Clinical and Experimental Pharmacology and Physiology (1999) 26, 596600

BRIEF REVIEW

HUMAN ERYTHROCYTE BUT NOT BRAIN ACETYLCHOLINESTERASE HYDROLYSES HEROIN TO MORPHINE

Asher Y Salmon,1 Zafrir Goren,2 Yaniv Avissar2 and Hermona Soreq1
Department of Biological Chemistry, Institute of Life Sciences, Hebrew University of Jerusalem, Givat Ram and 2Analytical Chemistry Laboratory, Division of Identication and Forensic Science, Israel Police, Jerusalem, Israel
1

SUMMARY
1. In human blood, heroin is rapidly hydrolysed by sequential deacylation of two ester bonds to yield rst 6-monoacetylmorphine (6-MAM), then morphine. 2. Serum butyrylcholinesterase (BuChE) hydrolyses heroin to 6-MAM with a catalytic efciency of 4.5/min per mol/L, but does not proceed to produce morphine. 3. In vitro, human erythrocyte acetylcholinesterase (AChE) hydrolyses heroin to 6-MAM, with a catalytic efciency of 0.5/min per mol/L under rst-order kinetics. Moreover, erythrocyte AChE, but not BuChE is capable of further hydrolysing 6-MAM to morphine, albeit at a considerably slower rate. 4. Both hydrolysis steps by erythrocyte AChE were totally blocked by the selective AChE inhibitor BW284c51 but were not blocked by the BuChE-specic inhibitor, iso-OMPA (tetraisopropylpyrophosphoramide). 5. The brain synaptic form of AChE, which differs from the erythrocyte enzyme in its C-terminus, was incapable of hydrolysing heroin. 6. Heroin suppressed substrate hydrolysis by antibodyimmobilized erythrocyte but not by brain AChE. 7. These ndings reveal a new metabolic role for erythrocyte AChE, the biological function of which is as yet unexplained, and demonstrate distinct biochemical properties for the two AChE variants, which were previously considered catalytically indistinguishable. Key words: acetylcholinesterase, brain, butyrylcholinesterase, erythrocytes, heroin, 6-monoacetylmorphine, morphine, opiates.

INTRODUCTION
Heroin (3,6-diacetylmorphine) is a worldwide leading cause of morbidity and mortality due to drug abuse. In certain countries, heroin is legally used for treating chronic pain and for other medical purposes.1 In the human bloodstream, heroin is rapidly hydrolysed to 6-monoacetylmorphine (6-MAM) and then into morphine.2 The serum enzyme butyrylcholinesterase (BuChE, acylcholine acylhydrolase, EC 3.1.1.8) has been shown to perform the rst but not the second step in this process.3,4 Moreover, physiological studies have demonstrated that heroin degradation occurs primarily in the micro-environment of erythrocytes, where BuChE is absent, but not in the serum.5,6 Two major esterases and a few non-specic esterases are present in human erythrocytes. These are arylesterase (EC.3.1.1.2) and acetylcholinesterase (AChE, acetylcholine acetylhydrolase, EC.3.1.1.7). Lockridge et al. demonstrated that serum BuChE, but not erythrocyte arylesterase, is capable of hydrolysing heroin.3 This left two questions unanswered: rst, can erythrocyte AChE hydrolyse heroin? and second, can it degrade it completely to morphine? Acetylcholinesterase, whose main catalytic activity is to hydrolyse the neurotransmitter acetylcholine, appears in three C-terminally distinct isoforms, derived from alternatively spliced AChE mRNA species.7 Therefore, nervous system AChE differs from red blood cell AChE in its C-terminal peptide. Because both 6-MAM and morphine, but not heroin, are physiologically active in the mammalian brain,8 another issue arises: whether the brain and red blood cell variants of AChE differ in their capacity to hydrolyse heroin and/or 6-MAM. To address these issues, we combined the use of highly puried AChE isoforms with high-pressure liquid chromatography (HPLC) for testing heroin degradation kinetics by the various isoforms of human AChE.

MATERIALS AND METHODS

Human AChE from erythrocytes (Sigma type XIII); human serum BuChE, human recombinant AChE (brain form), chromatographic standards (heroin hydrochloride, morphine sulphate, 6-MAM), 1,5-bis 4-allyldimethylammoniumphenyl pentan-3-one dibromide, (BW284c51), tetraisopropylpyrophosphoramide (iso-OMPA), acetylthiocholine (ATCh), butyrylthiocholine (BTCh) and other basic chemicals were all purchased from Sigma Chemical Co. (St Louis, MO, USA). Morphine-HCl and codeine sulphate were purchased from Teva Co. (Kfar-Saba, Israel). High-performance liquid chromatography reagents were purchased from Cavlo Ebra (Paris, France). Monoclonal mouse antihuman BuChE (no. 534) and antihuman AChE (no. 1011)7 were gratefully received from Dr B. Norgaard-Pedersen

Correspondence: Dr Hermona Soreq, Department of Biological Chemistry, Institute of Life Sciences, Hebrew University of Jerusalem, Givat Ram, Jerusalem 91904, Israel. Email: <soreq@shum.huji.ac.il> Presented as an invited lecture at the Annual Meeting of the Australian Physiological and Pharmacological Society (APPS), Brisbane, September October 1998. Received 5 January 1999; accepted 14 April 1999.

Heroin hydrolysis by human acetylcholinesterase

(Copenhagen, Denmark). Heroin HCl, 6-monoacetylmorphine (6-MAM) and 3-monoacetylmorphine (3-MAM) were puried and standardized prior to experimentation as previously described.

597

Opiates hydrolysis by ChEs

Opiates (85 mol/L) were incubated under agitation with 0.5 units per mL cholinesterase, in sodium phosphate buffer (0.1 mol/L, pH 7.4) at 37C. One unit of enzyme will hydrolyse 1 mol per min of BTCh or ATCh at pH 8.0 at 37C. Hydrolysis was terminated by removing 400 L aliquots from the reaction mixture at timed intervals, adding 17 L of 1 mmol/L H2SO4 and 200 L methanol and placing the mixture on ice, where the low temperature and pH suppressed both enzymatic and spontaneous hydrolysis.9

High-performance liquid chromatography analysis

A total of 10 nmol/L acetyl-codeine was added to each stopped reaction mixture as a high-powered liquid chromatography (HPLC) internal standard. The resultant mixtures were ltered through a 0.45 m membrane and injected into a Waters 717 auto-sampler (Milford, MA, USA) at 4C. Duplicates of 10 L samples were deposited onto MerckTM RP (Darmstadt, Germany) Select B-125 4 (5 m particle size) HPLC columns at 27C. The mobile phase gradient elution contained 2 mmol/L sulphuric acid in water, in acetonitrile and in methanol with a ow rate of 1.8 mL/min. Chromatograms were recorded at 230 nm with a Waters 996 photodiode array detector. Absorption spectra for each peak were recorded between 200 and 350 nm (1.2 nm resolution). Duplicate measurements of peak areas were processed by Millenium 2.1 data analysis (Waters). Michaelis-Menten constant (Km) values for heroin hydrolysis by ChEs was derived from double reciprocal Lineweaver-Burk plots over a substrate range of 324000 mol/L.

Immobilization of cholinesterases
Mouse antihuman serum BuChE or monoclonal antihuman AChE antibodies10 were absorbed to microtitre plates (Nunc, Roskilde, Denmark) at 0.5 g/mL in 0.1 mol/L carbonate buffer, pH 9.6, for at least 4 h at room temperature. Plates were then washed three times in phosphate-buffered saline (PBS)-T buffer (144 mmol/L NaCl, 20 mmol/L Na phosphate, pH 7.4, 0.05% Tween-20). Free binding sites on the surface of the microtitre plate wells were blocked with PBS-T for 1 h at 37C. Cholinesterases were then added at a concentration of 100 mIU/mL in PBS-T for at least 3 h at room temperature with agitation. Plates were washed three times with PBS-T before use.

Inhibition of cholinesterase activity by heroin

Heroin, 6-MAM, morphine and codeine at 650 mol/L in 180 L of Ellmans solution11 were added to antibody-immobilized enzymes12 for 5 min at room temperature, followed by three washes with PBS-T to remove free opiates. Rates of ATCh (1 mmol/L) or BTCh (10 mmol/L) hydrolysis were determined spectrophotometrically.11,12 Inhibition constant (Ki) values were determined by measuring cholinesterase activity in a solution containing 305000 mol/L concentrations of heroin, 6-MAM or morphine as inhibitors and ATCh or BTCh as substrates. Ki values were calculated as described previously, Ki IC50/(1 S/Km),13 following inhibition of substrate hydrolysis by 305000 mol/L opiate over a substrate range of 0.110 mmol/L.

RESULTS
Both isolated human serum BuChE and erythrocyte AChE were capable of hydrolysing heroin to 6-MAM, although at different rates (Fig. 1). Human BuChE (0.5 IU) hydrolysed the physiologically relevant initial amount of 85 nmol heroin in 1 mL (36 g) with a halflife (1/2) of 3.5 min, close to the reported in vivo enzyme/ substrate ratio and 1/2 value of heroin in blood of 25 min.6,14,15 In the present study, BuChE did not further hydrolyse 6-MAM to

morphine, in agreement with the ndings of Lockridge et al.,3 but unlike the ndings of Kamendulis et al.4 The BuChE Km for heroin was 110 mol/L with catalytic constant (kcat) of 540/min (Table 1). Under similar conditions erythrocyte AChE C-terminated with the peptide encoded by exon 5,7 hydrolysed 85 nmol heroin in 1 mL to 6-MAM with the longer 1/2 of 25 min. Erythrocyte AChE further displayed higher Km (620 mol/L) and a lower kcat (351/min) values for heroin hydrolysis than those of BuChE, reecting weaker afnity and/or catalytic efciency than those of BuChE. This was in line with the nine-fold difference in the catalytic efciencies of BuChE and erythrocyte AChE in hydrolysing heroin (4.5 and 0.5/min per mol/L, respectively). However, unlike BuChE, AChEE5 proceeded to hydrolyse 6-MAM to morphine, although at the low rate of 0.1 nmol/min (Fig. 1). Heroin hydrolysis by AChE-E5 was not affected by 10 mol/L of the selective BuChE inhibitor isoOMPA (data not shown). In contrast, no hydrolysis was observed in the presence of AChE-E5 and 10 mol/L of the selective AChE inhibitor BW284c51, except for the expected rate of spontaneous hydrolysis (0.07 nmol/min). This conrmed that it was red blood cell AChE that hydrolysed heroin, excluding the possibility of contaminating enzymes. Interestingly, no hydrolysis was detected when 6-MAM was added to erythrocyte AChE as a substrate, demonstrating that AChEE5 could degrade 6-MAM only when produced from heroin within its active site. In contrast, the non-natural metabolite 3-MAM was efciently hydrolysed to morphine, both by BuChE and AChE-E5 (data not shown). Recombinant AChE, consisting of the synaptic AChE form, abundant in the human brain and C-terminated by the exon 6-encoded peptide, did not signicantly hydrolyse heroin, 6MAM or 3-MAM (Fig. 1 and data not shown). Because both heroin and 6-MAM are much poorer cholinesterase substrates than the acylthiocholines, they could also be considered as inhibitors of thiocholine ester hydrolysis. Ki determinations for heroin and 6-MAM revealed a decreasing order of afnities for their interactions with BuChE, AChE-E5 and AChE-E6 (Table 1). Consistent with the observed relative rates of catalysis, the Ki values of BuChE and AChE-E5 for morphine were approximately 10-fold higher, reecting considerably weaker afnities than those for heroin and 6-MAM. To further explore the interaction(s) between various cholinesterase variants and opiate derivatives, we added heroin, 6-MAM, morphine or codeine (as a control) to solid phaseconjugated enzymes. Following 5 min incubation with 650 mol/L of the noted agents and washing to remove unbound drug, we tested the residual capacity of these enzymes to hydrolyse ATCh or BTCh using a spectrophotometric assay adapted for use with microtitre plates.16 When tested within 5 min of drug removal, codeine did not affect substrate hydrolysis by BuChE, AChE-E5 or AChE-E6. In contrast, heroin enhanced the activity of BuChE by 25% and inhibited the activity of AChE-E5 by 28%. There was almost no effect of heroin on substrate hydrolysis by AChE-E6, consistent with its inability to hydrolyse heroin. 6-Monoacetylmorphine inhibited the activity of BuChE, AChE-E5 and AChE-E6 by 30%, 25% and 22%, respectively (Table 1). All three enzymes displayed full activity within 60 min of drug removal.

DISCUSSION
Using puried native and recombinant variants of human cholinesterases,17 we have demonstrated that erythrocyte, but not

598

AY Salmon et al. morphine and that both heroin and 6-MAM affect the catalytic properties of erythrocyte AChE and serum BuChE. These ndings suggest that erythrocyte AChE hydrolyses heroin in human blood

the brain, AChE can hydrolyse heroin to 6-MAM at a physiologically relevant rate not much lower than that of serum BuChE; that erythrocyte AChE, but not serum BuChE, can further degrade 6-MAM to

Table 1.

Effects of heroin and its derivatives on the catalytic activities of cholinesterase variants Serum BuChE Enzyme variants Erythrocyte AChE-E5 620 351 0.55 765 727 656 752 52612 990.5 Brain AChE-E6 NS NS NS 1533 964 1302 781 5119 980.5

Heroin Km, mol/L* kcat/min Catalytic efciency/min per mol/L Ki (mol/L) % Remaining immobilized enzyme activity 6-MAM Ki (mol/L) % Remaining immobilized enzyme activity Morphine Ki (mol/L)* % Remaining immobilized enzyme activity Codeine

110 540 4.5 442 1254 304 704 63320 1004.1 No detectable effects

*Km values were derived from double reciprocal Lineweaver-Burk plots over a heroin concentration range of 324000 mol/L. Kcat values were based on observed Vmax values (for heroin) divided by the number of enzyme active sites as determined by substrate hydrolysis assays (as described in Materials and Methods) and using turnover numbers reported for human AChE (16) and BuChE (12). Ki values for inhibition of substrate hydrolysis by 305000 mol/L opiate over a substrate range of 0.110 mmol/L. One out of three reproducible experiments with triplicate measurements. Per cent inhibition of substrate (ATCh or BTCh) hydrolysis capacity as measured on antibody-immobilized enzyme following 5 min incubation with 100 L of 650 mol/L drug and subsequent washes. One out of three reproducible experiments with triplicate measurements. Activity of antibody-immobilized enzyme incubated without drug served as control and referred to as 100% activity (the activity of the different enzymes was almost the same (> 95%) as the initial activity before treatment). NS, non-signicant; BuChE, butyrylcholinesterase; AChE, acetylcholinesterase; BTCh, butyrylthiocholine; ATCh, acetylthiocholine.

Fig. 1. Heroin hydrolysis. (a) In vivo metabolism. The structures of heroin, 6monoacetylmorphine (6-MAM) and morphine are shown. Sites of enzyme hydrolysis are marked by arrows. Intravenous injection of heroin leads to rapid hydrolysis of the ester bond at position 3, yielding 6-MAM (1/2 25 min),6 followed by a slower rate hydrolysis of the ester bound at position 6, yielding morphine (1/2 30 min).2 (b) In vitro hydrolysis by cholinesterase variants. Cholinesterase variants encoded by the noted exons (E) differ in their entire sequence (between butyrylcholinesterase (BuChE) and acetylcholinesterase (AChE)) or in their Cterminal peptides, encoded by E5 or E6 (for red blood cells (RBC) and brain AChE, respectively). Enzymes were incubated with 85 mol/L heroin and levels of the noted opiate derivatives (, heroin; , 6-MAM; , morphine). Presented are residual fractions of heroin and accumulation of 6-MAM and morphine as a function of time (in different time scales for each enzyme). Calculated t2 values were 3.5 min for heroin hydrolysis by serum BuChE (left), 25 min for heroin hydrolysis by erythrocyte AChE-E5 (RBC, centre) and no hydrolysis for synaptic AChEE6 (right). Note morphine production only by AChE-E5 and the slow spontaneous hydrolysis of heroin into 6-MAM during the longterm incubation with AChE-E6, which correlates well with incubation in the absence of enzyme (up to 0.07 nmol/min). One out of two experiments with duplicates differed in less than 8%.

Heroin hydrolysis by human acetylcholinesterase and demonstrate the rst biochemical distinction between the previously indistinguishable AChE variants12 expressed in the brain and haematopoietic cells. While BuChE is well known for its wide range of substrates, AChEs specicity is much more limited, which is considered to contribute towards the high rate of ACh hydrolysis.18 In addition to their active site, both enzymes possess peripheral binding sites for substrates and inhibitors.12 The difference between the erythrocyte and brain AChE variants is in their distinct C-terminal domains,7 and ample biochemical evidence demonstrates that these C-terminal domains are only loosely associated with the globular enzyme core unit.18,19 This indicates that the alignment of the C-terminus along the protein core may affect heroin binding, perhaps by physical masking of a peripheral domain in only one of the two variants. The fact that 6-MAM cannot serve as a substrate but can serve as an inhibitor for AChE-E5 suggests that in order to be hydrolysed 6-MAM must be correctly positioned within the active site of AChE. The efcient hydrolysis of 3-MAM indicates that the 3-acetyl but not the 6-acetyl group is required for proper access of the molecule 6-MAM into the active site groove. In addition, the distinct interactions of the various opiate derivatives with cholinesterase variants may reect binding to the well-recognized peripheral site.20 Thus, the 3-acetyl group of the drug enables interaction with the enzyme through its peripheral site. This latter hypothesis is strengthened by the apparent similarity between the heroin-induced enhancement of BTCh hydrolysis by immobilized BuChE and the phenomenon of substrate activation previously reported for BuChE.2123 The inhibition of AChE-E5 by heroin is similarly parallel to the substrate inhibition characteristic of this enzyme.24 Both BuChE substrate activation and AChE substrate inhibition were attributed in these and other studies to a peripheral binding site(s).2124 The fact that heroin does not affect AChE-E6 hydrolysis, leads us to tentatively attribute the heroinAChE peripheral interaction to a site masked by the E6-derived C-terminus. In spite of the better catalytic prole of BuChE to hydrolyse heroin, most of the drug hydrolysis in vivo occurs in the erythrocyte fraction of human blood.5,6 Since heroin is a highly lypophilic molecule with a tendency to concentrate on cellular membranes, the concentration of heroin at the micro-environment of the erythrocyte membrane may reach much higher values than in the plasma. This would assist erythrocytic AChE to hydrolyse most of the administered heroin. Moreover, we now demonstrate that the second step of hydrolysis of 6-MAM into morphine can also be carried in this micro-environment. Red blood cells can therefore serve as efcient carriers of heroin and 6-MAM from the circulation towards the bloodbrain barrier, where these compounds penetrate into the brain to create the rapid onset of heroin effects. Within both the blood and the brain, the inhibition of BuChE catalysis by 6-MAM supports the notion of an inhibitory loop whereby 6-MAM production suppresses BuChEs capacity to hydrolyse heroin.3 In contrast, 6-MAM affects AChE much less and cannot in its free state serve as a substrate for either AChE variant. Therefore, penetrance of heroin or 6-MAM into the brain would protect them from the rapid degradation taking place in the circulation, consistent with previous in vivo reports.6 The distinct opiate hydrolysing capacities of the brain and erythrocyte AChE therefore shed new light on heroin metabolism in humans.

599

ACKNOWLEDGEMENTS
We would like to thank Dr B Norgaard-Pedersen, Copenhagen, for anti-AChE and BuChE antibodies. This research was supported by a grant to HS from the Israel Science Fund.

REFERENCES
1. Inturrisi CE, Max MB, Foley KM, Schultz M, Shin S-U, Houde RW. The pharmacokinetics of heroin in patients with chronic pain. New Engl. J. Med. 1984; 310: 121317. 2. Boerner V, Abbot S, Roe RL. The metabolism of morphine and heroin in man. Drug Metab. Rev. 1975; 4: 3973. 3. Lockridge O, Mattershaw-Jackson N, Eckerson HW, LaDu BN. Hydrolysis of diacetylmorphine (heroin) by human serum cholinesterase. J. Pharmacol. Exp. Ther. 1980; 215: 18. 4. Kamendulis LM, Brezezinski MR, Pindel EV, Bosron WF, Dean RA. Metabolism of cocaine and heroin is catalyzed by the same human liver carboxylesterases. J. Pharmacol. Exp. Ther. 1996; 279: 71317. 5. Smith DA, Cole WJ. Identication of an arylesterase as the enzyme hydrolysing diacetylmorphine (heroin) in human plasma. Biochem. Pharmacol. 1976; 25: 36770. 6. Owen JA, Nakatsu K. Diacetylmorphine (heroin) hydrolases in human blood. Can. J. Physiol. Pharmacol. 1983; 61: 8705. 7. Ben Aziz-Aloya R, Seidman S, Timberg R, Sternfeld M, Zakut H, Soreq H. Expression of a human acetylcholinesterase promoter-reporter construct in developing neuromuscular junctions of Xenopus embryos. Proc. Natl Acad. Sci. USA 1993; 90: 24715. 8. Inturrisi CE, Schultz M, Shin S, Umans JG, Angel C, Simon EJ. Evidence from opiate binding studies that heroin acts through its metabolites. Life Sci. 1983; 33: 7736. 9. Barrett DA, Dyssegaard AL, Shaw PN. The effect of temperature and pH on the deacylation of diamorphine in aqueous solution and in human plasma. J. Pharm. Pharmacol. 1992; 44: 6068. 10. Liao J, Mortensen V, Norgaard-Pedersen B, Koch C, Brodbeck U. Monoclonal antibodies against brain acetylcholinesterase which recognize the subunits bearing the hydrophobic anchor. Eur. J. Biochem. 1993; 215: 3340. 11. Ellman GL, Courtney D, Anders V Jr, Featherstone RM. A new and rapid colorimetric determination of acetylcholinesterase activity. Biochem. Pharmacol. 1961; 7: 8895. 12. Schwarz M, Loewenstein-Lichtenstein Y, Glick D, Liao J, NorgaardPedersen B, Soreq H. Successive organophosphate inhibition and oxime reaction reveals distinct responses of recombinant human cholinesterase variants. Mol. Brain Res. 1995; 31: 10110. 13. Loewenstein-Lichtenstein Y, Schwarz M, Glick D, Soreq H. Genetic predisposition to adverse consequences of anti-cholinesterases in atypical BCHE carriers. Nature Med. 1995; 1: 10825. 14. Martin WR, Fraser HF. A comparative study of physiological and subjective effects of heroin and morphine administered intravenously in post-addicts. J. Pharmacol. Exp. Ther. 1961; 133: 38899. 15. Nakamura G, Thornton JI, Noguchi TT. Kinetics of heroin deacylation in aqueous alkaline solution and in human serum and whole blood. J. Chromatogr. 1975; 110: 819. 16. Neville LF, Gnatt A, Loewenstein Y, Seidman S, Ehrlich G, Soreq H. Intramolecular relationships in cholinesterases revealed by oocyte expression of site-directed and natural variants of human BCHE. EMBO J. 1992; 11: 16419. 17. Soreq H, Ben-Aziz R, Prody CA et al. Molecular cloning and construction of the coding region for human acetylcholinesterase reveals a GC-rich attenuating structure. Proc. Natl Acad. Sci. USA 1990; 87: 968892.

600

AY Salmon et al.
22. Eriksson H, Agustinsson KB. A mechanistic model for butyrylcholinesterase. Biochim. Biophys. Acta 1979; 567: 16173. 23. Getley SJ. Activities of the enantiomers of cocaine and some related compounds as substrates and inhibitors of plasma butyrylcholinesterase. Biochem. Pharmacol. 1991; 41: 124954. 24. Haas R, Adams EW, Rosenberry MA, Rosenberry TL. Substrate-selective inhibition and peripheral site labeling of acetylcholinesterase by platinum (terpyridine) chloride. In: Shafferman A, Velan B (eds). Multidisciplinary Approaches to Cholinesterase Functions. Plenum Press, New York. 1992; 1319.

18. Sussman JL, Harel M, Frolow F et al. Atomic structure of acetylcholinesterase from Torpedo californica: A prototype acetylcholine binding protein. Science 1991; 253: 8729. 19. Taylor P, Radic Z. The cholinesterases: From genes to proteins. Ann. Rev. Pharmacol. Toxicol. 1994; 34: 281320. 20. Shafferman A, Velan B, Ordentlich A et al. Substrate inhibition of acetylcholinesterase: Residues affecting signal transduction from the surface to the catalytic center. EMBO J. 1992; 11: 36518. 21. Cauet G, Friboulet A, Thomas D. Horse serum butyrylcholinesterase kinetics: A molecular mechanism based on inhibition studies with dalensylaminoethyltrimethylammonium. Biochem. Cell. Biol. 1987; 65: 52935.