Chem. Res. Toxicol. 1994,7, 882-890
882
Conformations of Zinc Chelates of m s o - and ruc-2,3-DimercaptosuccinicAcid in Aqueous Solution Xiaojun Fang and Quintus Fernando* Department of Chemistry, University of Arizona, Tucson, Arizona 85721 Received June 14,1994@
ruc-2,3-Dimercaptosuccinicacid (DMSA) was found to be superior to meso-2,3-dimercaptosuccinc acid in mobilizing in vivo heavy metals such as Cd, Hg, and Pb. The disadvantage of using ruc-DMSA alone as a clinical antidote for heavy metal poisoning is that it causes a greater loss of endogenous zinc t h a n its meso isomer. The W e r e n c e between the two diastereoisomers of DMSA in t h e excretion of endogenous zinc has been rationalized on the basis of the differences in the conformations of their zinc complexes. The zinc complexes of ruc-DMSA in aqueous solution are more stable than t h e corresponding complexes of its meso isomer because the ruc-DMSA ligands always adopt staggered anti conformations, in which the electrostatic repulsion between two bulky carboxylate groups is minimized; in contrast, unlike in t h e crystal lattice, meso-DMSA ligands always adopt staggered gauche conformations in their zinc complexes. The conformations of various monomeric and dimeric zinc complexes with rucand meso-2,3-dimercaptosuccinic acid in aqueous solution were determined by IR spectroscopy a n d proton NMR spectroscopy as a function of zinc:ligand ratio, by proton NMR spectroscopy as a function of pD, a n d by variable-temperature I3C NMR spectroscopy. ruc-DMSA in ZnLZ6coordinates with zinc ion via two thiolate groups and one carboxylate group, whereas in ZnzLz4each ligand complexes two zinc ions by using one carboxylate group a n d one thiolate group in t h e ,%position to bind to the same zinc ion. I n contrast, meso-DMSA in z1iL2~-uses one thiolate group a n d two carboxylate groups to bind to t h e zinc ion, a n d in Zn2Lz4- each ligand binds to one zinc ion via two carboxylate groups a n d one thiolate group, a n d to t h e other zinc ion via the other thiolate group. I n ZnLz6- of meso-DMSA, one of the carboxylate groups in the ligand binds to t h e zinc ion in a n unsymmetrical bidentate manner.
Introduction Recently, lead exposure in young children has become the focus of a major public health effort. Lead blood levels as low as 10-15 pg/dL, and possibly lower, are linked with undesirable developmental outcomes in human fetuses and children (2). These effects include impaired neurobehavioral development and other possible effects on early development and growth. More specifically, such levels of Pb in blood may lead to impairment of the central nervous system such as delayed cognitive development, delayed learning (1-31, impaired hearing (41, reduced I& scores (1, 51, and neurobehavioral deficiencies (1). The Center for Disease Control of the Department of Health and Human Services has revised its safety Pb level in blood of children from 25 pg/dL in 1985 ( 6 )to 10 pg/dL in 1990 (7). The Clean Air Scientific Advisory Committee to the US.Environmental Protection Agency regards 10-15 pgIdL and below a s a nontoxic level (8). mes0-2~3-Dimercaptosuccinic acid (meso-DMSA),l which is approved by the U.S. Food and Drug Administration, appears to be, so far, the most promising chelating agent available for treatment of lead poisoning in humans. Recently, the synthesis and the properties of rac-2,3-dimercaptosuccinic acid (rac-DMSA)' have been studied (9),and it was shown t h a t the rac-DMSA-Pb2+ complex has a ~~
~
* To whom correspondence should be addressed.
Abstract published in Advance ACS Abstracts, October 15, 1994. Abbreviations: meso-DMSA, meso-2,3-dimercaptosuccinic acid; ratDMSA, rac-dimercaptosuccinic acid; TBA, tert-butyl alcohol; DSS, TMS, tetramethylsodium 2,2-dimethyl-2-silapentane-5-sulfonate; silane; TMA, tetramethylammonium; IR,infrared. @
0893-228x/94/2707-0882$04.50/0
higher formation constant than the meso-DMSA-Pb2+ complex.2 This implies t h a t ruc-DMSA could be a potentially more effective antidote for the treatment of lead poisoning. The disadvantage, however, of using racDMSA alone as a clinical antidote for heavy metal poisoning is t h a t it causes a greater loss of endogenous zinc than its meso isomer. A significant increase in urinary excretion of zinc by rac-DMSA was observed by Egorova et al. (IO)when identical doses of meso-DMSA and rac-DMSA were administered to rats. The drastic difference between the two DMSA isomers in mobilizing endogenous zinc was rationalized with speciation models and with formation constants of the zinc complexes of the two diastereoisomers of DMSA (22). On the basis of the speciation models, it was concluded that the racDMSA is more effective as a prophylactic for the prevention of lead poisoning and that the use of a zinc complex of ruc-DMSA will minimize the depletion of endogenous zinc (11). It is important to confirm, by a n independent method, the speciation model, which was proposed by using the results obtained from potentiometric studies on the basis of statistics (11). It is also important to understand, at a molecular level, the reasons for the significant difference in the depletion of endogenous zinc when different chelating agents or the stereoisomers of the same chelating agent are administered in the treatment of lead poisoning. A knowledge of the interaction between zinc and the diastereoisomeric DMSA molecules will also assist in designing derivatives of rac-DMSA which do not excrete endogenous zinc t o a significant X. Fang and Q . Fernando Department of Chemistry, University of Arizona, Tucson, AZ 85721;unpublished work, May, 1994.
0 1994 American Chemical Society
Chem. Res. Toxicol., Vol. 7, No.6,1994 883
Zinc-DMSA Complexes extent but are more effective in mobilizing in vivo heavy metals. A knowledge of the structure of zinc complexes with DMSA molecules will also provide important information on the mechanism of in vivo mobilization of endogenous zinc by DMSA. We report in this paper the results of experiments that were performed in order to explain, a t a molecular level, the difference in the depletion of endogenous zinc. We have presented experimental evidence, which confirms the speciation model proposed (ll),and we have determined the conformations of various dimeric and monomeric zinc complexes as well a s the protonated forms of rac- and meso-DMSA in aqueous solution.
Experimental Section Materials. Deuterium oxide, deuterated methanol, zinc
bromide (98+%), and tetramethylammonium hydroxide pentahydrate (99%)were purchased from Aldrich Chemical Co. Inc. (Milwaukee, WI). Caution:Tetramethylammonium hydroxide is a very strong base and is corrosive, has a strong ammonialike odor, and should be manipulated in a hood. tert-Butyl alcohol (TBA)l was purchased from EM (Gibbstown, NJ) and distilled before use; all organic solvents used were purchased from EM; rac-DMSA was synthesized in our laboratory as described previously (9); meso-DMSA was a gift from Johnson & Johnson Baby Products Co. (Skillman, NJ); and all other inorganic compounds used were purchased from Mallinckrodt, Inc. (Paris, KY), and were of analytical reagent grade. Caution: Gloves are necessary in the handling of all the above chemicals to avoid direct contact with skin. Synthesis of [(Me)4Nls[Zn(rac-DMSA)zl. The synthesis of the ion association complex was carried out under argon to prevent oxidation of the mercapto compounds to disulfides. Stoichiometric amounts of rac-DMSA (H4L), zinc nitrate, and tetramethylammonium hydroxide shown below were used in the synthesis:
2H4L
+ Zn(N03), + 8(Me),NOH [ZnL,16- + 8(Me),N+ + 2N03- + 8H,O
rac-DMSA (273.3 mg = 1.5 mmol) was dissolved in 10 mL of oxygen-free double-deionized HzO. A 2.72 mL aliquot (0.75 mmol) of 0.2763 M Zn(NO3)z solution, which was previously standardized with EDTA, was added, and finally 2.56 mL of 2.343 M (Me)aOH solution, which was standardized previously with standard HNO3 solution, was added slowly with stirring. The clear solution was allowed to react with continuous stirring for 30 min under an argon atmosphere. After the solvent was evaporated in a rotatory evaporator under vacuum, a white sticky precipitate was left behind. This precipitate was treated with a deaereated mixture of 50%v/v ethanol-acetone, filtered, and dried in vacuum. The purification procedure was repeated three times, and the final product was rinsed with ethyl ether. The stoichiometry and purity of the final product were determined by proton NMR spectroscopy. The compound was dissolved in deuterated methanol (99.8 atom % deuterium) containing 0.03% tetramethylsilane (TMS)' (v/v), and chemical shifts were measured with respect to the TMS signal. In the proton NMR spectrum of the purified compound (Figure 1)a large singlet was observed at 4.92 ppm and a small pentet at 3.31 ppm, which originated from the residual protons in the deuterated methanol solvent. The singlet at 3.25 ppm was attributed to the methyl protons in the tetramethylammonium cation, and the singlet at 3.35 ppm was attributed to the methine protons in the rac-DMSA. The small peaks around 1.2 and 3.5 ppm probably originated from a trace amount of solvent residues introduced in the purification process. The integrated peak ratio of tetramethylammonium protons to DMSA methine protons was 1:18, shown in the scaleup of the chemical shift region around 3.3 ppm (inset of Figure 1). These observations
I
>
.
3.40
,
.
I
.
3.35 3.30 3.15 Chemical S h R (ppm)
9.0
Si0
I
7:O
3.20
6.0 5.0 4:O 3.0 Chemical shift (ppm)
2.0
1:O
0:O
Figure 1. Proton NMR spectrum of [(Me)4Nle[Zn(rac-DMSA)~l
in methanol-d4.
supported the expected stoichiometry of the compound, [(Me).&[Zn(rac-DMSA)zl. NMR Measurements. Proton and l3C NMR spectra were obtained with either a Bruker WM-250 or Bruker AM-250 spectrometer. All experiments were carried out at ambient temperature (22 "C) unless specified in the text. The proton spectra were obtained in 5-mm NMR tubes, and the digital resolution was 0.365 Hz/pt. 13C spectra were obtained in a 5-mm NMR tube, and the digital resolution was 1.850 Hdpt. The chemical shifts in aqueous solution were measured with respect to TBA but reported with respect to sodium 2,2dimethyl-2-silapantane-5-sulfonate(DSS).' The advantage of using TBA is that its signal is independent of pH and ionic strength (12).-Under our experimental conditions 0.4-1% by volume of TBA was sufficient to yield a good reference signal at 1.01 ppm downfield from DSS, which differed from the reported value (13)by 0.22 ppm upfield. (A) Proton NMR Spectroscopy of (~uc-DMSA)~and (me~o-DMsA)~at Different L4-/Zn2+Ratios. A Zn(NO3)z stock solution in D2O was prepared by dissolving Zn(N03)z~HzO in DzO [the depletion of water of hydration in the solid Zn(NO3)z.nHzOwith DzO is difficult due to its hydrolysis] and standardized with EDTA. A titration solution was prepared in a 5-mm NMR tube by mixing a weighed amount of rac-DMSA or meso-DMSA, four equivalent amounts of NaOD solution, 10 pL of 10% TBA, and a calculated volume of DzO to bring the final volume to 300 pL. A calculated volume of the stock solution of zinc nitrate solution containing 10% of the total number of millimoles of rac-DMSA or meso-DMSA was added, and the proton NMR spectrum was acquired. The sequential addition of the zinc nitrate solution was continued until a zinc: ligand ratio 1.2:l was reached.
(B)Proton NMR Titration of Znzf-ruc-DMSA System with NaOD. A solution for proton NMR spectroscopy was prepared in a 5-mm NMR tube by mixing a weighed amount of rac-DMSA, a measured volume of zinc nitrate stock solution corresponding to half the number of millimoles of rac-DMSA, 10 pL of 10%TBA, and a calculated volume of D2O to bring the final volume to 300 pL. A volume of NaOD corresponding to 20% of the number of millimoles of rac-DMSA was added to the NMR tube, successively before the proton NMR spectra were acquired after swirling the solution in the tube. The pD values of the above solutions at the various points in the proton NMR titration were determined with the aid of a glass calomel electrode pair by carrying out a parallel experiment under exactly the same conditions in a small test tube. The glass electrode was calibrated in the same manner as described previously (91,and the pD values reported are pH meter readings and have not been converted to [D+]values in mom. (C) Variable-Temperature13CN M R of [(Me)ms[Zn(racDMSAM in Methanol-&. The variable-temperature 13CNMR
studies were carried out with a solution containing 25 mg of the compound in 0.7 mL of deuterated methanol containing 0.03% TMS (v/v). Peaks arising from methanol do not interfere with the peaks arising from the ion association complex in the
884 Chem. Res. Toxicol., Vol. 7, No. 6,1994
aliphatic carbon region. The carbonyl carbon signal was not detected in the background noise because the methyl carbon signal of the TMA cation was so intense that it raised the detection limit of the analog-to-digital converter, and the acquisition time was not set long enough to record the slow relaxation of the carbonyl carbon signal under our experimental conditions. The chemical shifts were recorded relative to the TMS signal. Infrared Spectroscopy of (~uc-DMSA)~-and (mesoDMSA)4- at Different L4-/Zn2+Ratios. Infrared (IR)’ spectra of aqueous solutions containing varying ratios of zinc to ligand were measured with a Nicolet Fourier transform IR spectrometer Model 510P at ambient temperature. A calcium fluoride absorption cell, which has a transmission range from 66 666 to 1110 cm-l, equipped with a thin Teflon spacer was employed, and the thickness of the cell, which was experimentally determined from the interference fringes (141, was 31 pm. Since liquid water possesses strong IR absorption bands near 3300 and 1600 cm-l (15,161, all solutions were prepared in DzO, which shifts the angular vibration frequency of 0-H from 1640 to 1210 cm-l as a result of the replacement of hydrogen atoms with deuterium. This left the carboxylate stretching vibration free of the solvent absorption band. The sample chamber was purged with dry air during the measurements, and all spectra of the IR titration solution were recorded using air as background. The resolution of the IR spectra was 2 cm-1. The spectra of IR titration solutions were plotted after subtraction of the spectrum of DzO containing 0.20 M HzO from the recorded spectra followed by subtraction of the water vapor spectrum. Stock solutions of 0.10 M tetrasodium DMSA of the two DMSA isomers were prepared by dissolving 0.40 mmol of the appropriate DMSA in 4.0 mL of DzO containing 0.40 M NaOD. L4-:zinc 1:lstock solutions of 0.10 M for the two DMSA isomers were prepared by dissolving 32.3 mg (0.143 mmol) of zinc bromide in 1.43 mL of 0.10M tetrasodium DMSA solution. IR spectroscopy studies of DMSA solutions containing varying L4-: zinc ratios were carried out by acquiring the IR spectra of solutions which were prepared by mixing 0.10 M stock ligand, L4-, with 0.10 M stock L4-:zinc 1:l solution in different ratios.
Fang and Fernando
A12
All
I
I
I
1
4.2
4.4
1 8 1 4 1 3.4 3.2 3.0
I
4.0
3.8
3.6
Chemical shift vs DSS (ppm)
Figure 2. Proton NMR spectra of the aliphatic region of 0.081 mmol of rac-DMSA, L4-, at zinc:ligand molar ratios ranging from 0.0 to 1.2 (spectra AO-A12); the increment of the zinc:ligand ratio is 0.1.
A0 A8
1
A7 A6
A)
A5 A4
~
A
Results
NMR Spectroscopy. The utility of proton NMR spectroscopy for probing the protonation sites of polyfunctional ligands and the conformations of metal complexes in aqueous solution has been amply demonstrated during the past three decades. Kinetic information on ligand exchange and mechanistic information may be extracted (17,18)from the line widths or bandshapes of the N M R signals at varying temperatures. In our studies the proton NMR spectra of rac- or meso-DMSA were recorded as a function of zincligand ratio. The completely deprotonated ligand, L4-, was titrated with zinc nitrate, and the spectra are plotted in Figure 2 for racDMSA and in Figure 3 for meso-DMSA to show the change in bandwidth in the course of the titration; the chemical shifts of methine proton peaks are plotted as a function of millimolar ratios of zinc and DMSA isomers (Figure 4). The possible reaction of DMSA ligands, L4-, with zinc ion in the course of addition of zinc nitrate solution is schematically illustrated in Figure 5. The DzO solution of zinc nitrate is acidic and can be reasonably represented as a mixture of ZnOD+ and D+ for a discussion o f the chemical shifts that are observed when aliquots of the zinc nitrate solution were added to the solution of L4-. The methine proton resonance of the L4- of rac-DMSA alone occurred as a singlet at 3.08 ppm (A0 in Figure 2). Addition of zinc nitrate [(a) in Figure 51 gave rise to another peak at 3.14 ppm (A1 in Figure 21, Le., a 0.06 ppm downfield shift as a result of the complexation. The
1
~
~
~
4.4
4.2
4.0
3.8
l 3.6
~ 3.4
l 3.2
~
l
3.0
Chemical shift vs DSS (ppm)
Figure 3. Proton NMR spectra of the aliphatic region of 0.081 mmol of meso-DMSA, L4-, at zinc:ligand molar ratios ranging from 0.0 to 1.0 (spectra AO-A10); the increment of the zinc: ligand ratio is 0.1.
methine protons in ZnLzOD7- underwent a slow exchange with the L4- and its protonated form, DL3-, and the uncomplexed ligand resonance was shiRed slightly downfield, supporting the formation of the DL3- and its fast exchange with the L4-. After the addition of 0.5 equiv of zinc nitrate [(b) in Figure 51, the resonance signal arising from the L4- and the DL3-completely disappeared, indicating that all the uncomplexed ligand was converted to ZnLZ6-, which gave rise to the singlet at 3.19 ppm (A5 in Figure 2). Further addition of zinc nitrate caused a downfield chemical shift and a significant broadening of the peak (A6-Al0 in Figure 2). This was attributed to the formation of ZnL2- [ ( c )in Figure 51, which was likely undergoing a ligand exchange with ZnLz6- at a n intermediate rate at the experimental temperature, relative to the NMR time scale (200 ms) of methine protons in the ligand. Therefore, no separate peaks were resolved €or ZnL2- and ZnLZ6-, but instead, peak coalescence was observed. After 0.75 equiv of zinc nitrate was added, a new peak, well separated from the resonance peak due to ZnL2- and ZnLZ6-, emerged at 4.20 ppm (A8 in Figure
~
I
Chem. Res. Toxicol., Vol. 7, No. 6, 1994 885
Zinc-DMSA Complexes 4.4
4.2 T
f, 4.0
8 v,
v,
Q 3.8
2 E
5 3.6 3 \
E
8
3.4
3.2
A7
A6
3.0 0.0
O
I
~
0.4
0.2
I
0.6
O
'
1.0
0.8
I
AS
~
A4 A3
;
mmol zinc/mmol DMSA
Figure 4. Chemical shift of the methine protons as a function
k/
A2
of zinc:ligand molar ratio. (B) Observed for the zinc-ruc-DMSA solutions; ( 0 )observed for the zinc-meso-DMSA solutions.
h l
a)
b)
L4-
ZnOD+ + D+ ZnbOD7ZnOD+
c)
e)
+
+
+ DL3.
+ DL"
A: .;:E
Lk
D+
L4-
2ZnLf
2ZnL''
+ D+ H : : Z : L z + ZnLr
ZnOD+
+
ZnOD' d)
+
-kZnLzOD7-
D+
+ ZnL,G'
HT
2Dzo Zn,L?
2 Zn2L: ZnOD'
+ D+ X
: : L : 3 -
+
Zn,L>
Figure 5. The proposed reactions of DMSA, L4-, with zinc nitrate after (a) 0.25 equiv of zinc added; (b) 0.5 equiv of zinc added; (c) 0.75 equiv of zinc added; (d) 1.0 equiv of zinc added; (e) 1.25 equiv of zinc added.
21, which corresponded to the formation of a dimeric complex, ZnzLz4- [(d) in Figure 51. A ligand exchange phenomenon was observable between ZnL2-, ZnL26-, and ZnzLZ6-, which resulted in peak broadening, and the resonance peak due to ZnL2- and ZnLz6- gradually diminished, leaving only one peak due to ZnzLz4- at 4.21 ppm (A10 in Figure 2). When more than 1equiv of zinc nitrate was added, the resonance peak shifted slightly downfield to 4.28 ppm (Figure 21, as a result of formation of the protonated dimeric complex DZn2L23-[(e) in Figure 51. The interaction of the meso-DMSA ligand, L4-, with zinc ion in the course of the addition of zinc nitrate solution is assumed to be similar to that of the ruc-DMSA ligand, a s illustrated in Figure 5. The methine proton resonance of the L4- alone occurred as a singlet a t 3.07 ppm (A0 in Figure 3). Addition of zinc nitrate yielded another peak at 3.59 ppm, i.e., a 0.52 ppm downfield shift as a result of the complexation (A1 in Figure 3). The proton resonance resulting from pure ligand completely disappeared upon addition of 0.5 equiv of zinc nitrate, indicating the formation of ZnLZ6- (A5 in Figure 3). Further addition of zinc nitrate up to 1.0 equiv caused a slight downfield shift of the peak to 3.72 ppm; however, a significant peak broadening occurred after 0.7 equiv of
4.4
,
l 4.2
,
I 4.0
I
I 3.8
A0
I
I 3.6
J
1 3.4
l
l 3.2
l
l
l
3.0
Chemical shift vs DSS (ppm)
Figure 6. Proton NMR spectra of the aliphatic region of 0.08 mmol of ruc-DMSA in the presence of 0.04 mmol of Zn2+ at Na0D:ruc-DMSA ratios ranging from 0.0 to 4.2 (spectra AOA21); the increment of the molar ratio of Na0D:ruc-DMSA is 0.2.
zinc nitrate was added and reached a maximum a t a zinc: ligand ratio around 0.8 (A7 and A8 in Figure 3). This might be an indication that only Zn2Lz4- was formed upon addition of zinc nitrate to ZnLz6- and that it underwent a n intermediate ligand or metal exchange with ZnLZ6-, resulting in a maximized peak broadening when the concentrations of both species were close to each other. There was no conclusive evidence, however, to exclude the formation of the monomeric complex, ZnL2-. A proton NMR titration of ruc-DMSA with NaOD, shown in Figure 6 , was carried out for ruc-DMSA in the presence of Zn2+a t a 1igand:zinc ratio of 2:l to confirm the formation of the zinc complexes of ruc-DMSA, which were predicted in potentiometric studies (11). The chemical shifts plotted as a function of pD values are shown in Figure 7. With the aid of the proton NMR results previously obtained from the titration of L4- with zinc nitrate and from the titration of &L with NaOD (91, the proton NMR results obtained from the titration of Zn2+-ruc-DMSA with NaOD are interpreted as follows: a t pD values greater than 12.5 only one sharp signal from the methine protons in ruc-DMSA was observed a t 3.15 ppm (shown in Figure 7, but not in Figure 61, which indicated that the ZnLzOD7-was the predominant species in solution; when pD was reduced, the resonance peak started to broaden and shift downfield as a result of the loss of the deuteroxy group of ZnLzOD7-and the subsequent formation of ZnL&; at pD about 7.3 (A21 in Figure 6), another resonance peak emerged a t 3.75 ppm which corresponds to the protonation of ZnLZ6-; a decrease in pD from 7.0 to 6.0 resulted in another new resonance peak a t 4.01-4.24 ppm which is direct evidence for the presence of dimeric complexes in this pD region. The change in the chemical shift and in the width of the methine proton peak located between 4.01 and 4.24 ppm implied that the protonated dimeric complexes may be present and undergo a n intermediate
886 Chem. Res. Toxicol., Vol. 7, No. 6, 1994
Fang and Fernando
200
4.4
56.3.
4.0
E
140
3.6 8 v)
120
3.2
‘
z 56.1 8 v)
s
E 56.0.
,p
E
SJ
2.8
56.2.
2
-.2
$j 5 5 . 9 .
$
2.4
55.8.
6 55.7.
2.0 0
2
4
6
8
10
12
14
55.6
-
PD
Figure 7. Plots of the chemical shift of the methine protons and the speciation curves, calculated by the “BEST” program, as a function of pD for solutions containing 0.08 mmol of rucDMSA and 0.04 mmol of Zn2+. Table 1. Chemical Shifts of the Methine Protons in Zinc Complexes of rac- and meso-DMSAvs DSS meso-DMSA
rac-DMSA
species L4- ZnLZ6- ZnzLz4- L4- ZnL6- DZnL6- ZnzL4- DZnzb3ppm 3.07 3.59 3.72 3.08 3.19 3.75 4.21 ca.4.28
ligand exchange among themselves as well as with the various protonated forms of the free ligand in this pD region. A further decrease in pD resulted in the disappearance of the two resonance peaks upfield and downfield, leaving only the middle resonance peak around 3.74 ppm which arose from the free ligand in its various stages of protonation as deduced from our previous proton NMR studies of ruc-DMSA (9). By ignoring the isotope effect of deuterium on the pK, values of ruc-DMSA, the distributions of various ionic and molecular species as a function of pD in the Zn2+-rm-DMSA system used in the NMR titration are calculated and shown in Figure 7 by using the speciation model and the stability constants obtained from potentiometric studies (11). The NMR results confirmed the model (11)because they do predict, as shown in Figure 7, t h a t ZnL2OD’- and ZnLf- predominate a t high pD values; DZnL25- predominates a t pD around 7.3; dimeric complexes are formed at pD between 6.0 and 7.0; and various protonated ligand species are predominant in the low pD region. By comparing the species distribution model and NMR results of the NaOD titration, an “abnormal” upfield shift from around 4.19 to around 3.69 ppm (i.e., the disappearance of the resonance peak around 4.2 ppm) was observed for the dimeric complex, D2Zn2Lz2-. This will be discussed later in the Discussion section. The chemical shifts of the methine protons of DMSA in some zinc complexes are summarized in Table 1. Variable-TemperaturelSC NMR. This experiment was originally designed to study the structures of the zinc-DMSA complexes on the basis of the splitting pattern of the 13C resonance peaks in the coordination complexes (19). I t was observed t h a t in the aliphatic region of [(Me)4Nl~ZnLzthere were only two resonance peaks at ambient temperature, which corresponded to the methyl carbon in the tetramethylammonium (TMA)l cation and the methine carbon in the ligand. As shown in Figure 8, both resonance peaks shifted upfield with a
55.5 250
I
1
I
I
260
270
280
290
300
Temperature (K)
Figure 8. 13C chemical shifts in the aliphatic region of [(Me)JVl6EZn(rac-DMSA)zlversus temperature: ( 0 ) methyl carbon of TMA cation; (0)methine carbon of Zn-rac-DMSA complex. decrease in temperature from 295-260 K, however, the expected splitting of the resonance peak in the ZnLZ6- of ruc-DMSA was unfortunately not observed even at 260 K. The upfield shift of methyl carbon in the TMA cation was caused by neutralization of the positive charge on the TMA cation. This indicates that the ion pair association equilibrium between the TMA cation and the ZnLz6complex anion shifts toward the formation of the ion pair, when the temperature decreases. As a result of ion pair association, a deshielding effect was expected on the methine carbon in ZnL2S-, and the methine carbon resonance in the ligand should, therefore, shift downfield; a n upfield shift was observed.instead. The extent of the shift of the methine carbon resonance with temperature (0.0063 ppm/K), however, is not as great as t h a t of the methyl carbon (0.01 ppm/K), (Figure 8). This “abnormal” upfield shift could be a result of change in the conformation of the complex with a decrease in temperature, but its exact origin is still in question. JR Spectroscopy. The correlation between the asymmetric stretching frequency (Y,) and symmetric stretching frequency ( Y ~ of ) a carboxyl group and their separation (Av)with the nature of the coordination of a carboxylate group has been well established (20) and has been used in the interpretation of the strength of the metal-oxygen bond (21,221 since the early 1960s. It has been summarized by Deacon et al. (21)that Av values 1200 cm-’ appear to be generally associated with unidentate coordination of a carboxylate group, in which only one oxygen atom of the carboxylate group participates in the metaloxygen bonding. The correlation between molecular structure and band intensity of the carbonyl group was demonstrated by Ramsay and Jones et al. (23,24) in the early 1950s; the use of band intensity, however, has not received significant attention since then, because its application is limited to bands free from overlap with other bands. The ratio of the intensity of the asymmetric stretching band to t h a t of the symmetric stretching band is about 0.8 (25)for a saturated sodium acetate aqueous solution, and about 1.4 (26) for a 50/50 (movmol) acetic acidlwater mixture. An increase in the ratio of intensity of the asymmetric band to that of the symmetric band
Zinc-DMSA Complexes
Chem. Res. Toxicol., Vol. 7, No. 6, 1994 887
I Abr = 0.7
.A
10
10
72M
9
8
8
A 5
1
1
~
I700
77'00
Id00
100
(
7400
'
"
'
I
.
13w
'
Wavenumbers(cm-?
Figure 9. IR spectra of 0.1 M zinc-ruc-DMSA solutions at zinc: ligand molar ratios ranging from 0.0 to 1.0; the increment of the zinc:ligand ratio is 0.1. Solid lines 0-10, represented as AOA10, are the experimental spectra; dot-dashed lines 1-5, represented as Bl-B5, are the calculated spectra for the ZnL9complex; broken lines 1-5 are the spectra resulting from the presence of the ligand in the solution; double-dot-dashed lines 6-10, represented as C6-Cl0, are the calculated spectra for the ZnzLz4- complex; dot-dashed lines 6-10 are the spectra resulting from the presence of ZnLZ6- in the solution.
relative to the ratio for the sodium carboxylate, in which the metal-oxygen bond is assumed to be completely ionic, can be an indication of an increase in the covalency of the metal-oxygen bond. By use of the PCDR software 3.10 supplied with the Nicolet FT-IR spectrometer Model 510P, i t is possible to obtain the spectrum of an unknown component from a spectrum of a multicomponent solution if the composition of the solution and the spectra of the other components are known. The frequencies and the intensity ratios of the asymmetric and symmetric bands have been used to elucidate the nature of the coordination bonds in the zinc complexes of DMSA. In our studies, the IR spectra of ruc- or meso-DMSA were obtained as a function of zinc: ligand ratio. The IR spectra, in the frequency range between 1740 and 1260 cm-', of the titration solutions in which the ratios of Zn:L range from 0 to 1 are shown in Figures 9 and 10. The asymmetry of the absorption bands in the spectra for the pure ruc-DMSA (L4-),ZnL&, and ZnzLz4-, respectively, originated from the presence of d- and 1-isomers of ruc-DMSA. The asymmetry of the absorption peak a t 1376 cm-l (C10 in Figure lo), for the ZnzLz4- of meso-DMSA, was attributed to the possible conformational isomers of ZnzLz4- of meso-DMSA. With successive addition of zinc ion to the ruc-DMSA solution the asymmetric band of ruc-DMSA shifted to a higher frequency and the symmetric band of the ligand shifted to a lower frequency (solid lines A0 through A10 in Figure 9), which indicated that new complexes were formed in the course of addition of zinc ions. As indicated by NMR results, before the zincligand ratio reached 0.5, only ZnLZ6- was formed in the titration solution. The IR spectra for ZnLz6- were obtained by subtraction of the ligand spectrum proportionally from the spectra obtained with the titration solutions a t zincligand ratios ranging from 0.1 to 0.5. The resulting spectra were the same,
1
1
1
1
~
7600
1
1
1
1
1
1
1
1500
1
1
1400
1
1
1
1
1
1
1300
Wavenumbers (cm-')
Figure 10. IR spectra of 0.1 M zinc-meso-DMSA solutions at zinc:ligand molar ratios ranging from 0.0 to 1.0; the increment of the zincligand ratio is 0.1. Solid lines 0-10, represented as AO-A10, are the experimental spectra; dot-dashed lines 1-5, represented as Bl-B5, are the calculated spectra for the ZnL26complex; broken lines 1-5 are the spectra resulting from the presence of the ligand in the solution; double-dot-dashed lines 6-10, represented as C6-Cl0, are the calculated spectra for the ZnzLz4- complex; dot-dashed lines 6-10 are the spectra resulting from the presence of ZnL26- in the solution.
confirming that ZnL$- of ruc-DMSA indeed was the only species formed in the solution. After more than 0.5 equiv of zinc ion was added, ZnzLz4- or ZnL2- may be formed. Since the stoichiometry of the two complexes was the same, their IR spectra could be obtained by subtraction of the ZnLZ6- spectrum (B5 in Figure 91, proportionally from the spectra obtained with the titration solutions a t zincligand ratios ranging from 0.6 to 1.0. The resulting spectra were the same with respect to the asymmetric bands, although the left shoulder of the peaks of the symmetric bands vaned slightly. This indicated that new complexes ZnzLz4- (or ZnL2-) of ruc-DMSA were formed in the solution. The separation between the asymmetric band and the symmetric band of the carboxylate groups in the ruc-DMSA alone (L4-) was 158 cm-l, and their intensity ratio was 1.08 (A0 in Figure 9). The separation between the asymmetric band and the symmetric band of the carboxylate groups in ZnLf of ruc-DMSA was 179 cm-l, and their intensity ratio was 1.26 (B5 in Figure 9). The increase in both the intensity ratio and the band separation indicated that the zinc ion in the complex was bound to the carboxylate group of ruc-DMSA. On the basis of a n increase in the intensity ratio by 0.18 and a n increase in the overall band separation by 21 cm-l, it was concluded that only one of the two carboxylate groups in ruc-DMSA participated in a unidentate coordination with the zinc ion in the complex. The separation between the asymmetric band and the symmetric band of the carboxylate groups in ZnzLZ4- of ruc-DMSA was 201 cm-l, and their intensity ratio was 2.39 (C10 in Figure 9). This indicated that both carboxylate groups in rac-DMSA participated in unidentate coordination with zinc ions in ZnzLZ4-. IR spectra obtained for the titration of meso-DMSA with zinc ion are shown in Figure 10. A changing pattern, similar to t h a t obtained with ruc-DMSA, of the asymmetric and symmetric bands with the addition of
1
1
888 Chem. Res. Toxicol., Vol. 7, No. 6, 1994 zinc ions was observed. The same approach was used as with ruc-DMSA to interpret the IR spectra of the titration solution at various zincligand ratios. The species ZnLz6- was formed before the zincligand ratio reached 0.5, and species ZnzLz4- (or ZnL2-) was formed in the titration solutions when the zincligand ratio was greater than 0.5. Similar to its racemic isomers, the separation between the asymmetric band and the symmetric band of the carboxylate groups in the meso-DMSA alone (L4-) was 159 cm-' and their intensity ratio was 1.06 (A0 in Figure 10). The separation between the asymmetric band and the symmetric band of the carboxylate groups in ZnLz6- of meso-DMSA was 193 cm-', and their intensity ratio was 1.35 (B5 in Figure 10). Since the Av value was very close to 200 cm-l, it was concluded that both carboxylate groups of meso-DMSA were bound to the zinc ion in the complex. The intensity ratio, however, is close to that observed for the ZnLz6- of ruc-DMSA, and the Av value was slightly less than 200 cm-l. On this basis it may be postulated that one of the two carboxylate groups of meso-DMSA in the complex may be bound to the zinc ion in an unsymmetrical bidentate manner, as shown in Figure 12 (I), in which the asymmetry of this bound carboxylate group is slightly reduced with respect to a carboxylate group which is bound in a unidentate manner. Similar to the corresponding complex of its racemic isomers, the separation between the asymmetric band and the symmetric band of the carboxylate groups in ZnzLz4-of meso-DMSA was 201 cm-l and their intensity ratio was 2.17 (C10 in Figure 10). This indicated t h a t both carboxylate groups in meso-DMSA participated in unidentate coordination with zinc ions in Z ~ Z L ~ There ~ - . was another small absorption peak that appeared a t 1288 cm-' upon addition of zinc ions to meso-DMSA (Al-A10 in Figure 101, and the peak started to shift t o higher frequency when the zincligand ratio became greater than 0.8. The intensity of this peak reached a maximum a t zinc:ligand ratio 0.5. The origin of this peak is unclear.
Discussion The formation of various monomeric complexes and dimeric complexes of meso- and ruc-DMSA has been shown by the results obtained with proton NMR spectroscopy and IR spectroscopy. The speciation model, previously determined (11 )for the zinc-ruc-DMSA system, has been confirmed by the results of proton NMR spectroscopy as a function of pD (Figure 7). Zinc-DMSA solution is a dynamic system in which there are many conformations of zinc complexes of DMSA. The complexation of ruc-DMSA with zinc ion is even more complicated than that of meso-DMSA because ruc-DMSA is a mixture of equal amounts of two enantiomeric isomers, i.e., (RP)DMSA and (S,S)-DMSA. Despite the complexity of the system, the NMR and IR spectra of zinc complexes are mainly influenced by the major conformations present in solution and can provide conformational information on the complexes. Potentiometry is another experimental probe which can determine whether thiolate groups in DMSA complexes are free or coordinated to a metal ion. In the DMSA ligand, L4-, there are two types of protonation sites, carboxylate groups and thiolate groups, the protonation constants (log K values) of which are typically less than 4.0 and greater than 9.0, respectively. If a protonation constant larger than 9.0 is obtained potentiometrically for a zinc complex, there must be a free
Fang and Fernando -0oc
,$. b, 0
0
(IV)
(V)
Figure 11. Proposed conformations of zinc-ruc-DMSA complexes in an aqueous solution. (I) ZnLz"; (11) DZnLz5-; (111) DzZnLz4-; (IV)Z ~ Z L ~ (V) ~ DZnzLz3-. -;
U'
-S
(1)
( 11 )
Figure 12. Proposed conformations of zinc-meso-DMSA complexes in an aqueous solution. (I) Znze-; (11) Zn2L#-. thiolate group present in the complex. If a protonation constant between 5.0 and 8.0 is obtained for a zinc complex, there is no free thiolate group available, and only coordinated thiolate groups are present in the complex. The decrease in the protonation constant of a coordinated thiolate group, compared with that for a free thiolate group in the zinc complexes, results when the protonation of the thiolate group is preceded by an energetically unfavorable coordination bond-breaking process. On the basis of NMR and IR results presented in the results section and with the aid of a molecular model, the representative and energetically favorable conformations of various zinc complexes of meso- and rucDMSA are shown in Figures 11 and 12. Although the conformation of Zn[(R&)-ruc-DMSAl[(S,S)-ruc-DMSAl purports to represent ZnLz6- [Figure 11 (I)], Zn[(R,R)-ruc-DMSA]z and Zn[(S,S)-ruc-DMSAlz may also be present in solution, and their conformations are very similar to the conformation of Zn[(RP)-rucDMSA][(S,S)-ruc-DMSA] shown in Figure 11. As shown in Figure 4, the net downfield shift resulting from the formation of ZnLf of rac-DMSA is only 0.11 ppm, which is smaller than that due to the protonation of a thiolate group, 0.14 ppm (91, or smaller than that due to the protonation of a carboxylate acid group, 0.20 ppm (12), or smaller than t h a t due to the formation of ZnL.2- with meso-DMSA, 0.53 ppm. This implies that the deshielding
Zinc-DMSA Complexes effect by Zn2+ is not the only effect responsible for the chemical shift change. The ZnLz6- most likely adopts a conformation in which the methine proton of the ligand is oriented in the shielding region of the carbonyl double bond, which, to some' extent, cancels out the deshielding effect of Zn2+. As illustrated in Figure 11(I), each ligand in the complex forms one five-membered ring in which one carboxylate group and its adjacent thiolate group are bound to the zinc ion, another five-membered ring in which both thiolate groups are bound to the zinc ion, and a six-membered ring in which one carboxylate group and one thiolate group, a t the P-position relative to the carboxylate group, are bound to the zinc ion. The unbound carboxylate group can rotate freely around the C-C single bond as it does in the free ligand (L4-). With the aid of a molecular model, it is found that the methine protons [Ha in Figure 11 (I)] are located out of the deshielding plane of the C=O bond of their contiguous carboxylate groups, where they are exposed to a n additional shielding effect (91, because the six-membered ring adopts a rigid boat conformation as a result of zincsulfur bridging through the thiolate group a t the a-position relative to the carboxylate group participating in the six-membered ring. This cancels out in part the deshielding effect of the zinc ion on the methine protons of the ruc-DMSA and accounts for the small downfield shift, 0.11 ppm, of the methine proton peak in the ZnLZ6- from t h a t in the pure ruc-DMSA ligand. The meso-DMSA ligand in ZnLZ6-, as shown in Figure 12 (II), uses two carboxylate groups and one thiolate group to form a fivemembered ring, a six-membered ring, and a sevenmembered ring. It is found with the aid of a molecular model that, in contrast to the ZnLZ6- of ruc-DMSA, the methine protons [Ha in Figure 12 (I)] are located in the deshielding plane of the C=O bond of its contiguous carboxylate groups, where they are exposed to a n additional deshielding effect, because the six-membered ring adopts a chair conformation to minimize the electrostatic repulsion between the oxygen of the carboxylate group and the negatively charged free thiolate group. This accounts for the drastic downfield shift, 0.52 ppm (Figure 4), of the methine proton peak in the ZnLz6- from that in the pure meso-DMSA ligand. The proposed conformations of ZnLZ6-of ruc- and meso-DMSA are also confirmed by the results obtained from potentiometric studies (22). The protonation constant (log K value) of ZnLZ6- of meso-DMSA is 10.82, which indicates that a free thiolate group must be available in the complex for protonation, whereas the protonation constant (log K value) of ZnLZ6- of ruc-DMSA is 7.59, which indicates that the thiolate groups in the complex are all bound to the zinc ion. As shown in Figure 11(N), each zinc ion in the dimeric complex ZnzLz4- of ruc-DMSA is tetrahedrally coordinated with two carboxylate groups and two thiolate groups, forming two six-membered rings of the same type. In each six-membered ring structure, one carboxylate group and one thiolate group, a t the ,&position relative to the carboxylate group of the same ligand, are bound to the same zinc ion. There are four six-membered rings formed in the complex, and all the six-membered rings adopt rigid boat conformations as a result of cross-linking via two central zinc ions between two ligands. It is evident, in a molecular model, that all the methine protons are located in the deshielding planes of the C=O bonds of the carboxylate groups in this conformation of ZnzLz4- of ruc-DMSA. This accounts for the drastic
Chem. Res. Toxicol., Vol. 7, No. 6, 1994 889 downfield shift, 1.02 ppm, observed after addition of one more zinc ion on ZnLf of ruc-DMSA (Figure 4). Although the conformation of Zn[(S,S)-ruc-DMSAlzrepresents ZnzLz4- in Figure 11(IV), Z~Z[(R,R)-~UC-DMSAIZ is likely present a t the same concentration, and their conformations are very similar to each other. The possibility of Znz[(S,S)-ruc-DMSA][(R,R)-ruc-DMSA] is excluded because of symmetry restriction and steric hindrance. If Znz[(S,SI-ruc-DMSAI[(R,R )-rac-DMSAlwas formed, a mirror plane would be expected to pass through the two zinc ions, and the zinc ions would be positioned a t the apices of pyramids; this, however, violates the tetrahedral coordination of the zinc ion in the complex. The zinc ion in the dimeric complex ZnzLz4- of mesoDMSA is also tetrahedrally coordinated with two carboxylate groups and two thiolate groups [Figure 12 (II)], but because of steric hindrance, each ligand uses two carboxylate groups and one thiolate group to bind to one zinc ion and uses the other thiolate group t o bind to the other zinc ion in the complex. Compared with the conformation of the ligand in the ZnLZ6- of meso-DMSA, there is not much change in the magnetic environment of the methine protons after the formation of ZnZLz4-, except t h a t the free thiolate group in the ZnLZ6- is coordinated with a zinc ion in the ZnzLz4-. This accounts for the small downfield shift, 0.13 ppm (Figure 41, of the methine proton peak, after addition of another zinc ion to ZnLf of the meso-DMSA. The absence of free thiolate groups in the conformations of the dimeric complexes of ZnzLZ4-of both ruc- and meso-DMSA is confirmed by the results obtained from potentiometric studies (22). The protonation constants (log K values) of ZnzLz4- of mesoand ruc-DMSA are 6.31 and 7.47, respectively, which indicates that all thiolate groups iLiboth complexes are coordinated to zinc ions. The appearance of new peaks a t 3.75 ppm (A21-Al9 in Figure 6), which are assigned to DZnLz5- by the speciation model (Figure 7), implies t h a t the conformation of the ligand in the complex DZn1z5- of ruc-DMSA may be similar to t h a t in the ligand, DzL2-, because the chemical shift of the methine protons in the D2L2- of rucDMSA is 3.67 ppm (9). It is postulated, therefore, t h a t in DZnLz5- the zinc ion is tetrahedrally bound to two ligands via one thiolate group and one carboxylate group from each ligand as shown in Figure 11 (11). The presence of the free thiolate group in the DZnlZ5-complex is also confirmed by its protonation constant (log K value), 10.26 (22). The conformation of DzZnLz4- [Figure 11(11111 is postulated on a reasonable assumption that another similar six-membered ring is formed upon deuteration of the free thiolate group in DZnLz5-. The enantiomeric identity of ruc-DMSA in the deuterated monomeric complexes is not clear, and it is likely that the complexes contain either (S,S)-ruc-DMSA or (R,R)-ruc-DMSA or both. The protonation constant (log K value) of ZnzLz4- is 7.47 (121, which implies that deuteration of ZnzLz4- only results in a simple deuteration of one of the bound thiolate groups in ruc-DMSA without causing a significant change in conformation. Therefore, the conformation of DZnzLz3- of ruc-DMSA should be very similar to that of ZnzLz4- of ruc-DMSA. The proposed conformation of DZnzLZ3-, shown in Figure 11 (V), is confirmed by the proton NMR results. It has been observed that the downfield shift of methine protons caused by deuteration of one adjacent thiolate group in DMSA was 0.14 ppm (9), but only a comparatively small downfield shift, 0.07
890 Chem. Res. Toxicol., Vol. 7, No.6,1994
ppm, was observed as a result of deuteration of one thiolate group in Zn2Lz4-. Compared with the rigid conformation of Zn2Lz4-, in which all the methine protons of the ligand are in the deshielding plane of the C=O bonds, a small flexibility is expected in the conformation of DZnzLz3- because one S-Zn bond is disrupted. "his induced flexibility in conformation lessens the deshielding effect of C=O bonds on the methine protons in DZnzLz3- compared with the protons in ZnzLz4- and accounts for the observed small downfield shift. Deuteration of DZnzLz3- of ruc-DMSA is assumed to take place a t one bound thiolate group, because deuteration of one bound carboxylate group would be indicated by a protonation constant (log K value) much smaller than 4.0. Formation of DzZn2Lz2- upon addition of one deuterium ion on DZn2Lz3- causes a significant change in conformation, because the protonation constant (log K value) of DZnzLz3-is small, between 4.26 and 3.96 (ll),compared with those of the other zinc complexes of DMSA. The additional deshielding effect, which is present in the ZnzLz4- and DZnzLz3-, of the C=O bonds on the methine protons is completely lost in the DzZnzLz2- since the complex completely loses the rigidity of the boat conformation present in ZnzLz4- as a result of the disruption of two S-Zn bonds. This accounts for the disappearance, rather than a further downfield shift, of the typical dimeric complex peak, when DZn2Lz4- is further deuterated a t pD around 5.0 (Figure 7). In a n analysis of the conformations of various zinc complexes of DMSA proposed in Figures 11and 12, it is found that, in all the zinc complexes of rac-DMSA in aqueous solution, the ligand always adopts staggered anti conformations, whereas in all the zinc complexes of mesoDMSA, the ligand always adopts staggered gauche conformations. The zinc complexes of ruc-DMSA in aqueous solution are more stable than the corresponding complexes of its meso isomer because the electrostatic repulsion between two bulky carboxylate groups is minimized in the staggered anti conformations. This accounts for the larger formation constants of zinc complexes of rac-DMSA than those of the corresponding complexes of its meso isomer and also explains, in part, why an increase in the excretion of endogenous zinc was observed in the rue-DMSA.
Acknowledgment. This work was supported in part by NIEHS Grant ES 03356. References (1) Davis, J. M., and Svensgaard, D. J. (1987)Low-level lead exposure and child development. Nature 329,297-300. (2)Mucshak, P., Davis, J., Crochetti, A., and Grant, L. (1989) Prenatal and postnatal effects of low-level lead exposure: integrated summary of a report to the US. congress on childhood lead poisoning. Enuiron. Res. 50, 11-36. (3)Needleman, H.L.,Shell, A., Bellinger, D., Leviton, A., and Allred, E. N. (1990)The long term effects of exposure to low doses of lead in childhood. N. Engl. J. Med. 322, 83-88. (4)Schwartz J., and Otto, D. (1987)Blood lead, hearing thresholds, and neurobehavioral development in children and youth. Arch.
Fang and Fernando Enuiron. Health 42, 153-160. ( 5 ) Winneke, G., Brockhaus, A., Ewers, U., Kramer, U., and Neuf,
M. (1990)Results from European multicenter study on lead neurotoxicity in children: implication for risk assessment. Neurotoxicol. Teratol. 12,553-559. (6)US.Center for Disease Control (1985)Preuenting Lead Poisoning in Young Children (No. 99-2230)US.Department of Health and Human Services, Atlanta, GA. (7)US.Center for Disease Control (1991)Preuentinghad Poisoning in Young Children, US. Department of Health and Human Services (Report), Atlanta, GA. (8)Agency for Toxic Substances and Diseases Registry, Public Health Service (1988) The Nature a n d Extent of Lead Poisonfng in Children in the United States: A Report to Congress, U.S. Department of Health and Human Services, Atlanta, GA. (9)Fang, X.,and Fernando, Q. (1994)Synthesis, structure, and properties of rac-2,3-dimercaptosuccinicacid, a potentially useful chelating agent for toxic metals. Chem. Res. Toxicol. 7,148-156. (10) Egorova, L. G., Okonishnikova, I. E., Nirenburg, V. L., and Postovskii, I. Y. (1972)Complex formation of certain metals with dimercaptosuccinic acid stereoisomers. Xhim. -Farm. Zh. 6, 1416. (11) Fang, X.,and Fernando, Q.(1994)A comparative study of mesoand rac-2,3-dimercaptosuccinic acids and their zinc complexes in aqueous solution. Chem. Res. Toxicol. (in press). (12)Sudmeier, J. L., and Reilley, C. N. (1964)Nuclear magnetic resonance studies of protonation of polyamine and aminocarboxylate compounds in aqueous solution. Anal. Chem. 36,1698-1706. (13)Letkeman, P., and Westmore, J. B. (1971)Metal-aminopolycarboxylic acid complexes. I. Studies of lead(I1) and cadmium(I1) complexes with diethylenetriaminepentaacetic acid in aqueous solution by proton magnetic resonance spectroscopy. Can. J. Chem. 49,2073-2086. (14)Grim, W.M.,111, Fateley, W. G., and Grasselli, J. G. (1984)Fourier Transform Infrared Spectroscopy (Theophanides, T., Ed.) pp 2542,D. Reidel Publishing Co., Dordrecht. (15) Dorsey, N. E. (1940)Properties of Ordinary Water Substance, p 341,Reinhold Publishing Corp., New York. (16)Gore, R. C., Barnes, R. B., and Petersen, E. (1949)Infrared absorption of aqueous solutions of organic acids and their salts. Anal. Chem. 21,382-386. (17)Pople, J.A., Schneider, W. G., and Bernstein, H. J. (1959)Highresolution Nuclear Magnetic Resonance, Chapter 10, McGrawHill, New York. (18)Caldin, E.F. (1964)Fast Reactions in Solution, Chapter 11,John Wiley & Sons, New York. (19)Sandstrom, J. (1982)Dynamic NMR Spectroscopy, Academic Press, New York. (20)Rivera, M., and Fernando, Q. (1992)Synthesis, structure, and stability of the zinc complex of the dimethyl ester of meso-2,3dimercaptosuccinic acid. Chem. Res. Toxicol. 5, 142- 147. (21)Deacon, G. B., and Phillips R. J. (1980)Relationships between the carbon-oxygen stretching frequencies of carboxylate complexes and the type of carboxylate coordination. Coord. Chem. Rev. 33,227-250. (22)Nakamoto, IC, Morimoto, Y., and Martell, A. (1961)Infrared spectra of aqueous solutions. I. Metal chelate compounds of amino acids. J . Am. Chem. SOC. 83,4528-4532. (23)Nakamoto, K.(1978)Infrared and Raman Spectra of Inorganic and Coordination Compounds, Part 111,John Wiley & Sons, New York. (24)Ramsay, D. A. (1952)Intensities and shapes of infrared absorption bands of substances in the liquid phase. J. Am. Chem. SOC.74, 72-80. (25)Jones, R. N., Ramsay, D. A., Kier, D. S., and Dorbriner, K. (1952) The intensities of carbonyl bands in the infrared spectra of steroids. Ibid. 74, 80-88. (26)Ito, K.,and Bernstein, H. J. (1956)The vibrational spectra of the formate, acetate, and oxalate ions. Can. J . Chem. 34, 170-178. (27)Hadzi, D., and Sheppard, N. (1953)The infrared absorption bands associated with the COOH and COOD groups in dimeric carboxylic acids. Proc. R. SOC. (London) A216, 247-266.
