‘H nmr Study of the Effectiveness of Various

Thiols for Removal of Methylmercury from

Hemolyzed Erythrocytes
Dallas L. Rabenstein, R. Stephen Reid, and Anvarhusein A. Isab
Department of Chemistry, University of Alberta, Canada

ABSTRACT

The effectiveness of eight thiol ligands for removing methylmercury (CH3Hg(II)) from its
glutathione and hemoglobin complexes in hemolyzed erythrocytes has been studied by IH nuclear
magnetic resonance spectroscopy. These complexes are the predominant methylmercury species in
human erythrocytes. The effectiveness was determined from the exchange-averaged chemical shift
of the resonance for the proton on the *carbon of the cysteinyl residue and from the intensity of the
resonance for the methylene protons of the glycine residue of reduced glutathione (GSH), both of
which provide a measure of the amount of glutathione in the CH3Hg(Il)-complexed form. The thiol
ligands were found to release GSH from its CH,Hg(lI) complex in the order 2, 3-dimercap
tosuccinic acid > mercaptosuccinic acid > cysteine > mercaptoacetic acid > Dpenicillamine >
2, 3-dimercaptopropanesulfonic acid > N-acetyl-D,L-penicillamine > D.L-homocysteine.

INTRODUCTION

Chelation therapy’ is one of several techniques used for detoxification in methylmercury

(CH,Hg(II)) poisoning [l-3]. Complexing agents that contain one or more sulfhydryl
groups are effective in this role, and a variety of such molecules have been tested,
generally for their effectiveness in preventing neurological damage and/or increasing the
rate of elimination of CH,Hg(II) from rats or mice [4-161. The mechanism of action is
presumed to involve a simple competition among the endogenous sulfhydryl groups and
those of the antidote molecules. By conversion of the methylmercury (CH,Hg(II)) from
a complex with an intracellular sulfhydryl molecule to a freely diffusible complex, the
rate of efflux of CH,Hg(II) from cells and organisms can be enhanced. However,

Address reprint requests to Dr. Dallas L. Rabenstein, Department of Chemistry, University of Alberta, Ed-
monton, Alberta, Canada T6G 2G2.
1 Even though it is well established that CHsHg(II) has little tendency to react with more than one donor
group, such treatment is still referred to as chelation therapy by analogy with similar treatments for poisoning
by other metals.

Journal of Inorganic Biochemistry 18,241-25 l(1983)

@ Elsevier Science Publishing Co., Inc. 1983 241
52 Vanderbilt Ave., New York, NY 10017 0162-0134/83/$3.OQ
242 D. L. Rabenstein, R. S. Reid, and A. A. Isab

because of the molecular complexity of the test animals, little is known about the details
of the mechanism of action at the molecular level.
The results of our recent ‘H nuclear magnetic resonance (nmr) study [ 171 of the
binding of CH,Hg(II) in intact human erythrocytes suggest it may be possible to obtain
such information by ‘H nmr. Information about the removal of CH,Hg(II) from the
erythrocyte is also of interest, since the erythrocyte is known to play a central role in
CH,Hg(II) toxicology by binding a sizable fraction of the body burden CH,Hg(II) and by
transporting it to other tissues [ 18-201.
As the first part of a study of the chemistry involved in the removal of CH,Hg(II) from
human erythrocytes, we have determined the effectiveness of several sulfhydryl-
containing molecules for extracting CH,Hg(II) from its complexes with the sulthydryl
ligands naturally present in hemolyzed erythrocytes, using as a measure of this
effectiveness the fraction of reduced glutathione (GSH) in the free fomr as determined by
‘H nmr. This parameter has relevance to other tissues as well, since GSH is known to be
abundant in cellular systems [2 I] and to be involved in the toxicology of CH,Hg(II). For
example, there is a dramatic dependence of the uptake of CH,Hg(II) by the kidneys.
erythrocytes, brain, and liver of rats on intracellular GSH levels 1221, and 20-357~ of the
CH,Hg(II) in rat liver cytosol [23] and 30% of rat cerebral soluble CH,Hg(II) 1341 is in a
GSH complex.

EXPERIMENTAL

Erythrocytes
Erythrocytes were obtained from venous blood that was collected in Vacutainers
(Becton, Dickinson and Co.) containing ethylenediaminetetraacetic acid (EDTA)
solution. The whole blood was centrifuged at 5000 rpm at 4°C for 15 min, the plasma
and buffy coat drawn off, and then the packed cells were washed two or three times by
suspension in an equal volume of a DzO solution of isotonic saline-glucose. The DzO
wash replaces most of the intracellular Hz0 with DzO. Packed erythrocytes were
hemolyzed by the freeze-thaw technique or by sonication.

Chemicals
CH,HgI (Alfa Division, Ventron Corp.) was converted to a solution of CH,HgOH by
reaction with Ag?O, as described previously [27]. The CH,HgOH concentration was
found by titration with NaCl[25] to be 0.054 M.
Mercaptoacetic acid (Eastman Kodak Co.), cysteine, penecillamine, 2-mercap-
tosuccinic acid, 2,3-dimercapto- 1-propanesulfonic acid, sodium salt (Aldrich Chemical
Co.), N-acetylpenicillamine, meso-2,3-dimercaptosuccinic acid (Sigma Chemical Co.),
and homocysteine (Nutritional Biochemicals Corp.) were used as received. Their
purities were determined by procedures described previously [26].

nmr Measurements

‘H nmr spectra were measured at 400 MHz on a Bruker WH-4OO/DS spectrometer

operating in the pulsed Fourier transform mode. Spectra were measured at 25°C on 0.5
Methylmercury in Hemolyzed Erythrocytes 243

ml samples of hemolyzed cells in 5 mm outer diameter (0.d.) nmr tubes. The free
induction decay was collected in 8K of data points, with zero filling to 16K before
Fourier transformation. Quadrature detection was used with a spectral width of 5000 Hz,
and generally 100 transients were collected for each spectrum. Chemical shifts are
reported relative to the methyl resonance of sodium 2,2-dimethyl-2-silapentane-5-
sulfonic acid (DSS), based on the resonance for the methylene protons of glycine having
a chemical shift of 3 540 ppm.
The interfering hemoglobin resonances were either completely eliminated by taking
advantage of the differences in the spin-spin relaxation times of ‘H resonances of small
molecules and macromolecules with the spin-echo Fourier transform (SEFT) method
(90” - rr - 180” - rr - aquisition), [27, 281 or partially suppressed by taking
advantage of the transfer of saturation throughout the envelope of protein resonance that
follows a selective saturation pulse somewhere on the protein envelope [29]. In the SEFI
measurements, a r2 value of 0.060 set was used. In the suppression by saturation
measurements, the selective presaturation pulse was applied at 8.1 ppm for 2 set,
followed by the 90” observation pulse at 5.1 ppm.

RESULTS AND DISCUSSION

In one model that has been suggested for the binding of CH,Hg(II) in erythrocytes, the
CH,Hg(II) is bound reversibly by sulfhydryl groups and is distributed among the various
sylfbydtyl-containing ligands according to a simple chemical equilibrium [30]. The two
main binding sites for CH,Hg(II) in human erythrocytes are the sulfhydryl groups of
GSH and hemoglobin [ 17, 311. The conditional formation constant2 for the CH,Hg(II)--
GSH complex is 3.5 X 1Ol’ at pH 7.4, while that for the CH,Hg(IIhhemoglobin
complex at pD 7.4 is 5.0 x lOlo [32]. These constants, when combined with the amounts
of GSH and hemoglobin in normal human erythrocytes, indicate that -60% of the
CH,Hg(II) in human erythrocytes will be complexed by GSH and 40% by hemoglobin.
This is consistent with experimental data which show that, at low CH,Hg(II) levels,
GSH is the favored binding site in human erythrocytes [ 17, 3 11.
The 400 MHz SEFT ‘H nmr spectrum of hemolyzed erythrocytes is shown in Figure 1.
Most of the hemoglobin resonances have been eliminated, giving a spectrum of sharp
resonances from the more abundant small molecules. Those of interest to the present
study are identified in Figure 1.
When CH,HgGH is added to the hemolyzed erythrocytes, resonances gl, 82, and g5 of
GSH are affected due to complexation of CH,Hg(II), as shown by spectra A-C in
Figures 2 and 3. Specifically, resonance g 1 is decreased in intensity because the
methylene protons of the glycine residue are nonequivalent in the CH,Hg(II)-GSH
complex and give an AB pattern. Resonance g2 disappears and resonance g5 shifts to
higher frequency due to exchange of GSH between its free and complexed forms3 The
fractional concentration of GSH in the free form, fr=, can be obtained from the

2 The conditional formation constant, Kfc, is a measure of the tendency for complexation to occur under a
particular set of conditions, e.g., at pH 7.4, and can be defined in terms of the reaction: CH,Hgf + Lf Z
CH3HgLr where CH,Hgf and Lf include ail free forms of CHsHg(I1) and the &and and CHfHgLr all forms
of the complex.
3 Resonance g2 shifts by 180 Hz and resonance g5 by 16 Hz when GSH is complexed by CH3Hg(II). The
observation of an exchange-averaged resonance for g5 but no resonance for g2 indicates that the average life-
time of a CH3HG(II)-CSH complex is in the range of 0.01-0.001 sec.
 g6
/
H3NCH2COi
-02C~CH2CH2CNM~HCNHCH~O2-
NH3+
t
g3

el c2
/
CH3

H2N - 4 - N - CH2COP

NH

95
96
i

I I I I I I I I I I

8 6 4 2 0
PPm
FIGURE 1. The 400MHz tHSEFT nmr spectrum of hemolyzed human erythrocytes that had been
washed with a D20 solution of isotonic saline-glucose before hemolysis. A delay time (between
pulses) of 0.060 set was used in the spin-echo pulse sequence.
Methylmercury in Hemolyzed Erythrocytes 245

I 1

3.5
.

wm
1

3.0
,

FIGURE 2. Portions of the SEFT nmr spectra of hemolyzed erythrocytes containing (A) 0.00 M
CHsHg(II), (B) 0.00162 M CH3Hg(II), (C) 0.00432 M CHaHg(II), (D) 0.0040 M CHaHG(II) and
0.000 M DMPS, (E) 0.0039 M CHsHg(I1) and 0.0020 M DMPS, and (F) 0.0039 M CH3Hg(II) and
0.0033 M DMPS.
246 D. L. Rabenstein, R. S. Reid, and A. A. Isab

I 1
,\
‘I / :
I ,I
’i

E :’ ‘\
\_,- \ ’

1 B I, I I, IS I I / / 1, I I J
4.65 4.60 4.55 4.50 4.65 4.60 4.55 4.50
ppm PPrn

FIGURE 3. Portions of the tH nmr spectra obtained for the same hemolyzed erythrocyte samples
as in Figure 2 by the method based on the suppression of the hemoglobin resonances by satura-
tion [29].

exchange-averaged chemical shift of resonance 85, Bohr with the equation

where 6, and 6,. are the chemical shifts of resonance g5 for free and CH,Hg(II)-
complexed GSH. For example, 100, 63, and 23% of the GSH is calculated to be in the
free form from chemical shift data in spectra A-C, respectively, of Figure 3.
The complexation of CH,Hg(II) by GSH is reversible, as indicated by the increase in
the intensity of resonance g 1 in spectra D-F of Figure 2 and by the shift of resonance g5
to lower frequency in spectra D-F of Figure 3 when 2,3-dimercaptopropanesulfonic acid
(DMPS) is added to the CH,Hg(II)-containing erythrocytes. From spectra D-F in Figure
3. 26, 62, and 70% of the GSH is calculated to be in the free form.
To compare the extent to which the sulfhydryl-containing ligands listed in Table 1
release GSH from its CH,Hg(II) complex in hemolyzed erythrocytes, 1.2 ml of 0.054 M
CH,HgOH was added to 15 ml of hemolyzed erythrocytes. From the chemical shift of
resonance g5, GSH was calculated to be 74% in the complexed form. DMPS was then
added to portions of the CH,Hg(II)-containing hemolysate to determine the concen-
tration required to give approximately 50% of the GSH in the free form. This
concentration (0.002 M) of the other dithiol ligands, and twice this concentration of the
monothiol ligands, was then added to portions of the CH,Hg(II)-containing hemolysate.
In Figure 4 are shown the spin-echo spectra for the CH,Hg(II)-containing hemolysate
before (A) and after (B-F) the addition of some of the sulthydryl-containing ligands that
Methylmercury in Hemolyzed Erythrocytes 247

have been tested in animal studies. By comparison of the intensity of resonance g 1 to that
of resonance el or C2, which are not affected by CH,Hg(II), it is immediately obvious
that the different ligands have widely differing abilities for releasing GSH from its
CH,Hg(II) complex. The fractional concentration of GSH in the free form, as calculated
from resonance g.5, is given in Table 1. Also, the intensity of resonance g 1 is compared
to the intensities of resonances e 1 and C2 in Table 1. Even though the complexation of
CH,Hg(II) by hemoglobin cannot be observed directly, it can be assumed that the
redistribution involves the hemoglobin-bound CH,Hg(II) as well, since hemoglobin
binds CH,Hg(II) less strongly than does GSH, and exchange of CH,Hg(II) between its
hemoglobin and GSH complexes is rapid on the nmr time scale [ 171.
As the conditional formation constant of the CH,Hg(II)-complex of the added
sulfhydryl ligand increases (Table I), the amount of GSH in the free form increases,
which is consistent with the distribution of CH,Hg(II) among the various sulfhydryl
ligands according to a simple chemical equilibrium. A conditional formation constant is
not given for the DMSA complex. however, NMR experiments indicate the binding of
CH,Hg(II) by DMSA at pH 7.4 to be stronger than by any of the other ligands listed in
Table 1, consistent with its position in the effectiveness order as determined by nmr.
It also is of interest to compare the effectiveness of the sulfhydryl ligands for releasing
GSH from its CH,Hg(II) complexes with their effectiveness as CH,Hg(II) antidotes in
animal studies. Aaseth and co-workers have obtained the following orders of
effectiveness, as judged by removal of CH,Hg(II) from the brain and blood of mice:
DMSA > MSA > NAPA > DMPS [I l] and NAPA > PSH > GSH - CYS [4]. DMSA
and MSA are the most effective in the animal studies and in the in vitro experiments in
Table 1. In the animal experiments, NAPA is more effective relative to PSH and DMPS
than would be predicted on the basis of the results in Table 1, which might reflect its
ability to penetrate cell membranes [ 111. As the second part of our study of the removal
of CH,Hg(II) from human erythrocytes, we intend to study the transport of sulfhydryl
ligands across the erythrocyte membrane using ‘H nmr techniques [33]. Presumably
CYS and GSH are not effective in animal studies because they are consumed by normal
metabolic processes.
In terms of the effectiveness of sulfhydryl ligands as CH,Hg(II) antidotes, it is perhaps
significant that the kinetic stability of their CH,Hg(II) complexes in hemolyzed
erythrocytes appears to differ, as indicated by the presence or absence of a resonance for
the protons of CH,Hg(II). In general, the binding of CH,Hg(II) to sulfhydryl ligands in
biological systems is quite labile, even though the thermodynamic stability of such
binding is large [2&34]. For example, the average lifetime of a CH,Hg(II)-GSH
complex in human erythrocytes is estimated to be less than 0.01 set [ 171. Also, exchange
of CH,Hg(II) between GSH and hemoglobin in human erythrocytes is fast, as indicated
by the absence of a resonance for the CH,Hg(II) protons.4 That portion of the spectrum
where the CH,Hg(II) resonance would be expected is shown in Figure 5 for CH,Hg(II)-

4 If exchange of CH3Hg(II) between GSH and Hb were slow, we would expect to observe a resonance for the
CH3Hg(II) complexed by GSH because a resonance is observed for CH3Hg(II) in aqueous solutions of GSH
and CH3Hg(II) and in erythrocytes exchange of CH3Hg(II) between GSH molecules is fast, as indicated by
the exchange-averaged resonance for 85. However, we would not expect to observe a resonance for that
CH3Hg(II) complexed by hemoglobin because it would presumably have the motional properties of the
hemoglobin molecule, and thus its protons would have a short effective T2 value similar to those of the
hemoglobin protons. Fast exchange of CH3Hg(II) between its GSH and hemoglobin complexes causes the
effective T2 value for the CHsHg(II) resonance to be equal to the weighted average of the Tz values of the
CHsHg(I1) protons in the two complexes, in which case the resonance is broad and not observable.
248 D. L. Rabenstein, R. S. Reid, and A. A. Isab

/ 1 / I I I ( I I I I I I I I I I I 1

3.5 3.0 3.5 3.0

mm ppm
FIGURE 4. Portions of the SEFT nmr spectra of hemolyzed erythrocytes containing (A) 0.0040 M
CH3Hg(II), (B) 0.0039 M CH3Hg(II) and 0.0036 M NAPA, (C) 0.0039 M CHsHg(II) and 0.0018 M
DMPS, (D) 0.0039 M CH3Hg(II) and 0.0036 M PSH, (E) 0.0039 M CH3Hg(II) and 0.0036 M MSA,
and (F) 0.0039 M CH3Hg(II) and 0.0018 M DMSA.
 %
TABLE 1. Experimental Data on Hemolyzed Erythrocytes Containing CH,Hg(II) and Sulfhydryl Ligands m
G
Concentration % of CSH in the References to 5
Sulfhydryl Ligand Abbreviation (MI Free Forma gl/elb gl/Czb log Kfc” animal studies i:
_ v
2,3-Dimercaptosuccinic acid DMSA 0.002 56 1.19 2.06 5,9,11, 12, 14,lS 5
Mercaptosuccinic acid MSA 0.004 55 1.19 2.07 11.68 11
LCysteine CYS 0.004 51 1.09 1.78 11.57 4,5
Mercaptoacetic acid MAA 0.004 47 1.00 1.84 11.47
DPenicillamine PSH 0.004 51 0.95 1.71 11.33 4-9
2,3_Dimercaptopropane sulfonic DMPS 0.002 47 0.87 1.69 11, 12, 16
acid
N-Acetyl-D,L-penicillamine NAPA 0.004 40 0.80 1.50 11.20 4,9-12
D,L-Homocysteine HCYS 0.004 40 0.78 1.33 11.15 6

b Ratio of intensities. Precision of values estimated at 28%.

C From ref. (261.
250 D. L. Rabenstein, R. S. Reid, and A. A. lsab

I I I I 1 1 I
1.0 0.5
fvm
FIGURE 5. Portions of the SEFT nmr spectra of hemolyzed erythrocytes containing 0.0039 M
CH3Hg(II) and (A)O.0036 M PSH, (B) 0.0018 M DMPS, (C) 0.0018 M DMSA, (D) 0.0036 M MAA,
and (F) 0.0036 M MSA.

” containing hemolyzed erythrocytes to which various sulfhydryl ligands have been

added. Resonances are observed for CH,Hg(II) complexed by PSH, DMSA. MAA, and
MSA, but not by DMPS. However, there is little correlation between the intensity of the
resonance and the conditional formation constant of the complex. For example, only a
very weak resonance is observed for the CH,Hg(iI) complexed by CYS (not shown in
Figure 5), even though its thermodynamic stability, as indicated by its conditional
formation constant, is larger than that for MAA. Because the observation of a resonance
for CH,Hg(Ii) depends not only on the thermodynamic stability of the complex but also
its kinetic stability, fast exchange of CH,Hg(II) between the drug molecule and
hemoglobin or GSH will give rise to a shorter T,. which causes the resonance to be
attenuated in the spin-echo spectrum (see footnote 4). The spectra in Figure 5 suggest
that the kinetic stability of the MSA and MAA complexes is greater than that of the
complexes with the other sulfhydryl ligands studied. We are presently studying the
ligand exchange kinetics of these complexes to determine if there are significant
differences in their kinetic stability.

This research was supported by the Natural Sciences and Engineering Research Council of Conada
through their Strategic Grants Program and by the University of Alberta. Financial support by an
I. W. h’iilam Scholarship (to R.S.R.) and by an Alberta He&age Foundation for Medical Research
Postdoctoral Fellowship (to A.A.I.) is gratefully acknowledged.
Methylmercury in Hemolyzed Erythrocytes 251

REFERENCES

1. F. Bakir, S. F. Damluji, L. Amin-Zaki, M. Murtadha, A. Khalidi, N.Y. Al-Rawi, S. Tikriti,

H. I. Dhahir, T. W. Clarkson, J. C. Smith, and R. A. Doherty,Science 181,230 (1973).
2. T. W. Clarkson, H. Small, and T. Norseth,Arch. Environ. Healrh 26, 173 (1973).
3. P. J. Kostyniak, T. W. Clarkson, R. V. Cestero, R. B. Freeman, and A. H. Abbasi,J. Pharma-
col. Exp. Ther. 192, 260 (1975).
4. J. Aaseth, Acta Pharmacol. Toxicol. 39, 289 (1976).
5. L. Magos, T. W. Clarkson, and J. Allen, Biochem. Pharmacol. 27,2203 (1978).
6. J. Aaseth,Acta Pharmacol. Toxicol. 36, 193 (1975).
I. T. Suzuki. S. Shishido, and N. Ishihara, Brit. J. Ind. Med. 33,88 (1976).
8. L. J. Zimmer and D. E. Carter, Life Sci. 23, 1025 (1978).
9. S. K. Tandon and L. Magos, Brit. J. Ind. Med. 37, 128 (1980).
10. J. Aaseth, A. Wannag, and T. Norseth, Acta Pharmacol. Toxicol. 39, 302 (1976).
11. J. Aaseth and E. A. H. Freidheim,Acla Pharmacol. Toxicol. 42, 248 (1978).
12. F. Planas-Bohne, J. Pharmacol. Exp. Ther. 217, 500 (1981).
13. M. Berlin, L.-G. Jerksell, and G. Nordberg,Acta Pharmacol. Toxicol23, 312 (1965).
14. J. A. Hughes and S. B. Sparber, Res. Commun. Chem. Pathol. Pharmacol. 22, 357 (1978).
15. L. Magos, Brit. J. Pharmacol. 56,479 (1976).
16. B. Gabard, Toxicol. Appl. Pharmacol. 38,415 (1976).
17. D. L. Rabenstein, A. A. Isab, and R. S. Reid, Biochim. Biophys. Acta, 696, 53 (1982).
18. J. K. Miettinen, T. Rahola, T. Hattula, K. Risanen, and M. Tillander, Ann. Clin. Res. 3, 116
(1971).
19. B. Aberg, L. Ekman, R. Falk, U. Greitz, G. Persson, and J. 0. Snihs, Arch. Environ. Health
19,478 (1969).
20. K. D. Lundgren, A. Svenson, and U. Ulfvarson, Stand. J. Clin. Lab. Invest 20, 164 (1967).
21. P. C. Jocelyn, Biochemistry of rhe SH Group, Academic Press, New York, 1972, p. 10.
22. R. J. Richardson and S. D. Murphy, Toxicol. AppZ. Pharmacol31,505 (1975).
23. S. Omata,.K. Sakimura, T. Ishii, and H. Sugano, Biochem. Pharmacol. 27, 1700 (1978).
24. D. J. Thomas and J. C. Smith, Toxicol. Appl. Pharmacol. 47,547 (1979).
25. D. L. Rabenstein and M. T. Fairhurst,J. Amer. Chem. Sot. 97, 2086 (1975).
26. R. S. Reid and D. L. Rabenstein, Can. J. Chem. 59, 1505 (1981).
27. F. F. Brown, I. D. Campbell, P. W. Kuchel, and D. L. Rabenstein,FEBSLett. 82, 12 (1977).
28. D. L. Rabenstein and T. T. Nakashima, Anal. Chem. 51, 1465A (1979).
29. D. L. Rabenstein, A. A. Isab, and D. W. Brown,J. Magn. Res. 41, 361 (1980).
30. J. F. White and A. Rothstein, Toxicol. Appl. Pharmacol. 26, 370 (1973).
31. A. Naganuma and N. Imura, Toxicol. Appl. Pharmacol. 47, 613 (1979).
32. R. S. Reid and D. L. Rabenstein, J. Amer. Chem. Sot. 104,6733 (1982).
33. K. M. Brindle, F. F. Brown, I. D. Campbell, C. Grathwohl, and P. W. Kuchel, Biochem J. 180,
37 (1979).
34. D. L. Rabenstein and C. A. Evans, J. Bioinorg. Chem. 8, 107 (1978).

Received March 24, 1982; accepted July 30, I982