pocket and have low 1-AR potencies. The 2-AR, on the other hand, seems far more sterically restricted than the 1-AR in the para-position. For instance, the addition of the short bulky substituent of compound 14 reduces the 1-AR potency by 1.42 log units when compared to the base compound, whereas this substituent reduces 2-AR potency by 2.11 log units. Using compound 16 as an example the differences between the two steric maps can be clearly illustrated. In the 1-AR model (gure 9A), the para-substituent of compound 16 does not impinge on the sterically restricted region (ie. yellow region), however, the terminal uorophenyl ring is embedded in the bulk preferring pocket (i.e. green region) and hence the 1-potency is almost restored to that of compound 1 (table II). By contrast, the para-substituent of compound 16 is clearly embedded in the sterically restricted region of the 2-AR model (gure 9B) and hence this compound has a much lower 2-pA2 than compound 1 (table II). 4.6. Predictive abilities of the CoMFA models The potencies of three test compounds (table VII), not in the training set, were predicted at rat 1- and 2-ARs using the nal models and the predict property function in the QSAR package of SYBYL (version 6.40). For the 1-AR the predicted potencies for two of the compounds were in close agreement with those obtained experimentally (residuals < 0.15). For the 2-AR, however, all three compounds were relatively poorly predicted (residuals 0.25 to 0.38) suggesting that our CoMFA model of the 2-AR is not sufficiently resolved.

set to 4 kcal/mol (q2 = 0.50; ONC = 5; table V). When 1 = 0, q2 was maximised when column ltering was set to 5 kcal/mol (q2 = 0.45; ONC = 2; table V). When the additional parameters were included in the analysis the highest q2 occurred when 1 = 90o, column ltering was set at 6 kcal/mol and both exibility and ESP were considered in addition to the steric and electrostatic elds generated within CoMFA (q2 = 0.60; ONC = 4; SE = 0.38). The nal model, without cross-validation, for the 2-AR was obtained using 4 ONC, column ltering set at 6 kcal/mol and included exibility and ESP. This model had an r2 value of 0.83, SE of 0.25 and an F (4, 17) of 20.71. The regression equation for exibility and ESP was: 2-pA2 = 2.05 + 0.07 exibility6.46 ESP. The relative contributions of the components were: steric, 0.73; electrostatic, 0.00; exibility, 0.11; and ESP, 0.161. Figure 8b displays the relationship between the calculated and measured pA2 values for the cross validated analysis and residual values are given in table VI. It is important to note that by using the column ltering values to maximise the q2 for the 1- and 2-AR models (column ltering = 13 and 6 kcal/mol, respectively) it effectively removes the electrostatic eld parameter from the CoMFA analysis. For both models this occurs at

929
Table V. CoMFA results for rat 2-ARs (n = 22). Analysis variables CoMFA CoMFA, exibilitya CoMFA, ESPb CoMFA, lengthc CoMFA, CoMFA, exibility, ESPd CoMFA, exibility, length CoMFA, exibility, CoMFA, ESP, length CoMFA, ESP, CoMFA, length, CoMFA, exibility, ESP, length CoMFA, exibility, ESP, CoMFA, exibility, length, CoMFA, ESP, length, CoMFA, exibility, ESP, length, CoMFA CoMFA, exibilitya CoMFA, ESPb CoMFA, lengthc CoMFA, CoMFA, exibility, ESP CoMFA, exibility, length CoMFA, exibility, CoMFA, ESP, length CoMFA, ESP, CoMFA, length, CoMFA, exibility, ESP, length CoMFA, exibility, ESP, CoMFA, exibility, length, CoMFA, ESP, length, CoMFA, exibility, ESP, length,
a

1 90 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 90o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o 0o
o

Column ltering 4 4 6 4 4 6 4 4 4 5 4 5 7 4 12 12 5 5 2 5 5 2 5 7 2 1 7 2 1 1 1 1

q2 0.501 0.574 0.586 0.499 0.364 0.602 0.498 0.387 0.575 0.454 0.284 0.587 0.430 0.372 0.473 0.455 0.445 0.462 0.453 0.463 0.351 0.454 0.479 0.538 0.459 0.386 0.494 0.491 0.387 0.440 0.451 0.475

ONC 5 4 4 6 4 4 4 5 5 3 3 4 5 3 2 2 2 4 5 4 5 5 4 9 5 6 9 5 6 10 6 6

SE 0.434 0.389 0.383 0.449 0.475 0.376 0.422 0.481 0.400 0.428 0.490 0.383 0.464 0.459 0.409 0.416 0.420 0.437 0.454 0.437 0.495 0.454 0.430 0.482 0.452 0.497 0.505 0.438 0.497 0.555 0.470 0.460

Flexibility of the para-substituent. bESP phenyl ring charge. cLength of the para-substituent. dBold indicates the options used for the nal model.

5. Discussion The aim of this study was to determine the optimal structural requirements of para-substituted N-isopropylphenoxypropanolamines to maximise antagonist potency and selectivity for 1-ARs. The study also provides information on the extent to which the CoMFA derived models of the rat 1- and 2-ARs differed for a series of para-substituted -AR antagonists. Although parasubstitution generally reduced antagonist potency at both 1- and 2-ARs, differences did exist between the two models. For both models, steric factors were the most important (0.88 for 1- and 0.73 for 2-ARs). The 1-AR also possesses a pocket which prefers bulky substituents. Long exible compounds that accessed this region had their 1-AR potency restored which is consistent with our previous work [10] that suggested a hydrophobic binding

site existed that was accessible to long para-substituents containing a ring system. Our data, however, suggests that although the pocket accommodated large, hydrophobic substituents, eg. compounds 15, 16 and 21, it will also accept large, less hydrophobic substituents such as compounds 17, 24 and 25 and long exible carbon chains like compound 22. The important characteristic is that the substituent is able to access this pocket, for example, shorter hydrophobic compounds that cannot reach the pocket, like compound 14, encounter steric hindrance and have low 1-potency. In addition, a number of other physical characteristics of the compounds seemed to inuence their potency at 1- and 2-ARs. For both 1- and 2-ARs, ESP was negatively correlated with potency suggesting that the more negative the phenyl ring charge the higher the 1- or

930

Figure 9. (A) CoMFA contour map of the nal model of the 1-AR with compound 16 embedded. Green areas show favoured areas where increasing bulk is associated with higher potency, yellow areas where increasing bulk is associated with lower potency and blue areas where positive charge is associated with higher potency. (B) CoMFA contour map of the nal model of the 2-AR with compound 16 embedded.

2-AR potency which supports the hypothesis of previous workers that an electron rich moiety is an essential pharmacophoric element for binding to -ARs [42, 43]. In addition, exibility was positively correlated with 2-AR potency and this may be related to the ability of more exible para-substituents to avoid sterically restricted regions in the 2-AR. Unfortunately, our synthetic program was designed to examine 1-AR potency and hence only compound 1 had high potency (pA2 > 7.0) at the rat 2-AR and higher affinity compounds are required before we can condently predict the structure of the 2-AR pharmacophore. We can say, however, that the 2-AR appears to be very sterically restricted around the para-position of the phenyl ring of AOPA antagonists.

On the whole, the potency of the training set of compounds were predicted better by the 1-AR model (r2 = 0.95; residuals < 0.22) than the 2-AR model (r2 = 0.83; residuals < 0.80). In particular, the potency of compound 8 was over estimated for the 2-AR. This may have been due to the existence of a negative charge preferring region in the region of the positively charged amine in the para-substituent. This, however, was not identied by CoMFA perhaps due to the small number of compounds examined that possess this characteristic. It is of interest that the OH-group of compound 9, which has an electronegative nature, has signicantly higher 2-AR potency than compound 8 and slightly higher activity than the uncharged substituent of compound 7 (table II), therefore, this region requires further study.

931
Table VI. CoMFA Actual, calculated and residual activities of the training set of compounds for 2-ARs. Compound 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Actual 7.47 6.43 6.47 6.34 6.56 5.94 6.23 5.39 6.58 5.72 6.43 6.77 5.97 5.36 5.93 5.64 6.38 5.69 5.40 5.49 5.68 5.80 Calculated 7.41 6.34 6.53 6.27 6.50 6.17 6.03 6.19 6.32 5.76 6.21 6.49 6.11 5.39 6.07 5.49 6.47 5.69 5.33 5.40 5.67 5.84 Residual 0.06 0.09 0.06 0.07 0.06 0.23 0.20 0.80 0.26 0.04 0.22 0.28 0.14 0.03 0.14 0.15 0.09 0.00 0.07 0.09 0.01 0.04 Table VII. Comparison of predicted and experimentally determined potencies at rat 1- and 2-ARs for the test compounds. Compound 23 24 25 23 24 25 Receptor 111222Actual 7.22 7.73 8.04 6.47 5.44 5.81 Predicted 7.52 7.88 7.94 6.85 5.80 6.06 Residual 0.30 0.15 0.10 0.38 0.36 0.25

the symbols of the elements, results obtained were within 0.4% of the theoretical values. Descriptions of the synthetic procedures for preparing compounds 6, 7, 9, 10, 15, 16, 18, 20 and 22 are outlined below. No attempts were made to maximise yields. 6.1.2. Materials Phenol, 4-methylphenol, 4-ethylphenol, 4-methoxyphenol, 4-benzyloxyphenol, 4-uorophenol, p-cresol, 4-aminophenol, 4-uorophenethyl bromide, 2-cyclohexyl-1-bromoethane, 2-(4-benzyloxyphenyl)-1-ethanol, benzyl bromide, 4-toluene sulfonylchloride, phenyl-nbutanoate, ethylchloroformate, bromoethane, iodomethane, 2-hydroxymethyl tetrahydrofuran, 2-methyl-propane-1ol, n-pentanol, isobutanol, cyclopentanol, cyclohexanol, chloroacetic acid, ethyl-4-hydroxyphenyl propionate, lithium aluminium hydride and dimethyl formamide (DMF) were purchased from Aldrich Chemicals. Sodium hydride and isopropylamine were obtained from Fluka Chemicals. Anhydrous aluminium chloride was obtained from Merck. Inorganic reagents were supplied by Ajax Chemicals. Analytical grade solvents, including carbon disulphide, were purchased from Rhone Poulenc Australia. (-)-Isoprenaline and carbachol were obtained from Sigma. Silica plates (5 10 cm, Silica F254) were purchased from Merck. Silica for column chromatography (100 , 50 mm) was supplied by Amicon Inc., MA 01923, USA. 6.1.3. General procedure 1 for the synthesis of 1phenoxy-2,3-epoxypropanes (B) The phenol A (20 mmol), KOH (1.23 g, 22 mmol) and epichlorohydrin (5.55 g, 60 mmol) in EtOH (60 mL) were stirred overnight at room temperature. The reaction mixture was evaporated to dryness and the residue was partitioned between ether (or EtOAc) and water. The organic layer was washed with aqueous NaOH (5%), followed by a water wash and then dried over Na2SO4. After the solvent was evaporated, the residue was freed from excess epichlorohydrin under vacuum and used

The two CoMFA models expand the results of earlier studies [1015] that high 1-selectivity and potency can be achieved by para-substitution of the phenyl ring. For para-substituted N-isopropylphenoxy-propanolamines it appears at least two factors contribute to 1-potency. These are the ability of the para-substituent to access a bulk preferring pocket and the phenyl ring charge. The latter is negatively correlated with potency. 6. Experimental section 6.1. Chemistry 6.1.1. General Melting points were determined using a manual Gallenkamp electrothermal apparatus (range 0200 C) in glass capillary tubes and are uncorrected. IR spectra were recorded on a Perkin-Elmer FT IR 1600. NMR spectra were recorded on a Varian Associates EM 360 spectrometer and are expressed in using TMS (tetramethyl silane) as reference. Mass spectra were recorded on a Finnigan 4000 series GC/MS Mass spectrometer or a Thermo Instruments GCQ Mass spectrometer using methane gas as the ionising medium for CI (chemical ionisation spectra). All spectra were consistent with the assigned structures. Where analyses are indicated only by

932 without further purication. All epoxides were homogenous by TLC (silica, eluent dichloromethane). 6.1.4. General procedure 2 for preparing N-isopropylphenoxypropanolamines (C) The crude 1-phenoxy-2,3-epoxypropane B (20 mmol) and isopropylamine (100 mmol, 5 molar excess) in EtOH (60 mL) were stirred overnight at room temperature. The reaction mixture was evaporated to dryness and the residue was dissolved in ethanol and then treated with excess ethereal HCl. The crude precipitate was recrystallised from a suitable solvent or puried by column chromatography (silica, eluent: CH2Cl2/MeOH/NH4OH 28%, 90:9:1). 6.1.5. 4-Ethoxyphenol (E) Ethyl bromide (6.0 g, 55 mmol), 4-benzyloxyphenol (10.11 g, 50 mmol), anhydrous K2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirred and reuxed in anhydrous acetone (250 mL) for 48 h. The solvent was evaporated and the residue was partitioned between dichloromethane and water. The organic layer was washed with water and dried over anhydrous MgSO4. The solvent was evaporated and the residual 4-ethoxyphenylbenzylether D (R = H) was recrystallised from EtOH. Yield = 8.2 g, 81.5%; m.p. = 6466 C; MS m/e 229 (M + 1); 1H-NMR (CDCl3) 1.58 (3H, t), 4.15 (2H, q), 5.17 (2H, s), 7.06 (5H, s), 7.58 (4H, s). The ether D (R = H) (7.2 g, 36 mmol) in EtOH (200 mL) was hydrogenated over 10% Pd/C (0.1 g) at room temperature and atmospheric pressure. When hydrogen absorption ceased, the reaction mixture was ltered and evaporated to dryness. The product E was recrystallised from EtOH. Yield = 4.26 g, 87.0%; m.p. = 64.565.5 C; MS m/e 139 (M + 1); 1H-NMR (CDCl3) 1.61 (3H, t), 4.19 (2H, q), 5.63 (1H, s), 6.93 (4H, s). 6.1.6. 1-(4-Ethoxyphenoxy)-3-isopropylamino-2-propanol (6) Prepared from E according to the general procedures 1 and 2, chromatographed and isolated as the hydrochloride. Yield = 3.66 g, 63.4%; m.p. = 87.588.5 C; MS m/e 254 (M + 1); 1H-NMR (CD3OD) 1.33 (6H, d), 1.62 (3H, t), 2.163.23 (6H, bm), 4.23 (2H, q), 7.10 (4H, s). Anal. (C14H23NO3.HCl) C, H, Cl, N. 6.1.7. 4-n-Butylphenol (Q) Phenyl-n-butanoate (82.0 g, 500 mmol) was added into a stirred suspension of aluminium chloride (75 g) in carbon disulphide (80 mL) to furnish, after work-up, a mixture of 2- and 4-hydroxyphenylpropiophenones (butyrophenones) (64.5 g, 78.6%) which were separated by repeated fractional distillation. The desired 4-hydroxyphenyl propiophenone P was obtained in overall 33.4% yield (27.4 g) and had the following physical data: b.p. = 148150 C (11.5 mm Hg); MS m/e 165 (M + 1); 1 H-NMR (CD3OD) 1.14 (3H, t), 2.392.82 (2H, m), 3.623.98 (2H, m), 6.97 (5H, dd, the phenolic hydrogen is hiding under the aromatic protons). Reduction of P (27.0 g, 166 mmol) was achieved by its portion wise addition into freshly prepared sodium amalgam (68.0 g of powdered Zn and 5.03 g of HgCl2 in 650 mL of methanol and 320 mL of concentrated HCl), evaporation of methanol, isolation and chromatography on silica (eluent CH2Cl2:hexane, 1:1). Q was produced in 62.5% yield (15.56 g); MS m/e 151 (M + 1); 1H-NMR (CD3OD) 1.08 (3H, t), 1.252.05 (4H, m), 2.65 (2H, t), 3.87 (5H, m), 6.89 (5H, dd, the phenolic hydrogen is hiding under the aromatic protons). 6.1.8. 1-(4-n-Butylphenoxy)-3-isopropylamino-2-propanol (7) Prepared from Q according to the general procedures 1 and 2, chromatographed and isolated as the hydrochloride. Yield = 4.35 g, 72.1%; m.p. = 70.573.0 C; MS m/e 266 (M + 1); 1H-NMR (CD3OD) 1.13 (3H, t), 1.75 (6H, d), 2.512.93 (3H, m), 3.233.83 (5H, bm), 4.064.31 (2H, m), 4.635.12 (2H, m), 6.777.28 (4H, dd). Anal. (C16H27NO2.HCl) C, H, Cl, N. 6.1.9. 4-(n-3-Hydroxypropyl)phenol (T) Ethyl-4-hydroxyphenylpropionate (54.5 g, 250 mmol) was reuxed with benzyl bromide (44.5 g, 30.9 mL, 260 mmol), anhydrous potassium carbonate (38.7 g, 270 mmol) and sodium iodide (2.0 g) in dried acetone for 48 h. Filtration of the cooled mixture and evaporation of the solvent produced a solid residue which was taken up in ethyl acetate, washed with diluted sodium hydroxide solution and water, dried, ltered and evaporated. The product, ethyl-4-benzyloxyphenylpropionate, R, was puried by chromatography (silica, eluent CH2Cl2) and isolated as an oil. Yield = 68.6 g, 89.1%; MS m/e 309 (M + 1); 1H-NMR (CDCl3) 1.16 (3H, t), 2.76 (2H, t), 3.82 (2H, t), 4.64 (2H, q), 5.06 (2H, s), 6.88 (4H, s), 7.26 (5H, s). Ethyl-4-benzyloxyphenylpropionate (61.6 g, 200 mmol) in anhydrous THF (200 mL) was added dropwise into a suspension of lithium aluminium hydride (8.0 g, 210 mmol) in THF (500 mL). The resultant mixture was reuxed for 2 h, cooled, the excess reagent decomposed by the dropwise addition of water and ltered. The ltrate was evaporated to dryness and the residue was puried by chromatography (silica, eluent CH2Cl2). The isolated product, 1-(4-benzyloxyphenyl)-3-propanol, S, was recrystallised from ethanol-ether. Yield = 32.1 g, 66.2%;

933 m.p. = 59.561.5 C; MS m/e 243 (M + 1); 1H-NMR (CDCl3) 0.72 (1H, bs), 0.871.23 (2H, m), 2.81 (2H, t), 3.87 (2H, t), 5.13 (2H, s), 7.11 (4H, dd), 7.52 (5H, s). A solution of 1-(4-benzyloxyphenyl)-3-propanol (24.2 g, 100 mmol) in ethanol (600 mL) was hydrogenated over 10% Pd/C at ambient temperature and pressure. Filtration of the catalyst and evaporation of the solvent furnished the desired phenol T which was puried by chromatography (silica, eluent CH2Cl2) and isolated as an oil. Yield = 12.81 g, 84.3%; MS m/e 153 (M + 1); 1H-NMR (CDCl3/CD3OD 1:1) 1.772.23 (2H, m), 2.77 (2H, m), 3.81 (2H, m), 7.02 (4H, dd). 6.1.10. 1-(4-(n-3-Hydroxypropyl)-phenoxy)-3-isopropylamino-2-propanol (9) Prepared from T according to the general procedures 1 and 2, chromatographed and isolated as the hydrochloride. Yield = 2.41 g, 39.7%; m.p. = 68.571.5 C; MS m/e 268 (M + 1); 1H-NMR (CD3OD) 1.62 (6H, d), 1.832.29 (2H, m), 2.613.12 (2H, m), 3.153.94 (6H, m), 4.074.36 (2H, m), 6.907.48 (4H, dd). Anal. (C15H25NO3.HCl) C, H, Cl, N. 6.1.11. 4-n-Propoxyphenol (F) n-Propyl bromide (6.8 g, 55 mmol), 4-benzyloxyphenol (10.11 g, 50 mmol), anhydrous K2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirred and heated to reux in anhydrous acetone (250 mL) for 48 h. The solvent was evaporated and the residue was partitioned between dichloromethane and water. The organic layer was washed with water and dried over anhydrous MgSO4. The solvent was evaporated and the residual 4-n-propoxyphenylbenzylether D (R = Me) was recrystallised from EtOH. Yield = 11.07 g, 83.2%; m.p. = 6769 C; MS m/e 243 (M + 1); 1H-NMR (CDCl3) 1.22 (3H, t), 1.702.23 (2H, m), 4.03 (2H, t), 5.14 (2H, s), 7.03 (5H, s), 7.52 (4H, s). The ether D (R = Me) (10.1 g, 42 mmol) in EtOH (200 mL) was hydrogenated over 10% Pd/C (0.1 g) at room temperature and atmospheric pressure. When hydrogen absorption ceased, the reaction mixture was ltered and evaporated to dryness. The product F was recrystallised from EtOH. Yield = 5.12 g, 80.2%; m.p. = 6769 C; MS m/e 153 (M + 1); 1H-NMR (CDCl3) 1.19 (3H, t), 1.772.26 (2H, m), 4.06 (2H, t), 6.73 (1H, s), 6.99 (4H, s). 6.1.12. 1-(4-n-Propoxyphenoxy)-3-isopropylamino-2propanol (10) Prepared from F according to the general procedures 1 and 2, chromatographed and isolated as the hydrochloride. Yield = 3.56 g, 58.6%; m.p. = 116.5119.0 C; MS m/e 268 (M + 1); 1H-NMR (CD3OD) 1.22 (3H, t), 1.72 (6H, d), 1.792.18 (2H, m), 3.273.78 (4H, m), 3.874.29 (4H, m), 4.556.07 (1H, m), 5.485.74 (1H, m), 6.92 (4H, s). Anal. (C15H25NO3.HCl) C, H, Cl, N. 6.1.13. 4-Cyclohexylethoxyphenol (G) 4-Cyclohexylethyl bromide (10.5 g, 55 mmol), 4-benzyloxyphenol (10.11 g, 50 mmol), anhydrous K2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirred and reuxed in anhydrous acetone (250 mL) for 48 h. The solvent was evaporated and the residue was partitioned between dichloromethane and water. The organic layer was washed with water and dried over anhydrous MgSO4. The solvent was evaporated and the residual 4-(4-cyclohexylethoxy)phenylbenzylether D (R = cylohexyl) was recrystallised from EtOH. Yield = 13.6 g, 81.4%; m.p. = 6870 C; MS m/e 311 (M + 1); 1H-NMR (CDCl3) 0.891.96 (13H, m), 3.654.02 (2H, m), 4.95 (2H, s), 6.72 (4H, s), 7.24 (5H, s). The ether D (R = cyclohexyl) (12.0 g, 36 mmol) in EtOH (200 mL) was hydrogenated over 10% Pd/C (0.1 g) at room temperature and atmospheric pressure. When hydrogen absorption ceased, the reaction mixture was ltered and evaporated to dryness. The product G was recrystallised from EtOH. Yield = 7.78 g, 85.6%; m.p. = 9193 C; MS m/e 221 (M + 1); 1H-NMR (CDCl3) 0.772.01 (13H, m), 3.654.21 (2H, m), 4.47 (1H, s), 6.73 (4H, s). 6.1.14. 1-(4-Cyclohexylethoxyphenoxy)-3-isopropylamino2-propanol (15) Prepared from G according to the general procedures 1 and 2, chromatographed and isolated as the hydrochloride. Yield = 4.91 g, 66.3%; m.p. = 131.0133.0 C; MS m/e 336 (M + 1); 1H-NMR (CD3OD) 0.91 (6H, d), 0.641.53 (11H, bm), 2.80 (4H, m), 3.87 (5H, m), 6.43 (4H, s). Anal. (C20H33NO3.HCl) C, H, Cl, N. 6.1.15. 4-Fluorophenethoxyphenol (H) 4-Fluorophenethyl bromide (11.16 g, 55 mmol), 4-benzyloxyphenol (10.11 g, 50 mmol), anhydrous K2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirred and reuxed in anhydrous acetone (250 mL) for 48 h. The solvent was evaporated and the residue was partitioned between dichloromethane and water. The organic layer was washed with water and dried over anhydrous MgSO4. After evaporation of the solvent, the residual 4-(4-uorophenethoxy)phenylbenzylether D (R = 4-uorophenyl) was recrystallised from EtOH. Yield = 15.2 g, 87.9%; m.p. = 113114 C; MS m/e 323 (M + 1); 1 H-NMR (CDCl3) 3.05 (2H, m), 4.12 (2H, m), 5.11 (2H, s), 6.88 (5H, s), 7.10 (4H, dd), 7.55 (5H, dd). The ether D (R = 4-uorophenyl) (13.84 g, 40 mmol) in EtOH (250 mL) was hydrogenated over 10% Pd/C

934 (0.1 g) at room temperature and atmospheric pressure. When hydrogen absorption ceased, the reaction mixture was ltered and evaporated to dryness. The product H was recrystallised from toluene. Yield = 7.12 g, 76.1%; m.p. = 166168 C; MS m/e 233 (M + 1); 1H-NMR (CDCl3) 2.45 (2H, m), 3.56 (2H, m), 4.67 (1H, s), 6.53 (4H, dd), 6.93 (5H, dd, the phenolic hydrogen is hiding under the aromatic protons). 6.1.16. 1-(4-(4-Fluorophenethoxy)-phenoxy)-3-isopropylamino-2-propanol (16) Prepared from H according to the general procedure, chromatographed and isolated as the hydrochloride. Yield = 11.71 g, 61.4%; m.p. = 136.0138.0 C; MS m/e 348 (M + 1); 1H-NMR (CD3OD) 1.16 (6H, d), 2.713.32 (6H, m), 3.714.22 (4H, m), 6.46 (4H, s), 6.73 (4H, dd). Anal. (C20H26FNO3.HCl) C, H, Cl, F, N. 6.1.17. 4-(2-Methylpropoxy)ethoxyphenol (M) 2-Methylpropoxyethyl bromide (9.96 g, 55 mmol), 4-benzyloxyphenol (10.11 g, 50 mmol), anhydrous K2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirred and reuxed in anhydrous acetone (250 mL) for 48 h. The solvent was evaporated and the residue was partitioned between dichloromethane and water. The organic layer was washed with water and dried over anhydrous MgSO4. The solvent was evaporated and the residual 4-(2-methylpropoxyethoxy)phenylbenzylether L (R = 2-methylpropyl) was puried by chromatography (silica, eluent CH2Cl2) and isolated as an oil, homogenous by TLC. Yield = 13.8 g, 83.6%; MS m/e 301 (M + 1); 1 H-NMR (CDCl3) 1.16 (6H, d), 1.872.35 (1H, m), 3.333.60 (2H, m), 3.774.35 (4H, m), 5.12 (2H, s), 6.94 (4H, dd), 7.52 (5H, s). The ether L (R = 2-methylpropyl) (10.8 g, 36 mmol) in EtOH (200 mL) was hydrogenated over 10% Pd/C (0.1 g) at room temperature and atmospheric pressure. When hydrogen absorption ceased, the reaction mixture was ltered and evaporated to dryness. The product M was puried by chromatography (silica, eluent CH2Cl2) and isolated as an oil, homogenous by TLC. Yield = 6.85 g, 86.5%; MS m/e 221 (M + 1); 1H-NMR (CDCl3) 1.18 (6H, d), 1.972.42 (1H, m), 3.383.64 (2H, m), 3.804.39 (4H, m), 6.96 (1H, s), 7.02 (4H, dd). 6.1.18. 1-4-(2-Methylpropoxyethoxyphenoxy)-3-isopropylamino-2-propanol (18) Prepared from M according to the general procedures 1 and 2, chromatographed and isolated as the hydrochloride. Yield = 4.00 g, 55.3%; m.p. = 89.591.3 C; MS m/e 326 (M + 1); 1H-NMR (CD3OD) 1.13 (6H, d), 1.75 (6H, d), 3.373.62 (6H, bm), 3.804.37 (7H, bm), 6.94 (4H, s). Anal. (C18H31NO4.HCl) C, H, Cl, N. 6.1.19. 4-(2-Tetrahydrofurylmethoxy)ethoxyphenol (N) 2-Bromomethyl tetrahydrofuran (11.5 g, 55 mmol), 4-benzyloxyphenol (10.11 g, 50 mmol), anhydrous K2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirred and heated to reux in anhydrous acetone (250 mL) for 48 h. The solvent was evaporated and the residue was partitioned between dichloromethane and water. The organic layer was washed with water and dried over anhydrous MgSO4. The solvent was evaporated and the residual 4-(2-tetrahydrofurylmethoxy)ethoxyphenylbenzyl ether L (R = 2-tetrahydrofurylmethyl) was puried by chromatography (silica, eluent CH2Cl2) and isolated as an oil, homogenous by TLC. Yield = 14.6 g, 80.9%; MS m/e 329 (M + 1); 1H-NMR (CDCl3) 1.732.29 (4H, m), 3.484.39 (9H, m), 5.12 (2H, s), 7.03 (4H, s), 7.55 (5H, s). The ether L (R = 2-tetrahydrofurylmethyl) (11.8 g, 36 mmol) in EtOH (200 mL) was hydrogenated over 10% Pd/C (0.1 g) at room temperature and atmospheric pressure. When hydrogen absorption ceased, the reaction mixture was ltered and evaporated to dryness. The product N was puried by chromatography (silica, eluent CH2Cl2) and isolated as an oil, homogenous by TLC. Yield = 6.85 g, 80.1%; MS m/e 238 (M + 1); 1H-NMR (CDCl3) 1.732.32 (4H, m), 3.574.52 (9H, m), 6.81 (1H, s), 6.91 (4H, s). 6.1.20. 1-4-(2-Tetrahydrofurylmethoxy)ethoxyphenoxy)-3-isopropylamino-2-propanol (20) Prepared from N according to the general procedures 1 and 2, chromatographed and isolated as the hydrochloride. Yield = 3.59 g, 48.4%; m.p. = 165.5167.0 C; MS m/e 336 (M + 1); 1H-NMR (CD3OD) 0.93 (6H, d), 1.732.32 (4H, m), 2.383.23 (6H, bm), 3.574.52 (9H, m), 6.93 (4H, s). Anal. (C19H31NO5.HCl) C, H, Cl, N. 6.1.21. 4-n-Pentyloxyethoxyphenol (O) n-Pentyloxyethyl bromide (10.72 g, 55 mmol), 4-benzyloxyphenol (10.11 g, 50 mmol), anhydrous K2CO3 (8.29 g, 60 mmol) and NaI (0.5 g) were stirred and reuxed in anhydrous acetone (250 mL) for 48 h. The solvent was evaporated and the residue was partitioned between dichloromethane and water. The organic layer was washed with water and dried over anhydrous MgSO4. The solvent was evaporated and the residual 4-(4-n-pentyloxyethoxy)phenylbenzyl ether L (R = n-pentyl) was puried by chromatography (silica, eluent CH2Cl2) and isolated as an oil, homogenous by TLC. Yield = 13.7 g, 79.4%; MS m/e 315 (M + 1); 1H-NMR (CDCl3) 1.03 (3H, t), 1.192.02 (8H, m), 3.534.22 (4H, m), 5.16 (2H, s), 6.87 (4H, s), 7.24 (5H, s).

935 The ether L (R = n-pentyl) (11.3 g, 36 mmol) in EtOH (200 mL) was hydrogenated over 10% Pd/C (0.1 g) at room temperature and atmospheric pressure. When hydrogen absorption ceased, the reaction mixture was ltered and evaporated to dryness. The product O was puried by chromatography (silica, eluent CH2Cl2) and isolated as an oil, homogenous by TLC. Yield = 7.52 g, 93.25%; MS m/e 225 (M + 1); 1H-NMR (CDCl3) 1.07 (3H, t), 1.222.07 (8H, m), 3.574.29 (4H, m), 6.87 (4H, s). 6.1.22. 1-(4-n-Pentyloxyethoxyphenoxy)-3-isopropylamino-2-propanol (22) Prepared from O according to the general procedures 1 and 2, chromatographed and isolated as the hydrochloride. Yield = 3.90 g, 52.1%; m.p. = 77.079.0 C; MS m/e 340 (M + 1); 1H-NMR (CD3OD) 1.13 (3H, t), 1.75 (6H, d), 1.362.10 (8H, bm), 3.264.04 (6H, m), 3.864.28 (4H, dt), 7.03 (4H, s). Anal. (C19H33NO4.HCl) C, H, Cl, N. 6.2. Pharmacological methods 6.2.1. Isolated tissue preparations Studies were carried out on Sprague-Dawley rat isolated atria and tracheal rings according to our method described previously [16]. All tissues were allowed to equilibrate for 45 min with Krebs Ringer physiological salt solution; the composition of which in mmol L1 was NaCl, 120; KCl, 5.6; MgSO4, 1.2; CaCl2, 2.5; KH2PO4, 1.4; NaHCO3, 25; glucose 11.2 and EGTA, 0.0025. Cumulative concentration-response curves were obtained for the non-selective -AR agonist (-)-isoprenaline in each preparation [16]. (-)-Isoprenaline was dissolved and diluted in 1 mg mL1 ascorbic acid to prevent oxidation. For the measurement of antagonist activity the appropriate agent was added to the organ bath at least 30 min after the rst control concentration-response curve was completed and allowed to equilibrate for 10 min before the next concentration-response curve established. The shift in this curve to the right was calculated as a pA2 value [39]. At least three concentrations of each antagonist were examined to verify the pA2. 6.2.1.1. Rat isolated spontaneously beating atria Rat hearts were removed from adult animals (200250 g) and placed in Krebs Ringer salt solution (pH 7.4) aerated with 5% CO2 in O2. The atria were dissected free of the ventricles and overlying tissue and placed in a 20 mL bath maintained at 37 C and connected to an isotonic transducer. A tension of 1 g was applied and chronotropic activity was amplied and recorded on a Grass Polygraph. 6.2.1.2. Rat isolated tracheal chains Trachea were excised from adult rats (200250 g), dissected free of overlying tissue and cut transversely into segments about 2 mm wide. Five segments were mounted in a 20 mL bath maintained at 37 C at a tension of 1 g. Relaxation of the segments by (-)-isoprenaline was recorded by an isotonic transducer connected to a Grass Polygraph after tone had been established by administration of 1 M carbachol (45 min prior to concentrationresponse curves). 6.3. Computational methods 6.3.1. Molecular modelling Modelling studies were performed using the SYBYL software package (Tripos Inc., version 6.40) on a Silicon Graphics Indigo 2 UNIX Workstation. The structure of the active S-isomer of compound 1 was constructed from standard bond lengths and bond angles using the sketch routine in SYBYL. The geometry of compound 1 was fully optimised in vacuo as previous studies [19] have identied that the propanolamine side-chain of aryloxypropanolamine (AOPA) type compounds can adopt common conformations in solid, theoretical gas and solution states. Optimisation was conducted using the AM1 Hamiltonian in MOPAC (version 6.0) [48], the keywords PRECISE, ESP and NOMM were specied. The keyword ESP calculates the electrostatic potential derived atomic charges, these charges have been used since they are reported to be reliable and highly correlated with ab initio ESP charges [49, 50]. 6.3.2. Molecular alignment Compounds 225 were constructed using the minimised structure of compound 1 as a starting point. The appropriate para-substituent was added using the add multiple atom function in SYBYL. The structure was then optimised using the AM1 Hamiltonian in MOPAC, with the keywords PRECISE, ESP and NOMM specied, however, 1 was xed at 0 or 90 (gure 1) and the bond, dihedral and valence angles of the AOPA core structure were xed at the optimised values for compound 1. Therefore, two series of compounds were created, one where 1 = 0 and the other where 1 = 90. After optimisation the compounds were superimposed over the entire N-isopropylphenoxypropanolamine structure (compound 1) using the linear least squares routine within SYBYL, as shown in gure 7. In addition to the optimisation and superimposition of each compound a number of other physical parameters were examined:

936 6.3.3. Phenyl ring charge calculations For each compound in table II (125), the AM1 ESP charges calculated for the six phenyl ring carbon atoms were added together to obtain the parameter phenyl ring charge. Various oxypropanolamine and para-substituent conformations (ie. different 1, 2 values, gure 1) were examined for their effect upon phenyl ring charge. Both 1 and 2 were varied independently in 90 steps through a full 360, giving a total of sixteen conformers examined for each compound, and the phenyl ring charge was calculated for each conformer. The phenyl ring charge did not vary greatly between conformers and therefore the value obtained when 1 = 90 and 2 = 0 was used in this study. All sixteen phenyl ring charge values calculated for a particular compound were within 0.04e of the value obtained when 1 = 90 and 2 = 0. 6.3.4. LogP calculations The logP values for the twenty ve compounds were calculated using the Pallas PrologP (version 1.1) program at Swinburne University of Technology, Hawthorn, Victoria, Australia. The values [51] for the parasubstituents were determined by subtracting the logP of compound 1 from the compound in question and are given in table II. 6.3.5. Length of the para-substituent The length was dened as the distance between the para-carbon of the phenyl ring and the most distant non-H atom in the para-substituent, hence the parasubstituent length of compound 1 is zero (table II). The chain length of the para-substituent (R-group, table II) was calculated for the relaxed, fully extended low energy conformation of each molecule. 6.3.6. Flexibility of the para-substituent A measure of the exibility was incorporated into the QSAR analysis by assigning an integer for the number of torsion angles which affect the conformation of the para-substituent and is given in table II. For example, the exibility of compound 4 is assigned the integer 1, as rotation of terminal methyl, hydroxyl and amine groups do not affect the conformation of the para-substituent. 6.3.7. Development of the CoMFA model CoMFA [41] was performed using the QSAR option in SYBYL. Both steric and electrostatic elds were considered. The probe atom had the properties of an sp3 carbon atom and a charge of +1.0. Cut-off values were SYBYL default values and the grid step size was 2.0 . For each group of compounds (ie. 1 = 0 or 90) the effects of changing column ltering values ( 1.016.0 kcal/mol) were examined. The CoMFA QSAR equations were generated using the PLS algorithm using the leave-oneout cross validation procedure, and the number of components with the lowest standard error of prediction value was selected as the ONC. The additional physical parameters were also included in the cross validated PLS analysis to determine whether they improved q2. The nal models, non cross-validated, for the 1- and 2-ARs were generated using CoMFA, the combination of additional parameters that maximised q2 and the ONC as determined above. Acknowledgements The authors would like to thank Ms. Leanne Styan for her technical assistance and Dr Margaret Wong for her helpful comments. This work was supported by a grant from the National Health and Medical Research Council of Australia. References
[1] [2] [3] [4] [5] Greenidge P.A., Pharm. Acta Helv. 68 (1994) 201210. Lafontan M., Belan M., J. Lipid Res. 34 (1993) 10571091. Ostrowski J., Kjelsberg M.A., Caron M.G., Lefkowitz R.J., Ann. Rev. Pharmacol. Toxicol. 32 (1992) 167183. Strosberg A.D., Prot. Sci. 2 (1993b) 11981209. Emorine L.J., Marullo S., Briend-Sutren M.M., Patey G., Tate K., Delavier-Klutchko C., Strosberg A.D., Science 245 (1989) 11181121. Frielle T.J., Collins S., Daniel K.W., Caron M.G., Lefkowitz R.J., Kobilka B.K., Proc. Natl. Acad. Sci. USA 84 (1987) 79207924. Kobilka B.K., Dixon R.A.F., Frielle T., Dohlman H.G., Bolanowski M.A., Sigal I.S. et al., Proc. Natl. Acad. Sci. USA 84 (1987) 4650. Donne-Op Den Kelder G.M., Bultsma T., Timmerman H., Rademaker B., J. Med. Chem. 31 (1988) 10691079. Linschoten M.R., Bultsma T., Ijzerman A.P., Timmerman H., J. Med. Chem. 29 (1986) 278286. Ball J.B., Nero T.L., Iakovidis D., Tung L., Jackman G., Louis W.J., J. Med. Chem. 35 (1992) 46764682. Baldwin J.J., Christy M.E., Denny G.H., Habecker C.N., Freedman M.B., Lyle P.A. et al., J. Med. Chem. 29 (1986) 10651080. Machin P.J., Hurst D.N., Bradshaw R.M., Blaber L.C., Burden D.T., J. Med. Chem. 27 (1984) 503509. Berthold R., Louis W.J., US Patent, US 4, 425, 362, (1984). Hoee M.L., Hastings S.G., Meyer R.F., Corey R.M., Holmes A., Stratton C.D., J. Med. Chem. 18 (1975) 148152. Large M.S., Smith L.H., J. Med. Chem. 25 (1982) 12861292. Louis S.N.S., Nero T.L., Iakovidis D., Jackman G.P., Louis W.J., Eur. J. Pharmacol. 367 (1999) 431435. Beckett A.H., Fortschr. Arzneim. Forsch. 1 (1959) 455530. Easson L.H., Stedman E., Biochem. J. 27 (1933) 12571266. Nero T.L., Louis W.J., Iakovidis D., J. Mol. Struct. (Theochem.) 285 (1993) 251272. Brandstrom A.E., Carlsson P.A.E., Carlsson S.A., Corrodi H.R., Ek L., Ablad B.A.H., US Patent, US 3928601, (1975).

[6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]

937
[21] Carlsson P.A.E., Brandstrom A.E., Lamm B.R., Ablad B.A.H., Carlsson S.A.I., Corrodi H.R., Ek L., DE Patent, DE 2020864, (1970). Crowther A.F., Gilman D.J., McLoughlin B.J., Smith L.H., Turner R.W., Wood T.M., J. Med. Chem. 12 (1969) 638642. Erez M., Shtacher G., Weinstock M., J. Med. Chem. 21 (1978) 982984. Ferrari G., Ferrini R., Casagrande C., Boll. Chim. Farm. 107 (1968) 234248. Jain P.C., Belleau B., Rastogi S.N., Anand N., Tripathi O.N., Srimal R.C., Indian J. Chem. 12 (1974) 10211027. Leigh T., GB Patent 1185046 (1970). Loevgren K., Hedberg A., Nilsson J., Lars G., J. Med. Chem. 24 (1981) 451454. Machin P.J., Hurst D.N., Bradshaw R.M., Blaber L.C., Burden D.T., Fryer A.D., Melarange R.A., Shivdasani C., J. Med. Chem. 26 (1983) 15701576. Malguzzi R., Giordano C., EP Patent, EP 50885 (1982). Manoury P.M., Binet J.L., Rousseau J., Lefevre-Borg F.M., Cavero I.G., J. Med. Chem. 30 (1987) 10031011. Rosell G., Mauleon D., Pujol M.D., Granados R., Arch. Farmacol. Toxicol. (1986) 6570. Wick A.E., Binet J., US Patent, US 4515814 (1985). Zaagsma J., Nauta W.T., J. Med. Chem. 17 (1974) 507513. Vogel A.I., Practical Organic Chemistry 3rd ed. Longmann, London, 1962 pp. 676962. Isagulyants V.I., Tisukova V.N., Favorskaya N.A., Oganesyan P.O., Chem. Abstracts 62 (1962) 14545a. [36] [37] [38] [39] [40] Van Rossum J.M., Hurkmans J.A.T.M., Wolters C.J.J., Arch. Int. Pharmacodyn. Ther. 143 (1963) 299330. Zaborowsky B.R., McMahan W.C., Griffin W.A., Norris F.H., Ruffolo Jr. R.R., J. Pharmacol. Methods 4 (1980) 165178. MicroCal Origin, MicroCal Software Inc., USA. Mackay D.J., Pharm. Pharmacol. 30 (1978) 312313. Baldwin J.J., Denny G.H., Hirschmann R., Freedman M.B., Ponticello G.S., Gross D.M., Sweet C.S., J. Med. Chem. 26 (1983) 950957. Cramer R.D., Paterson D.E., Bunce J.D., J. Am. Chem. Soc. 110 (1988) 59595967. Triggle D.J., in: Triggle D.J., Triggle C.R. (Eds.), Chemical pharmacology of the synapse, Academic Press, London, 1976 p. 233. Belleau B., Ann. NY Acad. Sci. 139 (1967) 580605. Dixon R.A.F., Sigal I.S., Strader C.D., Cold Spring Harbor Symposium Quant. Biology. 3 (1988) 487497. Lewell X.Q., Drug Design Discovery 9 (1992) 2948. Strader C.D., Candalore M.R., Hill W.S., Dixon R.A.F., Sigal I.S., J. Biol. Chem. 264 (1989) 1357213578. Strader C.D., Sigal I.S., Register R.B., Candelore M.R., Rands E., Dixon R.A.F., Proc. Natl. Acad. Sci. USA 84 (1987) 43844388. Dewar M.J.S., Zoebisch E.G., Healy E.F., Stewart J.J.P., J. Am. Chem. Soc. 107 (1985) 39023909. Besler B.H., Merz Jr. K.M., Kollman P.A., J. Computat. Chem. 11 (1990) 431439. Ferenczy G.G., Reynolds C.A., Richards W.G., J. Computat. Chem. 11 (1990) 159169. Hansch C., Leo A., Unger S.H., Kim K.H., Nikaitani D., Lien E.J., J. Med. Chem. 16 (1973) 12071216.

[22] [23] [24] [25] [26] [27] [28]

[41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51]

[29] [30] [31] [32] [33] [34] [35]

Eur. J. Med. Chem. 34 (1999) 939952 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

939

Original article

Protease inhibitors Part 3. Synthesis of non-basic thrombin inhibitors incorporating pyridinium-sulfanilylguanidine moieties at the P1 site
Andrea Scozzafava, Fabrizio Briganti, Claudiu T. Supuran*
Universit degli Studi, Laboratorio di Chimica Inorganica e Bioinorganica, Via Gino Capponi 7, I-50121 Florence, Italy (Received 15 March 1999; revised 17 May 1999; accepted 20 May 1999)

Abstract Using benzamidine and sulfaguanidine as lead molecules, three series of derivatives have been prepared by reaction of sulfaguanidine with pyrylium salts, with the pyridinium derivatives of glycine and with the pyridinium derivatives of -alanine, respectively. The new compounds were assayed as inhibitors of two serine proteases, thrombin and trypsin. The study showed that in contrast to the leads, possessing KIs around 100300 nM against thrombin, and 1 2001 500 nM against trypsin, respectively, the new derivatives showed inhibition constants in the range of 1550 nM against thrombin, whereas their affinity for trypsin remained relatively low. Derivatives of -alanine were more active than the corresponding Gly derivatives, which in turn were more inhibitory than the pyridinium derivatives of sulfaguanidine possessing the same substitution pattern at the pyridinium ring. Thus, the present study proposes two novel approaches for the preparation of high affinity, specic thrombin inhibitors: a novel S1 anchoring moiety in the already large family of arginine/amidine-based inhibitors, i.e., the SO2N=C(NH2)2 group, and novel non-peptidomimetic scaffolds obtained by incorporating alkyl-/aryl-substitutedpyridinium moieties in the hydrophobic binding site(s). The rst one is important for obtaining bioavailable thrombin inhibitors, devoid of the high basicity of the commonly used arginine/amidine-based inhibitors, whereas the second one may lead to improved water solubility of such compounds due to facilitated salt formation as well as increased stability at hydrolysis (in vivo). 1999 ditions scientiques et mdicales Elsevier SAS thrombin / trypsin / sulfaguanidine / pyridinium salts / pyridinium amino acid / non-basic thrombin inhibitor

1. Introduction Thrombin (EC 3.4.21.5) has become an important target for drug design in recent years, in the search for low molecular-weight potent and selective inhibitors with applications as diagnostic and therapeutic agents for the increasingly common thrombotic diseases [18]. Although a large number of potent active site-directed thrombin inhibitors, such as peptide aldehydes [9, 10], boronates [11], benzamidine- [2, 3, 12, 13] or arginine/ guanidine-derived [14] inhibitors are reported, none of them meet all the criteria needed for an ideal antithrombotic drug [2, 15]. Thus, the largest majority of the presently available low-molecular weight inhibitors, such as argatroban (MQPA) 1 [16], inogatran 2 [8], NAPAP 3 [17], 4-TAPAP 4 or its 3-amidino-isomer, 3-TAPAP 5 [2, 17] (gure 1), are poorly bioavailable, either due to their high basicity, connected with the presence of
*Correspondence and reprints

guanidino-/amidino moieties in their molecule, or are not absorbable orally, or are rapidly eliminated from the circulation, mainly due to their peptidic nature. Although recently some non-basic S1 anchoring groups have been incorporated in the molecules of some thrombin inhibitors [3, 7, 18], the presence of guanidino-/benzamidino moieties in such compounds is critical, since it is by means of the interaction of these highly polar groups with Asp 189, the central amino acid residue from the specicity pocket, that the enzyme-inhibitor adduct is initially formed (obviously, a lot of other secondary interactions are responsible for the formation of high affinity adducts between thrombin and its inhibitors) [35, 1214]. In order to exploit the intrinsically high affinity of guanidino-/benzamidino-containing inhibitors for the thrombin active site, but also to avoid undesired properties connected with their too high basicity, we propose here a novel approach for designing tight-binding such inhibitors, by using sulfanilylguanidino moieties as an-

940

Figure 1. Structures of serine protease inhibitors 14.

choring groups to the S1 specicity pocket. Obviously, the presence of the SO2 group in the neighbourhood of the guanidino moiety strongly reduces the basicity of the latter, presumably without precluding to the binding of inhibitors within the enzyme active site. In this paper we report the preparation and serine protease inhibitory properties (against human thrombin and human trypsin) of three series of compounds obtained by reaction of sulfaguanidine with pyrylium salts, with the pyridinium derivatives of glycine (prepared from Gly and pyrylium salts) and with the pyridinium derivatives of -alanine (obtained from -Ala and pyrylium salts), respectively. From the point of view of their thrombin inhibitory properties, as well as that of their specicity for thrombin over trypsin, some of our compounds showed inhibition constants of the same order of magnitude as those of the clinically used compounds, argatroban (MQPA) 1 [16], and inogatran 2 [8], in the 1550 nM range against thrombin, whereas maintaining a much lower trypsin affinity (inhibition constants around

1 2001 500 nM) as compared to the above-mentioned clinically used derivatives. 2. Chemistry Compounds prepared by reaction of di-, tri- or tetrasubstituted pyrylium salts with sulfaguanidine, of types A1A16, as well as the corresponding Gly derivatives of types B1B16 and -Ala derivatives C1C16 are shown in table I. Non-exceptional synthetic procedures have been used for the reactions of pyrylium salts with nucleophiles (for the preparation of compounds A, B, C (116) as well as the pyridinium amino acid intermediates 10 and 11) [19, 20], whereas for attaching the pyridinium-amino acyl moieties, the condensation reactions in the presence of carbodiimide derivatives has been used, as outlined in gure 2 [21, 22]. Sulfanilylguanidine 7 was reacted with di-, tri- or tetrasubstituted pyrylium salts 6 leading to the pyridinium

941
Table I. Pyridinium-sulfanilylguanidines (A1A16), pyridinium-methylcarboxamido- (B1B16) and pyridinium-ethylcarboxamidosulfanilylguanidines (C1C16) prepared in the present study, with their inhibition data against human thrombin and human trypsin.

Compound

R1

R2

R3

R4

KIa Thrombin Trypsin (nM) 1 290 80 1 180 90 1 440 105 1 120 75 1 100 62 1 165 63 1 200 85 1 265 102 1 240 77 1 260 80 1 210 104 1 340 120 1 170 96 1 950 130 1 950 140 1 300 90 1 210 91 1 120 60 1 320 95 1 100 72 1 020 60 1 150 45 1 175 75 1 250 88 1 200 105 1 210 65 1 175 90 1 210 60 1 100 55 1 900 60 1 905 65 1 265 90 1 140 87 1 100 100 1 290 97 1 055 73 1 010 55 1 105 102 1 100 79 1 210 75 1 165 50 1 200 60 1 155 70 1 200 85 1 130 50 1 380 45 1 520 50 1 210 40

A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 A13 A14 A15 A16 B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 B11 B12 B13 B14 B15 B16 C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16
a

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

Me i-Pr i-Pr Me Et n-Pr i-Pr Me Et n-Pr i-Pr n-Bu t-Bu Ph Ph Me Me i-Pr i-Pr Me Et n-Pr i-Pr Me Et n-Pr i-Pr n-Bu t-Bu Ph Ph Me Me i-Pr i-Pr Me Et n-Pr i-Pr Me Et n-Pr i-Pr n-Bu t-Bu Ph Ph Me

H H H H H H H H H H H H H H H Me H H H H H H H H H H H H H H H Me H H H H H H H H H H H H H H H Me

Me Me Me Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph H Me Me Me Me Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph H Me Me Me Me Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph Ph H Me

Me Me i-Pr Me Et n-Pr i-Pr Ph Ph Ph Ph Ph Ph Ph Ph Me Me Me i-Pr Me Et n-Pr i-Pr Ph Ph Ph Ph Ph Ph Ph Ph Me Me Me i-Pr Me Et n-Pr i-Pr Ph Ph Ph Ph Ph Ph Ph Ph Me

83 5 76 6 102 5 41 2 37 3 54 7 48 5 32 2 30 4 36 5 34 2 60 5 33 3 54 4 58 6 79 6 75 5 62 4 80 7 34 3 30 3 50 4 43 5 21 2 17 1 23 2 22 2 50 5 25 3 50 4 57 5 73 6 69 4 47 4 71 5 29 2 26 2 47 5 41 4 18 2 15 1 21 2 20 1 47 5 19 2 42 5 50 3 70 5

KI values were obtained from Dixon plots using a linear regression program [26], from at least three different assays.

942

Figure 2. Synthesis of derivatives A1A16, B1B16 and C1C16.

derivatives A1A16. Alternatively, reaction of pyrylium salts with Gly or -Ala afforded the pyridinium amino acid derivatives 10 and 11, which were coupled with 7 in the presence of EDCI or diisopropylcarbodiimide as condensing agents, leading to compounds B1B16, and C1C16, respectively. 3. Pharmacology Inhibition data against two serine proteases, human thrombin and human trypsin are shown in table I. The chromogenic substrate Chromozym TH (Ts-Gly-Pro-Argp-nitroanilide) was used in the assay, with the spectrophotometric method of Lottenberg et al. [23]. Inhibition data with the standard serine protease inhibitors 13 are also provided for comparison in table II. pKa values for the guanidino/amidino and sulfonamido moieties for some of the thrombin inhibitors reported here, as well as standard compounds, are shown in table III. 4. Results and discussion The lead molecule for obtaining novel types of thrombin inhibitors considered by us was benzamidine 12, one of the simplest of such compounds, which possesses an inhibition constant KI = 300 nM against human thrombin; moreover, the X-ray crystallographic structure for the

complex of benzamidine with this enzyme has recently been reported (PDB entry: 1DWB) [24]. From the X-ray data it was observed that the amidino moiety of the inhibitor is anchored to the S1 specicity pocket of the enzyme, interacting electrostatically and by means of hydrogen bonds with Asp 189. Several other van der Waals contacts between the inhibitor molecule and the enzyme were also evidenced [24]. Obviously, benzamidine is a weak thrombin inhibitor, since the binding energy is only gained due to the strong electrostatic interaction of the carboxylate of Asp 189 and the positively charged amidino moiety. On the other hand, as already mentioned in the introductory section, the amidino moiety possesses too high a basicity for allowing the formation of bioavailable enzyme inhibitors, and it thus
Table II. Inhibition data of two serine proteases with standard inhibitors 13 and 12, and sulfaguanidine 7. Compound 1 2 3 7 12
a

KI (nM)a Argatrobanb Inogatran NAPAP Sulfaguanidine Benzamidine Thrombin 19 2 15 1 6.5 0.05 95 4 300 5 Trypsin 540 11 690 24 1 350 120 450 6

KI values were obtained from Dixon plots using a linear regression program [26], from at least three different assays. bFrom ref. [5].

943
Table III. pKa data of serine protease inhibitors 13, 7, A9, B9 and C9. Compound pKa
a

1 2 3 7 A9 B9 C9

Argatrobanb Inogatranb NAPAP Sulfaguanidine

Guanidino/amidino moiety 12.5 12.3 12.6 8.4 8.1 8.3 8.5

SO2NH moiety 7.0 7.0 7.1 7.1

Figure 3. Tautomeric forms of compound 13.

a pKa values were determined in 30% Et-OH/water (v/v) as described in the Experimental section. bFrom ref. [8].

appeared of great interest to elaborate on non-basic variants of this interesting serine protease anchoring group. The sulfonylguanidino moiety appeared as an attractive candidate for such a purpose, since the presence of the SO2 moiety in the neighbourhood of the strong base, guanidine, should drastically weaken its basicity. Such modied anchoring groups should not presumably interfere with the binding of the inhibitor to the enzyme, since the hydrogen-bonding donor/acceptor properties, as well as the possibility to interact electrostatically with the enzyme for the compounds incorporating them, should not differ too much from those of the classical amidino-/ guanidino-based inhibitors of types 15 or 12. The sulfonyl-guanidines possess a large number of possible tautomeric forms, and this factor might also be a critical one for the binding of such a compound to thrombin. Thus, in previous work [25] we have shown that arylsulfonyl guanidines, including sulfaguanidine 7, possess moderate but specic thrombin inhibitory properties. Moreover, by means of AM1 and MOPAC calculations it was demonstrated that the tautomer of type 13A of benzenesulfonylguanidine is much more stable than the tautomer 13B (gure 3), a situation that seems to be important for binding to the enzyme [25]. Thus, we presume that the same is true for the pyridinium-based compounds reported here, i.e., that the symmetrical tautomers of type 13A are more stable than the corresponding non-symmetrical tautomers of type 13B. It is obvious from the above data that the symmetrical nature of the favoured tautomer should enable stronger interactions with the carboxylate moiety of Asp 189 and presumably, the formation of high affinity E-I adducts. Thus, three series of pyridinium containing sulfanilylguanidines A1A16, B1B16 and C1C16 were prepared in order to test the above-mentioned hypothesis (table I). These compounds were obtained by reactions of

pyrylium salts with sulfaguanidine, or alternatively, by condensation of sulfaguanidine with the pyridinium derivatives of glycine or -alanine, obtained from the two mentioned amino acids and pyrylium salts, by the original procedure of Balabans group [2732]. The following should be noted regarding the serine protease inhibition data of tables I and II with the new compounds and standard inhibitors: (i) the pyridinium derivatives A, B, C (116) reported here generally behave as stronger thrombin inhibitors as compared to the lead molecule from which they were derived, i.e., benzamidine 12 and sulfaguanidine 7. At the same time, their affinity for trypsin is relatively low, which constitutes a positive feature for the putative clinical use of such compounds; (ii) in the three subseries of investigated compounds, thrombin inhibitory properties increased from the pyridinium derivatives of sulfaguanidine A (116) to the corresponding pyridinium-Gly-derivatives B (116), with the pyridinium--Ala derivatives C(1-16) behaving as the most active inhibitors in the whole series of reported compounds (obviously, this discussion takes into account the same substitution pattern at the pyridinium ring for compounds in the three investigated subseries); (iii) the nature of R1R4 groups substituting the pyridinium ring was critical for the biological activity of the obtained compounds, similarly to the situation evidenced for the carbonic anhydrase sulfonamide inhibitors reported previously [19, 20]. Thus, tri- or tetraalkylpyridinium- as well as 2,6-di- or 2,4,6-triphenylpyridinium moieties were generally less effective than 2-alkyl-4,6-diphenyl-pyridinium groups in inducing strong thrombin inhibitory properties to the compounds incorporating them. Practically, the most active derivatives in all three subseries were those containing 2-alkyl4,6-diphenyl-pyridinium moieties, such as 2-methyl-; 2-ethyl-; 2-iso-propyl- or 2-tert-butyl-4,6-diphenylpyridinium groups. Replacing the 2-alkyl group mentioned above with a bulky phenyl one (such as in compounds A14, B14 or C14) or with a longer aliphatic

944 chain (n-butyl, such as in A12, B12 or C12) led to a drastic reduction of the thrombin inhibitory effects of the corresponding compounds. On the other hand, compounds possessing 2,6-dialkyl-4-phenyl-pyridinium moieties in their molecule (such as A, B, C (4 and 5)) possessed a behaviour intermediate between the strong inhibitors of the type A, B, C (8, 9, 11 and 13) and the relatively weak inhibitors of type A, B, C (13 and 1416). Anyhow, the best substitution for inducing strong thrombin inhibitory properties was that incorporating the 2-ethyl-4,6-diphenylpyridinium moiety in the molecules of the new derivatives. Some of the compounds containing this substitution pattern, such as B9 and C9 (but also the structurally-related compounds B8, B10, B11, C8, C10 and C11) showed thrombin inhibitory properties of the same order of magnitude as the clinically used derivatives argatroban 1 and inogatran 2, although they are less effective as compared to the very potent inhibitor NAPAP (table II). A special mention should be made regarding the fact that the new compounds reported here possess a much lower affinity for trypsin as compared to the standard inhibitors 13, which constitutes a highly desirable feature for a compound to be developed for clinical use. pKa values for the amidino/guanidino as well as sulfonamido moieties of some of the newly synthesized serine protease inhibitors and standard compounds such as inogatran, argatroban and NAPAP (table III) prove that the approach proposed here for reducing the basicity of such an enzyme inhibitor is a successful one. Thus, unlike the highly basic guanidines/amidines of type 13 (pKas around 12.312.6), sulfaguanidine 7 and its derivatives reported here (such as compounds A9, B9 or C9) have pKa values of the guanidino moiety around 8.18.5, being at least 104 times less basic than the previously mentioned derivatives. Furthermore, due to the presence of the sulfonyl moiety in their molecules, these compounds also possess a weakly acidic character, with another ionization step around the pKa value of 7, due to the loss of the SO2NH proton. These features should positively inuence the pharmacological prole of a thrombin inhibitor of the type described here. The strong thrombin inhibitory properties of some compounds reported in this study might be explained by taking into account the X-ray crystallographic structure of the enzyme as well as those of some of its adducts with guanidine/amidine-based inhibitors [4, 5, 33, 34]. Thus, it was shown that effective binding is achieved when a proline, a pipecolic acid or a similarly non-hydrophilic moiety is present in the P2 position, which allows favourable interactions with the enzyme S2 cavity, comprising among others, amino acid residues Trp 60D and Tyr 60A as well as when hydrophobic (generally aromatic: Ph, Ts; naphthyl) groups are present at P3, which allow strong interactions with the S3 site, comprising residues Leu 99; Trp 215 and Ile 174 among others [4, 5, 33, 34]. Some moieties present in the compounds prepared by us might possess just the required structural elements for the formation of high affinity adducts with thrombin. For example, for the strongest inhibitor reported in this paper, C9 (KI = 15 nM against thrombin), the CH2CH2CO moiety might interact with the S2 cavity, whereas the two phenyls substituting the pyridinium moiety probably bind within the aryl binding site (S3). Obviously, the sulfonylguanidino moiety of all these inhibitors probably lls the S1 specicity pocket, interacting with Asp 189, as discussed earlier. But another aspect might be important for explaining the relatively high affinity of this entire class of inhibitors for thrombin. Thus, around the entrance of the specicity pocket of this enzyme, ten negatively-charged amino acid residues are clustered [4, 5, 33, 34]. Some of these (such as Asp 189 and Glu 192) are directly involved in the substrate/ inhibitor recognition process, whereas some others might be crucial for driving or stabilizing the inhibitor within the active site [4]. In the case of cationic inhibitors, such as the compounds reported in the present paper, the presence of such a cluster of ten negatively-charged residues at the entrance of the active site might be extremely favourable for obtaining strong E-I adducts, due to the possibility of salt bridge formation between the cationic moiety of the inhibitor and the anionic groups of the enzyme. Associated with precise geometric requirements for binding to the S2 and S3 sites (mentioned above) our approach led to high affinity, less basic thrombin inhibitors. 5. Conclusion Three series of cationic sulfaguanidine derivatives have been prepared by reaction of sulfaguanidine with di-, tri- or tetrasubstituted pyrylium salts bearing alkyl, aryl or a combination of the two moieties in their molecule, and with the corresponding Gly-pyridinium and -Ala pyridinium derivatives, respectively. Qualitative SAR proved that the best activity for inhibiting thrombin was obtained for compounds bearing 2-alkyl4,6-diaryl- pyridinium moieties, and that the -Ala derivatives were more active than the corresponding Gly derivatives, which in turn were more active than the corresponding pyridinium-sulfaguanidines. The obtained compounds generally possessed a low affinity for trypsin, which might be considered a positive feature for the putative pharmacological development of such a throm-

945 bin inhibitor. Thus, our study proposes two novel approaches for the preparation of high affinity, specic thrombin inhibitors: 1) a novel S1 anchoring moiety of the arginine/amidine type, i.e., the SO2N=C(NH2)2 group; and 2) novel non-peptidomimetic scaffolds obtained by incorporating alkyl-/aryl-substituted-pyridinium moieties in the hydrophobic binding site(s). The rst approach is important for obtaining bioavailable thrombin inhibitors devoid of the high basicity of the commonly used arginine/amidine-based inhibitors, with some of the new derivatives proving to be 104 times less basic than the standard compounds in clinical use. The second one may lead to improved water solubility of such derivatives due to facilitated salt formation as well as increased in vivo stability at hydrolysis. 6. Experimental 6.1. Chemistry Melting points: heating plate microscope (not corrected); IR spectra: KBr pellets, 4004 000 cm1 PerkinElmer 16PC FTIR spectrometer; 1H-NMR spectra: Varian 300CXP apparatus (chemical shifts are expressed as values relative to Me4Si as standard); Elemental analysis ( 0.4% of the theoretical values, calculated for the proposed formulas, data not shown): Carlo Erba Instrument CHNS Elemental Analyzer, Model 1106. All reactions were monitored by thin-layer chromatography (TLC) using 0.25 mm precoated silica gel plates (E. Merck). Preparative HPLC was performed on a Dynamax-60A column (25 250 mm), with a Beckman EM-1760 instrument. The detection wavelength was 254 nm. Sulfaguanidine, triethylamine, carbodiimides, and amino acids used in the syntheses were commercially available compounds (from Sigma, Acros or Aldrich). Pyrylium salts were prepared as described in the literature [3032]. Acetonitrile, acetone, dioxane, ethyl acetate (E. Merck, Darmstadt, Germany) or other solvents used in the synthesis were doubly distilled and kept on molecular sieves in order to maintain them in anhydrous conditions. Inogatran was from Astra Hassle (Molndal, Sweden). Benzamidine, NAPAP, human thrombin, human trypsin and Chromozym TH were from Sigma Chem. Co. (St Louis, MO, USA). 6.1.1. General procedure for the preparation of compounds A (116) Method A: an amount of 0.21 g (1 mM) of sulfaguanidine 7 and the stoichiometric amount of pyrylium salt 6 and 140 L of triethylamine (1 mM) were dissolved/ suspended in 20 mL of absolute methanol. The mixture was reuxed for 30 min, then 0.45 mL of glacial acetic acid were added and reuxation was continued for another 2 h. The cold mixture was treated with 100200 mL of diethyl ether for the precipitation of the pyridinium salts A1A16 which were recrystallized from water with 25% perchloric acid. Method B: an amount of 0.60 g (2.9 mM) of sulfaguanidine 7 and 2.9 mM of pyrylium salt 6 were suspended in 5 mL of anhydrous methanol and poured into a stirred mixture of 14.5 mM of triethylamine and 5.8 mM of acetic anhydride. After 5 min of stirring, another 10 mL of methanol were added to the reaction mixture, which was heated to reux for 15 min. Then 14.5 mM of acetic acid was added and heating was continued for 25 h. The role of the acetic anhydride is to react with the water formed during the condensation reaction between the pyrylium salt and the aromatic amine, in order to shift the equilibrium towards the formation of the pyridinium salts of type A1A16. In the case of sulfaguanidine, this procedure is the only one which gave acceptable yields in pyridinium salts possessing 2-methyl groups. The precipitated pyridinium salts obtained were then puried by treatment with concentrated ammonia solution (which also converts the eventually unreacted pyrylium salt to the corresponding pyridine which is soluble in acidic medium), reprecipitation with perchloric acid and recrystallization from water with 25% HClO4. 6.1.1.1. 1-N-(4-Guanidinosulfonyl-phenyl)-2,4,6-trimethylpyridinium perchlorate A1 White crystals, m.p. 273275 C (yield of 34%); IR (KBr), cm1: 625, 740, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 040, 3 245, 3 335; 1H-NMR (TFA), ppm: 2.56 (s, 6H, 2,6-(Me)2); 2.81 (s, 3H, 4-Me); 7.357.85 (m, AABB, 4H, ArH from 1,4-phenylene); 8.10 (s, 2H, ArH, 3,5-H from pyridinium). Anal. C15H19N4O2S+ ClO4 (C, H, N, S). 6.1.1.2. 1-N-(4-Guanidinosulfonyl-phenyl)-2-iso-propyl4,6-dimethylpyridinium perchlorate A2 Pale yellow crystals, m.p. 255256 C (yield of 51%); IR (KBr), cm1: 625, 680, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 020, 3 235; 1H-NMR (TFA), ppm: 1.50 (d, 6H, 2Me from i-Pr); 2.70 (s, 3H, 6-Me); 2.83 (s, 3H, 4-Me); 3.48 (heptet, 1H, CH from i-Pr); 7.258.45 (m, AABB, 4H, ArH from 1,4-phenylene); 7.98 (s, 2H, ArH, 3,5-H from pyridinium). Anal. C17H23N4O2S+ ClO4 (C, H, N, S). 6.1.1.3. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-di-isopropyl-4-methylpyridinium perchlorate A3 Tan crystals, m.p. 201202 C (yield of 76%); IR (KBr), cm1: 625, 685, 820, 1 100, 1 175, 1 290, 1 345,

946 1 580, 1 675, 3 030, 3 250; 1H-NMR (TFA), ppm: 1.51 (d, 12H, 4Me from 2 i-Pr); 2.80 (s, 3H, 4-Me); 3.42 (heptet, 2H, 2CH from 2 i-Pr); 7.318.51 (m, AABB, 4H, ArH from 1,4-phenylene); 8.05 (s, 2H, ArH, 3,5-H from pyridinium). Anal. C19H27N4O2S+ ClO4 (C, H, N, S). 6.1.1.4. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-dimethyl4-phenylpyridinium perchlorate A4 White crystals, m.p. 280281 C (yield of 50%); IR (KBr), cm1: 625, 690, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 030, 3 260, 3 330; 1H-NMR (TFA), ppm: 2.58 (s, 6H, 2,6-(Me)2); 8.109.12 (m, 11H, ArH from 1,4phenylene, pyridinium and 4-Ph). Anal. C20H21N4O2S+ ClO4 (C, H, N, S). 6.1.1.5. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-diethyl-4phenylpyridinium perchlorate A5 Yellow crystals, m.p. 263265 C (yield of 37%); IR (KBr), cm1: 625, 765, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 040, 3 270, 3 360; 1H-NMR (TFA), ppm: 1.43 (t, 6H, 2 Me from ethyl); 2.82 (q, 4H, 2 CH2 from Et); 7.688.87 (m, 11H, ArH from 1,4-phenylene, pyridinium and 4-Ph). Anal. C22H25N4O2S+ ClO4 (C, H, N, S). 6.1.1.6. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-di-npropyl-4-phenylpyridinium perchlorate A6 Yellowish crystals, m.p. 215216 C (yield of 61%); IR (KBr), cm1: 625, 775, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 060, 3 220, 3 315; 1H-NMR (TFA), ppm: 1.01 (t, 6H, 2 Me from propyl); 1.70 (sextet, 4H, 2CH2 () from n-Pr); 2.80 (t, 4H, 2 CH2 () from n-Pr); 7.558.78 (m, 11H, ArH from 1,4-phenylene, pyridinium and 4-Ph). Anal. C24H29N4O2S+ ClO4 (C, H, N, S). 6.1.1.7. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-di-isopropyl-4-phenylpyridinium perchlorate A7 White crystals, m.p. 198199 C (yield of 24%); IR (KBr), cm1: 625, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 060, 3 270, 3 315; 1H-NMR (TFA), ppm: 1.45 (d, 12H, 4 Me from i-Pr); 2.95 (heptet, 2H, 2 CH from i-Pr); 7.928.97 (m, 11H, ArH from 1,4-phenylene, pyridinium and 4-Ph). Anal. C24H29N4O2S+ ClO4 (C, H, N, S). 6.1.1.8. 1-N-(4-Guanidinosulfonyl-phenyl)-2-methyl-4,6diphenylpyridinium perchlorate A8 White crystals, m.p. 262263 C (yield of 30%); IR (KBr), cm1: 625, 770, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 040, 3 245, 3 350; 1H-NMR (TFA), ppm: 2.72 (s, 3H, 2-Me); 7.558.73 (m, 16H, ArH from 1,4phenylene, pyridinium and 4,6-Ph2). Anal. C25H23N4O2S+ ClO4 (C, H, N, S). 6.1.1.9. 1-N-(4-Guanidinosulfonyl-phenyl)-2-ethyl-4,6diphenylpyridinium perchlorate A9 White-yellow crystals, m.p. 233234 C (yield of 39%); IR (KBr), cm1: 625, 700, 770, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 040, 3 250, 3 350; 1H-NMR (TFA), ppm: 1.50 (t, 3H, Me from ethyl); 2.97 (q, 2H, CH2); 7.408.57 (m, 16H, ArH from 1,4-phenylene, pyridinium and 4,6-Ph2). Anal. C26H25N4O2S+ ClO4 (C, H, N, S). 6.1.1.10. 1-N-(4-Guanidinosulfonyl-phenyl)-2-n-propyl4,6-diphenylpyridinium perchlorate A10 White crystals, m.p. 243244 C (yield of 36%); IR (KBr), cm1: 625, 700, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 030, 3 270, 3 350; 1H-NMR (TFA), ppm: 1.05 (t, 3H, Me from propyl); 1.93 (sextet, 2H, -CH2 from n-Pr); 2.93 (t, 2H, -CH2 from n-Pr); 7.388.53 (m, 16H, ArH from 1,4-phenylene, pyridinium and 4,6-Ph2). Anal. C27H27N4O2S+ ClO4 (C, H, N, S). 6.1.1.11. 1-N-(4-Guanidinosulfonyl-phenyl)-2-isopropyl-4,6-diphenylpyridinium perchlorate A11 White crystals, m.p. 183184 C (yield of 25%); IR (KBr), cm1: 625, 700, 770, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 040, 3 250, 3 360; 1H-NMR (TFA), ppm: 1.52 (d, 6H, 2 Me from i-propyl); 2.523.25 (m, 1H, CH from i-Pr); 7.338.60 (m, 16H, ArH from 1,4phenylene, pyridinium and 4,6-Ph2). Anal. C27H27N4O2S+ ClO4 (C, H, N, S). 6.1.1.12. 1-N-(4-Guanidinosulfonyl-phenyl)-2-n-butyl4,6-diphenylpyridinium perchlorate A12 White crystals, m.p. 244245 C (yield of 72%); IR (KBr), cm1: 625, 710, 770, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 060, 3 260, 3 345; 1H-NMR (TFA), ppm: 0.90 (t, 3H, Me from butyl); 1.102.15 (m, 4H, CH3-CH2-CH2-CH2 from n-Bu); 2.97 (t, 2H, -CH2 from n-Bu); 7.258.52 (m, 16H, ArH from 1,4-phenylene, pyridinium and 4,6-Ph2). Anal. C28H29N4O2S+ ClO4 (C, H, N, S). 6.1.1.13. 1-N-(4-Guanidinosulfonyl-phenylmethyl)-2tert-butyl-4,6-diphenylpyridinium perchlorate A13 White crystals, m.p. 197198 C (yield of 62%); IR (KBr), cm1: 625, 765, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 060, 3 270; 1H-NMR (TFA), ppm: 1.90 (s, 9H, t-Bu); 6.838.83 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C28H29N4O2S+ ClO4 (C, H, N, S).

947 6.1.1.14. 1-N-(4-Guanidinosulfonyl-phenyl)-2,4,6-triphenylpyridinium perchlorate A14 Yellow crystals, m.p. 234235 C (yield of 80%); IR (KBr), cm1: 625, 700, 770, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 030, 3 260, 3 350; 1H-NMR (TFA), ppm: 7.478.63 (m, 21H, ArH from 1,4-phenylene, pyridinium and 2,4,6-Ph3). Anal. C30H25N4O2S+ ClO4 (C, H, N, S). 6.1.1.15. 1-N-(4-Guanidinosulfonyl-phenyl)-2,6-diphenylpyridinium perchlorate A15 Yellow-orange crystals, m.p. 250252 C (yield of 36%); IR (KBr), cm1: 625, 705, 765, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 050, 3 260; 1H-NMR (TFA), ppm: 6.718.40 (m, 17H, ArH from 1,4phenylene, 2,6-Ph2 and 3,4,5-H from pyridinium). Anal. C24H20N4O2S+ ClO4 (C, H, N, S). 6.1.1.16. 1-N-(4-Guanidinosulfonyl-phenyl)-2,3,4,6tetramethylpyridinium perchlorate A16 White crystals, m.p. 256257 C (yield of 28%); IR (KBr), cm1: 625, 750, 1 100, 1 175, 1 290, 1 345, 1 580, 1 675, 3 040, 3 245, 3 330; 1H-NMR (TFA), ppm: 2.45 (s, 3H, 3-Me); 2.50 (s, 3H, 4-Me); 2.55 (s, 3H, 6-Me); 2.75 (s, 3H, 2-Me); 8.039.17 (m, 5H, ArH from 1,4phenylene and pyridinium 5-H). Anal. C16H21N4O2S+ ClO4 (C, H, N, S). 6.1.2. General procedure for the preparation of derivatives 10 and 11 An amount of 10 mM of amino acid (Gly or -Ala) was suspended/dissolved in 50 mL of anhydrous acetonitrile and the stoichiometric amount (10 mM) of pyrylium salt 6 and triethyl amine (10 mM, 1.47 mL) were added. The reaction mixture was heated at reux for 4 h, then 2.5 mL of glacial acetic acid were added and reuxation was continued for another 2 h. The obtained reaction mixture was treated as described above (Method A), in order to obtain the pure intermediates 10 and 11 (recrystallized from water with 25% perchloric acid). 6.1.3.General procedure for the preparation of compounds B and C (116) An amount of 1 mM of pyridinium-amino acid derivative 10 or 11 was dissolved/suspended in 25 mL of anhydrous acetonitrile or acetone, and then treated with 210 mg (1 mM) of sulfaguanidine 7 and 190 mg (1 mM) of EDCI. HCl or di-isopropyl-carbodiimide. The reaction mixture was magnetically stirred at room temperature for 15 min, then 30 L (2 mM) of triethylamine were added and stirring was continued for 16 h at 4 C. The solvent was evaporated in vacuo and the residue taken up in ethyl acetate (5 mL), poured into a 5% solution of sodium bicarbonate (5 mL) and extracted with ethyl acetate. The combined organic layers were dried over sodium sulfate and ltered, and the solvent removed in vacuo. Preparative HPLC (Dynamax-60A column (25 250 mm); 90% acetonitrile/8% methanol/2% water; ow rate of 30 mL/ min) afforded the pure compounds B and C (116) as colourless solids. 6.1.3.1. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2,4,6-trimethylpyridinium perchlorate B1 White-tan crystals, m.p. 280282 C (yield of 80%); IR (KBr), cm1: 625, 680, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 030, 3 250; 1H-NMR (TFA), ppm: 2.70 (s, 3H, 4-Me); 2.85 (s, 6H, 2,6(Me)2); 4.12 (s, 2H, Gly CH2); 7.138.41 (m, AABB, 4H, ArH from 1,4-phenylene); 8.00 (s, 2H, ArH, 3,5-H from pyridinium). Anal. C17H21N4O3S+ ClO4 (C, H, N, S). 6.1.3.2. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2-iso-propyl-4,6-dimethylpyridinium perchlorate B2 Light orange crystals, m.p. 205207 C (yield of 64%); IR (KBr), cm1: 625, 680, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 020, 3 235; 1H-NMR (TFA), ppm: 1.50 (d, 6H, 2Me from i-Pr); 2.80 (s, 3H, 6-Me); 2.90 (s, 3H, 4-Me); 3.48 (heptet, 1H, CH from i-Pr); 4.12 (s, 2H, Gly CH2); 7.258.43 (m, AABB, 4H, ArH from 1,4-phenylene); 7.98 (s, 2H, ArH, 3,5-H from pyridinium). Anal. C19H25N4O3S+ ClO4 (C, H, N, S). 6.1.3.3. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2,6-di-iso-propyl-4-methylpyridiniumperchlorate B3 Tan crystals, m.p. 202203 C (yield of 75%); IR (KBr), cm1: 625, 820, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 030, 3 250; 1H-NMR (TFA), ppm: 1.51 (d, 12H, 4Me from 2 i-Pr); 2.83 (s, 3H, 4-Me); 3.42 (heptet, 2H, 2CH from 2 i-Pr); 4.12 (s, 2H, CH2); 7.318.51 (m, AABB, 4H, ArH from 1,4-phenylene); 8.03 (s, 2H, ArH, 3,5-H from pyridinium). Anal. C21H29N4O3S+ ClO4 (C, H, N, S). 6.1.3.4. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2,6-dimethyl-4-phenylpyridinium perchlorate B4 Orange-red crystals, m.p. 240241 C (yield of 69%); IR (KBr), cm1: 625, 765, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 050, 3 265; 1H-NMR (TFA), ppm: 3.00 (s, 6H, 2,6-(Me)2); 4.12 (s, 2H, CH2); 7.218.51 (m, 11H, ArH from 1,4-phenylene, 4-Ph and 3,5-H from pyridinium). Anal. C22H23N4O3S+ ClO4 (C, H, N, S).

948 6.1.3.5. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2,6-diethyl-4-phenylpyridinium perchlorate B5 Tan crystals, m.p. 223224 C (yield of 53%); IR (KBr), cm1: 625, 770, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 060, 3 230; 1H-NMR (TFA), ppm: 1.55 (t, 6H, 2 Me from Et); 3.30 (q, 4H, 2 CH2 from Et); 4.12 (s, 2H, N+ -CH2); 7.088.63 (m, 11H, ArH from 1,4-phenylene, 4-Ph and 3,5-H from pyridinium). Anal. C24H27N4O3S+ ClO4 (C, H, N, S). 6.1.3.6. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2,6-di-n-propyl-4-phenylpyridinium perchlorate B6 Tan crystals, m.p. 218219 C (yield of 55%); IR (KBr), cm1: 625, 775, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 060, 3 240; 1H-NMR (TFA), ppm: 1.15 (t, 6H, 2 Me from Pr); 1.90 (sextet, 4H, 2 CH2 from Pr); 3.18 (t, 4H, 2 CH2 from Pr); 4.12 (s, 2H, N+-CH2); 7.108.50 (m, 11H, ArH from 1,4-phenylene, 4-Ph and 3,5-H from pyridinium). Anal. C26H31N4O3S+ ClO4 (C, H, N, S). 6.1.3.7. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2,6-di-iso-propyl-4-phenylpyridiniumperchlorate B7 Tan crystals, m.p. 210213 C (yield of 79%); IR (KBr), cm1: 625, 775, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 060, 3 240; 1H-NMR (TFA), ppm: 1.55 (d, 12H, 4 Me from i-Pr); 3.53 (heptet, 2H, 2 CH from i-Pr); 4.13 (s, 2H, N+-CH2); 7.238.65 (m, 11H, ArH from 1,4-phenylene, 4-Ph and 3,5-H from pyridinium). Anal. C26H31N4O3S+ ClO4 (C, H, N, S). 6.1.3.8. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2-methyl-4,6-diphenylpyridinium perchlorate B8 Yellow crystals, m.p. 254255 C (yield of 51%); IR (KBr), cm1: 625, 770, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 050, 3 250; 1H-NMR (TFA), ppm: 3.00 (s, 3H, 2-Me); 4.12 (s, 2H, CH2); 7.088.58 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C27H25N4O3S+ ClO4 (C, H, N, S). 6.1.3.9. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2-ethyl-4,6-diphenylpyridinium perchlorate B9 White crystals, m.p. 221223 C (yield of 84%); IR (KBr), cm1: 625, 705, 770, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 050, 3 250; 1H-NMR (TFA), ppm: 1.60 (t, 3H, Me from Et); 3.27 (q, 2H, CH2 from Et); 4.12 (s, 2H, N+-CH2); 7.088.60 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C28H27N4O3S+ ClO4 (C, H, N, S). 6.1.3.10. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2-n-propyl-4,6-diphenylpyridinium perchlorate B10 White-yellowish crystals, m.p. 200201 C (yield of 53%); IR (KBr), cm1: 625, 685, 770, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 080, 3 250; 1 H-NMR (TFA), ppm: 1.18 (t, 3H, Me from Pr); 2.10 (sextet, 2H, CH2 from n-Pr); 3.20 (t, 2H, CH2 from n-Pr); 4.12 (s, 2H, N+-CH2); 7.088.63 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C29H29N4O3S+ ClO4 (C, H, N, S). 6.1.3.11. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2-iso-propyl-4,6-diphenylpyridinium perchlorate B11 Tan crystals, m.p. 187188 C (yield of 62%); IR (KBr), cm1: 625, 710, 770, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 070, 3 250; 1H-NMR (TFA), ppm: 1.55 (d, 6H, 2 Me from i-Pr); 3.55 (heptet, 1H, CH from i-Pr); 4.10 (s, 2H, N+-CH2); 7.088.63 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C29H29N4O3S+ ClO4 (C, H, N, S). 6.1.3.12. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2-n-butyl-4,6-diphenylpyridinium perchlorate B12 Tan crystals, m.p. 198199 C (yield of 43%); IR (KBr), cm1: 625, 690, 770, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 080, 3 250; 1H-NMR (TFA), ppm: 0.93 (t, 3H, Me from n-Bu); 1.55 (sextet, 2H, CH2 from n-Bu); 2.05 (quintet, 2H, CH2 from n-Bu); 3.17 (t, 2H, CH2 from n-Bu); 4.12 (s, 2H, N+-CH2); 7.088.58 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C30H31N4O3S+ ClO4 (C, H, N, S). 6.1.3.13. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2-tert-butyl-4,6-diphenylpyridinium perchlorate B13 White crystals, m.p. 201203 C (yield of 54%); IR (KBr), cm1: 625, 705, 765, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 060, 3 270; 1H-NMR (TFA), ppm: 1.90 (s, 9H, t-Bu); 4.22 (s, 2H, CH2); 6.838.83 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C30H31N4O3S+ ClO4 (C, H, N, S).

949 6.1.3.14. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2,4,6-triphenylpyridinium perchlorate B14 Orange crystals, m.p. 234235 C (yield of 70%); IR (KBr), cm1: 625, 705, 770, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 050, 3 270; 1H-NMR (TFA), ppm: 4.09 (s, 2H, CH2); 6.708.56 (m, 21H, ArH from 1,4-phenylene, 2,4,6-Ph3 and 3,5-H from pyridinium). Anal. C32H27N4O3S+ ClO4 (C, H, N, S). 6.1.3.15. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2,6-diphenylpyridinium perchlorate B15 Yellow-orange crystals, m.p. 204206 C (yield of 40%); IR (KBr), cm1: 625, 705, 765, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 050, 3 260; 1 H-NMR (TFA), ppm: 4.13 (s, 2H, CH2); 6.718.40 (m, 17H, ArH from 1,4-phenylene, 2,6-Ph2 and 3,4,5-H from pyridinium). Anal. C26H22N4O3S+ ClO4 (C, H, N, S). 6.1.3.16. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylmethyl]-2,3,4,6-tetramethylpyridinium perchlorate B16 White-tan crystals, m.p. 253255 C (yield of 65%); IR (KBr), cm1: 625, 800, 1 100, 1 175, 1 290, 1 345, 1 535, 1 580, 1 640, 1 675, 3 030, 3 305; 1H-NMR (TFA), ppm: 2.60 (s, 3H, 4-Me); 2.77 (s, 3H, 3-Me); 2.87 (s, 6H, 2,6-(Me)2); 4.12 (s, 2H, CH2); 7.218.50 (m, AABB, 4H, ArH from 1,4-phenylene); 7.90 (s, 1H, ArH, 5-H from pyridinium). Anal. C18H23N4O3S+ ClO4 (C, H, N, S). 6.1.3.17. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2,4,6-trimethylpyridinium perchlorate C1 White crystals, m.p. 266267 C (yield of 84%); IR (KBr), cm1: 625, 680, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 060, 3 250, 3 330; 1H-NMR (TFA), ppm: 2.66 (s, 3H, 4-Me); 2.88 (s, 6H, 2,6(Me)2); 3.12 (t, 2H, CH2); 4.05 (t, 2H, CH2); 7.478.38 (m, 6H, ArH from 1,4-phenylene and 3,5-H from pyridinium). Anal. C18H23N4O3S+ ClO4 (C, H, N, S). 6.1.3.18. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2-iso-propyl-4,6-dimethylpyridinium perchlorate C2 White crystals, m.p. 244245 C (yield of 83%); IR (KBr), cm1: 625, 685, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 040, 3 255, 3 380; 1H-NMR (TFA), ppm: 1.47 (d, 6H, 2Me from i-Pr); 2.68 (s, 3H, 4-Me); 2.90 (s, 3H, 6-Me); 3.103.75 (m, 3H, CH from i-Pr + CH2); 4.03 (t, 2H, CH2); 7.338.35 (m, 6H, ArH from 1,4-phenylene and 3,5-H from pyridinium). Anal. C20H26N4O3S+ ClO4 (C, H, N, S). 6.1.3.19. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2,6-di-iso-propyl-4-methylpyridinium perchlorate C3 White crystals, m.p. 250251 C (yield of 76%); IR (KBr), cm1: 625, 685, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 040, 3 235, 3 410; 1H-NMR (TFA), ppm: 1.48 (d, 12H, 4Me from 2 i-Pr); 2.70 (s, 3H, 4-Me); 3.153.79 (m, 4H, 2CH from 2 i-Pr + CH2); 4.02 (t, 2H, CH2); 7.338.27 (m, 6H, ArH from 1,4phenylene and 3,5-H from pyridinium). Anal. C22H31N4O3S+ ClO4 (C, H, N, S). 6.1.3.20. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2,6-dimethyl-4-phenylpyridinium perchlorate C4 White crystals, m.p. 221223 C (yield of 72%); IR (KBr), cm1: 625, 690, 780, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 050, 3 280; 1H-NMR (TFA), ppm: 3.08 (s, 6H, 2,6-(Me)2); 3.15 (t, 2H, CH2); 4.03 (t, 2H, CH2); 7.558.37 (m, 11H, ArH from 1,4phenylene, 4-Ph and 3,5-H from pyridinium). Anal. C23H25N4O3S+ ClO4 (C, H, N, S). 6.1.3.21. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2,6-diethyl-4-phenylpyridinium perchlorate C5 White crystals, m.p. 227229 C (yield of 80%); IR (KBr), cm1: 625, 700, 780, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 060, 3 240, 3 335; 1H-NMR (TFA), ppm: 1.67 (t, 6H, 2 Me from Et); 3.153.80 (m, 6H, 2 CH2 from Et + CH2 from ethylene bridge); 4.07 (t, 2H, CH2 from ethylene bridge); 7.578.50 (m, 11H, ArH from 1,4-phenylene, 4-Ph and 3,5-H from pyridinium). Anal. C25H29N4O3S+ ClO4 (C, H, N, S). 6.1.3.22. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2,6-di-n-propyl-4-phenylpyridinium perchlorate C6 White crystals, m.p. 212214 C (yield of 63%); IR (KBr), cm1: 625, 685, 775, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 050, 3 255, 3 335; 1H-NMR (TFA), ppm: 1.23 (t, 6H, 2 Me from Pr); 2.03 (q, 4H, 2 CH2 from Pr); 3.073.75 (m, 6H, 2 CH2 from Pr + CH2 from ethylene bridge); 4.05 (t, 2H, CH2 from ethylene bridge); 7.558.43 (m, 11H, ArH from 1,4-phenylene, 4-Ph and 3,5-H from pyridinium). Anal. C27H33N4O3S+ ClO4 (C, H, N, S). 6.1.3.23. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2,6-di-iso-propyl-4-phenylpyridinium perchlorate C7 White crystals, m.p. 241243 C (yield of 69%); IR (KBr), cm1: 625, 685, 765, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 060, 3 270, 3 350; 1H-NMR

950 (TFA), ppm: 1.60 (d, 12H, 4 Me from i-Pr); 3.103.83 (m, 4H, 2 CH from i-Pr + CH2 from ethylene bridge); 4.13 (t, 2H, CH2 from ethylene bridge); 7.478.43 (m, 11H, ArH from 1,4-phenylene, 4-Ph and 3,5-H from pyridinium). Anal. C27H33N4O3S+ ClO4 (C, H, N, S). 6.1.3.24. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2-methyl-4,6-diphenylpyridinium perchlorate C8 White crystals, m.p. 233234 C (yield of 77%); IR (KBr), cm1: 625, 675, 775, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 050, 3 245, 3 435; 1H-NMR (TFA), ppm: 3.033.39 (m, 5H, 2-Me + CH2 from ethylene bridge); 4.06 (t, 2H, CH2 from ethylene bridge); 7.058.45 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C28H27N4O3S+ ClO4 (C, H, N, S). 6.1.3.25. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2-ethyl-4,6-diphenylpyridinium perchlorate C9 White crystals, m.p. 229230 C (yield of 54%); IR (KBr), cm1: 625, 685, 750, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 050, 3 220, 3 390; 1H-NMR (TFA), ppm: 1.72 (t, 3H, Me from Et); 2.903.78 (m, 4H, CH2 from Et + CH2 from ethylene bridge); 4.08 (t, 2H, CH2 from ethylene bridge); 6.888.47 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C29H29N4O3S+ ClO4 (C, H, N, S). 6.1.3.26. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2-n-propyl-4,6-diphenylpyridinium perchlorate C10 White crystals, m.p. 235236 C (yield of 59%); IR (KBr), cm1: 625, 705, 775, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 080, 3 255, 3 340; 1H-NMR (TFA), ppm: 1.32 (t, 3H, Me from Pr); 2.17 (sextet, 2H, CH2 from n-Pr); 2.823.66 (m, 4H, CH2 from n-Pr + CH2 from ethylene bridge); 4.09 (t, 2H, CH2 from ethylene bridge); 6.838.43 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C30H31N4O3S+ ClO4 (C, H, N, S). 6.1.3.27. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2-iso-propyl-4,6-diphenylpyridinium perchlorate C11 White crystals, m.p. 234235 C (yield of 73%); IR (KBr), cm1: 625, 700, 765, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 070, 3 250, 3 350; 1H-NMR (TFA), ppm: 1.70 (d, 6H, 2 Me from i-Pr); 3.15 (t, 2H, CH2 from ethylenic bridge); 3.504.03 (m, 1H, CH from i-Pr); 4.11 (t, 2H, CH2 from ethylenic bridge); 6.958.53 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C30H31N4O3S+ ClO4 (C, H, N, S). 6.1.3.28. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2-n-butyl-4,6-diphenylpyridinium perchlorate C12 White crystals, m.p. 207208 C (yield of 78%); IR (KBr), cm1: 625, 685, 7650, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 080, 3 255, 3 330; 1H-NMR (TFA), ppm: 1.15 (t, 3H, Me from n-Bu); 1.382.45 (m, 4H, 2 CH2 from n-Bu); 3.003.68 (m, 4H, CH2 from n-Bu + CH2 from ethylenic bridge); 4.10 (t, 2H, CH2 from ethylenic bridge); 7.028.43 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C31H33N4O3S+ ClO4 (C, H, N, S). 6.1.3.29. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2-tert-butyl-4,6-diphenylpyridinium perchlorate C13 White crystals, m.p. 220222 C (yield of 69%); IR (KBr), cm1: 625, 700, 765, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 060, 3 250, 3 370; 1H-NMR (TFA), ppm: 1.92 (s, 9H, t-Bu); 3.14 (t, 2H, CH2); 4.10 (t, 2H, CH2 from ethylene bridge); 6.908.77 (m, 16H, ArH from 1,4-phenylene, 4,6-Ph2 and 3,5-H from pyridinium). Anal. C31H33N4O3S+ ClO4 (C, H, N, S). 6.1.3.30. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2,4,6-triphenylpyridinium perchlorate C14 Yellow crystals, m.p. 213214 C (yield of 82%); IR (KBr), cm1: 625, 680, 770, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 050, 3 260, 3 335; 1H-NMR (TFA), ppm: 3.12 (t, 2H, CH2 from ethylene bridge); 4.05 (t, 2H, CH2 from ethylene bridge); 6.578.40 (m, 21H, ArH from 1,4-phenylene, 2,4,6-Ph3 and 3,5-H from pyridinium). Anal. C33H29N4O3S+ ClO4 (C, H, N, S). 6.1.3.31. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2,6-diphenylpyridinium perchlorate C15 Yellow crystals, m.p. 212214 C (yield of 16%); IR (KBr), cm1: 625, 700, 760, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 050, 3 240, 3 325; 1H-NMR (TFA), ppm: 3.07 (t, 2H, CH2); 4.13 (t, 2H, CH2 from ethylene bridge); 6.558.50 (m, 17H, ArH from 1,4phenylene, 2,6-Ph2 and 3,4,5-H from pyridinium). Anal. C27H24N4O3S+ ClO4 (C, H, N, S). 6.1.3.32. 1-N-[(4-Guanidinosulfonyl-phenyl)aminocarbonylethyl]-2,3,4,6-tetramethylpyridinium perchlorate C16 White crystals, m.p. 210211 C (yield of 51%); IR (KBr), cm1: 625, 680, 1 100, 1 175, 1 285, 1 345, 1 540, 1 580, 1 645, 1 675, 3 030, 3 245, 3 325; 1H-NMR

951 (TFA), ppm: 2.52 (s, 3H, 3-Me); 2.62 (s, 3H, 4-Me); 2.83 (s, 3H, 6-Me); 2.92 (s, 3H, 2-Me); 3.13 (t, 2H, CH2); 4.07 (t, 2H, CH2); 7.618.55 (m, 5H, ArH from 1,4phenylene + 5-H from pyridinium). Anal. C19H25N4O3S+ ClO4 (C, H, N, S). 6.2. Pharmacology 6.2.1. Enzyme assays: KI determinations Human thrombin and human trypsin were purchased from Sigma Chemical Co. (St. Louis, MO, USA); their concentrations were determined from the absorbance at 280 nm and the extinction coefficients furnished by the supplier. The activity of such preparations was in the range of 2 5003 000 NIH units/mg. The potency of standard and newly obtained inhibitors was determined from the inhibition of the enzymatic (amidolytic) activity of these serine proteases, at 21 C, using Ts-Gly-Pro-ArgpNA (Chromozym TH) from Sigma as substrate, by the method of Lottenberg et al. [23]. The substrate was reconstituted as a 4 mM stock in ultrapure water and brought to pH 4 with hydrochloric acid. Substrate concentrations were determined from absorbance at the isosbestic wavelength for the peptide-p-nitroanilide/pnitroaniline mixtures. Extinction coefficients of 8 270 L.mol1.cm1 in the used buffer (0.01 M Hepes, 0.01 M Tris, 0.1 M NaCl, 0.1% polyethylene glycol 6000; pH 7.80) were employed. The rate of p-nitroanilide hydrolysis was determined from the change in absorbance at 405 nm using an extinction coefficient for p-nitroaniline of 9 920 L.mol1.cm1 for the abovementioned reaction buffer. Measurements were made using a Cary 3 spectrophotometer interfaced with a PC. Initial velocities were thus estimated using the direct linear plot-based procedure as reported by Lottenberg et al. [23]. KIs were then determined according to Dixon, using a linear regression program [26]. The KI values determined are the means of at least three determinations. 6.2.2. pKa determination The half neutralization point was measured by titrating the organic acids/bases with 0.05 N NaOH and 0.05 N HCl in EtOH/water (30%, v/v), using a glass electrode, as described by Bell and Roblin [35] for the structurallyrelated antibacterial sulfonamides. Acknowledgements This research was nanced in part by the EU grant ERB CIPDCT 940051. Thanks are addressed to Dr M.A. Ilies for expert technical assistance. References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Shafer J.A., Curr. Opin. Chem. Biol. 2 (1998) 458465. Strzebecher J., Meier J., J. Enzyme Inhib. 9 (1995) 12. Strzebecher J., Prasa D., Hauptmann J., Vieweg H., Wikstrom P., J. Med. Chem. 40 (1997) 30913099. Pavone V., De Simone G., Nastri F., Galdiero S., Staiano N., Lombardi A., Pedone C., Biol. Chem. 379 (1998) 9871006. Engh R.A., Brandstetter H., Sucher G., Eichinger A., Baumann U., Bode W. et al., Structure 4 (1996) 13531362. Babine R.E., Bender S.L., Chem. Rev. 97 (1997) 13591472. Salemme F.R., Spurlino J., Bone R., Structure 5 (1997) 319324. Eriksson U.G., Renberg L., Bredberg U., Teger-Nilsson A.C., Regardh C.G., Biopharm. Drug Dispos. 19 (1998) 5564. Bajusz S., Szell E., Bagdy D., Barbas E., Horvath G., Dioszegi M. et al., J. Med. Chem. 33 (1990) 17291735. Krishnan R., Zhang E., Hakansson K., Arni R.K., Tulinsky A., Lim-Wilby M.S. et al., Biochemistry 37 (1998) 1209413103. Claeson G., Philipp M., Agner E., Scully M.F., Metternich R., Kakkar V.V., Desoyza T., Niu L.H., Biochem. J. 290 (1993) 309312. Strzebecher J., Prasa D., Bretschneider E., Bode W., Bauer M., Brandstetter H., Wikstrom P., Vieweg H., New developments in the eld of benzamidine-derived thrombin inhibitors, in: MullerBerghaus G., Madlener K., Blomback M., ten Cate J.W. (Eds.), DIC Pathogenesis, Diagnosis, and Therapy of Disseminated Intravascular Fibrin Formation, Excerpta Medica, Amsterdam, London, New York, Tokyo, 1993, pp. 183190. Lumma W.C., Witherup K.M., Tucker T.J., Brady S.F., Sisko J.T., Naylor-Olsen A.M. et al., J. Med. Chem. 41 (1998) 10111013. Semple J.E., Rowley D.C., Brunck T.K., Ha-Uong T., Minami N.K., Owens T.D. et al., J. Med. Chem. 39 (1996) 45314536. Sixma J.J., de Groot P.G., Thromb. Res. 68 (1992) 507512. Okamoto S., Kinjo K., Hijikata A., Kikumoto R., Tamao Y., Ohkubo K., Tonomura S., J. Med. Chem. 23 (1980) 827830. Strzebecher J., Markwardt F., Voigt B., Wagner G., Walsmann P., Thromb. Res. 29 (1983) 635642. Groutas W.C., Kuang R., Venkataraman R., Epp J.B., Ruan S., Prakash O., Biochemistry 36 (1997) 47394750. Supuran C.T., Manole G., Dinculescu A., Schiketanz A., Gheorghiu M.D., Puscas I., Balaban A.T., J. Pharm. Sci. 81 (1992) 716719. Supuran C.T., Scozzafava A., Ilies M.A., Iorga B., Cristea T., Chiraleu F., Banciu M.D., Eur. J. Med. Chem. 33 (1998) 577594. Anderson G.W., Zimmerman J.E., Callahan F.M., J. Am. Chem. Soc. 85 (1963) 3039. Sheehan J.C., Ledis S.L., J. Am. Chem. Soc. 95 (1973) 875879. Lottenberg R., Christensen U., Jackson C.M., Coleman P.L., Meth. Enzymol. 80 (1981) 341361. Banner D.W., Hadvary P., J. Biol. Chem. 266 (1991) 2008520093. Supuran C.T., Scozzafava A., Briganti F., Clare B.W., J. Med. Chem. in press. Hemker H.C., Willems G.M., Beguin S.A., Thromb. Haemostas. 56 (1986) 917. Balaban A.T., Toma C., Tetrahedron Suppl. 7 (1966) 17. Dinculescu A., Balaban A.T., Rev. Roum. Chim. 25 (1980) 15051528. Supuran C.T., Banciu M.D., Balaban A.T., Rev. Roum. Chim. 38 (1993) 199205. Balaban A.T., Dinculescu A., Dorofeenko G.N., Fischer G.W., Koblik A.V., Mezheritskii V.V., Schroth W., Pyrylium salts: synthe-

[12]

[13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30]

952
ses, reactions and physical properties, in: Katritzky A.R. (Ed.), Advances in Heterocyclic Chemistry, Academic Press, New York, 1982, pp. 8360. Supuran C.T., Pop E., Dinculescu A., Heterocycles 37 (1994) 667671. [32] [33] [34] [35] Neidlein R., Witerzens P., Monats. Chem. 106 (1975) 643648. Bode W., Turk D., Karshikov A., Protein Sci. 1 (1992) 426471. Stubbs M.T., Bode W., Thromb. Res. 69 (1993) 158. Bell P.H., Roblin R.O., J. Am. Chem. Soc. 64 (1942) 29052917.

[31]

Eur. J. Med. Chem. 34 (1999) 953966 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

953

Original article

Synthesis and antispasmodic activity of analogues of natural pterosins

Helen Sheridana*, Neil Frankishb, Ronan Farrella
a b

Department of Pharmacognosy, School of Pharmacy, Trinity College Dublin 2, Ireland Department of Pharmacology, School of Pharmacy, Trinity College Dublin 2, Ireland (Received 22 February 1999; accepted 31 May 1999)

Abstract The synthesis of an extensive range of analogues of natural pterosins using modied Heck coupling is reported. The smooth muscle relaxant activity of these compounds has been examimed. Several compounds with signicant smooth muscle relaxant activity have been identied. 1999 ditions scientiques et mdicales Elsevier SAS pterosins / indanones / indanes / smooth muscle relaxant activity

1. Introduction As part of an ongoing study into the synthesis and smooth muscle relaxant activity of sesquiterpene indanones related to the pterosin family of fern metabolites [1], we have developed an improved synthetic route to pterosin Z (1) [2], a metabolite of Pityrogramma and Pteris species [3, 4]. More recently we have shown that (1) exhibits potent smooth muscle relaxant activity [5]. We now report on the synthesis and activity of a number of indane and indanone analogues of pterosins in which a variety of three carbon side chains replace the hydroxyethyl side chain of the natural pterosins. 2. Chemistry Palladium catalysed Heck coupling [6] was used in this study for attaching a three carbon side chain to a range of indanone nucleii to yield a series of pterosin analogues. Methylacrylate (2) was identied as a suitable three carbon unit for coupling with bromoindanes (38), key intermediates used in earlier pterosin synthetic studies [1, 2]. The regioselectivity of the site of arylation on the double bond of methylacrylate has been reported to occur exclusively at the terminal alkene position giving rise to
*Correspondence and reprints

a single structural isomer [6, 7]. Coupling of 2 with 38 was catalysed by palladium acetate in the presence of triphenylphosphine and triethylamine and was carried out at 100 C in sealed ampoules (Expt. l). The lowest yields in the coupling reaction were consistently observed for the products 10, 13 and 14 while the highest yields were observed for the isopterosin type structure 12. The products of the initial coupling reactions 914 (table I) were characterised spectroscopically and in all cases were single geometric isomers. The 1H-NMR spectrum of 10, typical for the series, shows the presence of two sets of doublets downeld at 6.2 and 7.8 ppm, characteristic of the methylpropenoate side chain. The coupling constant of the doublets J (16 Hz) is characteristic of trans coupled protons and is observed for all compounds in the series. It is therefore conrmed that in the Heck coupling between methylacrylate and bromoindanes only the E isomer is formed. Compounds 914 were subjected to a range of chemical reactions in which the methyl propenoate side chain was modied (gure 1) to give rise to a series of pterosin analogues 1539 for which the change in smooth muscle relaxant activity could be measured relative to the change in structure. The structure of these compounds was established by spectroscopic means and the smooth muscle relaxant activity of 939 was measured using established methods.

954
Table I. Yield and phsical data for compounds 914. Compound 9 Yield % 60 Melting point 8687 C M+ 244.1100

10

45

9394 C

272.1407

11

61

124125 C

244.1100

12

73

oil

272.1413

13

45

4748 C

230.1307

14

41

oil

258.1620

3. Pharmacological results and discussion Addition of calcium (2.5 mM) to guinea-pig ileum bathed in high potassium (45 mM), calcium-free Krebs solution caused a contracture of the tissue which was sustained for a period greater than 40 min. The compounds (added cumulatively) caused a dose-dependant inhibition of calcium (2.5 mM) contractures of guineapig ileum (for examples see gure 2). The more potent compounds (12, 16, 20, 28 and 29), with EC50 values

ranging from 1.3 106 to 1.1 105 M, were more than 3 orders of magnitude less potent than the positive control, the calcium antagonist, nifedipine (gure 2) which had an EC50 value of 9.3 0.5 109 M. Comparison of the activities of the compounds was made at a single concentration of 105 M, with inhibitory activity ranging from 2075% (gures 35). In comparison, nifedipine, at a concentration of 108 M, inhibited calcium contractures by 48.1 2.3%. The most potent smooth muscle relaxant (gure 2) is the diol 39. The

955

Figure 1. Example of reactions carried out to modify the methylpropenoate side chain of pterosins.

activity of this compound is statistically greater (P < 0.05) than that of its close analogue 29. These two compounds (29 and 39) are structurally quite similar, having the three carbon side chain at the C-4 position and the aromatic methyl groups at the 3,5-positions. The next level of activity is shown by compounds 12, 20 and 28. These compounds are statistically (P < 0.05) less active than 29 and 39 but more active than 16 (P < 0.05). No clear structure activity relationship is apparent from the results of the 30 compounds analysed in this study, although it does appear that an unsaturated three carbon side chain at the C-4 position coupled with aromatic methyls at the 3,5- positions with oxygenation at C-1 (isopterosin form) enhances activity relative to the positioning of the carbon side chain at the C-6 position combined with the aromatic methyl substitution at the 5,7- positions (pterosin form). The smooth muscle relaxant activity (gure 2) of the most active compound 39 in this study (EC50 4.9 0.6 106 M) is signicantly lower (P < 0.05) than the activity we have recently reported for pterosin Z (1) (EC50 1.3 0.1 106 M) [5] and the

fungal indane 40 (EC50 2.9 1.6 106 M) [1]. However, the structural modications observed in this group of pterosin analogues has led to an increase in activity relative to the rst reported naturally occurring smooth muscle relaxants, onitin (EC50 1 104 M), onotisin (EC50 2 103 M) and otninoside (EC50 7 104M) [8, 9]. Pterosins have been shown to inhibit contractile responses of guinea-pig ileum by both histamine and acetylcholine [8], together with barium and potassium [9], suggesting a mechanism of action involving interference with calcium handling in the smooth muscle cell. This conclusion is supported by the results of this study, showing that the compounds inhibit calcium contractures of potassium-depolarised smooth muscle. However it has not been determined at this time whether such interference with calcium handling involves inhibition of extracellular calcium inux through membrane channels or interference with the calcium/calmodulin cascade of reactions within the cell.

956

Figure 1. (Continued).

4. Experimental protocols Melting points were determined on a Me-Opta hot stage and are uncorrected. Infra red spectra were recorded on a Nicolet 205 FT-IR. Ultraviolet spectra were recorded

on a Varian Carey 3E UV-visible spectrophotometer. Mass spectra were determined at 70 eV on an AEI MS 30 instrument. 1H-NMR spectra were recorded on a Bruker MSL 300 instrument at 300 MHz. 13C-NMR were recorded at 75.47 MHz. Deuteriochloroform was used as

957

Figure 1. (Continued).

solvent with SiMe4 as internal standard. TLCs were run on commercially pre-coated plates (Merck, Kieselgel

60F254). Merck Kieselgel 60 (9385) was used for column chromatography.

958

Figure 2. Effect of compounds 3 1083 105 M and Nifedipine, 1 10101 107 M (added cumulatively) on inhibition of calcium (2.5 mM) contractions of guinea-pig ileum suspended in high-potassium, calcium-free modied Krebs solution. Values are expressed as means SEM, n = 6.

4.1. Synthesis of bromoindanones 38. As outlined in reference [1] 4.1.1. General procedure for the coupling on methylacrylate (2) with bromoindanones (38) to synthesise products 914 A solution of indanone (2 mmol), palladium acetate (0.11 mmol), triphenyl phosphine (0.21 mmol) and methylacrylate (2) (5 mmol) in triethylamine (10 mL) was sealed in a glass ampoule (25 mL). The ampoule was shaken until the mixture was homogeneous and was then heated under pressure 15 lbs/in2 at 100 C for 48 h. After cooling, the ampoule was broken and the reaction mix-

ture was ltered and stirred on ice/HCl and extracted into EtOAc. The organic layer was washed, dried (Na2SO4) and evaporated under vacuum. The residue was puried by column chromatography on silica gel (eluant: pet. ether:ethyl acetate, 9:1). 4.1.1.1. E-2,3-Dihydro-5,7-dimethyl-1H-inden-1-one-6propenoic acid methyl ester (9) Prepared by coupling 6-bromo-5,7-dimethyl-1Hindan-1-one [1] with methacrylate. White solid (59.5%), m.p. 8687 C (EtOH). Found M+ 244.1109 (C15H16O3 requires 244.1099). max (KBr) 2 361, 1 707, 1 686 cm1; H (CDCl3): 2.41 (3H, s, CH3), 2.62 (2H, t J 6.0 Hz,

959

Figure 3. Effect of several compounds (1 105 M) and Nifedipine 1 108 M, on inhibition of calcium (2.5 mM) contractions of guinea-pig ileum suspended in high-potassium, calcium free modied Krebs solution. Values are expressed as means SEM, n = 6.

CH2), 2.65 (3H, s, CH3), 3.08 (2H, t J 6.0 Hz, CH2), 3.81 (3H, s, OCH3), 6.23 (1H, d J 16.4 Hz), 6.96 (1H, s, Ar-H), 7.84 (1H, d J 16.4 Hz); C 18.1(CH3), 21.0 (CH3), 25.94 (CH2), 36.72 (CH2), 51.72 (OCH3), 122.1 (CH), 128.3 (CH), 129.0 (ArC), 133.1 (ArC), 139.6 (ArC), 140.2 (ArCH), 144.0 (ArC), 155.3 (ArC), 167.2 (CO), 207.0 (CO); m/z 244 (M+, 100), 213 (59), 212 (66), 185 (58), 184 (40), 171 (43), 157 (32), 142 (66), 141 (44), 128 (30), 115 (25), 77 (19). 4.1.1.2. E-2,3-Dihydro-2,2,5,7-tetramethyl-1H-inden-1one-6-propenoic acid methyl ester (10) Prepared by coupling 6-bromo-2,3-dihydro-2,2,5,7tetramethyl-1H-inden-1-one [1] with methacrylate. White crystals (46%), m.p. 9394 C (EtOH). Found M+ 272.1416 (C17H20O3 requires 272.1413); max (KBr) 2 951, 1 712, 1 698 cm1; H (CDCl3) 1.19 (6H, s, 2CH3), 2.45 (3H, s, CH3), 2.59 (3H, s, CH3), 2.99 (2H, s, CH2), 3.81 (3H, s, OCH3), 6.23 (1H, d J 16.5 Hz), 7.01 (1H, s, Ar-H), 7.86 (1H, d J 16.4 Hz); C 18.2 (CH3), 21.0 (CH3), 25.3 (2CH3), 43.1 (CH2), 45.4 (C), 51.7(OCH3), 122.0 (CH), 128.9 (ArC), 131.3(ArC), 132.5 (ArCH), 140.4 (ArCH), 141.2 (ArC), 144.0 (ArC), 152.3 (ArC), 167.2 (CO), 211.3 (CO); m/z 272 (M+, 100), 257 (68) 211 (21), 213 (32), 195 (28), 170 (17), 128 (14).

4.1.1.3. E-2,3-Dihydro-4,6-dimethyl-1H-inden-1-one-5propenoic acid methyl ester (11) Prepared by coupling 5-bromo-2,3-dihydro-4,6dimethyl-1H-inden-1-one [1] with methacrylate. White crystals (61%), m.p. 124125 C (EtOH). Found M+ 244.1096 (C15H16O3 requires 244.1099); max (KBr) 2 955, 1 721, 1 704, 1 647cm1; H (CDCl3): 2.15 (3H, s, CH3), 2.18 (3H, s, CH3), 2.50 (2H, t J 5.5 Hz, CH2), 2.83 (2H, t J 5.5 Hz, CH2), 3.69 (3H, s, OCH3), 5.91 (1H, d J 16.4 Hz), 7.31 (1H, s, ArH), 7.64 (1H, d J 16.4 Hz); C 15.7 (CH3), 20.6 (CH3), 24.5 (CH2), 35.9 (CH2), 59.8 (OCH3), 121.7 (CH), 124.8 (CH), 133.0 (ArC), 135.6 (ArC), 135.8 (ArC), 140.0 (ArC), 142.4 (ArCH), 151.9 (ArC), 166.0 (CO), 206.3 (CO); m/z 244 (M+, 40), 229 (48), 213 (49), 212 (42), 184 (44), 171 (61), 143 (39), 142 (100), 141 (37), 128 (48), 115 (37). 4.1.1.4. E-2,3-Dihydro-2,2,4,6-tetramethyl-1H-inden-1one-5-propenoic acid methyl ester (12) Prepared by coupling 5-bromo-2,3-dihydro-2,2,4,6tetramethyl-1H-indan-1-one [1] with methacrylate. Pale oil (72%). Found M+ 272.1408 (C17H20O3 requires 272.1413) max (Film) 2 958, 1 714, 1 716, 1 644 cm1; H (CDCl3) 1.11 (6H, s, 2CH3), 2.18 (3H, s, CH3), 2.22 (3H, s, CH3), 2.77 (2H, s, CH2), 3.72 (3H, s, OCH3), 5.96

960

Figure 4. Effect of several compounds (1 105 M) and Nifedipine 1 108 M, on inhibition of calcium (2.5 mM) contractions of guinea-pig ileum suspended in high-potassium, calcium free modied Krebs solution. Values are expressed as means SEM, n = 6.

(1H, d J 16.5 Hz), 7.32 (1H, s, ArH), 7.70 (1H, d J 16.5 Hz); C 15.6 (CH3), 20.6 (CH3), 24.9 (2CH3), 41.7 (CH2), 45.1 (C), 51.4 (OCH3), 122.5 (CH), 124.7 (CH), 133.3 (ArC), 134.1 (ArC), 135.8 (ArC), 140.3 (ArC), 142.6 (ArCH), 149.0 (ArC), 166.1 (CO), 210.9 (CO); m/z 272 (M+, 45), 258 (19), 257 (100), 241 (28), 240 (23), 225 (19), 213 (13), 197 (17), 153 (10), 128 (14), 77 (5). 4.1.1.5. E-2,3-Dihydro-2,4-dimethyl-1H-inden-3-propenoic acid methyl ester (13) Prepared by coupling 3-bromo-2,3-dihydro-2,4dimethyl-1H-inden [1] with methacrylate. White solid (45%), m.p. 4748 C (EtOH). Found M+ 230.1300 (C15H18O2 requires 230.1307); max KBr 2 951, 1 721, 1 633, 1 161 cm1; H (CDCl3) 2.09 (2H, m J 7.5 Hz, CH2), 2.29 (3H, s, CH3), 2.34 (3H, s, CH3), 2.86 (2H, t J 7.5 Hz, CH2), 2.93 (2H, J 7.5 Hz, CH2), 3.85 (3H, s, OCH3), 6.06 (1H, d J 16.2 Hz, CH), 6.99 (1H, s, ArH), 7.90 (1H, d J 16.2 Hz); C 17.4 (CH3), 21.1 (CH3), 24.6 (CH2), 31.8 (CH2), 33.1 (CH2), 51.5 (OCH3), 123.0 (CH), 123.9 (CH), 131.6 (ArC), 132.1 (ArC), 134.8

(ArC), 141.4 (ArC), 144.2 (ArCH), 144.4 (ArC), 167.2 (CO); m/z 230 (M+, 67), 215 (39), 199 (100), 170 (67), 156 (19), 155 (57), 141 (15), 128 (23).

4.1.1.6. E-2,3-Dihydro-2,2,5,7-tetramethyl-1H-inden-6propenoic acid methyl ester (14) Prepared by coupling 6-bromo-2,3-dihydro-2,2,5,7dimethyl-1H-inden [1] with methacrylate. Pale oil (41%). Found M+ 258.1610 (C17H22O2 requires 258.1602). max (Film) 2 952, 1 723, 1 634 cm1; H (CDCl3): 1.16 (6H, s, 2CH3), 2.21 (3H, s, CH3), 2.32 (3H, s, CH3), 2.65 (2H, s, CH2), 2.73 (2H, s, CH2), 3.82 (3H, s, OCH3), 6.04 (1H, d J 16.2 Hz, CH), 6.87 (1H, s, ArH), 7.90 (1H, t J 16.2 Hz, CH); C 17.3 (CH3), 20.3 (CH3), 29.1 (2CH3), 39.2 (C), 45.8 (CH2), 47.9 (CH2), 51.5 (OCH3), 124.2 (CH), 129.8 (CH), 131.6 (ArC), 134.8 (ArC), 135.4 (ArC), 140.6 (ArC), 142.8 (ArCH), 143.7 (ArC), 167.3 (CO); m/z 258 (M+, 96), 243 (39), 228 (18), 227 (100), 212 (22), 211 (61), 199 (40), 197 (17), 171 (11), 157 (13), 153 (20), 143 (19), 128 (17).

961

Figure 5. Effect of several compounds (1 105 M) and Nifedipine 1 108 M, on inhibition of calcium (2.5 mM) contractions of guinea-pig ileum suspended in high-potassium, calcium free modied Krebs solution. Values are expressed as means SEM, n = 6.

4.1.2. General procedure for the preparation of propenoic acids (15, 19, 23, 27, 31 and 35) A solution of propenoic acid methyl ester (2.25 mmol) in HCl (5 M, 15mL) and THF (15 mL) was reuxed for 6 h. After this time the THF was evaporated in vacuo. The aqueous residue was then diluted with H2O (50 mL) and was extracted with EtOAc (2 30 mL). The organic layer was washed with 1 M NaOH (2 20 mL) and the aqueous layer retained. Following acidication with HCl the aqueous layer was extracted with ethyl acetate, washed with water, dried over anhydrous sodium sulphate and evaporated off under vacuum to give a crude residue which was puried by recrystallisation to yield the propenoic acids. 4.1.2.1. E-2,3-Dihydro-5,7-dimethyl-1H-inden-1-one-6propenoic acid (15) Prepared from 9. White crystalline powder, m.p. 219220 C. Found M+ 230.0953 (C14H16O3 requires 230.0943). max (KBr) 3 5003 200b, 1 725, 1 696 cm1;

H (DMSO-D6), 2.41 (3H, s, CH3), 2.51 (3H, s, CH3), 2.59 (2H, t J 6 Hz, CH2), 3.06 (2H, t J 6 Hz, CH2), 6.27 (1H, d J 16 Hz, CH), 7.10 (1H, s, CH), 7.72 (1H, d J 16 Hz, CH); C 17.7 (CH3), 20.5 (CH3), 25.5 (CH2), 36.5 (CH2), 123.6 (CH), 129.0 (C), 132.0 (C), 138.2 (CH), 139.1 (C), 140.2 (CH), 143.6 (C), 155.4 (C), 167.4 (CO), 206.5 (CO). m/z 230(M+, 96), 215 (28), 212 (41), 186 (18), 185 (82), 184 (29), 171 (37), 157 (68), 156 (19), 143 (37), 128 (41). 4.1.2.2. E-2,3-Dihydro-2,2,5,7-tetramethyl-1H-inden-1one-6-propenoic acid (19) Prepared from 10. Pale yellow needles (EtOH), m.p. 173174 C, yield 76%. Found M+ 258.1245 (C16H18O3 requires 258.1256). max (KBr) 3 4003 200b, 1 702, 1 679 cm1; H (CDCl3) 1.19 (6H, s, 2CH3), 2.45 (3H, s, CH3), 2.59 (3H, s, CH3), 3.01 (2H, s, CH2), 6.23 (1H, d J 16 Hz, CH), 7.01 (1H, s, CH), 7.93 (1H, d J 16 Hz, CH); C 18.2 (CH3), 21.0 (CH3), 25.3 (2CH3), 43.1 (CH2), 45.4 (C), 121.9 (CH), 128.6 (C), 131.3 (C), 132.6

962 (CH), 139.1 (C), 140.7 (C), 141.8 (CH), 144.3 (C), 152.6 (C), 170.7 (CO), 211.6 (CO); MS(EI) m/z 258 (M+, 82), 243 (91), 240 (12), 213 (38), 195 (31), 170 (15), 128 (16), 115 (10), 73 (13). 4.1.2.3. E-2,3-Dihydro-4,6-dimethyl-1H-inden-1-one-5propenoic acid (23) Prepared from 11. White powder (EtOH), m.p. 225226 C yield 95%. Found M+ 230.0934 (C14H14O3 requires 230.0943). max (KBr) 3 5003 000b, 1 722, 1 711 cm1; H (DMSO d6) 2.28 (3H, s, CH3), 2.32 (3H, s, CH3), 2.63 (2H, t J 6 Hz, CH2), 2.98 (2H, t J 6 Hz, CH2), 6.09 (1H, d J 16 Hz, CH), 7.37 (1H, s, CH), 7.71 (1H, d J 16 Hz, CH), 12.73 (1H, s, COOH); C 15.6 (CH3), 20.6 (CH3), 24.5 (CH2), 36.0 (CH2), 121.3 (CH), 126.4 (CH), 133.8 (C), 135.6 (C), 135.7 (C), 140.2 (C), 141.8 (CH), 152.4 (C), 166.9 (CO), 206.3 (CO); m/z 230 (M+, 100), 215 (61), 212 (44), 184 (32), 171 (29), 143 (62), 142 (100), 141 (35), 128 (45), 91 (73), 77 (23). 4.1.2.4. E-2,3-Dihydro-2,2,4,6-tetramethyl-1H-inden-1one-5-propenoic acid (27) Prepared from 12. White solid (EtOH), m.p. 181182 C yield 86%. Found M+ 258.1227 (C16H18O3 requires 258.1256). max (KBr) 3 4003 200b, 1 714, 1 702 cm1; H (CDCl3) 1.08 (6H, s, 2CH3), 2.16 (3H, s, CH3), 2.21 (3H, s, CH3), 2.75 (2H, s, CH2), 5.94 (1H, d J 16 Hz, CH), 7.30 (1H, s, CH), 7.70 (1H, d J 16 Hz, CH); C (CDCl3) 15.7 (CH3), 20.7 (CH3), 25.0 (2CH3), 41.9 (CH2), 45.4 (C), 122.7 (CH), 125.3 (CH), 133.6 (C), 134.1 (C), 136.1 (C), 140.7 (C), 143.5 (CH), 149.4 (C), 168.5 (CO), 211.9 (CO); m/z 258(M+, 100), 244 (28), 243 (100), 240 (28), 153 (20), 128 (35), 115 (23), 77 (12). 4.1.2.5. E-2,3-Dihydro-5,7-dimethyl-1H-indene-6-propenoic acid (31) Prepared from 13. Pale oil, yield 64%. Found M+ 216.1387 (C14H16O2 requires 216.1146); max (Film) 3 2002 900, 1 693, 1 428 cm1; H (CDCl3) 2.05 (2H, m, CH2), 2.25 (3H, s, CH3), 2.31 (3H, s, CH3), 2.83 (2H, t J 7.5 Hz, CH2), 2.86 (2H, t J 7.5 Hz, CH2), 6.01 (1H, d J 16 Hz, CH), 6.95 (1H, s, CH), 7.89 (1H, d J 16 Hz, CH); C 17.4 (CH3), 21.1 (CH3), 24.5 (CH2), 31.7 (CH2), 33.1 (CH2), 116.4 (CH), 123.8 (CH), 131.5 (C), 132.2 (C), 134.9 (C), 141.4 (C), 144.4 (CH), 145.2 (C), 169.0 (CO); m/z 216 (M+, 100) 198 (29), 183 (42), 143 (44), 91 (10), 77 (17). 4.1.2.6. E-2,3-Dihydro-2,2,4,6-tetramethyl-1H-indene-5propenoic acid (35) Prepared from 14. White solid (EtOH), m.p. 138139 C, yield 83%. Found M+ 244.1459 (C16H18O2 requires 244.1436). max (KBr) 3 000b, 1 689 cm1; H (CDCl3) 1.17 (6H, s, 2CH3), 2.26 (3H, s, CH3), 2.35 (3H, s, CH3), 2.68 (2H, s, CH2), 2.73 (2H, s, CH2), 6.07 (1H, d J 16 Hz, CH), 6.92 (1H, s, CH), 8.01 (1H, d J 16 Hz, CH); C 19.1 (CH3), 20.3 (CH3), 29.1 (2CH3), 39.4 (C), 45.7 (CH2), 48.9 (CH2) 124.4 (CH), 127.4 (C), 129.9 (CH), 132.6 (C), 136.0 (C), 140.8 (C), 143.2 (C), 144.9 (CH), 168.5 (CO); m/z 244 (M+, 100), 229 (51), 226 (22), 199 (23), 185 (22), 183 (42), 171 (24), 157 (20), 143 (23), 128 (25), 91 (15), 77 (15). 4.1.3. General procedure for the preparation of propanoic acid methyl esters (16, 20, 24, 28, 32 and 36) A solution of the corresponding propenoic acid methyl ester (2 mmol) in EtOH was stirred with H2 under Wilkinsons catalyst [RhCl(PPh3)3] at room temperature for 24 h. On completion, the mixture was ltered and the EtOH was removed under vacuum. The residue was puried by column chromatography (eluant pet. ether:ethyl acetate, 4:1) to yield the following: 4.1.3.1. 2,3-Dihydro-5,7-dimethyl-1H-inden-1-one-6-propanoic acid methyl ester (16) Prepared by reduction of 15. White solid (n-Hexane), m.p. 6667 C yield 86%. Found M+ 246.1262 (C15H18O3 requires 246.1256); max (KBr) 1 738, 1 703, 1 438 cm1; H (CDCl3) 2.33 (3H, s, CH3), 2.45 (2H, t J 8 Hz, CH2), 2.51 (3H, s, CH3), 2.61 (2H, t J 6 Hz, CH2), 2.95 (2H, t J 8 Hz, CH2), 2.97 (2H, t J 6 Hz, CH2), 3.66 (3H, s, OCH3), 6.89 (1H, s, CH); C 17.7 (CH3), 19.2 (CH3), 23.8 (CH2), 24.0 (CH2), 33.0 (CH2), 36.7 (CH2), 51.6 (OCH3), 131.8 (CH), 132.6 (C), 133.1 (C), 136.4 (C), 142.5 (C), 154.7 (C), 173.0 (CO), 210 (CO); m/z 246 (M+, 88), 214 (74), 186 (19), 173 (100), 159 (43), 144 (40), 129 (35), 115 (24), 91 (17). 4.1.3.2. 2,3-Dihydro-2,2,4,6-tetramethyl-1H-indene-5-propanoic acid methyl ester (20) Prepared by reduction of 19. White solid (n-Hexane), m.p. 5455 C yield 92%. Found M+ 274.1573 (C17H22O3 requires 274.1569); max (KBr) 1 731, 1 693, 1 441 cm1; H (CDCl3) 1.14 (6H, s, 2CH3), 2.30 (3H, s, CH3), 2.42 (2H, t J 9 Hz, CH2), 2.49 (3H, s, CH3), 2.83 (2H, s, CH2), 2.90 (2H, t J 9 Hz, CH2), 3.62 (3H, s, OCH3), 6.87 (1H, s, CH): C 17.6 (CH3), 19.2 (CH3), 23.9 (CH2), 25.4 (2CH3), 33.0 (CH2), 41.0 (CH2), 45.2 (C), 51.5 (OCH3), 130.7 (C), 132.0 (CH), 133.1 (C), 137.1 (C), 142.5 (C), 151.5 (C), 172.82 (CO), 211.6

963 (CO); m/z 274(M+, 100), 260 (17), 259 (97), 242 (27), 201 (56), 200 (29), 199 (23), 187 (14), 185 (21). 4.1.3.3. 2,3-Dihydro-4,6-dimethyl-1H-inden-1-one-5-propanoic acid methyl ester (24) Prepared by reduction of 23. White solid (n-Hexane), m.p. 9798 C, yield 86%. Found M+ 245.1255 (C15H18O3 requires 246.1256; max (KBr) 1 738, 1 708 cm1. H (CDCl3) 2.26 (3H, s, CH3), 2.32 (3H, s, CH3), 2.40 (2H, t J 9 Hz, CH2), 2.58 (2H, t J 6 Hz, CH2), 2.92 (2H, t J 6 Hz, CH2), 3.00 (2H, t J 9 Hz, CH2), 3.67 (3H, s, OCH3), 7.35 (1H, s, CH); C 14.3 (CH3), 19.9 (CH3), 24.9 (CH2), 25.2 (CH2), 32.7 (CH2), 36.2 (CH2), 51.7 (OCH3), 122.3 (CH), 133.6 (C), 134.8 (C), 135.8 (C), 144.2 (C), 152.6 (C), 172.8 (CO), 207.1 (CO); m/z 245 (M+, 65), 214 (59), 186 (22), 173 (100), 159 (29), 144 (25), 128 (22), 115 (17), 91 (16). 4.1.3.4. 2,3-Dihydro-2,2,4,6-tetramethyl-1H-inden-1one-5-propanoic acid methyl ester (28) Prepared by reduction of 27. Colourless oil yield 91%. Found M+ 274.1570 (C17H22O3 requires 274.1569). max (KBr) 1 739, 1 711 cm1; H (CDCl3) 1.15 (6H, s, 2CH3), 2.23 (3H, s, CH3), 2.32 (3H, s, CH3), 2.40 (2H, t J 9 Hz, CH2), 2.80 (2H, s, CH2), 3.00 (2H, t J 9 Hz, CH2), 3.66 (3H, s, OCH3), 7.45 (1H, s, CH); C 14.4 (CH3), 19.9 (CH3), 25.2 (2CH3), 25.3 (CH2), 32.6 (CH2), 42.1 (CH2), 45.3 (C), 51.7 (OCH3), 123.1 (CH), 133.0 (C), 133.5 (C), 136.0 (C), 144.5 (C), 149.6 (C), 172.9 (CO), 211.5 (CO); m/z 274 (M+, 76), 260 (26), 259 (100), 227 (17), 201 (26), 187 (20), 128 (10). 4.1.3.5. 2,3-Dihydro-4,6-dimethyl-1H-indene-propanoic acid methyl ester (32) Prepared by reduction of 31. White solid (Pet. ether:ether), m.p. 4243 C, yield 83%. Found M+ 232.1464 (C15H20O2 requires 232.1463). max (KBr) 2 952, 1 740, 1 194cm1. H (CDCl3) 2.09 (2H, quintet J 7 Hz, CH2), 2.29 (3H, s, CH3), 2.37 (3H, s, CH3), 2.49 (2H, t J 9 Hz, CH2), 2.91 (4H, 2t J 7 Hz, 2CH2), 3.02 (2H, t J 9 Hz, CH2), 3.76 (3H, s, OCH3), 6.97 (1H, s, CH); C 15.8 (CH3), 19.9 (CH3), 24.7 (CH2), 24.9 (CH2), 31.9 (CH2), 32.9 (CH2), 33.6 (CH2), 51.5 (OCH3), 123.8 (CH), 131.8 (C), 134.0 (C), 134.4 (C), 141.3 (C), 141.7 (C), 173.5 (CO); m/z 232 (M+, 69), 217 (3), 160 (36), 159 (100), 146 (25), 149 (59), 128 (13). 4.1.3.6. 2,3-Dihydro-2,2,4,6-tetramethyl-1H-inden-1one-5-propanoic acid methyl ester (36) Prepared by reduction of 35. Colourless oil, yield 87%. Found M+ 260.1754 (C17H24O2 requires 260.1776). max (neat) 2 952, 1 742, 1 195cm1; H (CDCl3) 1.24 (6H, s, 2CH3), 2.23 (3H, s, CH3), 2.35 (3H, s, CH3), 2.52 (2H, t J 8 Hz, CH2), 2.71 (2H, s, CH2), 2.80 (2H, s, CH2), 2.95 (2H, t J 8 Hz, CH2), 3.76 (3H, s, OCH3), 6.86 (1H, s, CH); C 18.6 (CH3), 18.7 (CH3), 25.6 (CH2), 29.3 (2CH3), 33.7 (CH2), 39.2 (C), 46.3 (CH2), 46.6 (CH2), 51.5 (OCH3), 129.5 (CH), 131.7 (C), 131.9 (C), 133.4 (C), 139.8 (C), 142.0 (C), 173.4 (CO); m/z 260 (M+, 77), 245 (3), 228 (29), 187 (76), 173 (100), 141 (11). 4.1.4. General procedure for the preparation of propanoic acids (17, 21, 25, 29, 33 and 37) A solution of the corresponding propanoic acid methyl ester (1.45mmol) in 4 M HCl (10 mL) and THF (20 mL) was reuxed for 6 h. On completion, the THF was removed under vacuum and the remaining mixture was poured onto iced water (50 mL). The aqueous mixture was extracted with EtOAc (2 50 mL). The organic layer was then washed with NaOH (2 20 mL) and the aqueous layer was acidied with HCl and extracted into EtOAc. The EtOAc was dried over anhydrous Na2SO4, ltered and evaporated under vacuum. The residue was puried by recrystallisation to yield the following; 4.1.4.1 2,3-Dihydro-5,7-dimethyl-1H-inden-1-one-6-propanoic acid (17) Prepared by HCl reux from 16. White solid (Pet. ether:ether), m.p. 154 C, yield 88%. Found M+ 232.1075 (C14H16O3 requires 232.1099). max (KBr) 3 300, 1 732, 1 162 cm1; H (CDCl3) 2.38 (3H, s, CH3), 2.55 (2H, t J 6 Hz, CH2), 2.57 (3H, s, CH3), 2.67 (2H, t J 6 Hz, CH2), 3.01 (2H, t J 8 Hz, CH2), 3.04 (2H, t J 8 Hz, CH2), 6.94 (1H, s, CH); C 17.79 (CH3), 19.25 (CH3), 23.71 (CH2), 24.07 (CH2), 33.07 (CH2), 36.80 (CH2), 131.99 (CH), 132.76 (C), 133.06 (C), 136.72 (C), 142.66 (C), 154.85 (C), 177.84 (CO), 207.83 (CO). m/z 232 (M+, 86), 214 (11), 173 (100), 159 (59), 149 (55), 144 (15), 128 (17), 91 (11), 71 (18). 4.1.4.2. 2,3-Dihydro-2,2,5,7-tetramethyl-1H-inden-1one-6-propanoic acid (21) Prepared by HCl reux from 20. White needles (EtOH), m.p. 8486 C, yield 70%. Found M+ 260.1408 (C16H20O3 requires 260.1412). max (KBr) 3 4002 800b, 1 737, 1 700 cm1. H (CDCl3); 1.20 (6H, s, 2CH3), 2.37 (3H, s, CH3), 2.51 (2H, t J 8 Hz, CH2), 2.57 (3H, s, CH3), 2.89 (2H, s, CH2), 2.97 (2H, t J 8 Hz, CH2), 6.95 (1H, s, CH); C 17.86 (CH3), 19.35 (CH3), 23.82 (CH2), 25.47 (2CH3), 33.21 (CH2), 41.17 (CH2), 45.47 (C), 130.84 (C), 132.12 (CH), 132.91 (C), 137.36 (C), 142.83 (C), 151.81 (C), 178.42 (CO), 212.26 (CO). m/z 260 (M+, 74), 246 (20), 245 (100), 187 (19), 185 (20), 157 (17), 143 (17), 128 (19), 91 (15), 77 (11).

964 4.1.4.3. 2,3-Dihydro-4,6-dimethyl-1H-inden-1-one-5-propanoic acid (25) Prepared by HCl reux from 24. White needles (EtOH), m.p. 209210 C, yield 93%. Found M+ 232.1102 (C14H16O3 requires 232.1099). max (KBr) 3 1002 800b, 1 731, 1 178 cm1; H (CDCl3) 2.33 (3H, s, CH3), 2.40 (3H, s, CH3), 2.69 (2H, t J 8 Hz, CH2), 2.68 (2H, t J 6 Hz, CH3), 3.00 (2H, t J 6 Hz, CH2), 3.09 (2H, t J 8 Hz, CH2), 7.44 (1H, s, CH); C 14.17 (CH3), 19.76 (CH3), 24.80 (CH2), 24.84 (CH2), 32.46 (CH2), 36.06 (CH2), 122.31 (CH), 133.45 (C), 134.68 (C), 135.70 (C), 143.90 (C), 152.90 (C), 177.25 (CO), 207.36 (CO); m/z 232 (M+, 61), 173 (100), 172 (35), 160 (25), 159 (38), 129 (20), 128 (22), 115 (20), 91 (18), 77 (13). 4.1.4.4. 2,3-Dihydro-2,2,4,6-tetramethyl-1H-inden-1one-5-propanoic acid (29) White cubes (n-Hexane), m.p. 124125 C, yield 94%. Found M+ 260.1405 (C16H20O3 requires 260.1412). max (KBr) 3 200, 1 709, 1 692 cm1; H (CDCl3) 1.19 (6H, s, 2CH3), 2.28 (3H, s, CH3), 2.37 (3H, s, CH3), 2.49 (2H, t J 8 Hz, CH2), 2.84 (2H, s, CH2), 3.06 (2H, t J 8 Hz, CH2), 7.41 (1H, s, CH); C 14.41 (CH3), 19.92 (CH3), 25.18 (CH2), 25.26 (2CH3), 32.79 (CH2), 42.26 (CH2), 45.30 (C), 123.27 (CH), 133.10 (C), 133.59 (C), 136.09 (C), 144.50 (C), 149.75 (C), 177.41 (CO), 211.78 (CO); m/z 260 (M+, 25), 245 (44), 149 (29), 139 (10), 125 (19), 119 (16), 97 (43), 77 (100). 4.1.4.5. 2,3-Dihydro-4,6-dimethyl-1H-inden-5-propanoic acid (33) Prepared by HCl reux from 32. White solid (EtOH), m.p. 112 C, yield 89%. Found M+ 218.1302 (C14H18O2 requires 218.1099). max (KBr) 2 953, 1 699, 1 297 cm1; H (CDCl3) 2.10 (2H, quin J 7 Hz, CH2), 2.31 (3H, s, CH3), 2.38 (3H, s CH3), 2.54 (2H, t J 9 Hz, 8 Hz, CH2), 2.91 (4H, m, 2CH2), 3.05 (2H, t J 9 Hz, CH2), 6.98 (1H, s, CH). C 15.78 (CH3), 19.92 (CH3), 24.72 (2CH2), 31.96 (CH2), 32.95 (CH2) 33.64 (CH2), 123.84 (CH), 131.82 (C), 134.00 (C), 134.17 (C), 141.40 (C), 141.79 (C), 179.12 (CO); m/z 218 (M+, 59), 160 (28), 159 (100), 145 (44), 128 (14), 115 (10). 4.1.4.6. 2,3-Dihydro-2,2,4,6-tetramethyl-1H-indene-5-propanoic acid (37) Prepared by HCl reux from 36. White cubes (EtOH), m.p. 6364 C, yield 86%. Found 246.1617 (C16H22O2 requires 246.1620). max (KBr) 3 200, 1 736, 1 170 cm1; H (CDCl3) 1.16 (6H, s, 2CH3), 2.16 (3H, s, CH3), 2.28 (3H, s, CH3), 2.48 (2H, t J 9 Hz, CH2), 2.64 (2H, s, CH2), 2.72 (2H, s, CH2), 2.89 (2H, t J 9 Hz, CH2), 6.80 (1H, s, CH); C 18.67 (2CH3), 25.36 (CH2), 29.28 (2CH3), 33.62 (CH2), 39.21 (C), 46.28 (CH2), 46.53 (CH2), 129.46 (CH), 131.60 (C), 131.90 (C), 133.53 (C), 139.96 (C), 142.02 (C), 178.99 (CO); m/z 246 (M+, 6), 228 (45), 187 (57), 186 (38), 185 (16), 173 (83), 85 (17), 72 (100). 4.1.5. General procedure for the preparation of 3 hydroxypropyl derivatives (18, 26, 30, 34 and 38) Freshly generated diborane was introduced under N2 to a solution of the appropriate propanoic acids (1.3 mmol) in dry THF (10 mL) at 0 C. After 15 min EtOH was added and the solvents were removed under vacuum. The residue was puried by column chromatography (petether, EtOAc; 9:17:3) to yield the corresponding alcohols. 4.1.5.1. ()-2,3-Dihydro-6-(3-hydroxypropyl)-5,7-dimethyl-1H-inden-1-ol (18) Prepared by diborane reduction of 17. White powder (n-hexane), m.p. 86 C, yield 63%. Found M+ 220.1457 (C14H20O2 requires 220.1463). max (KBr) 3 350b, 2 936, 1 465cm1; H (CDCl3) 1.74 (2H, m, CH2), 2.08 (1H, m, CH), 2.30 (3H, s, CH3), 2.36 (3H, s, CH3), 2.39 (2H, m, CH), 2.65 (2H, t J 8 Hz, CH2), 2.81 (1H, dq J 3,9 Hz, CH), 3.08 (1H, quin J 8 Hz, CH), 3.69 (2H, t J 6 Hz, CH2), 5.28 (1H, dd J 2,7 Hz, CH), 6.87 (1H, s, CH); C 17.8 (CH3), 18.8 (CH3), 26.2 (CH2), 28.7 (CH2), 32.2 (CH2), 35.0 (CH2), 62.7 (CH2), 75.4 (CH), 130.5 (CH), 132.2 (C), 133.7 (C), 136.4 (C), 140.6 (C), 142.8 (C); m/z 220 (M+, 50), 205 (26), 184 (26), 175 (30), 169 (36), 161 (41), 159 (39), 157 (100), 145 (35), 141 (24). 4.1.5.2. 2,3-Dihydro-6-(3-hydroxypropyl)-2,2,5,7-tetramethyl-1H-inden-1-one (22) LiAlH4 (15 mg, 0.39 mmol) was added to a solution of 20 (200 mg, 0.73 mmol) in dry Et2O (10 mL) and the mixture was stirred at 0 C for 10min. The reaction mixture was poured onto ice/HCl and extracted with EtOAc. The organic layer was evaporated and the residue was puried by column chromatography on silica gel eluant (pet. ether:EtOAc, 7:3), followed by preparative TLC using the same developer to give 22 as a pale oil yield 45%. Found M+ 246.1016 (C16H22O2 requires 246.1614). max (lm) 3 400, 1 705, 1 465 cm1; H (CDCl3) 1.21 (6H, s, CH3), 1.64 (2H, m, CH2), 2.37 (3H, s, CH3), 2.47 (2H, t J 8 Hz, CH2), 2.57 (3H, s, CH3), 2.89 (2H, s, CH2), 4.06 (2H, t J 7, CH2), 6.95 (1H, s, CH); C 17.8 (CH3), 19.3 (CH3), 24.0 (CH2), 25.5 (2CH3), 32.3 (CH2), 41.1 (CH2), 45.4 (C), 62.1 (CH2), 131.0 (C), 132.0 (CH), 133.2 (C), 137.1 (C), 142.6 (C), 151.6 (C), 215.0 (C=O). 246 (M+, 8), 245 (42), 201 (100), 200 (93).

965 4.1.5.3. ()-2,3-Dihydro-5-(3-hydroxypropyl)-4,6-dimethyl-1H-inden-1-ol (26) Prepared by diborane reduction of 25. Waxy solid (EtOH), yield 47%. M+ 220.1431 (C14H20O2 requires 220.1463). max (KBr) 3 400b, 2 954, 1 457cm1; H (CDCl3) 1.68 (2H, m, CH2), 1.71 (1H, m, CH), 2.09 (1H, m, CH), 2.24 (3H, s, CH3), 2.35 (3H, s, CH3), 2.70 (2H, t J 8 Hz, CH2), 2.77 (1H, m, CH), 2.98 (1H, quin J 8 Hz, CH), 3.70 (2H, t J 9 Hz, CH2), 4.89 (1H, dd J 4 Hz, 7 Hz, CH), 7.09 (1H, s, CH); C 15.5 (CH3), 20.1 (CH3), 25.8 (CH2), 29.4 (CH2), 31.9 (CH2), 32.2 (CH2), 62.7 (CH2OH), 83.4 (CH), 124.2 (CH), 131.9 (C), 134.3 (C), 138.6 (C), 139.9 (C), 140.9 (C); m/z 220 (M+, 67), 175 (100), 169 (19), 160 (16), 131 (33), 115 (35), 91 (29). 4.1.5.4. ()-2,3-Dihydro-5-(3-hydroxypropyl)-2,2,4,6tetramethyl-1H-inden-1-ol (30) Prepared by diborane reduction of 29. White needles (EtOH), m.p. 130131 C yield 85%. M+ 248.1773 (C16H24O2 requires 220.1776). max (KBr) 3 600b, 1 464 cm1. H (acetone D6) 0.98 (3H, s, CH3), 1.13 (3H, s, CH3), 2.17 (3H, s, CH3), 2.24 (3H, s, CH2), 2.28 (3H, s, CH3), 2.48 (1H, d J 15 Hz, CH), 2.66 (1H, d J 14 Hz, CH), 2.70 (2H, t J 8 Hz, CH2), 2.90 (1H, bs, OH), 3.62 (2H, t J 7 Hz, CH2), 3.97 (1H, d J 5 Hz, OH), 4.56 (1H, d J 5 Hz, CH), 6.94 (1H, s, CH); C 15.4 (CH3), 20.1 (CH3), 25.9 (CH2), 27.2 (CH2), 32.1 (CH2), 43.6 (C), 44.2 (CH2), 62.6 (CH2), 83.6 (CH), 123.8 (C), 132.0 (C), 134.4 (C), 138.3 (C), 138.7 (C), 141.4 (C); m/z 248 (M+, 94), 247 (28), 233 (15), 203 (46), 189 (100), 186 (32), 171 (21), 91 (14). 4.1.5.5. 2,3-Dihydro-5-(3-hydroxypropyl)-4,6-dimethyl1H-indene (34) Prepared by diborane reduction of 33. White needles (EtOH), m.p. 5152 C yield 64%. Found M+ 204.1508 (C14H20O requires 204.1514). max (KBr) 3 400b, 1 459cm1; H (CDCl3) 1.80 (1H, quin J 8 Hz, CH2), 2.08 (2H, quin J 8 Hz, CH2), 2.28 (3H, s, CH3), 2.35 (3H, s, CH3), 2.74 (2H, t J 8 Hz, CH2), 2.90 (4H, 2t J 8 Hz, 2CH2), 3.78 (2H, t J 8 Hz, CH2), 6.96 (1H, s, CH); C 15.8 (CH3), 20.0 (CH3), 24.7 (CH2), 25.7 (CH2), 32.0 (CH2), 32.4 (CH2), 32.9 (CH2), 63.0 (CH2), 123.7 (CH), 131.7 (C), 133.9 (C), 136.0 (C), 141.0 (C), 141.2 (C); m/z 204 (M+, 61), 160 (39), 159 (100), 146 (19), 145 (28), 129 (11). 4.1.5.6. 2,3-Dihydro-5-(3-hydroxypropyl)-2,2,4,6-tetramethyl-1H-indene (38) Prepared by diborane reduction of 37. Pale oil, yield 59%. Found M+ 232.1785 (C16H24O requires 232.1821). max (lm) 3 400b, 1 464 cm1. H (CDCl3) 1.18 (6H, s, 2CH3), 1.76 (2H, m, CH2), 2.21 (3H, s, CH3), 2.32 (3H, s, CH3), 2.60 (2H, t J 7 Hz, CH2), 2.66 (2H, s, CH2), 2.72 (2H, s, CH2), 3.76 (2H, t J 6 Hz, CH2), 6.86 (1H, s, CH); C 15.6 (CH3), 18.6 (CH3), 25.7 (CH2), 29.2 (2CH3) 32.5 (CH2), 39.1 (C), 46.3 (CH2), 47.1 (CH2), 62.8 (CH2), 123.9 (CH), 129.3 (C), 131.8 (C), 132.4 (C), 139.6 (C), 140.3 (C); m/z 232 (M+, 100), 187 (90), 173 (90), 159 (58), 128 (40). 4.1.5.7. 2,3-Dihydro-5-(3-hydroxypropyl)-2,2,4,6-tetramethyl-1H-inden-1-one (39) A solution of NaOCl (5 mL, 12.5% w/v) was added dropwise to a stirred solution of 30 (250 mg, 1 mmol) in glacial acetic acid (3 mL) at 0 C. The reaction mixture was stirred for 30 min and was then poured onto ice and extracted with EtOAc. The organic layer was washed with H2O, dried over Na2SO4 and evaporated under vacuo. The residue was puried by column chromatography on silica gel (eluant: pet. ether:EtOAc, 7:3) to yield 39 as a waxy solid yield 43%. Found M+ 246.1599 (C16H22O2 requires 246.1614). max (KBr) 3 400b, 1 712 cm1; H (CDCl3) 1.23 (6H, s, 2CH3), 1.70 (2H, m, CH2), 2.28 (3H, s, CH3), 2.38 (3H, s, CH3), 2.80 (2H, t J 8 Hz, CH2), 2.86 (2H, s, CH2), 3.77 (2H, t J 7 Hz, CH2), 7.43 (1H, s, CH); C 15.9 (CH3), 20.0 (CH3), 25.3 (2CH3), 25.8 (CH2), 31.7 (CH2), 43.6 (C), 44.2 (CH2), 62.9 (CH2), 123.0 (CH), 132.7 (C), 134.5 (C), 138.3 (C), 141.4 (C), 146.4 (C), 209.5 (C=O); m/z 246 (M+, 73), 231 (40), 215 (100), 128 (22). 4.2. Pharmacological methods Smooth muscle relaxant activity was assessed as described previously [1]. Guinea-pigs (250400 g) of either sex were killed by cervical dislocation and exsanguination. The abdomen was opened by midline incision and the ileum removed. The tissue was stored at 4 C in Krebs solution (composition (mM): NaCl 118, KCl 4.7, CaCl2 2.5, MgCl2 1.15, NaH2PO4 1.17, NaHCO3 25, glucose 14.4). Segments of ileum 2.5 cm in length were suspended in a high potassium calcium-free modied Krebs solution (composition (mM): NaCl 12.5, KCl 45, MgCl2 1.15, NaH2PO4 1.17, NaHCO3 25, glucose 11.1) at 37 C, gassed with 95% O2/5% CO2 under a resting tension of 1.5 g, from Grass FT.03 transducers tted with black springs to record contractions isometrically. Contractions were displayed on a Grass 79D oscillograph. Sustained (> 40 min) contactures were elicited by addition of CaCl solution sufficient to raise the calcium concentration in the bath to 2.5 mM. When contractures had reached a stable maximum, test compounds were administered at a single concentration of 1 105 M or in the case of sufficiently active compounds, they were added cumulatively (3 1083 105M). Compounds

966 were dissolved in 0.5% ethanol, addition of which (0.1 mL) induced an inhibition of calcium contractions of approximately 14%. However, this inhibition was temporary, and contractions returned to control levels within 5 min. In contrast, the inhibitory effect of all compounds was sustained, and subsequently all responses were measured at least 10 min following addition of compound. 5. Statistics Results are expressed as means SE, mean of % inhibition of calcium contractures at a single dose of 1 105 M. EC50 values were determined from the linear portion of individual dose/inhibition curves by regression analysis, and expressed as a mean SE, mean for each compound shown. Statistical analysis was performed using Students t-test for unpaired samples, with a probability value (P) of less than 0.05 being taken as signicant. n denotes the number of preparations used in that series of experiments. Acknowledgements We would like to acknowledge an EOLAS grant which supported this work. References
[1] [2] [3] Sheridan H., Lemon S., Frankish N., McCardle P., Higgins T., James J.P., Bhandari P., Eur. J. Med. Chem. 25 (1990) 603608. Farrell R., Kelleher F., Sheridan H., J. Nat. Prod. 59 (1996) 446447. Murakami T., Tanaka N., in: Herz W., Griesbeck H., Kirby G., Tamm C. (Eds.), Progress in the Chemistry of Organic Natural Products Vol 54, Springer-Verlag, Vienna, New York, 1988, pp. 5156 and 176195. Bardouille V.S., Mooto B., Hirotsu K., Clardy J., Phytochemistry 17 (1978) 275278. Sheridan H., Frankish N., Farrell R., Planta Medica 65 (1998) 271272. Cabri W., Candiani I., Acc. Chem. Res. 28 (1995) 2. Heck R., Palladium Reagents in Organic synthesis, Academic Press, London, 1985, p. 415. Ho S., Yang M., Wu T., Wang C., Planta Medica 51 (2) (1985) 148150. Yang M., Planta Medica 52 (1) (1986) 2527.

[4] [5] [6] [7] [8] [9]

Eur. J. Med. Chem. 34 (1999) 977989 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

977

Original article

Synthesis and pharmacological evaluation of carboxamide derivatives as selective serotoninergic 5-HT4 receptor agonists
Katsuhiko Itoh*, Koji Kanzaki, Tsuguo Ikebe, Takanobu Kuroita, Hideo Tomozane, Shuji Sonda, Noriko Sato, Keiichiro Haga, Takeshi Kawakita
Research Laboratories, Yoshitomi Pharmaceutical Industries Ltd., 955 Koiwai, Yoshitomi-cho, Chikujo-gun, Fukuoka 871-8550, Japan (Received 29 September 1998; revised 8 December 1998; accepted 9 December 1998)

Abstract A number of new carboxamide derivatives were synthesized. The affinity of these compounds for the serotoninergic 5-HT4 receptor was evaluated by use of radioligand-binding techniques. The agonistic activity was evaluated as the contractile effect of the ascending colon isolated from guinea-pigs. Among these compounds, 4-amino-5-chloro-2-methoxy-N-[1-[2-[(methylsulfonyl)amino]ethly]-4piperidinylmethyl]benzamide (24) showed a high affinity for the 5-HT4 receptor (Ki = 9.6 nM). Compound 24 displayed a higher affinity for 5-HT4 receptors than the other receptors, including, 5-HT3 and dopamine D2 receptors. In addition, compound 24 was conrmed to be a potent 5-HT4 receptor agonist (ED50 = 7.0 nM). An interaction model between compound 24 and 5-HT4 receptor was proposed. 1999 ditions scientiques et mdicales Elsevier SAS 5-HT4 receptor / 5-HT4 receptor agonist / structureactivity relationship / carboxamide / receptor model construction

1. Introduction Serotonin (5-HT) is a neurotransmitter responsible for a wide range of pharmacological reactions. Serotonergic receptors are now classied into four broad subtypes such as 5-HT1, 5-HT2, 5-HT3, and 5-HT4 receptors, and clones for additional subtypes termed 5-HT5, 5-HT6, and 5-HT7 receptors have been identied [1]. Among diverse 5-HT receptors, we have investigated 5-HT4 function. Activation of the 5-HT4 receptor mediates widespread effects in the central and peripheral nervous systems [2]. Therefore, selective 5-HT4 receptor ligands would be useful for elucidating the physiological function of this receptor. Agonists for the 5-HT4 receptor known to date are indole derivatives such as 5-methoxytryptamine, benzamide derivatives such as cisapride [35], benzimidazolone derivatives such as BIMU8 [6], and benzoate derivatives such as ML 10302 [7] (gure 1). However, nonselective affinity or lability of these compounds prevents pure pharmacological characterization of this receptor in vivo. For example, 5-methoxytryptamine activates all 5-HT receptor subtypes except the 5-HT3 receptor [8].
*Correspondence and reprints

Cisapride and BIMU8 show obvious affinity for the 5-HT3 receptor [9, 10]. While, ML 10302 is a selective 5-HT4 receptor agonist, it shows limited pharmacological activities in vivo because of possible hydrolysis at its benzoate moiety. Recently, some ketone derivatives were reported to show a high affinity for the 5-HT4 receptor and to be selective 5-HT4 partial agonists [11]. In addition, the original 5-TH4 receptor mapping by the active analogue approach by using several 5-HT4 receptor antagonists and inactive ligands was reported [12]. One of our purposes is to supply a useful tool for elucidating the physiological function of the 5-HT4 receptor. In the course of our synthetic study on 5-HT3 receptor antagonists [13], it has been already claried that a benzoxazine-8-carboxamide derivative 1 (gure 2) shows a weak affinity for the 5-HT4 receptor. This result lead us to design selective 5-HT4 receptor agonists having novel chemical structure and chemical stability. Thus, compound 1, a leading compound for this study, could further be optimized at four points: optimization of (1) the structure of the cycloamine moiety, (2) the distance between the amide nitrogen in the carboxamide moiety and the basic nitrogen in the cycloamine moiety,

978

BIMU 8 Figure 1. Chemical structures of 5-HT4 receptor agonists.

ML 10302

(3) the variety and number of the substituents on the aromatic moiety, and (4) the structure of the side chain on the cycloamine moiety. Herein, we describe the design, synthesis, and pharmacological evaluation of carboxamide derivatives as selective 5-HT4 receptor agonists. Interaction between putative 5-HT4 receptor models and agonistic ligand is also discussed here. 2. Chemistry The general synthetic procedure used in this study is illustrated in gure 3. By the mixed anhydrides method (method A) or the 1-ethyl-3-[3-(dimethylamino)propyl] carbodiimide hydrochloride (WSC) method (method B), carboxylic acids (2a2h) were coupled with appropriate

amines (4a, 4b and 6a6f) to afford target benzamides (724), on which the synthetic data are listed in table I. Among the acidic starting materials, 3,4-dihydro-1,4benzoxazine-8-carboxylic acid (2a) was prepared by our method reported previously [14], and 2,3-dihydro-5chlorobenzofuran-7-carboxlic acid (2b) [15] and 5-chloro-2-methoxy-4-methylaminobenzoic acid (2h) [16] were prepared by known procedures. As for the other group of starting materials, pyrrolidine-3-methanol (3) was condensed with phthalimide by the Mitsunobu reaction [17] to afford a corresponding phthalimide derivative, which gave 1-benzyl-3-pyrrolidinylmethylamine (4a) through a hydrazinolysis (gure 4). 1-Substituted-4piperidinylmethylamines (6a6d) were prepared from their corresponding isonipecotamides (5a5d) by lithium aluminium hydride reduction (gure 5). 2-(1-Benzyl-4piperidiyl)ethylamine (6f) and [1-(2-methansufonylamino)ethyl]-4-piperidylmethylamine (6g) were prepared using our procedures [18]. 3. Pharmacological data and discussion 3.1. Structureactivity relationships The affinity of compounds 724 for the 5-HT4 receptor was determined as their ability to inhibit the binding of [3H]GR113808 to the receptor. Their affinities for 5-HT3 and dopamine D2 receptors were similarly evaluated by using [3H]granisetron and [3H]spiperone as radioligand, respectively. Here, membrane preparations of the stria-

Figure 2. Chemical structure of compound 1.

979

Figure 3. The general synthetic procedures used in this study. (a) method A: i-BuOCOCl, Et3N, AcOEt or method B: WSC, HOBt, DMF.

tum of guinea-pigs, the cerebral cortex of rats, and the striatum of rats were used for 5-HT4, 5-HT3, and D2 receptor binding assays, respectively. Either agonistic or antagonistic activity of these compounds was evaluated as the contractile ability of the ascending colon of guinea-pigs. Pharmacological data on compounds 1 and 724, 4-amino-N-(1-benzyl-4-piperidinyl)-5-chloro-4-methoxy-

benzamide (clebopride), and 5-HT are listed in table II, where clebopride and 5-HT are reference compounds. As concerns the cycloamine part, we chose not the bicycloamine group (tropane) but the monocycloamine group (pyrrolidine and piperidine) to obtain the compounds having a higher affinity for 5-HT4 receptor than for 5-HT3

Figure 4. Condensation of 3 by the Mitsunobu reaction. (a) PPh3, phthalimide, DEAD, THF; (b) NH2NH2H2O, EtOH.

Figure 5. Lithium aluminium hydride reduction of 5a5d. (a) LiA1H4, THF.

980
Table I Physicochemical properties for compounds 724.

Compound 7

Ar

n 0

R Bn
a

Method A

Mp (C) 9495

Formula C21H24CIN3O2

Bna

232233

C22H26CIN3O2C4H4O4f

Bna

186187

C22H26CIN3O2C2H4O4g

10

Bna

9798

C23H28CIN3O2

11

Bna

122124

C24H30CIN3O2C2H2O41/2EtOH 1/2H2Og

12

Bna

176178/DEC

C22H25CIN2O2C2H2O4g

13

Bna

161163

C20H24CIN3O2C2H2O4g

14

Bna

155156

C21H26CIN3O2

15

Bna

135137

C22H28CIN3O2C2H2O4g

16

Bna

169170

C21H26N2O2C2H2O4g

17

Bna

9093

C21H25CIN2O2C2H2O4 1/4H2Og

18

Bna

110112/DEC

C20H24CIN3O2C2H2O4 4/5H2Og

981
Table I (Continued) Compound 19 Ar n 1 X R Bn
a

Method B

Mp (C) 160161

Formula C21H27N3O2C2H2O4g

20

Bna

148149

C22H28CIN3O2C4H4O4f

21

Etb

115116

C16H24N3O2

22

Buc

179181

C18H28CIN3O2C2H2O4g

23

Hexd

113115

C20H32CIN3O2

24

MSAEe

177178/DEC

C17H27CIN4O4S2C2H2O4g

Bn, benzyl; bEt, ethyl; cBu, butyl; dHex, hexyl; eMSAE, (methylsulfonylamino)ethyl; ffumalate; goxalate.

receptor [7]. As a result, monocycloamine derivatives (7 and 8) in benzoxazine series showed a low affinity for 5-HT3 receptor. Furthermore, piperidine derivatives (10 and 14) showed a higher affinity for 5-HT4 receptor than pyrrolidine derivatives (9 and 13). Next, we investigated the distance requirements between the carbonyl group in the amide bond and the basic nitrogen in the cyclic amine of pyrrolidine and piperidine derivatives. Here, two benzoxazinecarboxamide derivatives (7 and 8) and clebopride, where the hydrogen atom on their carboxamide nitrogen is replaced by a 1-benzyl3-pyrrolidinyl or 1-benzyl-4-piperidinyl group, showed affinities for 5-HT3, 5-HT4 and D2 receptors. On the other hand, compounds 9, 10 and 1214, where the hydrogen atom on their carboxamide nitrogen is replaced by a 1-benzyl-3-pyrrolidinylmethyl or 1-benzyl-4-piperidinylmethyl group, showed a higher affinity for the 5-HT4 receptor than for the 5-HT3 receptor, and showed a low affinity for the D2 receptor. Compounds 11 and 15, where the hydrogen atom on their carboxamide nitrogen is replaced by a 2-(1-benzyl-4-piperidinyl)ethyl group, exhibited only low affinities for these kinds of receptors. The results indicate that the optimum distance for selective affinity for the 5-HT4 receptor was one carbon [9].

Compound 10 was conrmed to be a 5-HT4 receptor antagonist (pKB = 7.9). Compound 12, a benzofuran analogue of benzoxazine derivative 10, was also conrmed to behave as a 5-HT4 receptor antagonist (pKB = 8.2) although they showed a high affinity for the receptor. However, compounds 13 and 14, both of which have a classic 4-amino-5-chloro-2-methoxybenzamide skeleton, were conrmed to be a 5-HT4 receptor agonist. We next focused on the effect of each substituent (4-amino, 5-chloro and 2-methoxy group) on the phenyl ring in benzamide derivatives. A simple 2-methoxybenzamide derivative 16 showed a low affinity to the receptor. Although 5-chloro-2-methoxybenzamide derivative 17 exhibited a 5-HT4 receptor affinity similar to compound 14, the former never behaved as a agonist for the receptor. On the other hand, such an affinity was practically undetectable for the 4-amino-3-chlorobenzamide derivative 18, a 2-desmethoxy analogue of compound 14. The moderate affinity of compound 16 and the low affinity of compound 18 indicate that the 2-methoxy group is essential for binding to the 5-HT4 receptor. Compound 17 showed a similar affinity to compound 14. This result suggests that the 5-chloro group helped ligands to bind to the receptor complementarily. The 4-Amino-2-

982
Table II Pharmacological data of compound 1, compounds 724 and two reference compounds. Binding affinities, Ki (nM)a Compound 1 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 clebopride 5-HT
a

Contractile effects in guinea pigs ascending colon D2 > 1000b 12 67 > 1000b > 1000b > 1000b NT > 1000b > 1000b > 1000b 880 NTc NTc NTc NTc NTc NTc NTc > 1000b 63 EC
50

5-HT4 280 160 530 19 0.93 530 1.7 20 6.7 > 1000b 66 8.1 > 1000b 110 100 28 8.8 6.7 9.6 92 130

5-HT3 0.27 10 21 95 120 > 1000b 340 160 290 470 > 1000b > 1000b > 1000b > 1000b > 1000b > 1000b > 1000b > 1000b > 1000b 230

(nM)d

Maximal response (%)e

NTc NTc NTc > 10000 > 10000 NT > 10000 47 20 NTc NTc > 10000 NTc 150 900 61 13 130 7.0 NTc 35

21 18 20 8.9 16 15 20 27 21

Each value is the mean from triplicate assays in a single experiment; bIC50 value; cnot tested; dEC50 values were determined by linear regression; e% of contraction to methacoline at 30 M.

methoxybenzamide derivative 19, a 5-deschloro analogue of compound 14, showed a potent 5-HT4 receptor agonistic activity, but showed only a moderate affinity for the receptor. These results indicate that the 4-amino group played an important role in 5-HT4 receptor agonistic activity. In addition, this consideration is supported by the facts, which 2,3-dihydro-4-amino-5-chlorobenzofuran carboxamides showed 5-HT4 receptor agonistic activity [19], while 4-desamino-2,3-dihydro-5-chlorobenzofuran carboxamide 12 showed low 5-HT4 receptor agonistic activity. Although 5-chloro-2-methoxy-4methylaminobenzamide derivative 20, an N-methyl analogue of compound 14, also exhibited a moderate affinity, it showed low 5-HT4 receptor agonistic activity. Consequently, all the three substituents on the phenyl ring are essential for both affinity and agonistic activity. Finally, we examined the inuence of the side chains of N-substituted piperidinylmethyl group on the interaction with the 5-HT4 receptor. Ethyl derivative 21 showed a moderate affinity for the 5-HT4 receptor. Butyl derivative 22 was almost equal to compound 14 in the affinity, hexyl derivative 23 was equal to compound 14 in the affinity, but with reduced 5-HT4 receptor agonistic activity. When the substituent was a polar methylsulfonylamino group, agonistic activity increased. Compound 24 was a more

potent agonist at the 5-HT4 receptor than other compounds (14, 2123). Therefore, compound 24 was assayed for its binding ability for several kinds of receptors. The results were as follows (IC50 value (nM), ligand): 1 > 1 000, [3H]prazosin; 5-HT1A > 1 000, [3H]8-OHDPAT; 5-HT2 > 1 000, [3H]ketanserin; MACh > 1 000, [3H]QNB. Thus Compound 24 represented a selective 5-HT4 receptor agonist. 3.2. Computational modelling study 3.2.1. 5-HT4 receptor model construction A 5-HT4 receptor model was constructed in order to obtain rational structureactivity relationships and to clarify the interaction between ligands and the receptor. The 5-HT4 receptor is a member of the G-protein coupled receptor (GPCR). Generally, GPCRs are assumed to consist of seven transmembrane helices and eight intra and extracellular loops and their models are constructed from coordinates of bacteriorhodopsin [2027]. The credibility of the model is evaluated by the consistency with structureactivity relationships and mutational data. In this study, affinity for the guinea-pig 5-HT4 receptor was measured. Therefore the receptor model should be constructed from the sequence of guinea-pig. But the guinea-

983 pig sequence has not been reported yet. Although sequences of receptors are not identical among animals, amino acids involved in ligand recognition are most likely to be conserved. Therefore, in the agonist recognition site, the conserved amino acids between human and rat 5-HT4 receptors should be conserved in guinea-pig 5-HT4 receptors. The 5-HT4 receptor model was constructed with the residues in the human 5-HT4 receptor and the conserved residues were used for speculation of receptor-ligand interaction. The seven transmembrane regions of the 5-HT4 receptor was determined from a hydropathy plot [28]. To determine sequence alignment of bacteriorhodopsin and the 5-HT4 receptor, over 70 GPCRs such as dopamine, adrenergic, muscarine receptor, and rhodopsin were included in the study. 3.2.2. Docking of compound 24 into the 5-HT4 receptor model Mutational studies for the 5-HT4 receptor have not been reported yet. Therefore, the structureactivity relationships for this series of compounds and mutational data for the 5-HT1A and 5-HT2A receptors [2933] were used for the docking study of compound 24 into the receptor model. From the structureactivity relationships, the 2-methoxy, 5-chloro, and 4-amino groups on the phenyl ring, and the methylsulfonylamino group in the piperidinomethyl moiety were dened to be essential for both affinity and agonistic activity. These groups are assumed to interact with certain residues in the 5-HT4 receptor. From some mutational data, Asp308, Thr312, Ser507, Ala510, and Phe619 are dened to be important for agonist binding and receptor activation. Asp308 is supposed to form a salt bridge with the ammonium nitrogen of the ligand. Thr312 and Ser507 are supposed to form a hydrogen bond. Ala510 is supposed to perform a hydrophobic interaction with N-alkyl substituents in some tryptamine derivatives. Phe619 is assumed to recognise an aromatic ring of the ligand. From these structureactivity relationships and mutational data, an interaction model between the 5-HT4 receptor and compound 24 was proposed : (gure 6) the ammonium nitrogen in the piperidine ring forms a salt bridge with Asp308. The 2-methoxy group forms an intramolecular hydrogen bond with the NH group in the benzamide and constrains the conformation of the benzamide group. The 5-chloro group performs a hydrophobic interaction with Ala510. The 4-amino group forms a hydrogen bond with Ser507. The methylsulfonylamino group forms a hydrogen bond with Thr312. Furthermore, aromatic residues such as Phe515, Phe612, Trp616, Phe619, and Phe620 are presumed to locate near the ligand and they are supposed to form an aromatic binding pocket or stabilize the positive charge of the cationic amine. As for Trp409, there may be a hydrogen bond to the CO group in the benzamide.

Figure 6. Stereoview of the interaction of compound 24 with the 5-HT4 receptor.

984 4. Conclusion We describe the synthesis and biological evaluation of a series of carboxamides as selective 5-HT4 receptor agonists. Based on our results, we propose that 4-amino5-chloro-2-methoxy-N-[(4-piperidinyl)methyl]benzamide with a polar group at the 1-position on piperidine moieties are necessary for the selective agonistic activity for 5-HT4 receptors. Compound 24 represents a selective 5-HT4 agonist and would be a useful tool for probing the 5-HT4 receptor function in vivo. Detailed analyses of structureactivity relationships for the side chain moieties will be reported in due course. 5. Experimental protocols 5.1. Chemistry All melting points were measured in open capillaries and uncorrected. Proton nuclear magnetic resonance (1H-NMR) spectra were recorded on Jeol JNM-EX270 spectrometers and chemical shifts are expressed in ppm with tetramethylsilane (TMS) as an internal standard. Signal multiplicities are represented by s (singlet), d (doublet), t (triplet), q (quartet), br-s (broad singlet) and m (multiplet). Mass spectra (MS) were taken on Jeol JMS-O1SG spectrometers. Elementary analysis was performed for C, H and N, and were within 0.4% of the calculated values. Silica-gel plates (Merck F254) and silica gel 60 (Merck, 70230 mesh) were used for analytical and preparative column chromatography, respectively. 5.1.1. General procedure for the preparation of 711 (method A) 5.1.1.1. N-(1-Benzyl-3-pyrrolidinyl)-6-chloro-3,4dihydro-4-methyl-2H-1,4-benzoxazine-8-carboxamide 7 Isobutyl chloroformate (1.2 g, 9.7 mmol) was added to a mixture of 2a (2.0 g, 8.8 mmol), triethylamine (2.0 g, 19 mol), and ethyl acetate (40 mL) at 10 C. The mixture was stirred below 5 C for 30 min and a solution of 4b (1.5 g, 8.8 mmol) in ethyl acetate (10 mL) was added with stirring at 10 C. Stirring was continued at the same temperature for 30 min. The resulting mixture was added to water and extracted with ethyl acetate. The extract was washed with brine, and dried over anhydrous magnesium sulfate. After evaporation in vacuo, the residue was chromatographed on silica gel (CHCl3:MeOH = 10:1). Recrystallized from ethyl acetate/diisopropylether to give 7 (0.50 g, 15%); 1H-NMR (CDCl3) : 1.701.82 (2H, m), 2.64 (2H, d, J = 11 Hz), 2.90 (3H, s), 3.35 (2H, t, J = 5.6 Hz), 3.55 (1H, d, J = 11 Hz), 3.70 (1H, d, J = 11 Hz), 4.35 (2H, t, J = 5.6 Hz), 4.444.80 (1H, m), 6.62 (1H, d, J = 3.0 Hz), 7.207.30 (5H, m), 7.39 (1H, d, J = 3.0 Hz), 7.95 (1H, br-s); MS m/z: 385 (M+). 5.1.1.2. N-(1-Benzyl-4-piperidinyl)-6-chloro-3,4-dihydro-4-methyl-2H-1,4-benzoxazine-8-carboxamide fumalate 8 Similarly to 7, 8 was prepared starting from 2a (3.4 g, 15 mmol), triethylamine (3.3 g, 33 mmol), isobutyl chloroformate (2.3 g, 17 mmol), ethyl acetate (70 mL), and 6e (3.0 g, 15 mmol). The resulting oil was transformed into fumalate and recrystallized from ethanol to give 8 (3.4 g, 55%); 1H-NMR (DMSO-d6) : 1.54 (2H, dd, J = 12, 24 Hz), 1.82 (2H, d, J = 12 Hz), 2.202.39 (2H, t, J = 12 Hz), 2.80 (2H, d, J = 12 Hz), 2.87 (3H, s), 3.29 (2H, t, J = 4.0 Hz), 3.50 (2H, s), 3.703.83 (1H, m), 4.29 (2H, t, J = 4.0 Hz), 6.60 (2H, s), 6.75 (1H, d, J = 3.0 Hz), 6.82 (1H, d, J = 3.0 Hz), 7.227.38 (5H, m), 7.98 (1H, d, J = 7.2 Hz); MS m/z: 399 (M+). 5.1.1.3. N-[(1-Benzyl-3-pyrrolidinyl)methyl]-6-chloro3,4-dihydro-4-methyl-2H-1,4-benzoxazine-8-carboxamide oxalate 9 Similarly to 7, 9 was prepared starting from 2a (1.8 g, 7.9 mmol), triethylamine (1.6 g, 16 mmol), isobutyl chloroformate (1.1 g, 7.9 mmol), ethyl acetate (30 mL), and 4a (1.5 g, 7.9 mmol). The resulting oil was transformed into oxalate and recrystallized from methanol to give 9 (2.4 g, 62%); 1H-NMR (DMSO-d6) : 1.602.22 (2H, m), 2.642.75 (2H, m), 2.802.92 (2H, m), 2.90 (3H, s), 3.17 (2H, d, J = 5.9 Hz), 3.55 (2H, s), 3.783.95 (1H, m), 4.21 (2H, t, J = 4.0 Hz), 4.29 (2H, t, J = 4.0 Hz), 6.74 (1H, d, J = 3.0 Hz), 6.78 (1H, d, J = 3.0 Hz), 7.257.60 (5H, m), 8.28 (1H, t, J = 6.0 Hz); MS m/z: 399 (M+). 5.1.1.4. N-[(1-Benzyl-4-piperidinyl)methyl]-6-chloro3,4-dihydro-4-methyl-2H-1,4-benzoxazine-8-carboxamide 10 Similarly to 7, 10 was prepared starting from 2a (1.2 g, 5.3 mmol), triethylamine(0.80 g, 8.0 mmol), isobutyl chloroformate (0.67 g, 5.3 mmol), ethyl acetate (25 mL), and 6a (1.0 g, 5.3 mmol). The resulting solid was recrystallized from ethyl acetate to give 10 (1.4 g, 64%); 1 H-NMR (CDCl3) : 1.231.32 (2H, m), 1.55 (1H, br-s), 1.73 (2H, d, J = 11 Hz), 2.002.42 (2H, m), 2.80 (2H, d, J = 11 Hz), 2.90 (3H, s), 3.30 (2H, t, J = 5.6 Hz), 3.35 (2H, d, J = 11 Hz), 3.50 (2H, s), 4.35 (2H, t, J = 5.6 Hz), 6.64 (1H, d, J = 3.0 Hz), 7.207.30 (5H, m), 7.42 (1H, d, J = 3.0 Hz), 7.69 (1H, br-s); MS m/z: 413 (M+).

985 5.1.1.5. N-[2-(1-Benzyl-4-piperidinyl)ethyl]-6-chloro3,4-dihydro-4-methyl-2H-1,4-benzoxazine-8-carboxamide oxalate 11 Similarly to 7, 11 was prepared starting from 2a (1.0 g, 4.6 mmol), triethylamine (1.0 g, 10 mmol), isobutyl chloroformate (0.70 g, 5.1 mmol), ethyl acetate (50 mL), and 6f (1.0 g, 4.6 mmol). The resulting oil was transformed into oxalate and recrystallized from ethanol to give 11 (0.56 g, 24%); 1H-NMR (DMSO-d6) : 1.321.60 (4H, m), 1.84 (2H, d, J = 12 Hz), 2.80 (2H, d, J = 12 Hz), 2.84 (2H, t, J = 11 Hz), 2.87 (3H, s), 3.29 (5H, m), 4.17 (2H, s), 4.28 (2H, t, J = 4.6 Hz), 6.73 (1H, d, J = 2.7 Hz), 6.82 (1H, d, J = 2.7 Hz), 7.417.49 (5H, m), 8.12 (1H, t, J = 5.2 Hz); MS m/z: 427 (M+). 5.1.2. General procedure for the preparation of 1224 (method B) 5.1.2.1. 4-Amino-N-[(1-benzyl-3-pyrrolidinyl)methyl]-5chloro-2-methoxybenzamide oxalate 13 A mixture of 2c (1.6 g, 7.9 mmol), 4a (1.5 g, 7.9 mmol), 1-hydroxybenzotriazole (HOBt) (1.1 g, 8.1 mmol), and dimethylformamide (50 mL) was stirred under ice-cooling for 1 h and then 1-ethyl-3-[3(dimethylamino)propyl] carbodiimide hydrochloride (WSC) (2.3 g, 8.4 mmol) was added at the same temperature. Stirring was continued overnight at room temperature. After evaporation, 5% aqueous sodium bicarbonate was added to the residue and extracted with ethyl acetate. The extraction was washed with brine, and was dried over anhydrous magnesium sulfate. After evaporation in vacuo, the residue was dissolved in ethanol, an alcoholic solution of oxalic acid (2.5 equiv.) was added. The precipitates were collected and recrystallized from ethanol to give 13 (1.9 g, 52%); 1H-NMR (DMSO-d6) : 1.602.22 (2H, m), 2.642.75 (2H, m), 2.802.92 (3H, m), 3.17 (2H, d, J = 5.9 Hz), 3.55 (2H, s), 3.80 (3H, s), 4.23 (2H, br-s), 6.48 (1H, s), 7.257.55 (5H, m), 7.62 (1H, s), 8.02 (1H, t, J = 6.0 Hz); MS m/z: 373 (M+). 5.1.2.2. N-[(1-Benzyl-4-piperidinyl)methyl]-5-chloro3,4-dihydro-1,4-benzofuran-7-carboxamide oxalate 12 Similarly to 13, 12 was prepared starting from 2b (0.50 g, 2.5 mmol), 6a (0.49 g, 2.5mmol), HOBt (0.33 g, 2.5 mmol), dimethylformamide (20 mL), and WSC (0.47g, 2.5 mmol). The resulting oil was transformed into oxalate and recrystallized from isopropanol to give 12 (0.89 g, 77%), 1H-NMR (CDCl3) : 1.652.02 (6H, m), 2.452.60 (2H, br-s), 3.16 (2H, t, J = 6.0 Hz), 3.30 (2H, br-s), 3.58 (2H, br-s), 4.18 (2H, br-s), 4.72 (2H, t, J = 6.0 Hz), 7.307.42 (5H, m), 7.62 (1H, t, J = 6.0 Hz), 8.82 (1H, d, J = 2.0 Hz); MS m/z: 384 (M+). 5.1.2.3. 4-Amino-N-[(1-benzyl-4-piperidinyl)methyl]-5chloro-2-methoxybenzamide 14 Similarly to 13, 14 was prepared starting from 2c (1.0 g, 5.0 mmol), 6a (1.0 g, 5.0 mmol), HOBt (0.86 g, 6.4 mmol), dimethylformamide (30 mL), and WSC (1.1 g, 5.7 mmol). The resulting solid was recrystallized from ethyl acetate to give 14 (1.4 g, 72%); 1H-NMR (CDCl3) : 1.351.55 (5H, m), 2.00 (2H, d, J = 12 Hz), 2.90 (2H, d, J = 12 Hz), 3.30 (2H, t, J = 5.6 Hz), 3.50 (2H, s), 3.92 (3H, s), 4.38 (2H, br-s), 7.227.38 (5H, m), 7.70 (1H, br-s), 8.10 (1H, s); MS m/z: 387 (M+). 5.1.2.4. 4-Amino-N-[2-(1-benzyl-4-piperidinyl)ethyl]-5chloro-2-methoxybenzamide oxalate 15 Similarly to 13, 15 was prepared starting from 2c (0.93 g, 4.6 mmol), 6f (1.0 g, 4.6 mmol), HOBt (0.65 g, 0.48 mmol), dimethylformamide (30 mL), and WSC (0.92 g, 4.8 mmol). The resulting oil was transformed into oxalate and recrystallized from ethanol to give 15 (0.56 g, 24%); 1H-NMR (DMSO-d6) : 1.351.55 (5H, m), 1.85 (2H, d, J = 12 Hz), 2.75 (2H, t, J = 6.8 Hz), 3.193.30 (4H, m), 3.88 (3H, s), 4.14 (2H, s), 5.91 (2H, br-s), 6.47 (1H, s), 7.417.47 (5H, m), 7.66 (1H, s), 7.89 (1H, t, J = 5.9Hz); MS m/z: 401 (M+). 5.1.2.5. N-[(1-Benzyl-4-piperidinyl)methyl]-2-methoxybenzamide oxalate 16 Similarly to 13, 16 was prepared starting from 2d (0.74 g, 4.9 mmol), 6a (1.0 g, 4.9 mmol), HOBt (0.73 g, 5.4 mmol), dimethylformamide (30 mL), and WSC (1.0 g, 5.4 mmol). The resulting oil was transformed into oxalate and recrystallized from ethanol to give 16 (1.5 g, 71%); 1H-NMR (DMSO-d6) : 1.44 (2H, dd, J = 12, 24 Hz), 1.701.80 (1H, m), 1.82 (2H, d, J = 12 Hz), 2.78 (2H, t, J = 12 Hz), 3.173.24 (4H, m), 3.85 (3H, s), 4.15 (2H, s), 6.76 (1H, d, J = 8.0 Hz), 7.01 (1H, d, J = 7.2 Hz), 7.437.50 (6H, m), 7.54 (1H, dd, J = 2.0, 7.2 Hz), 8.22 (1H, t, J = 6.0 Hz); MS m/z: 338 (M+). 5.1.2.6. N-[(1-Benzyl-4-piperidinyl)methyl]-5-chloro-2methoxybenzamide oxalate 17 Similarly to 13, 17 was prepared starting from 2e (0.82 g, 4.4 mmol), 6a (0.90 g, 4.4 mmol), HOBt (0.66 g, 4.8 mmol), dimethylformamide (30 mL), and WSC (0.93 g, 4.8 mmol). The resulting oil was transformed into oxalate and recrystallized from ethanol to give 17 (0.27 g, 16%); 1H-NMR (CDCl3-CD3OD) : 1.69 (2H, dd, J = 12, 24 Hz), 1.93 (2H, d, J = 12 Hz), 2.00 (1H, m), 2.78 (2H, br-s), 3.32 (2H, d, J = 6.6 Hz), 3.48 (2H, d, J = 12 Hz), 3.96(3H, s), 4.24 (2H, s), 6.97 (1H, d, J = 10 Hz), 7.41 (1H, dd, J = 2.7, 10 Hz), 7.507.62 (5H, m), 8.00 (1H, d, J = 2.7 Hz), 7.93 (1H, br-s); MS m/z: 372 (M+).

986 5.1.2.7. 4-Amino-N-[(1-benzyl-4-piperidinyl)methyl]-5chlorobenzamide oxalate 18 Similarly to 13, 18 was prepared starting from 2f (0.76 g, 4.4 mmol), 6a (0.90 g, 4.4 mmol), HOBt (0.66 g, 4.8 mmol), dimethylformamide (30 mL), and WSC (0.93 g, 4.8 mmol). The resulting oil was transformed into oxalate and recrystallized from ethanol to give 18 (0.41 g, 21%); 1H-NMR (CDCl3-CD3OD) : 1.74 (2H, dd, J = 12, 24 Hz), 1.90 (2H, d, J = 12 Hz), 1.98 (1H, m), 2.74 (2H, br-s), 3.24 (2H, d, J = 6.6 Hz), 3.44 (2H, d, J = 12 Hz), 4.26 (2H, s), 6.76 (1H, d, J = 10 Hz), 7.327.48 (5H, m), 7.54 (1H, dd, J = 2.7, 10 Hz), 7.79 (1H, d, J = 2.7 Hz), 7.93 (1H, br-s); MS m/z: 357 (M+). 5.1.2.8. 4-Amino-N-[(1-benzyl-4-piperidinyl)methyl]-2methoxybenzamide oxalate 19 Similarly to 13, 19 was prepared starting from 2g (0.82 g, 4.9 mmol), 6a (1.0 g, 4.9 mmol), HOBt (0.73 g, 5.4 mmol), dimethylformamide (30 mL), and WSC (1.0 g, 5.4 mmol). The resulting oil was transformed into oxalate and recrystallized from ethanol to give 19 (0.49 g, 23%); 1H-NMR (DMSO-d6) : 1.43 (2H, t, J = 12 Hz), 1.70 (1H, br-s), 1.77 (2H, d, J = 12 Hz), 2.78 (2H, t, J = 12 Hz), 3.18 (2H, t, J = 5.9 Hz), 3.25 (2H, d, J = 12 Hz), 3.80 (3H, s), 4.16 (2H, br-s), 6.16 (1H, dd, J = 2.0, 8.0 Hz), 6.22 (1H, J = 2.0 Hz), 7.46 (5H, m), 7.59 (1H, d, J = 8.0 Hz), 7.90 (1H, t, J = 6.0 Hz); MS m/z: 353 (M+). 5.1.2.9. N-[(1-Benzyl-4-piperidinyl)methyl]-5-chloro-2methoxy-4-methylaminobenzamide fumalate 20 Similarly to 13, 20 was prepared starting from 2h (1.2 g, 5.6 mmol), 6a (1.1 g, 5.6 mmol), HOBt (0.84 g, 6.2 mmol), dimethylformamide (30 mL), and WSC (1.2 g, 6.2 mmol). The resulting oil was transformed into fumalate and recrystallized from ethanol/acetone to give 20 (0.31 g, 14%); 1H-NMR (DMSO-d6) : 1.27 (2H, dd, J = 12, 21 Hz), 1.59 (1H, m), 1.65 (2H, d, J = 12 Hz), 2.21 (2H, t, J = 11 Hz), 2.83 (3H, d, J = 4.7 Hz), 2.95 (2H, d, J = 11 Hz), 3.18 (2H, t, J = 6.0 Hz), 3.69 (2H, s), 3.93 (3H, s), 6.50 (1H, q, J = 4.7 Hz), 6.22 (1H, s), 6.59 (2H, s), 7.317.70 (5H, m), 7.93 (1H, t, J = 6.0Hz); MS m/z: 401 (M+). 5.1.2.10. 4-Amino-5-chloro-N-[(1-ethyl-4-piperidinyl) methyl]-2-methoxybenzamide 21 Similarly to 13, 21 was prepared starting from 2c (2.0 g, 9.9mmol), 6b (1.4 g, 9.9 mmol), HOBt (1.7 g, 13 mmol), dimethylformamide (30 mL), and WSC (2.3 g, 1.2 mmol). The resulting solid was recrystallized from ethyl acetate to give 21 (2.0 g, 62%); 1H-NMR (CDCl3) : 1.14 (3H, t, J = 12 Hz), 2.203.05 (7H, m), 2.20 (2H, q, J = 12 Hz), 2.92 (2H, d, J = 12 Hz), 3.32 (2H, t, J = 6.6 Hz), 3.88 (3H, s), 4.38 (2H, br-s), 6.23 (1H, s), 7.68 (1H, br-s), 8.03 (1H, s); MS m/z: 325 (M+). 5.1.2.11. 4-Amino-N-[(1-butyl-4-piperidinyl)methyl]-5chloro-2-methoxybenzamide oxalate 22 Similarly to 13, 22 was prepared starting from 2c (0.40 g, 2.0 mmol), 6c (0.31 g, 2.0 mmol), HOBt (0.29 g, 2.2 mmol), dimethylformamide (10 mL), and WSC (0.42 g, 2.2 mmol). The resulting oil was transformed into oxalate and recrystallized from ethanol to give 22 (0.26 g, 29%); 1H-NMR (CD3OD) : 1.23 (3H, t, J = 8.5 Hz), 1.351.42 (2H, m), 1.62 (2h, br-s), 1.701.85(2H, m), 1.95 (1H, br-s), 1.97 (2H, d, J = 13Hz), 2.90 (2H, br-s), 3.07 (2H, t, J = 8.5 Hz), 3.31 (2H, br-s), 3.35 (2H, br-s), 3.91 (3H, s), 6.23 (1H, s), 7.78 (1H, s); MS m/z: 353 (M+). 5.1.2.12. 4-Amino-5-chloro-N-[(1-hexyl-4-piperidinyl) methyl]-2-methoxybenzamide 23 Similarly to 13, 23 was prepared starting from 2c (1.7 g, 8.4 mmol), 6d (1.5 g, 8.4 mmol), HOBt (1.5 g, 11 mmol), dimethylformamide (30 mL), and WSC (2.1 g, 11 mmol). The resulting solid was recrystallized from ethyl acetate/diisopropylether to give 23 (2.1 g, 66%); 1 H-NMR (CDCl3) : 0.83 (3H, t, J = 12 Hz), 1.202.05 (15H, m), 3.30 (2H, t, J = 12, 24 Hz), 3.92 (2H, d, J = 12 Hz), 3.33 (2H, t, J = 6.6 Hz), 3.90 (3H, s), 4.38 (2H, br-s), 6.28 (1H, s), 7.70 (1H, br-s), 8.10 (1H, s); MS m/z: 381 (M+). 5.1.2.13. 4-Amino-5-chloro-N-[[2-[(methylsulfonyl) amino]ethyl]-4-piperidinyl]-2-methoxybenzamide oxalate 24 Similarly to 13, 24 was prepared starting from 2c (2.0 g, 9.9 mmol), 6g (2.3 g, 9.9 mmol), HOBt (1.6 g, 12 mmol), dimethylformamide (50 mL), and WSC (2.3 g, 12 mmol). The resulting oil was transformed into oxalate and recrystallized from ethanol to give 24 (1.3 g, 26%); 1 H-NMR (DMSO-d6) : 1.202.00 (5H, m), 2.603.80 (13H, m), 2.96 (3H, s), 3.79 (3H, s), 4.38 (2H, br-s), 6.42 (1H, s), 7.60 (1H, s), 7.94 (1H, t, J = 6.0 Hz); MS m/z: 418 (M+). 5.1.3. 1-Benzyl-3-pyrrolidinylmethylamine 4a A mixture of 3 (5.0 g, 23 mmol), triphenylphosphine (6.6 g, 25 mmol), phthalimide (3.7 g, 25 mmol) and tetrahydrofuran (30 mL) was stirred under ice-cooling and then diethyl azodicarboxylate (3.6 mL, 23 mmol) was added at the same temperature. Stirring was continued overnight at room temperature. After evaporation, the residue was dissolved in ethyl alcohol and then hydrazine monohydrate (3.3 mL, 68.4 mmol) was added. The mixture was reuxed for 2 h. After cooling, the reaction

987 mixture was ltered by suction though celite and to the ltrate was added diethylether. The precipitate was ltered off and ltrate was extracted with 10% hydrochloric acid. The aqueous layer was basied by 10% NaOH and extracted with chloroform. The extraction was washed with brine, and was dried over anhydrous magnesium sulfate. After evaporation in vacuo to give 4a (2.8 g, 56%); 1H-NMR (CDCl3) : 1.101.29 (2H, m), 1.381.52 (1H, m), 1.55 (2H, t, J = 7.2 Hz), 1.63 (2H, d, J = 6.6 Hz), 1.92 (2H, dt, J = 1.0, 12 Hz), 2.50 (2H, br-s), 2.72 (2H, t, J = 7.2 Hz), 2.82 (2H, d, J = 12 Hz), 3.61 (2H, s), 7.127.23 (5H, m); MS m/z: 218 (M+). 5.1.4. General procedure for the preparation of 6a6d 5.1.4.1. 4-Aminomethyl-1-benzylpipeidine 6a A solution of 5a (5.7 g, 26 mmol) in tetrahydrofuran (50 mL) was added dropwise to a suspension of lithium aluminium hydride (LiAlH4) (2.0 g, 52 mmol) in tetrahydrofuran (50 mL) at 0 C, and the mixture was heated at 45 C with stirring for 4 h. After cooling, water was added dropwise to the mixture at a temperature below 0 C to destroy the excess LiAlH4. After ltration, the ltrate was dried over anhydrous magnesium sulfate, and after evaporation in vacuo, the residue was distilled under reduced pressure to give 6a (5.0 g, 94%), b.p. 125 C/ 1.0 mmHg; 1H-NMR (CDCl3) : 1.28 (2H, dd, J = 4.0, 12 Hz), 1.30 (2H, br-s), 1.351.50 (1H, m), 1.72 (2H, d, J = 6.6 Hz), 1.92 (2H, dt, J = 2.0, 6.6 Hz), 2.64 (2H, d, J = 6.6 Hz), 2.87 (2H, d, J = 6.6Hz), 3.00 (2H, t, J = 6.6 Hz), 7.187.35 (5H, m); MS m/z: 142 (M+). The other piperidines (6b6d) were prepared in a similar manner. 5.2. Pharmacology 5.2.1. Radioligand binding assays 5.2.1.1. 5-HT4 receptor Male Hartley guinea-pigs (Japan SLC, Ltd., Shizuoka, Japan) were sacriced by cervical dislocation and the striatum was separated from each brain. The striatum was homogenized in 15 volumes of 50 mmol/L ice-cold HEPES buffer (pH 7.4) with Polytron PT-10 and then centrifuged at 35 000 g for 20 min. The resulting pellet was resuspended in the HEPES buffer and nally diluted to the appropriate concentration for assay (6 mg wet weight per assay tube). This suspension was used as the tissue preparation. Assay tubes contained 50 L of HEPES buffer or a solution of the test agents, 50 L solution of [3H]GR113808 (Amersham International, UK) to give a nal concentration of 0.1 nmol/L and 900 L of tissue preparation. Each tube was incubated for 30 min at 37 C and the reaction was terminated by rapid ltration through a Whatmann GF/B lter (presoaked in 0.01% v/v polyethyleneimine) followed by washing with 1 4 mL of ice-cold HEPES buffer. Then the lter was placed in 3 mL of scintillator and the radioactivity was determined by scintillation counting in a Beckman model LS3801 scintillation counter. Non specic binding was dened in the presence of unlabelled GR113808 to give a nal concentration of 1 mol/L. The IC50 value was determined by non-linear regression of the displacement curve, and the Ki value was calculated according to the formula (Ki = IC50/(1 + L/Kd)), where L is the concentration of radioligand and Kd is the dissociation constant of the radioligand. 5.2.1.2. 5-HT3 receptor [3H]Granisetron binding assays were performed according to the method of Nelson and Thomas [34]. Male Wistar rat (Japan SLC, Ltd., Shizuoka, Japan) cerebral cortex was homogenized in 20 volumes of 0.32 mol/L sucrose and the centrifuged at 1 000 g for 10 min. The supernatant was centrifuged at 40 000 g for 15 min. The pellet was suspended in 20 volumes of HEPES buffer (50 mmol/L, pH 7.4) and the suspension was incubated at 37 C for 10 min and centrifuged at 40 000 g for 15 min. The pellet was washed and centrifuged (40 000 g for 15 min). The nal pellet was resuspended in 30 volumes of HEPES buffer and used as tissue homogenate. The binding assay consisted of 50 mol/L of [3H]Granisetron, 50 L of displacing drugs and 900 L of tissue homogenate. Following a 30 min incubation at 25 C, the assay mixture was rapidly ltered under reduced pressure through Whatman GF/B glass lters which had been presoaked in 0.1% polyethyleneimie. Filters were washed immediately with 3 3 mL of ice-cold Tris-HCl buffer (50 mM, pH 7.4). ICS 205930 (100 mmol/L) was used for the determination of nonspecic binding. 5.2.1.3. D2 receptor [3H]Spiperone binding assays were performed according to the method of Crees et al. Male Wistar rat (Japan SLC, Ltd., Shizuoka, Japan) striatal membrane was homogenized in 100 volumes of ice-cold Tris-HCl buffer (50 mmol/L, pH 7.7) and centrifuged (500 g, 10 min, 0 C). The supernatant was centrifuged at 50 000 g for 15 min. The pellet was suspended in 100 volumes of ice-cold Tris-HCl buffer (50 mmol/L, pH 7.7) and recentrifuged (500 g, 10 min, 0 C). The nal pellet was resuspended in 150 volumes (50 mmol/L, pH 7.7) containing 120 mmol/L NaCl, 5 mmol/L KCl, 2 mmol/L CaCl2, 1 mmol/L MgCl2, 1.1 mmol/L ascorbic acid and 10 mol/L pargyline, and incubated at 37 C for 10 min.

988 A portion of this membrane suspension (900 mol/L) was placed in a tube, and 50 mol/L of either test compound or vehicle solution was added, followed by 50 L of [3H]Spiperone (40 Ci/mmol) at a nal concentration of 0.2 nmol/L. The tubes were incubated at 37 C for 20 min and ltered through Whatman GF/B glass lters, which were then washed three times with 3 mL of Tris-HCl buffer (50 mmol/L, pH 7.7). Sulpiride (100 mol/L) was used for the determination of nonspecic binding. The radioactivity trapped on the lters was measured by liquid scintillation spectrometry. 5.2.2. Contractile effects 5.2.2.1. 5-HT4 receptor agonism Male Hartley guinea-pigs (Japan SLC, Ltd., Shizuoka, Japan) were killed by cervical dislocation and the ascending colon (a 10 cm segment starting 1 cm from the caecum) was removed. The longitudinal muscle layer was separated from the underlying circular muscle and divided into four segments of about 2.5 cm. Four muscle strip preparations were individually mounted vertically for isotonic measurement into a tissue bath containing 10 mL Tyrode solution. Only 5-HT was tested in the Tyrode solution containing methysergide (1 mol/L) and granisetron (1 mol/L) to inhibit responses mediated by 5-HT2 and 5-HT1-like and 5-HT3 receptors, respectively. This solution was kept at 37 C and gassed with 95% O2, 5% CO2. The strips were subjected to a preload of 1 g and allowed to stabilize for 20 min. After stabilization, the response of the longitudinal muscle to 30 mol/L methacholine was measured. Agonist concentration-effect curves were constructed using sequential dosing, leaving 15 min between doses. A 15 min dosing cycle was required to prevent desensitization. The agonist was left in contact with a preparation until the response had reached a maximum, the preparation was washed. Forty minutes was left between the determination of concentration-effect curves. GR113808 (10 nmol/L) were incubated for 10 min before repeating agonist concentration-effect curves. After each determination of concentration-effect curve, 30 mol/L of methacholine was added to the tissue bath again. All responses were expressed as a percentage of the mean of the two contractions induced by 30 mol/L methacholine. The EC50 value, the concentration causing 50% of the maximal response, was determined by linear regression analysis. 5.2.2.2. 5-HT4 receptor antagonism 5-HT4 receptor antagonism was expressed in the form of pKB value on the contractile response to 5-MeOT. 5.3. Construction of receptor model 5.3.1. Sequence The sequence in this study, except for the 5-HT4 receptor, was from the Swiss-Prot Protein Sequence Data Bank [35]. The coordinates of Bacteriorhodopsin (entry 1BRD) [36] were from the Protein Data Bank [37] at Brookhaven National Laboratory. Modelling was achieved with the molecular package SYBYL 6.2 [38]. The interactive modelling and display were performed on a Silicon Graphics IRIS INDIGO/Elan 4 000 computer. Five main steps were used: stretch of -helical structure where putative transmembrane region is longer than bacteriorhodopsin, amino acid substitution, local geometry optimization, docking of ligand, and side-chain rotation to minimize overlaps between helices. Some manual adjustments were made to remove bad steric interactions in geometry optimization. In Energy minimization procedure, the backbone was aggregated until the RMS gradient was less than 0.05 (kcal/mol 2). A dielectric constant of compound 24 was calculated where a cut off distance of 8 was used and no solvent molecules were included in the calculation. Acknowledgements We thank Mrs. F. Matsugaki for some of the biological results. We also thank Dr. M. Terasawa and Dr. K. Adachi for helpful discussion. References
[1] [2] [3] [4] [5] Alexander S.H., Peters J.A., TPiS Receptor and Ion Chanenel Nomenclature Supplement, Elsevier, 1998, pp. 4648. Eglen R.M., Wong E.F., Dumuis A., Bockaert J., Trends Pharmacol. Sci. 16 (1995) 391398. Craig D.A., Clarke D.E., J. Pharmacol. Exp. Ther. 252 (1990) 13781386. Elswood C.J., Bunce K.T., Humphrey P.P.A., Eur. J. Pharmacol. 196 (1991) 149155. Briejer M.R., Akkerman L.S., Meulemans A.L., Lefebvre R.A., Schuurkes L.J., Naunyn-Schmiedergs Arch. Pharmacol. 347 (1993) 464470. Rizzi C.A., Coccini T., Onori L., Manzo L., Tonini M., J. Pharmacol. Exp. Ther. 261 (1992) 412419. Langlois M., Zhang L., Yang D., Brmont B., Shen S., Manara L., Croci T., Bioorg. Med. Chem. Lett. 4 (1994) 14331436. Dumuis A., Bouhelal R., Sebben M., Cory R., Bockaert J., Mol. Pharmacol. 34 (1988) 880887. Flynn D.L., Zabroski D.L., Becker D.P., Nosal R., Villamil C.L., Gullikson G.W., Moummi C., Yang D., J. Med. Chem. 35 (1992) 14871491. Turconi M., Schiantarelli P., Borsini F., Rizzi C.A., Ladinsky H., Donetti A., Drugs Future 16 (1991) 10111026.

[6] [7] [8] [9]

[10]

989
[11] Clark R.D., Jahangir A., Langston J.A., Weinhardt K.K., Miller A.B., Leung E., Eglen R.M., Bioorg. Med. Chem. Lett. 4 (1994) 24772480. Lpez-Rodrguez M.L., Morcillo M.J., Benha B., Rosado M.L., J. Comput.-Aided Mol. Des. 11 (1997) 589599. Kuroita T., Sakamori M., Kawakita T., Chem. Pharm. Bull. 44 (1996) 756764. Kuroita T., Ikebe T., Murakami S., Takehara S., Kawakita T., Bioorg. Med. Chem. Lett. 5 (1995) 12451250. Tahara T., Hayano K., Murakami S., Fukuda T., Setoguti M., Ikeda K., Marubayasi N., Chem. Pharm. Bull. 38 (1990) 16091615. Iwanami S., Takasima M., Hirata Y., Hasegawa O., Usuda S., J. Med. Chem. 24 (1981) 12241230. Mitsunobu O., Synthesis (1981) 128. Kawakita T., Kuroita T., Murozono T., Hakira H., WO 95/26953 Chem. Abstr. 124 (1996) 14913. Fancelli D., Caccia C., Fornaretto M.G., McArthur R., Severino D., Vaghi F., Varasi M., Bioorg. Med. Chem. Lett. 6 (1996) 236266. Hibert M.F., Trumpp-Kallmeyer S., Bruinvels A., Hoack J., Mol. Pharmacol. 40 (1991) 815. Ijzerman A.P., Van Gallen P.M., Jacobson K.A., Drug Design Dev. 9 (1992) 4967. Trumpp-Kallmeyer S., Hoack J., Bruinvels A., Hibert M., J. Med. Chem. 35 (1992) 34483462. Yamamoto Y., Kamiya K., Terao S., J. Med. Chem. 36 (1993) 820825. Nordvall G., Hacksell U., J. Med. Chem. 36 (1993) 976976. [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] Glennon R.A., Dukat M., Westkaemper R.B., Ismaiel A.M., Izzarelli D.G., Parker E.A., Mol. Pharmacol. 49 (1996) 198206. Teeter M.M., Froimowitz M.B., Durand C.J., J. Med. Chem. 37 (1996) 28742888. Kotoyianni M., Deweese C., Penzotti J.E., Lybrand T.P., J. Med. Chem. 39 (1996) 44064420. Kyte J., Doolittle R.F., J. Mol. Biol. 157 (1982) 105132. Begonia Y.H., Karschin A., Branchek T., Davidson N., Lester H.A., FEBS. Lett. 312 (1992) 259262. Johnson M.P., Loncharich R.J., Baez M., Nelson D.L., Mol. Pharmacol. 45 (1993) 277286. Almaula N., Ebersole B.J., Zhang D., Weinstein H., Sealfon S.C., J. Biol. Chem. 271 (1996) 1467214675. Almaula N., Ebersole B.J., Ballesteros J.A., Weinstein H., Sealfon S.C., Mol. Pharmacol. 50 (1996) 3442. Choudhary M.S., Craigo S., Roth B.L., Mol. Pharmacol. 43 (1993) 755761. Nelson D.R., Thomas D.R., Biochem. Pharmacol. 38 (1989) 16931695. Bairoch A., Boeckmann B., Nucleic Acids Res. 20 (1992) 20192022. Henderson R., Baldwin J.M., Ceska T.A., Zemlin F., Beckmann E.D., J. Mol. Biol. 213 (1990) 899929. Bernstein F.C., Koetzle T.F., Williams G.B., Meyer E.F. Jr., Brice M.D., Rodgers J.R., Kennard O., Shimanouchi T., Tasumi M., J. Mol. Biol. 112 (1977) 535542. Tripos E & S, 6548 Clayton Road, St. Louis, MO 62117. SYBYL molecular modelling software.

[12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24]

[38]

Eur. J. Med. Chem. 34 (1999) 991996 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

991

Preliminary communication

Synthesis and antioxidant activity of new tetraarylpyrroles

Jacques Lehuda, Bernard Fauconneaub, Laurence Barrierb, Marina Ourakowa, Alain Pirioub, Jean-Michel Vierfonda*
a

Laboratoire de Chimie Organique, Facult de Pharmacie, BP 199, 34, rue du Jardin des Plantes, F-86005 Poitiers, France b Centre dtudes et de Recherche sur les Xnobiotiques, EA 1223, Facult de Pharmacie, BP 199, 34, rue du Jardin des Plantes, F-86005 Poitiers, France (Received 26 January 1998; revised 18 January 1999; accepted 11 March 1999)

Abstract The synthesis and in vitro antioxidant activity of 17 new tetraarylpyrroles are investigated by 2 tests highly documented in the literature: capability to prevent Fe2+-induced lipid peroxidation on microsomes, which is a membrane preparation rich in polyunsaturated fatty acids, and direct scavenging effect on a stable free radical, 1,l-diphenyl-2-picryl-hydrazyl (DPPH). For the Fe2+-induced microsomal lipid peroxidation system, the results show that molecules which possess 2-pyrazinyl or 2-pyridyl in the 3- and 4-positions on the pyrrole ring are the most efficient. Introduction of methoxy groups on the phenyl ring in the 2- and 5-positions increases the effects but the higher activity is obtained with 2-furyl or 2-thienyl. The only compounds which possess a direct scavenger effect on trapping the stable free radical DPPH are those which have 2-pyridyl in the 3- and 4-positions and 2-furyl or 2-thienyl in the 2- and 5-positions. 1999 ditions scientiques et mdicales Elsevier SAS tetraarylpyrrole / antioxidant activity / radical scavenging effect / microsomes / DPPH

1. Introduction Formation of reactive oxygen species and the ensuing oxidation of biological molecules is a well recognized mechanism of tissue damage in many pathological situations, such as, inammation, stroke, acute myocardial infarction and the subsequent reperfusion phase. Numerous natural or synthetic antioxidant compounds have been tested with success in various disease models as well as in clinics [1]. In the literature, it is well known that pyrrolic compounds such as polypyrroles, poly- or hetero-arylpyrroles present an electronic delocalization, conferring to these molecules electric conductor and oxidizable properties [2, 3]. Furthermore, pyrrolic structures such as benzoylpyrrole-3-acetic acids were tested in vitro by examining their effects on lipid peroxidation using rat hepatic microsomes and as hydroxyl radical scavengers [4].

Recently we reported results on six pyrrolic compounds synthesized in our laboratory showing interesting antioxidant activities [5]. Consequently, it seems relevant to test large series of molecules bearing different substituents on the pyrrole ring: phenyl rings (with or without methoxy groups) and aromatic heterocycle rings in order to have an electronic delocalization and consequently, potent antioxidant activities. 2. Chemistry The synthesis of compounds 2aq (table I) was realized in two steps (gure 1): metalation of methylazine from lithium diisopropylamine and condensation with an aromatic nitrile to obtain an imine-enamine 1 moisture sensitive (rst step), and oxidation of imine-enamine l by Pb(OAc)4 to give tetraarylpyrroles 2 (second step) [6]. The structure is given in table I. 2.1. Antioxidant activity The antioxidant potency of compounds 2aq was assessed by two tests commonly used in the literature

*Correspondence and reprints

992
Table I. Structure and antioxidant effects of tetraarylpyrroles. IC50 values are expressed as means SEM.

R 2a 2b 2c 2d 2e 2f 2g 2h 2i 2j 2k 2l 2m 2n 2o 2p 2q Trolox
a

Ar1 2-pyrazinyl 2-pyrazinyl 2-pyrazinyl 2-pyrazinyl 2-pyrazinyl 2-quinoxalinyl 2-pyridyl 4-pyridyl 2-pyridyl 2-pyridyl 2-pyridyl 2-pyridyl 4-pyridyl 2-pyridyl 2-pyridyl 2-pyridyl 2-pyridyl

Ar2 phenyl 2-pyridyl 4-pyridyl 2-furyl 2-thienyl phenyl phenyl phenyl 2-methoxyphenyl 3-methoxyphenyl 4-methoxyphenyl 3,4,5-methoxyphenyl 4-methoxyphenyl 2-furyl 2-furyl 2-thienyl 3-thienyl

Yields (%) 90 77a 52a 98a 90a 59a 50b 37b 53c 36c 67c 25c 29c 17.5b 88d 53b 32b
a

IC50 microsomes (M) 320 16 > 1 000 > 1 000 19.0 9.6 8.0 1.1 > 1 000 67.2 6.7 > 1 000 28 3 63 4 12.3 4.0 26.2 8.5 100 12 11.0 1.9 61.1 5.0 2.7 1.1 12.5 2.2 5.0 0.3

IC50 DPPH (M) > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 200 50 > 1 000 410 127 > 1 000 10.1 0.5

H H H H H H H H H H H H H H CH3 H H

: based on isolated imine-enamine. b: based on aromatic nitrile. c: with TMEDA and heating, based on aromatic nitrile. d: methylation from 2n.

bearing on this topic: capability to prevent Fe2+-induced lipid peroxidation on microsomes, which is a membrane preparation rich in polyunsaturated fatty acids, and direct

scavenging effect on a stable free radical, 1,l-diphenyl-2picryl-hydrazyl (DPPH). The results were compared to those observed with Trolox (water soluble vitamin E analogue), a reference antioxidant. 3. Results and discussion 3.1. Effect on lipid peroxidation Three distinctive steps can be distinguished in lipid peroxidation: initiation, propagation and termination reactions. Lipid peroxidation may be induced, for example, by radical species which are sufficiently reactive to abstract a hydrogen atom from the unsaturated fatty acids. This is the starting point for the lipid radical chain propagation reaction. Thus, many molecules of lipids may be oxidized to lipid hydroperoxides for every initiation event. The propagation cycle is broken by termination reactions which result in the destruction of free radicals. Termination reactions occur when two radical species combine to form non-radical nal products [7].

Figure 1. Synthesis of tetraarylpyrroles.

993 A chain breaking antioxidant has the ability to donate hydrogen radicals and contributes to stop the chain reaction. It reduces the primary radical by a one electron reduction to a non-radical chemical species and, as a consequence, is transformed into an oxidized antioxidant radical. However, this antioxidant must be more than just a hydrogen atom donor; it needs also to form a radical of such low reactivity that no further reactions with lipids can occur. Peroxyl radicals are the major chain-propagating species in the process of lipid peroxidation in membranes [8]. Thus, a standard test for antioxidants is the action of a substance in inhibiting peroxidation of membranes such as microsomes. The thiobarbituric acid (TBA) test is widely used in microsomal systems but it is essential to ensure that the apparent antioxidant effect of an added compound is not due to the interference with the TBA test itself [9]. Consequently, relevant controls are necessary to check the absence of such an interference. Peroxidation can be accelerated by adding iron salts to membranes, e.g. Fe2+. In metal-ion dependent systems, an added antioxidant might act not only by scavenging peroxyl radicals but also by binding iron ions and stopping them from accelerating peroxidation. However, these two possibilities can be distinguished easily. The efficient compounds tested in this study acted as chain breaking antioxidants because at low concentrations, a lag period into the peroxidation process was observed (data not shown), corresponding to the time taken for the antioxidant to be consumed, whereas metal-binding antioxidants give a constant inhibition throughout the reaction [9]. The effects of the compounds on lipid peroxidation induced by FeCl2 are shown in table I. When Ar2 is constant, i.e. phenyl ring, and if Ar1 is 2-pyridyl (2g), the activity of the molecule is higher than that observed with Ar1 = 2-pyrazinyl (2a). The activity disappears if Arl is 4-pyridyl (2h) or 2-quinoxalinyl (2f). If the phenyl ring in the Ar2 position is substituted, i.e. 4-methoxyphenyl, the higher activity of 2-pyridyl compared with that of 4-pyridyl in Arl is conrmed (2k versus 2m). When they are in the Arl position, the tested substituents can be classied as follows: 2-pyridyl > 2-pyrazinyl > 4-pyridyl > 2-quinoxalinyl. Consequently, molecules with Ar1 = 2-pyridyl or 2-pyrazinyl show the greatest interest and the synthesis of molecules with different groups in Ar2 positions is performed to improve the antioxidant effect. When methoxy groups are added in the 2, 3, or 4 position on the phenyl ring in Ar2, the antioxidant activity is increased (2i, 2j, 2k versus 2g), with a major activity for 4-methoxyphenyl (2k). This activity may be correlated to the introduction of electron donor substituents which stabilize the generated radical during oxidation. The electronic mesomeric donor effect is of the same type when methoxy groups are added in the 2 or 4 position on the Ar2 phenyl ring (i.e. 2i or 2k), but 2k has a better activity, probably due to the steric bulk observed with 2-methoxyphenyl (2i). The substituent 3-methoxyphenyl has a lower electronic effect, and consequently is less active than 2i and 2k. But methoxy tri-substitution in the 3, 4, 5 positions on the phenyl ring (2l) decreases 2 times the activity of 2k. The antioxidant tests performed with compounds bearing heterocyclic substituents in the Ar2 position make it possible to classify them into two groups: those with 2and 4-pyridyl which are inactive (2b and 2c) and those with pentagonal heterocycles (2d, 2e, 2n, 2p, 2q) which seem to be of major importance because the antioxidant activity is at least as great as that of the more active phenylmethoxylated compounds. The pentagonal substituents have a donor mesomeric effect on a conjugated system, whereas pyridyl has an attractor effect. Substituent 3-thienyl (2q) is less active than 2-thienyl (2p). On microsomes oxidized by Fe2+, 2p (Arl = 2-pyridyl and Ar2 = 2-thienyl; IC50 = 2.7 1.1 M) is the most efficient antioxidant. Its activity is higher than that obtained with Trolox (5.0 0.3 M), a reference antioxidant. These results show the major importance of the pyrrol ring in antioxidant effects since methylation of 2n (i.e. 2o) decreases about 6 times the activity of this compound. This may be explained by the difficulty in forming a free radical with a pentasubstituted pyrrole. 3.2. Effect on DPPH The DPPH test provided information about the reactivity of the tested compounds with a stable free radical. Because of its odd electron, the DPPH radical showed a strong absorption band at 515 nm in visible spectroscopy (a deep purple colour). As this electron is paired off in the presence of a free radical scavenger, absorption vanishes and the resulting decoloration is stoichiometric with respect to the number of electrons taken up. This bleaching of DPPH absorption, which occurs when the odd electron of the radical is paired, is thus representative of the capacity of the compounds to scavenge free radicals independently of any enzymatic activity. Only 2p and 2n have a radical scavenging activity, with a major activity for 2n, whereas that of Trolox is considerably higher. It means that these compounds possess a direct trapping effect (table I). It is noteworthy that these molecules (i.e. 2p and 2n) also have a good activity with a microsome system.

994 Nevertheless, 2e has no activity on DPPH, while it has good activity on Fe2+-induced microsomal lipid peroxidation. This result shows that this molecule doesnt possess a direct scavenger effect, at least on this radical. This effect is improved when a 2-pyridyl is substituted in the Ar1 position (2e versus 2p). In conclusion, our study provides evidence that several tetraarylpyrroles exhibit interesting antioxidant properties mainly expressed by their capacity to inhibit Fe2+induced microsomal lipid peroxidation. These effects may be useful in the treatment of pathologies in which free radical oxidation plays a fundamental role. The early trials on animals are encouraging, since toxicity tests have been performed on female Swiss mice (1820 g, Iffa-Credo, lArbresle, France). At the present time, the only compound tested is 2a and is responsible for no mortality when orally administered at the dose of 1g/kg (batch of 10 animals). 4. Experimental protocols Melting points were measured by using a Ker apparatus and are uncorrected. The 1H-NMR spectra were recorded on Varian EM 360 and Bruker 200 A.C. spectrometers. IR spectra were realized on a Unicam SP 1100 and on an ATI Mattson Genesis spectrometer. Elemental analyses were performed on a Perkin Elmer 240 apparatus. Acetonitrile was dried over Na2SO4. Tetrahydrofuran (THF) was dried and extemporaneously distillated over sodium. Diisopropylamine and tetramethylethylenediamine (TMEDA) were dried over BaO and distillated. Butyl lithium used was a Merck 1.6 M solution in hexane. The synthesis of compounds 2ah was previously described [6]. The 2nq pyrroles were prepared according to the precedent method without isolation of the imine-enamine intermediate. 4.1. General procedure for the methoxylated phenylpyrroles (2im) preparation of TMEDA 1/4; v/v) was added at 40 C. The mixture was heated for 2 h to reuxing THF afterwards. After cooling to 20 C, 1 mL of water was added, with strong and quick stirring, the mixture was stored on anhydrous sodium sulphate. After ltration and evaporation of solvents, the residue (non-isolated imine-enamine) was dissolved in acetonitrile (100 mL) plus chloroform (50 mL) and cooled to 40 C. The lead tetraacetate (2.22 g; 5 mmol) was introduced at 40 C with stirring. After 15 min, the mixture was heated to 20 C and stirred for 1 h. Then, 12 mL of aqueous saturated sodium carbonate was added. The precipitate was washed out with acetonitrile and chloroform. The organic phase was dried over anhydrous sodium sulphate, evaporated and puried by chromatography on a silica gel column, elution with ethyl acetate. 4.1.1. 1H-2.5-di-(2-methoxyphenyl)-3.4-di-(2-pyridyl) pyrrole 2i Yellow powder (2.29 g, 53%); m.p. 179 C; IR (KBr) (cml): 3 480, 1 580, 783, 751; 1H-NMR (CDCl3) 10.2 (1H, s, NH), 8.4 (2H, d, J = 45 Hz), 7.56.6 (14H, m), 3.7 (6H, s); Anal. C28H23N3O2 (C, H, N). 4.1.2. 1H-2.5-di-(3-methoxyphenyl)-3.4-di-(2-pyridyl) pyrrole 2j Pale yellow powder (1.55 g, 36%); m.p. 165 C; IR (KBr) (cm1): 3 389, 1 608, 1 558, 871, 852, 784, 748; 1 H-NMR (CDCl3) 8.8 (1H, s, NH), 8.45 (2H, m), 7.56.6 (14H, m), 3.6 (6H, s); Anal. C28H23N3O2 (C, H, N). 4.1.3. 1H-2.5-di-(4-methoxyphenyl)-3,4-di-(2-pyridyl) pyrrole 2k Pale yellow powder (2.90 g, 67%); m.p. 128 C; IR (KBr) (cml): 3 210, 1 590, 834, 792, 744; 1H-NMR (CDCl3) 8.7 (1H, s, NH), 8.4 (2H, d, J = 45 Hz), 7.56.6 (14H, m), 3.7 (6H, s); Anal. C28H23N3O2 (C, H, N). 4.1.4. 1H-2.5-di-(3,4,5-trimethoxyphenyl)-3,4-di-(2pyridyl) pyrrole 2l Brown powder (1.39 g, 25%); m.p. > 260 C; IR (KBr) (cml): 3 357, 1 584, 842, 793, 747, 699; 1H-NMR (CDCl3) 9.0 (1H, s, NH), 8.5 (2H, d, J = 45 Hz), 7.66.9 (6H, m), 6.6 (4H, s), 3.8 (6H, s), 3.65 (12H, s); Anal. C32H31N3O6 (C, H, N). 4.1.5. 1H-2.5-di-(4-methoxyphenyl)-3,4-di-(4-pyridyl) pyrrole 2m White powder (1.25 g, 29%); m.p. > 260 C; IR (KBr) (cm1): 3 350, 1 550, 900, 645; 1H-NMR (CDCl3) 8.75 (1H, s, NH), 8.2 (4H, d, J = 45 Hz), 7.156.6 (12H, m), 3.7 (6H, s); Anal. C28H23N3O2 (C, H, N).

Imines were prepared according to the general procedure [6]. 13.8 mL (22 mmol) of a solution of n-butyllithium (1.6 M) in hexane was added, via syringe, to a stirred solution of diisopropyl-amine (2.2 g; 22 mmol) solved in 30 mL of dry THF plus 10 mL of dry TMEDA, under 0 C nitrogen atmosphere. After 20 min at 0 C, the solution was cooled to 40 C and a solution of 2-methylpyrazine (1.88 g; 20 mmol) in 5 mL of THF was slowly added. Then, the solution was stirred for 45 min at 40 C and a solution of methoxylated aromatic nitrile (10 mmol) in 5 mL of mixed solvent (THF 3/4 plus

995 4.1.6. 1H-2.5-di-(2-furyl) 3,4-di-(2-pyridyl)pyrrole 2n Pale yellow powder (0.62 g, 17.5%); m.p. 172 C; IR (KBr) (cml): 3 430, 1 589, 1 562, 803, 790, 733; 1 H-NMR (CDCl3) 9.4 (1H, s, NH), 8.6 (2H, m), 7.56.9 (8H, m), 6.3 (4H, m); Anal. C22H15N3O2 (C, H, N). 4.1.7. 1H-1-methyl-2.5-di-(2-furyl)-3,4-di-(2-pyridyl) pyrrole 2o Brown powder; m.p. 180 C; IR (KBr) (cml): 3 134, 1 587, 1 561, 909, 750; 1H-NMR (CDCl3) 8.4 (2H, m), 7.56.8 (8H, m), 6.3 (4H, m), 3.6 (3H, s); Anal. C23H17N3O2 (C, H, N). 4.1.8. 1H-2.5-di-(2-thienyl)-3.4-di-(2-pyridyl)pyrrole 2p Pale yellow powder (2.04 g, 53%); m.p. 230 C; IR (KBr) (cm1): 3 250, 1 590, 840, 750, 690; 1H-NMR (CDCl3) 8.7 (1H, s, NH), 8.5 (2H, m), 7.56.8 (12H, m); Anal C22H15N3S2 (C, H, N). 4.1.9. 1H-2.5-di-(3-thienyl)-3.4-di-(2-pyridyl)pyrrole 2q Pale yellow powder (1.23 g, 32%); m.p. 243 C; IR (KBr) (cm1): 3 420, 1 590, 1 560, 865, 782, 749, 693, 616; 1H-NMR (CDCl3) 9.6 (1H, s, NH), 8.4 (2H, m), 7.56.8 (12H, m) Anal. C22H15N3S2 (C, H, N). 4.2. Procedures concerning antioxidant activities 4.2.1. Chemicals All chemicals were purchased from the Sigma Chemical Co. (St. Louis, MO, USA) except Trolox which was obtained from Aldrich (St Quentin Fallavier, France). The powders corresponding to the molecules 2aq were kept under nitrogen. For the biological tests, the compounds were dissolved extemporaneously in ethanol and the different dilutions were performed in ethanol. 4.2.2. Inhibitory effect on lipid peroxidation Male Sprague-Dawley rats (200250 g, Iffa-Credo, lArbresle, France) had been deprived of food overnight (16 h). The rats were anaesthetized by inhalation of ethyl ether. Livers were perfused, rapidly isolated and minced thoroughly with scissors. The minced tissue was washed with ice-cold 0.15 M KCl and homogenized with 5 volumes of ice-cold 0.15 M KCl, using a Teon-glass potter homogenizer (3 000 rpm for 2 min). Microsomal fractions were isolated in Tris-HCl 0.05 M/KCl 0.15 M, pH 7.4, by removal of the nuclear fraction at 800 g for 15 min, removal of the mitochondrial fraction at 15 000 g for 15 min and sedimentation at 105 000 g for 30 min. Pellets were washed twice in Tris-HCl 0.05 M/KCl 0.15 M, pH 7.4 buffer by centrifugation, with subsequent sedimentation at 125 000 g for 15 min. Microsome pellets were resuspended in the same buffered solution at 5 mg protein/mL and stored at 80 C for a maximum of 1 month. The protein content was determined by the method of Bradford [10], using bovine albumin as a standard. For the test, microsomal fractions were thawed just before use and were diluted with 0.05 M Tris-HCl, pH 7.4, containing 0.15 M KCl. The nal protein concentration in the incubation mixture amounted to 0.75 mg/mL. Microsomes were pre-incubated with different concentrations of compounds in a shaking water-bath at 37 C for 10 min. Then, lipid peroxidation was initiated with 10 M FeCl2 and the samples were incubated at 37 C for 30 min. After action of thiobarbituric acid, the absorbance was measured at 532 nm for thiobarbituric acid reactive substance (TBARS) determination [11], versus control containing the same quantity of ethanol but without the compound studied. The measurements were performed in triplicate. The inhibition of lipid peroxidation was expressed as a percentage and the inhibition concentration 50% (IC50) was obtained from the inhibition curve by graphical determination. 4.2.3. Radical scavenging effect Free radical scavenging capacity of the compounds was determined using DPPH [12]. An ethanol DPPH solution (0.1 mM) was mixed with different concentrations of compound and the absorbance change at 515 nm was measured 10 min later with a spectrophotometer (Uvikon 940, Kontron) versus control containing the same quantity of ethanol, but without compound studied. The measurements were performed in triplicate. The inhibition of coloration was expressed as a percentage and the IC50 was obtained from the inhibition curve, by graphical determination. Acknowledgements The authors wish to thank R. Pontcharraud and S. Mairesse-Lebrun for excellent technical assistance. References
[1] [2] [3] [4] [5] Rice-Evans C.A., Diplock A.T., Free Rad. Biol. Med. 15 (1993) 7796. Niziurski-Mann R.E., Scordilis-Kelley C., Tae-Lane L., Carlin R.T., J. Am. Chem. Soc. 115 (1993) 887891. Demopoulos V.J., Rekka E., J. Pharm. Pharmacol. 46 (1994) 740744. Demopoulos V.J., Rekka E., Retsas S., Pharmazie 45 (1990) 403407. Fauconneau B., Lehuede J., Karge E., Vierfond J.M., Piriou A., Klinger W., Exp. Toxicol. Pathol. 48 SII (1996) 116120.

996
[6] [7] [8] [9] Lehud J., Mettey Y., Vierfond J.M., Synth. Comm. 26 (1996) 793802. Gutteridge J.M.C., Clin. Chem. 41 (1995) 18191828. Halliwell B., Gutteridge J.M.C., Free Radicals in Biology and Medicine, second edition, Clarendon Press, Oxford, UK, 1989. Halliwell B., Free Rad. Res. Com. 9 (1990) 132. [10] [11] [12] Bradford M.M., Anal. Biochem. 72 (1976) 248254. Wilbur K.M., Bernheim F., Shapiro O.W., Arch. Biochem. Biophys. 24 (1949) 305313. Tamura A., Sato T., Fuhu T., Chem. Pharm. Bull. 38 (1990) 255257.

Eur. J. Med. Chem. 34 (1999) 9971001 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

997

Preliminary communication

Synthesis and anti-inammatory activity of N-substituted 2-oxo-2H-1benzopyran-3-carboxamides and their 2-iminoanalogues

Igor E. Bylov*, Maksym V. Vasylyev, Yaroslav V. Bilokin
Department of Organic Chemistry, Ukrainian Academy of Pharmacy, 310002 Kharkov, Ukraine (Received 12 November 1999; revised 6 May 1999; accepted 7 May 1999)

Abstract A series of N-arylsubstituted 2-imino-2H-1-benzopyran-3-carboxamides 3a and b and 2-oxo-2H-1-benzopyran-3-carboxamides 4ah were synthesized and evaluated for their anti-inammatory activity in carrageenan-induced rat paw oedema assays and in acetic acid-induced peritonitis tests in albino rats. The resulting products were found to be active anti-inammatory agents and their effects were comparable to that of piroxicam as the reference compound. In the consideration of the efficacy of the compounds in these assays, 2-imino/oxo-2H-1-benzopyran-3-carboxamides 3a and b and 4ah were further studied at graded doses for their acute toxicity (ALD50) in albino mice and were essentially non-toxic at the highest dose tested. 1999 ditions scientiques et mdicales Elsevier SAS coumarins / benzopyrans / amides / Knoevenagel condensation / anti-inammatory activity

1. Introduction Compounds comprising a coumarin (2-oxo-2H-1benzopyran) backbone have a wide range of biological activities. Thus, among the natural and synthetic coumarin derivatives there are compounds possessing antimicrobial [1], antitumour [2], antiviral [3] and other [4] activities. Moreover, studies of the hydroalcohol extract of Justicia pectoralis (Eha) and its main constituents, coumarin (Cou) and umbelliferone (Umb), showed analgesic and anti-oedema activities on acetic acid-induced writhing in mice and on the carrageenan end dextran paw oedema in rats [5]. The Eha, Cou and Umb presented a signicant anti-oedema effect in the carrageenan model but only Cou decreased the rat paw volume in the dextran model. Anti-inammatory activity of coumarins isolated from Santolina oblongifolia Boiss was also reported [6]. The isolated coumarins identied as 7-methoxycoumarin (herniarin), 6,7-dihydroxycoumarin (aesculetin), 6-methoxy-7-glucosidylcoumarin (scopolin), and 6-hydroxy-7-methoxycoumarin (scopoletin) showed marked activity as inhibitors of eicosanoid-release from
*Correspondence and reprints Present address: Department of Organic Chemistry, The Weizmann Institute of Science, Rehovot 76100, Israel.

ionophore-stimulated mouse peritoneal macrophages. It was also revealed [7] that compounds containing a benzopyran moiety were potent and selective inhibitors of cyclooxygenase (COX). Of the 3-substituted coumarin derivatives, our attention was called to their N-substituted amide derivatives, since marked anti-inammatory activity of structurally related N-substituted amides of 4-hydroxy-2-quinolone-3-carboxylic acids has been reported [8]. Also taking into consideration the fact that such anti-inammatory drugs as mefenamic and meclofenamic acids [4] constitute derivatives of aromatic amino acids, N-substituted coumarin-3-carboxamide derivatives containing aromatic amino acid residues were selected as targets for our anti-inammatory studies. 2. Chemistry The general synthetic strategy employed to prepare N-substituted 2-imino/oxo-2H-1-benzopyran-3-carboxamide derivatives was based on Knoevenagel condensation [9, 10] of active methylene compounds. As shown in gure 1, 2-imino-2H-1-benzopyran-3-N-R-carboxamides 3a and b were prepared by condensing N-substituted cyanoacetamides 1a and b and salicylic aldehyde 2 to form the expected imino compounds using piperidine as

998 condensed with an equivalent amount of salicylic aldehydes 2af to produce the desired coumarin derivatives 4ch utilizing piperidine as a catalyst in ethanol at room temperature. An alternative approach for synthesis of N-substituted coumarin-3-carboxamides of type 4, based on rearrangements of 2-imino-2H-1-benzopyran-3carboxamides under the action of anthranilic acid as N-nucleophile, has been recently developed [14] in our laboratory. Physicochemical data for the compounds 3a and b and 4ah are given in table I.
Figure 1. Synthesis of N-arylsubstituted 2-imino/oxo-2H-1benzopyran-3-carboxamides 3a and b and 4a and b.

3. Pharmacology The compounds synthesized were evaluated for their anti-inammatory activity in carrageenan-induced rat paw oedema assays [15] and in acetic acid-induced peritonitis tests in albino rats [16]. 2-Imino/oxo-2H-1benzopyran-3-carboxamides 3a and b and 4ah were further studied for their acute toxicity [17]. Piroxicam was used as a control compound. 4. Results and discussion Pharmacological results on anti-inammatory activities and acute toxicity of the benzopyran derivatives 3a and b and 4ah and piroxicam are summarized in table II. In carrageenan-induced rat paw oedema assays, the compounds 4g, 3b, 4b, and 4h were found to be the most active anti-inammatory agents at the graded dose 10 mg/kg po and exhibited 54 6.50%, 51 5.05%, 48 6.51%, and 47 7.13% inhibition of inammation, respectively. Their effects were comparable to that of

a catalyst in ethanol at room temperature [11, 12]. 2-Oxo2H-1-benzopyran-3-N-R-carboxamides 4a and b were obtained by acidic hydrolysis of the corresponding imino analogues 3a and b employing a mixture of ethanol/ water/ 32% hydrochloric acid and reuxing for 1 h. The synthesis of coumarin-3-(N-2-carboxyphenyl) carboxamides 4ch was carried out as shown in gure 2. The precursor 2-carboxymalonanilic acid ester 5 [13] was

Figure 2. Synthesis of N-arylsubstituted 2-oxo-2H-1-benzopyran-3-carboxamides 4ch.

Table I. Physicochemical data of the synthesized N-arylsubstituted 2-imino/oxo-2H-1-benzopyran-3-carboxamides 3a and b and 4ah. Compound 3a 3b 4a 4b 4c 4d 4e 4f 4g 4h
a

R1 H H H H H 6-OCH3 8-OCH3 6-NO2 6-Cl 8-allyl

R 2-CO2CH3 4-CO2C2H5 2-CO2CH3 4-CO2C2H5 2-CO2H 2-CO2H 2-CO2H 2-CO2H 2-CO2H 2-CO2H

Molecular formula C18H14N2O4 C19H16N2O4 C18H13NO5 C19H15NO5 C17H11NO5 C18H13NO6 C18H13NO6 C17H10N2O7 C17H10ClNO5 C20H15NO5

Yield (%) 74 82 87 85 67 59 67 63 71 54

Recryst. solvent i-PrOH i-PrOH i-PrOH i-PrOH AcOH BuOH BuOH BuOH i-PrOH MeCN

M.p. (C) 137138 222224 205208 246247 275276a 124125 140142 147150 232235 249251

Lit. [21] m.p. for 4c: 279 C.

999
Table II. Anti-inammatory activities and acute toxicity of the benzopyran derivatives 3a and b and 4ah and piroxicam. Acute toxicity Compound Approximate LD50 (mg/kg) (in mice) po 3a 3b 4a 4b 4c 4d 4e 4f 4g 4h Piroxicam
a

Anti-inammatory activity carrageenan-induced paw oedema acetic acid peritonitis

ip > > > > > > > > > > > 700 700 700 700 700 700 700 700 700 700 1 000a 19 2.11 51 5.05 44 7.21 48 6.51 38 1.01 30 6.96 41 6.42 35 7.11 54 6.50 47 7.13 57 6.61

(Mean % inhibition SE) 10 mg/kg po 40 6.43b 32 6.89 39 5.09 35 1.02 42 7.20 18 4.63 32 7.01b 29 5.58b 35 4.88b 31 3.31b 29 7.24b

> 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 > 1 000 360a

Data from ref. [22]; bP < 0.05 Students t test versus controls.

piroxicam, the reference compound, which showed 57 6.61% inhibition at the same dose level. At the same dose level (10 mg/kg po), in acetic acid peritonitis tests, all compounds exhibited moderate to good anti-inammatory activity. The most active compounds 4c, 3a, and 4a showed 42 7.20%, 40 6.43%, and 39 5.09% inhibition of inammation, respectively. In this test, piroxicam revealed a protection of 29 7.24%. All tested compounds were essentially non-toxic at the highest dose graded.

5. Conclusion The products synthesized were found to be active anti-inammatory agents and their effects were comparable to that of piroxicam as the reference compound. The most active compounds 3b, 4a, 4b and 4g have been marked for further detailed pharmacological studies to be evaluated for COX-2 and COX-1 inhibition in microsomal and cellular assays.

6. Experimental protocols 6.1. Chemistry Melting points (C) were measured with a Bchi melting point apparatus and were uncorrected. Thin layer

chromatography (TLC) was performed on aluminium sheets precoated with silica gel (Merck, Kieselgel 60 F-254). 1H-NMR spectra were recorded on a Varian WXR-400 spectrometer in DMSO-d6 using TMS as an internal standard (chemical shifts in ppm), but a study on isomerization of benzopyran-2-imines in DMSO-d6 has to be mentioned [18]. OCallaghan et al. [18] revealed that when unsubstituted 2-imino-2H-1-benzopyran-3-carboxamide was dissolved in DMSO-d6, NMR spectra showed that a mixture of 2-imino-2H-1benzopyran-3-carboxamide and the isomeric 2-cyano-3(2-hydroxyphenyl)prop-2-ene-1-carboxamide was present and other benzopyran-2-imines behaved similarly and the degree of isomerization varied considerably, depending on the nature and position of the substituents presented. In our case, isomerization of N-arylsubstituted 2-imino-2H-1-benzopyran-3-carboxamides 3a and b did not occur in dimethyl sulfoxide-d6 at room temperature and only starting materials were present. Mass spectra (MS) were obtained with a Finnigan MAT-4615B spectrometer at an ionization potential of 70 eV. Infrared spectra (IR) were recorded in KBr pellets on an IBM 486 PC computer-controlled Specord M-80 spectrometer. Elemental analyses were performed at the Microanalysis Laboratory, Kharkov State University, and the combustion analyses of all compounds synthesized indicated by the symbols of the elements were within 0.4% of theoretical values. The N-substituted cyanoacetamides 1a and b, which are key intermediates for synthesis of the benzopyran-2-imines 3a and b, were prepared according

1000 to reported methods [19, 20] from ethyl cyanoacetate and methyl anthranilate or ethyl 4-aminobenzoate. 6.1.1. N-arylsubstituted 2-imino-2H-1-benzopyran-3carboxamides 3a and b To a well-stirred solution of N-substituted cyanoacetamides 1a and b (4 mmol) in 15 mL of ethanol, was added an equivalent amount of salicylic aldehyde 2 and a few drops of piperidine as a catalyst. The reaction mixture was stirred at room temperature for 2 h. The products, which precipitated in the course of the reactions were ltered and recrystallized from the proper solvent. Yields and physicochemical data of the synthesized N-substituted 2-imino-2H-1-benzopyran-3-carboxamides 3a and b are listed in table I. 3a: 1H-NMR: 3.88 (s, 3H, CH3); 7.137.25 (m, 3H, ArH); 7.487.58 (m, 2H, ArH); 7.70 (d, 1H, J = 7.9 Hz, ArH); 7.88 (d, 1H, J = 7.9 Hz, ArH); 8.42 (d, 1H, J = 8.41 Hz, ArH); 8.50 (s, 1H, 4-CH); 8.79 (s, 1H, C=NH); 12.94 (s, 1H, NH). MS m/z 322 (M+). IR (KBr), cm1: m 3 300 (NH), 3 207 (NH), 3 041 (CH), 1 715 (C=O), 1 678 (C=O), 1 641 (C=N), 1 606 (C=C). Anal. C18H14N2O4 (C, H, N). 3b: 1H-NMR: 1.31 (t, 3H, J = 6.9 Hz, CH3); 4.30 (q, 2H, J = 6.9 Hz, CH2); 7.278.04 (m, 8H, ArH); 8.59 (s, 1H, 4-CH); 9.32 (s, 1H, C=NH); 13.16 (s, 1H, NH). MS m/z 336 (M+). IR (KBr), cm1: m 3 315 (NH), 2 976 (CH), 1 704 (C=O), 1 688 (C=O), 1 644 (C=N), 1 593 (C=C). Anal. C19H16N2O4 (C, H, N). 6.1.2. N-arylsubstituted 2-oxo-2H-1-benzopyran-3carboxamides 4a and b A solution of the corresponding 2-iminobenzopyran derivatives 3a and b (4 mmol) in 1520 mL of a mixture of ethanol/water/ 32% hydrochloric acid (30:1:1, v/v/v) was reuxed with vigorous stirring for 1 h. After cooling to room temperature the products precipitated were ltered and recrystallized from the appropriate solvent. Yields and physicochemical data of the synthesized N-substituted 2-oxo-2H-1-benzopyran-3-carboxamides 4a and b are listed in table I. 4a: 1H-NMR: 3.92 (s, 3H, CH3); 7.208.08 (m, 7H, ArH); 8.61 (d, 1H, J = 8.2 Hz, ArH); 9.04 (s, 1H, 4-CH); 12.16 (s, 1H, NH). MS m/z 323 (M+). IR (KBr), cm1: m 3 248 (NH), 3 042 (CH), 1 735 (C=O), 1 713 (C=O), 1 668 (C=O), 1 608 (C=C). Anal. C18H13NO5 (C, H, N). 4b: 1H-NMR: 1.38 (t, 3H, J = 7.0 Hz, CH3); 4.34 (q, 2H, J = 7.0 Hz, CH2); 7.427.54 (m, 2H, ArH); 7.728.06 (m, 6H, ArH); 9.00 (s, 1H, 4-CH); 10.88 (s, 1H, NH). MS m/z 337 (M+). IR (KBr), cm1: m 3 250 (NH), 2 984 (CH) 1 704 (C=O), 1 671 (C=O), 1 596 (C=C). Anal. C19H15NO5 (C, H, N). 6.1.3. N-arylsubstituted 2-oxo-2H-1-benzopyran-3carboxamides 4ch To a well-stirred solution of 2-carboxymalonanilic acid ester 5 [13] (4 mmol) in 10 mL of ethanol was added an equivalent amount of salicylic aldehydes 2af and a few drops of piperidine as a catalyst. The reaction mixture was stirred at room temperature for ca. 1 day and then poured into water. The products precipitated were ltered and recrystallized from the suitable solvent. Yields and physicochemical data of the synthesized N-substituted 2-imino-2H-1-benzopyran-3-carboxamides 4ch are listed in table I. 4c: 1H-NMR: 7.15 (dd 1H, J = 8.0, 8.0 Hz, ArH); 7.39 (m, 2H, ArH); 7.56 (dd, 1H, J = 8.0, 8.0 Hz, ArH); 7.73 (dd, 1H, J = 7.9, 7.9 Hz, ArH); 7.94 (d, 1H, J = 8.2 Hz, ArH); 8.05 (d, 1H, J = 8.2 Hz, ArH); 8.65 (d, 1H, J = 8.3 Hz, ArH); 8.85 (s, 1H, 4-CH); 13.52 (br s, 1H, NH). MS m/z 309 (M+). IR (KBr), cm1: m 3 266 (NH), 3 032 (CH), 1 731 (C=O), 1 696 (C=O), 1 673 (C=O), 1 608 (C=C). Anal. C17H11NO5 (C, H, N). 4d: 1 H-NMR: 3.95 (s, 3H, OCH3); 7.147.32 (m, 4H, ArH); 7.42 (d, 1H, J = 8.2 Hz, ArH); 8.01 (d, 1H, J = 8.0 Hz, ArH); 8.60 (d, 1H, J = 8.0 Hz, ArH); 8.89 (s, 1H, 4-CH); 13.20 (s, 1H, NH). MS m/z 339 (M+). IR (KBr), cm1: m 3 287 (NH), 2 952 (CH), 1 726 (C=O), 1 694 (C=O), 1 675 (C=O), 1 614 (C=C). Anal. C18H13NO6 (C, H, N). 4e: 1H-NMR: 3.95 (s, 3H, OCH3); 7.157.44 (m, 4H, ArH); 7.52 (d, 1H, J = 8.3 Hz, ArH); 8.00 (d, 1H, J = 7.9 Hz, ArH); 8.64 (dd, 1H, J = 8.0, 8.0 Hz, ArH); 8.90 (s, 1H, 4-CH); 13.12 (s, 1H, NH). MS m/z 339 (M+). IR (KBr), cm1: m 3 291 (NH), 2 947 (CH), 1 730 (C=O), 1 689 (C=O), 1 677 (C=O), 1 604 (C=C). Anal. C18H13NO6 (C, H, N). 4f: 1H-NMR: 7.21 (dd, 1H, J = 8.0, 8.0 Hz, ArH); 7.60 (dd, 1H, J = 8.0, 8.0 Hz, ArH); 7.71 (d, 1H, J = 8.2 Hz, ArH); 8.02 (d, 1H, J = 7.9 Hz, ArH); 8.46 (d, 1H, J = 8.3 Hz, ArH); 8.72 (d, 1H, J = 7.9 Hz, ArH); 8.95 (s, 1H, 5-CH); 9.10 (s, 1H, 4-CH); 13.21 (s, 1H, NH). MS m/z 354 (M+). IR (KBr), cm1: m 3 295 (NH), 3 085 (CH), 1 754 (C=O), 1 724 (C=O), 1 683 (C=O), 1 618 (C=C). Anal. C17H10N2O7 (C, H, N). 4g: 1 H-NMR: 7.20 (t, 1H, J = 7.9 Hz, ArH); 7.607.74 (m, 3H, ArH); 7.988.06 (m, 2H, ArH); 8.12 (s, 1H, 5-CH); 9.03 (s, 1H, 4-CH); 12.47 (s, 1H, NH). MS m/z 345, 343 (M+). IR (KBr), cm1: m 3 389 (OH + NH), 3 047 (CH), 1 736 (C=O), 1 689 (C=O), 1 671 (C=O), 1 607 (C=C). Anal. C17H10ClNO5 (C, H, N). 4h: 1H-NMR: 3.63 (m, 2H, CH2CH=CH2); 5.14 (m, 2H, CH2CH=CH2); 6.03 (m, 1H, CH2CH=CH2); 7.13 (dd, 1H, J = 8.0, 8.0 Hz, ArH); 7.32 (dd, 1H, J = 7.9, 7.9 Hz, ArH); 7.55 (m, 2H, ArH); 7.78 (d, 1H, J = 8.0 Hz, ArH); 8.03 (d, 1H, J = 7.9 Hz, ArH); 8.76 (dd, 1H, J = 7.9, 7.9 Hz, ArH); 8.98 (s, 1H, 4-CH); 12.53 (s, 1H, NH) MS m/z 349 (M+). IR (KBr), cm1: m 3 224 (NH), 2 942 (CH), 1 724 (C=O),

1001 1 684 (C=O), 1 657 (C=O), 1 600 (C=C). Anal. C20H15NO5 (C, H, N). 6.2. Pharmacology 6.2.1. Anti-inammatory activity 6.2.1.1. Carrageenan-induced rat hind paw oedema test Experiments were carried out on groups of ve Sprague-Dawley rats (140160 g). The tested compounds and reference drug were administered orally (po) in 0.5% methylcellulose solution in water and 1 h later 0.1 mL of 1% carrageenan solution was injected under the plantar aponeurosis of the right hind paw of the rat by the method of Winter et al. [15]. The volume of the paw was measured before and 3 h after carrageenan treatment by a mercury plethysmometer. Anti-inammatory activity was given as percentage of inhibition of oedema in treated groups compared with controls and was calculated according to the formula: % inhibition = 100 [1 (Vt/Vc)] where Vt is the mean increase in paw volume of the rats treated with tested compounds and Vc is the mean increase in paw volume of the control group of rats. 6.2.1.2. Acetic acid peritonitis assay This test was performed according to the procedure of Arrigoni-Martelli [16]. Groups of ve rats were administered intraperitoneally (ip) with 10 mL/kg of 0.5% acetic acid solution 1 h after oral administration of the tested compounds. After 30 min, the rats were killed with diethyl ether and peritoneal exudate was collected and measured. The anti-exudate response was expressed as the inhibition percentage in comparison to the vehicletreated control: % inhibition = 100 [1 (Vt/Vc)] where Vt is the mean volume of the peritoneal exudate in treated rats and Vc is the mean volume of the peritoneal exudate in vehicle-treated rats. 6.2.1.3. Toxicity studies All tested compounds were investigated for their acute toxicity and approximate lethal dose (ALD50). Albino mice (either sex) weighing 2025 g were used for the study. ALD50 values were determined by observing mortality within 24 h after drug administration [17]. 6.2.1.4. Statistical calculation Data are expressed as mean SE. The Students t test was applied to determine the signicance of the difference between the control and the treated groups. The Acknowledgements We are grateful to Dr S.N. Kovalenko (Department of Organic Chemistry) for initial suggestions and helpful discussions. We thank Dr S.M. Drogovoz (Department of Pharmacology) for providing the Departments facilities for pharmacological tests. We are indebted to Dr I.V. Ukrainets (Department of Pharmaceutical Chemistry) for his generous gift of the compound 5. References
[1] [2] [3] [4] [5] [6] [7] [8] [9] Maxwell A., Mol. Microbiol. 9 (1993) 681686. Dexeus F.H., Logothetis C.J., Sella A., Fitz K., Amato R., Reuben J.M., Dozier N., J. Clin. Oncol. 8 (1990) 325329. Zembower D.E., Liao S.Y., Flavin M.T., Xu Z.Q., Stup T.L., Buckheit R.W. et al., J. Med. Chem. 40 (1997) 10051017. Negwer M. (Ed.), Organic-Chemical Drugs and Their Synonyms, Academie-Verlag, Berlin, 1987. Lino C.S., Taveira M.L., Viana G.S.B., Matos F.J.A., Phytother. Res. 11 (1997) 211215. Silvan A.M., Abad M.J., Bermejo P., Sollhuber M., Villar A., J. Nat. Prod. 59 (1996) 11831185. Naka Y., Yamada I., Ochi H., Matsuura M., Nagamatsu Y., Oe T., Bioorg. Med. Chem. Lett. 5 (1995) 959962. Ukrainets I.V., Gorokhova O.V., Taran S.G., Turov A.V., Chem. Heterocycl. Compd. (NY) 30 (1994) 12111213. Tietze L.F., Beifuss U., in: Trost B.M., Fleming I. (Eds.), Comprehensive Organic Synthesis Vol. 2, Pergamon Press, Oxford, 1991, pp. 341394. Jones G., Org. React. (NY) 15 (1967) 204599. Schiemenz G.P., Chem. Ber. 95 (1962) 483486. Czerney P., Hartmann H., J. Prakt. Chem. 323 (1981) 691693. Ukrainets I.V., Bezugly P.A., Treskach V.I., Taran S.G., Gorokhova O.V., Tetrahedron 50 (1994) 1033110338. Bilokin Y.V., Kovalenko S.N., Bylov I.E., Chernykh V.P., Heterocycl. Commun. 4 (1998) 257260. Winter C.A., Risley E.A., Nuss G.W., Proc. Soc. Exp. Biol. Med. 111 (1962) 544547. Arrigoni-Martelli E., Boll. Chim. Farm. 107 (1968) 205217. Swinyard E.A., Brown W.C., Goodman L.S., J. Pharmacol. Exp. Ther. 106 (1952) 319330. OCallaghan C.N., McMurry T.B.H., OBrein J.E., J. Chem. Soc. Perkin Trans. 2 (1998) 425429. Ukrainets I.V., Taran S.G., Bezugly P.A., Kovalenko S.N., Turov A.V., Marusenko N.A., Khim. Geterotsikl. Soedin. (1993) 12231226. Huang C.K., Wu F.Y., Ai Y.X., Bioorg. Med. Chem. Lett. 5 (1995) 24232428. Selim M.R., Aly F.M., Bendair A.H., Abu-Shanab F.A., J. Indian Chem. Soc. 69 (1992) 688690. Wiseman E.H., Chang Y.H., Lombardino J.G., Arzneim. -Forsch. 26 (1976) 13001303.

difference in results was considered to be signicant when P < 0.05.

[10] [11] [12] [13] [14] [15] [16] [17] [18] [19]

[20] [21] [22]

Eur. J. Med. Chem. 34 (1999) 10031008 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1003

Short communication

Research on heterocyclic compounds, XLI. 2-Phenylimidazo[1,2-b]pyridazine-3-acetic derivatives: synthesis and anti-inammatory activity
Antonia Sacchia, Sonia Laneria*, Francesca Arenaa, Enrico Abignentea, Marina Gallitellia, Michele Damicob, Walter Filippellib, Francesco Rossib
a Dipartimento di Chimica Farmaceutica e Tossicologica, Facolt di Farmacia, Universit di Napoli Federico II, Via Domenico Montesano 49, I-80131 Napoli, Italy b Istituto di Farmacologia e Tossicologia, Facolt di Medicina e Chirurgia, II Universit di Napoli, Via Costantinopoli 16, I-80138 Napoli, Italy

(Received 21 January 1999; accepted 21 April 1999)

Abstract The synthesis of a group of 2-phenylimidazo[1,2-b]pyridazine-3-acetic esters and acids is described. The structures of the new compounds are supported by 1H-NMR spectra. These compounds were tested in vivo for their anti-inammatory, analgesic and ulcerogenic activity. All new compounds showed remarkable anti-inammatory action in the carrageenan rat paw oedema (one third of that for indomethacin) but no signicant analgesic activity in the acetic acid writhing test together with negligible ulcerogenic action, and were also found to be lacking inhibitory activity on cyclooxygenase in vitro. 1999 ditions scientiques et mdicales Elsevier SAS imidazo[1,2-b]pyridazines / anti-inammatory activity / analgesic activity / ulcerogenic activity / cyclooxygenase inhibition

1. Introduction In the context of our research on the structure-activity relationships and mode of action of bicyclic imidazoderivatives with anti-inammatory and analgesic activity, we synthesized a series of 2-phenylimidazo[1,2-b]pyridazine-3-carboxylic acids 1 (gure 1) which showed high analgesic activity in the acetic acid writhing test in mice, low or no anti-inammatory activity in the carrageenaninduced rat paw oedema and low ulcerogenic action on the rat gastric mucosa [1, 2]. We then synthesized three series of analogues, namely 2-methylimidazo[1,2-b]pyridazine-3-carboxylic acids 2 [3], imidazo[1,2-b]pyridazine-2-acetic acids 3 [4] and imidazo[1,2-b]pyridazine-2-carboxylic acids 4 [5]. In comparison with the rst series of acids 1, compounds 2 and 3 showed a lower analgesic activity, whereas the same activity reached the lowest level in the last series (acids 4). However, the pharmacological prole was the same in all four groups of compounds, since
*Correspondence and reprints

Figure 1. Imidazo[1,2-b]pyridazine acidic derivatives.

the analgesic activity was always coupled with low or no anti-inammatory and ulcerogenic action. In consideration of the fact that the highest analgesic activity was shown by the rst type of compounds, i.e. 2-phenylimidazo[1,2-b]pyridazine-3-carboxylic acid 1, and in order to extend the study of the structure-activity relationships, we have synthesized a new series of 2-phenyl derivatives, with only one structural change, i.e. the replacement of the carboxylic moiety in position 3 by an acetic group.

1004 prepared using as starting material a mixture (nearly 1:1) of 3-amino-6-chloro-5-methylpyridazine (5f) and 3-amino-6-chloro-4-methylpyridazine (5g): this mixture was obtained from 3,6-dichloro-4-methylpyridazine [8] following the Mori procedure [9]. After reaction with 6, the products 7f and 7g were separated by column chromatography. The above mixture of the amines 5f and 5g afforded the corresponding dehalogenated mixture of 3-amino-5-methylpyridazine (5d) and 3-amino-4-methylpyridazine (5e) by catalytic hydrogenation. The reaction of 6 with the mixture 5d5e afforded the esters 7d and 7e. The correct structural assignments to these products were performed by 1H-NMR spectra (table I), and are in accordance with the literature data, in particular with the chemical shifts found by Kobe et al. [10] for H-6, H-7, H-8 in the imidazo[1,2-b]pyridazine. 3. Pharmacology The new esters 7ag and acids 8ag were tested in vivo using the acetic acid writhing test in mice and carrageenan induced rat paw oedema to study analgesic and anti-inammatory activity. Higher doses were administered to rats to study the irritative and ulcerogenic action on the mucosa of the stomach and small intestine. Indomethacin was used in all tests as reference drug. These three tests should allow us not only to extend the SAR study to another series of imidazopyridazines, but also to obtain some information about their mechanism of action: they should display a similar level of activity in all three tests if they are inhibitors of prostaglandin biosynthesis. In order to unequivocally resolve this question, some new compounds (the more and less active ones) were also subjected to two different cyclooxygenase activity assays in vitro [11, 12]. 4. Results and discussion The anti-inammatory activity displayed by the compounds under examination is reported in table II. The esters 7c, 7d and 7g, and the acids 8a, 8c, 8f and 8g are the most active compounds with comparable levels of activity, as can be seen on the basis of reported values of ED50 (approximately one third of that for indomethacin). The results obtained in the acetic acid writhing test are quite different (table III), in fact all compounds show weak or no activity. The gastrointestinal irritative and ulcerogenic action (table IV) was almost completely absent in all compounds. There is no parallelism between the results of all three in vivo tests. This is certainly not the pharmacological

Figure 2. Synthetic method for 2-phenylimidazo[1,2-b]pyridazine-3-acetic derivatives.

2. Chemistry The required compounds were prepared using the synthetic method depicted in gure 2. The reaction in dry ethanol of 3-aminopyridazines 5ag with ethyl 3-benzoyl-3-bromopropionate 6, prepared ad hoc, afforded the corresponding ethyl 2-phenylimidazo[1,2b]pyridazine-3-acetates 7ag. These esters were converted into the respective acids 8ag by acid hydrolysis. Esters 7a, 7b and 7c were prepared by reacting 6 with 3-aminopyridazine 5a, 3-amino-6-chloropyridazine 5b, and 3-amino-6-methoxypyridazine 5c, respectively, prepared ad hoc following the methods described by Wermuth et al. [6] and Steck et al. [7]. The couples of isomers 7de and 7fg were obtained with the same method previously employed to obtain the corresponding isomeric couples in the preceding series of imidazo[1,2-b]pyridazines 14 [5]. The couple 7fg was

1005
Table I. Yields, m.p. and 1H-NMR spectral data of 2-phenylimidazo[1,2-b]pyridazine-3-acetic derivatives.
Compounds1 yield % 32 20 40 40 40 30 30 55 55 50 55 60 60 60 m.p. C 103105 108110 104105 110112 121123 113115 125127 > 200 > 200 > 200 > 200 > 200 > 200 > 200
1

: H-6 8.50 (dd)

H-7 7.20 (dd) 7.00 (d) 6.65 (d) 7.19 (d)

H-8 7.85 7.90 8.70 7.95 (dd) (d) (d) (d)

J6, 7 4.5

J6, 8 Hz 2.0

J7, 8 10 10 10

H-NMR in ppm2 Ph 7.70 7.70 8.60 7.50 7.60 7.65 7.80 7.60 7.90 8.67 7.60 7.65 7.90 7.82 (m) (m) (m) (m) (m) (m) (m) (m) (m) (m) (m) (m) (m) (m) 7.60 7.40 8.40 7.40 7.50 7.50 7.50 7.55 7.68 8.42 7.45 7.50 7.50 7.45

CH2 (s) (m) 4.10 (m) 4.10 (m) 4.10 (m) 4.10 (m) 4.10 (m) 4.20 (m) 7.30 (m) 4.20 (m) 4.05 (m) 4.10 (m) 4.10 (m) 4.10 (m) 4.10 (m) 4.10 (m) 4.20

Ethyl 4.00 4.05 4.05 4.08 4.05 4.10 4.10 (q) (q) (q) (q) (q) (q) (q) 1.00 1.20 1.10 1.20 1.30 1.30 1.30 (t) (t) (t) (t) (t) (t) (t)

Substituents

7a 7b 7c 7d 7e 7f 7g 8a 8b 8c 8d 8e 8f 8g
1 2

8.25 (d) 8.15 (d)

2.4 4.3

7.70 (s) 8.49 (dd) 6.85 7.20 7.21 6.75 (s) (dd) (d) (d) 7.80 7.95 8.77 8.04 (dd) (d) (d) (d) 4.2 2.0 10 10 10

6-OCH3: 3.90 (s) 7-CH3: 2.60 (s) 8-CH3: 2.80 (s) 7-CH3: 2.50 (s) 8-CH3: 2.70 (s)

8.35 (d) 8.25 (d)

2.3 4.3

7.20 (d) 7.75 (s) 6.80 (s)

6-OCH3: 3.90 (s) 7-CH3: 2.65 (s) 8-CH3: 2.55 (s) 7-CH3: 2.55 (s) 8-CH3: 2.68

All compounds were analysed for C, H, N (also for Cl when present): found values were within 0.4% compared with theoretical values. Solvents: CDCl3 for 7ag, CD3OD for 8ag.

Table II. Anti-inammatory activity by the carrageenan rat paw oedema test. Compound 7a 7b 7c Dose mg/kg p.o. 40 40 20 40 80 20 40 80 40 40 20 40 80 10 20 40 40 10 20 40 40 40 10 20 40 10 20 40 5 7.5 10 % Oedema inhibition relative to control at the: 1st h 2nd h 3rd h 17 7 62 17 52 28 63 62 17 0 5 27 57 30 47 53 14 30 58 83 67 33 22 66 86 22 58 86 3 16 39 26 14 53 26 72 14 56 78 26 20 17 34 75 31 42 68 31 20 42 66 51 51 33 65 81 28 42 72 40 33 55 29 19 34 47 80 24 53 84 47 43 20 37 82 22 33 73 33 22 41 55 43 43 33 51 76 22 41 76 38 49 67 4th h 44 21 37 59 87 42 51 90 44 44 37 51 77 32 47 66 24 24 34 64 25 38 22 55 75 25 34 75 35 55 79 ED50, mg/kg (ducial limits) 3rd h 4th h 35.1 (29.441.8) 36.5 (32.141.7) 28.9 (24.034.8)

7d

7e 7f 7g

33.2 (27.040.7) 21.4 (16.727.5) 27.8 (22.634.2) 19.8 (17.023.1) 23.0 (1.633.7) 6.7 (4.88.7)

8a

8b 8c

31.4 (24.540.1) 17.8 (14.721.5) 21.9 (18.925.4) 7.0 (4.510.8)

8d 8e 8f

8g

IMA

1006
Table III. Analgesic activity by the acetic acid writhing test in mice. Compound Dose mg/kg p.o. % Decrease of mean no. of writhes in 25 min after treatment relative to control 22.5 4.5 34.4 8.8 32.3 4.0 25.1 17.2 8.9 26.5 10.8 9.3 29.6 15.8 51.2

7a 7b 7c 7d 7e 7f 7g 8a 8b 8c 8d 8e 8f 8g IMA

40 40 40 40 40 40 40 40 40 40 40 40 40 40 5

prole expected from compounds acting as inhibitors of the prostaglandin biosynthesis. In order to investigate this aspect of the question, our compounds were tested in vitro for their cyclooxygenase-inhibiting activity. The most active compounds 7c, 7d, 8c, 8f and 8g were tested for their cyclooxygenase-inhibiting activity by measuring the rate of conversion of [1-14C]arachidonic acid into PGE2 in the microsomal fraction of mucosa preparation of rabbit distal colon after incubation with test compounds, following the method previously reported [13]. All compounds were found to be devoid of inhibitory activity, i.e. 0.79.0% relative to control, compared with 9092% of indomethacin at the same concentration (10 M).
Table IV. Incidence of gastrointestinal lesions in rats. Compound Dose mg/kg p.o. 100 100 100 100 100 100 100 100 100 100 5 Remarks at 6 h after treatment: % animal with: hyperaemia ulcers 20 10 30 30 20 30 40 20 30 30 80 0 0 0 0 0 0 0 0 0 0 50

The second test was carried out by measuring the rate of conversion of exogenous arachidonic acid into PGE2 in the rat medullary and cortical kidney microsomes. Inhibition of microsomal PGE2 production was measured by radioimmunoassay as reported previously [13]. This test conrmed the above results: no compound showed signicant activity (< 12% relative to control), except indomethacin (9094%) at the same concentration (10 M). Therefore in the present case the remarkable antiinammatory activity shown by these compounds was found to be independent of the cyclooxygenase inhibition. It should be noted that we recently came to the same conclusion for a group of imidazo[1,2-a]pyrimidine analogues [14]. From the point of view of our studies on structureactivity relationships of imidazo[1,2-b]pyridazine acidic derivatives 14 (gure 1), the most signicant nding is that the 2-phenylimidazo[1,2-b]pyridazine-3-acetic derivatives (esters 7 and acids 8) showed weak analgesic activity and proved to be completely lacking ulcerogenic action, whereas showed a signicant anti-inammatory activity. The pharmacological prole resulting from the above data for these 2-phenylimidazo[1,2-b]pyridazine3-acetic derivatives is different from that previously observed for 14 series in which the analgesic activity was clearly prevailing over the anti-inammatory action, so further investigation will be necessary in order to explain the activity of these imidazo-derivatives, which is probably due to the multiple mechanism of action differently inuenced by the structural changes. 5. Experimental protocols 5.1. Chemistry Thin layer chromatography by precoated silica gel plates (Merck 60 F254) was used to control the course of reactions and purity of products: all compounds were designated as pure when they showed a single spot after elution with chloroform/methanol mixture (95:5); detection of components was made by UV light and/or treatment with iodine vapors. Preparative separations were performed in columns packed with silica gel from Farmitalia Carlo Erba (RS, J mm 0.05:0.20). Melting points were determined with a Koer hot stage microscope and are uncorrected. Elemental analyses indicated by the symbols of the elements were within 0.4% of the theoretical values. The 1H-NMR spectra were recorded using a Bruker AMX- 500 spectrometer equipped with a Bruker X-32 computer; chemical shift values are reported

7a 7b 7c 7f 7g 8a 8b 8c 8f 8g IMA

1007 in units (ppm) relative to tetramethylsilane used as internal standard. 5.1.1. Ethyl 3-benzoyl-3-bromopropionate 6 A solution of 3-benzoylpropionic acid (19.6 g, 0.1 mol) in 100 mL of dry ethanol with concentrated H2SO4 (5 mL) was reuxed for 7 h. After cooling, ethanol was removed in vacuo, and the residue dissolved in diethyl ether and extracted with NaHCO3 saturated solution. The organic extract was washed with water, dried on Na2SO4 and evaporated in vacuo to obtain ethyl 3-benzoylpropionate as an oil in 75% yield. A solution of ethyl 3-benzoylpropionate in 100 mL of diethyl ether was added slowly with an equimolar amount of bromine at 0 C. The solution was stirred at room temperature for 1 h. The organic solution was washed three times with NaHCO3 saturated solution, dried on Na2SO4 and evaporated in vacuo to obtain the required compound 6 as an oil in 65% yield. 5.1.2. Ethyl-2-phenylimidazo[1,2-b]pyridazine-3-acetates 7a and 7cg General procedure: a mixture of the starting 3-aminopyridazine and ethyl 3-benzoyl-3-bromopropionate 6 (molar ratio 1:1.5) in ligroine was reuxed for 3 h. After cooling, the mixture was ltered and the ltrate was extracted with 10% aqueous HCl. The aqueous layer was separated and adjusted to pH 78 with NaHCO3 to obtain the precipitation of the ester which was then recrystallized from n-hexane. This procedure allowed us to obtain the esters 7a and 7c. In the case of the isomeric mixtures 7de and 7fg, obtained starting from the mixtures of the isomeric amines 5de and 5fg, respectively, the crude product precipitated was subjected to column chromatographic separation, eluting with n-hexane/diethyl ether mixtures with increasing percentage of ether. The single products obtained from the columns were then recrystallized from n-hexane. 5.1.3. Ethyl 6-chloro-2-phenylimidazo[1,2-b]pyridazine3-acetate, 7b A mixture of a 3-amino-6-chloropyridazine 5b and ethyl 3-benzoyl-3-bromopropionate 6 (molar ratio 1:1.5) in anhydrous ethanol was reuxed for 10 h. After cooling, ethanol was removed in vacuo, and the residue treated with NaHCO3 saturated solution and extracted with CHCl3. The organic extract was washed with water, dried on Na2SO4 and evaporated in vacuo to obtain the required product 7b, which was recrystallized from n-hexane. 5.1.4. 2-Phenylimidazo[1,2-b]pyridazine-3-acetic acids 8ag General procedure: a mixture of 10 mmol of each ethyl ester and 40 mL of 10% aqueous HCl was reuxed for 23 h. After cooling, the solution was adjusted to pH 45 with NaHCO3 to obtain the precipitation of the acid which was recrystallized from ethanol. 5.2. Pharmacology As regards the experiments carried out in vivo, test compounds were administered orally by gavage in 1% methylcellulose suspension. In the oedema and writhing test each compound was rst tested at 40 mg/kg. If a signicant activity was observed, lower and/or higher doses were administered in order both to study the dose-dependence of the pharmacological activity and to calculate ED50 values, when possible. Gastric ulcerogenic action was studied in rats which were treated orally with higher doses (100 mg/kg). Indomethacin was included in all tests for comparison purposes (IMA in tables IIIV). 5.2.1. Anti-inammatory activity The paw oedema inhibition test [15] was used on rats. Groups of 5 rats of both sexes (body weight 150200 g), pregnant females excluded, were given a dose of a test compound. After 30 min, 0.2 mL of 1% carrageenan suspension in 0.9% NaCl solution was injected subcutaneously into the plantar aponeurosis of the hind paw and the paw volume was measured by a water plethysmometer Socrel and then measured again 1, 2, 3 and 4 h later. The mean increase of paw volume at each time interval was compared with that of a control group (5 rats treated with carrageenan, but not treated with test compounds) at the same time intervals and percentage inhibition values were calculated. The experimental results are listed in table II. 5.2.2. Analgesic activity Acetic acid writhing test [16] was used on mice. Groups of 5 mice of both sexes (body weight 2025 g), pregnant females excluded, were given a dose of a test compound. After 30 min the animals were injected intraperitoneally with 0.25 mL/mouse of 0.5% acetic acid solution and writhes were counted during the following 25 min. The mean number of writhes for each experimental group and percentage decrease compared with the control group (5 mice not treated with test compounds) were calculated. The experimental results are listed in table III.

1008 5.2.3. Ulcerogenic action Groups of 10 rats of both sexes (body weight 200220 g), pregnant females excluded, fasted for 24 h, were given an oral dose of a test compound, except the control group. All animals were sacriced 6 h after dosing and their stomachs and small intestine were macroscopically examined to assess the incidence of hyperaemia and ulcers. The experimental results are listed in table IV.
[2] [3] [4] [5] [6] [7] [8] Abignente E., Arena F., Luraschi E., Saturnino C., Marmo E., Berrino L., Donnoli D., Farmaco 45 (1990) 10751087. Abignente E., Arena F., Luraschi E., Saturnino C., Rossi F., Berrino L., Cenicola M.L., Farmaco 47 (1992) 931944. Luraschi E., Arena F., Sacchi A., Laneri S., Abignente E., DAmico M., Berrino L., Rossi F., Farmaco 50 (1995) 349354. Luraschi E., Arena F., Sacchi A., Laneri S., Abignente E., Avallone L., DAmico M., Berrino L., Rossi F., Farmaco 52 (1997) 213217. Wermuth C.G., Bourguignon J.J., Schlewer G., Gies J.P., Schoenfelder A., Melikian A. et al., J. Med. Chem. 30 (1987) 239249. Steck E.A., Brundage R.P., Fletcher L.T., J. Am. Chem. Soc. 76 (1954) 32253226. Steck E.A., Brundage R.P., Fletcher L.T., J. Am. Chem. Soc. 76 (1954) 44544457. Mori K., Yakugaku Zasshi, 82 (1962) 303-309; Chem. Abs., 58 (1963) 3427h. Kobe J., Stanovnik B., Tisler M., Tetrahedron 24 (1968) 239243. Calderaro V., Parrillo C., Giovane A., Greco R., Matera M.G., Berrino L., Rossi F., J. Pharmacol. Exp. Ther. 263 (1992) 579587. Gans K.R., Galbraith W., Roman R.J., Haber S.B., Kerr J.S., Schdmit W.K., Smith C., Hewes W.E., Ackerman N.R., J. Pharmacol. Exp. Ther. 254 (1990) 180187. Abignente E., Sacchi A., Laneri S., Rossi F., DAmico M., Berrino L., Calderaro V., Parrillo C., Eur. J. Med. Chem. 29 (1994) 279286. Laneri S., Sacchi A., Gallitelli M., Arena F., Luraschi E., Abignente E., Filippelli W., Rossi F., Eur. J. Med. Chem. 33 (1998) 163170. Winter C.A., Risley E.A., Nuss G.W., Proc. Soc. Exp. Biol. Med. 111 (1962) 544547. Davies J.E., Kellet D.N., Pennington J.C., Arch. Int. Pharmacodyn. Ther. 221 (1976) 274282.

Acknowledgements The NMR spectra were performed at the Centro di Ricerca Interdipartimentale di Analisi Srtumentale, Universit di Napoli Federico II, Naples. The assistance of staff is much appreciated. This research has been nancially supported by the Italian Ministry of University and Scientic and Technological Research (MURST, Rome).

[9] [10] [11] [12]

[13] [14]

References
[15] [1] Abignente E., Arena F., Luraschi E., De Caprariis P., Marmo E., Vitagliano S., Donnoli D., Res. Commun. Chem. Pathol. Pharmacol. 67 (1990) 4354. [16]

Eur. J. Med. Chem. 34 (1999) 10091018 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1009

Short communication

2-Substituted indazoles. Synthesis and antimicrobial activity

Tams Lrnda*, Bla Kocsisb, Levente Emdyb, Pal Sohrc
b a Department of Medical Chemistry, University Medical School, H-7601 Pcs, POB 99, Hungary Department of Medical Microbiology and Immunology, University Medical School, H-7601 Pcs, POB 99, Hungary c Department of General and Inorganic Chemistry, Lornd Etvs University, H-1518 Budapest, POB 32, Hungary

(Received 24 November 1998; revised 4 May 1999; accepted 7 May 1999)

Abstract 2-Isothiocarbamoyl substituted fused pyrazolines and their S-alkyl derivatives were prepared as potentially antimicrobial agents. Conventional methods were used to synthesize the novel derivatives starting from cyclic unsaturated ketones and thiosemicarbazide under acidic catalyst. These cyclizations yielded only one diastereoisomer of 3-H, 3a-H cis. The alkylations were performed applying alkyl halides. The structures of the new compounds, including congurations and conformations, were elucidated by NMR spectroscopy, also making use of 2D-HSC, DEPT and DNOE measurements. The S-alkyl derivatives were evaluated for activity against Gram-negative and Gram-positive bacteria and their in vitro toxicity was determined on HeLa cells. The structure-activity relationship was also studied. 1999 ditions scientiques et mdicales Elsevier SAS indazole / alkylation / stereostructure by NMR / antibacterial effect / in vitro toxicity

1. Introduction Some thiosemicarbazides are known antibacterial compounds, like thiosemicarbazones of 5-nitrofurfurylideneacetone [1] and dodecanone [2]. The 1-methylindole-2,3dione 3-thiosemicarbazone, called Metisazone, was used as an antiviral agent [3]. Because of their relatively high toxicity these agents are not widely used drugs. Similarly, some members of the family of the fused pyrazolines and pyrazoles show antimicrobial effects [46]. This prompted us to nd an effective antimicrobial agent of relatively low toxicity having a thioamide moiety attached to a pyrazoline ring. Recently we have published the synthesis of several bi- and tricyclic pyrazolines starting from unsaturated ketones and hydrazine derivatives [79]. Some of these compounds can be considered as cyclic thiosemicarbazide derivatives too. Our aim was to prepare water-soluble thiosemicarbazide derivatives for microbiological investigations. Therefore, 3,5diarylidene-1-methyl-4-piperidones were also used as starting ketones.

2. Chemistry In order to obtain potentially antibacterial compounds starting from 2,6-dibenzylidenecyclohexanone, 3,5diarylidene-1-methyl-4-piperidones or 2-arylidene-1tetralones and semicarbazide or thiosemicarbazides, ten new bi- and tricyclic pyrazolines (4b, 8ad and 11bf) have been prepared (gures 13). The cyclizations performed with thiosemicarbazides under acidic conditions yielded only one diastereoisomer of 3-H, 3a-H cis. While the reactions with semicarbazide afforded the mixture of the 3-H, 3a-H cis and trans diastereoisomers, which have been separated. The structure and the relative conguration of the compounds have been determined by 1H-NMR and 13C-NMR spectroscopic methods. To increase the water solubility of the 2-isothiocarbamoyl pyrazolines they have been alkylated with alkyl halides. These reactions gave the corresponding S-alkyl derivatives (5, 6ab, 9ad and 12ag) depicted in gures 13. From 6a hydroiodide, the free base 6c was liberated to provide a compound of better solubility for NMR study. At the alkylation of 8cd pyrazolo[4,3c]pyridines the corresponding hydrochloride was used to avoid the quaternarization of the pyridine ring. On the

*Correspondence and reprints

1010

Figure 2. Synthesis of hexahydropyrazolo[4,3-c]pyridines. Figure 1. Synthesis of substituted hexahydroindazoles.

contrary, allylic halides, like allyl bromide and benzyl chloride, yielded the 13 and 14 N-alkyl derivatives (gure 4). N-alkylation was observed at thioamides with alkylating agents favouring the carbonium ion formation [10]. The structures of the N-alkylated and S-alkylated derivatives were proven by their 1H- and 13 C-NMR spectra. Two of the S-alkyl derivatives (6a and 9a) were prepared via an alternative route using S-methylthiosemicarbazide hydroiodide and the corresponding ,-unsaturated ketone. The products of this one-step synthesis were identical in every respect with those prepared in a two-step method. The physical data of the novel compounds are displayed in table I. The 1H- and 13C-NMR data on the new compounds are listed in tables II and III. As the spectra inequivocally prove the expected constitutions, the following discussion is focused on the steric aspects of the structures, on the congurations and conformations. Two problems have to be considered: the cis or trans conguration of the H-3, H-3a hydrogens and the hindered rotation of the carbamide or thiocarbamide moiety. Concerning the rst problem, the spectral data of the cis-trans pair 8a and b can serve as a starting point. The vicinal H-3, H-3a coupling is 9.3 and 11.5 Hz for the

isomers. While the values are in accordance with the expected ratio 3Jcis > 3Jtrans [11], the difference is far too small for a rm differentiation of the cis or trans congurations in the case of single compounds without their counterparts. Moreover, because of broadened signals, it was not possible to determine the value of this coupling constant for all compounds and in most cases the splitting was between the two values (about 1011 Hz) measured for the pair 8a and b. On the contrary, the 13 CNMR eld effect [12, 13] arising in the cis isomers is a rm base to identify the C-3,3a congurations. The C-3a chemical shifts for the cis and trans isomers 8a and b are 48.4 and 55.4 ppm. The signicant upeld shift (8.2 ppm) for C-3a in 8a and the similar C-3a values (48.451.0 ppm) measured for the other compounds proved their cis-conguration unambiguously. Signal splitting or broadening in the 1H- and 13C-NMR of most of the salts investigated refers to the equilibrium of rotamers or other differences in the structures. Because of the absence of these phenomena in the spectra of bases, the origin can also be found in different protonation sites: besides exocyclic, the basic cyclic sp2-N can also be protonated. The double splitting of H-8 and C-7a signals for 12-type salts supports this latter assumption, while splitting or broadening of H-3 and C-3 signals can be

1011

Figure 4. N-alkyl derivatives 13, 14 and monocyclic pyrazole derivative 15.

Figure 3. Synthesis of tricyclic pyrazolines.

interpreted by hindered rotation only. Probably, the above facts have to arise from hindered rotation and/or site isomers of protonated molecules, depending on the structure as well as medium effects (solvent, concentration, temperature, pH, etc.). Two further aspects of stereostructures should be considered: the geometrical isomerism of the benzylidene moiety and the conformation of the partly saturated condensed ring. As regards the geometrical isomers, the Z-isomer (S-cis Ar and N-1) is to be precluded due to unfavourable steric structure. The S-cis (to N-1) position of the hydrogen in the exocyclic =CH- group reveals a high downeld shift of its 1H-NMR signal (this appears to overlap with the multiplets of the phenyl hydrogens) due to anisotropic deshielding of the non-bonded electron pair on N-1 [14]. The condensed, partly saturated six-membered ring can exist in two preferred conformations. These chairand boat-like forms with out-of-plane C-5 and C-7a atoms are both distorted approaching the sofa-

arrangement with coplanar C-7a. It is to be noted that the inversion of the six-membered ring hardly inuences the envelope form of the pyrazoline ring (with an out-ofplane C-3 atom) bearing the Ar group in quasi-equatorial position, cf. Ref [7]. Nevertheless, the quartet-like signal of H-4ax (for 4b, 5, 6ac, 11c, 13 and 14) refers to three large H,H-splittings, one of which is due to geminal coupling [15] 4ax,4eq and the two others to 3a,4ax and 4ax,5ax diaxial vicinal interactions [16, 17] which can be present only in the chair-form. This conformation is also preferred in the case of 8a and c, but the H-4ax signal shows a triplet splitting in the absence of H-5ax here. Due to signal-overlap, it was not possible to identify the multiplicity of the H-4ax signal for 9a and d, but in the TFA-d solution of 9a it is separated with triplet-like multiplicity and the two large splits are proof of the same conformation. In the 1H-NMR spectra of 12a,c and g the signals are broadened and consequently, it was not possible to identify the multiplicities. It is conceivable that the signal-broadening is the consequence of a slow ring inversion combined with the hindered rotation of the protonated thiocarbamide group. That is, the conformationally homogeneous quasi-rigid system of the other compounds is substituted by a exible equilibrium of ring isomers for 9-type derivatives. Again, this equilibrium can be inuenced in 12g by the incorporated sulphur in position 5. 3. Biology Antibacterial activity tests of the synthesized compounds were carried out on twenty cultures of Staphylococcus aureus and Escherichia coli strains isolated from different clinical samples (see Experimental protocols). Eight of the most effective antibacterial compounds (5,

1012
Table I. Physical data of compounds 4b, 5, 6ac, 8ad, 9ad, 11bf, 12ag and 1315. Compound 4b 5 6a 6b 6c 8ac 8bc 8c 8c HCl 8d 8d HCl 9a 9b 9c 9d 11b 11c 11d 11e 11f 12a 12b 12c 12d 12e 12f 12g 13 14 15
a

General formulaa C22H23N3S C15H19N3SHI C22H23N3SHI C23H25N3SHI C22H23N3S C21H22N4O C21H22N4O C21H22N4S C21H22N4SHCl C23H26N4O2S C23H26N4O2SHCl C22H24N4S 2HCl C24H28N4O2S 2HCl C23H27N5S 3HCl C23H27IN4SHI C18H16ClN3S C18H16BrN3S C18H15Cl2N3S C20H21N3O2S C16H15N3S2 C19H19N3SHI C19H18ClN3SHI C19H18BrN3SHI C19H17Cl2N3SHI C21H23N3O2SHI C17H17N3S2HI C18H17N3S2HI C24H25N3SHBr C25H23N3SHCl C17H17N3SHI

M.p. (C) 155 (dec., methanol) 172 (dec., acetone) 180183 (acetone) 242 (dec., acetone) 160 (dec., methanol) 163165 (methanol) 130 (dec., methanol) 174177 (methanol) 213216 (methanol) 181 (dec., methanol) 189 (dec., methanol) 200204 (ethanol) 195 (dec., ethanol) 114 (dec., ethanol) 198202 (ethanol) 225227 (methanol) 232233 (methanol) 118 (dec., methanol) 130132 (methanol) 190 (dec., methanol) 158162 (acetone) 168171 (acetone) 160 (dec., acetone) 170173 (acetone) 158160 (acetone) 140143 (acetone) 183 (dec., acetone) 176180 (acetone) 192195 (acetone) 116 (dec., acetone)

Yield (%) 62 58 79 47b 92 33 6 70 60 39 42 47 39 38 42 90 93 74 40 80 65 41 43 65 69 71 36 75 49 25

Method A, C A B, C B B B A A A A A A A A A A

The analytical values were within 0.4% of the theoretical values for C, H and N. b36% of 4b recovered. cIR (KBr) 1 685 cm1.

6a, 9a, 9d and 12ad) were examined for determination of their minimum inhibitory concentration (MIC) values by the test tube dilution method (see Experimental protocols). In these experiments we used standard reference strains. The MIC values of these ve standard strains were tested on eight well-known antibiotics to compare our compounds with antibiotics used in therapy as well. In addition, the MIC values of compound 9a were investigated on ve E. coli strains cultivated from urine. In vitro cytotoxicity tests of the compounds on HeLa cell-line OHIO, was carried out on microplates (see Experimental protocols). 4. Results and discussion Our aim was to study the structure-biological activity relationship for these fused pyrazolines by varying the

type and size of the ring system, the aromatic substituent at position 3 and the substituents on the isothiocarbamoyl group. 4.1. Antibacterial results Our work provided us with information about the structure-antimicrobial activity relationship connected to the class of fused pyrazolines (table IV). Opposed to the Metisazone they are ineffective without an S-alkyl substituent, like 8c. An arylidene substituent or a third (aromatic) ring is needed for effectiveness (see 6a and 12a versus 5), in addition, without fused ring the pyrazolines are ineffective, like 15. Replacement of the 5-CH2 group with sulfur in the benz[g]indazole series (12g) removed the antimicrobial effect. This can be explained in part by the different conformation of compounds 12a and 12g and the different electron structure. As for the

1013
Table II. 1H-NMR data (chemical shifts, in ppm, TMS = 0 ppm, and coupling constants in Hz) of compounds 4b, 5, 6ac, 8ad, 9a, d, 11bf, 12ag and 1315 in CDCl3 or DMSO-d6 solutiona at 250 MHzb. Compoundc 4b 5 6a 6b 6c 8a 8b 8c 8d 9a 9d 11b 11c 11d 11e 11f 12a 12b 12c 12d 12e 12f 12g 13 14 15
a

H-3 d (1H)d 6.08 5.75 5.87 6.15 5.70 5.60 4.82 6.09 6.02 6.07 6.12 6.10 6.09 6.08 5.93 6.26 5.95 5.96g 6.02 6.0g 5.86 6.27 6.0g 6.05 5.75 5.95

H-3a m (1H)e 3.45 3.75 3.9g 3.85 3.45 3.70 3.05k 3.77 3.74 4.70 4.7k 3.67 3.70 3.7g 3.73 3.77 4.18 4.13 4.18 4.14 4.07 4.10 4.50 3.95 3.98

NMe s (3H)f 3.18 3.25 2.25 2.25 2.25 2.27 2.70 3.15 4.00 4.42

SMe s (3H) 2.57 2.65 2.45 2.28 2.76 2.66l 2.67 2.63 2.67 2.64 2.63 2.64 2.64 2.67

NH 1 or 2 s (2H)g,h i 8.9 9.5 9.2 9.7 9.2 6.5 5.65 6.5 6.2 7.1 i 10.05 12.4 9.5 10.0 6.3 7.1 6.2 7.1 6.3 7.1 7.75 7.95 7.74 8.02 9.2 9.7 9.15 9.75 9.2 9.7 9.2 9.8 9.10 9.65 9.25 9.70 9.5 10.0 8.35 8.9 i 9.15 9.75

Solvent was CDCl3 for 4b, 6c, 8ad and 11bd, TFA-d for 14. Compounds 6b and 13 were measured in CDCl3 and DMSO-d6 (13 in DMSO-d6 at 100 C), 9a in DMSO-d6 and TFA-d. bMeasuring frequency was 500 MHz for 8d, 11b and df, 12b and dg and 15. Further signals: CH2 (position 47), 28 signal (6H for 4b, 6ac, 13, 8H for 5, 2H for 12g and 15 and 4H for all other compounds): 0.44.7, OCH3, s (3H): 3.78 and 3.82 (8d), 3.67 (6H, 11e), 3.69 and 3.71 (12e), ArH + C(sp2)H, 16 ms [11H or 5H (5), 8H (11c and 12c), 9H (12a and g), 10H (15) and 14H (14)]: 6.88.1, allylic group (13), C(sp3)H2: 4.10, d and 4.25, dqa, C(sp2)H2: 5.20, d and 5.30, d, C(sp2)H: 5.65 br and 5.90 br (doublet signals due to rotamers). cAssignments were proved by DR- (8a), 2D-COSY (8d, 11f) and DNOE-measurements (11c). A sample containing ca. 5% of cis-isomer was measured for compound 14, H-3: 5.90 for the minor isomer. dJ: 11.0 (4b, 6a and 8c), 9.4 (6b in DMSO-d6, 8a), 10 (6c, 9a and d, 13 and 14), 11.5 (8b), 10.7 (8d, 11b and d), 8.3 (11c), 10.5 (11e), 10.2 (11f), broadened signal (5, 12a,c and g and 15). Doubled signals due to rotamers with the second d at 6.05 (6a and 12b), 6.62 (9a), 6.25 (9d), 6.1 (12a), 6.00 (12e) and 6.35 (12f).edt (5, 6c, 8a and c, 12a, c and g), ddd (11c, 13 and 14), broadened (6a). fd for 4b (J: 4.8), 9a (J: 8.5) and 13 (J: 7), NCH2 (intensity 2H) for 13 and 14 (N-allyl and N-benzyl group), two singlets (23H) for 9d with the second signal at 3.19. gBroad signals. hIntensity: 1H (5 and 6ac and for the signal at 12.4 in 9a). The split to two signals due to rotamers for 5, 6a, 8c, 9a and d, 11bf, 12ag, 13 and 15. The intensity ratio is ca. 1:1, for 5, 6a, 9d, 12a and d and 15 ca. 2:1 for 12e and f ca. 5:2. iNot identiable. kIn overlap with the CH2-signals. l Doubled signal with the second singlet at 2.67.

3-aryl group, the best effect has been shown with the 4-halogen substituted derivatives (12bd). With respect to the S-alkyl group, the replacement of the nonpolar methyl group removed the antibacterial effect as in the case of 9c. This effect requires an N-unsubstituted thioamide moiety (see 6b). 9a proved to be the only effective compound against both Gram-negative and Grampositive strains. Its quaternarization decreased its effect both against the Gram-negative and Gram-positive strains.

4.2. Conclusion We have synthesized a series of 2-isothiocarbamoyl substituted bi- and tricyclic pyrazolines and their S-alkyl derivatives as potent antibacterial compounds. For this class of compounds a certain substitution pattern is necessary for the optimal antibacterial effect. Thus the 7-benzylidene-3,3a,4,5,6,7-hexahydro-5-methyl-2S-methylthiuronyl-3-phenyl-2H-pyrazolo[4,3-c]pyridine dihydrochloride (9a) is an acceptably potent compound

1014
Table III. 13C-NMR chemical shifts (TMS= 0 ppm) of compounds 4b, 5, 6ac, 8ad, 11bf, 12ag and 15 in CDCl3 or DMSO-d6 solutiona at 63 or 126 MHzb. Compound 4b 5 6a 6b 6c 8a 8b 8c 8d 11b 11c 11d 11e 11f 12a 12b 12c 12d 12e 12f 12gk 15
a

C3 67.7 65.7f 66.7g 67.5 66.2 62.6 66.5 66.5 66.3 66.6 66.4 66.1 67.2 63.7 66.5g 66.5g 66.1f 66.3 67.6g 64.2 67.3 63.6

C3a 49.7 50.7 51.0 50.6 49.9 48.4 55.4 48.8 48.9 49.2 48.9 49.1 49.1 49.3 49.8 50.6 49.7 50.6 50.6 50.7 49.5 44.5

C4 28.2 27.1e 27.8 27.8 28.0 55.7e 56.6e 55.8e 54.3 29.3 29.1 29.2 29.4 29.3 28.3 29.2 28.3 29.2 29.2 29.1 26.5

C5 23.6 22.9 22.7 22.9 23.7 24.3 24.1 24.3 24.7 24.0 23.8 24.7 23.9 24.7 24.6 23.9

C6 25.5 24.7 25.0 24.7e 25.5 54.2e 56.0e 54.3e 56.0 140.3 140.1 140.2 141.1 141.2 141.4 142.4 141.4 142.3 142.4 142.4 137.8

C7 137.3 27.6e 135.0 135.1 136.5 135.2 142.9 135.0 ?f 126.5 126.3 126.3 127.7 ?f 126.0 126.8 125.9 126.5 126.8 126.0 123.7

C7a 159.0 164f 163.7 159.0 155.4e 153.2 157.2 158.0 159.0 157.6 157.5 157.6 157.1 157.0 163.0g 163.3 163.2g 164f ?f 164f 159.4 162.5

C8 127.3e 126.2 127.2e 127.2e 127.7 129.6i 127.7 129.7 126.9 126.7 126.9 125.7 125.8 127.0i 127.8 127.0 127.8 126.1 127.7 126.4e

C(sp2)Xc 176.4 171f 165.6g 176.1 159.9e 154.3 152.8 176.0 175.9 176.1 175.8 176.1 176.6 176.6 165.0 164.8 165f 165f 164.4g 165f,g 164.8 165.1

XMed 31.1 14.5f 14.2g 25.4e 12.8 45.3 45.1 45.5 45.5 14.5g 15.5g 14.7f 15.4 14.8 15.5 14.6 14.6

Solvent was CDCl3 for 4b, 6b and c, 8a,c and d and 11bd and DMSO-d6 for 5, 6a, 8b, 11e and f, 12ag and 15. Compounds 8d, 11b and df, 12b and df and 15 were measured at 126 MHz. Assignments were supported by DEPT (for 8d, 11b and df, and 12b and dg) and 2D-HSC (for 8d and 11e and f) measurements. bFuther lines, OMe: 55.2 and 55.3 (8d), 56.3 and 56.4 (11e), 56.4 and 56.6 (12e). Aromatic carbons (because of poor quality of the 13C-NMR spectra due to hindered rotation and bad solubility it was not possible to measure exact chemical shifts for some broad and weak lines of these carbons in the cases of 6a and b, 12b and df and 15), 3-phenyl/aryl or in 11f and 12f 3-(2-thienyl) and the conjugated phenyl/aryl groups in the side chain and in position 3 (for 15): C-1: 135.0137.3 (non-conjugated rings), 126130 (cr: conjugated rings), 141.8 (11f), C-2,6: 125.0130.8, C-3,4,5: 126.0130.0, except for the following cases: 8d: C-3,5: 114.1, C-2,6: 131.7 (cr), C-4: 159.0, 159.7 (cr), 11b and 12b: C-4: 133.5, 11c and 12c: C-3,5: 131.7, 132.6, C-4: 121.4, 121.9, 11d and 12d: C-3: 132.9, 132.2, C-4: 131.8, C-5: 130.7, 11e and 12e: C-1: 131.2, C-2: 111.0, C-3: 149.6, C-4: 148.7, 149.7, C-5: 112.6, 113.0, C-6: 118.4, 11f: C-3,5: 125.2, 125.5. Condensed benzene ring in 11bf and 12ag (the numbering is given in the gure), C-9: 125.1126.6; C-10: 131.2134.1; C-11: 129.0130.4. cThiocarbamide (X = S) for 4b, 5, 6ac, 8c and d, 11bf, 12ag and 15, carbamide (X = O) for 8a and b. dX = N for 4b, 6b and 8ad, X = S for 5, 6a and c, 12ag and 15. eInterchangeable assignments (with an aromatic line for C-8 in 4b, 6b and c and 12g). fBroadened signal due to hindered rotation of the N-acyl group. Broadening was also observed for some aromatic carbon lines in 12b, d and f and 15. gSplit signals due to hindered rotation with the second line at 67.4 (C-3), 166.7 (Csp2S) and 14.8 (SMe) for 6a, 66.9 (C-3), 163.4 (C-7a) and 14.8 (SMe) for 12a, 67.0 (C-3) and 14.9 (SMe) for 12b, 163.9 (C-7a) for 12c, 67.0 (C-3) and 164.9 (= CS) for 12e and z 167f for 12f. iTwo overlapping lines, one originates from an aromatic carbon. kMeasurement temperature was 70 C.

inhibiting both Gram-positive and Gram-negative bacteria. 12c was the best agent against the Gram-positive strains (cf. table V; MIC values 12.525 g/mL.) Either a fused aromatic ring or a planar arylidene substituent is also crucial. It is very important that the thioamide moiety can not be substituted. In comparison with antibiotics commonly used in therapy our most effective compounds show similar or slightly less antibacterial activity (tables V and VI). This class of compounds having relatively low toxicity (cf. table VII; LDt50 > 250 g/mL) could be new potent antibacterial agents. Resistance to antimicrobial drugs is increasing all over the world. Both Gramnegative and Gram-positive strains are involved in this

process. New types of compounds with antimicrobial activity could diminish this negative tendency. 5. Experimental protocols 5.1. Chemistry Melting points were determined on a Boetius apparatus and are uncorrected. Microanalyses were carried out at the Central Research Laboratory, University Medical School, Pcs. IR spectra were run in KBr discs on a Bruker IFS-55 FT-IR spectrometer controlled by Opus 2.0 software. 1H- and 13C-NMR spectra were recorded in

1015
Table IV. Microbial screening of the antimicrobial effects of the novel compounds. Compounda 5 6a 6b 8a 8cb 9a 9b 9c 9d 12a 12b 12c 12d 12e 12f 12g 15
a

Number of S. aureus inhibited / tested 1/20 19/20 0/20 0/20 0/20 15/20 0/20 0/20 20/20 20/20 18/20 19c/20 19/20 0/20 0/20 0/20 0/20

Number of E. coli inhibited / tested 0/20 0/20 0/20 0/20 0/20 13/20 0/20 0/20 0/20 0/20 0/20 0/20 0/20 0/20 0/20 0/20 0/20

used in the DR experiments. DEPT spectra [18] were run in a standard way [19], using only the = 135 pulse to separate the CH/CH3 and CH2 lines phased up and down, respectively. For DNOE measurements [20, 21] the standard Bruker microprogram DNOEMULT.AU to generate NOE was used. The 2D-HSC spectra [22] were obtained by using the standard Bruker pulse program XHCORRD.AU. Analytical thin layer chromatography (TLC) was applied to monitor the reactions using precoated plates (Silica gel 60 F-254, Merck), and spots were visualized with UV light. The synthesis of some starting tricyclic pyrazolines (3, 4a, 11a and 11g,) [7, 8], 2-isothiocarbamoyl-3,5-diphenyl-2-pyrazoline [23], 2-arylidene1-tetralones [24, 25, 26 and unpublished data], 2-benzylidene-1-thiochroman-4-one [27] and 3,5-diarylidene-1-methyl-4-piperidones [28] has been reported earlier. The analytical values were within 0.4% of the theoretical values for C, H and N. 5.1.1. General procedure for the preparation of 2-isothiocarbamoyl- or 2-carbamoyl- substituted compounds (4b, 8ad and 11bf) The mixture of thiosemiocarbazide or semicarbazide (30 mmol) and the corresponding unsaturated ketone (10 mmol) was reuxed in ethanol (110 mL) containing 9% concentrated hydrochloric acid until the disappearance of the starting unsaturated ketone. The reaction mixture was cooled down, the precipitate was ltered and washed with cold ethanol and water until neutral. As

Concentration, 50 g/mL. bUsed as a hydrochloride. cThe effect was bactericidal against some strains.

different solutions (tables II and III) in 5 mm tubes at room temperature, on a Bruker WM-250 FT-spectrometer equipped with an Aspect 2000 computer at 250.13 (1H) and 62.89 (13C) MHz, respectively, using the deuterium signal of the solvent as the lock and TMS as internal standard. Conventional CW irradiation of 0.15 W was
Table V. Minimum inhibitory concentration (MIC) values.

Concentration g/mL STRAINS S. aureus NIH HUNGARY 118 003 S. saprophyticus NIH HUNGARY 120 008 M. luteus ATCC 9341 > 200 5 9d 200 6a 12a 5 9d 100 50 9a 12b 9a 12a 9a 9d 12a 9d 5, 6a, 9a, 9d, 12a, 12b, 12c, 12d 9a 9a 9a 5 6a 12a 25 12c 12d 12c 12d 5 6a 12b 12d 12d 6a 12c 12b 12c 12.5

B. subtilis ATCC 6633 E. coli ATCC 25922 E. coli from 12 urine 35

12b 12c

1016
Table VI. Minimum inhibitory concentration (MIC) values of standard commercial antibiotics measured on standard bacterial strains. Concentration (g/mL) STRAINS S. aureus NIH HUNGARY 118 003 S. saprophyticus NIH HUNGARY 120 008 E. coli ATCC 25922 M. luteus ATCC 9341 > 100 T 50 25 12.5 6.25 C 3.12 1.56 CRO 0.78 CXM 0.39 G A A 0.20 < 0.20 OXA P G OXA P CRO

CRO

C CXM C T G

OXA P

CXM A

B. subtilis ATCC 6633

C CXM

CRO

CXM CRO A OXA P G T OXA P

P = penicillin-G; OXA = oxacillin; T = oxytetracycline; A = ampicillin; CRO = ceftriaxone; CXM = cefuroxime; C = chloramphenicol; G = gentamicin.

for the preparation of cis-7-Benzylidene-2-carbamoyl3,3a,4,5,6,7-hexahydro-5-methyl-3-phenyl-2H-pyrazolo[4,3-c]pyridine (8a), at the end of the reaction the solution was made alkaline. The precipitate separated was ltered and washed with cold ethanol. Diastereoisomers 8a and 8b were separated by fractional recrystallization from methanol.

5.1.2. General procedure for the preparation of 2-isothiocarbamoylpyrazolo[4,3-c]pyridine hydrochlorides (8c HCl and 8d HCl) 5.51 mmol of the free base (8c or d) was dissolved in ethanol and 5 mL of 6 N HCl in ethanol was added. The salt separated was ltered off and recrystallized from methanol. 5.1.3. General procedure for the alkylation of 2-isothiocarbamoyl-substituted compounds (3, 4ab and 11a-g) Method A: an alkyl halide (9.3 mmol) was added to the solution of the 2-thiocarbamoyl compound (8.4 mmol) in anhydrous ethanol (100 mL). The reaction mixture was reuxed for 2 h with the exclusion of moisture till the disappearance of the starting 2-isothio-carbamoyl compound. The solvent was removed in vacuo and the residue was recrystallized from acetone. 5.1.4. cis-7-Benzylidene-3,3a,4,5,6,7-hexahydro-2-S-methylthiuronyl-3-phenyl-2H-indazole (6c) 6a (0.50 g, 1 mmol) was dissolved in ethanol (100 mL) and it was treated with concentrated ammonia solution (5 mL). The solution was poured into water and the separated precipitate was ltered off. It was recrystallized from methanol.

Table VII. In vitro cytotoxicity test of compounds on a HeLa cell-line. Compound 5 6a 9a 9d 12a 12b 12c 12d Cell death % at 250 g/mL concentration 87 35 90 73 72 29 8 17 LDt50 M/L 155.7 > 510.8 22.2 96.7 278.2 > 516.7 > 473.3 > 482.4

1017 5.1.5. General procedure for the alkylation of 2-isothiocarbamoyl-substituted compounds (8ad) Method B: alkyl halide (1.93 mmol) was added to the solution of the hydrochloride of the 2-isothiocarbamoyl compound (1.75 mmol) in anhydrous ethanol (65 mL). The reaction mixture was reuxed for 2.5 h with the exclusion of moisture until the disappearance of the starting 2-isothiocarbamoyl compound. The reaction mixture was made alkaline by using concentrated ammonia solution (0.7 mL) and poured into water. The precipitate was ltered off and washed with water until neutral. After drying the product was dissolved in ethanol and triturated with HCl gas. 5.1.6. General method for the preparation of S-methyl derivatives from dibenzylidene ketones and S-methylthiosemicarbazide hydroiodide (6a and 9a) Method C: the unsaturated ketone (20 mmol) and S-methylthiosemicarbazide hydroiodide (30 mmol) were dissolved in the mixture of ethanol (300 mL) and concentrated hydrochloric acid (10 mL). After 12 h boiling the reaction mixture was cooled down. The precipitate was ltered off, washed with cold ethanol and water until neutral. It was recrystallized from the mixture of acetone and methanol. The salt formed in the case of 9a was converted to the corresponding base that was treated with HCl to yield a dihydrochloride. These samples were in every respect identical with the product of the alkylation. 5.2. Biology 5.2.1. Microbial screens Twenty each of S. aureus and E. coli isolates (the most common representatives of Gram-positive and Gramnegative bacterial pathogens) of various clinical sources were selected for screening the antimicrobial effect of the compounds. The strains were maintained on nutrient agar medium. The test compounds were dissolved in nutrient broth (Difco) medium at a concentration of 50 g/mL, and 2 L of a Nutrient Broth starter culture of the bacterial strain to be tested was added to achieve a nal inoculum of ca. 5 105 colony forming units per mL [25]. The cultures were incubated overnight at 37 C. Inhibition was shown by no change in optical density. Nutrient broth medium without the compound served as control. Loopfuls of nutrient broth cultures were plated on nutrient agar to show if the effect of the compounds was bacteriostatic or bactericidal [29]. 5.2.2. Determination of minimum inhibitory concentration (MIC) value by test tube dilution method Eight of the most effective antibacterial compounds (5, 6a, 9a, 9d and 12ad) were tested on reference strains: S. aureus NIH Hungary 118003, Staphylococcus saprophyticus NIH Hungary 120008, Micrococcus luteus ATCC 9341, Bacillus subtilis ATCC 6633, E. coli ATCC 25922. The MIC values of these ve standard strains were tested on eight antibiotics commonly used in therapy: penicillin-G (P), Biogal, Debrecen, Hungary; oxacillin (OXA) and ampicillin (A), Bristol Myers Squibb Co., USA; oxytetracycline (T) and gentamicin (G), Chinoin, Budapest, Hungary; ceftriaxone (CRO) F, Hoffmann-La Roche AG, Basel, Switzerland; cefuroxime (CXM), GlaxoWellcome, Greenford, UK; chloramphenicol (C), EGIS, Budapest, Hungary. In addition the 9a compound was investigated on ve E. coli strains isolated from urine. The test conditions were similar as mentioned earlier. The exceptions: the compounds in 200 g/mL concentrations were diluted in medium containing 2.5% DMSO. Using backwards dilution of DMSO all tubes contained the same concentration of DMSO. Control tubes without compounds were used to check the effect of DMSO. Double dilution series of compounds were made in test tubes. After inoculation and 24 h incubation at 37 C, the MIC values were obtained from the lowest concentration of compound where the tubes remained clear, where the bacterial growth was inhibited. All experiments were performed in triplicate [29]. 5.2.3. In vitro cytotoxicity tests of compounds on HeLa cell-line OHIO A microplate technique was used. The DMEM1 (Sigma, Missouri, USA) growth medium contained 10% foetal bovine serum (Sigma, Missouri, USA). The tested compounds were solved in growth medium containing 2.5% DMSO. Double dilution series were set up from 250 g/mL concentrations. 1.5 104 cells/well were incubated for 24 h at 37 C. Cells killed by compounds were washed out, living cells were xed and stained by crystal violet in methanol. The remnants of stain were washed out. One hundred L of 1% SDS/well solved the cells under slow shaking (20 min). The cell lysates were measured in a Dynatech MR7000 photometer at 595 nm. Controls without compounds (100% cell) were used in determination of cell death percent. A graph of cell death percent vs. concentration was drawn. From this graph the LDt50 values were determined [30]. Acknowledgements We are indebted to Mrs Gabriella Nmeth, Mr Zoltn Nagy and Mr Attila Frjes for their valuable technical assistance and to Mrs Ildik Meleg and Mr Antal Kovcs for computer formulation of the manuscript.

1018 References
[1] [2] [3] [4] [5] [6] Tornetta B., Arcoria A., Boll. Sci. Fac. Chim. Ind. Bologna 14 (1956) 5051. Manowitz M., Walter G., J. Pharm. Sci. 54 (1965) 650. Bauer D.J., Sadler P.W., Br. J. Pharmacol. 15 (1960) 101110. Searle G.D., and Co. (1975) Brit. 1, 382, 773; Chem. Abstr. (1975) 83 p58829g. Hamilton R.W., (1976) U.S. 3, 940, 418 Chem. Abstr. (1976) 85 p5623v. Ramalingam K., Thyvelikakath G.X., Berlin K.D., Chesmut R.W., Brown R.A., Durham N.N., Ealick S.E., Van der Helm D., J. Med. Chem. 20 (1977) 847850. Lrnd T., Szab D., Fldesi A., Prknyi L., Klmn A., Neszmlyi A., J. Chem. Soc. Perkin Trans. 1 (1985) 481486. Tth G., Szllsy ., Lrnd T., Knya T., Szab D., Lvai A., J. Chem. Soc. Perkin Trans. 2 (1989) 319323. Szllsy ., Tth G., Lrnd T., Knya T., Aradi F., Lvai A., J. Chem. Soc. Perkin Trans. 2 (1991) 489493. Walter W., Krohn J., Chem. Ber. 102 (1969) 37863794. Sohr P., Cyclopentane derivatives and their heteroanalogues, in: Nuclear Magnetic Resonance Spectroscopy Vol. 2, CRC Press, Boca Raton, Florida, 1983, pp. 2223. Sohr P., The effect of structural parameters on the shielding of carbon nuclei, in: Nuclear Magnetic Resonance Spectroscopy Vol. 2, CRC Press, Boca Raton, Florida, 1983, pp. 154155. Grant D.M., Cheney B.V., J. Am. Chem. Soc. 89 (1967) 5315. Sohr P., Fused and quasiaromatic heterocyclic systems; analogues containing more nitrogens, in: Nuclear Magnetic Resonance Spectroscopy Vol. 2, CRC Press, Boca Raton, Florida, 1983, pp. 8990. Sohr P., Geminal couplings, in: Nuclear Magnetic Resonance Spectroscopy Vol. 1, CRC Press, Boca Raton, Florida, 1983, pp. 5560. Karpus M.J., J. Chem. Phys. 30 (1959) 11. [17] [18] [19] Karpus M.J., J. Chem. Phys. 33 (1960) 1842. Pegg D.T., Doddrell D.M., Bendall M.R., J. Chem. Phys. 77 (1982) 2745. Bendall M.R., Doddrell D.M., Pegg D.T., Hull W.E., High Resolution Multipulse NMR Spectra Editing and DEPT. Bruker, Karlsruhe (1982). Sohr P., Gated decoupling. Determination of NOE, and the absolute intensities of 13C-NMR signals, in: Nuclear Magnetic Resonance Spectroscopy Vol. 1, CRC Press, Boca Raton, Florida, 1983, pp. 196197. Sanders J.K.M., Mersch J.D., Prog. Nucl. Magn. Reson. 15 (1982) 353. Ernst R.R., Bodenhausen G., Wokaun A., Principles of Nuclear Magnetic Resonance in One and Two Dimensions, Clarendon Press, Oxford, 1987 pp. 471479. Balaban A.T., Zugravescu I., Avramovici S., Silhan W., Monatsh. Chem. 101 (1970) 704708. El-Rayyes N.R., Al-Jawhary A., J. Heterocycl. Chem. 23 (1968) 135140. Al-Nakib T.M., Bezjak V., Meegan M.J., Chandy R., Eur. J. Med. Chem. 25 (1990) 455462. Al-Nakib T.M., Perjsi P., Varghese R., Meegan M.J., Med. Princ. Pract. 6 (1997) 1421. Lvai A., Schg J.B., Pharmazie 34 (1979) 749. Krapcho J., Turk C.F., J. Med. Chem. 22 (1979) 207210. Sahm D.F., Washington J.A., Antibacterial susceptibility tests: dilution methods, in: Ballows A., Hausler Jr. W.J., Herrmann K.L., Isenberg H.D., Shadomy H.J. (Eds.), Manual of Clinical Microbiology 5th ed., ASM, Washington DC, 1991, pp. 11051116. Schlager S.I., Adams A.C., Use of dyes and radioisotopic markers in cytotoxicity tests, in: Langone J.J., Vunakis H.V., (Eds.), Methods in Enzymology, Vol. 93, Immunochemical Techniques Part F, Conventional Antibodies, Fc Receptors and Cytotoxicity, Academic Press Inc., San Diego, USA, (1983), pp. 233245.

[20]

[21] [22]

[7] [8] [9] [10] [11]

[23] [24] [25] [26] [27] [28] [29]

[12]

[13] [14]

[30]

[15]

[16]

Eur. J. Med. Chem. 34 (1999) 10231034 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1023

Original article

Synthesis and structure-activity relationship study of the new set of trypsin-like proteinase inhibitors
Pavol Zlatoidsky, Tibor Maliar*
Drug Research Institute, SK-90001 Modra, Slovak Republic (Received 8 October 1998; revised 2 April 1999; accepted 6 April 1999)

Abstract A new set of 25 trypsin-like proteinase inhibitors was prepared and the inhibiting activity on trypsin, thrombin, plasmin and urokinase was measured. The structure-activity relationship is discussed. High inhibiting activities were observed in 4-guanidinobenzoic acid esters only. The replacement of this moiety for N-formamidinyl-isonipecotic acid or an arginine moiety caused almost total loss of the activity. In the series of 4-guanidinobenzoic acid esters, any important inuence of the ester-groups reactivity was observed. The trypsin-thrombin selectivity in the compounds with the guanidine-remote carboxylic function was also observed. 1999 ditions scientiques et mdicales Elsevier SAS trypsine-like proteinases / inhibitors of proteinases / guanidine derivatives

1. Introduction Trypsin-like proteinases are serine proteinases for which hyperactivity can cause a variety of damage to health [1]. Recently, the activity of the urokinase type plasminogen activator (uPA) was discovered as an important starting proteinase in the proteolytic cascade of tumour invasion and metastases [2]. Inhibitors of these enzymes are of interest as potential therapeutics in various diseases. Recently, we have reported a set of 4-guanidinobenzoic acid esters with considerable inhibiting activity on trypsin [3]. Continuing in this research we have synthesised a further set of compounds to demonstrate the inuence of either ester group reactivity or guanidinebearing function moiety. 2. Chemistry Synthesis of the compounds 2ah was performed via selective esterication of the hydroxyaromatic acid with the corresponding halogen derivative, and followed by esterication of the produced compounds 1ah with
*Correspondence and reprints

4-guanidinobenzoic acid mesylate via the dicyclohexyl carbodiimide (DCC) method as described [3]. Reaction of the 4-aminobenzoic acid with chloroacetyl morpholine in the presence of triethylamine in acetonitrile gave a moderate yield of the corresponding ester 1l. After diazotation and nucleophile substitution with potassium ethyl dithiocarbonate and amonolysis of the intermediate, the thiole 1m was produced. This was esteried with 4-GBA/DCC to get 2l. Compound 2i was prepared by hydrogenolysis of benzylester 2c, compounds 2j and 2k were prepared by acidic hydrolysis of the tert. butyl esters 2g and 2h (gure 1). Compounds 4ae were prepared via activation of the free hydroxyaromatic acid with N-ethyl-N-(3dimethylaminopropyl)-carbodiimide hydrochloride (EDI) followed by aminolysis of the activated complex with the corresponding amine. The hydroxyaromatic amides 3ae were esteried by 4-guanidinobenzoic acid mesylate and DCC (gure 2). Synthesis of compounds 7a and b is mentioned in gure 3. N-chloroacetyl morpholine was reacted with potassium ethyl dithiocarbonate in ethanol. Following amonolysis with ammonia yielded N-mercaptoacetyl morpholine 5a. BOC-glycine was activated with ethyl chloroformate and reacted with morpholine. After acidic

1024

Figure 1. i: Halogen-CH2COY/MeCN, TEA. ii: 4-Guanidinobenzoic acid mesylate/DCC/pyridide/RT. iii: H2/Pd-C. iv: CF3COOH/ DCM/anisole. v: NaNO2, HCl, 2.KSCSOEt, 3.NH3.

hydrolysis of the BOC group, the free N-aminoacetyl morpholine was reacted with 4-acetoxybenzoyl chloride. Then, one-pot cleavage of the protective acetyl group via aqueous methanolic sodium carbonate gave 6a, as well as 6b which was prepared via the same reactions. Esterication by 4-guanidinobenzoic acid mesylate/DCC gave 7a and b (gure 3).

The sodium salt of 4-mercaptoanisole was the starting material for synthesis of compounds 10 and 11. It reacted with ethyl chloroacetate to give the corresponding ester which was deprotected with hydrobromic acid, both on phenolic and carboxylic hydroxyl, to give the acid 8. The acid 8 was esteried via the usual method with chloroacetyl morpholine to give compound 9. This was esteri-

1025 ed to get compound 10. After oxidation of 10 with hydrogen peroxide-acetic acid, compound 11 was prepared (gure 4). N1-BOC-N6,8 di-Z-Arginine was coupled with compounds 1a and 1d to get esters 12a and b. After hydrogenolysis of both Z-protective groups, compounds 13a and b were obtained (gure 5). Isonipecitic acid was protected by the BOC group to get 14 and coupled with compounds 1a and 1b via the DCC method to give esters 15a and b. After acidic cleavage of the BOC protective group, the free amines were reacted with cyanamide to give compounds 16a and b (gure 6 ). 3. Results and discussion All inhibiting activities are presented in table I. There are no remarkable differences in the inhibiting activity on trypsin, thrombin or urokinase in the rst group of compounds 2ah. After the benzene ring was replaced by the naphthalene ring, the increase of the selectivity for trypsin/thrombin is observable (compare

Figure 2. i: EDI/amine/TEA/MeCN. ii: 4-Guanidinobenzoic acid mesylate/DCC/pyridine.

Figure 3. i: 1. TEA, DCM, 2. Na2CO3, MeOH, H2O. ii: 4-Guanidinobenzoic acid mesylate/DCC/pyridine.

1026

Figure 4. i: 1. ClCH2COOEt/NaOET/EtOH, 2. HBr/AcOH. ii: N-Chloroacetyl morpholine/TEA/MeCN. iii: 4-Guanidinobenzoic acid mesylate/DCC/pyridine. iv: H2O2/AcOH.

2a versus 2e). The extremely high activity (picomolar range) of 2c on plasmin is remarkable. Tert. butyl esters 2g and 2h didnt show any remarkable activity. But after removal of the ester group, compounds 2i, 2j and 2k show remarkable selectivity for trypsin over thrombin. Introduction of the pyridine ring instead of the benzene ring to increase the ester function reactivity didnt lead to a considerable change in the activity, similarly to the compound 2l. Introduction of the thioester moiety was intended to increase the ester reactivity towards nucleophiles. No effect of the change of sulde (10, no M+) and sulfone 11 (M effect and proposed enhancement of the nucleophile reactivity) has been observed. Mild effects of conjugation on compounds of group 4 were observed (compare compounds 4a and 4b with 4d and 4e). Reduced activity was observed in the compound 7a (compare e.g. with 2a) but the thioester 7b shows the similar activity to the oxygen analogue 2a. The com-

pounds with an arginine moiety 13a, 13b and with an N-formamidinoisonipecotic acid moiety 16a and 16b are inactive. It is possible to conclude that the activity is not dependent on the ester function reactivity in nucleophilic substitution but only on the rate of hydrolysis of the 4-guanidinobenzoyl-trypsine complex which is the commonly accepted mechanism of the inhibiting mechanism of serine proteinases [4]. 4. Experimental protocols Melting points were measured on a Boetius microscope and are uncorrected. NMR spectra were run on a Varian 200 (200 MHz) using TMS as internal standard. All chemicals were supplied from Merck and Aldrich, solvents from Microchem (Slovakia) and enzymes and substrates from Sigma. HPLC was performed on a Pye-Unicam system using Tessek C-18 columns (25

1027

Figure 5. i: 1a or 1d/DCC, DCM/MeCN. ii: H2/Pd-C/EtOH/HCl.

0.25 cm) for analytical purposes and Labio C-18 (25 5 cm) for semipreparations. The gradient MeCN-H2O (each contained 0.05% of triuoroacetic acid) 090% was used at a ow rate of 1 mL/min (analytical) over 50 min, or 10 mL/min (semipreparations). GC-MS analyses were run on Carlo Erba 1106. N-Chloroacetyl-N-methylpiperazine hydrochloride and 4-guanidinobenzoic acid mesylate were prepared according to the described procedure [3], as well as N2BOC-N6, N8-di-Z-Arginine [5]. 4.1. N-chloroacetyl glycine benzyl ester Glycine benzylester tosylate (16.85 g, 0.05 mol) were dissolved in 250 mL of dry dichloromethane and 26 mL (0.1 mol) of dry triethylamine were added. The mixture was cooled to 0 C, and 3.8 mL of chloroacetyl chloride in 15 mL of dry dichloromethane were added dropwise. The mixture was stirred for 1 h at room temperature, extracted with water, sat. NaHCO3 and brine, dried over Na2SO4, evaporated and used without further purication. The analytical sample was puried by high vacuum (0.0005 torr) molecular distillation (bath temperature, 80 C), to give the solidifying oil. Yield: 11.3 g, 74% (M.p. after solidication, 3335 C). MS: M+-H+ = 241,

Figure 6. i: BOC2O/Na2CO3/wt. dioxane. ii: 1a or 1b/DCC/ DMAP/MeCN-DCM. iii: 1. EtOAc/HCl, 2. cyanamide/concd./ HCl/EtOH.

1028
Table I. Inhibiting activities. IC50 (nM) of compounds 2a-l, 7a, 7b, 10, 11, 13a, 13b, 16a, 16b on trypsin, thrombin, plasmin and urokinase. Compound 2a 2b 2c 2d 2e 2f 2g 2h 2i 2j 2k 2l 4a 4b 4c 4d 4e 7a 7b 10 11 13a 13b 16a 16b Trypsin 10 10 8 65 26 5 32 375 10 19 700 8 32 140 30 13 13 75 17 15 15 4.4 105 4.72 105 1.3 104 1.7 104 Thrombin 35 94 12 475 1 030 40 78 435 726 2 557 8 310 34 447 1 018 192 249 218 321 44 22 64 NT NT NT NT Plasmin 13 28 0.05 1 163 446 8 336 5 025 38 381 8 015 36 343 1 450 79 76 49 129 8 507 291 NT NT NT NT Urokinase 75 55 32 130 715 106 710 1 005 43 213 2 235 11 117 130 NT 45 21 384 22 31 33 NT NT NT NT

NT = not tested.

1a: m.p. 185186 C (iPr2O/EtOAc), yield: 1.48 g, 56%. Elemental anal. for C13H15NO5 (C, H, N). 1H-NMR (CDCl3): 3.433.46 (2H), 3.553.64 m (6H), 4.95 s (2H), 6.74 d (2H), J = 7.56 Hz), 7.81 d (2H, J = 7.55 Hz). 1b: m.p. 201203 C (iPr2O/EtOAc), yield: 1.33 g, 48%. Elemental anal. for C14H18N2O4 (C, H, N). 1HNMR (CDCl3): 2.54 s (3H), 2.55 m (2H), 3.50 m (4H), 3.65 m (2H), 6.78 d (2H), 7.68 d (2H). 1c: m.p. 65 C (MeOH/iPr2O), yield: 1.12 g, 46%. Elemental anal. for C18H17NO6 (C, H, N). 1H-NMR (CDCl3): 3.5 bs (NH), 4.04 d (2H), 4.75 s (2H), 5.22 s (2H), 6.88 d (2H, J = 7.45 Hz), 7.43 m (5H), 7.97 d (2H, J = 7.44 Hz). 1d: m.p. 7980 C (iPr2O/EtOAc), yield: 1.36 g, 48%. Elemental anal. for C16H22N2O4 (C, H, N). 1H-NMR: 2.54 s (3H, NCH3), 2.63 t (2H, J = 5.79 Hz), 2.89 m (2H), 3.00 t (2H, J = 5.81 Hz), 3.45 m (6H), 6.75 d (2H, J = 7.36 Hz), 7.43 d (2H, J = 7.38 Hz). 1e: m.p. 165167 C (hexane/EtOAc), yield: 1.61 g, 51%. Elemental anal. for C17H17NO5 (C, H, N). 1H-NMR (CDCl3): 3.65 m (8H), 4.89 s (2H), 7.247.31 m (broad, 4H), 7.56 dd (1H), 7.807.86 m (4H). 1f: m.p. 212214 C (iPr2O/MeOH), yield: 1.75 g, 63%. Elemental anal. for C15H18O5 (C, H). 1H-NMR (DMSO): 3.75 m (4H), 4.28 m (4H), 4.97 m (2H), 6.53 d (1H), 8.91 dd (1H), 8.24 d (1H, J = 6.56 Hz). 1g: m.p. 134 C (hexane/EtOAc), yield: 1.80 g, 65%. Elemental anal. for C15H18O5 (C, H). 1H-NMR (CDCl3): 1.52 s (9H), 4.66 s (2H), 6.19 d (1H, J = 9.78 Hz), 6.27 d,

other peaks: 192 (M+-CH2Cl), 164 (M+-ClCH2CO), 150 (M+-tropylium), 91 (tropylium). 1H-NMR (CDCl3): 2.38 s (2H, ClCH2), 3.68 d (2H, CH2NH), 3.88 s (2H, CH2O), 5.85 bs (1.3 H, NH), 6.496.56 m (5H, H arom). 4.2. N-chloroacetyl morpholine Chloroacetyl chloride (7.5 mL, 0.1 mol), dissolved in 50 mL of dry ether, was added dropwise to the stirred solution of morpholine (22 mL, 0.2 mol) in 300 mL of dry ether at 20 C. The mixture was stirred for 30 min at room temperature, the morpholinium chloride was ltered and washed with ether, the ltrate was evaporated and the oily residue distilled in vacuo. B.p. 6769 C/0.6 torr. Yield: 15.2 g, 93%. MS: 162 (M+-H+), other peaks: 114 (M+-ClCH2), 86 (morpholinyl). 1H-NMR: 2.22 s (2H, ClCH2), 3.65 m (4H), 4.44 m (2H) and 4.56 m (2H). 4.3. Typical procedure for preparation of hydroxyderivatives 1ah was described in [3] Physical-analytical data for each:

Figure 7. i. Nchloroacetyl morpholin, MeCN, TEA, ii. 1. NaNO2, HCL, 2. KSCS (OEt), 3. wt. NH3, EtOH, iii. 4guanidinobenzoic acid/DCC/pyridine.

1029 6.84 d (2H, J = 7.55 Hz), 6.99 bs (1H), 7.33 d (2H, J = 7.56 Hz) 7.61 d (1H, J = 7.78 Hz). 1h: m.p. 129 C (iPr2O/EtOAc), yield: 1.19 g, 39%. Elemental anal. for C16H19NO5 (C, H, N). 1H-NMR (CDCl3): 1.59 s (9H), 2.28 s (2H), 4.28 s (2H, OCH2), 6.18 m (1H), 6.29 dd (1H), 7.36 dd (1H), 7.77 m (1H), 8.65 bs (1H, NH). 4.4. Typical procedure for preparation of hydroxyamides 3ae 4-hydroxyaromatic acid (0.01 mol) was suspended in acetonitrile (25 mL) and N-(3-dimethylaminopropyl)N-ethylcarbodiimide (1.92 g, 0.01 mol) were added at once. The acid dissolved immediately. Then, triethylamine (1.34 mL 0.01 mol) and the amine (0.01 mol) were added and the mixture was stirred for 18 h. After evaporation of volatile compounds in vacuo, the residue was distributed between water (50 mL) and ethyl acetate (20 mL) and the water phase was twice more extracted with ethyl acetate. Joint organic extracts were washed with brine, dried (Na2SO4) and evaporated in vacuo. The residue was recrystallised. 3a: m.p. 130 C (iPr2O/EtOAc), yield: 1.79 g, 76%. Elemental anal. for C13H17NO3 (C, H, N). 1H-NMR (CDCl3): 2.56 t (2H, J = 5.55 Hz), 2.87 t (2H, J = 5.56 Hz), 3.33 m (2H), 3.50 m (2H), 6.75 d (2H, J = 7.56 Hz,) 7.03 d (2H, J = 7.55 Hz), 9.9 bs (1H). 3b: Syrup. Data for fumarate: m.p. 141142 C (THF), yield: 2.88 g 79%. Elemental anal. for C18H24N2O6 (C, H, N). 1H-NMR (DMSO): 2.18 s (3H), 2.25 t (2H), 2.52 t (2H), 2.66 m (4H), 3.40 m (4H), 6.59 s (2H, fumarate), 6.45 d (2H, J = 7.59 Hz), 7.01 d (2H, J = 7.60Hz). 3c: m.p. 6567 C (iPr2O/Hexane), yield: 2.72 g, 84%. Elemental anal. for C15H20N2O6 (C, H, N). 1H-NMR (CDCl3): 2.36 t (2H, J = 5.55 Hz), 3.83 t (2H, J = 5.57 Hz), 3.36 m (2H), 3.76 m (6H), 5.01 s (2H), 6.75 d (2H, J = 7.28 Hz), 7.18 d (2H, J = 7.26 Hz). 3d: m.p. 213 C (hexane/EtOAc), yield: 1.51 g, 65%. Elemental anal. for C13H15NO3 (C, H, N). 1H-NMR (CDCl3): 3.70 m (8H), 6.63 (1H, J = 10.27 Hz), 6.86 d (2H J = 7.45 Hz), 7.41 d (2H, J = 7.44 Hz), 7.67 d (1H, J = 10.26 Hz). 3e: m.p. 230235 C decomp. (hexane/EtOAc), yield: 1.79 g, 71%. Elemental anal. for C16H15NO2 (C, H, N). 1 H-NMR (CDCl3): 4.56 s (2H), 6.31 d (1H, J = 11.0 Hz), 7.1 d (2H, J = 7.67 Hz), 7.267.29 m (5H), 7.52 d (2H, J = 7.67 Hz), 7.69 d (1H, J = 10.9 Hz). 4.5. N-(Mercaptoacetyl)-morpholine (5b) Chloroacetyl morpholin (2.95 g, 0.02 mol) was dissolved in 35 mL of absolute ethanol and potassium ethyldithiocarbonate (4 g, 0.025 mol) was added at once. The mixture was stirred under N2 and reuxed for 6 h, cooled and 35 mL of water ammonia were added and stirred overnight. The mixture was extracted three times with dichloromethane, dried (Na2SO4) and evaporated in vacuo. After distillation under reduced pressure, the pale yellow oil was obtained. B.p. 7274 C/0.45 torr, yield: 2.7 g, 86%. GC-MS: 161 (M+), others: 114 (Nmorpholinylcarbonyl), 86 (morpholinyl), 46 (CH2=S+). 1 H-NMR (CDCl3): 1.11 s (1H), 2.36 s (2H), 3.753.79 m (8H). 4.6. Typical procedure for the preparation of compounds 6a and b In 35 mL of dry dichloromethane, N-(aminoacetyl)morpholin or N-(mercaptoacetyl)-morpholin 5b (0.02 mol) were dissolved, triethylamine (2.7 mL, 0.02 mol) were added and cooled to 10 C. Then, the solution of 4-acetoxybenzoyl chloride 2 (3.95 g, 0.02 mol) in 35 mL of dry DCM were added dropwise at 10 C. The mixture was stirred for 2 h at room temperature, evaporated in vacuo and the residue was dissolved in 30 mL of methanol. Then, 30 mL of sat. Na2CO3 were added and the mixture was stirred for an additional 1 h. It was partly evaporated in vacuo, extracted three times with ethyl acetate, washed with brine, dried (Na2SO4), evaporated in vacuo and recrystallised. 6a: m.p. 196197 C (EtOAc/hexane), yield: 4.1 g, 77%. Elemental anal. for C13H16N2O4 (C, H, N). 1HNMR: 3.453.59 m (8H), 4.26 d (2H, J = 6.96 Hz, CH2NH), 6.79 d (2H, J = 7.56 Hz), 7.10 d (2H, J = 7.56 Hz). 6b: m.p. 153154 C (EtOAc/hexane), yield: 3.8 g, 72%. Elemental anal. for C13H15NO3S (C, H, N, S, % S calcd. 12.08, found 12.49). 1H-NMR: 3.383.63 m (8H), 3.79 s (2H, CH2S), 6.83 d (2H, J = 7.49 Hz), 7.21 d (2H, J = 7.50 Hz). 4.7. Ethyl methoxyphenyl thioacetate Sodium (2.3 g, 0.1 mol) was dissolved in 250 mL of absolute ethanol and 4-mercaptoanisole (14 g, 0.1 mol) were added. After being stirred for 1 h at room temperature under N2, 11.4 g (0.1 mol) of ethyl chloroacetate was added. The mixture was stirred and reuxed under N2 for 6 h, evaporated in vacuo and the residue was distributed between water (100 mL) and ether (50 mL). The water phase was twice more extracted with ether, washed with brine, dried (Na2SO4), evaporated in vacuo and distilled in vacuo. B.p. 7274 C/0.5 torr, yield: 13.9 g, 88%. GC-MS: 226 (M+), other peaks: 153 (MeO-C6H4S=CH2), 107 (MeO-Ph+), 46 (CH2=SH+). 1H-NMR

1030 (CDCl3): 1.11 t (3H, ethyl), 3.43 d (2H, SCH2), 4.11 s (3H, CH3O), 4.76 q (2H, ethyl), 6.66 d (2H, J = 7.34 Hz), 7.24 d (2H, J = 7.34 Hz). 4.8. 4-hydroxyphenyl-1-thioacetic acid (8) Ethyl-4-methoxyphenyl thioacetate (12.9 g, 0.057 mol) was dissolved in 50 mL of glacial acetic acid and 25 mL of conc. hydrobromic acid was added. The mixture was gently reuxed under N2 for 25 min, cooled and evaporated to dryness. The residue was crystallised from water-ethanol. M.p. 266267 C. Yield: 5.8 g, 56%. Elemental anal. for C8H8O3S (C, H, N, S, % S calc. 17.40%, found 17.86%). 1H-NMR (DMSO): 3.76 s (2H, CH2), 6.87 d (2H, J = 7.88 Hz), 7.87 d (2H, J = 7.87 Hz), 10.11 bs, (1.1H, COOH). 4.9. 4-aminobenzoyloxy-acetylmorpholin (1l) The compound was prepared according to the procedure described for compounds 1ah, starting from 1.25 g (0.01 mol) of 4-aminobenzoic acid. M.p. 177 C (MeOH/ diisopropyl ether), yield: 0.9 g, 36%. Elemental anal. C13H16NO4 (C, H, N). 1H-NMR (DMSO): 3.73 m (2H), 3.83 m (6H), 4.76 s (2H), 6.55 d (2H, J = 7.23 Hz), 7.43 d (2H, J = 7.34 Hz). 4.10. N-(4-Mercaptobenzoyloxy)-acetyl morpholine (1m) N-(4-aminobenzoyl)-oxyacety morpholine 1l (1.77 g, 0.0072 mol) was dissolved in 20 mL of water and 1.5 mL of conc. HCl was added. The mixture was cooled to 0 C, and 0.55 g (0.008 mol) of sodium nitrite in 5 mL water was added dropwise at 0 C. Then, the mixture was stirred for 15 min at the same temperature and 1.28 g (0.008 mol) of potassium ethyl dithiocarbonate in 10 mL of water was added portionwise. The mixture was stirred for 10 min and heated gradually to 60 C, and then stirred for 2 h at room temperature. The precipitated compound was dissolved by addition of 20 mL of dioxane and 20 mL of water ammonia was added and stirred for 2 h more. The mixture was evaporated partly in vacuo, extracted with four portions of ethyl acetate, washed with sat. NaHCO3, brine, dried (Na2SO4), evaporated in vacuo and recrystallised from diisopropyl ether and methanol. M.p. 165166 C. Yield: 0.97 g, 48%. Elemental anal. C13H15NO4S (C, H, N, S, % S calc. 11.39, found 11.10). 1 H-NMR (CDCl3): 1.17 s (1.3H, SH), 3.56 m (2H), 3.93 (6H), 4.55 s (2H, CH2O), 6.14 d (2H, J = 6.93 Hz), 7.25 d (2H, J = 7.01 Hz). 4.11. Preparation of compounds 2ah, 2l, 4ae, 7a, 7b and 10 (typical procedure) 4-Guanidinobenzoic acid mesylate (1.38 g, 0.005 mol) was dissolved in 40 mL of dry pyridine and 0.005 mol of the hydroxycompound were added. The mixture was cooled to 0 C, and DCC (1.03 g 0.005 mol) were added at once. The mixture was stirred overnight under chlorcalcium tube at room temperature, precipitated dicyclohexyl urea was ltered, washed with two small portions of pyridine and the ltrate was poured into 350 mL of ether. The precipitated product was dissolved in methanol (3040 mL) and passed through the column of 50 g of neutral Al2O3, which was washed with methanol (ca. 50 mL). The solution was evaporated in vacuo to ca. 10 mL volume, and it was poured into 50 mL of dry ethyl acetate. The product was ltered, washed with ethyl acetate, dried in vacuo and the purity was monitored by HPLC. 2a: amorphous, yield: 1.06 g, 43%. Elemental anal. for C22H26N4O9S (C, H, N, S) % S calc. 6.14, found 5.79. 1 H-NMR (DMSO): 2.98 s (3H), 3.73 m (8H), 5.06 s (2H), 7.21 d (2H), 7.24 d (2H), 7.86 d (2H), 7.99 d (2H). 13 C-NMR (DMSO): 41.67, 44.36, 62.09, 65.96 (OCH2), 122.43, 122.57, 124.89, 127.20, 129.47, 131.52, 131.14, 155.45, 162.33, 163.57, 169.65. 2b: amorphous, yield: 1.04 g, 39%. Elemental anal. for C23H29N5O8S (C, H, N, S). 1H-NMR (DMSO): 2.96 s (3H), 3.34 m (2H), 3.36 s (3H), 3.55 m (6H), 4.99 s (2H), 7.11 d (2H), 7.23 d (2H), 7.76 d (2H), 8.06 d (2H). 13 C-NMR (DMSO): 38.46, 45.36, 45.76, 47.31, 47.26, 51.31, 58.42, 122.49, 124.8, 127.1, 151.1, 131.3, 151.6, 141.5, 155.6, 163.5, 164.6, 171.3. 2c: after semi-preparative HPLC purication, 250 mg of the compound was puried. Amorphous, Elemental anal. for C27H29N4O10S (C, H, N, S). 1H-NMR (DMSO): 2.43 s (3H), 3.46 s (2H, OCH2Ar), 3.84 s (2H, NHCH2), 5.53 s (2H, OCH2CO), 6.90 d (2H, J = 7.55 Hz), 7.29 m (7H), 7.96 d (2H, J = 7.56 Hz), 8.25 d (2H, J = 8.91 Hz), 9.72 bs (3H). 13 C-NMR (DMSO): 39.50 (mesylate), 40.97 (OCH2Ar), 61.31 (OCH2CO), 115.4, 117.0, 120.1, 122.6, 125.3, 126.6, 127.6, 128.5, 128.2, 129.7, 130.3, 134.9, 155.4, 163.2, 164.9, 165.3, 170.6. 2d: amorphous, yield: 1.40 g, 45%. Elemental anal. for C27H30N5O8S. 1H-NMR (DMSO): 2.37 s (3H, mesylate), 2.39 s (3H, CH3N), 2.75 t (2H, J = 4.78 Hz, CH2), 2.90 t (2H, J = 4.78 Hz, CH2), 3.66 m (8H), 4.78 s (2H, OCH2), 6.90 d (2H, J = 8.76 Hz), 7.10 d (2H, J = 7.56 Hz), 7.32 d (2H, J = 8.77 Hz), 7.90 d (2H, J = 8.76 Hz). 13 C-NMR (DMSO): 29.4 (CH2), 34.7 (CH2), 39.4 (mesylate), 45.6, 54.11, 54.2, 61.3 (OCH2), 119.0, 121.7,

1031 121.8, 122.46, 129.18, 130.87, 137.8, 149.1, 153.7, 156.5, 164.6, 171.7. 2e: amorphous, yield: 1.17 g, 45%. Elemental anal. for C24H28N4O9S (C, H, N, S) % S calc. 5.84, found 5.41. 1 H-NMR (DMSO): 2.44 (3H, mesylate), 3.45 m (8H), 5.05 s (1H), 7.27.32 m (5H), 7.517.53 m (2H), 7.777.79 m (1H), 8.64 m (1H). 13C-NMR (DMSO): 38.9 (mesylate), 40.7, 43.7, 60.4 (OCH2), 66.3, 118.3, 119.2, 124.6, 126.2, 126.9, 129.5, 129.9, 130.5, 131.6, 133.7, 136.48, 143.8, 146.3, 155.8, 165.8, 1265.9, 167.4. 2f: amorphous, yield: 1.19 g, 45%. Elemental anal. for C21H25N5O9S (C, H, N, S). 1H-NMR (DMSO): 2.36 s (3H, mesylate), 3.553.74 2d (8H), 5.05 s (2H), 6.57 d (1H), 7.36 d (2H, J = 8.76 Hz), 7.65 d (2H, J = 8.77 Hz), 8.04 dd (1H), 8.29 d (1H). 13C-NMR (DMSO): 39.5 (mesylate), 43.43, 46.15, 62.68 (OCH2), 67.5, 67.6, 111.3, 120.4, 121.5, 124.3, 130.34, 131.3, 135.4, 141.8, 141.9, 155.7, 160.3, 165.8, 174.6. 2g: amorphous, yield: 1.03 g, 39%. Elemental anal. for C25H29N3O9S (C, H, N, S). 1H-NMR, (DMSO): 1.49 s (9H, tBuO), 2.46 (3H, mesylate), 4.66 s (2H, OCH2), 6.69 d (1H, J = 11.35 Hz, CH=), 7.19 d (2H, J = 7.45 Hz), 7.45 d (2H, J = 8.1 Hz), 7.71 d (1H, J = 11.36 Hz), 7.89 d (2H, J = 7.98 Hz), 8.17 d (2H, J = 7.44 Hz). 13C-NMR (DMSO): 27.65, 39.65 (MsOH), 61.14 (OCH2), 81.58, 117.3, 122.43, 122.6, 124.7, 130.5, 131.5, 131.9, 142.1, 143.3, 152.2, 155.6, 163.8, 165.6, 166.8. 2h: amorphous, yield: 1.12 g, 40%. Elemental anal. for C25H30N4O9S (C, H, N, S) % S calc. 5.70, found 5.39. 1 H-NMR (DMSO): 1.42 s (9H, tBuO), 2.38 s (3H, MsOH), 4.58 s (2H, OCH2), 6.90 d (2H, J = 7.34 Hz), 7.12 dd (1H), 7.367.40 m (2H), 7.86 d (2H, J = 7.83 Hz), 8.88 s (1H, NH ind.). 13C-NMR (DMSO): 27.5 (tBu), 36.6 (MsOH), 37.2 (CH2Ar), 56.6 (OCH2), 111.1, 118.69, 119.3, 121.6, 121.8, 121.9, 123.5, 123.9, 125.8, 131.6, 136.3, 141.4, 142.3, 155.1, 161.3, 163.8, 170.9. 2l: amorphous, yield: 1.29 g, 48%. Elemental anal. for C22H26N4O8S2 (C, H, N, S). 1H-NMR (DMSO): 2.34 s (3H, MsOH), 3.56 m (2H), 3.87 m (6H), 7.23 d (2H, J = 7.56 Hz), 7.36 d (2H, J = 7.80 Hz), 7.85 d (2H, J = 7.55 Hz), 7.98 d (2H, J = 7.58 Hz). 13C-NMR (DMSO): 36.7 (MsOH), 45.5, 46.1, 59.1 (OCH2), 65.5, 116.3, 117.5, 117.9, 118.0, 118.2, 121.3, 138.3, 139.8, 155.8, 160.2, 162.1, 165.3. 4a: amorphous, yield: 1.19 g, 46%. Elemental anal. for C24H28N4O7S (C, H, N, S). 1H-NMR (DMSO): 2.36 s (3H, MsOH), 2.64 t (2H, J = 8.96 Hz), 2.86 t (2H, J = 7.77 Hz), 3.43 m (8H), 7.11 d (2H, J = 6.98 Hz), 7.17 d (2H, J = 7.85 Hz), 7.37 d (2H, J = 7.00 Hz), 8.04 d (2H, J = 7.85 Hz). 13C-NMR (DMSO): 33.3 (CH2), 33.7 (CH2), 39.7 (MsOH), 40.4, 66.1 (OCH2), 121.6, 122.0, 122.7, 129.4, 131.2, 138.9, 148.9, 149.5, 154.4, 164.2, 170.0. 4b: amorphous, yield: 1.14 g, 44%. Elemental anal. for C24H31N5O6S (C, H, N, S). 1H-NMR (DMSO): 2.37 s (3H, Me-N), 2.46 s (3H, MsOH), 2.73 t (2H, J = 5.58 Hz), 2.85 t (2H, J = 5.53 Hz), 3.32 m (6H), 3.64 m (2H), 7.19 d (2H, J = 7.45 Hz), 7.36 d (2H, J = 8.24 Hz), 7.45 d (2H, J = 7.46 Hz), 8.17 d (2H, J = 8.22 Hz). 13C-NMR (DMSO): 29.9 (CH2), 33.6 (CH2), 39.3 (MsOH), 43.1 (MeN), 44.1, 52.9, 53.1, 121.5, 122.5, 122.7, 125.3, 129.5, 138.1, 148.7, 155.5, 164.1, 170.1. 4c: amorphous, yield: 1.17 g, 44%. Elemental anal. for C24H31N4O8S (C, H, N, S). 1H-NMR (DMSO): 2.38 s (3H, MsOH), 2.56 t (2H, J = 5.87 Hz), 2.89 t (2H, J = 5.88 Hz), 3.28 bs (2H), 3.87 m (6H), 4.22 d (2H, CH2NH), 6.66 bs (1H, NH), 6.87 d (2H, J = 7.81 Hz, 7.12 d (2H, J = 7.56 Hz), 7.87 d (2H, J = 7.87 Hz), 7.99 d (2H, J = 7.53 Hz), 10.11 bs (2H, NH), 11.1 bs (1.5H, NH). 13 C-NMR (DMSO): 30.1 (CH2), 31.7 (CH2), 38.9 (MsOH), 40.0, 45.7, 50.2, 53.3, 61.1 (OCH2), 120.1, 122.21, 123.1, 123.8, 125.7, 126.6, 129.3, 138.3, 141.5, 151.3, 158.7, 161.6, 170.0. 4d: amorphous, yield: 0.88 g, 38%. Elemental anal. for C20H26N4O7S (C, H, N, S), % S calc. 6.87, found 6.45. 1 H-NMR (DMSO): 2.38 s (3H, MsOH), 3.60 m (8H), 6.67 d (1H, J = 10.26 Hz), 7.23 d (2H, J = 7.83 Hz), 7.36 d (2H, J = 7.22 Hz), 7.66 d (1H, J = 10.3 Hz), 8.06 d (2H, J = 7.28 Hz), 8.11 d (2H, J = 7.81 Hz). 13C-NMR (DMSO): 39.91 (MsOH), 42.5, 45.7, 66.2, 66.3, 118.13, 122.3, 122.5, 122.7, 122.8, 129.3, 131.4, 132.9, 140.7, 147.1, 151.6, 154.7, 164.1, 164.5. 4e: amorphous, yield: 1.22 g, 48%. Elemental anal. for C25H26N4O6S (C, H, N, S). 1H-NMR (DMSO): 2.39 s (3H, MsOH), 2.75 s (2H, CH2NH), 6.72 d (1H, J = 10.3 Hz), 7.27 m (5H), 7.62 d (2H, J = 7.80 Hz), 7.71 d (2H, J = 7.21 Hz), 8.08 d (1H, J = 10.4 Hz), 8.17 d (1H, J = 7.85 Hz), 8.24 d, (2H, J = 7.23 Hz). 13C-NMR (DMSO): 39.5 (MsOH), 44.37, 116.7, 122.1, 122.3, 123.3, 123.4, 125.1, 128.3, 128.44, 129.3, 129.8, 130.2, 140.9, 145.3, 155.8, 168.8, 178.2. 7a: amorphous, yield: 1.28 g, 49%. Elemental anal. for C22H27N5O8S (C, H, N, S). 1H-NMR (DMSO): 2.41 s (3H, MsOH), 3.48 m (2H), 3.46 m (6H), 5.62 d (2H, CH2NH), 7.11 d (2H, J = 7.77 Hz), 7.49 d (2H, J = 7.42 Hz), 7.58 d (2H, J = 7.77 Hz), 8.11 d (2H, J = 7.39 Hz). 13 C-NMR (DMSO): 33.3 (MsOH), 40.1, 41.8, 47.5, (NHCH2), 66.1, 114.8, 122.7, 124.8, 129.1, 131.5, 141.3, 148.5, 155.4, 160.2, 165.1, 167.9. 7b: amorphous, yield: 1.00 g, 38%. Elemental anal. for C21H26N4O8S2 (C, H, N, S). 1H-NMR (DMSO): 2.41 s (3H, MsOH), 3.40 m (2H), 3.57 m (6H), 4.14 s (2H, SCH2), 7.39 d (2H, J = 7.33 Hz), 7.59 d (2H, J = 7.87 Hz), 8.05 d (2H, J = 7.34 Hz), 8.19 d (2H, J = 7.87 Hz). 13 C-NMR (DMSO): 39.5 (MsOH), 42.1, 45.9, 52.7

1032 (SCH2), 66.4, 115.3, 122.4, 128.5, 129.2, 130.7, 131.6, 133.4, 155.9, 162.0, 163.6, 165.2. 10: amorphous, yield: 1.01 g, 37%. Elemental anal. for C23H28N4O9S2 (C, H, N, S). 1H-NMR (DMSO): 2.42 s (3H, MsOH), 3.37 m (2H), 3.63 m (6H), 4.02 s (2H, SCH2), 4.87 s (2H, OCH2), 7.21 d (2H, J = 7.45 Hz), 7.32 d (2H, J = 8.00 Hz), 7.51 d (2H, J = 7.56 Hz), 8.13 d (2H, J = 8.02 Hz). 13C-NMR (DMSO): 39.7 (MsOH), 41.5, 44.18, 44.27 (SCH2), 62.2 (OCH2), 65.8, 116.1, 122.6, 122.9, 129.7, 130.9, 133.3, 134.4, 149.1, 155.3, 164.0, 164.6, 168.9. 4.12. Preparation of compound 2i In the common apparatus for hydrogenation under normal pressure, was dissolved 850 mg (1.12 mmol) of the compound 2c in 10 mL of methanol, and 21 mg of 10% Pd on charcoal were added. The hydrogenation proceeded for 4 h, the catalyst was ltered, washed with methanol, the methanolic solution was partly evaporated, and the product was precipitated with an excess of ethyl acetate. After purication by semi-preparative HPLC and lyophilisation, 420 mg of the product was gained as an amorphous powder (yield: 74%). Elemental anal. for C20H22N4O10S. (C, H, N, S). 1H-NMR (DMSO): 2.80 s (3H, MsOH), 3.61 d (2H, NHCH2), 4.75 s (2H, OCH2), 7.33 d (2H, J = 7.11 Hz), 7.45 d (2H, J = 7.34 Hz), 8.06 d (2H, J = 7.09 Hz), 8.21 d (2H, J = 7.14 Hz), 11.01 bs (3.5H, NH and COOH). 13C-NMR (DMSO): 20.4 (MsOH), 41.4 (NHCH2), 62.4 (OCH2), 114.4, 122.1, 122.3, 124.2, 126.8, 130.9, 131.1, 131.4, 132.9, 142.2, 155.9, 163.5, 164.2, 165.9, 172.7. 4.13. Preparation of compounds 2j and 2k (general procedure) The tert. butyl ester 2g or 2h (0.005 mol) was dissolved in 40% triuoroacetic acid in dry DCM (10 mL), containing 2 mL of anisole, and stirred for 15 min. The volatile compounds were removed in vacuo and the oily residue was treated with ethyl acetate. The precipitate was ltered and puried by semi-preparative HPLC. After lyophilisation of the appropriate fraction, amorphous product was obtained. 2j: yield: 354 mg, 56%. Elemental anal. for C22H20N3O9S (C, H, N, S) % S calc. 6.38, found 6.80. 1 H-NMR (DMSO): 2.42 s (3H, MsOH), 4.70 s (2H, OCH2), 6.78 d (1H, J = 10.3 Hz), 7.36 d (2H, J = 6.91 Hz), 7.71 d (2H, J = 7.87 Hz), 7.84 d (1H, J = 10.1Hz), 7.94 d (2H, J = 6.92 Hz) 8.28 d (2H, J = 7.87 Hz), 9.9 bs (4H, COOH and NH). 13C-NMR: 39.9 (MsOH), 60.6 (OCH2), 117.5, 122.5, 122.7, 122.8, 129.8, 130.8, 130.9, 131.0, 131.5, 131.8, 140.9, 144.5, 155.3, 161.7, 165.6, 169.1. 2k: amorphous, yield: 372 mg, 59%. Elemental anal. for C21H22N4O9S (C, H, N, S). 1H-NMR (DMSO): 2.38 s (3H, MsOH), 2.56 t (2H, J = 5.87 Hz), 2.89 t (2H, J = 5.88 Hz), 3.28 bs (2H), 3.87 m (6H), 4.22 d (2H, CH2N), 6.66 bs (1H, NH), 6.87 d (2H, J = 7.84 Hz), 7.12 d (2H, J = 7.56 Hz), 7.87 d (2H, J = 7.87 Hz), 7.99 d (2H, J = 7.54 Hz), 10.11 bs (2H), 11.10 bs (1H). 13C-NMR (DMSO): 30.1 (CH2), 31.7 (CH2), 38.9 (MsOH), 40.0, 45.7, 50.2, 53.3, 61.11 (OCH2), 120.1, 122.2, 123.1, 123.8, 125.7, 126.6, 129.3, 138.3, 141.5, 151.3, 158.7, 161.6, 170.00. 4.14. Preparation of compound 11 Compound 10 (1.45 g, 2.55 mmol) were dissolved in 15 mL of glacial acetic acid and cooled to 0 C. Then, 4.5 mL of 30% water hydrogen peroxide were added and the mixture was stirred overnight at room temperature. The mixture was evaporated in vacuo (bath temp. max. 30 C), the gummy residue dissolved in methanol and precipitated with ethyl acetate. 1.22 g (87%) of the crude compound was gained. 350 mg of this compound were puried by semi-preparative HPLC and proceeded to the analyses and inhibiting activities. Elemental anal. C23H28N3O11S (C, H, N, S). 1H-NMR (DMSO): 2.42 s (3H, MsOH), 3.36 m (4H), 3.56 m (4H), 4.87 s (2H), 4.87 s (2H), 7.23 d (2H, J = 7.86 Hz), 7.36 d (2H, J = 8.11 Hz), 7.87 d (2H, J = 7.88 Hz), 8.01 d (2H, J = 8.09 Hz). 13 C-NMR (DMSO): 25.4 (MsOH), 41.6, 60.1 (OCH2), 62.4 (SO2CH2), 65.8, 122.4, 122.6, 122.8, 122.9, 124.9, 126.1, 133.9, 141.9, 155.5, 162.4, 163.7, 164.8. 4.15. 1-BOC-isonipecotic acid 14 Isonipecotic acid (12.9 g, 0.1 mol) was dissolved in water (200 mL) containing 10.6 g (0.1 mol) of sodium carbonate and the mixture was cooled to 0 C with stirring. Then, the solution of di-tert. butyl dicarbonate in 200 mL of dioxane was slowly added at the same temperature. The mixture was stirred overnight at room temperature, evaporated in vacuo to ca. 200 mL of the volume, diluted with 200 mL of water, acidied with an excess of the solid citric acid and extracted with three portions of ethyl acetate. The ethyl acetate solution was washed with sat. sodium carbonate, brine, dried (NA2SO4) and evaporated in vacuo.The product was recrystallised from hexane-EtOAc. M.p. 146147 C, yield: 19.9 g, 86%. Elemental anal. for C11H19NO4 (C, H, N). 1H-NMR (CDCl3): 1.031.45 m (4H), 2.79 s (9H, BOC), 3.11 m (4H), 4.33 m (1H, CH-N).

1033 4.16. Preparation of compounds 12a and b and 15a and b (typical procedure) The corresponding acid (0.001 mol) was dissolved in the mixture of dichloromethane-acetonitrile (both dry) (1:1 v/v, 20 mL) and the corresponding hydroxy derivatives (0.001 mol) were added. After both compounds were dissolved, the solution was cooled to 20 C under chlorcalcium cover and then DCC (2.01 g, 0.001 mol) was added at once. The mixture was stirred overnight and it was allowed to reach room temperature. The dicyclohexyl urea was removed by ltration, washed with two small portions of dichloromethane, the ltrate was evaporated and the product was crystallised. 12a: m.p. 145146 C (iPr2O-EtOAc). Yield: 490 mg, 63%. Elemental anal. for C40H47N5O12 (C, H, N). 1HNMR (CDCl3): 1.48 s (9H, BOC), 1.77 m (2H, CH2 Arg), 3.87 t (2H, CH2 Arg.), 3.914.01 m (12H), 4.13 m (1H, CH-N), 4.9 s (2H, CH2O), 5.13 s (2H, CH2O), 5.25 s (2H, CH2O), 7.23 d (2H, J = 7.46 Hz), 7.257.36 m (10H, C6H5), 8.11 d (2H, J = 7.50 Hz). 13C-NMR (DMSO): 24.8 (CH2 Arg), 28.2 (BOC), 33.9 (CH2N Arg), 42.1 (CH2 Arg), 44.1 and 45.1 (OCH2Ph), 49.1 (CH-N), 53.5 (OCH2), 61.7, 66.32, 69.0, (morpholine), 80.1 (BOC), 115.43, 121.48, 127.82, 128.1, 128.3, 128.6, 129.1, 129.2, 129.3, 129.4, 131.5, 132.1, 134.3, 136.7, 154.4, 155.6, 155.7, 160.4, 161.6, 165.9, 170.7. 12b: m.p. 6970 C (iPr2O/EtOAc). Yield: 555 mg, 68%. Elemental anal. for C42H51N5O12 (C, H, N). 1HNMR (CDCl3): 1.47 s (9H, BOC), 2.70 t (2H, J = 5.7 Hz, Ph-CH2CH2-), 2.91 t (2H, J = 5.6 Hz, PhCH2CH2-), 5.53 m (2H, CH2 Arg), 3.653.75 m (8H), 4.03 m (1H), 4.70 s (2H, OCH2), 5.11 s (2H, OCH2), 5.25 s (2H, OCH2), 6.90 d (2H, J = 7.9 Hz), 7.18 d (2H, J = 7.8 Hz), 7.227.26 m (10 H). 13C-NMR (DMSO): 24.9 (CH2Arg), 28.2 (BOC), 30.1, 33.9, 35.4, 42.1, 44.1, 44.9, 49.1, 53.4, 61.1, 66.6, 66.9, 70.1, 79.9, 115.4, 121.3, 127.7, 127.9, 128.1, 128.2, 128.3, 128.4, 128.6, 128.7, 129.1, 129.2, 134.6, 136.8, 138.1, 148.8, 155.4, 155.7, 160.4, 163.7, 163.8, 165.1. 15a: m.p. 111112 C (Hexane/EtOAC). Yield: 280 mg, 58%. Elemental anal. for C24H32N2O8 (C, H, N). 1 H-NMR (CDCl3): 1.08 and 1.09 dd (1H), 1.4 s (9H, BOC), 2.00 and 2.03 dd (1H), 2.92 dd (1H), 3.45 bs (2H), 3.58 m (3H), 3.70 bs (6H), 4.04 d (1H, J = 5.45 Hz), 4.90 s (2H, CH2O), 7.15 d (2H, J = 7.62 Hz), 8.16 d (2H, J = 7.66 Hz). 13C-NMR (DMSO): 27.8, 28.4, 41.5, 48.9, 81.1, 127.8, 128.1, 128.5, 132.3, 133.1, 138.7, 154.2, 165.3, 166.2, 172.4. 15b: m.p. 126127 C (iPr2O-EtOAc). Yield: 273 mg, 56%. Elemental anal. for C25H35N3O7 (C, H, N). 1HNMR (CDCl3): 1.43 s (9H), 1.72 dd (1H), 1.89 dd (1H), 2.0 dd (1H), 2.50 bs (2H), 3.33 s (3H, CH3N), 3.38 m (3H), 3.65 m (6H), 4.11 dd (1H), 4.96 s (2H, OCH2), 7.14 d (2H, J = 7.41 Hz), 8.10 d (2H, J = 7.41 Hz). 13C-NMR (DMSO): 28.3 (BOC), 34.7, 41.5, 41.8, 42.9, 43.1, 54.5, 61.8, 79.7 (BOC), 121.4, 121.6, 127.0, 131.5, 154.5, 164.7, 165.3, 172.5. 4.17. Preparation of compounds 13a and b (typical procedure) In the standard apparatus for hydrogenation under normal pressure, were placed, 0.0005 mol of the compound 16a or 16b dissolved in 10 mL of methanol, and 5 l of the 99% MeSO3H were added, followed by 50 mg of the 10% Pd on charcoal, and the hydrogenation proceeded over 2 h. The catalyst was removed by ltration, washed with methanol, the methanolic solution was partly evaporated in vacuo (to ca. 45 mL) and the product was precipitated with an excess of diisopropyl ether, ltered and puried via semi-preparative HPLC. 13a: m.p. amorphous. Yield: 121 mg (after HPLC purication). Elemental anal. for C25H39N5O11S (CHNS). 1H-NMR (DMSO): 1.81 m (2H, CH2 Arg), 2.87 s (9H, BOC), 3.25 t (2H, CH2 Arg), 3.37 m (2H, CH2N Arg), 3.59 bs (3H), 3.76 m (5H), 4.78 m (1H, CH-N), 4.87 s (2H, OCH2), 6.71 d (2H, J = 7.46 Hz), 7.82 d (2H, J = 7.49 Hz). 13C-NMR (DMSO): 25.7, 26.12 (BOC), 35.7, 39.4, 42.1, 43.4, 46.4, 47.1, 48.1 (MsOH), 50.4, 62.6, 81.1, 115.3, 118.4, 133.7, 142.5, 160.3, 168.4, 170.1, 176.9. 13b: amorphous. Yield: 134 mg (after HPLC purication). Elemental anal. for C27H43N5O11S (C, H, N, S). 1 H-NMR (DMSO): 1.18 m (3H), 1.51 s (9H), 2.36 t (2H, J = 6.1 Hz), 2.71 t (2H, J = 6.0 Hz), 3.33 m (2H), 3.7813.75 m (6H), 3.91 m (2H), 4.2 m (1H), 4.81 s (2H), 7.11 d (2H, J = 7.11 Hz), 7.26 d (2H, J = 7.13 Hz), 10.11 bs (3H). 13C-NMR (DMSO): 24.9, 29.7 (BOC), 31.3, 33.9, 35.6, 42.1, 44.1, 44.9, 50.1, 56.4, 60.3 (OCH2), 81.3 (BOC), 118.3, 121.6, 127.3, 136.6, 151.1, 155.6, 161.4, 168.6, 171.2. 4.18. Preparation of compounds 16a and b (typical procedure) The BOC protected compound 16a or 16b (0.005 mol) was dissolved in 10 mL of ethyl acetate and cooled to 0 C. Then, 5 mL of 5 M HCl in EtOAc was added dropwise over 5 min. The precipitate appeared immediately. The mixture was standing for another 30 min, the precipitate was ltered, dried in vacuo and dissolved in 10 mL of ethanol. Cyanamide (0.27 g, 0.0065 mol) and conc. HCl (2 mL) were added and the mixture was stirred overnight. The product was precipitated after the reaction

1034 mixture was kept at 20 C for 24 h. The precipitate was ltered, washed with isopropanol, dried and puried by semi-preparative HPLC. 16a: amorphous. Yield: 127 mg (after HPLC purication). Elemental anal. for C20H26N4O6Cl (C, H, N, Cl, % Cl calc. 7.83, found 8.11). 1H-NMR (DMSO): 1.92 m (2H), 2.15 m (2H), 2.93 m (2H), 3.283.61 m (9H), 5.44 m (1H), 5.05 s (2H, CH2O), 7.39 d (2H, J = 7.75 Hz), 8.09 d (2H, J = 7.73 Hz), 9.109.19 bs (4H). 13 C-NMR (DMSO): 22.7, 40.7, 41.6, 44.3, 44.4, 56.6, 61.1, 61.4, 65.9, 110.5, 115.3, 120.1, 126.2, 131.8, 157.3, 160.1, 160.4, 162.12, 165.2. 16b: amorphous. Yield: 111 mg (after HPLC purication). Elemental anal. for C20H32N5O5Cl2 (C, H, N, Cl, % Cl calc. 14.38, found 14.79). 1H-NMR (DMSO): 2.61 m (2H), 2.81 m (2H), 2.813.51 m (10H), 3.51s (3H), 4.51 m (1H), 5.11 s (2H, OCH2), 7.53 d (2H, J = 7.63 Hz), 8.04 d (2H, J = 7.38 Hz). 13C-NMR (DMSO): 16.19, 24.3, 37.7 (MsOH), 38.6, 40.7, 41.9, 46.5, 46.6, 46.9, 57.0, 61.9, 122.2, 127.1, 130.9, 131.6, 154.2, 164.1, 164.7, 171.6. 4.19. Inhibiting activity assays Inhibiting activities were measured according to the described procedure [3] using the following substrates: Bz-Arg-pNa for trypsin, Bz-Phe-Val-Arg-pNa for thrombin, D-Val-Leu-Lys-pNa for plasmin and Z-Val-Gly-ArgpNa for urokinase. References
[1] [2] [3] [4] [5] De Clerck Y.A., Imren S., Eur. J. Cancer 30a (14) (1994) 21802206. Das S., Mukopadhyay P., Acta Oncologica 33 (8) (1994) 859876. Zlatoidsky P., Maliar T., Eur. J. Med. Chem. 31 (1996) 895899. Ganu W.S., Shaw E., J. Med. Chem. 42 (1981) 698700. Jetten M., Peters C.A.M., Van Nispen J.W.F.M., Ottenheijm H.J.C., Tetrahedron Lett. 32 (1991) 60256028.

Eur. J. Med. Chem. 34 (1999) 10351042 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1035

Original article

Depsides as non-redox inhibitors of leukotriene B4 biosynthesis and HaCaT cell growth. 1. Novel analogues of barbatic and diffractaic acid
Sunil Kumar KCa, Klaus Mllerb*
b a Institut fr Pharmazie, Pharmazeutische Chemie I, Universitt Regensburg, D-93040 Regensburg, Germany Westflische Wilhelms-Universitt Mnster, Institut fr Pharmazeutische Chemie, Hittorfstrae 58 62, D-48149 Mnster, Germany

(Received 25 March 1999; accepted 7 May 1999)

Abstract A series of barbatic and diffractaic acid analogues has been synthesized and evaluated as inhibitors of leukotriene B4 (LTB4) biosynthesis and as antiproliferative agents. The 4-O-demethyl barbatic and diffractaic acid derivatives were among the most active compounds in both assays. In particular, ethyl 4-O-demethylbarbatate was the most potent LTB4 biosynthesis inhibitor of this series, with an IC50 value in the submicromolar range. Because the compounds did not show appreciable reactivity against a stable free radical, 2,2-diphenyl-1-picrylhydrazyl, and did not produce appreciable amounts of deoxyribose degradation as a measure of their potency to generate hydroxyl radicals, a simple redox effect could not explain their biological activity. Also, there was no nonspecic cytotoxicity as documented by the activity of lactate dehydrogenase released from the cytoplasm of keratinocytes, which was in the control range. 1999 ditions scientiques et mdicales Elsevier SAS barbatic acid / diffractaic acid / antiproliferative activity / keratinocytes / lactate dehydrogenase release / leukotriene biosynthesis

1. Introduction Depsides are a distinct class of lichen-derived compounds which are formed by condensation of two or more hydroxybenzoic acids whereby the carboxyl group of one molecule is esteried with a phenolic hydroxyl group of a second molecule. One of the most common secondary metabolites of many lichen species [1] is the didepside diffractaic acid (1, gure 1). This compound has been shown by several groups to exhibit antiviral [2], antitumour [3], analgesic and antipyretic [4] properties. Among several structurally dissimilar lichen-derived metabolites isolated from Parmelia species, we have identied 1 as a non-redox inhibitor of the biosynthesis of leukotriene B4 (LTB4) in bovine polymorphonuclear leukocytes (PMNL) [5]. Leukotrienes are derived from the biotransformation of arachidonic acid through the action of 5-lipoxygenase (5-LO) and play an important role in a variety of pathophysiological states in man, particularly those involving inammation [6]. Furthermore, we found that 1 is also a potent antiproliferative
*Correspondence and reprints

Figure 1. Structures of diffractaic acid (1) and barbatic acid (2).

agent against the growth of human keratinocytes [7]. These combined inhibitory actions against 5-LO and keratinocyte cell growth suggested a benecial effect against inammatory and hyperproliferative skin diseases such as psoriasis, since both pathological features were targeted. As part of our continuing search for agents suitable for the treatment of inammatory and hyperproliferative skin

1036 diseases such as psoriasis, we have synthesized several novel analogues of 1 and its congener barbatic acid (2), modied at the carboxylic acid function and the 4-methoxy group in the benzoyloxy moieties, to explore the effect of increased lipophilicity and some redox properties on the biological activity of the compounds. The redox properties were evaluated in terms of reactivity against the stable free radical 2,2-diphenyl-1picrylhydrazyl (DPPH), in order to evaluate the antioxidant potential, and deoxyribose degradation was determined as a measure of hydroxyl radical formation [8]. The ability of the novel compounds to inhibit the growth of human keratinocytes was evaluated in HaCaT cells [9], and inhibition of LTB4 biosynthesis was assayed in bovine polymorphonuclear leukocytes [8].

2. Chemistry

Unambiguous syntheses of the lichen depsides 1 and 2 have been reported [10]. The mononuclear precursors for the novel depsides were obtained from ethyl 2,4dihydroxy-3,6-dimethylbenzoate(3)followingthemethodology of Elix [10] which proved to be particularly suitable for the large scale preparation of this starting material (gure 2). Esters 4a4e were directly obtained from 3 by transesterication in the presence of the corresponding sodium alkoxides and alcohols [10]. Alkylation of 3 with a one molar proportion of dimethyl sulfate or the pertinent benzyl chlorides in the presence of potassium carbonate, selectively yielded the corresponding 4-methoxy derivative 5 or 4-benzyloxy derivatives 9a and b. Methylation of the second hydroxy group of these derivatives with dimethyl sulfate gave the 2-methoxy derivatives 6 and 10a and b. Subsequent hydrolysis of the esters 5, 6, 9a and b, and 10a and b yielded the requisite acids 7, 8, 11a and b, and 12a and b, respectively, for the A-ring of the desired depsides. Depside formation between these acids and the phenolic esters 3 and 4ae was achieved by treatment with triuoroacetic anhydride in anhydrous toluene and yielded the barbatic and diffractaic acid analogues 13af, 14af, 15af and 17af, 18af, 19af, respectively (gure 3). Hydrogenolytic cleavage of the benzyl ethers 14ae and 18ae over palladium/ carbon produced the phenolic analogues 16ae and 20ae, respectively, whereas benzyl esters 14f and 18f were cleaved to the acids 16f and 20f, respectively. Analogously, benzyl esters 13f and 17f yielded the parent compounds 2 and 1, respectively.

Figure 2. Reagents: (a) Na, ROH, , 24 h; (b) Me2SO4, K2CO3, acetone, , 24 h; (c) 4-RC6H4CH2Cl, K2CO3, acetone, , 24 h; (d) KOH, H2O, DMSO, 90 C.

3. Biological results and discussion Inhibition of LTB4 biosynthesis by the barbatic and diffractaic acid analogues was determined in bovine polymorphonuclear leukocytes. As shown in table I, activity of barbatic acid (2), with an IC50 value of 7.8 M, was similar to that of 1. The biological activity was reduced when the free carboxylic acid groups were esteried. The esters 13af of 2 were either moderately active or inactive even at concentrations up to 20 M. Also, esterication of 1 resulted in less active or inactive compounds (17a and cf), except for 17b, where activity was retained. In general, introduction of a 4-benzyl or 4-methoxybenzyl group into 1 and 2, as in 14af, 15af, 18af, and 19af dramatically reduced inhibitory action against LTB4 biosynthesis. Most of these compounds were inactive at 20 M. Exceptions were the

1037 Moreover, the results obtained from the deoxyribose assay (table I) also suggest that hydroxyl radicals are not involved in the mechanism of enzyme inhibition by the novel depsides. The deoxyribose assay is a sensitive test for the production of hydroxyl radicals [12]. The release of 2-thiobarbituric acid reactive material is expressed as malondialdehyde (MDA) and reects a measure for hydroxyl-radical generation. However, we did not observe any deoxyribose degradation from depsides, even for compounds 16af with three phenolic hydroxyl groups. In vitro antiproliferative activities were determined in 24-well culture dishes against the growth of HaCaT cells. This nontransformed human cell line can be used as a model for highly proliferative epidermis [13]. The parent compounds 1 and 2 were potent inhibitors of cell growth with IC50 values of 2.6 and 4.1 M, respectively (table I). While the esters 13ac of 2 showed comparable or slightly improved activity, the corresponding esters 17ac of 1 were inactive. However, butyl ester 17e displayed potent antiproliferative activity. Furthermore, among the 4-benzylated derivatives of 1 were also some antiproliferative active agents. Similar to the results obtained in the LTB4 assay, 4-O-demethylation of barbatic and diffractaic acid (16af and 20af, respectively) generally produced active compounds, although there were no improvements as compared to the parent depsides. There is little or no correlation between the in vitro antiproliferative activity of the compounds and their ability to inhibit LTB4 biosynthesis. With respect to both features, well-balanced representatives are found among the esters of diffractaic acids and the 4-O-demethylated analogues of 1 and 2. Unfortunately, the potent LTB4 biosynthesis inhibitor 20b is inactive at 20 M. The most potent inhibitor of keratinocyte growth, ethyl diffractate (13b), is also an inhibitor of LTB4 biosynthesis. Likewise, the most potent inhibitor of LTB4 biosynthesis, ethyl 4-O-demethylbarbatate (16b), also displays antiproliferative activity. Keratinocytes were also tested for their susceptibility to the action of the most potent depsides on plasma membrane integrity. As a measure of cytotoxicity, release of lactate dehydrogenase into the culture medium was determined [14]. In these experiments, all potent inhibitors of keratinocyte growth showed values in the control range, documenting that their activity was due to cytostatic rather than cytotoxic effects. This may be advantageous as compared with the topical antipsoriatic agent anthralin, which is known to induce inammation of the healthy skin surrounding a psoriatic lesion. As a result of the strong hydroxyl radical generating activity of this

Figure 3. Reagents: (a) (CF3CO)2O, toluene, room temperature; (b) Pd/C, EtOAc, room temperature. R1, R2, and R3 are dened in table I.

4-benzyloxydiffractates 14c, e and f of 2 which showed comparable activity. Since a major class of leukotriene biosynthesis inhibitors often contains a hydroxylated aromatic ring [11], we have speculated that demethylation of the 4-methoxy group of 1 and 2 might improve their activity. As expected, 4-O-demethyl barbatic and diffractaic acid derivatives 16af and 20af, respectively, inhibited LTB4 biosynthesis with IC50 values in the low micromolar range. With the exception of the free acids 16f and 20f, these analogues approached the potency of their respective parent compounds or were even more potent than these. In particular, the ethyl esters 16b and 20b were the most potent inhibitors of LTB4 biosynthesis of this series. Potency of compound 16b, with an IC50 of 0.8 M, was comparable to that of the standard inhibitor nordihydroguaiaretic acid. One chemical feature of many inhibitors of LTB4 biosynthesis is their ability to remove free radicals, since the conversion of arachidonic acid into LTB4 is an oxidative process. Therefore, we have evaluated the depsides for their ability to react with the stable free radical DPPH to give the reduced 2,2-diphenyl-1picrylhydrazine. Table I shows that no appreciable amount of reduced hydrazine was formed by these compounds, documenting their lack of reactivity against stable free radicals. This suggests that a simple redox effect does not explain their activity in the LTB4 assay. Rather, activity appears to be due to specic enzyme interaction.

1038
Table I. Redox properties, inhibition of LTB4 biosynthesis, antiproliferative activity and cytotoxicity against HaCaT cells of barbatic and diffractaic acid derivatives.

kDPPHa Compound 1 2 13a 13b 13c 13d 13e 13f 14a 14b 14c 14d 14e 14f 15a 15b 15c 15d 15e 15f 16a 16b 16c 16d 16e 16f 17a 17b 17c 17d 17e 17f 18a 18b 18c 18d 18e 18f
a

DD (OH)b 0.12 0.07 0.09 0.01 0.16 0.04 0.18 0.06 0.21 0.03 ND ND 0.08 0.01 ND ND 0.19 0.03 0.23 0.01 0.11 0.03 0.26 0.01 ND ND ND 0.19 0.01 ND ND 0.21 0.02 0.02 0.01 0.12 0.01 0.16 0.01 0.72 0.01f 0.01 0.01 0.19 0.01 0.21 0.01 0.06 0.02 ND ND ND ND 0.44 0.02f ND ND ND ND

LTB4c IC50 (M) 7.6 7.8 18.6 11.3 10.5 > 20 > 20 19.2 > 20 > 20 9.0 16.2 9.0 7.9 > 20 > 20 > 20 15.0 > 20 > 20 2.1 0.8 5.7 2.6 5.0 14.0 13.2 5.3 19 > 20 > 20 > 20 > 20 18.3 > 20 > 20 > 20 > 20

AAd IC50 (M) 2.6 4.1 4.8 1.9 3.2 > 20 > 20 > 20 19.0 16.7 > 20 > 20 9.0 8.4 > 20 > 20 > 20 > 20 > 20 > 20 8.2 8.4 3.6 8.0 3.8 8.2 > 20 > 20 > 20 14.0 4.1 > 20 > 20 > 20 ND ND ND ND

LDHe (mU) 136 148 148 143 143 ND ND ND 150 142 ND ND 150 140 ND ND ND ND ND ND 143 149 149 146 140 140 ND ND ND 170f 167f ND ND ND ND ND ND ND

(M

s )

Me Me Me Me Me Me Me Me PhCH2 PhCH2 PhCH2 PhCH2 PhCH2 PhCH2 4-MeOPhCH2 4-MeOPhCH2 4-MeOPhCH2 4-MeOPhCH2 4-MeOPhCH2 4-MeOPhCH2 H H H H H H Me Me Me Me Me Me PhCH2 PhCH2 PhCH2 PhCH2 PhCH2 PhCH2

Me H H H H H H H H H H H H H H H H H H H H H H H H H Me Me Me Me Me Me Me Me Me Me Me Me

H H Me Et Prop CHMe2 Bu CH2Ph Me Et Prop CHMe2 Bu CH2Ph Me Et Prop CHMe2 Bu CH2Ph Me Et Prop CHMe2 Bu H Me Et Prop CHMe2 Bu CH2Ph Me Et Prop CHMe2 Bu CH2Ph

0.63 0.01 0.14 0.07 0.93 0.01 0.31 0.05 0.29 0.01 ND ND 0.45 0.13 ND ND 0.55 0.04 0.78 0.01 0.51 0.01 0.96 0.01 ND ND ND 0.72 0.01 ND ND 0.34 0.01 0.52 0.07 0.91 0.60 0.43 0.14 0.67 0.09 0.81 0.03 0.85 0.03 0.57 0.01 0.93 0.01 ND ND ND ND 0.97 0.02 ND ND ND ND

Reducing activity against 2,2-diphenyl-1-picrylhydrazyl with equimolar amounts of test compound. bDeoxyribose degradation as a measure of hydroxyl-radical formation. Indicated values are mol of malondialdehyde/mmol of deoxyribose released by 75 M of test compound (controls < 0.1). cInhibition of LTB4 biosynthesis in bovine PMNL. Inhibition was signicantly different with respect to that of the control; n = 3 or more, P < 0.01. Nordihydroguaiaretic acid was used as the standard inhibitor (IC50 = 0.4 M) [8]. dAntiproliferative activity against HaCaT cells. Inhibition of cell growth was signicantly different with respect to that of the control, n = 3, P < 0.01. eActivity of LDH (mU) release in HaCaT cells after treatment with 2 M test compound (n = 3, SD < 10%). fValues are signicantly different with respect to vehicle control (P < 0.05). ND = not determined. gPositive control [8].

1039
Table I. (continued). kDPPHa Compound 19a 19b 19c 19d 19e 19f 20a 20b 20c 20d 20e 20f anthraling R1 4-MeOPhCH2 4-MeOPhCH2 4-MeOPhCH2 4-MeOPhCH2 4-MeOPhCH2 4-MeOPhCH2 H H H H H H R2 Me Me Me Me Me Me Me Me Me Me Me Me R3 Me Et Prop CHMe2 Bu CH2Ph Me Et Prop CHMe2 Bu H (M1 s1) ND ND ND ND ND ND 0.93 0.01 0.64 0.01 0.89 0.05 0.97 0.01 0.51 0.01 0.24 0.03 24.2 4.2f ND ND ND ND ND ND 0.18 0.01 0.10 0.02 0.01 0.01 0.09 0.01 0.16 0.01 0.15 0.01 2.89 0.14f DD (OH)b LTB4c IC50 (M) > 20 > 20 > 20 > 20 > 20 > 20 7.8 1.4 8.5 5.8 7.8 11.0 37.0 AAd IC50 (M) > 20 > 20 ND ND ND ND 9.0 > 20 > 20 7.2 > 20 9.8 0.7 LDHe (mU) ND ND ND ND ND ND 138 ND ND 168 ND 114 294f

agent [15], LDH release by anthralin signicantly exceeded that of the vehicle control. In conclusion, barbatic acid analogues were consistently more active against the biosynthesis of LTB4 and the growth of HaCaT keratinocytes than the corresponding diffractaic acid analogues. Though this may be related to their additional phenolic hydroxyl group, determination of the antioxidant and pro-oxidant potential of the compounds did not reveal any appreciable redox activity. Barbatic acid analogue 16b has been identied as a potent non-redox inhibitor of LTB4 biosynthesis which also displays antiproliferative activity against keratinocyte growth. 4. Experimental protocols 4.1. Chemistry 4.1.1. General For analytical instruments and methods see reference [16]. Compounds 13 and 58 were prepared as described [10]. 4.1.2. Propyl 2,4-dihydroxy-3,6-dimethylbenzoate 4b Sodium (0.5 g, 21.73 mmol) was dissolved in absolute propanol (50 mL) and stirred at room temperature. Then 3 (3 g, 14.28 mmol) was added to the solution and reuxed under nitrogen for 24 h. The solution was cooled, acidied with cold 10% HCl and extracted with ether (3 100 mL). The combined organic phase was dried over MgSO4 and evaporated. The crude product was puried by ash chromatography (SiO2, CH2Cl2) to give colourless crystals; FTIR 3 396, 2 980, 2 957,

1 630 cm1; 1H-NMR (250 MHz, CDCl3) 12.14 (s, 1H), 6.20 (s, 1H), 4.97 (s, 1H), 4.30 (t, J = 6.5 Hz, 2H), 2.48 (s, 3H), 2.10 (s, 3H), 1.80 (m, 2H), 1.03 (t, J = 7.4 Hz, 3H); Anal. (C12H16O4) C, H. Analogously, compounds 4a and ce were prepared from 3 (table II). 4.1.3. Ethyl 2-hydroxy-4-(4-methoxybenzyloxy)-3,6dimethylbenzoate 9b A suspension of 3 (6 g, 28.57 mmol), anhydrous potassium carbonate (11.25 g, 81.39 mmol) and 4-methoxybenzylchloride (4.47 g, 28.57 mmol) in dry acetone (75 mL) was reuxed for 24 h, then cooled, acidied with cold 10% HCl and extracted with ether (3 200 mL). The combined organic phase was washed with water, dried over MgSO4 and evaporated. The crude product was puried by column chromatography (SiO2) using hexane/ethyl acetate (9:1) to afford colourless crystals; FTIR 3 438, 1 720, 1 636 cm1; 1H-NMR (CDCl3) 11.91 (s, 1H), 7.35 (d, J = 9.5 Hz, 2H), 6.93 (d, J = 9.5 Hz, 2H), 6.34 (s, 1H), 5.03 (s, 2H), 4.38 (q, J = 7.1 Hz, 2H), 3.82 (s, 3H), 2.52 (s, 3H), 2.11 (s, 3H), 1.41 (t, J = 7.1 Hz, 3H). Anal. (C19H22O5) C, H. 4.1.4. Ethyl 2-methoxy-4-(4-methoxybenzyloxy)-3,6dimethylbenzoate 10b A suspension of 9b (2.20 g, 6.67 mmol), anhydrous potassium carbonate (2.76 g, 20 mmol) and dimethyl sulfate (0.6 mL, 6.67 mmol) in dry acetone (50 mL) was reuxed for 24 h, then cooled, acidied with cold 10% HCl and extracted with ether (3 200 mL). The combined organic phase was washed with water, dried over MgSO4 and evaporated. The crude product was puried by column chromatography (SiO2) to afford a colourless

1040
Table II. Chemical data of starting materials, barbatic and diffractaic acid derivatives. Compounda 4a 4b 4c 4d 4e 9a 9b 10a 10b 11a 11b 12a 12b 13a 13b 13c 13d 13e 13f 14a 14b 14c 14d 14e 14f 15a 15b 15c 15d 15e 15f 16a 16b 16c 16d 16e 16f 17a 17b 17c 17d 17e 17f 18a 18b 18c 18d 18e 18f 19a 19b 19c 19d 19e 19f Formulab C10H12O4 C12H16O4 C12H16O4 C13H18O4 C16H16O4 C18H20O4 C19H22O5 C19H22O4 C20H24O5 C16H16O5 C17H18O5 C17H18O4 C18H20O5 C20H22O7 C21H24O7 C22H26O7 C22H26O7 C23H28O7 C26H26O7 C26H26O7 C27H28O7 C28H30O7 C28H30O7 C29H32O7 C32H30O7 C27H28O8 C28H30O8 C29H32O8 C29H32O8 C30H30O8 C33H32O8 C19H20O7 C20H22O7 C21H24O7 C21H24O7 C22H26O7 C18H18O7 C21H24O7 C22H26O7 C23H28O7 C23H28O7 C24H30O7 C27H28O7 C27H28O7 C28H30O7 C29H32O7 C29H32O7 C30H34O7 C33H32O7 C28H30O8 C29H32O8 C30H34O8 C30H34O8 C31H32O8 C34H34O8 M.p. (C) 144; ref. [18] 145 135; ref. [19] 139 88; ref. [19, 20] 92 120; ref. [19, 20] 123 118; ref. [10] 113 67; ref. [10] 6768 80 oil; ref. [17] oil oil 167; ref. [10] 165167 198 120; ref. [17] 122 136 166; ref. [17] 170 186; ref. [21] 189 137; ref. [20] 138139 125; ref. [20] 128129 134; ref. [20] 133 132; ref. [10] 136138 137; ref. [22] 133134 148 108 124 114 128; ref. [10] 131132 98 107 88 97 119 103 108; ref. [23] 108112 142 146 123 153 172; ref. [10] 136138 133; ref. [24] 127128 144; ref. [25] 141144 126; ref. [19] 127 124; ref. [19] 127 115; ref. [19] 115 123; ref. [10] 119 124 124 141 113 142 118; ref. [17] 118120 130 98 122 117 111 89 Yield (%) 84 93 62 88 55 65 85 91 86 84 92 71 76 77 71 89 64 90 69 63 73 73 65 61 78 69 73 65 68 63 71 87 95 90 90 89 91 69 73 82 66 88 83 74 67 63 74 63 69 82 73 62 80 77 65 Solventc,d (vol%) MCc MCc MCc MCc MCc H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (1 H/EAc (1 H/EAc (1 H/EAc (1 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAd H/EAd H/EAd H/EAd H/EAd H/EAd H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 H/EAc (9 Anal.e C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, C, H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H

+ + + + + + + + + + + + + + + + + + + + + + + + + +

1); 1); 1) 1) 1) 1) 1) 1) 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1);

Md Md

M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd

+ + + + + + + + + + + + + + + + + +

1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1); 1);

M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd M/Cd

1041
Table II. (continued). Compounda 20a 20b 20c 20d 20e 20f
a

Formulab C20H22O7 C21H24O7 C22H26O7 C22H26O7 C23H28O7 C19H20O7

M.p. (C) 161 170 143 124 112 203; ref. [17] 207209

Yield (%) 98 93 97 94 91 95

Solventc,d (vol%) H/EAd H/EAd H/EAd H/EAd H/EAd H/EAd

Anal.e C, C, C, C, C, C, H H H H H H

All compounds were obtained as colourless crystals except where stated otherwise. bAll new compounds displayed 1H-NMR and FTIR consistent with the assigned structure. cEluant used for column chromatography. dSolvent for recrystallization; C = chloroform; EA = ethyl acetate; H = hexane; M = methanol; MC = methylene chloride. eElemental analyses were within 0.4% of calculated values except where stated otherwise.

oil; FTIR 3 423, 1 700, 1 638 cm1; 1H-NMR (CDCl3) 7.32 (d, J = 9.5 Hz, 2H), 6.93 (d, J = 9.5 Hz, 2H), 6.53 (s, 1H), 4.98 (s, 2H), 4.39 (q, J = 7.1 Hz, 2H), 3.82 (s, 3H), 3.76 (s, 3H), 2.29 (s, 3H), 2.13 (s, 3H), 1.38 (t, J = 7.1 Hz, 3H). Anal. (C20H24O5) C, H. 4.1.5. 2-Hydroxy-4-(4-methoxybenzyloxy)-3,6-dimethylbenzoic acid 11b A solution of aqueous potassium hydroxide (2.86 g in 7 mL H2O, 51.0 mmol) was added to a solution of 9b (3.00 g, 8.82 mmol) in DMSO (40 mL) and heated on a water bath for 2.5 h (TLC control). The solution was cooled to room temperature, diluted with excess water (100 mL), acidied with cold 10% HCl, and extracted with ether (3 100 mL). The combined organic phase was washed with water (3 200 mL), dried over MgSO4 and evaporated. The residue was puried by column chromatography (SiO2) using hexane/ethyl acetate (1:1) to afford colourless crystals; FTIR 3 053, 1 700, cm1; 1 H-NMR (CDCl3, DMSO-d6) 12.49 (s, 1H), 7.30 (d, J = 9.5 Hz, 2H), 6.99 (d, J = 9.5 Hz, 2H), 6.32 (s, 1H), 5.09 (s, 2H), 3.82 (s, 3H), 2.55 (s, 3H), 2.13 (s, 3H), Anal. (C17H18O5) C, H. Analogously, 12b was prepared from 10b (table II). 4.1.6. General procedure for the condensation of benzoic acids with phenolic esters 4.1.6.1. Ethyl 4-(4-benzyloxy-2-hydroxy-3,6-dimethylbenzoyloxy)-2-hydroxy-3,6-dimethylbenzoate 14b A solution of 11a (136 mg, 0.5 mmol) and 3 [10] (107 mg, 0.5 mmol) in anhydrous toluene (2 mL) and triuoroacetic anhydride (0.5 mL) was stirred at room temperature for 2.5 h (TLC control). The solvent was removed under reduced pressure and the residue was puried by column chromatography (SiO2) using hexane/ ethyl acetate (9:1). The product was recrystallized from MeOH/CHCl3 to give colourless crystals; FTIR 3 430,

2 965, 2 900, 1 665, 1 618 cm1; 1H-NMR (CDCl3) 11.99 (s, 1H), 11.53 (s, 1H), 7.327.47 (m, 5H), 6.51 (s, 1H), 6.44 (s, 1H), 5.17 (s, 2H), 4.44 (q, J = 7.1 Hz, 2H), 2.66 (s, 3H), 2.55 (s, 3H), 2.17 (s, 3H), 2.08 (s, 3H), 1.43 (t, J = 7.13 Hz, 3H); Anal. (C27H28O7) C, H. Analogously, 13af were prepared from 7 [10] and 3 and 4ac; 14a and cf were prepared from 11a [10] and 4ac; 15af were prepared from 11b and 3 and 4ac; 17af were prepared from 8 [10] and 3 and 4ac; 18af were prepared from 12a [17] and 3 and 4ac; 19af were prepared from 12b and 3 and 4ac (table II). 4.1.7. General procedure for hydrogenolysis 4.1.7.1. Ethyl 4-(2,4-dihydroxy-3,6-dimethylbenzoyloxy)2-hydroxy-3,6-dimethylbenzoate 16b A suspension of 14b (112 mg, 0.25 mmol) and 10% palladium/carbon (25 mg) in dry ethyl acetate (2 mL) was stirred in H2 for 2 h (TLC control). The suspension was then ltered through celite, and the ltrate was evaporated. The residue was puried by column chromatography (SiO2) using hexane/ethyl acetate (9:1) to give colourless crystals; FTIR 3 456, 2 943, 1 665, cm1; 1 H-NMR (CDCl3) 12.00 (s, 1H), 11.71 (s, 1H), 6.50 (s, 1H), 6.30 (s, 1H), 5.29 (s, 1H), 4.43 (q, J = 7.1 Hz, 2H), 2.61 (s, 3H), 2.54 (s, 3H), 2.12 (s, 3H), 2.08 (s, 3H), 1.43 (t, J = 7.1 Hz, 3H); Anal. (C20H22O7) C, H. Analogously, 16a and cf were prepared from 14a and cf; 20af were prepared from 18af; 1 and 2 were prepared from 17f and 13f, respectively (table II). 4.2. Biological assay methods The procedures for the biological assays presented in table I were described previously in full detail: determination of the reducing activity against 2,2-diphenyl-1picrylhydrazyl [8], deoxyribose degradation [8], inhibi-

1042 tion of LTB4 biosynthesis [8], inhibition of HaCaT cell proliferation [9], and release of LDH into culture medium [14]. Acknowledgements We thank Mr K. Ziereis for his excellent technical assistance. S.K. KC thanks the German Academic Exchange Service for a scholarship. References
[1] [2] Culberson C.F., Chemical and Botanical Guide to Lichen Products, The University of North Carolina Press, Chapel Hill, 1969. Neamati N., Hong H., Mazumder A., Wang S., Sunder S., Nicklaus M.C., Milne G.W.A., Proksa B., Pommier Y., J. Med. Chem. 40 (1997) 942951. Yamamoto Y., Miura Y., Kinoshita Y., Higuchi M., Yamada Y., Murakami A., Ohigashi H., Koshimizu K., Chem. Pharm. Bull. 43 (1995) 13881390. Okuyama E., Umeyama K., Yamazaki M., Kinoshita Y., Yamamoto Y., Planta Med. 61 (1995) 113115. Kumar K.C.S., Mller K., J. Nat. Prod. 62 (1999) in press. Brooks C.D.W., Summers J.B., J. Med. Chem. 39 (1996) 26292654. Kumar K.C.S., Mller K., J. Nat. Prod. 62 (1999) in press. Mller K., Grster D., Piwek S., Wiegrebe W., J. Med. Chem. 36 (1993) 40994107. [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] Mller K., Leukel P., Ziereis K., Gawlik I., J. Med. Chem. 37 (1994) 16601669. Elix J.A., Norfolk S., Aust. J. Chem. 28 (1975) 11131124. Fitzsimmons B.J., Rokach J., in: Rokach J. (Ed.), Leukotrienes and Lipoxygenases, Elsevier, Amsterdam, 1989, pp. 427502. Halliwell B., Grootveld M., Gutteridge J.M.C., Methods Biochem. Anal. 33 (1988) 5990. Boukamp P., Petrussevska R.T., Breitkreutz D., Hornung J., Markham A., Fusenig N.E., J. Cell Biol. 761 (1988) 761771. Mller K., Huang H.S., Wiegrebe W., J. Med. Chem. 39 (1996) 31323138. Mller K., Biochem. Pharmacol. 53 (1997) 12151221. Mller K., Reindl H., Gawlik I., Eur. J. Med. Chem. 33 (1998) 969973. Elix J.A., Chester D.O., Gaul K.L., Parker J.L., Wardlaw J.H., Aust. J. Chem. 42 (1989) 11911199. Whalley W.B., J. Chem. Soc. (1949) 32783280. Fujikawa, Yakugaku Zasshi 60 (1940) 473478, cited from Beilstein Commander. Fujikawa, Shimamura, Tarui, Yakugaku Zasshi 61 (1941) 491495, cited from Beilstein Commander. Robertson A., Stephenson R.J., J. Chem. Soc. (1932) 16751681. Elix J.A., Mahadevan I., Wardlaw J.H., Arvidsson L., Jrgensen P.M., Aust. J. Chem. 40 (1987) 15811590. Yamamoto Y., Nishimura K.I., Kiriyama N., Chem. Pharm. Bull. 24 (1976) 18531859. Asahina Y., Fuzikawa F., Ber. Dtsch. Chem. Ges. 65 (1932) 583584. Asahina Y., Hashimoto A., Ber. Dtsch. Chem. Ges. 66 (1933) 641649.

[3]

[4] [5] [6] [7] [8]

Eur. J. Med. Chem. 34 (1999) 10431051 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1043

Original article

Preparation and local anaesthetic activity of benzotriazinone and benzoyltriazole derivatives

Giuseppe Caliendoa*, Ferdinando Fiorinoa, Paolo Griecoa, Elisa Perissuttia, Vincenzo Santagadaa, Rosaria Melib, Giuseppina Mattace Rasob, Angelina Zanescoc, Gilberto De Nuccic
Dipartimento di Chimica Farmaceutica e Tossicologica, Universit degli Studi di Napoli Federico II Via D. Montesano 49, 80131 Napoli, Italy b Dipartimento di Farmacologia Sperimentale, Universit degli Studi di Napoli Federico II Via D. Montesano 49, 80131 Napoli, Italy c Cartesius Analytical Unit, Department of Pharmacology, Biomedical Sciences Institute, University of Sao Paulo, Sao Paulo, Brazil (Received 19 April 1999; revised 6 July 1999; accepted 15 July 1999)
a

Abstract Two sets of benzotriazinone and benzoyltriazole derivatives were prepared and tested for local anaesthetic activity in comparison with lidocaine. Several of the prepared compounds exhibited a fairly good activity comparable or superior to that of lidocaine. The presence of a benzotriazinone or a benzoyltriazole moiety as an aromatic system was quite protable for both the intensity and duration of activity. The acute toxicity in mice of the four most potent compounds of the series was also assessed. Compound 1b, which has an anaesthetic activity comparable to that of lidocaine, was also characterized by a more favourable therapeutic index. All compounds were tested in vitro to evaluate their negative chronotropic action in isolated rat right atria. 1999 ditions scientiques et mdicales Elsevier SAS benzotriazinone / benzoyltriazole / local anaesthetic / negative chronotropic action

1. Introduction In previous articles we described the synthesis of two sets of N-[2-(alkylamino)ethyl]benzotriazol-X-yl-acetamides and of N-[2-(alkylamino)ethyl]benzotriazol-X-ylisobutyramides designed as local anaesthetic agents [1, 2]. The compounds of two sets fulll the requirements of the pharmacofore scheme proposed by Lfgren [3] because they are provided with an aromatic system, an intermediate chain and an aminic moiety ionizable under physiological pH. They have been assayed in vivo for their local anaesthetic activity. Some of the investigated compounds showed anaesthetic activities comparable with or higher than those exhibited by the reference drug lidocaine. Considering these results and the important role played by the aromatic system in the interaction with a corresponding hydrophobic region, in order to increase the

activity and to evaluate the inuence on activity by the different aromatic systems, the benzotriazole nucleus was replaced by a 1,2,3-benzotriazin-4(3H)-one and by a 4-benzoyl-1,2,3-triazole ring, thereby achieving general structures 1, 2 and 3 (gure 1). The N-alkylamino groups of the intermediate acetylacetamido side chain are those previously reported [1, 2] displaying the highest anaesthetic activity: dimethylamino, diethylamino, pyrrolidine and piperidine substituents. Herein we report the synthesis of a series of 1,2,3benzotriazinones (1ad) and of 4-benzoyl-1,2,3-triazole (2ad and 3ad) derivatives (tables I and II). The synthesized compounds were rst tested as local anaesthetics with different in vivo assays and, successively considering the relationships between local anaesthetic (Na+ channel block) and antiarrhythmic activities [4, 5], all compounds were preliminary tested to evaluate their negative chronotropic action in rat isolated right atria.

*Correspondence and reprints

1044

Figure 1. General structure of considered compounds.

Table I. Physicochemical properties of 1,2,3-benzotriazinone derivatives 1ad.

Formula*

MW

Compound

M.p. (C)

Cryst.** Solvent

Yield (%)

C13H17N5O2

275.31

1a

145146

a+b

60

C15H21N5O2

303.37

1b

155156

55

C15H19N5O2

301.35

1c

180181

a+b

66

C16H21N5O2

315.38

1d

169170

52

*Satisfactory microanalyses obtained: C, H, N values are within 0.4% of theoretical values. **Crystallization solvents: a) diethyl ether; b) ethyl alcohol.

2. Chemistry Compounds of general formula 1 reported in gure 1 were synthesized as illustrated in gure 2. They derived from condensation of a 1,2,3-benzotriazin-4(3H)-one ring (4) with ethyl bromoacetate in butan-2-one in the presence of potassium carbonate. The ethyl acetate deriva-

tives (5) were converted into the amide derivatives (1) by reaction in methyl alcohol solution with the appropriate amines. The nal compounds 1ad, reported in table I, were puried by cromatography on a silica gel column and further by crystallization from an appropriate solvent (yields ranging between 5266%).

1045
Table II. Physicochemical properties of 4-benzoyl-1,2,3-triazole derivatives 2ad and 3ad.

1-substituted 4-benzoyl-1,2,3-triazoles R Formula* MW Compound** M.p. (C) Yield (%)

2-substituted 4-benzoyl-1,2,3-triazoles Compound** M.p. (C) Yield (%)

C15H19N5O2

301.35

2a

132133

70

3a

9697

60

C17H23N5O2

329.40

2b

136137

40

3b

99100

45

C17H21N5O2

327.39

2c

151153

50

3c

9899

55

C18H23N5O2

341.41

2d

159161

40

3d

9495

50

*Satisfactory microanalyses obtained: C, H, N values are within 0.4% of theoretical values. **All compounds were crystallized by methyl alchol/diethyl ether.

The 4-benzoyl-1,2,3-triazole derivatives (2ad and 3ad) were synthesized following the steps outlined in gure 3. The parent compound, 4-benzoyl-1,2,3-triazole (8), was obtained in 95% yield by a modied method already described [6, 7]. Oxidation of the phenylethynylcarbinol

Figure 2. Synthesis of 1ad.

(6) with chromium trioxide and concentrated sulphuric acid produced phenyl ethynyl ketone (7) which was successively treated with NaN3 in anhydrous dimethylacetamide providing the expected compound. Reaction of 4-benzoyl-1,2,3-triazole with ethyl bromoacetate and potassium carbonate in butan-2-one gave a mixture of the triazole isomers 1- and 2- with an overall yield of 80% [8]. The two isomers were separated by column chromatography on silica gel using diethyl ether/nhexane (7:3 v/v) as eluent. The faster moving 2- isomer, ethyl 4-(benzoyl)-1,2,3-triazole-2-acetate (10), was collected with a higher yield (65%) with respect to the slower moving 1-substituted isomer (9) (35%). The ethyl acetate derivatives (9 and 10) were converted into the corresponding amide derivatives 2 and 3 by reaction in methyl alcohol solution with the appropriate amines. All the nal products (2ad and 3ad) reported in table II were further puried by crystallization from a mixture of diethyl ether and methyl alcohol (yields ranging between 4070%). The synthesized compounds listed in tables I and II were characterized by 1H-NMR spectroscopy whose data were fully consistent with the described structures.

1046

Figure 3. Synthesis of 4-benzoyl-1,2,3-triazole derivatives.

1 H-NMR of compounds 2ad and 3ad (CDCl3) differentiated clearly between 1- and 2- substituted 4-benzoyl-1,2,3-triazole derivatives. In fact there is a difference in chemical shift values among the protons in the 5- position of the benzoyltriazole ring in the series of 1- and 2- isomers. The triazole proton of the 1- isomer appears always as a singlet at lower eld with respect to the position of the same proton of analogues 2-substituted (chemical shift values ranging between 8.458.70 and 8.208.35 , respectively) according with literature [8, 9]. Physicochemical data of all the nal compounds are summarized in tables I and II.

ble IV). The ip acute toxicity and the therapeutic index of the more active compounds were also determined (table V). Successively the synthesized compounds were tested in vitro to evaluate their negative chronotropic action in rat isolated right atria (table VI). The synthesized compounds were always compared for their activity with lidocaine, taken as reference drug.

4. Results and discussion Table III summarizes the results of the surface and inltration anaesthesia assays performed on the 1,2,3benzotriazinone (1ad) and on the 4-benzoyl-1,2,3triazole (2ad and 3ad) derivatives. As for the 1,2,3benzotriazinone derivatives (1ad), in both tests, the results indicate that the structures endowed with highest activity are 1b and 1d, with values similar to that measured for lidocaine.

3. Pharmacology The compounds reported in tables I and II were tested in vivo for their anaesthetic activity by different tests: corneal anaesthesia in the rabbit, mouse tail anaesthesia (table III) and rat sciatic nerve block anaesthesia (ta-

1047
Table III. Rabbit corneal and mouse tail anaesthetic activities. Compound 1a 1b 1c 1d 2a 2b 2c 2d 3a 3b 3c 3d lidocaine HClc
a

Corneal anaesthetica inactive 74.3 9.3 12.1 3.9 73.8 8.9 inactive 4.7 1.5 3.2 2.1 30.1 5.6 22.5 4.5 105.0 3.5 48.5 9.1 92.3 8.0 100

Mouse tail anaestheticb 5.4 0.69 2.6 1.2 4.0 3.0 3.0 3.6 3.0 1.4 3.0 1.4 0.68 ( ( ( ( ( ( ( ( ( ( ( ( ( 0.38)102 0.31)102 0.43)102 0.42)102 0.36)102 0.37)102 0.33)102 0.35)102 0.29)102 0.25)102 0.31)102 0.28)102 0.39)102

Table V. Acute toxicity in mouse (LD50) and therapeutic index of selected compounds 1b, 1d, 3b and 3d.
Compound 1b 1d 3b 3d lidocaine HCl
a

LD50a

IC50

Therapeutic indexb LD50/IC50 M M M M M 8.83 4.37 6.21 3.19 8.23

6.09 ( 5.24 ( 8.69 ( 4.46 ( 5.60 (

0.49)102 0.38)102 0.53)102 0.35)102 0.39)102

M M M M M

0.69 ( 1.2 ( 1.4 ( 1.4 ( 0.68 (

0.31)102 0.42)102 0.25)102 0.28)102 0.39)102

Molar concentration of the injected solution (v = 0.2 mL/20 g body weight). bEvaluated as ratio between LD50 and IC50 (mouse tail test) expressed in mg/kg.

All compounds were in aqueous solution at 2% concentration. The values expressed as % of the anaesthetic activity of lidocaine (= 100), are means SE of three determinations. bIC50 values expressed as mol/L. cLidocaine hydrochloride was used for comparison.

As for the 4-benzoyl-1,2,3-triazole derivatives (2ad and 3ad), the results indicate that the nature and the position of the intermediate acetylacetamido side chain on the benzoyltriazole nucleus have signicant inuence on activity. In both tests the 2-substituted isomers appear to be more active than the 1-substituted derivatives. Compounds 3b and 3d stand out in the data set for their high activities (these values are comparable to those determined for lidocaine). It appears that the position and
Table IV. Duration of local anaesthetic activity in rat sciatic nerve block. Compound 1% 1a 1b 1c 1d 2a 2b 2c 2d 3a 3b 3c 3d lidocaine HCl
a

Durationa (min) 2% 45 ( 17.0) 160 ( 18.2) n.t. 160 ( 15.3) 72 ( 6.9) 63 ( 6.8) n.t. 72 ( 7.0) 150 ( 17.2) 210 ( 16.9) 50 ( 6.5) 180 ( 20.3) 117 ( 11.0)

the nature of the side chain affect surface and inltration activities in a similar way. For all considered compounds, with respect to the nature of the intermediate acetylacetamido side chain, analogues with a terminal diethylamino group or a piperidine ring displayed the highest activity in accordance with previous results [1, 2]. Reducing the size from a six to a ve-membered ring resulted in an improvement of anaesthetic activity. In order to evaluate the duration of the local anaesthetic activity, additional investigations were conducted by rat sciatic nerve block assay. In accordance with the previous results, when a 2% solution was used, as reported in table IV, several compounds (1b, 1d, 3a, 3b, and 3d) exhibited a better performance in blocking the rat sciatic nerve with respect to lidocaine (160, 160, 150, 210 and 180 min for motor activity recovering, respectively, versus 117 min for lidocaine). Finally, the acute toxicity and therapeutic index of the more active compounds in all anaesthetic tests (1b, 1d,
Table VI. Maximum response to different compounds in rat isolated right atria compared to lidocaine. Compound 1a 1b 1c 1d 2a 2b 2c 2d 3b 3c 3d lidocaine Maximum response (beats/min) 15 6 42 8 52 8 106 27 33 13 55 12 72 5 107 27 67 9 135 31 152 12 180 35 n 6 6 6 5 6 6 6 6 6 6 6 6

25 ( 10.5) 77 ( 6.8) inactive 90 ( 3.6) 40 ( 15.0) 35 ( 11.0) inactive 40 ( 12.0) 75 ( 7.0) 117 ( 8.0) 25 ( 12.3) 70 ( 6.6) 65 ( 10.7)

In vivo duration of local anaesthetic activity in rat sciatic nerve block (each rat received 0.2 mL of 1% and 2% anaesthetic solution, n.t. = not tested. The values are means SE of three determinations.

1048 3b, and 3d) were determined in mice. From the corresponding LD50 and LD50/IC50 values shown in table V it is evident that 1b is characterized by the most favourable ratio between toxicity and mouse tail anaesthetic activity (its therapeutic index is slightly higher than that measured for lidocaine). These studies showed that the presence of a piperidine ring as a terminal N-alkylamino group gives an increase in toxicity, particularly evident on compound 3d. In conclusion, a comparison between the pharmacological prole of benzotriazinone and benzoyltriazole derivatives versus that of the above mentioned benzotriazole counterpartes [1, 2] clearly shows that the replacement of the benzotriazole moiety, found in the preceeding derivatives, by a benzotriazinone or by a benzoyltriazole ring, seems to have a little inuence on the anaesthetic activity. These results might be ascribed to an analogue lipophilic, steric, and electronic complementarity between the benzotriazole residue compared to the benzotriazinone or to the benzoyltriazole moiety. As regards the results of chronotropic activity reported in table VI, all compounds induced a negative chronotropic effect in rat isolated right atria at high doses, in a concentration range greater than M. The intrinsic activity of different compounds in decreasing spontaneous beating rate in isolated right atria was distinct. Lidocaine was used as the pattern drug to compare the intrinsic activity of all compounds; in fact it is well known that lidocaine has local anaesthetic as well antiarrhythmic actions In series 1, compound 1d determines a larger magnitude of maximum response compared to drugs 1b, 1c, and 1a, whereas drug 1a was less effective in producing a negative chronotropic response among these compounds (gure 4 and table VI). In series 2, compound 2d had higher intrinsic activity compared to compounds 2a, 2b, and 2c (gure 5). The intrinsic activity of compounds 2a and 2b were similar. The intrinsic activity of compound 3d was the best of all studied compounds. Compound 3c showed similar 3d intrinsic activity even if less, whereas compound 3b induced the lowest activity of the series (gure 6 and table VI). A comparison between in vivo anaesthetic activity and in vitro negative chronotropic activity data displayed a similar pattern. In fact, clear negative chronotropic action was displayed by compound 1d and 3d, which bear a piperidine ring, but also compound 2d, which was poorly active on anaesthetic activity, showed an appreciable negative chronotropic action.

Figure 4. Negative chronotropic response for the compounds 1a, 1b, 1c, 1d, and lidocaine in isolated right atria from rats. Data are means SEM for 56 experiments.

5. Experimental protocols 5.1. Chemistry Melting points were determined on a Koer hot stage apparatus and are uncorrected. Structures described were supported by 1H-NMR spectra and microanalytical data. 1 H-NMR spectra were recorded on a Bruker WM 500 spectrometer using DMSO and CDCl3 as solvent; chemical shifts () are reported as follows: s, singlet; d,

Figure 5. Negative chronotropic response for the compounds 2a, 2b, 2c, 2d, and lidocaine in isolated right atria from rats. Data are means SEM for 6 experiments.

1049 5.1.2. General procedure for the preparation of compounds 1ad To 0.01 mol of 3-carbethoxymethyl-1,2,3-benzotriazin-4(3H)-one (5) dissolved in anhydrous methanol (50 mL) was added the appropriate amine (0.01 mol) dropwise. The reaction mixture was kept under reux with magnetic stirring for 812 h and monitored by TLC until the starting material had disappeared. After cooling the solvent was removed under reduced pressure and the residue was puried by silica-gel column chromatography using mixtures of diethyl ether/methanol 9:1 v/v. Further purication was obtained by crystallization from the appropriate solvent (table I). Yields ranging between 5266%. Spectral data of title compound 1a: 1H-NMR (CDCl3), = 8.33 (d, 1H, Ar-H, J = 7.5 Hz), 8.15 (d, 1H, Ar-H, J = 7.5 Hz), 7.94 (t, 1H, Ar-H, J = 7.5 Hz), 7.78 (t, 1H, Ar-H, J = 7.5 Hz), 5.11 (s, 2H, CH2-CO), 3.39 (t, 2H, NH-CH2, J = 7.5 Hz), 2.47 (t, 2H, CH2-N, J = 7.5 Hz) and 2.24 ppm (s, 6H, N(CH3)2). Similar 1H-NMR data occur in all derivatives of general formula 1. 5.1.3. Phenyl ethynyl ketone (7) A solution of chromium trioxide (0.10 mol) in water (30 mL) and concentrated sulphuric acid (8.5 mL) was slowly added to a stirred and cooled solution of phenylethynylcarbinol (6) (0.15 mol) in acetone (50 mL). The reaction mixture was carried out at 4 C under a nitrogen atmosphere. After stirring for 7 h, water was added to dissolve the precipitated chromium salts and the product was extracted with chloroform. Evaporation of the organic solution gave a yellow solid which was crystallized from aqueous methanol to give 16.6 g (85%) of 7 as pale yellow needles. The physical data are in agreement with those given in ref. [6]. 5.1.4. 4-benzoyl-1H-1,2,3-triazole (8) To a stirred and heated solution of NaN3 (0.10 mol) in anhydrous dimethylacetamide (80 mL), phenyl ethynyl ketone 7 (0.10 mol) dissolved in anhydrous dimethylacetamide (80 mL) was slowly added. The reaction mixture was kept at 100 C for 2 h. After stirring for a further 12 h at room temperature, evaporation of the solvent under reduced pressure gave a liquid residue which was diluted with water. The aqueous layer was acidied (pH = 5) with 10% HCl and extracted with ether (3 200 mL). The combined organic layers were dried and evaporated to give a solid residue which was puried by crystallization from ethyl alcohol: 16.4 g (95%) of 8. The physical data are in agreement with those given in ref. [7].

Figure 6. Negative chronotropic response for the compounds 3b, 3c, 3d, and lidocaine in isolated right atria from rats. Data are means SEM for 6 experiments.

doublet; t, triplet; m, multiplet. The spectra obtained conrmed the proposed structures. All pure compounds gave a satisfactory analysis (C, H, N) within 0.4% of theoretical values. Analytical TLC was performed on precoated silica-gel (0.2 mm GF 254, E Merck); the spots were located by UV (254 nm) light. Evaporation was performed in vacuo. Anhydrous sodium sulfate was used as a drying agent. Crude products were routinely passed through columns of silica gel (0.050.20 mm, Carlo Erba). Analytical data, melting points and crystallization solvents are reported in tables I and II. 5.1.1. 3-Carbethoxymethyl-1,2,3-benzotriazin-4(3H)-one (5) To a magnetically stirred solution of 1,2,3benzotriazin-4(3H)-one (4) (0.01 mol) and ethyl bromoacetate (0.01 mol) in butan-2-one (50 mL) was added K2CO3 (0.01 mol). The reaction mixture was heated under reux for 10 h and monitored by TLC. After cooling, the butan-2-one was removed under reduced pressure and the residue was diluted with H2O and extracted with CHCl3. The organic layer was dried and evaporated to dryness. The crude product 5 was puried by crystallization from methyl alcohol/diethyl ether. Yield 68%, m.p. 114116 C.1H-NMR (CDCl3) = 8.37 (d, 1H, Ar-H, J = 7.5 Hz), 8.20 (d, 1H, Ar-H, J = 7.5 Hz), 7.99 (t, 1H, Ar-H, J = 7.5 Hz), 7.82 (t, 1H, Ar-H), 5.19 (s, 2H, -NCH2), 4.28 (m, 2H, OCH2), and 1.30 ppm (t, 3H, CH3).

1050 5.1.5. 1-carbethoxymethyl-4-benzoyl-1H-1,2,3-triazole (9) and 2-carbethoxymethyl-4-benzoyl-2H-1,2,3-triazole (10) Anydrous K2CO3 (0.04 mol) was added to a solution of 4-benzoyl-1,2,3-triazole (8) (0.04 mol) and ethyl bromoacetate (0.04 mol) in 50 mL of butan-2-one. The mixture was reuxed for 12 h and monitored by TLC. After cooling, the butan-2-one was removed under reduced pressure and the residue was diluted with H2O and extracted with CHCl3. The organic layer was washed with 2 M NaOH, dried and evaporated to dryness. The obtained residue, containing 1- and 2-substituted isomers was nally fractionated by column chromatography using diethyl ether/n-hexane, 7:3 v/v, as eluent. Further purication of the isolated 1- and 2- isomers by crystallization from diethyl ether gave the nal products 9 and 10. Characterization by 1H-NMR spectra showed that the rst compound to be eluted was the 2-substituted 1,2,3-triazole (compound 10, yield 65%, m.p. 5557 C), whereas the 1-substituted isomer was eluted successively (compound 9, yield 35%, m.p. 132133 C). 9: 1H-NMR (DMSO), = 9. (s, 1H, H-triaz), 8.50 (d, 2H, Ar-H, J = 7.5 Hz), 7.68 (t, 1H, Ar-H, J = 7.5 Hz), 7.55 (t, 2H, Ar-H J = 7.5 Hz), 5.53 (s, 2H, CH2-CO), 4.27 (q, 2H, CH2O, J = 7.5Hz) and 1.28 (t, 2H, CH3, J = 7.5 Hz). 10: 1HNMR (DMSO), = 8 (s, 1H, H-triaz), 8.58 (d, 2H, Ar-H, J = 7.5 Hz), 7.70 (t, 1H, Ar-H, J = 7.5 Hz), 7.58 (t, 2H, Ar-H, J = 7.5 Hz), 5.61 (s, 2H, CH2-CO), 4.26 (q, 2H, CH2O, J = 7.5 Hz) and 1.25 (t, 2H, CH3, J = 7.5 Hz). 5.1.6. General procedure for the preparation of compounds 2ad and 3ad To 0.01 mol of the appropriate ethyl-4-benzoyl-1,2,3triazolacetate derivative (9 or 10) dissolved in anhydrous methanol was added dropwise the appropriate amine (0.01 mol). The reaction mixture was kept under reux with magnetic stirring for 812h and monitored by TLC, until the starting material had disappeared. After cooling, the solvent was removed by ltration under reduced pressure and the residue was puried by silica gel column chromatography using methanol as eluent. Further purication was obtained by crystallization from a mixture of methyl alcohol/diethyl ether. Yields ranging between 40 and 70%. Spectral data of the title compound 2a: 1HNMR (CDCl3), = 8 (s, 1H, H-triaz), 8.27 (d, 2H, Ar-H, J = 7.5 Hz), 7.62 (t, 1H, Ar-H, J = 7.5 Hz), 7.51 (t, 2H, Ar-H, J = 7.5 Hz), 5.14 (s, 2H, CH2-CO), 3.35 (t, 2H, NH-CH2, J = 7.5 Hz), 2.40 (t, 2H, CH2-N(CH3)2, J = 7.5 Hz) and 2.17 ppm (s, 6H, N(CH3)2). Similar 1H-NMR data occur in all derivatives of the general formula 2. Spectral data of the title compound 3a: 1H-NMR (CDCl3), = 8.31 (s, 1H, H-triaz), 8.27 (d, 2H, Ar-H, J = 7.5 Hz), 7.62 (t, 1H, Ar-H, J = 7.5 Hz), 7.52 (t, 2H, Ar-H, J = 7.5 Hz), 5.28 (s, 2H, CH2-CO), 3.33 (t, 2H, NH-CH2, J = 7.5 Hz), 2.35 (t, 2H, CH2-N(CH3)2, J = 7.5 Hz) and 2.15 ppm (s, 6H, N(CH3)2). Similar 1H-NMR data occur in all derivatives of the general formula 3. 5.2. Pharmacology 5.2.1. Corneal anaesthesia Local anaesthetic activity was evaluated in male New Zealand rabbits (Harlan-Nossan, Correzzana, Milan, weighing 2.42.8 kg) as local surface anaesthesia [10], by determining every 3 min the number of stimuli to the cornea, effected rhythmically with a Freys horse-hair, necessary to produce the blink reex. If the reex did not occurr after 100 stimulations, anaesthesia was considered total. At the beginning of the experiment care was taken to ascertain that this reex was normal in both eyes of the rabbits. All compounds were dissolved in 0.1 N HCl and the solution buffered to pH 67. The aqueous solutions (2%) of the compounds studied were dropped onto the conjunctival sac so that the space between the eyelids contained a clearly visible lm of solution for the set time of 3 min. Lidocaine solution (2%) was used for comparison. 5.2.2. Mouse tail anaesthesia Male Swiss mice (Harlan-Nossan, Correzzana, Milan, weighing 1820 g) were used. The test was performed according to the method of Bianchi [11] in which the aqueous anaesthetic solution (0.1 mL) is injected subcutaneously about 1 cm from the base of the tail. Fifteen minutes after injection, the pain reex of all injected animals was tested by applying a small artery clip to the zone where the compound was injected. The proportion of animals which did not show the usual pain reex within 30 s was noted for each dose. Lidocaine solutions were used for comparison. IC50 values were calculated for each compound by probit analysis using a computer program [12]. 5.2.3. Rat sciatic nerve block This test was performed according to Al-Saadi and Sneider [13] to determine conduction anaesthesia and its duration. Triplicate sets of three groups of three male Wistar rats (Harlan-Nossan, Correzzana, Milan, weighing 180200 g) were used. Each rat received an injection (0.2 mL) of the aqueous anaesthetic solution (1% and 2%) into the posterior side of the femur head. A positive effect of the drug resulted in a complete loss of motor control of the injected limb. In order to assess the duration of the effect, the animals were observed from the

1051 time of onset of the motor paralysis, at 10 min intervals over time, up to the rst sign of motor activity. 5.2.4. Acute toxicity The ip acute toxicity of the most active compounds 1b, 1d, 3b, and 3d was determined in male Swiss mice (Harlan-Nossan, Correzzana, Milan, weighing 1820 g) 7 days after treatment. LD50 values were calculated for each compound by probit analysis using a computer program [12]. 5.2.5. Chronotropic activity: functional assay using isolated rat right atria This test was performed according to Hughes and Smith [14]. Wistar rats (250300 g) were anesthetized with halothane and the hearts were rapidly removed and placed in oxygenated Krebs Henseleit buffer (KHB). The right atria were removed and mounted in a water jacketed tissue chamber (20 mL volume) containing KHB, pH 7.37.5, at 37 C and gassed with 95% O2/5% CO2. The composition of the KHB was (mM): NaCl, 124; KCl, 4.75; MgSO4, 1.30; CaCl2, 2.25; NaHCO3, 25.0; NaH2PO4, 0.6; dextrose, 10.0; sodium ascorbate, 0.3. 5.2.6. Construction and analyses of concentrationresponse curves Concentration-response curves for all compounds were constructed by the cumulative variation of agonist concentration at one-half log unit increments [15]. All concentration-response data were evaluated for a t to a logistic function in the form: E = Emax/((1 + (10c/10x)n) + U), where E is the increase in rate above basal; Emax is the maximum response that the agonist can produce; c is the logarithm of the EC50, the concentration of agonist that produces half-maximal response; x is the logarithm of the concentration of agonist; the exponential term, n, is a curve tting parameter that denes the slope of the concentration-response line, and is the response observed in the absence of added agonist. Nonlinear regression analyses to determine the parameters Emax, log EC50 and n were done using GraphPad Prism (GraphPad Software, San Diego, CA) with the constraint that = 0. Acknowledgements The NMR spectral data were provided by Centro di Ricerca Interdipartimentale di Analisi Strumentale, Universit degli Studi di Napoli Federico II. The assistance of staff is gratefully appreciated. References
[1] Caliendo G., Di Carlo R., Greco G., Grieco P., Meli R., Novellino E., Perissutti E., Santagada V., Eur. J. Med. Chem. 30 (1995) 603608. Caliendo G., Greco G., Grieco P., Mattace Raso G., Meli R., Novellino E., Perissutti E., Santagada V., Eur. J. Med. Chem. 31 (1996) 99110. Lfgren N., Studies on Local Anaesthetics, Hoeggstrom, Stockholm, 1948. Morgan P.H., Mathison I.W., J. Pharm. Sci. 65 (1976) 635648. Heymans L., Le Therizien L., Godfroid J., J. Med. Chem. 23 (1980) 184193. Bowden K., Heilbron I.M., Jones E.R.H., Weedon B.C.L., J. Chem. Soc. Part I (1946) 3945. Nesmeyanov A.N., Rybinskaya M.I., Dolk. Akad. Nauk. SSSR 158 (1964) 408410. Biagi G., Ferretti M., Livi O., Scartoni V., Lucacchini A., Mazzoni M., Il Farmaco Ed. Sc. 41 (1986) 388400. Livi O., Biagi G., Ferretti M., Lucacchini A., Barili P.L., Eur. J. Med. Chem. -Chim. Ther. 18 (1983) 471475. Regnier J., Bull. Sci. Pharm. 30 (1923) 580586. Bianchi C., Br. J. Pharmacol. 11 (1956) 104106. Tallarida R.J., Murray R.B., Manual of Pharmacological Calculations with Computer Programs, 2nd Edition, Springer, New York, 1981. Al Saadi D., Sneider W.E., Arzneim. -Forsch. 41 (1991) 195198. Hughes I.E., Smith J.A., J. Pharm. Pharmacol. 30 (1978) 124126. Van Rossum J.M., Arch. Int. Pharmacodyn. Ther. 143 (1963) 29933.

[2]

[3] [4] [5] [6] [7] [8] [9] [10] [11] [12]

[13] [14] [15]

Eur. J. Med. Chem. 34 (1999) 10531060 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1053

Original article

Synthesis and in vitro antitumour evaluation of benzothiazole-2-carbonitrile derivatives

Valrie Bnteaua, Thierry Bessona*, Jrme Guillarda, Stphane Lonceb, Bruno Pfeifferc
a

Laboratoire de Gnie Protique et Cellulaire, UPRES 2001, Ple Sciences et Technologie, Universit de La Rochelle, avenue Marillac, F-17042 La Rochelle cedex 1, France b Institut de Recherche Servier, 11, rue des Moulineaux, 92150 Surennes, France c A.D.I.R., 1, rue Carle Hbert, 92415 Courbevoie cedex, France (Received 20 May 1999; revised 7 July 1999; accepted 15 July 1999)

Abstract Novel benzothiazole derivatives have been synthesised via the corresponding imino-1,2,3-dithiazoles. The cytotoxicity of some of these polyheterocyclic compounds was studied. Our results show that 2-cyano derivatives exhibit a medium in vitro antitumour activity. 1999 ditions scientiques et mdicales Elsevier SAS imino-1,2,3-dithiazoles / antitumour activity / benzothiazoles / benzodioxines

1. Introduction The benzothiazole ring is present in various marine or terrestrial natural compounds which have useful biological activities [14]. Because we are interested in heterocyclic systems with potential pharmacological value, we decided to synthesise new benzothiazole derivatives which are related to synthetic thiazoles which have shown antitumour activity [5]. Novel dioxinobenzothiazoles were also prepared with the aim to enhance the antiproliferative activity. In this paper, we describe the biological evaluation of 4,7-dimethoxybenzothiazoles and dioxinobenzothiazoles prepared via N-arylimino-1,2,3-dithiazoles which have proved to be highly versatile intermediates in heterocyclic chemistry [68]. 2. Chemistry 2.1. 4,7-Dimethoxybenzothiazoles 4,5-Dichloro-1,2,3-dithiazolium chloride 1 is a pale greenish yellow solid, insoluble in organic solvents. It is
*Correspondence and reprints

completely stable in a dry inert atmosphere but reacts slowly with moisture to form 4-chloro-1,2,3-dithiazol-5one. This compound is readily prepared from chloroacetonitrile and disulfur dichloride [9], reacts rapidly with 2,5-dimethoxy-aniline, in dichloromethane at room temperature, to give the stable N-arylimino-1,2,3dithiazoles 2 in high yield (84%). Pyrolysis of these imines gave 2-cyanobenzothiazoles 3 by cyclisation of the ortho carbon onto sulfur with liberation of the other sulfur atom and hydrogen chloride [10] (gure 1). Removal (hydrolysis and decarboxylation) of the cyano group in the thiazole ring was performed by vigorous heating of compound 3a in concentrated hydrochloric acid. The decyanated benzothiazole 4 was isolated in good yield (70%) (gure 2). Using standard conditions for the transformation of cyano groups into carboxylic acid or amido groups [11], the starting thiazole 3a was treated with aqueous sodium hydroxide to give the acid 5, or with sulfuric acid to provide the amide 6, in quite good yields. The substituted amide 7 was prepared by treatment of the acid 5 with N,N-dimethylethylenediamine in the presence of 1-(3-dimethylaminopropyl)-3-ethyl carbodiimide hydrochloride (EDCI) and 1-hydroxybenzotriazole (HOBT), a method commonly used for the coupling of amino acids (gure 2).

1054

Figure 1. Reactions and conditions: a) pyridine, dichloromethane, 15 C, 3 h (2a, 40%) or room temperature, 3 h (2b, 84%); b) toluene, sealed tube, reux, 4 h (3a, 70%); c) PyHBr3, pyridine, reux, 1.5 h (3b, 72%).

Decyanation of compound 3b, using the conditions described above, afforded the amino decyanated benzothiazole 8, in which the amino group was easily reprotected, to give 9, by treatment with acetic anhydride in the presence of pyridine. The quinone 10 was obtained by oxidation of 3b with cerium ammonium nitrate (gure 3). Starting from the commercially available isothiocyanate 11, the 2-alkoxyderivative 13 was synthesised by a Jacobson process via the intermediate thiocarbamate 12 (gure 4). In this case, an electron releasing group is present in the 2-position of the benzothiazole ring. 2.2. Dioxinobenzothiazoles [12] The chemistry of the salt 1 described above also allows a rapid access to the dioxinobenzothiazoles 1518, from

the starting 6-amino-2,3-dihydro-1,4-benzodioxin (gure 5). The bromination (NBS, CCl4, AIBN)debromination (NaI, acetone) sequence previously described in several syntheses of benzodioxins [13], lead to the expected products 19 and 20 (gure 5), whilst heating of the brominated intermediate 21 in pyridine, in the presence of one equivalent of copper iodide (CuI), allowed an alternative access to the linear compound 15 (the use of the standard method afforded the angular brominated derivative 22) (gure 6). 3. Pharmacology Fifteen compounds were evaluated in vitro for their antiproliferative activity using the murine L1210 leukaemia cell line [14]. The results expressed as IC50

Figure 2. Reactions and conditions: a) conc. HCl, reux, 8 h, 70%; b) NaOH 10%, 80 C, 3.5 h, 64%; c) conc. H2SO4, room temperature, 2 h, 72%; d) N,N-dimethylethylenediamine, EDCI, HOBT, DMF, 0 C, 4 days, 43%.

1055

Figure 3. Reactions and conditions: a) conc. HCl, reux, 1.5 h, 83%; b) CAN, CH3CN, H2O, room temperature, 15 min, 56%; c) Ac2O, pyridine, room temperature, 12 h, 56%.

(concentration reducing the cell proliferation by 50%) are reported in table I. Cell cycle perturbations induced by the most active compounds (IC50 < 20 M) were also investigated.

4. Results and discussion The 2-cyano-4,7-dimethoxybenzothiazole derivatives 3a and 3b were rst evaluated and found practically

Figure 4. Reactions and conditions: a) PrOH, NaH, 83%; b) K3FeCN6, NaOH, 28%.

Figure 5. Reagents and conditions: a) PyHBr3 (1 eq.), DMF/pyridine, reux, 90 min, 58% (29% for 15 and 29% for 16); b): HCl, reux, 2.5 h, 80% (17) and 65% (18); c) i) NBS/AIBN, CCl4, h, reux, 10 h; ii) NaI, acetone, reux, 1.5 h, 69% (19) and 27% (20); d) Br2 (1 eq.), CH3COOH, room temperature, 2 h, 98%.

1056 results already published on dioxinocoumarins [15, 16]. The dioxinobenzothiazole 19 was found able to block L1210 cells in the G2 + M phase of the cell cycle. 5. Conclusion
Figure 6. Reagents and conditions: a) PyHBr3 (1 eq.), DMF/ pyridine, reux, 2 h, 25%; b) CuI, pyridine, reux, 90 min, 55%.

equipotent on cell proliferation with IC50s of, respectively, 20.6 M and 25.2 M. Also, both were found to be able to almost totally block the cells in the G2 + M phase of the cell cycle. Suppression of the 2-cyano substituent (compounds 4 and 9) or its replacement by a 2-carboxy (compound 5), a 2-aminocarbonyl (compound 6), a 2-[2(N,N-dimethylamino)ethylaminocarbonyl] (compound 7) or a 2-propoxy (compound 13) substituent led to inactive derivatives. Compound 10 which is a quinone derivative was found signicantly more active that its 4,7dimethoxy counterpart 3b with an IC50 of 5M versus 25.2 M. This compound was unfortunately devoid of any specic effect on the cell cycle. All the 2-cyano-dioxinobenzothiazoles have also shown a relatively interesting antiproliferative activity. As for the 4,7-dimethoxybenzothiazoles, removal of the cyano substituent present in the 2-position of the dioxinobenzothiazole ring (compound 17) involved the lost of any activity (IC50 > 100 M). The unsaturated compounds 19 and 20 were found signicantly more active than their saturated conterparts 15 and 18, with IC50s of, respectively, 18.2 M and 31.7 M, conrming the

In conclusion, we have described the synthesis of novel benzothiazoles and dioxinobenzothiazoles, among which the 2-cyano derivatives exhibit interesting in vitro antitumour activity. Presence of unsaturation in the dioxin moiety, in combination with the cyano group in the 2-position of the thiazole ring, did not really involve any specic effect on the cell cycle. Our results suggest that introduction of the thiazolo-2-carbonitrile ring into more extended and more complexe heterocyclic moieties could open the door to promising applications. 6. Experimental protocols 6.1. Chemistry Melting points were determined using a Koer banc and are uncorrected. IR spectra were recorded on a Perkin-Elmer Paragon 1000PC instrument. 1H- and 13CNMR were recorded on a JEOL JNM LA400 (400 Mhz) spectrometer (Centre Commun dAnalyse, Universit de La Rochelle); chemical shifts () are reported in parts per million (ppm) downeld from tetramethylsilane (TMS), which was used as internal standard. Mass spectra were recorded on a Varian MAT311 in the Centre de Mesure Physiques de LOuest (C.R.M.P.O.), Universit de Rennes. Chromatography was carried out on silica gel 60

Table I. Characteristics and pharmacological activity of synthesised compounds. Compound 3a 3b 4 5 6 7 9 10 13 15 16 17 19 20 22

a

M.p. (C) 174 173 108 > 230 240 144 154 242 73 178 156 142 176 150 250

Formula C10H8N2O2S C12H11N3O3S C9H9NO2S C10H9NO4S C10H10N2O3S C14H19N3O3S C11H12N2O3S C10H5N3O3S C12H15NO3S C10H6N2O2S C10H6N2O2S C9H7NO2S C10H4N2O2S C10H4N2O2S C10H5BrN2O2S

IC50 () 20.6 25.2 > 100 > 100 > 100 > 100 > 100 5 > 100 58.4 42.3 > 100 18.2 31.7 80.8

% of L1210 cells in the G2 + M phasea (M) 64% (50) 77% (100) n.e.b n.e. n.e. n.e. n.e. non specic n.e. n.e.b n.e. n.e. 41% (50) n.e. n. e.

24% of untreated control cells were in the G2 + M phase of the cell cycle. bn.e.: not evaluated (for IC50 > 30 M).

1057 at medium pressure and the sample mixtures were applied to the column preadsorbed onto silica. Light petroleum refers to the fraction b.p. 4060 C. Further solvents were used without purication. Thin-layer chromatography was performed on Merck Kieselgel 60 F254 aluminium backed plates. Spectral data for compounds 1422 are consistent with assigned structures as previously described in ref. [12]. 6.1.1. 2,5-Dimethoxy-N-(4-chloro-5H-dithiazol-5-ylidene)aniline 2a Under an inert atmosphere, 4,5-dichloro-1,2,3dithiazolium chloride (15.42 g, 73.7 mmol) and pyridine (10.8 mL, 134 mmol) were added to a stirred solution of 2,5-dimethoxyaniline (10.3 g, 67 mmol) in dichloromethane (150 mL) cooled at 15 C. After 2 h, the mixture was ltered and the solvent removed in vacuo. The crude residue was puried by column chromatography (eluent: light petroleum/ethyl acetate: 9/1) to afford compound 2a as an orange oil (7.8 g, 40%); IR (lm): 2 938 and 2 833 (CH3), 1 586 (C=N), 1 498, 1 464, 1 278, 1 223, 1 195, 1 165, 1 044 cm1; 1H-NMR (CDCl3): 3.78 (s, 3H, OCH3), 3.82 (s, 3H, OCH3), 6.69 (d, 1H, J = 2.9 Hz, 6-H), 6.75 (dd, 1H, J = 2.9 Hz and J = 8.8 Hz, 4-H), 6.93 (d, 1H, J = 8.8 Hz, 3-H); 13 C-NMR (CDCl3): 56.65, 57.21, 106.11, 112.91, 114.43, 141.71, 144.68, 148.48, 154.97, 161.02; MS (EI): m/z 288 (M+), 188, 148; HRMS: calc. for C10H9ClN2O2S2: 287.9794, found: 287.9801. 6.1.2. 4-Acetamido-2,5-dimethoxy-N-(4-chloro-5H-1,2,3dithiazol-5-ylidene) aniline 2b Under an inert atmosphere, 4,5-dichloro-1,2,3dithiazolium chloride (220 mg, 1.04 mmol) was added to a stirred solution of 4-acetamido-2,5-dimethoxyaniline (200 mg, 0.95 mmol) in dichloromethane (10 mL). After 1 h, pyridine (110 mg, 1.39 mmol) was added and the mixture stirred for 15 min. The solvent was removed in vacuo and the crude residue puried by column chromatography (eluent: dichloromethane/ethyl acetate: 4/1) to afford compound 2b (275 mg, 84%) as an orange powder; m.p. 155 C; IR (KBr): 3 506 (NH), 1 671 (C=O), 1 483, 1 402, 1 214, 1 041, 862 cm1; 1H-NMR (CDCl3): 2.22 (s, 3H, NHCOCH3), 3.85 (s, 3H, OCH3), 3.87 (s, 3H, OCH3), 6.72 (s, 1H, 6-H), 7.80 (s, 1H, NH), 8.28 (s, 1H, 3-H); 13C-NMR (CDCl3): 25.01, 56.20, 56.24, 102.98, 104.85, 126.62, 133.91, 141.64, 143.81, 147.80, 158.88, 168.24; MS (EI): m/z 345 (M+), 252 (MCNSCl), 195; HRMS: calc. for C12H12ClN3O3S2: 345.0009, found: 345.0016. 6.1.3. 4,7-Dimethoxybenzothiazole-2-carbonitrile 3a In a closed system, a solution of 2a (871 mg, 3.02 mmol) in toluene (5 mL) was heated in an oil bath at 200210 C for 8 h. After cooling, the toluene was removed in vacuo. The crude residue was puried by column chromatography (eluent: dichloromethane/light petroleum: 6/4) and recrystallised from ethanol to afford 3a (471 mg, 70%) as yellow needles; m.p. 174 C; IR (KBr): 2 229 (nitrile), 1 595 (C=N), 1 501, 1 454, 1 280, 1 140, 1 094, 1 046, 969 cm1; 1H-NMR (CDCl3): 3.98 (s, 3H, OCH3), 4.05 (s, 3H, OCH3), 6.93 (s, 2H, 5-H and 6-H); 13C-NMR (CDCl3): 56.36, 56.49, 108.10, 108.44, 112.97, 126.32, 135.55, 143.75, 147.59, 148.91; MS (EI): m/z 220 (M+), 205 (M+CH3), 191; HRMS: calc. for C10H8N2O2S: 220.0306, found: 220.0303. 6.1.4. 6-Acetamido-4,7-dimethoxybenzothiazole-2-carbonitrile 3b Under an inert atmosphere, a mixture of 2b (1.60 g, 4.63 mmol) and pyridinium perbromide (1.63 g, 5.10 mmol) in pyridine (20 mL) was heated at reux for 2.5 h. After cooling, pyridine was removed in vacuo and the crude residue puried by column chromatography (eluent: dichloromethane/ethyl acetate: 95/5). Recrystallisation from ethanol afforded compound 3b (927 mg, 72%) as yellow needles; m.p. 173 C; IR (KBr): 3 390 (NH), 2 226 (nitrile), 1 683 (C=O), 1 600 (C=N), 1 523, 1 448, 1 420, 1 240, 1 132, 1 050 cm1; 1H-NMR (CDCl3): 2.30 (s, 3H, NHCOCH3), 3.96 (s, 3H, OCH3), 4.08 (s, 3H, OCH3), 7.86 (bs, 1H, NH), 8.34 (s, 1H, 5-H); 13 C-NMR (CDCl3): 25.20, 56.68, 59.88, 101.65, 112.78, 128.05, 132.37, 132.79, 134.56, 139.33, 151.25, 168.62; MS (EI): m/z 277 (M+), 220 (M+CNOCH3); HRMS: calc. for C12H11N3O3S: 277.0521, found: 277.0524. 6.1.5. 4,7-Dimethoxybenzothiazole 4 A suspension of compound 3a in concentrated hydrochloric acid (20 mL) was heated at reux for 8 h. The mixture was cooled at 0 C, neutralised to pH 8 using saturated aqueous sodium hydrogen carbonate and extracted with dichloromethane. The combined extracts were dried over MgSO4 and the solvent removed in vacuo. After recrystallisation from hexane, compound 4 (177 mg, 70%) was obtained as colourless needles; m.p. 108 C; IR (KBr): 3 074 (CHunsat.), 2 959 (CH3), 1 594 (C=N), 1 500, 1 456, 1 439, 1 345, 1 263, 1 184, 1 150 cm1; 1H-NMR (CDCl3): 3.97 (s, 3H, OCH3), 4.03 (s, 3H, OCH3), 6.78 (d, 1H, J = 8.6 Hz, Harom.), 6.86 (d, 1H, J = 8.6 Hz, Harom.), 8.91 (s, 1H, 2-H); 13C-NMR (CDCl3): 55.29, 55.63, 104.81, 106.39, 123.95, 144.11,

1058 147.59, 152.49, 152.53; MS (EI): m/z 195 (M+), 180 (M+CH3), 166; HRMS: calc. for C9H9NO2S: 195.0354, found: 195.0361. 6.1.6. 4,7-Dimethoxybenzothiaole-2-carboxylic acid 5 A suspension of compound 3a (204 mg, 0.93 mmol) in 10% aqueous sodium hydroxide (10 mL) was heated at 80 C for 3.5 h. After cooling to room temperature, the mixture was poured onto iced water and acidied to pH 1 using 10% aqueous hydrochloric acid. The yellow precipitate that separated was ltered under vacuum and dried. The product was puried by column chromatography (eluent: hexane/ethyl acetate/methanol: 1/1/0.5 then methanol). Recrystallisation from water afforded compound 5 (141 mg, 64%) as yellow needles; m.p. 230 C (dec); IR (KBr): 3 367 (broad: OH), 2 838 (CH3), 1 632 (C=O), 1 499, 1 383, 1 263, 1 189, 1 097, 1 044, 974 cm1; 1H-NMR (DMSO-d6): 3.87 (s, 3H, OCH3), 3.88 (s, 3H, OCH3), 6.90 (s, 2H, Harom.); 13C-NMR (DMSO-d6): 55.92, 56.11, 106.04, 107.53, 126.56, 144.48, 147.54, 148.19, 162.07, 169.85; MS (EI): m/z 209 (MCH2O), 195 (MCO2), 180 (MCO2, CH3), 166; HRMS could not be measured due to the absence of the molecular pic. 6.1.7. 4,7-Dimethoxybenzothiazole-2-carboxamide 6 A solution of compound 3a (703 mg, 3.19 mmol) in concentrated sulfuric acid (5 mL) was stirred at room temperature for 2 h. After cooling at 0 C, the mixture was basied using 10% aqueous sodium hydroxide and then extracted with ethyl acetate. The combined extracts were dried over MgSO4. Removal of the solvent in vacuo followed by a recrystallisation from ethanol gives compound 6 (548 mg, 72%) as amber needles; m.p. 240 C; IR (KBr): 3 406 (NH), 1 682 (C=O), 1 600 (C=N), 1 558, 1 505, 1 118, 799 cm1; 1H-NMR (CDCl3): 3.97 (s, 3H, OCH3), 4.04 (s, 3H, OCH3), 5.73 (bs, 1H, NH), 6.84 (d, 1H, J = 8.6 Hz, Harom.), 6.89 (d, 1H, J = 8.6 Hz, Harom.), 7.41 (bs, 1H, NH); 13C-NMR (DMSO-d6): 55.98, 56.15, 107.50, 108.33, 126.58, 143.89, 147.44, 148.33, 161.17, 163.60; MS (EI): m/z 238 (M+), 223 (M+CH3), 209, 192, 180; HRMS: calc. for C10H10N2O3S: 238.0412, found: 238.0415. 6.1.8. 4,7-Dimethoxybenzothiazole-2-[2-(N,N-dimethylamino)ethyl]carboxamide 7 Under an inert atmosphere, 1-(3-dimethylaminopropyl)-3-ethyl-carbodiimide hydrochloride (1.22 g, 6.47 mmol), 1-hydroxybenzotriazole (840 mg, 6.47 mmol) and N,N-dimethylethylenediamine (0.7 mL, 6.47 mmol) were added to a stirred solution of compound 5 (703 mg, 2.94 mmol) in DMF (20 mL) at 0 C. The mixture was stirred for 4 days, allowing the temperature to increase slowly. Water was then added with cooling and the crude material extracted with ethyl acetate. The combined extracts were washed with water (three times) and dried over MgSO4. After removal of the solvent in vacuo, the product was puried by column chromatography (eluent: dichloromethane/methanol: 95/5) and recrystallised from hexane/ethanol to afford compound 7 (393 mg, 43%) as pale yellow needles; m.p. 144 C; IR (KBr): 3 168 (NH), 2 950 (CH3), 2 824 (CH2), 1 662 (C=O), 1 540, 1 276, 1 188, 1 143, 1 090, 1 045 cm1; 1H-NMR (CDCl3): 2.24 (s, 6H, N(CH3)2), 2.50 (t, 2H, J = 6.25 Hz, CH2N), 3.54 (q, 2H, J = 6.25 Hz, NHCH2), 3.91 (s, 3H, OCH3), 3.99 (s, 3H OCH3), 6.77 (d, 1H, J = 8.7 Hz, Harom.), 6.82 (d, 1H, J = 8.7 Hz, Harom.), 7.73 (bs, 1H, NH); 13C-NMR (CDCl3): 37.47, 45.29 (2), 56.05, 56.34, 57.86, 106.40, 107.07, 128.11, 144.29, 148.29, 148.39, 159.89, 163.77; MS (EI): m/z 309 (M+), 58 ((CH3)2N=CH2+); HRMS: calc. for C14H19N3O3S: 309.1147, found: 309.1155. 6.1.9. 6-Amino-4,7-dimethoxybenzothiazole 8 A solution of compound 3b (93 mg, 0.34 mmol) in concentrated hydrochloric acid (7 mL) was heated at reux for 1.5 h. After cooling at room temperature, the mixture was basied to pH 8 using saturated aqueous sodium hydrogen carbonate and extracted with dichloromethane. The combined extracts were washed with water and brine, and dried over MgSO4. The solvent was removed in vacuo and the crude residue was puried by column chromatography (eluent: light petroleum/ethyl acetate: 1/1) to afford compound 8 (59 mg, 83%) as a brown powder; m.p. 153 C; IR (KBr): 3 440/3 310 (NH2), 1 614 (C=N), 1 500, 1 463, 1 398, 1 237, 1 043, 987 cm1; 1H-NMR (CDCl3): 3.86 (s, 3H, OCH3), 3.99 (s, 3H, OCH3), 6.40 (s, 1H, 5-H), 8.60 (s, 1H, 2-H); 13 C-NMR (CDCl3): 56.09, 58.75, 98.02, 128.61, 133.24, 137.35, 137.48, 148.11, 150.38; MS (EI): m/z 210 (M+), 195 (M+CH3), 154 (M+C3H4O). 6.1.10. 6-Acetamido-4,7-dimethoxybenzothiazole 9 Acetic anhydride (2.2 mL, 23.33 mmol) was added to a stirred solution of 8 (265 mg, 1.26 mmol) in pyridine (15 mL). The mixture was strirred overnight at room temperature. Then, water was added, the product was extracted with ethyl acetate, puried by column chromatography (eluent: ethyl acetate/light petroleum: 7/3) and recrystallised from dichloromethane/light petroleum to afford compound 9 (179 mg, 56%) as pale yellow needles; m.p. 154 C; IR (KBr): 3 296 (NH), 2 963, 1 668 (C=O), 1 606 (C=N), 1 532, 1 463, 1 386, 1 254, 1 222, 1 042 cm1; 1H-NMR (CDCl3): 2.25 (s, 3H, NHCOCH3), 3.93 (s, 3H, OCH3), 4.04 (s, 3H, OCH3),

1059 7.78 (bs, 1H, NH), 8.18 (s, 1H, 5-H), 8.79 (s, 1H, 2-H); 13 C-NMR (CDCl3): 25.09, 56.35, 59.74, 100.80, 126.26, 128.94, 135.58, 140.60, 149.89, 150.81, 168.42; MS (EI): m/z 252 (M+), 237 (M+CH3), 195; HRMS: calc. for C11H12N2O3S: 252.0569, found: 252.0567. 6.1.11. 6-Acetamido-2-cyano-4,7-dioxobenzothiazole 10 A solution of cerium ammonium nitrate (2.59 g, 4.72 mmol) in water (10 mL) was added to a stirred solution of 3b (524 mg, 1.89 mmol) in acetonitrile (20 mL) over 10 min. After 15 min the mixture was extracted with ethyl acetate, the combined extracts were dried over MgSO4 and the solvents removed in vacuo. The crude residue was puried by column chromatography (eluent: dichloromethane/ethyl acetate: 95/5) to afford compound 10 (261 mg, 56%) as orange needles; m.p. 242 C; IR (KBr): 3 266 (NH), 3 108 (CHunsat.), 2 238 (nitrile), 1 708 (C=O), 1 522, 1 427, 1 331, 1 194, 1 141, 1 014 cm1; 1H-NMR (CDCl3): 2.32 (s, 3H, NHCOCH3), 7.94 (s, 1H, 5-H), 8.13 (bs, 1H, NH); 13 C-NMR (CDCl3): 25.04, 111.15, 115.47, 138.29, 140.03, 142.88, 153.75, 169.02, 175.34, 178.42; MS (EI): m/z 247 (M+), 205; HRMS: calc. for C10H5N3O3S: 247.0052, found: 247.0046. 6.1.12. 4,7-Dimethoxy-2-propyloxybenzothiazole 13 A mixture of 12 (698 mg, 2.73 mmol) and 30% aqueous sodium hydroxide (2.9 mL, 21.84 mmol) in ethanol (3 mL) was added dropwise to a solution of K3FeCN6 (3.6 g, 10.62 mmol) in water (5 mL) heated at 85 C. After 1.5 h, the mixture was allowed to cool to room temperature and extracted with dichloromethane. The combined extracts were dried over MgSO4 and the solvents removed in vacuo. The crude residue was puried by column chromatography (eluent: light petroleum/ethyl acetate: 7/1). Recrystallisation from hexane afforded compound 13 (191 mg, 28%) as colourless needles; m.p. 73 C; IR (KBr): 2 964 and 2 833 (CH2/CH3), 1 534, 1 500, 1 467, 1 382, 1 340, 1 262, 1 097, 1 051 cm1; 1H-NMR (CDCl3): 1.05 (t, 3H, J = 7 Hz, CH3), 1.86 (sextuplet, 2H, J = 7 Hz, CH2), 3.89 (s, 3H, OCH3), 3.96 (s, 3H, OCH3), 4.57 (t, 2H, J = 7 Hz, CH2), 6.63 (d, 1H, J = 8.5 Hz, Harom.), 6.77 (d, 1H, J = 8.5 Hz, Harom.); 13C-NMR (CDCl3): 10.29, 22.20, 55.93, 56.35, 73.68, 103.90, 107.25, 121.18, 139.71, 146.32, 147.86, 173.04; MS (EI): m/z 253 (M+), 211, 196; HRMS: calc. for C12H15NO3S: 253.0773, found: 253.0774. 6.2. Antiproliferative activity L1210 cells (murine leukaemia) provided by the NCI, Frederik, USA were cultivated in RPMI 1640 medium (Gibco) supplemented with 10% foetal calf serum, 2 mM L-glutamine, 100 units/mL penicillin, 100 g/mL streptomycin, and 10 mM HEPES buffer (pH = 7.4). Cytotoxicity was measured by the microculture tetrazolium assay as described in ref. [14]. Cells were exposed to graded concentrations of the compounds for 48 h and results expressed as IC50 (concentration which reduced by 50% the optical density of treated cells with respect to untreated controls). For the cell cycle analysis, L1210 cells (2.5 105 cells/mL) were incubated for 21 h with various concentrations of the compounds, then xed by 70% ethanol (v/v), washed and incubated in PBS containing 100 g/ mL RNAse and 25 g/mL propidium iodide for 30 min at 20 C. For each sample, 1 104 cells were analysed on an ATC3000 ow cytometer (Brucker, France) using an argon laser (Spectra-Physics) emitting 400 mW at 488 nm. The uorescence of propidium iodide was collected through a 615 nm long-pass lter. Data are displayed as linear histograms and results are expressed as the percentage of cells found is the G2 + M phase of the cell cycle. Acknowledgements We thank the Communaut de Villes de lAgglomration de La Rochelle, the society ADIR (Groupe SERVIER) and the Comit de CharenteMaritime de la Ligue Nationale Contre le Cancer for nancial support. References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Gunawardana G.P., Kohmoto S., Gunasekara S.P., McConnel O.J., Koehn F.E., J. Am. Chem. Soc. 110 (1988) 48564858. Gunawardana G.P., Kohmoto S., Burres N.S., Tetrahedron Lett. 30 (1989) 43594362. Gunawardana G.P., Koehn F.E., Lee A.Y., Clardy J., He H.Y., Faulkner J.D., J. Org. Chem. 57 (1992) 15231526. Carroll A.R., Scheuer P.J., J. Org. Chem. 55 (1990) 44264431. Shi D.F., Bradshaw T.D., Wrigley S., McCall C.J., Lelieveld P., Fichtner I., Stevens M.F.G., J. Med. Chem. 39 (1996) 33753384. Rees C.W., J. Heterocycl. Chem. 29 (1992) 639651. Besson T., Emayan K., Rees C.W., J. Chem. Soc. Perkin Trans. 1 (1995) 20972102. Besson T., Guillaumet G., Lamazzi C., Rees C.W., Synlett. (1997) 704706. Appel R., Janssen H., Siray M., Knoch F., Chem. Ber. 118 (1985) 16321643. English R.F., Rakitin O.A., Rees C.W., Vlasova O.G., J. Chem. Soc. Perkin Trans. 1 (1997) 201205. March J., in: Advanced Organic Chemistry, 4th edition, WileyInterscience Publication, 1992, p. 887.

1060
[12] Besson T., Guillard J., Tetrahedron 55 (1999) 51395144. In this paper the multistep synthesis of such compounds was transposed to a focused microwave oven. Besson T., Ruiz N., Coudert G., Guillaumet G., Tetrahedron 51 (1995) 31973204. Lonce S., Prez V., Casabianca-Pignde M.R., Anstett M., Bisagni E., Atassi, G., Invest. New Drugs 14 (1996) 169180. Csik G., Besson T., Coudert G., Guillaumet G., Nocentini S., J. Photochem. Photobiol. B: Biol. 19 (1993) 119124. Csik G., Ronto G., Nocentini S., Averbeck S., Averbeck D., Besson T., Coudert G., Guillaumet, G., J. Photochem. Photobiol. B: Biol. 24 (1994) 129139.

[15] [16]

[13] [14]

Eur. J. Med. Chem. 34 (1999) 10611070 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1061

Original article

Synthesis and aldose reductase inhibitory activity of a new series of 5-[[2-(-carboxyalkoxy)aryl]methylene]-4-oxo-2thioxothiazolidine derivatives
Makoto Murata*, Buichi Fujitani, Hiroyuki Mizuta
Department of Chemistry I. Discovery Research Laboratories, Dainippon Pharmaceutical Company Ltd., Enoki 33-94, Suita/Osaka 564-0053, Japan (Received 20 April 1999; accepted 15 July 1999)

Abstract A new series of 5-[[2-(x-carboxyalkoxy)aryl]methylene]-4-oxo-2-thioxothiazolidine derivatives was synthesized and evaluated for their potency as aldose reductase inhibitors (ARIs). Their activities were examined in terms of their inhibitory effect on rat lens aldose reductase in vitro and in terms of the preventive effect on sorbitol accumulation in the sciatic nerve of streptozotocin (STZ)-induced diabetic rats in vivo. Of these compounds, some of the naphthylmethylene thiazolidine derivatives were comparable to Zenarestat in the inhibitory potency in vitro and in vivo. In particular, compound 30 was 1.5 times more potent than Zenarestat in the in vivo activity, and had an adequate potency for clinical development. 1999 ditions scientiques et mdicales Elsevier SAS 5-[[2-(x-carboxyalkoxy)aryl]methylene]-4-oxo-2-thioxothiazolidine derivatives / aldose reductase inhibitory activity / sorbitol accumulation inhibition

1. Introduction Aldose reductase (AR), which is the rate limiting enzyme of the polyol pathway, is implicated in diabetic complications. Since AR has a low substrate affinity for glucose, the activity of the polyol pathway is very low at normal physiological glucose concentrations [1]. However, when there is elevated glucose concentration in diabetes, excessive sorbitol production from glucose by AR is thought to cause cellular damage as a result of osmotic imbalance [2]. This effect leads to the development of diabetic complications such as neuropathy, retinopathy, nephropathy, and cataract formation [36]. Recently, numerous compounds have been selected as potential aldose reductase inhibitors (ARIs) [711], whose representatives are shown in gure 1. These compounds possess an acidic proton which is attached to the imidic nitrogen or an acetic acid moiety in the mol*Correspondence and reprints

ecule [12, 13]. Thus, we gave our attention to the compounds having a 2-(-carboxyalkoxy)aryl moiety and synthesized those compounds. In this paper, we report the synthesis and AR inhibitory activity of the 5-[[2-(carboxyalkoxy)aryl]methylene]-4-oxo-2-thioxo-thiazolidine derivatives 1 shown in gure 1. 2. Chemistry The 4-substitued-2-formylphenoxyacetic acid derivatives were prepared according to the method of Emmott and Livingstone [14] with slight modications, as shown in gure 2. The 5-substitued-salicylaldehydes 24 were alkylated with ethyl bromoacetate in the presence of nely ground potassium carbonate, followed by hydrolysis under the alkaline condition to give 4-substitued 2-formylphenoxyacetic acids 57. The -(1-formyl-2-naphthyloxy)alkanoic acid derivatives 910 were prepared in a similar manner as described above.

1062

Figure 1. Potential aldose reductase inhibitors.

The 3-(1-formyl-2-naphthyloxy)propionic acid 11 was prepared by the alkylation of the aldehyde 8 with -propiolactone in the presence of sodium hydroxide [15]. The 5-(2-carboxymethoxybenzylidene)thiazolidine derivatives 1324 were prepared according to the method of Baranov and Komaritsa [16], as shown in gure 3. Condensation of the 4-substitued-2-formylphenoxyacetic acids 12, 5, 6, or 7 with rhodanine under reuxing, in the presence of 2 eq. anhydrous sodium acetate in acetic acid, gave compound 13, 16, 19, or 22. Moreover, condensation of those with N-methylrhodanine gave compound 14, 17, 20, or 23, and with rhodanine-3-acetic acid likewise gave compound 15, 18, 19, or 24, respectively. The 5-[(2--carboxyalkoxy)naphthylmethylene]thiazolidine derivatives 2530 were prepared in a similar manner as described above.

The conguration of the exocyclic double bond of those thiazolidine derivatives were determined by 1HNMR and 13C-NMR spectroscopy according to the method of Fresneau et al. and Isida et al. [17, 18]. The 2:3 mixture of compounds 28 and 31 were obtained by the photoirradation (uorescent lamp) of 28 in methanol, as shown in gure 4. However, attempted separation of 28 and 31 failed since smooth reisomerization (31 to 28) occured during separation. 1H-NMR and 13 C-NMR experiments on the mixture of compound 28 and 31 disclosed the following results. (1) Although the C5 proton of compound 28 showed a signal at 8.26 due to the anisotropic effect by the carbonyl group (C4=O) of the thiazolidine ring, the same proton signal of 31 appeared at 7.91 in 1H-NMR spectroscopy. (2) Although the coupling constant value of the C4 carbon of compound 28 had J = 6.5 Hz, the same value of 31 had

1063

Reagents: a) K2CO3/toluene, tris(dioxa-3,6-heptyl)amine, BrCH2CO2Et; b) 1 N NaOH/dioxane; c) -propiolactone, NaOH/H2O.

Figure 2. Synthesis of compounds 57, 9, 10 and 11.

J = 12.4 Hz in 13C-NMR spectroscopy. From these results, we conrmed that the exocyclic double bond of compound 28 had Z conguration. Moreover, we postulated that the conguration of other thiazolidine compounds had Z form on the basis of NMR data.

4. Results and discussion

3. Pharmacology The AR inhibitory activity of the synthesized compounds was assessed by measuring the inhibition of the enzymatic activity in a partially puried rat lens preparation [1]. The inhibitory activity was expressed as the concentration (nM) of the test compound which inhibited the activity of AR by 50% (IC50). The in vivo AR inhibitory activity of the test compounds were also evaluated by measuring their ability to inhibit the sorbitol accumulation in the sciatic nerve of STZ-induced diabetic rats [19].

The in vitro AR inhibitory activity of the benzylidene thiazolidine derivatives 1324 are shown in table I. When R1 is hydrogen or a carboxymethyl group, introduction of the substituent R resulted in an increase of AR inhibitory activity by 1 order of the potency as judged from IC50 values (i.e. compound 13 vs. 16, 19 or 22; or compound 15 vs. 18, 21 or 24). When R1 is a methyl group however, introduction of the substituent R (i.e. compound 14 vs. 17, 20 or 23) had no inuence on the AR inhibitory activity. Compounds 14 and 1923, which inhibited the rat AR in vitro at IC50 values of 10 nM order, were equipotent compared with the reference compound Zenarestat. Furthermore, compounds 16, 17 and 18 were 2 times more potent than Zenarestat. However, these compounds were inactive in inhibiting the sorbitol accumulation in the sciatic nerve of STZ-induced diabetic rats at 100 mg/kg p.o. (table II).

1064

Figure 3. Synthesis of thiazolidine derivatives 1324 and 2530.

The in vitro AR inhibitory activity of the naphthylmethylene thiazolidine derivatives 2530 are shown in table III. When R1 is hydrogen, the increase in the number of methylene groups gave a remarkable decrease of the in vitro activity (i.e. compound 25 and 26 vs. 27). When R1 is a methyl group, the increase in the number of

methylene groups gave a moderate decrease of the in vitro activity (i.e. compound 28 vs. 29 vs. 30). Of these compounds, compound 25 was equipotent to Zenarestat, and compound 26 was 3 times more potent than Zenarestat.

Figure 4. Photoirradiation of derivative 28.

1065
Table I. Chemical and biological data, in vitro, of benzylidene thiazolidine derivatives. AR inhibition in vitro Compound 13 14 15 16 17 18 19 20 21 22 23 24 Zenarestat
a

M.p. (C) > 250 220221 > 250 > 250 243246 204208 > 250 237239 209212 226228 215216 213215

Recryst. solvent EtOH-H2O AcOH-H2O EtOH-H2O (CH3)2CO-H2O (CH3)2CO-H2O AcOH-H2O (CH3)2CO-H2O (CH3)2CO-H2O AcOH-H2O MeOH-H2O (CH3)2CO-H2O AcOH-H2O

Formula C12H9NO4S0.2H2O C13H11NO4S2 C14H10NNaO6S2H2O C12H8BrNO4S2 C13H10BrNO4S2 C14H10BrNO6S2 C12H8ClNO4S2 C13H10ClNO4S2 C14H10ClNO6S20.5H2O C13H11NO5S20.25H2O C14H13NO5S2 C15H13NO7S2

rat lens ARa IC50 (nM)b 560 27 170 17 16 18 29 18 21 38 19 86 36

AR: aldose reductase. bThe concentration of test compounds required for 50% inhibition of AR.

The in vivo inhibitory activity of the naphthylmethylene thiazolidine derivatives on the sorbitol accumulation in the sciatic nerve of STZ-induced diabetic rats is shown in table II. When R1 is hydrogen, the increase in the number of methylene groups gave a moderate decrease of the in vivo activity (i.e. compound 25 vs. 26). When R1 is a methyl group, the increase in the number of methylene groups unexpectedly gave a moderate increase of the in vivo activity. Of these, compound 25 had equipotent in vivo activity to Zenarestat. Of particular interest is that the in vivo activity of compound 30 was more potent than that of Zenarestat, although it was 20 times less potent

than Zenarestat in vitro. Therefore, we postulate that compound 30 possessed good oral absorption and efficient tissue penetration properties because of the increased number of methylene groups. In conclusion, we have reported in this article 5-[[2(-carboxyalkoxy)aryl]methylene]-4-oxo-2-thioxothiazolidine derivatives having biological activities which were comparable to the in vitro and in vivo inhibitory activities of Zenarestat. Of these, compound 30 was 1.5 times more potent than Zenarestat in the in vivo activity, and had an adequate potency for clinical development.

Table II. Biological data, in vivo, of benzylidene thiazolidine and naphthylmethylene thiazolidine derivatives. Aldose reductase inhibition Compound 13 16 17 18 25 26 28 29 30 Zenarestat Dose (mg/kg) 100 100 100 100 100 100 30 100 100 100 in vivo % inhibitiona 5.8 3.1 NS NS NS 20.1 4.8 8.0 4.8 11.7 5.8 8.6 7.1 34.6 6.4 21.8 7.2

a Percent inhibition of sorbitol accumulation in the sciatic nerve of streptozotocin-induced diabetic rats. Test compounds were orally given at the single dose indicated. Values are mean SEM; mean of 4-6 rats. NS, no signicant inhibition.

1066
Table III. Chemical and biological data, in vitro, of naphthylmethylene thiazolidine derivatives. AR inhibition in vitro Compound 25 26 27 28 29 30 Zenarestat
a

M.p. (C) 231234 212216 195198 206209 165168 9395

Recryst. solvent AcOH-H2O AcOH-H2O AcOEt-hexane (CH3)2CO-H2O (CH3)2CO-H2O AcOEt-hexane

Formula C16H11NO4S2 C17H13NO4S2 C18H15NO4S2 C17H13NO4S2 C18H15NO4S2 C19H17NO4S20.25H2O

rat lens ARa IC50 (nM)b 33 11 1 100 180 630 700 36

AR: aldose reductase. bThe concentration of test compounds required for 50% inhibition of AR.

5. Experimental protocols 5.1. Chemistry Melting points (m.p.) were determined on a Yanagimoto micro-melting apparatus and are uncorrected. Proton nuclear magnetic resonance (1H-NMR) spectra were obtained on a Varian Gemini-200 spectrometer with tetramethylsilane as an internal standard. Chemical shifts are reported in values from internal tetramethylsilane. Splitting patterns are designated as follows: s, singlet; d, doublet; dd, double doublet; t, triplet; q, quartet; br s, broad singlet; m, multiplet. Coupling constants are reported in hertz (Hz). Infrared (IR) spectra were recorded on a Shimazu FTIR-8200PC spectrophotometer. Elemental analyses were carried out on a Perkin-Elmer 2400 element analyzer and results obtained for specied elements were within 0.4% of the theoretical values. Visualization was accomplished with UV light. Unless otherwise noted, all commercially available materials were used without further purication. 5.1.1. Typical procedure for the preparation of 2formylphenoxy acetic acids 57 5.1.1.1. (4-Chloro-2-formylphenoxy)acetic acid 6 5-Chlorosalicylaldehyde (3; 6.3 g, 40 mmol) was dissolved in dry toluene (90 mL) and nely ground potassium carbonate (5.6 g, 40 mmol) were added. After the mixture was heated at 100 C for 4 h, tris(dioxa-3,6heptyl)amine (1.3 mL, 4.1 mmol) and ethyl bromoacetate (5.6 mL, 50 mmol) were added. The mixture was heated at 100 C for 6 h, ltered through Celite and washed with toluene. The ltrate was washed with water, and brine, dried (anhydrous MgSO4), and evaporated. The residue was dissolved in dioxane (50 mL), and 1 N NaOH (50 mL) was added. The mixture was heated at 100 C for

1 h. After cooling to room temperature, the reaction mixture was diluted with water (150 mL) and then acidied with concentrated HCl on an ice bath, and the precipitated solid was collected by ltration. The solid was recrystallized from acetone/H2O to give the title compound as a solid (5.3 g, 62%). M.p. 174175 C (lit. 173174 C) [20]. The following compounds were prepared in the same way as described above: (4-Bromo-2-formylphenoxy)acetic acid 5 (prepared from 5-bromosalicylaldehyde 2; recrystallized from acetone/H2O). M.p. 172174 C (lit. 174176 C) [20]. (2-Formyl-4-methoxyphenoxy)acetic acid 7 (prepared from 5-methoxysalicylaldehyde 4; recrystallized from acetone/H2O). M.p. 156159 C (lit. 157159 C) [20].

5.1.2. (1-Formyl-2-naphthyloxy)acetic acid 9 2-Hydroxy-1-naphthylaldehyde (8; 32 g, 0.2 mol) was dissolved in dry toluene (200 mL) and nely ground potassium carbonate (15 g, 0.11 mol) was added. After the mixture was heated at 100 C for 4 h, tris(dioxa-3,6heptyl)amine (4 mL, 12.5 mmol) and ethyl bromoacetate (28 mL, 0.25 mmol) were added. The mixture was heated at 100 C for 4 h, ltered through Celite and washed with toluene. The ltrate was washed with water and brine, dried (anhydrous MgSO4), and evaporated. The residue was dissolved in dioxane (250 mL) and 1 N NaOH (250 mL) was added. The mixture was heated at 100 C for 1 h. After cooling to room temperature, the reaction mixture was diluted with water (700 mL) and then acidied with concentrated HCl on an ice bath, and the precipitated solid was collected by ltration. The solid was recrystallized from AcOEt/n-hexane to give the title compound as a solid (15 g, 65%): M.p. 176179 C (lit. 176177 C) [14].

1067 5.1.3. 4-(1-Formyl-2-naphthyloxy)butyric acid 10 Compound 10 was synthesized in the same way as in 5.1.2. Yield 34%. M.p. 179180 C (recrystallized from AcOEt/n-hexane); 1H-NMR (DMSO-d6) d 2.002.18 (2H, m), 2.47 (2H, t, J = 7.5 Hz), 4.35 (2H, t, J = 7.5 Hz), 7.407.70 (3H, m), 7.908.00 (1H, m), 8.27 (1H, d, J = 9 Hz), 9.079.17 (1H, m), 10.7 (1H, s), 12.18 (1H, br s); IR (KBr, cm1) 1 705 (CO), 1 672 (CO). Anal. calcd. for C15H14O4: C 69.76; H 5.46; found: C 69.65; H 5.40. 5.1.4. 3-(1-Formyl-2-naphthyloxy)propionic acid 11 A solution of -propiolactone (7.6 g, 0.1 mol) in water (10 mL) was added dropwise to a stirred solution of 2-hydroxy-1-naphthylaldehyde (8; 17.2 g, 0.1 mol) and NaOH (4 g, 0.1 mol) in water (50 mL) at 100 C, and the mixture was stirred for a further 30 min at 100 C. After cooling to room temperature, the reaction mixture was acidied with concentrated HCl on an ice bath, and then extracted three times with diethyl ether. The combined ether extracts were shaken three times with saturated NaHCO3. The solid, which was precipitated on acidication of the alkaline solution, was ltered, washed with water, dried, and recrystallized from AcOEt/n-hexane to give the title compound as a solid (4.2 g, 17%). M.p. 157159 C (recrystallized from AcOEt/n-hexane); 1HNMR (DMSO-d6) d 2.84 (2H, t, J = 6 Hz), 4.52 (2H, t, J = 6 Hz), 7.427.52 (1H, m), 7.587.72 (2H, m), 7.928.02 (1H, m), 8.30 (1H, d, J = 9 Hz), 9.079.17 (1H, m), 10.72 (1H, s), 12.48 (1H, br s); IR (KBr, cm1) 1 705 (CO), 1 672 (CO). Anal. calcd. for C14H12O4: C 68.85; H 4.95; found: C 68.74; H 4.88. 5.1.5. 5-(5-Bromo-2-carboxymethoxybenzylidene)-4-oxo2-thioxothiazolidine 16 A mixture of 4-Bromo-2-formylphenoxyacetic acid (5; 6.0 g, 23.2 mmol), rhodanine (3.7 g, 27.8 mmol) and anhydrous sodium acetate (3.8 g, 46.3 mmol) in acetic acid (80 mL) was reuxed for 17 h. Upon cooling, the reaction mixture was diluted with water and stirred at room temperature for a further 1 h. The solid was ltrated, and added to dilute HCl, then the mixture was stirred at room temperature for 1 h. After ltration, the solid was recrystallized from acetone/H2O to give the title compound as a yellow solid (5.9 g, 68%). M.p. > 250 C; 1H-NMR (DMSO-d6) d 4.90 (2H, S), 7.06 (1H, d, J = 9 Hz), 7.49 (1H, d, J = 2.5 Hz), 7.63 (1H, dd, J = 2.5, 9 Hz), 7.76 (1H, s), 13.20 (1H, br s), 13.88 (1H, br s); IR (KBr, cm1) 1 714 (CO). Anal. calcd. for C12H8BrNO4S2: C 38.51; H 2.15; N 3.74; S 17.14; Br 21.35; found: C 38.57; H 1.92; N 3.73; S 17.29; Br; 21.50. 5.1.6. 5-(2-Carboxymethoxybenzylidene)-4-oxo-2-thioxothiazolidine 13 Compound 13 was synthesized in the same way as in 5.1.5. Yield 32%. M.p. > 250 C (recrystallized from aq. EtOH); 1H-NMR (DMSO-d6) d 4.88 (2H, S), 7.037.18 (2H, m), 7.387.52 (2H, m), 7.90 (1H, s), 13.20 (1H, br s), 13.74 (1H, br s); IR (KBr, cm1) 1 705 (CO), 1 672 (CO). Anal. calcd. for C12H9NO4S20.2H2O: C 48.21; H 3.17; N 3.74; S 21.45; found: C 48.49; H 2.93; N 4.74; S 21.27. 5.1.7. 5-(2-Carboxymethoxybenzylidene)-3-methyl-4oxo-2-thioxothiazolidine 14 Compound 14 was synthesized in the same way as in 5.1.5. Yield 77%. M.p. 220221 C (recrystallized from AcOH/H2O); 1H-NMR (DMSO-d6) d 3.43 (3H, s), 4.90 (2H, s), 7.047.20 (2H, m), 7.437.55 (2H, m), 8.07 (1H, s), 13.18 (1H, br s); IR (KBr, cm1) 1 753 (CO), 1 689 (CO). Anal. calcd. for C13H11NO4S2: C 50.47; H 3.58; N 4.53; S 20.73; found: C 50.17; H 3.56; N 4.62; S 21.00. 5.1.8. [5-(2-Carboxymethoxybenzylidene)-4-oxo-2-thioxothiazolidin-3-yl]-acetic acid mono sodium salt 15 Compound 15 was synthesized in the same way as in 5.1.5. Yield 11%. M.p. > 250 C (recrystallized from aq. EtOH); 1H-NMR (DMSO-d6) d 4.50 (2H, s), 4.69 (2H, s), 6.957.15 (2H, m), 7.417.51 (2H, m), 8.06 (1H, s); IR (KBr, cm1) 1 716 (CO). Anal. calcd. for C14H10NNaO6S2H2O: C 42.75; H 3.07; N 3.56; S 16.30; Na 5.84; found: C 42.82; H 3.07; N 3.59; S 16.01; Na 5.54. 5.1.9. 5-(5-Bromo-2-carboxymethoxybenzylidene)-3methyl-4-oxo-2-thioxothiazolidine 17 Compound 17 was synthesized in the same way as in 5.1.5. Yield 67%. M.p. 243246 C (recrystallized from acetone/H2O); 1H-NMR (DMSO-d6) d 3.41 (3H, s), 4.93 (2H, s), 7.07 (1H, d, J = 9 Hz), 7.53 (1H, d, J = 2.5 Hz), 7.65 (1H, dd, J = 2.5, 9 Hz) 7.91 (1H, s), 13.24 (1H, br s); IR (KBr, cm1) 1 705 (CO). Anal. calcd. for C13H10BrNO4S2: C 40.22; H 2.60; N 3.61; S 16.52; Br 20.58; found: C 40.15; H 2.61; N 3.49; S 16.58; Br 20.55. 5.1.10. [5-(5-Bromo-2-carboxymethoxybenzylidene)-4oxo-2-thioxo-thiazolidin-3-yl]-acetic acid 18 Compound 18 was synthesized in the same way as in 5.1.5. Yield 55%. M.p. 204208 C (recrystallized from AcOH/H2O); 1H-NMR (DMSO-d6) d 4.75 (2H, s), 4.95 (2H, s), 7.09 (1H, d, J = 9 Hz), 7.60 (1H, d, J = 2.5 Hz), 7.96 (1H, dd, J = 2.5, 9 Hz) 7.96 (1H, s), 13.36 (2H, br s); IR (KBr, cm1) 1 712 (CO). Anal. calcd. for C14H10BrNO6S2: C 38.90; H 2.33; N 3.24; S 14.84; Br 18.49; found: C 38.73; H 2.39; N 3.15; S 14.59; Br 18.26.

1068 5.1.11. 5-(5-Chloro-2-carboxymethoxybenzylidene)-4oxo-2-thioxothiazolidine 19 Compound 19 was synthesized in the same way as in 5.1.5. Yield 62%. M.p. > 250 C (recrystallized from acetone/H2O); 1H-NMR (DMSO-d6) d 4.90 (2H, s), 7.14 (1H, d, J = 9 Hz), 7.44 (1H, d, J = 2.5 Hz), 7.52 (1H, dd, J = 2.5, 9 Hz) 7.84 (1H, s), 13.24 (1H, br s), 13.88 (1H, br s); IR (KBr, cm1) 1 733 (CO), 1 683 (CO). Anal. calcd. for C12H8ClNO4S2: C 43.71; H 2.45; N 4.25; S 19.45; Cl 10.75; found: C 43.79; H 2.42; N 4.19; S 19.12; Cl 10.75. 5.1.12. 5-(5-Chloro-2-carboxymethoxybenzylidene)-3methyl-4-oxo-2-thioxothiazolidine 20 Compound 20 was synthesized in the same way as in 5.1.5. Yield 58%. M.p. 237239 C (recrystallized from acetone/H2O); 1H-NMR (DMSO-d6) d 3.43 (3H, s), 4.93 (2H, s), 7.14 (1H, d, J = 9 Hz), 7.42 (1H, d, J = 2.5 Hz), 7.54 (1H, dd, J = 2.5, 9 Hz) 7.93 (1H, s), 13.24 (1H, br s); IR (KBr, cm1) 1 705 (CO). Anal. calcd. for C13H10ClNO4S2: C 45.42; H 2.93; N 4.07; S 18.65; Cl 10.31; found: C 45.43; H 2.89; N 4.02; S 18.92; Cl 10.11. 5.1.13. [5-(5-Chloro-2-carboxymethoxybenzylidene)-4oxo-2-thioxothiazolidin-3-yl]-acetic acid 21 Compound 21 was synthesized in the same way as in 5.1.5. Yield 64%. M.p. 209212 C (recrystallized from AcOH/H2O); 1H-NMR (DMSO-d6) d 4.75 (2H, s), 4.95 (2H, s), 7.15 (1H, d, J = 9 Hz), 7.49 (1H, d, J = 2.5 Hz), 7.55 (1H, dd, J = 2.5, 9 Hz) 7.97 (1H, s), 13.36 (1H, br s); IR (KBr, cm1) 1 716 (CO). Anal. calcd. for C14H10ClNO6S20.5H2O: C 42.38; H 2.79; N 3.53; S 16.16; Cl 8.93; found: C 42.66; H 2.65; N 3.45; S 15.83; Cl 8.76. 5.1.14. 5-(2-Carboxymethoxy-5-methoxybenzylidene)-4oxo-2-thioxothiazolidine 22 Compound 22 was synthesized in the same way as in 5.1.5. Yield 8%. M.p. 226228 C (recrystallized from MeOH/H2O); 1H-NMR (DMSO-d6) d 3.78 (3H, s), 4.83 (2H, s), 6.88 (1H, d, J = 2.5 Hz), 7.03 (1H, d, J = 9 Hz), 7.08 (1H, dd, J = 2.5, 9 Hz), 7.87 (1H, s), 13.10 (1H, br s), 13.84 (1H, br s); IR (KBr, cm1) 1 751 (CO), 1 695 (CO). Anal. calcd. for C13H11NO5S20.25H2O: C 47.34; H 3.51; N 4.25; S 19.44; found: C 47.39; H 3.19; N 4.38; S 19.22. 5.1.15. 5-(2-Carboxymethoxy-5-methoxybenzylidene)-4oxo-2-thioxothiazolidine 23 Compound 23 was synthesized in the same way as in 5.1.5. Yield 35%. M.p. 215216 C (recrystallized from acetone/H2O); 1H-NMR (DMSO-d6) d 3.43 (3H, s), 3.80 (3H, s), 4.83 (2H, s), 6.92 (1H, d, J = 2.5 Hz), 7.05 (1H, d, J = 9 Hz), 7.10 (1H, dd, J = 2.5, 9 Hz) 8.03 (1H, s), 13.12 (1H, br s); IR (KBr, cm1) 1 712 (CO). Anal. calcd. for C14H13NO5S2: C 49.55; H 3.86; N 4.13; S 18.90; found: C 49.36; H 3.59; N 4.03; S 18.73. 5.1.16. [5-(2-Carboxymethoxy-5-methoxybenzylidene)-4oxo-2-thioxo-thiazolidin-3-yl]-acetic acid 24 Compound 24 was synthesized in the same way as in 5.1.5. Yield 40%. M.p. 213215 C (recrystallized from acetone/H2O); 1H-NMR (DMSO-d6) d 3.80 (3H, s), 4.75 (2H, s), 4.84 (2H, s), 6.97 (1H, d, J = 2.5 Hz), 7.06 (1H, d, J = 9 Hz), 7.12 (1H, dd, J = 2.5, 9 Hz), 8.07 (1H, s), 13.30 (2H, br s); IR (KBr, cm1) 1 716 (CO). Anal. calcd. for C15H13NO7S2: C 46.99; H 3.42; N 3.65; S 16.73; found: C 47.09; H 3.26; N 3.75; S 16.67. 5.1.17. 5-(2-Carboxymethoxy-1-naphthylmethylene)-4oxo-2-thioxothiazolidine 25 Compound 25 was synthesized in the same way as in 5.1.5. Yield 50%. M.p. 231234 C; 1H-NMR (DMSOd6) d 5.03 (2H, s), 7.40 (1H, d, J = 9 Hz), 7.437.53 (1H, m), 7.577.67 (1H, m), 7.827.90 (1H, m), 7.928.00 (1H, m), 8.08 (1H, d, J = 9 Hz), 8.09 (1H, s), 13.20 (1H, br s), 13.70 (1H, br s); IR (KBr, cm1) 1 712 (CO), 1 662 (CO). Anal. calcd. for C16H11NO4S2: C 55.64; H 3.21; N 4.06; S 18.57; found: C 55.74; H 3.49; N 3.97; S 18.49. 5.1.18. 5-[2-(2-carboxyethoxy)-1-naphthylmethylene]-4oxo-2-thioxothiazolidine 26 Compound 26 was synthesized in the same way as in 5.1.5. Yield 31%. M.p. 212216 C (recrystallized from AcOH/H2O); 1H-NMR (DMSO-d6) d 2.78 (2H, t, J = 6 Hz), 4.48 (2H, t, J = 6 Hz), 7.427.66 (2H, m), 7.57 (1H, d, J = 9 Hz), 7.787.86 (1H, m), 7.948.02 (1H, m), 8.00 (1H, s), 8.11 (1H, d, J = 9 Hz), 12.38 (1H, br s), 13.66 (1H, br s); IR (KBr, cm1) 1 716 (CO), 1 683 (CO). Anal. calcd. for C17H13NO4S2: C 56.81; H 3.65; N 3.90; S 17.84; found: C 56.55; H 3.82; N 3.71; S 17.49. 5.1.19. 5-[2-(3-carboxypropoxy)-1-naphthylmethylene]4-oxo-2-thioxothiazolidine 27 Compound 27 was synthesized in the same way as in 5.1.5. Yield 58%. M.p. 195198 C (recrystallized from AcOEt/n-hexane); 1H-NMR (DMSO-d6) d 1.932.10 (2H, m), 2.40 (2H, t, J = 7.5 Hz), 4.31 (2H, t, J = 7.5 Hz), 7.417.66 (2H, m), 7.55 (1H, d, J = 9 Hz) 7.797.87 (1H, m), 7.938.00 (1H, m), 8.05 (1H, s), 8.10 (1H, d, J = 9 Hz), 12.15 (1H, br s), 13.68 (1H, br s); IR (KBr, cm1) 1 705 (CO). Anal. calcd. for C18H15NO4S2: C 57.89; H 4.05; N 3.75; S 17.17; found: C 57.94; H 4.11; N 3.71; S 16.88.

1069 5.1.20. 5-(2-carboxymethoxy-1-naphthylmethylene)-3methyl-4-oxo-2-thioxothiazolidine 28 Compound 28 was synthesized in the same way as in 5.1.5. Yield 60%. M.p. 206209 C; 1H-NMR (DMSOd6) d 3.43 (3H, s), 5.02 (2H, s), 7.42 (1H, d, J = 9 Hz), 7.437.53 (1H, m), 7.577.67 (1H, m), 7.827.91 (1H, m), 7.938.01 (1H, m), 8.10 (1H, d, J = 9 Hz), 8.26 (1H, s), 13.22 (1H, br s); IR (KBr, cm1) 1 757 (CO), 1 668 (CO). Anal. calcd. for C17H13NO4S2: C 56.81; H 3.65; N 3.90; S 17.84; found: C 56.67; H 3.54; N 3.76; S 18.14. 5.1.21. 5-[2-(2-carboxyethoxy)-1-naphthylmethylene]-3methyl-4-oxo-2-thioxothiazolidine 29 Compound 29 was synthesized in the same way as in 5.1.5. Yield 28%. M.p. 165168 C (recrystallized from acetone/H2O); 1H-NMR (DMSO-d6) d 2.78 (2H, t, J = 6 Hz), 3.43 (3H, s), 4.48 (2H, t, J = 6 Hz), 7.437.66 (2H, m), 7.59 (1H, d, J = 9 Hz), 7.797.87 (1H, m), 7.948.02 (1H, m), 8.13 (1H, d, J = 9 Hz), 8.18 (1H, s), 12.38 (1H, br s); IR (KBr, cm1) 1 716 (CO). Anal. calcd. for C18H15NO4S2: C 57.89; H 4.05; N 3.75; S 17.17; found: C 57.77; H 4.21; N 3.56; S 16.70. 5.1.22. 5-[2-(3-carboxypropoxy)-1-naphthylmethylene]3-methyl-4-oxo-2-thioxothiazolidine 30 Compound 30 was synthesized in the same way as in 5.1.5. Yield 43%. M.p. 9395 C (recrystallized from AcOEt/n-hexane); 1H-NMR (DMSO-d6) d 1.802.20 (2H, m), 2.40 (2H, t, J = 7.5 Hz), 3.44 (3H, s), 4.30 (2H, t, J = 7.5 Hz), 7.427.66 (2H, m), 7.56 (1H, d, J = 9 Hz), 7.807.88 (1H, m), 7.948.01 (1H, m), 8.12 (1H, d, J = 9 Hz), 8.23 (1H, s), 12.14 (1H, br s); IR (KBr, cm1) 1 716 (CO), 1 695 (CO). Anal. calcd. for C19H7NO4S2 0.25H2O: C 58.22; H 4.50; N 3.57; S 16.36; found: C 58.31; H 4.35; N 3.60; S 16.67. 5.2. Pharmacology 5.2.1. Preparation of aldose reductase Rat lens AR was prepared according to the method of Hayman and Kinoshita [1] with slight modications. The lenses were homogenized in 250 mM phosphate buffer (pH 7.4) containing 2 mM mercaptoethanol at 04 C, and the homogenate was centrifuged at 20 000 g for 30 min. The supernatant was subjected to a 4060% ammonium sulfate fractionation. The resultant precipitate was dissolved in 5 mM phosphate buffer (pH 7.4) containing 2 mM mercaptoethanol and used for enzyme assay. One unit (U) of AR was dened as the enzyme activity which oxidizes 1 mol of NADPH in 1 min under the assay conditions described below. 5.2.2. Inhibition of aldose reductase in vitro The reaction mixture consisted of 100 mM phosphate buffer (pH 6.5), 0.2 mM NADPH, 1.5 mM D, L-glyceraldehyde, 0.4 M lithium sulfate, 7.0 mU/mL of enzyme and test compounds at various concentrations. The reaction mixture was incubated at 37 C, and the absorbance at 340 nm was measured with a spectrophotometer (Model 150-20, Hitachi Ltd., Japan). The enzyme activity was estimated on the basis of its decrease in the absorbance over a period of 1 min. The concentration of compounds required for 50% inhibition of enzyme activity (IC50) was estimated graphically from the log concentration-inhibition curve. 5.2.3. Inhibition of sorbitol accumulation in vivo Male Wistar rats (200250 g) were rendered diabetic by an intravenous injection of streptozotocin (40 mg/kg), which had been freshly dissolved in physiological saline. After 7 days, the rats were divided into various groups with 46 animals/group, and orally given test compounds suspended in 0.5% tragacanth, or an equivalent volume of 0.5% tragacanth. The rats were sacriced 4 h after the administration of the test compounds. Tissue sorbitol levels were determined according to the method of Clements et al. [17] with slight modications. The sciatic nerve sample (3060 mg) was quickly dissected from the hind limb, placed into water (1.0 mL/ 40 mg of tissue), heated in a boiling bath for 2 min, and then homogenized with a Polytron instrument in 6% perchloric acid (1 mL/10 mg of tissue). The homogenate was centrifuged at 1 050 g for 15 min at 4 C. The supernatant was neutralized with 2 M K2CO3 and used as tissue extract for the assaying of sorbitol. Sorbitol was assayed by an enzymatic method in which sorbitol dehydrogenase catalyses the stoichiometric conversion of NAD by sorbitol to a uorogenic product, NADH. The reaction mixture contained 30 mM glycine buffer (pH 9.4), 1.3 mM NAD, 1.3 U/mL sorbitol dehydrogenase and 1.0 mL of the tissue extract in a total volume of 3 mL. After the mixture was allowed to stand for 60 min at room temperature, the uorescence intensity was measured at 365 nm excitation wavelength and 430 nm emission wavelength using a uorospectrophotometer (F3000, Hitachi Ltd., Japan). The sorbitol concentration was quantitated by comparison with standards of sorbitol. The sorbitol content in the sciatic nerve of each animal was expressed as nmole/wet weight. The activity of test compounds was expressed as the percent inhibition of sorbitol accumulation at a given dose, which was calculated according to the following equation: % inhibition = (S T) / (S N) 100

1070 where S is the sorbitol content in the sciatic nerve of untreated diabetic control rats, T is the sorbitol content in the sciatic nerve of diabetic rats given test compounds and N is the sorbitol content in the sciatic nerve of age-matched non-diabetic control rats. Acknowledgements We thank Dr S. Naruto and Dr T. Kadokawa for encouragement throughout this work, and members of the analytical section for elemental analyses and spectral measurements. We also thank Ms. M. Furuichi for measurements of pharmacological studies.
[6] [7] [8] Varma S.D., Schocket S.S., Richards R.D., Invest. Ophthalmol. Vis. Sci. 18 (1979) 237241. Canal N., Comi G., Trends Pharmacol. Sci. 6 (1985) 328330. Sestanj K., Bellini F., Fung S., Abraham N., Treasurywala A., Humber L., Simard-Duquesne N., Dvornik D., J. Med. Chem. 27 (1984) 255256. Kikkawa R., Hatanaka I., Yasuda H., Kobayashi N., Shigeta Y., Terashima H., Morimura T., Tsuboshima M., Diabetlogia 24 (1983) 290292. Mizuno K., Kato N., Matsubara A., Nakano K., Kurono M., Metabolism 41 (1992) 10811086. Ao S., Shingu Y., Kikuchi C., Takano Y., Nomura K., Fujiwara T. et al., Metabolism 40 (1991) 7787. Kador P.F., Kinoshita J.H., Sharpless N.E., J. Med. Chem. 28 (1985) 841849. Lee Y.S., Pearlstein R., Kador P.F., J. Med. Chem. 37 (1994) 787792. Emmott P., Livingstone R., J. Chem. Soc. (1957) 31443148. Gresham T.L., Jansen J.E., Shaver F.W., Bankert R.A., Beears W.L., Prendergast M.G., J. Am. Chem. Soc. 71 (1949) 661663. Baranov S.N., Komaritsa I.D., Khim. Geterotsikl. Soedin. Akad. Nauk Latv. SSR 1 (1965) 6973. Fresneau P., Cussac M., Morand J.M., Szymonski B., Tranqui D., Leclerc G., J. Med. Chem. 41 (1998) 47064715. Ishida T., In Y., Inoue M., Ueno Y., Tanaka C., Tetrahedron Lett. 30 (1989) 959962. Clements R.S., Morisson Jr. A.D., Winegrad A.I., Science 166 (1969) 10071008. Hullar T.L., Failla D.L., J. Med. Chem. 12 (1969) 420424.

[9]

[10] [11] [12] [13] [14] [15]

References
[1] [2] [3] [4] [5] Hayman S., Kinoshita J.H., J. Biol. Chem. 240 (1965) 877882. Kador P.F., Akagi Y., Kinoshita J.H., Metabolism 35 (1986) (suppl. 1) 1519. Gabbay K.H., New Engl. J. Med. 288 (1973) 831836. Akagi Y., Kador P.F., Kuwabara T., Kinoshita J.H., Invest. Ophthalmol. Vis. Sci. 24 (1983) 15161519. Cohen M.P., Metabolism 35 (1986) (suppl. 1) 5559.

[16] [17] [18] [19] [20]

Eur. J. Med. Chem. 34 (1999) 10711076 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1071

Short communication

Synthesis and antimycobacterial activity of some isonicotinoylhydrazones

Maria T. Coccoa*, Cenzo Congiua, Valentina Onnisa, Maria C. Pusceddub, Maria L. Schivob, Alessandro De Logub
b a Dipartimento di Tossicologia, Universit di Cagliari, Viale A. Diaz 182 - 09126 Cagliari, Italy Dipartimento di Scienze Chirurgiche e Trapianti dOrgano, Sezione di Microbiologia e Virologia, Universit di Cagliari, Via Palabanda 14 - 09123 Cagliari, Italy

(Received 15 February 1999; accepted 8 June 1999)

Abstract A series of isonicotinoylhydrazones 2 were prepared by addition of some aryloxyacetonitriles with isonicotinoylhydrazine in basic medium. These compounds have been further reacted with pyridinecarboxaldehydes to give the corresponding pyridylmethyleneamino derivatives 35. The new synthesized hydrazones and their pyridylmethyleneamino derivatives were tested for their activity against mycobacteria, Gram-positive and Gram-negative bacteria. The cytotoxicity was also tested. Several compounds showed a good activity against Mycobacterium tuberculosis H37Rv and some isonycotinoylhydrazones 2 showed a moderate activity against a clinically isolated M. tuberculosis which was isoniazid resistant. 1999 ditions scientiques et mdicales Elsevier SAS aminohydrazone derivatives / synthesis / antimycobacterial activity

1. Introduction At present, tuberculosis is considered, by the World Health Organisation, to be the most important chronic communicable disease in the world [1, 2, 3]. Over the past decade, tuberculosis has re-emerged both in industrial and developing countries. The emergence of AIDS, decline of socioeconomic standards and a reduced emphasis on tuberculosis control programs contribute to the diseases resurgence in industrialised countries [4]. In most developing countries, although the disease has always been endemic, its severity has increased because of the global HIV endemic and extensive social restructuring due to rapid industrialisation and conicts. Further contributing to the increased morbidity is the emergence of new strains of M. tuberculosis resistant to some or all current antitubercular drugs [5, 6]. Among these, isoniazid (INH), an inexpensive and relatively safe drug, continues to be well established for the treatment of tuberculosis. The mechanism of action of INH, as well as the mechanism conferring INH resistance, are complex and not completely understood. However, several studies
*Correspondence and reprints

suggest that INH inhibits the biosynthesis of cell wall mycolic acids, thereby making the mycobacteria susceptible to reactive oxygen radicals and other environmental factors [7]. INH is active against the Mycobacterium complex (Mycobacterium tuberculosis, M. bovis, M. africanum and M. microti) at MICs ranging from 0.0250.05 g/mL [8], while at a higher concentration (MIC of 500 g/mL) it inhibits the growth of other microorganisms such as the opportunists Klebsiella, Serratia and Enterobacter. For these reasons the antimycobacterial pharmacophore moiety of INH is introduced in several molecules to improve their activity against Mycobacteria. On the other hand aminohydrazone derivatives, structurally correlated to INH, have been described for their in vitro antimycobacterial activity and some of these compounds exhibited inhibitory activity toward a human strain of M. avium resistant to the primary drugs INH, rifampicin and ooxacin [9, 10]. As a part of our studies on aminohydrazone derivatives [11, 12], we became interested in a new series of 2-amino-2-(isonicotinoylhydrazono)ethyl aryl ethers, assuming that the isonicotine hydrazonic moiety is an important pharmacophore to antimycobacterial activity. Here, we report the

1072
Table I. Chemical and spectral data of compounds 2.
Compound 2a 2b 2c 2d 2e 2f 2g 2h 2i 2j 2k 2l R H 2-CH3 3-CH3 4-CH3 2-OCH3 3-OCH3 4-OCH3 2-Cl 3-Cl 4-Cl 4-NO2 Yield (%) 77 74 68 66 72 99 68 68 64 99 77 M.p. (C) IR (Nujol) (recryst. solvent) (cm1) 174176 (CH3CN) 178180 (EtOH) 162164 (2-PrOH) 208210 (CH3CN) 175177 (EtOH) 158160 (CH3CN) 193195 (CH3CN) 188190 (EtOH) 180182 (EtOH) 236238 (CH3CN) 228230 (EtOH) 4-NHCOCH3 73 191193 (EtOH) 3 410, 3 020, 1 590, 3 270, 1 660, 3 170, 1 680, 3 270, 1 680, 3 260, 3 020, 3 270, 1 675, 3 360, 1 590, 3 410, 1 690, 3 410, 1 680, 3 240, 1 680, 3 360, 3 160, 1 680, 3 220, 1 660, 3 280, 1 680, 1 570 3 310, 1 605 3 060, 1 650 3 080, 1 650 3 100, 1 670 3 120, 1 635 1 675, 1 570 3 320, 1 620 3 080, 1 585 3 080, 1 640 3 280, 3 080, 1 640 3 060, 1 595
1

H-NMR (DMSO-d6/TMS) (ppm)

4.50 (s, 2H, CH2), 6.63 (s, 2H, NH2), 6.977.24 (m, 5H, Ar ), 7.72, 8.63 (m, 4H, Py), 10.11 (br s, 1H, NH) 2.17 (s, 3H, CH3), 4.51 (s, 2H, CH2), 6.59 (s, 2H, NH2), 6.827.10 (m, 4H, Ar), 7.73, 8.63 (m, 4H, Py), 10.13 (br s, 1H, NH) 2.23 (s, 3H, CH3), 4.49 (s, 2H, CH2), 6.61 (s, 2H, NH2), 6.757.13 (m, 4H, Ar), 7.73, 8.60 (m, 4H, Py), 10.11 (br s, 1H, NH) 2.17 (s, 3H, CH3), 4.46 (s, 2H, CH2), 6.60 (s, 2H, NH2), 6.727.06 (m, 4H, Ar), 7.72, 8.62 (m, 4H, Py), 10.09 (br s, 1H, NH) 3.73 (s, 3H, CH3), 4.47 (s, 2H, CH2), 6.586.94 (m, 6H, Ar + NH2), 7.72, 8.63 (m, 4H, Py) 10.11 (s, 1H, NH) 3.68 (s, 3H, CH3), 4.49 (s, 2H, CH2), 6.497.18 (m, 6H, Ar + NH2), 7.72, 8.61 (m, 4H, Py), 10.10 (brs, 1H,NH) 3.64 (s, 3H, CH3), 4.44 (s, 2H, CH2), 6.60 (s, 2H, NH2), 6.806.94 (m, 4H, Ar), 7.72, 8.62 (m, 4H, Py), 10.08 (br s, 1H, NH) 4.58 (s, 2H, CH2), 6.62 (s, 2H, NH2), 6.947.38 (m, 4H, Ar), 7.72, 8.64 (m, 4H, Py), 10.16 (br s, 1H, NH) 4.54 (s, 2H, CH2), 6.64 (s, 2H, NH2), 6.957.31 (m, 4H, Ar), 7.72, 8.63 (m, 4H, Py), 10.12 (br s, 1H, NH) 4.51 (s, 2H, CH2), 6.64 (s, 2H, NH2), 7.017.31 (m, 4H, Ar), 7.72, 8.63 (m, 4H, Py), 10.11 (br s, 1H, NH) 4.70 (s, 2H, CH2), 6.74 (s, 2H, NH2), 7.187.73 (m, 4H, Ar), 8.20, 8.63 (m, 4H, Py), 10.16 (br s, 1H, NH) 1.95 (s, 3H, CH3), 4.46 (s, 2H, CH2), 6.61 (s, 2H, NH2), 6.907.42 (m, 4H, Ar), 7.72, 8.62 (m, 4H, Py), 9.77 (s, 1H, NH), 10.09 (br s, 1H, NH).

synthesis of these aminohydrazone derivatives and the evaluation of their activity against mycobacteria. Activity against Gram-positive and Gram-negative bacteria as well as their cytotoxicity were also evaluated.

3. Microbiology Antimycobacterial activity was investigated against M. tuberculosis H37Rv ATCC 25584, M. avium ATCC 19421 and M. Fortuitum ATCC 9820, and M. tuberculosis resistant to isoniazid (INH-R) isolated from a patient with an active clinical infection treated at the Universit di Cagliari. The antimicrobial activity of compounds 25 was also evaluated against Gram-positive and Gramnegative bacteria isolated from clinical specimens and Candida albicans ATCC E10931. At the same time, cell cytotoxicity of all compounds was tested in vitro on Vero cells.

2. Chemistry Isonicotinoylhydrazones 2 and their pyridylmethyleneamino derivatives 3, 4 and 5 described in this study are shown in tables I and II, and a reaction sequence for their preparation is outlined in gure 1. The starting aryloxyacetonitriles 1 were prepared according to known procedures [13]. Aryloxyacetonitriles 1 were converted to hydrazones 2 by addition of isonicotinoylhydrazine in the presence of catalytic amounts of sodium ethoxide in anhydrous ethanolic solution. Heating compounds 2 with pyridinecarboxaldehyde in ethanolic solution in the presence of piperidine afforded pyridylmethyleneamino derivatives 35. The structures for all new obtained compounds 25 were determined by examining their IR and 1H-NMR spectra as well as by elemental analyses (tables I and II).

4. Results and discussion With the exception of compound 2f which showed toxicity at concentrations higher than 62.5 g/mL, the tested compounds exhibited high values of maximum non-toxic dose (MNTD) on Vero cells, ranging from 250 g/mL for 4j, to 5001 000 g/mL for the other members of the series.

1073
Table II. Chemical and spectral data of compounds 3, 4 and 5.
Compound 3a 4a 5a 5b 4c 5c 4d 5d 4e 5e 4f 5f 4g 5g 4h 5h 4i 5i 4j 5j 4k 5k R H H H 2-CH3 3-CH3 3-CH3 4-CH3 4-CH3 2-OCH3 2-OCH3 3-OCH3 3-OCH3 4-OCH3 4-OCH3 2-Cl 2-Cl 3-Cl 3-Cl 4-Cl 4-Cl 4-NO2 4-NO2 Pyr Yield (%) M.p. (C) IR (Nujol) (recryst. solvent) (cm1) 168169 (Benzene) 164165 (2-PrOH) 4-pyridyl 63 4-pyridyl 48 3-pyridyl 57 4-pyridyl 59 3-pyridyl 64 4-pyridyl 56 3-pyridyl 68 4-pyridyl 67 3-pyridyl 50 4-pyridyl 52 3-pyridyl 72 4-pyridyl 60 3-pyridyl 60 4-pyridyl 50 3-pyridyl 51 4-pyridyl 67 3-pyridyl 20 4-pyridyl 51 3-pyridyl 65 4-pyridyl 62 190191 (Benzene) 162163 (Toluene) 124125 (Toluene) 174175 (Toluene) 152153 (Toluene) 159160 (Toluene) 134135 (Toluene) 129130 (Toluene) 137138 (Toluene) 157158 (Toluene) 123124 (Toluene) 129130 (Toluene) 173174 (Toluene) 184185 (Toluene) 139140 (Toluene) 139140 (Toluene) 164165 (Ethyl acetate) 159160 (Toluene) 189190 (2-PrOH) 184185 (2-PrOH) 3 050, 1 590, 3 180, 1 615, 1 550 3 070, 1 595 3 360, 1 620, 3 060, 1 540 3 060, 1 540 3 020, 1 600, 3 060, 1 560 3 080, 1 630, 1 630, 3 160, 1 640, 3 140, 1 620, 3 150, 1 620, 3 100, 1 620, 3 350, 1 610, 3 340, 1 620, 1 540 3 060, 1 590, 3 060, 1 590, 3 100, 1 630, 1 550 1 600, 1 610, 1 540 3 030, 1 600, 1 620, 3 020, 1 590 1 610, 1 620, 1 620, 1 540 1 620, 3 020, 1 590 1 600 3 060, 1 590 3 080, 1 600 3 020, 1 580 1 630, 1 595 1 630, 1 550 3 020, 1 590, 1 620, 1 550 1 620, 1 540 3 020, 1 610, 1 550
1

H-NMR (DMSO-d6 /TMS) (ppm)

2-pyridyl 42 3-pyridyl 55

4.74 (s, 2H, CH2), 6.64 (s, 1H, CH), 6.937.81 (m, 10H, Ar), 8.54 (s, 1H, NH), 8.63 (m, 3H, Py) 4.81 (s, 2H, CH2), 6.75 (s, 1H, CH), 6.938.62 (m, 13H, Ar), 8.68 (s, 1H, NH) 4.80 (s, 2H, CH2), 6.69 (s, 1H, CH), 6.938.64 (m, 13H, Ar), 8.71 (s, 1H, NH) 2.13 (s, 3H, CH3), 4.81 (s, 2H, CH2), 6.70 (s, 1H, CH), 6.838.64 (m, 13H, Ar + NH) 2.24 (s, 3H, CH3), 4.81 (s, 2H, CH2), 6.78 (s, 1H, CH), 6.808.68 (m, 13H, Ar + NH) 2.23 (s, 3H, CH3), 4.79 (s, 2H, CH2), 6.70 (s, 1H, CH), 6.788.65 (m, 12H, Ar), 8.71 (s, 1H, NH) 2.19 (s, 3H, CH3), 4.77 (s, 2H, CH2), 6.74 (s, 1H, CH), 6.868.74 (m, 13H, Ar + NH) 2.22 (s, 3H, CH3), 4.78 (s, 2H, CH2), 6.70 (s, 1H, CH), 6.888.66 (m, 12H, Ar), 8.72 (s, 1H, NH) 3.74 (s, 3H, CH3), 4.80 (s, 2H, CH2), 6.79 (s, 1H, CH), 6.878.65 (m, 13H, Ar + NH) 3.72 (s, 3H, CH3), 4.76 (s, 2H, CH2), 6.71 (s, 1H, CH), 6.838.64 (m, 13H, Ar + NH) 3.66 (s, 3H, CH3), 4.79 (s, 2H, CH2), 6.519.13 (m, 14H, Ar + NH + CH) 3.66 (s, 3H, CH3), 4.79 (s, 2H, CH2), 6.708.63 (m, 13H, Ar + CH), 8.71 (s, 1H, NH) 3.67 (s, 3H, CH3), 4.75 (s, 2H, CH2), 6.75 (s, 1H, CH), 6.838.64 (m, 12H, Ar), 8.67 (s, 1H, NH) 3.65 (s, 3H, CH3), 4.73 (s, 2H, CH2), 6.68 (s, 1H, CH), 6.818.64 (m, 12H, Ar), 8.69 (s, 1H, NH) 4.92 (s, 2H, CH2), 6.78 (s, 1H, CH), 6.958.65 (m, 13H, Ar + NH) 4.90 (s, 2H, CH2), 6.71 (s, 1H, CH), 6.958.63 (m, 12H, Ar), 8.68 (s, 1H, NH) 4.88 (s, 2H, CH2), 6.76 (s, 1H, CH), 6.968.73 (m, 13H, Ar + NH) 4.88 (s, 2H, CH2), 6.72 (s, 1H, CH), 6.978.66 (m, 12H, Ar), 8.74 (s, 1H, NH) 4.84 (s, 2H, CH2), 6.75 (s, 1H, CH), 7.008.62 (m, 12H, Ar), 8.67 (s, 1H, NH) 4.82 (s, 2H, CH2), 6.69 (s, 1H, CH), 6.998.64 (m, 12H, Ar), 8.71 (s, 1H, NH) 5.04 (s, 2H, CH2), 6.79 (s, 1H, CH), 7.208.63 (m, 12H, Ar), 8.74 (s, 1H, NH) 5.02 (s, 2H, CH2), 6.73 (s, 1H, CH), 7.198.64 (m, 12H, Ar), 8.78 (s, 1H, NH)

1 620, 1 580, 1 500 1 600, 1 580, 1 540

The tested compounds showed activity against mycobacteria with MIC values ranging from 6.25 g/mL to over 100 g/mL. MIC values as well as the results of cytotoxicity assays are reported in table III. The most active compounds against M. tuberculosis H37Rv were 2d, 2j and 2k (MIC 6.25 g/ml). However, a signicant

activity was also exhibited by compounds 2c, 2f, 2g and 2i (MIC 12.5 g/mL) and 2b, 2h and 4a (MIC 25 g/ mL). The most effective compounds in the inhibition of growth of M. tuberculosis INH-R were 2g, 2j and 2k (MIC 25 g/mL). All the tested compounds showed a higher activity against M. tuberculosis H37Rv than M.

1074

Figure 1. Synthetic pathways to compounds 2, 3, 4 and 5.

tuberculosis INH-R, with the exception of compounds 4j, 5g and 5j that were inactive against M. tuberculosis H37Rv (MIC > 100 g/mL) but showed a weak activity against M. tuberculosis INH-R (MIC 50g/mL). Isonycotinoylhydrazones 2g, 2j and 2k were the most active compounds in the whole series, with potencies superior to INH against M. tuberculosis INH-R. Although none of the tested compounds exhibited a signicant activity against M. fortuitum, we observed a poor activity by 2g and 2i (MIC 100 g/mL) against M. avium. Compounds 25 were inactive at inhibiting the growth of the Gram-negative species tested. A weak activity against Staphyloccocus aureus was shown by compounds 2h, 2i and 2j (MIC 100 g/mL) and by compound 3a against C. albicans (MIC 100 g/mL). From a rst examination of these results, it appears that compounds 2, containing the NH2 group, show better activity against M. tuberculosis H37Rv and M. tuberculosis INH-R with respect to derivatives 3, 4 and 5. Further studies to acquire more information about structureactivity relationships are in progress in our laboratories.

5.1.1. 2-Amino-2-(isonicotinoylhydrazono)ethyl aryl ethers 2 General Procedure: to a stirred solution of NaOEt (0.005 mol) in dry EtOH (5 mL) the appropriate aryloxyacetonitrile 1 (0.05 mol) was added. The mixture was stirred at room temperature for 1 h and after cooling to 0 C, an ethanolic solution of INH (6.8 g, 0.05 mol) was added dropwise. The resulting solution was stirred at the
Table III. Cytotoxicity and antimycobacterial activity of compounds 25. Com- MNTD pound (g/mL) Vero cells 2b 2c 2d 2f 2g 2h 2i 2j 2k 2l 4a 4j 4k 5c 5g 5i 5j INH 1 000 1 000 500 62.5 500 500 1 000 500 1 000 1 000 1 000 250 1 000 1 000 500 500 500 1 000 MIC (g/mL) M. tuberculosis M. tuberculosis M. avium H37Rv INH-R ATCC 19421 ATCC 25584 25 12.5 6.25 12.5 12.5 25 12.5 6.25 6.25 50 25 > 100 50 50 > 100 50 > 100 0.06 100 50 50 100 25 100 50 25 25 > 100 50 50 100 > 100 50 50 50 50 > 100 > 100 > 100 > 100 100 > 100 100 > 100 > 100 > 100 > 100 > 100 > 100 > 100 > 100 > 100 > 100 100

5. Experimental protocols 5.1. Chemistry Melting points were determined on a Koer hot stage and are uncorrected. IR spectra were recorded on Nujol mulls between salt plates in a Perkin-Elmer 398 spectrophotometer. 1H-NMR spectra were recorded on a Varian Unity 300 spectrometer. Elemental analyses were carried out with a Carlo Erba Model 1106 Elemental Analyzer.

1075 same temperature for 2 h and then allowed to reach room temperature and to stand overnight. The formed precipitate was collected by ltration and recrystallised from the solvent shown in table I. 5.1.2. 2-(Pyridylmethyleneamino)-2-(isonicotinoylhydrazono)ethyl aryl ethers 3, 4 and 5 General procedure: to a solution of compounds 2 (1.85 mmol) and 2-, 3- or 4-pyridinecarboxaldehyde (1.85 mmol) in EtOH (15 mL), a few drops of piperidine were added. After heating at reux for 6 h the solvent was removed in vacuo and the residue treated with H2O (20 mL) and extracted with CHCl3 (3 10 mL). The organic layers were dried (Na2SO4) and the solvent evaporated to give the crude 3, 4 and 5 that were puried by crystallisation from the solvent indicated in table II. 5.2. Microbiology For the antimicrobial and cytotoxicity studies, compounds were dissolved in DMSO at 10 mg/mL and kept at 20 C. The working solutions were prepared in the same medium used for the tests. To avoid interference by the solvent [14], the highest DMSO concentration was 1%. The MICs of various compounds against Grampositive, Gram-negative and C. albicans were determined by a standard broth serial dilution method [15, 16]. 5.2.1. Antibacterial assay The antimicrobial activity of compounds 25 was evaluated against six Gram-positive and ve Gramnegative species isolated at the Sezione di microbiologia e virologia, Dipartimento di Scienze Chirurgiche e Trapianti dOrgano, from clinical specimens obtained from patients treated at the Universit di Cagliari. In particular, compounds 25 were tested against Staphylococcus aureus (isolated from urine), S. epidermidis (isolated from urine by suprapubic aspirate), Streptococcus agalactiae (isolated from vaginal swab), S. faecalis (isolated from urine), Bacillus licheniformis (isolated from blood), B. subtilis (isolated from eye swab), Escherichia coli (isolated from urine), Pseudomonas aeruginosa (isolated from urine), Salmonella typhi (isolated from faeces), Proteus mirabilis (isolated from urine) and Klebsiella pneumoniae (isolated from urine). Tests with Grampositive and Gram-negative bacteria were carried out in Mueller Hinton broth (Difco). The compounds were diluted in the test medium to obtain nal concentrations ranging from 1000.19 g/mL. Tubes were inoculated with 1 105 cells/mL and were incubated at 37 C for 18 or 24 h. The effects on the growth of mycobacteria was investigated against M. tuberculosis H37Rv ATCC 25584, M. avium ATCC 19421, M. fortuitum ATCC 9820 and M. tuberculosis resistant to isoniazid (INH-R) isolated from a patient with an active clinical infection. The determination of MICs against mycobacteria were carried out by the two-fold agar dilution method [17] using 7H11 agar (Difco Laboratories) containing compounds 25 at concentrations that ranged between 1000.19 g/mL, on which 100 L of the test bacterial suspension were spotted. Suspensions to be used for drug susceptibility testing were prepared from 7H9 broth cultures supplemented with 10% OADC (oleic acid-albumin-dextrosecatalase) enrichment (Difco Laboratories) and 0.05% (v/v) Tween 80 to avoid clumping. Cells were then washed, suspended in saline, shaken and sonicated in a bath type ultrasonicator (output power 80 W) until visibile clumps were disrupted (usually from 1530 s). Suspensions were then diluted in saline to a turbidity of no. 1 McFarland (M. tuberculosis) or no. 0.5 McFarland (M. avium and M. fortuitum) and then diluted to obtain inocula of 3 105 cells per well of M. tuberculosis and 1.5 104 cells per well of M. avium and M. fortuitum. The MICs of the compounds were determined after 7 (M. fortuitum) or 21 (slow growers) days of cultivation at 37 C in a CO2 (5% CO2/95% humidied air) incubator. 5.2.2. Antifungal assay For the evaluation of the antifungal activity, Candida albicans ATCC E10931 was employed. Antifungal activity against C. Albicans ATCC E10231 was evaluated in yeast extract peptone dextrose medium (Difco) [18]. 5.2.3. In vitro cytotoxicity assay Cell cytotoxicity of compounds 25 was tested in vitro by two methods. In the rst method, RPMI 1640 medium (Gibco) with 2% foetal calf serum (FCS, Gibco) alone, or RPMI 1640 with 2% FCS containing compounds at concentrations ranging from 1 00062.5 g/mL, were inoculated onto cultures of Vero cells in 6 well tissue culture plates. The cells were observed daily for 6 days for any sign of cell cytotoxicity compared with the controls. In the second method, a cell viability assay previously reported [19, 20] was used. Monolayers of Vero cells in 96 multiwell plates were incubated with the testing compounds at concentrations of 1 00062.5 g/ mL in RPMI 1640 with 5% FCS for 48 h and the medium replaced with 50 L of 1 mg/mL solution of 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) in RPMI without phenol red. Cells were incubated at 37 C for 3 h, the untransformed MTT removed and 0.04 N HCl isopropanolic solution (50 L) was added to each well. After a few minutes at room temperature to ensure that all crystals were dissolved, the plates were

1076 read using an automatic plate reader with a 650 nm test wavelength and a 690 nm reference wavelength.
[7] [8] Sherman D.R., Mdluli K., Hickey M.J., Arain T.M., Morris S.L., Barry III C.E., Kendall Stover C., Science 272 (1996) 16411643. Mandell G.L., Sande M.A., in: Goodman and Gilmans The Pharmacological Basis of Therapeutics, 8th Edition, Pergamon Press, New York, 1984, p. 1147. Ban E., Mamolo M.G., Vio L., Predominato M., J. Chemother. 5 (1993) 164. Mamolo M.G., Vio L., Ban E., Farmaco 51 (1996) 6570. Bernard A.M., Cocco M.T., Congiu C., Onnis V., Piras P.P., J. Heterocycl. Chem. 34 (1997) 12831290. Bernard A.M., Cocco M.T., Congiu C., Onnis V., Piras P.P., Synthesis (1998) 317320. McManus J.R., Herbst R.M., J. Org. Chem. 24 (1959) 14641467. Jagannath C., Reddy V.M., Gangadharam P.R., J. Antimicrob. Chemother. 35 (1995) 381390. Trhupp L.D., in: Lorian V. (Ed.), Antibiotics in Laboratory Medicine, Williams & Wilkins, Baltimore, 1986, pp. 93150. Washington II J.A., Sutter V.L., in: Lennette E.H., Spaulding E.H., Truant J.I. (Eds.), Manual of Clinical Microbiology, American Society of Microbiology, Washington, 1980, pp. 533544. Hopewell P., Cynamon M., Starke J., Iseman M., OBrein R., Clin. Infec. Dis. 15 (suppl 1) (1992) S282S295. Pfaller M.A., Rinaldi M.G., Galgiani J.N., Bartlett M.S., Body B.A., Espinel-Ingroff A. et al., Antimicrob. Agents Chemother. 34 (1990) 16481654. Denizot F., Lang R., J. Immunol. Methods 89 (1986) 271277. Mosmann T., J. Immunol. Methods 65 (1983) 5563.

Acknowledgements Financial support from the Ministero della Universit e Ricerca Scientica e Tecnologica is gratefully acknowledged.

[9] [10] [11] [12] [13] [14]

References
[1] [2] [3] [4] [5] [6] Bloom B.R., Murray C.J.L., Science 257 (1992) 10551064. World Health Organization, Bridging the gaps: the world health report, World Health Organization, Geneva, 1995. World Health Organization report on TB epidemic, Global TB programme, World Health Organization, Geneva, 1997. Barnes P., Blotch A.B., Davidson B.T., Snyder Jr. D.E., N. Engl. J. Med. 324 (1991) 16441650. Snyder Jr. D.E., Roper W.L., N. Engl. J. Med. 326 (1992) 703705. Freiden T.E., Sterling T., Pablos-Mendez A., Kilburn J.O., Cauthen J.O., Dooley S.W., N. Engl. J. Med. 328 (1993) 521526.

[15] [16]

[17] [18]

[19] [20]

Eur. J. Med. Chem. 34 (1999) 10771083 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1077

Short communication

Structure-activity relationships of quaternary protoberberine alkaloids having an antimalarial activity

Kinuko Iwasaa*, Yumi Nishiyamaa, Momoyo Ichimarua, Masataka Moriyasua, Hye-Sook Kimb, Yusuke Watayab, Takao Yamoric, Turuo Takashid, Dong-Ung Leee
Kobe Pharmaceutical University, 4-19-1 Motoyamakita, Higashinada-ku, Kobe 658-8558, Japan Faculty of Pharmaceutical Sciences, Okayama University, Tsushima, Okayama 700-8530, Japan c Division of Experimental Chemotherapy, Cancer Chemotherapy Center, 1-37-1 kamiikebukuro, Toyoshima-ku, Tokyo 170-8455, Japan d Institute of Molecular and Cellular Biosciences, University of Tokyo, 1-1-1 yayoi, bunkyo-ku, Tokyo 113-0000, Japan e Department of Biochemistry, College of Natural Science, Dongguk University, Kyongju Kyongbuk 780-714, Korea
b a

(Received 20 April 1999; revised 28 June 1999; accepted 30 June 1999)

Abstract Seventeen quaternary protoberberine alkaloids related to berberine 1 were tested for antimalarial activity in vitro against Plasmodium falciparum and structure-activity relationships are proposed. The activity of the protoberberine alkaloids was inuenced by the type of the oxygen substituents on rings A, C and D and the position of the oxygen functions on ring D. The position of the oxygen functions on ring D and the type of the oxygen substituents at the C-13 position (ring C) strongly inuenced the activity. Shifting the oxygen functions at C-9 and C-10 to C-10 and C-11 on ring D resulted in a signicant increase in the activity. Compounds bearing a methylenedioxy function at C-2 and C-3 (ring A) or C-9 and C-10 (ring D) showed higher activity than those which have methoxy groups at the same positions. Introduction of a methoxy group into the C-1 position (ring A) decreased the activity. Replacement of a hydroxy group at C-2 or C-3 (ring A) by a methoxy group led to a reduction in the activity. Displacement of a hydroxy function at C-13 (ring C) by the oxygen substituents such as OMe, OEt, OCOOEt, and OCON(Me)2 reduced the activity. In the same replacement at C-9 (ring D), the activity depended upon the type of the oxygen function. Six protoberberines displayed more potent activity than berberine 1. The activity decreased in the order: 10, 11, 17 and 18 > 7 and 8 > 1. 1999 ditions scientiques et mdicales Elsevier SAS in vitro antimalarial activity / structure-activity relationships / protoberberinium salts

1. Introduction In Japan, and other Asian countries, berberine 1 and the extracts of Phellodendri cortex (Obaku in Japanese) and Coptidis rhizoma (Oren in Japanese) are used in the treatment of diarrhoea and other gastrointestinal diseases. Berberine and its relatives exhibit several types of biological activity [1]. We found that some of the 8- and 13-alkylderivatives of berberine 1 and palmatine 2 possessed antimicrobial and antimalarial activities [26]. We have recently shown that a 13-hydroxy derivative of 1 had similar antimalarial activity to that of 1 [6].

In the present study, the antimalarial activity of 17 protoberberinium salts, in which the type and/or position of the oxygen substituents such as hydroxy, methylenedioxy, methoxy, ethoxycarbonyloxy, and N, N-dimethylcarbamoyloxy functions on the aromatic rings A, C and D are different from 1 or 2 (table I), were examined using the selectivity indices as an indication of the activity. The results on the study of the structure-antimalarial activity relationships of the tested alkaloids are discussed. 2. Chemistry The protoberberinium salts 313 and 1720 have been prepared or isolated from natural sources previously [2, 5, 7, 8]. Palmatrubine 14 was prepared according to the method applied to the synthesis of 5 [2]. The carboethoxy

*Correspondence and reprints

1078
Table I. Protoberberinium salts. Table III. Mass spectral data and melting points of the protoberberinium salts (15, 16, 21 and 22).
Compound 15 16 21 22 M.p. C (dec) 160166 210214 155158 194200 Formula C22H20NO6 C22H21N2O5 C23H22NO7 C23H23N2O6 LSIMS m/z [M Cl]+ 394 393 424 423 HR-LSIMS calcd. found 394.1289 393.1449 424.1395 423.1556 394.1310 393.1458 424.1404 423.1565

R1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 H H H H H OMe H H H H H H H H H H H H H H H H

R2

R3

R4 OMe OMe OMe OMe OH OMe OCH2O OCH2O OCH2O H H OMe OMe OH OCOOEt OCON(Me)2 OMe OMe OMe OMe OMe OMe

R5 OMe OMe OMe OMe OMe OMe

R6 H H H H H H H H H OMe OMe H H H H H H H H H H H

R7 H H Me Me H H H Me Me H Me H Me H H H OH O OMe OEt OCOOEt OCON(Me)2

OCH2O OMe OMe OCH2O OMe OMe OCH2O OCH2O OCH2O OCH2O OMe OMe OCH2O OCH2O OMe OH OH OMe OMe OMe OCH2O OCH2O OCH2O OCH2O OCH2O OCH2O OCH2O OCH2O

and N,N-dimethylcarbamoyl derivatives of berberrubine 5 and 13-hydroxyberberine 17 (15 and 21, and 16 and 22, respectively) were prepared by the reaction with ethyl chloroformate and N,N-dimethylcarbamoyl chloride, respectively, in the presence of triethylamine or pyridine. 1 H-NMR (table II), LSIMS (table III), and HR-LSIMS (table III) data were consistent with the assigned structures. 3. Results and discussion Protoberberinium salts 622 were tested in vitro against human malaria Plasmodium falciparum FCR-3. The antimalarial activity of each compound was determined as a percentage reduction. The compound concentration required to inhibit cell growth by 50% was expressed as IC50 (table IV). From the evaluation of the toxicity of the compounds for mammalian cells, the concentration causing a 50% reduction of cell growth (GI50) of mouse mammary FM3A cells, a model of the host, was determined (table IV). The GI50/IC50 ratios for the compounds were calculated as selectivity indices (table IV). These ratios were used as an evaluation of antimalarial activity. The results are presented in table IV.

OMe OMe OMe OMe OMe OMe OMe OMe OMe OMe OMe OMe OMe

Table II. 1H-NMRa data of the protoberberinium salts 1416, 21 and 22.
C-9 or C-13 OCOOCH2CH3 14 C-9 or C-13 H-5 OCON(CH3)2 3.25 2H, t (6.0) 1.43 3H, t (7.0) 3.26 2H, t (6.5) 3.09 3.31 3.26 3H, s 3H, s 2H, t (6.5) 1.31 3H, t (7.0) 3.21 2H, t (6.0) 2.98 3.42 3.20 3H, s 3H, s 2H, t (6.0) 4.87 2H, t (6.0) 4.94 2H, t (6.5) 4.95 2H, t (6.5) 4.95 overlap 4.92 overlap 4.00 3H, s 3.92 3H, s 4.05 3H, s 4.11 3H, s 4.09 3H, s 4.13* 3H, s 4.12* 3H, s 4.24* 3H, s 4.22* 3H, s 6.11 2H, s 6.11 2H, s 6.12 2H, s 6.11 2H, s 7.00 1H, s 6.97 1H, s 6.97 1H, s 7.03 1H, s 7.00 1H, s 7.58 1H, s 7.69 1H, s 7.68 1H, s 7.67 1H, s 7.63 1H, s 7.96 7.71 1H, d (9.0) 1H, d (9.0) 8.24* 8.21* 1H, d (9.0) 1H, d (9.0) 8.19* 8.18* 1H, d (9.0) 1H, d (9.0) 8.19* 8.01* 1H, d (9.0) 1H, d (9.0) 8.17 7.96 1H, d (9.0) 1H, dd (9.0, 0.5) 8.65 1H, s 8.82 1H, s 8.80 1H, s 9.73 1H, s 9.79 1H, s 9.73 1H, s 9.88 1H, s 9.84 1H, d (0.5) H-6 2-OMe 3-OMe 10-OMe 9-OMe OCH2O H-4 H-1 H-11 H-12 H-13 H-8

15 4.40 2H, q (7.0) 16

21 4.31 2H, q (7.0) 22

Coupling constants (Hz in parentheses). *Assignments may be interchanged.

1079
Table IV. In vitro antimalarial activity of protoberberinium salts 122. 50% growth inhibition (M) Plasmodium falciparum mouse mammary cells FCR-3 FM3A IC50 GI50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Quinine
a

Selectivity indexes mouse mammary cells/ Plasmodium falciparum GI50/IC50 > 44b > 4.0b > 5.2b > 3.5b 2.3b 20 56 > 54 > 28 >106 > 52 > 25 > 16 5.6 2.0 > 22 > 68 > 53 12 2 25 7 910

Distribution of selectivity indexes human tumour cells/ Plasmodium falciparum the number (selectivity indexes: GI50a/IC50) < 150 150300 > 300 27 38 38 38 38 23 7 4 (> 370)

0.27 6.4b 2.1b 7.0b 2.2b 0.60 0.25 0.26 0.50 0.18

> 12 > 27b > 11b > 25b 5.0b 12 14 > 14 > 14 > 19
b

2 (300400) 6 (> 400)

0.29 0.67 1.0 0.90 2.2 1.9 0.56 0.78 2.4 0.084 0.68 5.3 0.11

> 15 > 17 > 16 5.0 4.0 > 41 > 38 > 41 20 0.16 17 37 100

1 (300400) 5 (400550) 19 (> 634)

38 (> 454)

20 (< 38)

17 (> 1100)

GI50 values presented in table V. bThese data have been previously reported [6].

Introduction of a methoxy group at the C-1 position (ring A) of 1 caused a decrease in antimalarial activity (compare 6 with 1) (table IV). Replacement of methoxy groups at the C-9 and C-10 positions (ring D) of 1 or 3 by a methylenedioxy group increased the activity (compare 7 with 1 or 8 with 3) (table IV) and the same replacement at the C-2 and C-3 positions (ring A) of 9 resulted in an increase in the activity (compare 9 with 8) (table IV). Shifting methoxy groups at C-9 and C-10 of 1 or 3 to C-10 and C-11 caused a signicant increase in the activity (compare 10 with 1 or 11 with 3) (table IV). Displacement of a methoxy group at C-3 (ring A) of 2 by a hydroxy function increased the activity and the same replacement at the C-2 position (ring A) of 4 also increased the activity (compare 12 with 2 and 13 with 4) (table IV). Substitution of a methoxy group at C-9 (ring D) of 2 by a hydroxy group increased the activity (compare 14 with 2) (table IV), though the same displacement at C-9 of 1 resulted in a decrease in the activity (compare 5 with 1) as shown in the previous data [6]. Replacement of a hydroxy group at C-9 (ring D) of 5 by

OCOOEt did not notably change its activity (compare 15 with 5) (table IV), whereas replacing a hydroxy function at C-9 of 5 by OCON(Me)2 caused an increase in the activity (compare 16 with 5) (table IV). We have found that the inhibitory effect of 13hydroxyberberine 17 was comparable to that of berberine 1 [6]. However, the inhibitory effects of phenolbetaine form 18 have been reported to be much less than that of 1 [9]. Thus, a comparison of the inhibitory effect between 17 and 18 was carried out. As a result, both 17 and 18 displayed a similar inhibitory effect in vitro. Substitution of a hydroxy group at the C-13 position of 17 by OMe or OEt or OCOOEt or OCON(Me)2 showed a reduction in the activity (compare 1922 with 17) (table IV). Among the tested salts, 7, 8, 10, 11, 17 and 18 exhibited much more potent activity (selectivity index over 50) than berberine 1. The selectivity indexes (table IV) of 1, 7, 10 and 17 were also determined using the 50% growth-inhibitory concentration (GI50, in table V) derived from the results [10] of in vitro cytotoxity tests on 38 human tumour

1080
Table V. Cytotoxicities of protoberberinium salts 15, 7, 10 and 17 against various human cell lines. Human tumour cell line Breast HBC-4 BSY-1 HBC-5 MCF-7 MDA-MB-231 CNS U251 SF-268 SF-295 SF-539 SNB-75 SNB-78 Colon HCC2998 KM-12 HT-29 WiDr HCT-15 HCT-116 Lung NCI-H23 NCI-H226 NCI-H522 NCI-H460 A549 DMS273 DMS114 Melanoma LOX-IMVI Ovarian OVCAR-3 OVCAR-4 OVCAR-5 OVCAR-8 SK-OV-3 Renal RXF-631L ACHN Stomach St-4 MKN1 MKN7 MKN28 MKN45 MKN74
a b

Compounds / Cytotoxicity (GI50 in M)a 1 4.4 5.2 30 24 50 6.9 7.7 4.8 23 30 35 20 4.3 41 42 b 30 8.2 8.1 5.8 12 41 5.9 3.9 33 6.2 31 39 17 b 59 b b 76 26 14 28 16 2 76 27 b b b 55 48 b b b b b 71 b b b b b 58 28 b b b 31 b 46 b b b b b b b b b 95 b 53 3 5.1 2.8 29 10 47 5.7 14 1.8 17 30 37 16 4.6 31 34 b 27 7.1 14 4.3 16 77 4.2 2.1 25 5.2 22 29 14 b 64 b 83 42 32 13 28 16 4 33 12 99 50 b 76 91 b 81 b 55 96 35 b b b b 62 40 28 b b 40 25 b 34 b b b b b b b b 84 76 b b 5 35 23 39 20 56 28 30 27 25 22 26 51 27 43 35 50 32 48 19 37 11 22 25 42 32 44 47 31 40 49 26 17 36 38 31 27 35 31 7 19 10 11 12 39 39 27 b 37 5.9 52 38 7.6 79 53 b 37 18 36 12 19 48 20 6.0 40 21 36 31 32 b b b b 93 4.2 11 24 9.8 10 17 7.1 79 27 b 35 27 b b 42 b 44 33 b b b b 81 b 16 b b 56 20 b 29 77 b 91 b b b b b 39 b 34 96 17 b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b Quinine b 35 b 55 12 b b b 37 b b 76 b 41 b 43 87 29 100 86 b 52 85 31 b b 72 b b b 79 30 74 38 79 b 71

The cytotoxicity GI50 values are the concentrations corresponding to 50% growth inhibition. GI50 values were not presented previously [10]. GI50 value is > 100 M.

cell lines (seven lung, six colon, six CNS, six stomach, ve ovarian, ve breast, two renal, and one melanoma), the hosts for FM3A cells. The selectivity indexes of

inactive 25 and the antimalarial drug quinine were also determined as the control experiments. 13Hydroxyberberine 17 and compound 10 displayed a

1081 signicant activity, coptisine 7 less potent activity, berberine 1 much less activity, and compounds 25 inactivity. 4. Conclusion From the structure-activity point of view, some features can be pointed out. On the basis of the results derived from examinations with protoberberinium salts, it appears that the position and the type of the oxygen substituents on rings A, C and D had an inuence on the antimalarial activity. The activity was strongly inuenced by the position of the oxygen functions on ring D and the type of the oxygen substituents at the C-13 position (ring C). Compounds with the oxygen functions at C-10 and C-11 on ring D displayed more activity than those which bear the same substituents at C-9 and C-10. Compounds bearing a methylenedioxy function at C-2 and C-3 or C-9 and C-10 on ring A or D showed higher activity than those which have methoxy groups at the same position. Introduction of a methoxy group into the C-1 position on ring A caused a decrease in the activity. Replacement of a hydroxy group at C-2 or C-3 on ring A by a methoxy group led to a reduction in the activity. Displacement of a hydroxy function at C-13 on ring C by an OMe, OEt, OCOOEt or OCON(Me)2 function decreased the activity. In the same replacement at C-9 on ring D, the activity depended on the type of the oxygen substituents. Among the potent protoberberinium salts under investigation, the activity decreased in the order: 10, 11, 17 and 18 > 7 and 8 > 1. The most active compound may be comparable with the antimalarial drug quinine in activity. Studies on in vivo antimalarial activity of the most potent alkaloids are in progress. 5. Experimental protocols 5.1. Chemistry Melting points were determined on a Yanako Micromelting Point apparatus and are uncorrected. 1H-NMR spectra were recorded on a Varian VXR-500 (500 MHz) spectrometer using tetramethylsilane as an internal standard and CD3OD as solvent. Mass spectra were determined on a Hitachi M-4100 instrument. The secondary ion mass spectra (LSIMS) were measured using glycerol as matrix. Preparative HPLC and HPLC analyses were performed on a Hitachi M-6250 Intelligent Pump (6 mL/ min) and a Hitachi M-6200 Intelligent Pump (1 mL/min), respectively, and a Hitachi L-4000 UV detector (280 nm). Analyses were made on a Cosmosil 5C18-AR reversedphase column (20 i.d. 250 mm or 4.6 i.d. 150 mm) eluting with H2O (0.05% TFA)/MeOH (0.05% TFA), A/B, initial (30% of B), 20 min (100% of B). Berberine 1 and palmatine 2 were purchased. Protoberberinium salts 3 [7], 4 [7], 5 [2], 6 [5], 7 [5], 911 [5], and 1720 [5] had previously been prepared. Corysamine 8 was obtained by oxidation of tetrahydrocorysamine [11] isolated from Corydalis incisa Pers. Columbamine 12 [8] and dehydrocorybulbine 13 [8] were natural products isolated from C. turtschaninovii Besser forma yanhusuo Y.H. Chou and C.C. Hsu. 5.1.1. Preparation of palmatrubine 14 Palmatine (1 g) was heated at 240250 C in a dry oven under vacuum (2030 mm Hg) for 10 min. The crude product was recrystallized from EtOH to give 14 (870 mg, 90%), m.p. 288295 C (dec.); LSIMS m/z 338. For 1H-NMR data see table II. 5.1.2. Preparation of the ethoxycarbonyl derivative of berberrubine 15 A solution of ClCOOEt (1.2 g) in CHCl3 (3 mL) was added dropwise to a mixture of berberrubine 5 (300 mg) [2] and (Et)3N (900 mg) in CHCl3 (50 mL). The mixture was stirred at room temperature for 30 min and evaporated in vacuo. (Me)2CO was added to the residue and the resulting crystals were collected, which recrystallized from MeOH-(Me)2CO to give 15 (229 mg, 64%). For 1H-NMR, LSIMS, and HR-LSIMS data and melting point see tables II and III. 5.1.3. Preparation of N,N-dimethylcarbamoyl derivative of berberrubine 16 A solution of (Me)2NCOCl (0.7 mL) in CHCl3 (2 mL) was added dropwise to a mixture of berberrubine 5 (200 mg) and (Et)3N (0.5 mL) in CHCl3 (50 mL). The mixture was stirred at room temperature for 5 h. To the reaction mixture (Me)2NCOCl (0.5 mL) was further added and the mixture was allowed to stir at room temperature overnight. (Me)2NCOCl (0.5 mL) was added to the solution which was further stirred overnight. Solvent was removed under reduced pressure and (Me)2CO was added to the residue. The resulting crystals were recrystallized from MeOH-(Me)2CO to give 16 (196 mg, 82%). For 1H-NMR, LSIMS, and HR-LSIMS data and melting point see tables II and III. 5.1.4. Preparation of ethoxycarbonyl derivative of 13hydroxyberberine 21 A solution of ClCOOEt (0.7 mL) in CHCl3 (2 mL) was added dropwise to a mixture of 13-hydroxyberberine 17 (200 mg) [5] and (Et)3N (0.5 mL) in CHCl3 (20 mL). The

1082 mixture was stirred at room temperature for 1 h and evaporated in vacuo. (Me)2CO was added to the residue and the resulting crystals were recrystallized from H2OMeOH to give 21 (100 mg, 46%). For 1H-NMR, LSIMS, and HR-LSIMS data and melting point see tables II and III. 5.1.5. Preparation of N,N-dimethylcarbamoyl derivative of 13-hydroxyberberine 22 A solution of (Me)2NCOCl (0.5 mL) in CHCl3 (2 mL) was added dropwise to a mixture of 13-hydroxyberberine 17 (50 mg) and (Et)3N (0.5 mL) in CHCl3 (20 mL). The mixture was stirred at room temperature overnight. The reaction was followed by HPLC. To the solution 3 drops of pyridine was added and the mixture was allowed to stir at room temperature for 3 days. The solvent was evaporated under reduced pressure. The residue was puried by preparative HPLC to give 22 (20 mg, 36%). For 1HNMR, LSIMS, and HR-LSIMS data and melting point see tables II and III. 5.2. In vitro antimalaria screening Acknowledgements 5.2.1. Parasites and mammalian cells Plasmodium falciparum (ATCC 30932, FCR-3 strain) was used in our study. P. falciparum was cultivated by a modication of the method of Trager and Jensen [12]. Mouse mammary tumour FM3A cells (wild-type, subclone F28-7) in culture were used as a control for mammalian cell cytotoxicity [13]. 5.2.2. Drug testing The following procedures were used for assaying antimalarial activity [6, 14, 15]. Five microlitres of each solution were added to individual wells of a 24 well dish. Erythrocytes with 0.3% parasitaemia were added to each well containing 995 L of culture medium to give a nal haematocrit level of 3%. The plates were incubated at 37 C for 72 h in a 5% CO2/5% O2/90% N2 incubator. To evaluate the antimalarial activity of the test compound, a total of 1 104 erythrocytes/L of thin blood lm were examined under microscopy. The IC50 value refers to the concentration of the compound necessary to inhibit the increase in parasite density at 72 h by 50% of control. 5.2.3. Toxicity to mammalian cells FM3A cells grew with a doubling time of about 12 h. Prior to exposure to drugs, cell density was adjusted to 5 104 cells/mL. A cell suspension of 995 L was dispensed into the test plate, and compounds were added to individual wells of a 24 well dish. The plates were incubated at 37 C in a 5% CO2 atmosphere for 48 h. Cell This work was supported in part by a Grants-in-Aid for Scientic Research on Priority Areas (08281105) from the Ministry of Education, Science, Culture and Sports, Japan. References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] Bhukuni D.S., Jain S., in: Brossi A. (Ed.), The Alkaloids Vol. 28, Academic Press, New York, 1986, pp. 95171. Iwasa K., Kamigauchi M., Ueki M., Taniguchi M., Eur. J. Med. Chem. 31 (1996) 469478. Iwasa K., Kamigauchi M., Sugiura M., Nanba H., Planta Med. 63 (1997) 196198. Iwasa K., Lee D.U., Kang S.I., Wiegrebe W., J. Nat. Prod. 61 (1998) 11501153. Iwasa K., Nanba H., Lee D.U., Kang S.I., Planta Med. 64 (1998) 14. Iwasa K., Kim H.S., Wataya Y., Lee D.U., Eur. J. Med. Chem. 33 (1998) 6569. Iwasa K., Kondoh Y., Kamigauchi M., J. Nat. Prod. 58 (1995) 379391. 27th Symposium on Natural Drug Analysis, Higashiosaka, Osaka, (1998); p. 69. Jonathan L.V., Daniel L.K., J. Med. Chem. 31 (1988) 10841087. Yamori T., Jpn. J. Cancer Chemother. 25 Supplement II (1998) 373381. Nonaka G., Okabe H., Nishioka I., Takao N., J. Pharm. Soc. Jpn. 93 (1973) 8793. Trager W., Jensen J.B., Science 193 (1976) 673675. Yoshioka A., Tanaka S., Hiraoka O., Koyama Y., Hirota Y., Ayusawa D., Seno T., Garrett C., Wataya Y., J. Biol. Chem. 262 (1987) 82358241.

numbers were measured using a microcell counter CC130 (Toa Medical Electric Co., Japan). The GI50 value refers to the concentration of the compound necessary to inhibit the increase in cell density at 48 h by 50% of control. Selectivity refers to the mean of the GI50 value for FM3A cells per the mean of the IC50 value for P. falciparum. 5.2.4. Selective toxicity using human tumour cells The in vitro cytotoxicity assay was carried out according to the method [10, 16] of a modied National Cancer Institute protocol [17]. We screened eight compounds, 15, 7, 10, and 17 against a panel of 38 human cancer cell lines (seven lung, six CNS, six colon, six stomach, ve breast, ve ovarian, two renal, and one melanoma). The 50% growth-inhibitory concentration (GI50) for any single cell line is an index of cytotoxicity or cytostasis. The selective toxicity was estimated from the GI50/IC50 ratio between the malaria parasites and human tumour cells which served as a model host.

1083
[14] [15] Kim H.S., Miyake H., Arai M., Wataya Y., Parasitol. Int. 47 (1998) 5967. Takaya Y., Kurumada K., Takeuji Y., Kim H.S., Shibata Y., Ikemoto N., Wataya Y., Oshima Y., Tetrahedron Lett. 39 (1998) 13611364. [16] [17] Yamori T., Jpn. J. Cancer Chemother. 24 (1997) 129135. Weinstein J.N., Myers T.G., OConnor P.M., Friend S.H., Fornace J.R.A.J., Kohn K.W. et al., Science 275 (1997) 343349.

Eur. J. Med. Chem. 34 (1999) 10851091 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1085

Short communication

Synthesis and CNS activity of tricyclic theophylline derivatives. 8-substituted imidazo[2,1-f]theophyllines

Maciej Pawowskia*, Anna Drabczynskab, Jacek Katlabia, Maria Gorczycaa, Danuta Malecc, Jerzy Modzelewskic
b a Department of Pharmaceutical Chemistry, Collegium Medicum of Jagiellonian University, 9 Medyczna St., 30-688 Krakw, Poland Department of Chemical Technology of Drugs, Collegium Medicum of Jagiellonian University, 9 Medyczna St., 30-688 Krakw, Poland c Department of Pharmacodynamics School of Medicine, 4 Staszica St., 20-081 Lublin, Poland

(Received 26 April 1999; revised 13 June 1999; accepted 15 July 1999)

Abstract Based on previously described results of pharmacological in vitro and in vivo tests of some series of pyrimido- and diazepino-[2,1-f]theophylline derivatives, N8-unsubstituted, N8-benzyl and N8-arylpiperazino-alkyl derivatives of imidazo[2,1-f]theophyllines were synthesized and tested for their CNS activity. It has been shown that tricyclic [2,1-f]theophyllines possess sedative and analgesic activity. N8-Phenylpiperazinopropyl-1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophylline 6 showed signicant anti-serotonin and long-lasting hypothermic effects. N8-Benzyl-1,3,6,8-tetrahydroimidazo-7-on[2,1-f]theophylline 8 possess anticonvulsant properties. 1999 ditions scientiques et mdicales Elsevier SAS imidazo[2,1-f]theophyllines / CNS activity of fused [2,1-f]theophyllines / arylpiperazinoalkylimidazo[2,1-f]theophyllines acting on CNS

1. Introduction In this paper we would like to summarize the results of our investigation on chemical and pharmacological properties of derivatives in which the third heterocyclic ring is bound at position 7,8- of theophylline. In our initial studies we found that some benzyl or dialkylamino-alkyl derivatives of 1,3,6,7,8,9-hexahydropyrimido[2,1-f]theophyllines and 1,3,6,7-tetrahydro-(9H)-pyrimid-8-on[2,1f]theophyllines (gure 1) changed the pharmacological prole of the mother compound (theophylline) into sedative, hypothermic, analgesic and neuroleptic-like activity [13]. These unexpected results seem to be related to the presence of a heterocyclic six membered ring, as well as to the basic alkylamino side chain substituted at the N9 position. Moreover, in preliminary screening it appeared that among others, the most active are the derivatives which contain arylpiperazino-propyl or its butyl homologue as an N9-substituent. This observation prompted us to synthesize the series of new compounds with seven, six
*Correspondence and reprints

Figure 1. Structure of 9-substituted 1,3,6,7,8,9-hexahydro(9H)-pyrimido[2,1-f]theophyllines and 1,3,6,7-tetrahydro(9H)-pyrimid-8-on[2,1-f]theophyllines.

and ve membered rings bound at position 7,8- of theophylline and a piperazino-alkyl substituent at the N8, N9 or N10 position, to study the inuence of ring size and kind of substituent in fused heterocyclic moieties on CNS activity of investigated compounds. The enlargement of the third heterocyclic ring lead to 1,3-diazepino- and 1,3-diazepin-9-on-[2,1-f]-theophyllines [4] (gure 2).

1086

Figure 2. Structure of 10-substituted 1,3,6,7,8,9-hexahydro(10H)-1,3-diazepino[2,1-f]theophyllines and 1,3,6,7,8,10hexahydro-1,3-diazepin-9-on[2,1-f]theophyllines.

Both investigated groups exhibit high CNS activity. Particularly, N10-phenyl-piperazinopropyl-1,3,6,7,8,9hexahydro-(10H)-1,3-diazepino[2,1-f]theophylline showed strong analgesic and sedative activity, but only slight hypothermic and antiamphetamine effects [4]. A similar mode of action was observed in the group of pyrimidin-8-on[2,1-f]theophyllines with a double bond between C6C7 in the third ring [5] (gure 3). Particularly, the N9-phenylpiperazinopropyl-derivate of 1,3-dihydro-(9H)-pyrimid-8-on[2,1-f]theophylline induced hypothermia and strongly decreased the spontaneous locomotor activity and amphetamine induced hyperactivity [5]. Taking into account that various 1-aryl(heteroaryl)substituted piperazines have been reported to possess agonistic activity on serotonin receptors [6], we also investigated, in behavioral tests, the compounds with a 1-(2-pyrimido)-piperazinoalkyl substituent connected with a non-lactam tricyclic system [7] and some of their lactam analogues (1 and 2) (gure 4). Following our research to verify the inuence of the third ring in tricyclic componds on the pharmacological prole of CNS-activity, we undertook the synthesis of N8-unsubstituted (3) and 8-arylpiperazinoalkyl derivatives (47) of 1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophylline. In order to compare CNS-activity, previously obtained N9-benzyl-1,3,6,7-tetrahydro-(9H)-pyrimid-8-on[2,1-f] theophylline [1] with its analogue in which the ve member lactam ring is fused in the 7,8- position of theophylline, N8-benzyl-1,3,6,8-tetrahydroimidazol-7on[2,1-f]theophylline 8 was synthesized.

Figure 4. Structure of 1-(2-pyrimido)-piperazinoalkyl derivatives of tricyclic non-lactam and lactam system.

2. Chemistry 9-{3-[4-(2-Pyrimidynyl)-1-piperazinyl]-propyl}-1,3,6, 7-tetrahydro-(9H)-pyrimid-8-on[2,1-f]theophylline (1) and 10-{3-[4-(2-pyrimidynyl)-1-piperazinyl]-propyl}-1, 3,6,7,8,10-hexahydro-1,3-diazepin-9-on[2,1-f]theophylline (2) were prepared according to the procedures described previously [6]. In the reaction of 7--bromoethyl-8-bromotheophylline [8] with ammonia 8-unsubstituted 1,3,6,7tetrahydro-(8H)-imidazo[2,1-f]theophylline, (3) was obtained. Arylpiperazino-alkyl derivatives of 1,3,6,7tetrahydro-(8H)-imidazo[2,1-f]theophylline (47) were obtained in a several step synthesis (gure 5). 8--Hydroxyethyl-1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophylline (I) and its homologue (Ia) were prepared by condensation of 7--bromoethyl-8-bromotheophylline with 2-aminoethanol or 3-aminopropan-1ol. Compounds I and Ia, upon bromination with PBr3 in CHCl3, yielded II and IIa. Aminolysis of compound II and IIa with a double amount of phenylpiperazine or 2-pyrimidyl-piperazine in anhydrous toluene in the presence of K2CO3 gave the compounds 47. N8-Benzyl-1,3,6,8-tetrahydroimidazol-7-on[2,1-f]theophylline 8 was synthesized via cyclization of 8-benzylaminotheophylline-7-acetic acid [9] in boiling acetic anhydride (gure 6). 3. Pharmacological results and discussion The compounds 18 were evaluated for their inuence on the CNS. Their acute toxicity assessed in mice as the LD50 (condence limits) was as follows: Compound 1: 235 mg/kg (178310)

Figure 3. Structure of 9-substituted 1,3-dihydro-(9H)pyrimidin-8-on-[2,1-f]theophyllines.

1087

Figure 7. The inuence of compounds (0.1 LD50) on spontaneous locomotor activity of mice. Figure 5. Arylpiperazino-alkyl derivatives of 1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophylline (47).

Figure 6. N8-Benzyl-1,3,6,8-tetrahydroimidazol-7-on[2,1-f] theophylline 8.

Table I. The inuence of the compounds (0.1 LD50) on the head twitch responses of mice receiving L-5HTP (180 mg/kg i.p.). Compound Tylose 1 2 3 4 5 6 7 8 Mean SE number of head twitch responses 11.4 2.9 18.0 4.3 14.5 4.0 15.0 4.2 3.6* 1.3 12.3 5.5 1.8** 0.7 9.3 3.2 12.7 3.7

The compounds were injected 60 min before L-5HTP. * P < 0.05, ** P < 0.01 (Students t-test).

Compound 2: 560 mg/kg (478655) Compound 3: 320 mg/kg (269381) Compound 4: 175 mg/kg (117263) Compound 5: 320 mg/kg (264387) Compound 6: 540 mg/kg (439664) Compound 7: 2 000 mg/kg Compound 8: 2 000 mg/kg The CNS effects of all these substances (1/10 LD50) were studied in several tests, and it was observed that substances 2, 3, 4, 6, 7 and 8 induced marked sedative effects in the locomotor activity test (gure 7). Compound 6 had the strongest action, which reduced the activity of mice by about 90% (1/10 and 1/20 LD50). These doses of substance 6 also antagonized amphetamine-induced hyperactivity in mice, but did not change apomorphine-induced stereotypy or haloperidolinduced catalepsy in rats. Compounds 4 and 6 also had antiserotonin effects, as they diminished 5HTP-induced head-twitch reactions of mice (table I). Hypothermic action was observed after the administration of compounds 1, 6, 7 and 8 (gure 8). The most apparent and long-lasting effects were observed for compounds 6 and 8. Analgesic action in the hot plate test was induced in mice by nearly all substances, except for compound 8 (table II). These effects were of rather a short duration (up to 90 min after injection). Compound 8 had anticonvulsant action in the pentetrazole test in mice, as it prevented the tonic phase of the seizures and lethality of the animals. Reserpine-induced hypothermia was slightly reversed by substances 1, 3 and 4, and this action suggests some antidepressant properties of these three substances. How-

1088

Figure 8. The inuence of compounds (0.1 LD50) on body temperature of normothermic mice.

ever, we observed neurotoxic effects in the rota-rod test in mice following injection of nearly all compounds tested. 4. Conclusion Analysing the results of our preliminary investigations on chemical and pharmacological properties of tricyclic theophylline derivatives from the point of view of structure-activity relationships, it may be concluded that: 1. In contrast to mother theophylline, all investigated tricyclic compounds generally exhibited a strong sedative action, particularly those which contained the basic aminoalkyl side chain, substituted on the third bound ring at N8 of imidazole, N9 of pyrimidine and N10 of diazepine. Their sedative properties were conrmed in a number of behavioral tests (in mice). 2. The replacement of the methylene group with a carbonyl one in pyrimido[2,1-f]theophyllines or diaze-

pino[2,1-f]theophyllines (lactam structure) did not have a signicant inuence on the direction of pharmacological activity. 3. The potency of sedative action depends signicantly on the kind of N-substituted basic side chain. Variation of the alkylamino substituents conrmed unequivocally that the most apparent CNS activity was exerted by the compounds with phenylpiperazinopropyl- or phenylpiperazinobutyl- substituents. 4. Length reduction of the alkylene chain between dialkylamino substituents and tricyclic substituents from three or four to two carbon atoms resulted mostly in an increase of toxicity. 5. The replacement of the phenylpiperazinoalkyl substituent with an (-4-pyrimidinyl)-piperazino-alkyl side chain generally decreased the activity and revealed some neurotoxic effects.

Table II. The inuence of the tested compounds on the reactivity of mice in the hot plate test. Compound 0.1 LD50 Vehicle 1 2 3 4 5 6 7 8 Reaction time (s) of mice after injection (min) 30 8.9 0.6 14.5*** 0.9 13.0*** 0.7 10.8 0.6 14.7*** 1.0 13.4** 1.1 13.0** 1.0 14.7*** 1.1 10.1 0.6 60 9.3 0.6 13.4** 0.8 11.7 0.9 13.8** 1.1 12.3* 0.8 12.9* 1.0 14.3** 1.2 13.2*** 1.0 10.9 0.7 90 9.5 0.7 15.1** 1.2 10.8 0.6 10.4 0.6 13.8* 1.1 11.2 0.7 12.3 1.2 10.6 0.6 9.3 0.5 120 9.1 0.7 11.5 0.9 10.4 0.5 10.8 0.7 13.4* 1.2 11.2 0.8 11.5 0.9 11.2 0.8 10.2 0.6

* P < 0.05; ** P < 0.02; *** P < 0.001 (Students t-test).

1089 6. All phenylpiperazinoalkyl substituted tricyclic [2,1f]theophyllines reduced the amphetamine induced hyperactivity. Pyrimido[2,1-f]theophyllines with phenylpiperazinoalkyl side chains showed mainly analgesic and neuroleptic-like properties. The pyrimidinyl-piperazino analogues exhibited agonistic activity on 5-HT1A receptors. Their diazepino homologues exerted apparent analgesic activity. 7. Some of previously described tricyclic theophylline derivatives, evaluated for the affinities for 5-HT receptors, were classied as postsynaptic antagonists or partial agonists of 5-HT1A receptors [6]. Pyrimido[2,1-f-]theophyllines were also tested for afnity to adenosine receptors [10]. They were classied as adenosine receptor subtype A1 antagonists. Some of the presented data on tricyclic [2,1-f-]theophyllines summarized in this article remains open to further investigation. 5. Experimental protocol 5.1. Chemistry Melting points were determined on a Boetius apparatus and are uncorrected. UV spectra were obtained on the UV-VIS Lambda 12 spectrophotometer (Perkin Elmer) (conc. 5 105 M/dm3 in methanol). IR spectra were determined on the Specord M 80 spectrophotometer (Carl Zeiss Jena) in KBr pellets. 1H-NMR spectra were obtained on a Brucker AC200 F spectrometer in CDCl3 (except I measured in DMSO) solution using TMS as internal standard. Chemical shifts are reported in units. Coupling constants are reported in Herz. Elemental analyses indicated by the symbols were within 0.4% of the theoretical value. TLC was performed on Merck plates (Kieselgel 60 F254 with S = benzene:acetone: methanol (1:1:1) + 2 drops NH4OH. 5.1.1. 8--Hydroxyethyl-1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophylline I A mixture of 7--bromoethyl-8-bromotheophylline (5.0 g; 0.014 mol), 2-aminoethanol (5 cm3; 0.08 mol) and 40 cm3 of ethanol was reuxed for 5 h. After refrigeration the product was ltered off and recrystallized. Yield 69%; m.p. 238240 C (ethanol). Rf 0.74, UV max nm 299.4 (log e 4.30). 1H-NMR 3.17 (s, 3H, N3CH3), 3.34 (s, 3H, N1CH3), 3.303.36 (m, 2H, CH2OH), 3.573.64 (t, 2H, N8CH2), 3.913.98 (t, 2H, C7H2), 4.054.12 (t, 2H, C6H2), 4.824.86 (t, 1H, OH). IR (cm1) 3 300 (OH), 2 8002 950 (CH2)n, 1 650 (CO), 1 450 (CH2N). Anal. C11H15O3N5; C, H, N. 5.1.2. 8--Hydroxypropyl-1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophylline Ia The title compound was obtained similarly to compound I using 3-amino-1-propanol and isobutanol as a solvent. The mixture was reuxed for 10 h. After refrigeration, the separated product was recrystallized. Yield 54%; m.p. 186188 C (isopropanol). Rf 0.74, UV max nm 298.5 (log e 4.19). 1H-NMR 1.801.90 (m, 2H, CH2CH2CH2) 3.34 (s, 3H, N3CH3), 3.48 (s, 3H, N1CH3), 3.493.55 (m, 2H, CH2OH), 3.663.71 (t, 2H, N8CH2), 3.923.98 (t, 2H, C7H2), 4.224.28 (t, 2H, C6H2). IR (cm1) 3 3003 450 (OH), 2 8002 950 (CH2)n, 1 650 (CO), 1 450 (CH2N). Anal. C12H17O3N5; C, H, N. 5.1.3. 8-(-Bromoethyl)-1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophylline II and 1,3-8-(-bromopropyl)1,3,6,7-tetrahydro-(8H)-imidazo[2,1-f]theophylline IIa These two compounds were obtained similarly. To 0.01 mol of compound I or Ia, fourfold amounts of PBr3 in 25 cm3 of CHCl3 was added and reuxed for 5 h. Then the solvent and the excess of PBr3 were distilled off under reduced pressure and to the residue, water was added. The mixture was alkalized with 10% NaOH (to pH 8). The crude products were separated and recrystallized. Compound II. Yield 92%; m.p. 210212 C (ethanol). Rf 0.87, UV max nm 298.5 (log e 4.20). 1H-NMR 3.36 (s, 3H, N3CH3), 3.51 (s, 3H, N1CH3), 3.603.65 (t, 2H, CH2Br), 3.763.81 (t, 2H, N8CH2), 4.024.08 (t, 2H, C7H2), 4.244.31 (t, 2H, C6H2). IR (cm1) 2 8002 950 (CH2)n, 1 650 (CO), 1 450 (CH2N). Anal. C11H14O2N5Br; C, H, N. Compound IIa. Yield 94%; m.p. 162164 C (50 ethanol). Rf 0.87, UV max nm 299.4 (log e 4.11). 1 H-NMR 2.212.32 (m, 2H, CH2CH2CH2) 3.37 (s, 3H, N3CH3), 3.52 (s, 3H, N1CH3), 3.473.54 (t, 2H, CH2Br), 3.883.95 (t, 2H, C7H2), 4.054.12 (t, 2H, C6H2). IR (cm1) 2 8002 950 (CH2)n, 1 650 (CO), 1 450 (CH2N). Anal. C12H16O2N5Br; C, H, N. 5.1.4. 1,3,6,7-Tetrahydro-(8H)-imidazo[2,1-f]theophylline 3 A mixture of 7--bromoethyl-8-bromotheophylline (4.0 g; 0.011 mol), 5 cm3 of 25% ammonia and 15 cm3 of ethanol was heated in an autoclave at 180190 C for 5 h. Then the solution was evaporated and the residue was reuxed with isopropanol to separate the inorganic salts. The hot mixture was ltered off and the ltrate was distilled off under reduced pressure. The crude product was recrystallized from ethanol; m.p. 270272 C, yield 60%. Compound 3 was obtained by Eckstein et al. [11] using another method.

1090 5.1.5. General procedure for compounds 47 Compounds 47 were obtained by aminolysis of compounds II or IIa with double the amount of phenylpiperazine or 2-pyrimidynyl-piperazine. The reaction was carried out in anhydrous toluene in the presence of a stoichiometric amount of K2CO3 and reuxed for 5 h. The inorganic salt was ltered off from the hot mixture and the ltrate was refrigerated. The crystalline products were ltered off and recrystallized. 5.1.5.1. 8-{2-[4-(Phenyl)-1-piperazinyl]-ethyl}-1,3,6,7,tetrahydro-(8H)-imidazo[2,1-f]theophylline 4 Yield 52%; m.p. 164166 C (ethanol + butyl acetate). Rf 0.80, UV max nm 244.0 (log e 4.09), 296.7 (log e 4.23). 1H-NMR 2.702.75 (m, 10H, CH2N(CH2)4N), 3.19 (t, J = 4.8, 2H, N8CH2), 3.37 (s, 3H, N3CH3), 3.51 (S, 3H, N1CH3), 3.99 (t, J = 7.7, 2H, C7H2), 4.22 (t, J = 7.8, 2H, C6CH2), 6.866.93 (m, 3H, 3,4,5-phenyl), 7.26 (m, 2H, 2,6-phenyl). Anal. C21H27O2N7, C, H, N. 5.1.5.2. 8-{2-[4-(2-Pyrimidynyl)-1-piperazinyl]-ethyl}1,3,6,7,-tetrahydro-(8H)-imidazo[2,1-f]theophylline 5 Yield 91%; m.p. 194195 C (toluene). Rf 0.50, UV max nm 242.0 (log e 4.14), 301.0 (log e 3.99). 1H-NMR 2.552.70 (m, 10H, CH2N(CH2)4N), 3.37 (s, 3H, N3CH3), 3.51 (s, 3H, N1CH3), 3.78 (t, J = 5.0, 2H, N8CH2), 3.354.03 (m, 2H, C7H2), 4.184.26 (m, 2H, C6CH2), 6.48 (m, J = 4.7, 1H, 5-pyrimidyl), 8.30 (d, J = 4.8, 2H, 4,6-pyrimidyl). Anal. C19H25O2N9, C, H, N. 5.1.5.3. 8-{2-[4-(Phenyl)-1-piperazinyl]-propyl}-1,3,6,7,tetrahydro-(8H)-imidazo[2,1-f]theophylline 6 Yield 79%; m.p. 192194 C (toluene). Rf 0.75, UV max nm 246.0 (log e 4.09), 299.0 (log e 4.09). 1H-NMR 1.841.92 (m, 2H, CH2CH2CH2), 2.442.63 (m, 10H, CH2N(CH2)4N), 3.19 (t, J = 4.7, 2H, N8CH2), 3.35 (s, 3H, N3CH3), 3.47 (s, 3H, N1CH3), 3.86 (t, J = 8.0, 2H, C7H2), 4.20 (t, J = 8.0, 2H, C6CH2), 6.846.93 (m, 3H, 3,4,5-phenyl), 7.24 (t, J = 7.8, 2H, 2,6-phenyl). Anal. C22H29O2N7, C, H, N. 5.1.5.4. 8-{2-[4-(2-Pyrimidynyl)-1-piperazinyl]-propyl}1,3,6,7,-tetrahydro-(8H)-imidazo[2,1-f]theophylline 7 Yield 79%; m.p. 192194 C (toluene). Rf 0.59, UV max nm 242.0 (log e 4.15), 301.0 (log e 3.94). 1H-NMR 1.841.92 (m, 2H, CH2CH2CH2), 2.432.53 (m, 10H, CH2N(CH2)4N), 3.35 (s, 3H, N3CH3), 3.41 (t, J = 4.7, 2H, N8CH2), 3.49 (s, 3H, N1CH3), 3.753.92 (m, 2H, C7H2), 4.21 (t, J = 8.0, 2H, C6CH2), 6.47 (t, J = 4.8, 1H, 5-pyrimidyl), 8.30 (d, J = 4.8, 2H, 4,6-pyrimidyl). Anal. C20H27O2N9, C, H, N. 5.1.5.5. 8-Benzyl-1,3,6,8-tetrahydroimidazol-7-on[2,1f]theophylline 8 A mixture of 8-benzylaminotheophyllino-7-acetic acid (3.43 g; 0.01 mol) and 15 cm3 of acetic anhydride was reuxed for 3 h. After cooling, the product was separated by ltration, washed with a small amount of methanol and recrystallized. Yield 50%, m.p. 216 218 C (methanol). Rf 0.92 (chloroform:methanol: 25% NH4OH (8:2:0.25)), UV max nm 300.0 (log e 4.01) (conc. 5 105 M/dm3 in chloroform). 1H-NMR 3.33 (s, 3H, N3CH3), 3.58 (s, 3H, N1CH3), 4.71 (s, 2H, N8CH2C6H5), 4.93 (s, 2H, C6CH2), 7.337.48 (m, 5H, C6H5). Anal. C16H15O3N5, C, H, N. 5.2. Pharmacology The experiments were performed on male Swiss mice (1926 g), and on Wistar rats (180200 g), kept at ambient temperature on a 12 h light/dark schedule, with free access to food and water before the experiments. The experimental groups consisted of 58 animals each. All investigated compounds were injected i.p., 1 h before the test, as a suspension in 0.5% tylose solution. The doses corresponding to 0.1 LD50 of the compounds were administered in all tests. The acute toxicity of the compounds was assessed in mice as the LD50 calculated on the basis of mortality within 48 h [12]. Spontaneous locomotor activity and amphetamineinduced hyperactivity in mice were observed in a photoresistor actometer for 30 min. Analgesic effects were assessed in the hot plate (56 C) test in mice, and body temperature was measured with a thermistor thermometer in the rectum of mice. Antipsychotic action was evaluated in the tests of amphetamine (5 mg/kg s.c.)-induced hyperactivity of mice and apomorphine (3 mg/kg s.c.)induced stereotypy in rats. Antiparkinsonian effects were measured by investigation of the inuence of the compounds on haloperidol (1 mg/kg)-induced catalepsy in rats. Anticonvulsant activity was assessed in pentetrazole (90 mg/kg s.c.)-induced seizures in mice. The antidepressant action was evaluated using the test of reserpine-induced hypothermia in mice (reserpine 2.5 mg/kg s.c., 20 h before injection of the investigated substances, and then body temperature was measured at 30 min intervals for 3 h). The inuence of the compounds on serotonin neurotransmission was evaluated in the head-twitch responses of mice following 5HTP injection [13, 14]. Neurotoxic properties were observed in the rota-rod test [15, 16] in mice. Statistical analysis was carried out using the Students t-test (convulsions) and Mann-Whitney U test (catalepsy, stereotypy).

1091 References
[1] [2] [3] [4] [5] [6] Drabczynska A., Pawowski M., Gorczyca M., Malec D., Mod zelewski J., Pol. J. Pharmacol. Pharm. 41 (1989) 385394. Pawowski M., Drabczynska A., Gorczyca M., Malec D., Mod zelewski J., Pol. J. Pharmacol. Pharm. 43 (1991) 6170. Drabczynska A., Pawowski M., Gorczyca M., Malec D., Mod zelewski J., Pol. J. Pharmacol. Pharm. 44 (1992) 487503. Pawowski M., Drabczynska A., Gorczyca M., Malec D., Mod zelewski J., Acta Pol. Pharm. 51 (1994) 385391. Pawowski M., Drabczynska A., Gorczyca M., Malec D., Mod zelewski J., Pharmazie 50 (1995) 453456. Chojnacka-Wjcik E., Kodzinska A., Drabczynska A., Pawowski M., Charakchieva-Minol S., Chon G., Gorczyca M., Eur. J. Med. Chem. 30 (1995) 587592. Malec D., Modzelewski J., Drabczynska A., Pawowski M., Gorc [8] [9] [10] [11] [12] [13] [14] [15] [16] zyca M., Pol. J. Pharmacol. Pharm. 47 (1995) 169173. Cacace F., Masironi R., Ann. Chim. 46 (1956) 806809. Cacace F., Crisera R., Zifferero M., Ann. Chim. 46 (1956) 99102. Geis U., Grahner B., Pawowski M., Drabczynska A., Gorczyca M., Mller C.E., Pharmazie 50 (1995) 333336. Eckstein M., Drabczynska A., Synthesis 8 (1979) 581584. Litcheld L.T., Wilcoxon F., J. Pharmacol. Exp. Ther. 96 (1949) 99113. Corn S.J., Pickering R.W., Werner B.T., Br. J. Pharmacol. 20 (1963) 106120. Moore N.A., Tye N.C., Axton M.S., Risius F.C., J. Pharm. Exp. Ther. 262 (1992) 545551. Gross F., Tripod J., Meir R., Schweiz. Med. Wochschr. 85 (1955) 305309. Dar M.S., Wooles W.R., Life Sci. 39 (1986) 14291437.

[7]

Eur. J. Med. Chem. 34 (1999) 10931100 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1093

Short communication

Synthesis and antimycobacterial activity of some coupling products from 4-aminobenzoic acid hydrazones
S. Gniz Kkgzela, Sevim Rollasa*, q Ilkay Kkgzela, Muammer Kirazb
a

University of Marmara, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, 81010 Haydarpasa, Istanbul, Turkey b University of Istanbul, Faculty of Medicine, Center for Research and Application of Culture Collections of Microorganisms, apa, Istanbul, Turkey (Received 25 February 1999; revised 2 July 1999; accepted 15 July 1999)

Abstract Various 2,3,4-pentanetrione-3-[4-[[(5-nitro-2-furyl/pyridyl/substituted phenyl)methylene]hydrazinocarbonyl]phenyl]hydrazones 3aj were synthesized by the reactions of acetylacetone with the diazonium salts of 4-aminobenzoic acid-[(5-nitro-2-furyl/pyridyl/substituted phenyl)methylene]hydrazides 2aj at 05 C. The structures of these compounds were determined using spectral data. All the synthesized compounds were evaluated for their antimycobacterial activity against Mycobacterium fortuitum ATCC 6841 and Mycobacterium tuberculosis H37Rv. Of the compounds screened, 2e, 2i, 3e and 3g were found to be active against M. fortuitum at an MIC value of 32 g/mL. Compound 3a, which exhibited > 90% inhibition in the primary screen at 12.5 g/mL against M. tuberculosis H37Rv, was the most promising derivative for antituberculosis activity. Results obtained from the level II screening showed that the actual MIC and IC50 values of 3a were 3.13 and 0.32 g/mL, respectively. The same compound was also tested against Mycobacterium avium, which was observed not to be susceptible to 3a. 1999 ditions scientiques et mdicales Elsevier SAS hydrazide-hydrazones / coupling products / antimycobacterial activity / cytotoxicity

1. Introduction Antituberculosis and antibacterial effects have been shown with various 4-aminobenzoic acid substituted benzalhydrazones [1]. Furthermore, the coupling products starting from diazonium salts of 4-aminobenzoic acid hydrazide substituted benzalhydrazones with indole have also been reported to possess promising antibacterial and antitubercular activities [2]. These observations prompted us to synthesize some novel coupling products of 4-aminobenzoic acid hydrazide substituted benzalhydrazone derivatives. Earlier reports indicated that certain compounds bearing aromatic amine functions, such as sulfaguanidine [3], PABA [4], benzocaine [4], sulfanilamide [5], 4-aminobenzoic acid hydrazide [6], oxadiazole [7, 8], thiadiazole [7], quinazolinone [7], 4-aminobenzoic acid hydrazide substituted benzalhydrazones [2], substituted
*Correspondence and reprints

anilines [9, 10], 1,2,4-triazoline-3(2H)-thiones [11, 12], 1,2,4-triazoline-3(2H)-one [12], 1,3,4-oxadiazoline2(3H)-thione [11, 12], 3,5-dimethyl-1H-pyrazole [11] and 4-aminobenzoic acid-[(4-uorophenyl)methylene]hydrazide [12] might be selected to couple their diazonium salts with compounds possessing an active hydrogen. 2. Chemistry 4-Aminobenzoic acid hydrazide 1 was prepared by the reaction of ethyl 4-aminobenzoate with hydrazine hydrate. 4-Aminobenzoic acid-[(5-nitro-2-furyl/pyridyl/ substituted phenyl)methylene]hydrazides 2aj were synthesized by condensation of 1 with appropriate aldehydes [13] as original compounds, except 2a [14] and 2c [15], which were previously reported elsewhere. The diazonium salts of 4-aminobenzoic acid-[(5-nitro-2-furyl/ pyridyl/substituted phenyl)methylene]hydrazide were then coupled with acetylacetone in ethanol (50%) con-

1094

Figure 1. Synthetic route to compounds 3aj.

taining sodium acetate. The resulting coupling products were formulated as 2,3,4-pentanetrione-3-[4-[[(5-nitro-2furyl/pyridyl/substituted phenyl)methylene]hydrazinocarbonyl]phenyl]hydrazones 3aj in view of spectral ndings (gure 1). Spectral investigations also revealed that the aromatic primary amine function present at 1 was not affected during the synthesis of 2aj, although no protective precautions were taken, probably because the -NH2 moiety of hydrazide function is more nucleophilic than the Ar-NH2 group. The novel compounds 3aj gave satisfactory elemental analyses. Their structures were established using spectral data (UV, IR, 1H-NMR and EI-MS). UV spectra of the synthesized compounds showed three absorption maxima at 236249, 265303 and 369402 nm values. The lack of absorption bands between 332360 nm [16, 17] and above 400 nm [18], which could be attributed to an azo function, together with observation of bands at around 369402 nm, indicated that compounds 3aj were in a hydrazone form [19]. The IR spectra of 3aj exhibited the NH band of hydrazone, the C=O band of ketone and hydrazone

arising from stretching vibrations at expected wavenumbers according to the literature [3, 4]. In the 1H-NMR spectra, methyl protons of the ketone moiety gave singlets at 2.52, 2.62 ppm (CDCl3, only for compound 3b) or 2.422.46 ppm (DMSO-d6, for compounds 3a, 3cj). The azomethine proton (-CH=N-) and NH proton of the hydrazide hydrazone N1H (-CONHN=CH-) exhibited the expected singlets at 8.328.71 ppm and 11.4913.52 ppm, respectively. Unlike the others, which were dissolved in DMSO-d6, the N1H proton of compound 3b resonated at 9.09 ppm, probably due to the use of CDCl3 as solvent [20]. The 1 H-NMR spectra of 3b in CDCl3 and 3cd and 3fj in DMSO-d6 displayed the hydrazone N2H(-CH=NNH-) protons at 13.7114.58 ppm. The hydrazone NH protons of 3a and 3e were observed to exchange with deuterium in DMSO-d6 [12]. In addition, in the 1H-NMR spectra of compounds 3aj, signals arising from the >CHN=Nstructure at 3.004.00 ppm [3, 4] were not observed. This nding also supported the idea that the structures of these compounds might be given in hydrazone form. The phenolic proton of compound 3g displayed a singlet at 9.69 ppm.

1095

Figure 2. Mass fragmentation pattern of 3a, 3i and 3j.

EI-Mass spectra of three selected prototypes, 3a, 3i and 3j conrmed their molecular weights and displayed characteristic fragment ions as shown in gure 2. The major fragmentation pathway appeared by the cleavage of -CONH-N= bonds of a hydrazide-hydrazone moiety. Initial loss of an acetyl fragment (m/z 43) followed by expulsion of CH3COCN from the molecular ion and the subsequent acquirement of a hydrogen radical afforded the fragment ions at m/z 282 and 308 for compounds 3i and 3j, respectively. The same type of fragment ion also appeared at m/z 274 in the EI-mass spectrum of compound 3a by the cleavage of >C=NNH- bonds of hydrazone followed by protonation. This latter ion exhibited the expected fragmentation pattern of a hydrazidehydrazone structure [13]. 3. Results and discussion The synthesized compounds 2aj and 3aj were evaluated for in vitro antimycobacterial activity against M. fortuitum ATCC 6841 using the microdilution method [2124]. Of the screened compounds, 2e and the related coupling product 3e, compounds 2i and 3g were found as active as tobramycin against the examined

strain. The remaining compounds showed less activity (64 g/mL) or no considerable effect (> 128 g/mL) as shown in table I. Compounds 2aj and 3aj were also tested for in vitro antituberculosis activity against M. tuberculosis H37Rv using the BACTEC 460 radiometric system [25, 26]. Rifampicin was used as the standard in the tests. Primary antituberculosis activity screening results of these compounds can be seen in table II. Compounds 2b, 2i, 3e and 3h were inactive against M. tuberculosis H37Rv, whereas remaining compounds showed varying degrees of inhibition in the primary screen. The compounds which exhibited < 90% inhibition in the primary screen (MIC > 12.5 g/mL) were not evaluated further. Compound 3a effecting > 90% inhibition in the primary screen at 12.5 g/mL was re-tested at lower concentrations against M. tuberculosis H37Rv to determine the actual minimum inhibitory concentration in a broth microdilution assay using alamar blue. The MIC was dened as the lowest concentration inhibiting 99% of the inoculum. Compound 3a was also tested against M. avium, a naturally drug-resistant opportunistic pathogen, using the same technique. Clarithromycin was used as the standard in this assay. However, no inhibition was observed with

1096
Table I. Antimycobacterial activity of 2aj and 3aj. Compound MIC values (g/mL) M. fortuitum ATCC 6841 64 64 64 64 64 64 64 > 128 32 32 > 128 > 128 64 32 64 > 128 Ar

2a 3a 2b 3b 2c 3c 2d 3d 2e 3e 2f 3f 2g 3g 2h 3h

primary amine functions were masked. The aim of this study was not only to screen the antimycobacterial activity of 2aj and the resulting 3aj, but also to observe the inuence of coupling 2aj with ethyl acetoacetate on this activity. As a matter of fact, series 3aj was found to be more active against M. tuberculosis H37Rv than the corresponding 2aj in most cases, as shown in table II. Even in the case of 3a, in which activity possibly arises from the presence of the 5-nitro-2-furanyl moiety, the coupling product 3a was found to be more active, with 95% inhibition (MIC = 3.13 g/mL), than the corresponding amine 2a which inhibited the growth of M. tuberculosis H37Rv by only 70%. From the above data it can be concluded that 3a could be a leading compound for further development. Structural modications, in which ketone functions present in 3aj are replaced with different moieties in order to minimize toxicological effects, will be the subject of our continuing studies. 4. Experimental protocols 4.1. Chemistry Benzocaine, 4-dimethylamino benzaldehyde and acetylacetone were purchased from Merck. All other chemicals were purchased from Sigma. Melting points were determined using a Bchi-530 melting point apparatus, uncorrected. IR spectra: (KBr, , cm1) Perkin-Elmer FTIR 1600 spectrophotometer. UV Spectra: (ethanol) max (log e) Shimadzu UV 2100S spectrophotometer. lH-NMR: (DMSO-d6, CDCl3, TMS as internal standard, chemical shifts, , in ppm) Brker AC 200 L spectrometer. EI-MS: Kratos MS-9/50 spectrometer (for compounds 3i and 3j) and ZabSpec EI+ Magnet (for compound 3a) at 70 eV. Elemental analyses: Carlo-Erba 1106 instrument. 4.1.1. Preparation of aromatic primary amines 2aj Compounds 2aj were prepared as described previously [1315]. M.p.s: for compound 2a (271274 C, lit. [14] 273274 C); for compound 2c (218 C, lit. [15] 218219 C). 4.1.2. Synthesis of the coupling products 3aj To a cooled solution of compounds 2aj (0.01 mol), in 2 mL of hydrochloric acid (37%), an ice-cold solution of 10 mL of sodium nitrite (10%) was added. The reaction mixture was then poured into a mixture of 1 mL of acetylacetone and 50 g of sodium acetate in ethanol (50%) by vigorous stirring. This mixture was allowed to stand in a refrigerator for 24 h. Precipitated solid was collected, washed with water, dried and washed with an appropriate solvent to give 3aj.

2i 3i 2j 3j Tobramycin

32 > 128 64 > 128 32

compound 3a at 12.5 g/mL, whereas clarithromycin exhibited 98% inhibition at 2 g/mL. Level II assay results of 3a are given in table III. Compound 3a was found to be active against M. tuberculosis H37Rv at 3.13 g/mL. The same compound was also tested for cytotoxicity (IC50) in VERO cells at concentrations equal to and greater than the MIC for M. tuberculosis H37Rv. The IC50 value was found at a concentration level of 0.32 g/mL for compound 3a. The selectivity index (SI = IC50/MIC) was calculated as 0.1, showing that this compound not only displayed a considerable antituberculosis activity, but also had a remarkable cytotoxicity. Compounds 2aj, whose synthesis was previously reported [13], was the starting point of our new work in order to obtain new coupling products in which aromatic

1097
Table II. Primary antituberculosis activity screen results of 2aj and 3aj. Compound MIC values (g/mL) M. tuberculosis H37Rv > 12.5 < 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 > 12.5 0.25 Inhibition (%) Ar

2a 3a 2b 3b 2c 3c 2d 3d 2e 3e 2f 3f 2g 3g 2h 3h 2i 3i 2j 3j Rifampicin

70 95 16 16 35 6 20 11 23 13 21 52 26 25 56 76 98

4.1.2.1. 2,3,4-pentanetrione-3-[4-[[(5-nitro-2-furyl)methylene]hydrazinocarbonyl]phenyl]hydrazone 3a Yield: 65%, m.p. 230 C (methanol). UV (ethanol): max (log e) = 303 (4.1611), 380 (4.5789); IR (KBr): = 3 4503 420 cm1 (H2O), 3 331 (NH), 1 672 (C=O, ketone), 1 625 (C=O, hydrazone); EI-MS (70 eV): m/z (%) = 385 (14.15) [M+], 274 (55.00), 273 (8.41), 262 (22.41),

259 (8.64), 244 (17.14), 231 (4.33), 219 (3.00), 206 (2.50), 120 (33.74), 92 (7.44), 78 (90.77), 65 (3.88), 63 (100); 1H-NMR (DMSO-d6): (ppm) = 2.46 (s, 6H, COCH3), 7.647.99 (m, 6H, Ar-H), 8.39 (s, 1H, CH=N), 11.9012.11 (s, 1H, -CONHN=CH-). Analysis: C17H15N5O6.H2O (% calculated/found) 50.62/50.21 (C); 4.24/3.54 (H); 17.36/18.15 (N).

Table III. Level II antituberculosis activity assay result of 3a. Compound 3a Rifampicin MIC value (g/mL) M. tuberculosis H37Rv 3.13 0.125 IC50 (g/mL) 0.32 SI (IC50/MIC) 0.1

1098 4.1.2.2. 2,3,4-pentanetrione-3-[4-[[(4-pyridyl)methylene]hydrazinocarbonyl]phenyl]hydrazone 3b Yield: 48%, m.p. 246 C (ethanol). UV (ethanol): max (log e) = 241 (4.1912), 295 (4.3144), 369 (4.6061); IR (KBr): = 3 200 cm1 (NH), 1 690 (C=O, ketone), 1 650 (C=O, hydrazone); 1H-NMR (CDCl3): (ppm) = 2.52 (s, 3H, COCH3), 2.62 (s, 3H, COCH3), 7.478.01 (m, 8H, Ar-H), 8.67 (s, 1H, CH=N), 9.09 (s, 1H, -CONHN=CH-), 14.58 (s, 1H, -CH=NNH-). Analysis: C18H17N5O3 (% calculated/found): 61.53/61.05 (C); 4.84/4.73 (H); 19.93/ 19.68 (N). 4.1.2.3. 2,3,4-pentanetrione-3-[4-[[(phenyl)methylene]hydrazinocarbonyl]phenyl]hydrazone 3c Yield: 75%, m.p. 199204 C (ethanol). UV (ethanol) max (log e) = 290 (4.2762), 369 (4.5297); IR (KBr): = 3 260 cm1 (NH), 1 684 (C=O, ketone), 1 642 (C=O, hydrazone); 1H-NMR (DMSO-d6): (ppm) = 2.46 (s, 6H, COCH3), 7.278.11 (m, 9H, Ar-H), 8.46 (s, 1H, CH=N), 11.74 (s, 1H, -CONHN=CH-), 13.71 (s, 1H, -CH=NNH-). Analysis: C19H18N4O3 (% calculated/ found): 65.13/64.30 (C); 5.18/5.20 (H); 15.99/15.41(N). 4.1.2.4. 2,3,4-pentanetrione-3-[4-[[(2-uorophenyl)methylene]hydrazinocarbonyl]phenyl]hydrazone 3d Yield: 67%, m.p. 214 C (ethanol). UV (ethanol): max (log e) = 233 (4.1269), 297 (4.3132), 371 (4.6158); IR (KBr): = 3 4503 400 cm1 (H2O), 3 200 (NH), 1 680 (C=O, ketone), 1 625 (C=O, hydrazone); 1H-NMR (DMSO-d6): (ppm) = 2.46 (s, 6H, COCH3), 7.288.00 (m, 8H, Ar-H), 8.71 (s, 1H, CH=N), 11.69 (s, 1H, -CONHN=CH-), 13.76 (s, 1H, -CH=NNH-). Analysis: C19H17FN4O3H2O (% calculated/found): 59.06/59.15 (C); 4.95/4.82 (H); 14.50/15.02 (N). 4.1.2.5. 2,3,4-pentanetrione-3-[4-[[(4-uorophenyl)methylene]hydrazinocarbonyl]phenyl]hydrazone 3e Yield: 41%, m.p. 218 C (ethanol). UV (ethanol): max (log e) = 285 (4.5454), 371 (4.8213); IR (KBr): = 3 5003 450 cm1 (H2O), 3 284 (NH), 1 684 (C=O, ketone), 1 625 (C=O, hydrazone); 1H-NMR (DMSO-d6): (ppm) = 2.42 (s, 6H, COCH3),7.307.94 (m, 8H, Ar-H), 8.47 (s, 1H, CH=N), 11.81 (b, H, -CONHN=CH-). Analysis: C19H17FN4O3H2O (% calculated/found): 59.06/58.57 (C); 4.95/4.88 (H), 14.50/14.47(N). 4.1.2.6. 2,3,4-pentanetrione-3-[4-[[(4-nitrophenyl)methylene]hydrazinocarbonyl]phenyl]hydrazone 3f Yield: 14%, m.p. 284 C (ethanol). UV (ethanol): max (log e) = 236 (4.1306), 265 (4.0161) sh, 372 (4.5728); IR (KBr): = 3 5003 450 cm1 (H2O), 3 330 (NH), 1 672 (C=O, ketone), 1 637 (C=O, hydrazone); 1H-NMR (DMSO-d6): (ppm) = 2.46 (s, 6H, COCH3), 7.708.31 (m, 8H, Ar-H), 8.56 (s, 1H, CH=N), 12.09 (b, 1H, -CONHN=CH-), 13.74 (b, 1H, -CH=NNH-). Analysis: C19H17N5O5H2O (% calculated/found): 56.43/56.38 (C); 4.48/4.14 (H);17.32/17.18 (N). 4.1.2.7. 2,3,4-pentanetrione-3-[4-[[(4-hydroxyphenyl)methylene]hydrazinocarbonyl]phenyl]hydrazone 3g Yield: 38%, m.p. 239 C (ethanol). UV (ethanol): max (log e) = 240 (4.4540), 275 (4.0277) sh, 372 (4.8302); IR (KBr): = 3 500 cm1 (OH), 3 295 (NH), 1 672 (C=O, ketone), 1 642 (C=O, hydrazone); 1H-NMR (DMSO-d6): (ppm) = 2.46 (s, 6H, COCH3), 6.838.00 (m, 8H, Ar-H), 8.36 (s, 1H, CH=N), 9.69 (s, 1H, Ar-OH), 13.52 (b, H, -CONHN=CH-), 14.28 (b, H, -CH=NNH-). Analysis: C19H18N4O42H2O (% calculated/found): 56.71/56.07 (C); 5.51/4.80 (H); 13.92/14.68 (N). 4.1.2.8. 2,3,4-pentanetrione-3-[4-[[(4-hydroxy-3-methoxyphenyl)methylene]hydrazinocarbonyl]phenyl]hydrazone 3h Yield: 69%, m.p. 178 C (ethanol). UV (ethanol): max (log e) = 237 (4.1504), 293 (4.1156) sh, 372 (4.4509); IR (KBr): = 3 500 cm1 (OH), 3 307 (NH), 1 678 (C=O, ketone), 1 625 (C=O, hydrazone); 1H-NMR (DMSO-d6): (ppm) = 2.46 (s, 6H, COCH3), 3.83 (s, 3H,-OCH3), 6.847.31 (m, 3H, Ar-H), 7.66 (d, 2H, o-NH, J = 8.6 Hz), 7.96 (d, 2H, m-NH, J = 8.6 Hz), 8.34 (s, 1H, CH=N), 9.48 (s, 1H, Ar-OH), 11.56 (b, 1H, -CONHN=CH-), 13.69 (b, H, -CH=NNH-). Analysis: C20H20N4O54H2O (% calculated/found): 51.28/51.40 (C); 5.98/4.76 (H); 11.96/ 11.84 (N). 4.1.2.9. 2,3,4-pentanetrione-3-[4-[[(4-dimethylaminophenyl)methylene]hydrazinocarbonyl]phenyl]hydrazone 3i Yield: 22%, m.p. 228 C (ethanol). UV (ethanol): max (log e) = 240 (4.0491), 384 (4.4508); IR (KBr): = 3 4503 400 cm1 (H2O), 3 225 (NH), 1 678 (C=O, ketone), 1 625 (C=O, hydrazone); EI-MS (70 eV): m/z (%) = 393 (66.18) [M+], 351 (1.88), 350 (2.42), 308 (1.04), 282 (14.09), 281 (1.71), 231 (100), 163 (29.73), 162 (15.75), 146 (47.29), 120 (50.04), 92 (10.39), 43 (48.83); 1 H-NMR (DMSO-d6): (ppm) = 2.46 (s, 6H, COCH3), 2.98 (s, 6H,-N(CH3)2), 6.77 (d, 2H, o-N(CH3)2, J = 8 Hz), 7.55 (d, 2H, m-N(CH3)2, J = 8 Hz), 7.68 (d, 2H, o-NH, J = 8.3 Hz), 7.97 (d, 2H, m-NH, J = 8.3 Hz), 8.32 (s, 1H, CH=N), 11.49 (s, 1H, -CONHN=CH-), 13.80 (s, 1H, -CH=NNH-). Analysis: C21H23N5O3H2O (% calculated/found): 61.31/61.39 (C); 6.08/6.08 (H);17.03/ 17.45 (N).

1099 4.1.2.10. 2,3,4-pentanetrione-3-[4-[[3-(4-dimethylaminophenyl)-2-propenylidene]hydrazinocarbonyl]phenyl]hydrazone 3j Yield: 47%, m.p. 225 C (ethanol). UV (ethanol): max (log e) = 249 (4.4045), 402 (4.8812); IR (KBr): = 3 177 cm1 (NH), 1 672 (C=O, ketone), 1 649 (C=O, hydrazone); EI-MS (70 eV): m/z (%) = 419 (7.39) [M+], 376 (3.98), 308 (25.37), 231 (46.63), 188 (9.57), 172 (30.68), 145 (4.34), 136 (7.60), 120 (100), 92 (20.39), 65 (12.30), 43 (32.25); 1H-NMR (DMSO-d6): (ppm) = 2.46 (s, 6H, COCH3), 3.23 (s, 6H, -N(CH3)2), 6.71 (d, 2H, o-N(CH3)2, J = 8.6 Hz), 6.766.96 (m, 2H, -CH=CH-), 7.43 (d, 2H, m-N(CH3)2, J = 8.6 Hz), 7.65 (d, 2H, o-NH, J = 8.6 Hz), 7.94 (d, 2H, m-NH, J = 8.6 Hz), 8.33 (d, 1H, -CH=N), 11.46 (s,1H, -CONHN=CH-), 13.72 (s, H, -CH=NNH-). Analysis C23H25N5O3H2O (% calculated/found): 63.14/63.69 (C); 6.22/5.95 (H); 16.01/15.57(N). 4.2. Microbiological procedures 4.2.1. In vitro evaluation of antimycobacterial activity against M. fortuitum Preparation of Mycobacterial inoculum required a few modications due to the difficulty of obtaining a homogenous suspension of M. fortuitum in the broth used. Four or ve colonies of M. fortuitum which were previously grown in tryptic soy agar (TSA) after 72 h of incubation at 30 C were collected by means of a swab and suspended in 4.5 mL of Mueller-Hinton broth enriched with Tween 80 (0.2%). Following the inclusion of 45 glass beads, this mixture was whirled using a vortex-mixer to ensure a good suspension. The density of this culture was then adjusted to a turbidity equivalent to that of a 0.5 McFarland standard and nally, the adjusted culture was diluted with sterile water so that, after inoculation, each microplate well had an inoculum size of 1.5 105 cfu/mL. Antimycobacterial testing of all compounds was carried out in Mueller-Hinton broth enriched with Tween 80 (0.2%) at pH 7.3. Tobramycin, which is an active antibiotic against rapidly growing mycobacteria, was selected as the standard drug. Quality control strains used in the present study were E. coli and S. aureus. Tobramycin was used at a concentration range of 40.03 g/mL against the quality control strains. The standard drug and test compounds were dissolved in water and DMSO, respectively, and were diluted with the broth used. The concentration intervals were 320.5 g/mL and 1281 g/mL for the standard drug and the test compounds, respectively. Microplate wells, containing 100 l of broth with Tobramycin or the test compounds, were then inoculated with 10 l of M. fortuitum suspension whose preparation is described above. Sheep-blooded agar was used for the purity control. After incubation for 72 h at 30 C, the last microplate well with no growth of microorganism was recorded to represent the MIC expressed in g/mL [2124]. 4.2.2. In vitro evaluation of antimycobacterial activity against M. tuberculosis H37Rv and M. avium A primary screen was conducted at 12.5 g/mL (or molar equivalent of highest molecular weight compound in a series of congeners) against M. tuberculosis H37Rv in BACTEC 12B medium using the BACTEC 460 radiometric system. Compounds effecting < 90% inhibition in the primary screen (MIC > 12.5 g/mL) were not evaluated further. Compounds demonstrating at least 90% inhibition in the primary screen were re-tested at lower concentration (MIC) in a broth microdilution assay with alamar Blue. The MIC was dened as the lowest concentration inhibiting 99% of the inoculum. These compounds were also tested against M. avium, a naturally drug-resistant opportunistic pathogen, using the same technique. Concurrent with the determination of MICs, compounds were tested for cytotoxicity (IC50) in VERO cells at concentrations equal to and greater than the MIC for M. tuberculosis H37Rv. After 72 h exposure, viability was assessed on the basis of cellular conversion of MTT into a formazan product using the Promega CellTiter 96 Non-radioactive Cell Proliferation Assay. 4.2.3. BACTEC radiometric method of susceptibility testing Inocula for susceptibility testing were either from a positive BACTEC isolation vial with a growth index (GI) of 500 or more, or a suspension of organisms isolated earlier on a conventional medium. The culture was well mixed with a syringe and 0.1 mL of a positive BACTEC culture was added to each of the vials containing the test drugs. The drug vials contained rifampicin (0.25 g/mL). A control vial was inoculated with a 1:100 dilution of the culture. A suspension equivalent to a McFarland No. 1 standard was prepared in the same manner as a BACTEC positive vial, when growth from a solid medium was used. Each vial was tested immediately on a BACTEC instrument to provide CO2 in the headspace. The vials were incubated at 37 C and tested daily with a BACTEC instrument. When the GI in the control reads at least 30, the increase in GI ( GI) from the previous day in the control was compared with that in the drug vial. The following formula was used to interpret results: GI control > GI drug = susceptible GI control < GI drug = resistant

1100 If a clear susceptibility pattern (the difference of GI of control and the drug bottle) was not seen at the time the control GI is 30 the vials were read for 1 or 2 additional days to establish a denite pattern of GI differences.
[9] [10] [11] [12] Tiwari M.M., Agarwal M., Saxena V.K., Bajpai S.K., Joshi M.M., Indian J. Pharm. Sci. 57 (3) (1995) 113116. Singh V., Srivastava V.K., Palit G., Shanker K., Arzneim.-Forsch. Drug Res. 42 (II) (1992) 993996. Pabuuoglu M.V., Rollas S., J. Pharm. Univ. Mar. 7 (1) (1991) 3949. Kkgzel S.G., Rollas S., Erdeniz H., Kiraz M., Eur. J. Med. q Chem. 34 (2) (1999) 153160. Kmrc S.G., Rollas S., lgen M., Gorrod J.W., evikbas A., Boll. q q Chim. Farmaceutico 134 (7) (1995) 375379. Casagrande C., Canova M., Ferrari G., Farmaco (Pavia), Ed. Sci. 20 (8) (1965) 544556. Chem. Abstr. 64 (1966) 9658d. Grammaticakis P., Bull. Soc. Chim. France, (1957) 12421254. Malik W.U., Garg H.G., Arora V., J. Pharm. Sci. 60 (11) (1971) 17381740. Garg H.G., Arora V., J. Pharm. Sci. 61 (1) (1972) 130132. Elguero J., Jacquier R., Tarrago G., Bull. Soc. Chim. France 9 (1966) 29812989. Ergen N., Rollas S., Demir S., zdemir F., J. Fac. Pharm. Istanbul 11 (1975) 183192. Ergen N., Rollas S., apan G., Dogan N., zger Y., Pharmazie 44 (8) (1989) 573574. Hawkins J.E., Wallace Jr. R.J., Brown B.A., in: Balows A., Hausler Jr. W.J., Herrmann K.L., Isenberg H.O., Shadami H.J. (Eds.), Manual of Clinical Microbiology, 5th ed., American Society for Microbiology, Washington, D.C., 1991, pp. 11381152. NCCLS (1990). Methods for Dilution Susceptibility Tests for Bacteria That Grow Aerobically, 2nd ed. Approved Standard. NCCLS document M7-A2. NCCLS, Villanova, Pa. NCCLS (1991). Performance Standards for Antimicrobial Susceptibility Testing. 3rd Informational Supplement. M 100-S3. NCCLS, Villanova, Pa. Swenson J.M., Thornsberry C., Silcox V.A., Antimicrob. Agents Chemother. 22 (1982) 186192. Inderleid C.B., in: Lorian V. (Ed.), Antibiotics in Laboratory Medicine, 3rd ed. Williams and Wilkins, Baltimore, 1991, p. 134. Karali N., Terzioglu N., Grsoy A., Arzneim.-Forsch. Drug Res. 48 (II) (1998) 7758763.

Acknowledgements We thank Dr Joseph A. Maddry from the Tuberculosis Antimicrobial Acquisition and Coordinating Facility (TAACF), National Institute of Allergy and Infectious Diseases Southern Research Institute, GWL Hansens Disease Center, Colorado State University, Birmingham, Alabama, USA, for the in vitro evaluation of antimycobacterial activity using M. tuberculosis H37Rv and M. avium.

[13] [14] [15] [16] [17] [18] [19] [20]

References
[21] [1] [2] [3] [4] [5] [6] [7] [8] Patel J.M., Dave M.P., Langalia N.A., Thaker K.A., J. Indian Chem. Soc. 61 (8) (1984) 718720. Jani M.K., Undavia N.K., Trivedi P.B., J. Indian Chem. Soc. 67 (7) (1990) 601602. Ergen N., Rollas S., J. Fac. Pharm. Istanbul 10 (1974) 7786. Ergen N., Rollas S., J. Fac. Pharm. Istanbul 11 (1975) 823. Rollas S., Ergen N., Bankaoglu G., Sarioglu E., J. Fac. Pharm. Istanbul 18 (1982) 107115. Kalsi R., Pande K., Bhalla T.N., Barthwal J.P., Gupta G.P., Parmar S.S., J. Pharm. Sci. 79 (4) (1990) 317320. Pande K., Kalsi R., Bhalla T.N., Barthwal J.P., Indian J. Pharm. Sci. 51 (1) (1989) 1821. Kalsi R., Pande K., Barthwal J.P., Indian J. Chem. Sect. B 27 B (2) (1988) 197199.

[22]

[23]

[24] [25] [26]

Eur. J. Med. Chem. 34 (1999) 11011108 1999 ditions scientiques et mdicales Elsevier SAS. All rights reserved

1101

Laboratory note

Synthesis and pharmacological properties of novel benzamide derivatives acting as ligands to the 5-hydroxytryptamine 4 (5-HT4) receptor
Katsuhiko Itoh*, Hideo Tomozane, Hidetoshi Hakira, Shuji Sonda, Kiyoshi Asano, Masatake Fujimura, Noriko Sato, Keiichiro Haga, Takeshi Kawakita
Research Laboratories, Yoshitomi Pharmaceutical Industries Ltd., 955 Koiwai, Yoshitomi-cho, Chikujo-gun, Fukuoka 871-8550, Japan (Received 12 April 1999; revised 15 June 1999; accepted 16 June 1999)

Abstract A series of 4-amino-5-chloro-2-methoxy-N-(1-substituted piperidin-4-ylmethyl)benzamides was synthesized as novel gastroprokinetic agents. The affinity of these compounds for the 5-hydroxytryptamine 4 (5-HT4) receptor was evaluated. Among these compounds, 4-amino-5-chloro-2-methoxy-N-[1-[5-(1-methylindol-3-ylcarbonylamino)pentyl]piperidin-4-ylmethyl]benzamide (3f, Y-34959) showed a higher affinity for the 5-HT4 receptor (Ki = 0.30 nmol/L) than for other receptors, and was conrmed to be a potent 5-HT4 receptor agonist having contractile effects in the isolated guinea-pig ascending colon (EC50 = 1.2 nmol/L). In dogs, compound 3f increased gastroprokinetic motility of both the gastric antrum and the ascending colon. In addition, this effect on the colon was inhibited by azasetron, a selective 5-HT3 receptor antagonist, demonstrating that the effect of gastroprokinetic agents having 5-HT3 receptor antagonism on the colon were reduced compared with that of selective 5-HT4 receptor agonists. 1999 ditions scientiques et mdicales Elsevier SAS 5-HT4 receptor agonist / 5-HT4 receptor agonism / 5-HT3 receptor antagonism / N-[1-[5-(1-methylindol-3-ylcarbonylamino)pentyl]piperidin-4-ylmethyl]benzamide / gastrointestinal motility

1. Introduction Activation of the 5-hydroxytryptamine 4 (5-HT4) receptor mediates a wide variety of effects in the central and peripheral nervous systems [1]. Benzamides (metoclopramide [2], cisapride [3], etc.) are used clinically as gastrointestinal motility stimulants. Recently, this gastroprokinetic effect is thought to be mediated by 5-HT4 receptor agonism [4]. In addition, we also found the contractile potency in the isolated guinea-pig ascending colon and the binding affinity for the 5-HT4 receptor were well correlated among these benzamides [5]. On the other hand, these benzamides show antagonistic activity against various receptors such as dopamine D2, 5-HT2 and 5-HT3 receptors [69], and the antagonism should reduce the gastroprokinetic effect of 5-HT4 receptor agonism and/or cause unfavourable side effects. For example, it is well known that the 5-HT3 receptor

antagonists induce constipation, and D2 receptor antagonists cause central nervous system effects such as depression and extrapyramidal syndrome [10]. Therefore, the search for selective 5-HT4 receptor agonists would be an important goal to develop novel useful gastrointestinal motility stimulants. In the course of our synthetic studies on 5-HT4 receptor agonists, we have found that the essential framework for selective 5-HT4 receptor agonism was 4-amino-5-chloro-2-methoxy-N-(piperidin-4-ylmethyl)benzamide with the polar side chain at the 1- position on the piperidine ring (compound 1 has a methylsulfonylaminoethyl group as a polar side chain as shown in gure 1) [11]. Modications of the polar side chain of compound 1 led to the discovery of a novel gastroprokinetic agent, 4-amino-5-chloro-2-methoxy-N-[1-[5(1-methylindol-3-ylcarbonylamino)pentyl]piperidin-4ylmethyl]benzamide (3f, Y-34959). Compound 3f was conrmed to be a selective 5-HT4 receptor agonist. Herein, we describe the synthesis and the pharmacological data of the novel selective 5-HT4 receptor agonist.

*Correspondence and reprints

1102 of [3H]GR113808 to the receptor. The agonistic activity of the candidate was evaluated as the contractile ability of the ascending colon in guinea-pigs. In in vivo studies, the gastrointestinal motility index of the candidate in conscious dogs was measured. Binding affinities for the 5-HT4 receptor of 1, 3a3f, and 5-HT are listed in table I, where 5-HT is the reference compound. Phenylsulfonylamine derivative 3a showed a higher affinity for the 5-HT4 receptor than methylsulfonylamine derivative 1. Next, we investigated the inuence of other polar side chains. A benzamide group was selected, regarding the group of side chains, because an amide group is considered to be a bioisoster of the sulfonylamine group [13] and the introduction of the amide group is synthetically easier. Among the benzamide derivatives (3b3d), an N-pentylbenzamide derivative 3d showed the highest affinity for the 5-HT4 receptor indicating that a bulky group is needed as a substituent on the piperidine ring at the 1- position and that a large pocket for the substituent should exist at the receptor. This consideration is supported by the Lpez-Rodrguez report [14], which described a 5-HT4 receptor mapping using an active analogue approach. Next, the phenyl group was replaced by a 1methylindole group. N-butyl-1-methylindole-3-carboxamide derivative 3e showed a higher affinity than the corresponding compound 3c. N-pentyl-1-methylindole-3carboxamide derivative 3f showed a subnanomolar affinity. These results suggest that the 1-methylindole-3carboxamide group may interact with some amino acid

Figure 1. Chemical structure of compound 1.

2. Chemistry The general synthetic procedure used in this study is illustrated in gure 2. 1-Substituted-4-(tert-butoxycarbonylaminomethyl)piperidine derivatives (2a2f) [12] were deprotected with hydrogen chloride in dioxane, affording the corresponding amines, which were coupled with 4-amino-5-chloro-2-methoxybenzoic acid using 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide hydrochloride (WSC HCl) and 1-hydroxybenzotriazole (HOBt) in the presence of triethylamine (TEA) to afford target benzamides (3a3f), whose data for physicochemical properties and 5-HT4 receptor affinities are listed in table I. 3. Pharmacological data and discussion The affinity of compounds 3a3f for the 5-HT4 receptor was determined as their ability to inhibit the binding

Figure 2. Synthesis of benzamides 3a3f. a: HCl in 1,4-dioxane. b: TEA, 4-amino-5-chloro-2-methoxybenzoic acid, WSC HCl, HOBt, DMF.

1103
Table I. Physicochemical properties and 5-HT4 receptor affinities of compounds 1 and 3a3f.

Compound 1

R CH3

X SO2

n 2

M.p. (C) 177178

Mass (m/z) 418

Formula C17H27ClN4O4S 2 C2H2O4c

Binding afnity, Ki (nmol/L)b 9.6

3a

SO2

Amorphous solid

480

C22H29ClN4O4S 3/4H2O

2.6

3b

CO

Amorphous solid

458

C24H31ClN4O3 H2O

9.4

3c

CO

165168

472

C25H33ClN4O3

4.0

3d

CO

122124

486

C26H35ClN4O3 1/2H2O

1.7

3e

CO

Amorphous solid

525

C28H36ClN5O3 H2O

1.6

3f

CO

161163

539

C29H38ClN5O3 HCl

0.3

5-HT
a

130

Elementary analyses were performed for C, H and N and were within 0.4% of the calculated values for formulae shown. bEach value is the mean from triplicate assays in a single experiment. coxalate.

residues in the pocket of the 5-HT4 receptor. Compound 3f was devoid of signicant affinities for other receptors, and produced concentration dependent contractions of the guinea-pig ascending colon (table II). Therefore, compound 3f was conrmed to be a selective 5-HT4 receptor agonist. Finally, we examined effects of compound 3f on gastrointestinal motility in conscious dogs in a postprandial state. When compound 3f was given intravenously to dogs at 0.01 mg/kg, the motility of the gastric antrum and ascending colon was rapidly enhanced, as shown in

gure 3, and the motility index of the gastric antrum and ascending colon increased. In addition, we tested the inuence of azasetron [15], a selective 5-HT3 receptor antagonist, on the stimulated gastrointestinal motility by a selective 5-HT4 receptor agonist compound 3f. Azasetron inhibited the increase in the colonic motility index caused by compound 3f. In contrast, azasetron did not inuence the increase of the antral motility index (gure 4). These results demonstrated that gastroprokinetic agents which have 5-HT3 receptor antagonism would be less effective in increasing the motility of the ascending

1104
Table II. Binding proles and 5-HT4 receptor agonistic activity of compound 3f. Binding affinitya, D2 rat striatum > 1 000b
a

Ki (nmol/L) 5-HT2 [3H]ketanserin 110 5-HT3 5-HT4 guinea-pig striatum

5-HT1A rat hippocampus > 1 000b

5-HT4 receptor agonistic activity EC50 (nmol/L)c Maximal response (%)d 1.2 0.3 6.3 0.1

rat cerebral cortex rat striatum > 1 000b

[3H]spiperone [3H]8-OH-DPAT

[3H]granisetron [3H]GR113808 0.3

Each value is the mean from triplicate assays in a single experiment. bIC50 value. cEC50 values (mean SE) was determined by linear regression. dA percentage (mean SE) of the contraction caused by methacholine (30 mol/L).

Figure 3. Typical tracings of the effect of compound 3f on gastrointestinal motility in conscious dogs in a postprandial state.

colon than selective 5-HT4 receptor agonists and that selective 5-HT4 receptor agonists should be able to enhance both upper and lower gastrointestinal motility. 4. Conclusion We described the synthesis, and affinity for the 5-HT4 receptor, of a series of benzamides. Among them, 4-amino-5-chloro-2-methoxy-N-[1-[5-(1-methylindol-3-

ylcarbonylamino)pentyl]piperidin-4-ylmethyl]benzamide (3f, Y-34959) was found to be a selective 5-HT4 receptor agonist. Compound 3f increased the gastrointestinal motility of both the gastric antrum and the ascending colon in conscious dogs in a postprandial state. Azasetron, a selective 5-HT3 receptor antagonist, inhibited the increase of the colonic motility index caused by compound 3f without inuencing the increase of the gastric motility index. Based on our results, we proposed that the selec-

1105

Figure 4. Effect of azasetron on the increase in the gastrointestinal motility of gastric antrum and ascending colon caused by compound 3f.

tive 5-HT4 receptor agonists were novel gastrointestinal motility stimulants which can enhance both upper and lower gastrointestinal motility with few side effects. 5. Experimental protocols 5.1. Chemistry All melting points were measured in open capillaries and are uncorrected. Proton nuclear magnetic resonance (1H-NMR) spectra were recorded on JEOL JNM-EX270 spectrometers and chemical shifts are expressed in ppm with tetramethylsilane (TMS) as an internal standard. Signal multiplicities are represented by s (singlet), d (doublet), t (triplet), q (quartet), br-s (broad singlet) and m (multiplet). Mass spectra (MS) were taken on JEOL JMS-O1SG spectrometers. Elementary analysis was performed for C, H, N and were within 0.4% of the calculated values. Silica-gel plates (Merck F254) and silica gel 60 (Merck, 70230 mesh) were used for analytical and preparative column chromatography, respectively. 5.1.1. General procedure for the preparation of 3a3f 5.1.1.1. 4-Amino-5-chloro-2-methoxy-N-[1-[5-(1-methylindol-3-ylcarbonylamino)pentyl]piperidin-4-ylmethyl] benzamide hydrochloride 3f A mixture of 2f (5.7 g, 13 mmol) and 4 mol/L hydrogen chloride in 1,4-dioxane (Aldrich) (60 mL) was stirred

under ice-cooling for 1 h. After evaporation, the residue was washed with CHCl3, and the resulting compound was dissolved in 80 mL of dimethylformamide (DMF). To the solution were added triethylamine (TEA) (5.7 mL, 41 mmol), 4-amino-5-chloro-2-methoxybenzoic acid (2.6 g, 13 mmol) and 1-hydroxybenzotriazole (HOBt) (1.9 g, 14 mmol). The reaction mixture was stirred at room temperature for 1 h and then 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide hydrochloride (WSC) (2.7 g, 14 mmol) was added under ice-cooling. Stirring was continued overnight at room temperature. After evaporation, 5% aqueous sodium bicarbonate was added to the residue and extracted with CHCl3. The extract was washed with brine and dried over anhydrous magnesium sulfate. After evaporation in vacuo, the residue was chromatographed on silica gel eluting with CHCl3MeOHNH4OH (100:10:1) and treated with an alcoholic solution of hydrogen chloride. The precipitates were collected and recrystallized from ethanol to give 3f (3.1 g, 43%); 1H-NMR (DMSO-d6) : 1.251.41 (2H, m), 1.471.62 (4H, m), 1.671.95 (4H, m), 2.802.93 (2H, m), 2.96 (2H, br-s), 3.17 (2H, br-s), 3.26 (2H, dd, J = 5.9, 13 Hz), 3.403.46 (2H, m), 3.82 (3H, s, CH3N), 3.83 (3H, s, CH3O), 5.95 (2H, s, Ar-NH2), 6.50 (1H, s, Ar-3-H), 7.13 (1H, t, J = 6.6 Hz, ind-5-H), 7.20 (1H, t, J = 6.6 Hz, ind-6-H), 7.46 (1H, d, J = 7.9 Hz, ind-7-H), 7.66 (1H, s, ind-2H), 7.93 (1H, t, J = 5.3 Hz, CONHCH2), 8.02 (1H, s, Ar-6-H), 8.14 (1H, d, J = 7.9 Hz, ind-4-H), 10.20 (1H, br-s, NHCO).

1106 5.1.1.2. 4-Amino-5-chloro-2-methoxy-N-[1-[2-(phenylsulfonyl)aminoethyl]piperidin-4-yl]benzamide 3a Similarly to 3f, 3a was prepared starting from 2a (1.6 g, 3.9 mmol), hydrogen chloride in 1,4-dioxane (11 mL), TEA (1.1 mL, 12 mmol), 4-amino-5-chloro-2methoxybenzoic acid (0.78 g, 3.9 mmol), HOBt (0.58 g, 4.3 mmol), DMF (20 mL), and WSC (0.82 g, 4.3 mmol) to give 3a (0.55 g, 29%); 1H-NMR (CDCl3) : 1.20 (2H, dd, J = 11, 15 Hz), 1.531.67 (3H, m), 1.86 (2H, t, J = 9.9 Hz), 2.36 (2H, t, J = 5.9Hz), 2.59 (2H, d, J = 11 Hz), 2.98 (2H, t, J = 6.6 Hz), 3.31 (2H, t, J = 6.0 Hz), 3.92 (3H, s, CH3O), 4.39 (2H, br-s, Ar-NH2), 6.31 (1H, s, Ar-3-H), 7.487.60 (3H, m, Ar-H), 7.60 (1H, t, J = 6.0 Hz, CONHCH2), 7.84 (2H, dd, J = 2.0, 6.6 Hz, Ar-H), 8.11 (1H, s, Ar-6-H). 5.1.1.3. 4-Amino-N-[1-(3-benzoylaminopropylpiperidin)4-ylmethyl]-5-chloro-2-methoxybenzamide 3b Similarly to 3f, 3b was prepared starting from 2b (1.0 g, 2.7 mmol), hydrogen chloride in 1,4-dioxane (11 mL), TEA (1.1 mL, 8.1 mmol), 4-amino-5-chloro-2methoxybenzoic acid (0.54 g, 2.7 mmol), HOBt (0.41 g, 3.0 mmol), DMF (15 mL), and WSC (0.58 g, 3.0 mmol) to give 3b (0.10 g, 11%), 1H-NMR (CDCl3) : 1.44 (2H, dd, J = 12, 22 Hz), 1.72189 (5H, m), 2.13 (2H, t, J = 11 Hz), 2.68 (2H, t, J = 6.0 Hz), 3.15 (2H, d, J = 12 Hz), 3.32 (2H, t, J = 6.6 Hz), 3.56 (2H, dd, J = 3.4, 12 Hz), 3.86 (3H, s, CH3O), 4.42 (2H, s, Ar-NH2), 6.29 (1H, s, Ar-3-H), 7.377.44 (3H, m, Ar-H), 7.74 (1H, t, J = 6.0 Hz, CONHCH2), 7.84 (2H, dd, J = 2.0, 8.0 Hz, Ar-H), 8.10 (1H, s, Ar-6-H), 8.28 (1H, br-s, CONH). 5.1.1.4. 4-Amino-N-[1-(4-benzoylaminobutyl)piperidin4-ylmethyl]-5-chloro-2-methoxybenzamide 3c Similarly to 3f, 3c was prepared starting from 2c (0.86 g, 2.2 mmol), hydrogen chloride in 1,4-dioxane (9 mL), TEA (0.90 mL, 6.6 mmol), 4-amino-5-chloro-2methoxybenzoic acid (0.44 g, 2.2 mmol), HOBt (0.32 g, 2.4 mmol), DMF (15 mL), and WSC (0.46 g, 2.4 mmol). The resulting solid was recrystallized from ethanolisopropanol to give 3c (0.20 g, 19%); 1H-NMR (CDCl3CD3OD) : 1.321.56 (4H, m), 1.641.89 (5H, m), 2.13 (2H, t, J = 11 Hz), 2.68 (2H, t, J = 6.0 Hz), 3.21 (2H, d, J = 12 Hz), 3.39 (2H, t, J = 6.6 Hz), 3.43 (2H, t, J = 6.6 Hz), 3.91 (3H, s, CH3O), 4.45 (2H, br-s, Ar-NH2), 6.35 (1H, s, Ar-3-H), 7.337.42 (3H, m, Ar-H), 7.74 (1H, t, J = 6.0 Hz, CONHCH2), 7.78 (2H, dd, J = 2.0, 8.0 Hz, Ar-H), 8.01 (1H, s, Ar-6-H). 5.1.1.5. 4-Amino-N-[1-(5-benzoylaminopentyl)piperidin4-ylmethyl]-5-chloro-2-methoxybenzamide 3d Similarly to 3f, 3d was prepared starting from 2d (1.7 g, 4.2 mmol), hydrogen chloride in 1,4-dioxane (17 mL), TEA (1.7 mL, 13 mmol), 4-amino-5-chloro-2methoxybenzoic acid (0.84 g, 4.2 mmol), HOBt (0.62 g, 4.6 mmol), DMF (20 mL), and WSC (0.88 g, 4.6 mmol). The resulting solid was recrystallized from ethyl acetate to give 3d (1.4 g, 72%); 1H-NMR (CDCl3) : 1.311.49 (4H, m), 1.501.85 (7H, m), 1.91 (2H, t, J = 12 Hz), 2.32 (2H, t, J = 7.8 Hz), 2.92 (2H, d, J = 12 Hz), 3.30 (2H, t, J = 6.0 Hz), 3.34 (2H, dd, J = 7.3, 13 Hz), 3.88 (3H, s, CH3O), 4.46 (2H, s, Ar-NH2), 6.30 (1H, s, Ar-3-H), 6.36, (1H, br-s, CONH), 7.337.42 (3H, m, Ar-H), 7.70 (1H, br-s, CONHCH2), 7.77 (2H, dd, J = 7.3 Hz, Ar-H), 8.09 (1H, s, Ar-6-H). 5.1.1.6. 4-Amino-5-chloro-2-methoxy-N-[1-[4-(1-methylindol-3-ylcarbonylamino)butyl]piperidin-4-ylmethyl] benzamide 3e Similarly to 3f, 3e was prepared starting from 2e (1.4 g, 3.2 mmol), hydrogen chloride in 1,4-dioxane (14 mL), TEA (1.3 mL, 9.6 mmol), 4-amino-5-chloro-2methoxybenzoic acid (0.66 g, 3.2 mmol), HOBt (0.47 g, 3.5 mmol), DMF (30 mL), and WSC (0.67 g, 3.5 mmol) to give 3e (0.63 g, 38%); 1H-NMR (CDCl3) : 1.24 (2H, dd, J = 11, 15 Hz), 1.301.72 (9H, m), 1.91 (2H, t, J = 9.9 Hz), 2.38 (2H, t, J = 6.6 Hz), 2.93 (2H, d, J = 11 Hz), 3.29 (2H, t, J = 6.6 Hz), 3.50 (2H, dd, J = 6.0, 12 Hz), 3.81 (3H, s, CH3N), 3.87 (3H, s, CH3O), 4.35 (2H, br-s, Ar-NH2), 6.20 (1H, br-s, CONH), 6.26 (1H, s, Ar-3-H), 7.217.34 (2H, m, ind-5,6-H), 7.63 (1H, s, ind-2-H), 7.70 (1H, t, J = 3.2 Hz, CONHCH2), 7.89 (1H, d, J = 8.0 Hz, ind-7-H), 7.92 (1H, d, J = 8.0 Hz, ind-4-H), 8.10 (1H, s, Ar-6-H). 5.2. Pharmacology 5.2.1. 5-HT4 receptor binding Male Hartley guinea-pigs (Japan SLC, Ltd., Shizuoka, Japan) were killed by cervical dislocation and the striatum was separated from each brain. The striatum was homogenized in 15 volumes of 50 mmol/L ice-cold HEPES buffer (pH 7.4) with Polytron PT-10 and then centrifuged at 35 000 g for 20 min. The resulting pellet was resuspended in the HEPES buffer and nally diluted to the appropriate concentration for assay (6 mg wet weight per assay tube). This suspension was used as the tissue preparation. Assay tubes contained 50 mL of HEPES buffer or a solution of the test agents, 50 mL solution of [3H]GR113808 (Amersham International, UK) to give a nal concentration of 0.1 nmol/L and 900 mL of tissue preparation. Each tube was incubated for 30 min at 37 C and the reaction was terminated by rapid ltration through a Whatmann GF/B lter (presoaked in 0.01% v/v polyethyleneimine) followed by washing with 1 4 mL of ice-cold HEPES buffer. Then

1107 the lter was placed in 3 mL of scintillant and the radioactivity was determined by scintillation counting in a Beckman model LS3801 scintillation counter. Non specic binding was dened in the presence of unlabelled GR113808 to give a nal concentration of 1 mol/L. The IC50 value was determined by non-linear regression of the displacement curve, and the Ki value was calculated according to the formula (Ki = IC50/(1 + L/Kd)), where L is the concentration of radioligand and Kd is the dissociation constant of the radioligand. 5.2.2. Binding to other receptors The binding studies of D2 [16], 5-HT1A [17], 5-HT2 [18], and 5-HT3 receptor [19] were carried out according to the previously published methods. 5.2.3. 5-HT4 receptor agonism Four male Hartley guinea-pigs (Japan SLC, Ltd., Shizuoka, Japan) were killed by cervical dislocation and the ascending colon was removed. The longitudinal muscle layer was separated from the underlying circular muscle. Each muscle strip preparation of about 2.5 cm was individually mounted vertically for isotonic measurement into a tissue bath containing 10 mL Tyrode solution. This solution was kept at 37 C and gassed with 95% O2, 5% CO2. The strips were subjected to a preload of 1 g and allowed to stabilize for 20 min. After stabilization, the response of the longitudinal muscle to 30 mol/L methacholine was measured. Agonist concentration-effect curves were constructed using sequential dosing, leaving 15 min between doses. A 15 min dosing cycle was required to prevent desensitization. The agonist was left in contact with a preparation until the response had reached a maximum, the preparation was then washed. Forty minutes was left between the determination of concentration-effect curves. GR113808 (10 nmol/L) was incubated for 10 min before repeating the agonist concentration-effect curves. After each determination of a concentration-effect curve, 30 mol/L of methacholine was added to the tissue bath again. All responses were expressed as a percentage of the mean of the two contractions induced by 30 mol/L methacholine. The EC50 value, the concentration causing 50% of the maximal response, was determined by linear regression analysis. 5.2.4. Gastrointestinal motility Three mongrel dogs were used. Under pentobarbital anaesthesia, the abdomen was opened, and strain gauge transducers (F-12SSH, Star Medical, Tokyo, Japan) were sutured onto the serosa of the stomach and colon in a manner to detect circular muscle contraction. The transducers were placed at the greater curvature of the gastric antrum, 5 cm proximal to the pyloric ring, and at the colon, 10 cm distal to the ileo-colic junction. The animals were then allowed more than 2 weeks to recover from the surgery. To measure the gastrointestinal motor activity, the signals from the gastric antrum and colon were recorded on a recording system (ESC-2000, Star medical, Tokyo, Japan) every 100 ms by a telemetry system (DAS-800T, Star medical, Tokyo, Japan). The motility index of contractile activity, shown as an area under contraction wave, was calculated both in the gastric antrum and in the colon by means of a program (Peaks, AD Instruments, Australia). Measurements of gastrointestinal motility were performed in a postprandial state. Approximately 150 min after feeding of a dry type meal (200 g), azasetron (0.1 mg/kg) or vehicle was given to the dog intravenously. Ten minutes later, test compound was injected intravenously. The motility index during 10 min before and after injection of the test compound was determined, and data were expressed as a difference between the motility index before and that after the injection of the test compound. Investigation consisted of six experiments in three dogs. Statistical analysis of data was performed by means of the Dunnett method.

Acknowledgements We thank Mrs F. Matsugaki for some of the biological results. We also thank Dr M. Terasawa and Dr K. Adachi for helpful discussion.

References
[1] [2] [3] [4] [5] [6] [7] [8] [9] Eglen R.M., Wong E.F., Dumuis A., Bockaert J., Trends Pharmacol. Sci. 16 (1995) 391398. Harrington R.A., Hamilton C.W., Brogden R.N., Linkewich J.A., Romankiewicz J.A., Heel R.C., Drugs 25 (1983) 451494. McCallum R.W., Prakash C., Campoli-Richards D.M., Goa K.L., Drugs 36 (1988) 652681. Bockaert J., Fozard J.R., Dumuis A., Clarke D.E., Trends Pharmacol. Sci. 13 (1992) 141145. Sato N., Asano K., Haga K., Arita M., Jpn. Pharmacol. Ther. 25 (1997) 383387. Turconi M., Schiantarelli P., Borsini F., Rizzi C.A., Ladinsky H., Donetti A., Drugs Future 16 (1991) 10111026. Cushing D.J., Cohen M.L., J. Pharmacol. Exp. Ther. 264 (1993) 193199. Yoshida N., Omoya H., Furukawa K., Ito T., Karasawa T., Arch. Int. Pharmacodyn. Ther. 300 (1989) 5167. Karasawa T., Yoshida N., Furukawa K., Omoya H., Ito T., Eur. J. Pharmacol. 183 (1990) 2181.

1108
[10] [11] Fernandez A.G., Roberts D.J., Drugs Future 16 (1991) 885892. Itoh K., Kanzaki K., Ikebe T., Kuroita T., Tomozane H., Sonda S., Sato N., Haga K., Kawakita T., Eur. J. Med. Chem. 34 (1999) 977989. Kawakita T., Kuroita T., Murozono T., Hakira H., WO 95/26953 Chem. Abstr. 124 (1996) 14913. Thornber C.W., Chem. Soc. Rev. 8 (1979) 563580. Lpez-Rodrguez M.L., Morcillo M.J., Benha B., Rosado M.L., J. Comput. -Aided Mol. Des. 11 (1997) 589599. [15] [16] [17] [18] [19] Kawakita T., Kuroita T., Yasumoto M., Sano M., Inaba K., Fukuda T., Tahara T., Chem. Pharm. Bull. 40 (1992) 624630. Creese I., Schneider R., Snyder S.H., Eur. J. Pharmacol. 46 (1977) 377381. Hall M.D., El Mestikawy S., Emerit M.B., Pichat L., Hamon M., Gozlan H., J. Neurochem. 44 (1985) 16851696. Leysen J.E., Niemegeers C.E., Van Nueten J.M., Laduron P.M., Mol. Pharmacol. 21 (1982) 301314. Nelson D.R., Thomas D.R., Biochem. Pharmacol. 38 (1989) 16931695.

[12] [13] [14]