A Comparative Study of meso- and ruc-2,3-Dimercaptosuccinic Acids

The conformations of meso-2,3-dimercaptosuccinic acid (meso-DMSA) in aqueous solution have been postulated from proton NMR titrations. The complexes ...
2 downloads 0 Views 1MB Size
Chem. Res. Toxicol. 1994, 7, 770-778

770

A Comparative Study of meso- and ruc-2,3-DimercaptosuccinicAcids and Their Zinc Complexes in Aqueous Solution Xiaojun Fang and Quintus Fernando* Department of Chemistry, University of Arizona, Tucson, Arizona 85721 Received March 28, 1994@ The conformations of meso-2,3-dimercaptosuccinicacid (meso-DMSA) in aqueous solution have been postulated from proton NMR titrations. The complexes formed with zinc(I1) and the ligands ruc-DMSA and meso-DMSA have been postulated from potentiometric titrations of solutions containing varying ratios of zinc:ligand. The complex formation behavior of MCDMSA with zinc(I1) is dramatically different from t h a t of meso-DMSA. These differences are reflected in the complex formation constants of the zinc(I1) species and their distributions in solution as a function of zinc:ligand ratios and as a function of pH. On the basis of these results we have predicted that rue-DMSA is more effective than meso-DMSA in mobilizing lead in vivo and that the 1:2 zinc complex of rac-DMSA can be used effectively for the treatment of heavy metal poisoning because endogenous zinc will not be depleted by this chelation treatment.

Introduction Cadmium is one of the most troublesome toxic metals because, unlike other toxic metals, it is excreted very slowly from the body and has an extremely long biological half-life (1). At the present time there is no acceptable chelating agent for the treatment of cadmium poisoning. An assessment by Jones et al. (2) of several chelating agents, which included 2,3-dimercaptopropanol (British anti-Lewisite; BAL),l 2,3-dimercaptopropane-l-sulfonic acid (unithiol),l meso-2,3-dimercaptosuccinicacid (mesoDMSA),' D-penicillamine, diethyldithiocarbamate, EDTA,l and diethylenetriaminepentaaceticacid (DTPA),' for their effectiveness in mobilizing cadmium, revealed that meso-DMSA and unithiol are likely to be the most useful ligands for the treatment of acute cadmium intoxication by complexing extracellular cadmium and promoting its excretion in the urine. BAL appears to be effective in reducing mortality in cases of acute exposure to cadmium, by directing the cadmium away from certain sensitive loci, and also in chronic cadmium poisoning, by mobilizing both hepatically and renally deposited cadmium even after the initiation of metallothionein synthesis (3, 4). The toxicity of BAL (LD50 = 0.9 mmolkg) is caused by the neutrality of the cadmium complex of BAL, which promotes its accumulation in the kidney where it enhances the nephrotoxicity of cadmium by causing severe renal damage. The utility of BAL, therefore, for the treatment of cadmium poisoning is severely limited. The toxicity of meso-DMSA, which is a n analog of BAL, is much lower [LD50 = 13.73 mmolkg (511, as a result of the presence of two carboxylic acid groups in the molecule which increase its water solubility. meso-

* To whom correspondence should be addressed.

Abstract published in Advance ACS Abstracts, October 1, 1994. Abbreviations: BAL, 2,3-dimercaptopropanol;meso-DMSA, meso2,3-dimercaptosuccinic acid; DTPA, diethylenetriaminepentaacetic acid; FDA, U.S.Food and Drug Administration; rac-DMSA, rac-2,3dimercaptosuccinic acid; TBA, tert-butyl alcohol; DSS, sodium 2,2dimethyl-2-silapentane-5-sulfonate; unithiol, 2,3-dimercaptopropane1-sulfonic acid; PMI, plasma mobilizing index; RPMI, relative plasma mobilizing index; (R,R)-DMSA, (R,R)-2,3-dimercaptosuccinic acid; (S,S)-DMSA, (S,S)-2,3-dimercaptosuccinic acid. @

DMSA has also been shown to be effective in the treatment of mercury (6,7) and lead (8,9) poisoning and has been approved by the U S . Food and Drug Administration (FDA). In vivo studies of the efficacy of mesoDMSA in the treatment of cadmium poisoning have shown that meso-DMSA mobilizes the hepatic deposits and reduces the renal uptake of cadmium effectively when given immediately after ingestion of the metal (10, 11). It has been predicted by Jones et al. (2) on the basis of a chemical speciation model that meso-DMSA will also mobilize endogenous zinc to about the same extent as EDTA. Graziano et al. (91,however, have observed that the use of meso-DMSA did not have a marked effect on the urinary excretion of zinc; in contrast, a comparable dose of NazCaEDTA caused a thousandfold increase in zinc excretion. DMSA exists in two diastereoisomeric forms, meso and rueemo. Despite the increasing use of meso-DMSA in the US.a s a n antidote for lead poisoning, the disadvantage of using meso-DMSA for the treatment of heavy metal poisoning in humans is that meso-DMSA and its metal chelates are slightly soluble in water (12). On the other hand, the rac-2,3-dimercaptosuccinicacid (ruc-DMSA)l is very soluble in water and in strongly acidic solutions. This striking difference in the solubility of the two diastereoisomers makes the ruc-DMSA a potentially more effective antidote for the treatment of heavy metal poisoning. Egorova et al. (13) and Okonishnikova (14, 15) reported in their toxicological studies with the isotopes HgZo3,Cd115, Co60, and Z d 5 that, with identical doses of the ruc- and meso-DMSA, the ruc-DMSA invariably increases the elimination of mercury and cadmium in comparison with the meso-DMSA; zinc excretion, however, was also found to have been increased. The comparatively higher toxicity of ruc-DMSA [LDso = 10.84 mmolkg (5)],than that of meso-DMSA, could be attributed to the high extent of excretion of endogenous zinc by rac-DMSA. It is important to understand the reasons for these dramatic differences in the depletion of endogenous zinc when different chelating agents or the

Q893-228x/94l27Q7-Q77Q$Q4.5QlQ0 1994 American Chemical Society

Zinc-DMSA

Complexes

stereoisomers of the same chelating agent are administered in the treatment of heavy metal poisoning. A knowledge of the chemistry of zinc complex formation with the diastereoisomers of DMSA can help in selecting the conditions under which these chelating ligands are employed for the treatment of heavy metal poisoning. We report in this paper the results of comparative experiments that we have performed with meso- and racDMSA. We have determined the pKls of meso-DMSA and of rac-DMSA, we have determined the conformations of meso-DMSA in aqueous solution a t W e r e n t pH values; we have proposed that a series of complexes are formed in aqueous solutions containing Zn2+and meso- or racDMSA, we have calculated the distribution of these zinc complexes in solution as a function of pH, and we have determined their formation constants. On the basis of the distribution of these zinc complexes formed with meso- and rac-DMSA, we have rationalized why racDMSA complexes and excretes endogenous zinc to a greater extent than its meso isomer.

Experimental Section Materials. Deuterium oxide was purchased from Aldrich Chemical Co. (Milwaukee, WI);tert-butyl alcohol ("BA)' was purchased from EM Science (Gibbstown, NJ) and distilled before use; rac-DMSA was synthesized in our laboratory as described previously (16); meso-2,3-dimercaptosuccinicacid was a gift from Johnson & Johnson Baby Products Co. (Skillman, NJ); all other inorganic compounds used were purchased from Mallinckrodt (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. Potentiometry. The apparatus used for potentiometric measurement of hydrogen ion concentration has been described in our previous work (16). Standard NaOH or KOH solutions were prepared from 50% (w/w) NaOH solution or 45% (w/w) KOH solution and CO2-free double-deionized H2O. The prepared base solutions were kept under a nitrogen atmosphere to prevent the absorption of C02. The carbonate content, determined as described by M d l l e t al. (171,was found to be less than 1.5% in the NaOH solution and 0.7% in the KOH solution; the molarities of NaOH and KOH solutions were determined with potassium acid phthalate with phenolphthalein a s indicator. The concentration of a stock HN08 solution, prepared from concentrated HNo3 and CO2-free double-deionized H20, was determined with the standard NaOH or KOH. The concentration of a stock Zn2+solution, prepared from Zn(NO3)2.xH2O and COz-free double-deionized H20, was determined by EDTA titration. The DMSA stock solutions were prepared by placing a desired amount of the compound, 2 equiv of KOH, if meso-DMSA stuck solution was being prepared, and CO2-free double-deionizedH2O in a 50 mL volumetric flask. The concentrations of the DMSA solutions were calculated from the initial weight of the ligands and then confirmed by NaOH or KOH titration with methyl red as indicator. The 1.00M NaN03 stock solution was prepared by dissolving 42.50 g of N a 0 3 in 500 mL of CO2-free double-deionized HzO. Potentiometric measurements of hydrogen ion concentration were performed in a 150 mL glass-jacketed vessel provided with a magnetic stirrer and a tightly fitting rubber stopper. The latter was equipped with inlet and outlet tubes for nitrogen, a buret for delivery of NaOH or KOH solution, a glass electrode, a salt bridge, made of 2% agarose gel containing saturated KCl, which was connected to a saturated calomel reference electrode, and a thermometer. During the titration the tip of the buret was positioned right below the surface of the solution; replacement of the NaOH or KOH in the tip of the buret with the titration solution was not observed, because most of the equilibria were established within a minute. For those titration points for which longer times were needed to stabilize the pH

Chem. Res. Toxicol., Vol. 7, No. 6, 1994 771 readings, the buret was raised above the surface of the solution during the waiting period. The temperature in the titration vessel was controlled by circulation of thermostated water through the jacket, by using a VWR 1160 refrigerated circulator. To minimize the effect of nitrogen flow on the evaporation of the titration solution, the nitrogen was saturated with water vapor by passing the gas through a ca. 0.1M N a 0 3 solution. A Beckman @-72 pH meter, equipped with an electrode pair, was f i s t calibrated with NIST buffers (pH = 4 and 7). Titration of a 9.448 x lom3M HNo3 solution in the presence of 0.10 M N a 0 3 or KN03 gave an experimental conversion curve of -log [H+l vs measured pH. The validity of the conversion curve was checked by the inherent relationships in the Gran plots (18) obtained in both acidic and basic regions. These include the equality of the absolute values of the x-intercepts, which represent the amount of strong acid present in the initial solution titrated, and the equality of the slopes, which represent the molarity of strong base added, of the Gran plots obtained from both acidic and basic regions. The validity of the conversion curve was also checked by a comparison of pKw values determined using the conversion curve with the pKw values calculated using the activity coefficients of hydrogen ion and hydroxide ion, which were determined from the extended form of the Debye-Huckel equation. The calculated pKw value is 13.80 (20),and the experimentally determined pKw value is

13.80f 0.01. The titration solutions were prepared according to one of the following two procedures: (a) A calculated volume of Con-free deionized water was placed in a titration vessel into which a desired volume of the stock solution of DMSA was transferred, followed by addition of a desired volume of standard solution of HN03 (if necessary) and a desired volume of zinc nitrate stock solution, when zinc complexes were titrated, and finally a desired amount Of KN03was weighed and placed in the titration vessel to maintain the ionic strength equal to 0.10 & 0.01 throughout the titration. (b) A constant volume of the stock solution of DMSA and 5 mL of 1 M N a 0 3 solution were transferred into a 50 mL volumetric flask, a desired volume of standard solution of HNo3 (if necessary) was added, followed by addition of a desired volume of zinc nitrate stock solution, when zinc complexes were titrated, and then the solution was diluted to the mark; 40 mL of this solution was transferred to the titration vessel for potentiometric titration. By changing the volume of zinc nitrate stock solution added, a series of titration curves a t M e r e n t ligand:Zn2+stoichiometric ratios were obtained. The potentiometric titrations were performed a t 25.0 "C. The experiments with pure ligand were repeated several times to estimate the uncertainties in the pK. values.

N M R Titration of meso-DMSAwith Sodium Deuteroxide (NaOD). A stuck solution of NaOD in deuterium oxide was prepared by dissolving NaOH in DzO, evaporating all solvent under vacuum, and redissolving it in DzO. This procedure was repeated to ensure that all the hydrogen in NaOH was substituted by deuterium. The concentration of the NaOD stock solution was determined by titration with the standard HNo3 solution. Proton NMR spectra were obtained with a Bruker AM-250 spectrometer using deuterium in the solvent as a lock material. All experiments were carried out a t ambient temperature (ca. 22 "C). The proton spectra were obtained in 5-mm NMR tubes, the observation frequency was 250.134 MHz, and the digital resolution was 0.365 Hdpt. The chemical shifts in aqueous solution were measured with respect to TBA but reported with (DSS).' respect to sodium 2,2-dimethyl-2-silapentane-5-sulfonate The advantage of using TBA is that its signal is independent of pH and ionic strength (20).Under our experimental conditions, 0.4-1% by volume of TBA was sufficient to yield a good reference signal a t 1.01 ppm downfield from DSS, which differed from the reported value (21)by 0.22ppm upfield.

772 Chem. Res. Toxicol., Vol. 7, No. 6, 1994

Fang and Fernando

l2 1 10 -

8-

7

z W 0

3

I

I

I

I

I

i

2 3 4 5 equivalents of OH-added Figure 1. Potentiometric titration curves of 1.295 x M meso-DMSAonly (0);and of 1.206x M meso-DMSAin the presence of Zn2+at 6.026 x M (2:l)(B), 1.205 x M M (1:2)(A),and 6.024 x (1:l)(x), 2.410 x M (15)(v) at T = 25.0 "C and ionic strength QL) = 0.10.The solid lines are the titration curves simulated by PKAS for meso-DMSA, and by BEST for meso-DMSA in the presence of zinc ion. 0

6-

1

The solution for proton NMR was prepared by weighing 14.6 mg of previously vaccum-dried meso-DMSA in a 5-mm NMR tube followed by the addition of 10 p L of 10% (v/v)TBA in DzO and 290 p L of DzO. A suspension was formed in the tube, but it was completely dissolved after 2 equiv of NaOD were added. A series of proton NMR spectra of the resulting NaZDz(mesoDMSA) were obtained after the successive addition of 0.2equiv of NaOD. The pD of the solution, after each addition of 0.2 equiv of NaOD, was determined separately by carrying out a parallel experiment in a small test tube. Instead of recording proton NMR spectra, the pD of each solution was measured using a glass electrode calibrated with two standard buffers (pH = 4 and 7); the pH meter readings were not converted to -log [D+].

Results Protonation Constants of meso- and ruc-DMSA. The potentiometric titration points (filled-in circles) and the simulated titration curve (solid line), calculated with the aid of the program "PKAS" (17)for meso-DMSA, are shown in Figure 1. The four pKa values of the mesoDMSA, calculated by using the program PKAS, are 2.45 f 0.01, 3.44f 0.01, 9.65f 0.01, and 11.89 f 0.01. The potentiometric titration points (filled-in circles) and the simulated titration curve (solid line), calculated with the aid of the program PKAS for rac-DMSA, are shown in Figure 2. The four pKa values of the rac-DMSA, calculated by using the program P U S , are 2.36 f 0.01, 3.87 f 0.04,9.42f 0.02,and 12.51 f 0.03, respectively. These pKa values of rac-DMSA are slightly different from the values reported previously (16)because corrections for the effect of carbonate contamination in the strong base solution were not made in the earlier report. Stability Constants of Zinc Chelates of meso- and ruc-DMSA. The potentiometric titration points for mesoand rac-DMSA in the presence of Zn2+at various DMSA

0

1

2 3 4 equivalents of OH-added

5

Figure 2. Potentiometric titration curves of 1.412 x 10-3 M ruc-DMSA only (0);and of 1.232 x 10-3 M ruc-DMSA in the presence of Zn2+at 6.024 x lo-* M (2:l)(m), 1.205 x 10-3 M (1:l)( x ) , 2.410x M (1:2)(A),and 6.024 x M (15)(VI; and of 1.201 x M ruc-DMSA in the presence of 4.016 x lo-* M Zn2+(1:3)(0with + inside);and of 9.745 x M PUCDMSA in the presence of 3.258 x M Zn2+(1:3)(0with x inside) at T = 25.0 "C and ionic strength QL) = 0.10.The solid lines are the titration curves simulated by PKAS for ruc-DMSA, and by BEST for ruc-DMSA in the presence of zinc ion.

Zn2+ratios are shown in Figures 1 and 2, respectively. The complexation of zinc with meso-DMSA has been studied by Jones et al. (2) and Harris et al. (22). It was concluded by both groups, after the analysis of potentiometric titration curves by nonlinear least-squares methods and as a consequence of the crossover and nonsuperimposability of the formation function curves, that a variety of polynuclear complexes were formed in aqueous solution, although there were differences in the speciation models proposed by the two groups. The complexation of zinc with rac-DMSA has not been investigated quantitatively. In the systems containing the completely deprotonated DMSA ligand, L4-, and Zn2+ ion, the equilibria governing complex formation are described by the following equation, on the assumption that no protonated metal complexes are formed:

pM2+ + qL4- =+ (MpLq)2p-49

(1)

where p and q are integers and each complex has a n overall stability constant ppP,defined by:

The formation function, ii, which can be used to determine the average stoichiometric composition of a metalligand complex, is defined by the ratio of the total number of moles of complexed ligand to the total number of moles of metal species present in the solution. By assuming that no polynuclear complexes were formed in the system (Le., p = l),the formation function, ii, can be simplified by the equation:

Chem. Res. Toxicol., Vol. 7, No.6,1994 773

Zinc-DMSA Complexes 2*5]

(3)

fi=

:, I

CM I

where CMrepresents the analytical concentration of the metal ion, Zn2+. The value of the formation function, ii, a s well as the free ligand concentration, [L4-l, a t any point on the titration curve can be calculated from known or experimentally measurable quantities by using the following equations:

a *

I

* ’ I

’ *

a

I



0.5

01 4

where C w represents the analytical concentration of the ligand, and K I ,Kz, K3,and K4 are the first, second, third, and fourth protonation constants, respectively, of DMSA. All concentrations in the above equations, in square brackets, are expressed in moles per liter. I t was postulated at first that ZnL2- and ZnLf were the chelated species formed in the system. In order to prove that these zinc complexes were formed with ruc-DMSA, titrations of ruc-DMSA in the presence of Zn2+and ligand in a ratio of 3:l were carried out and the titration curves were analyzed by means of eqs 4-6. The formation curves were obtained by plotting the formation function, ii, versus p[L4-], the values of negative log of the free ligand concentration. There are several important features in the formation curves, in Figure 3: First, the inflection in the curve occurred a t an ii value of 0.57 rather than 1.0 as would be the case if two complexes, ZnL2- and ZnL$-, were formed, with well-separated formation constants. The inflection a t ii < 1 led us to the conclusion that a protonated form of ZnL2- was formed. Second, the inflections at ii ca. 1.85 were observed as the values of -log [L4-l decreased, in both formation function curves with a decrease in ii value for the curve obtained with the higher initial concentration (filled-in circles in Figure 3) and a n increase in ii value for the curve with the lower initial concentration (filledin squares in Figure 3). This observation led us to propose that protonated ZnLZ6- species were formed in the system. The “abnormal” decrease in ii values with higher initial concentration (filled-in circles in Figure 3) a t low values of -log [L4-l, corresponding to pH values between 8.7 and 10.6, probably implies that the species Z n L P undergoes a n internal configuration exchange equilibrium. In the case of the low initial concentration (filled-in squares in Figure 3), there are less uncoordinated thiolate groups available in ZnLZ6- for protonation to form a significant amount of HMLz5-. Uncoordinated carboxylate groups may exist, however, but no protonation would occur between pH 8.7 and 10.6. In the case of high initial concentration (filled-in circles in Figure 31, the internal configuration equilibrium shifts toward the side which produces ZnLZ6- with more uncoordinated thiolate groups which can be protonated between pH 8.7

I

I

I

I

8

12

16

20

-log[ L4-1 Figure 3. Formation curves of ruc-DMSA-Zn2+ (3:l):( 0 )[L4-] = 1.201 x M and [Z2+] = 4.016 x M; (B) [L4-] = 9.745 x M and [Z2+]= 3.258 x M.

and 10.6 to form a significant amount of HMLz5-. This can result in an “abnormal” decrease in apparent ii values, defined by eq 3 a t low values of -log [L4-l along the formation curve. The net effect of the change in protonation sites in ZnLf- on the fi value is a decrease in the ii value caused by the invalidity of eq 3, since all protonated metal complexes were not taken into account in the numerator of eq 3. The shape of the formation curve is dependent on the initial concentration, and the curves are obviously not superimposable; this suggests that polynuclear complexes are also formed in the zincruc-DMSA system. Since protonated as well as polynuclear complexes are formed in the zinc-meso-DMSA and zinc-ruc-DMSA systems, eqs 3-5 are no longer valid, and no simple procedures exist for proving the presence of such complexes and for determining the formation constants of these complexes. The computer program “BEST” was therefore used (17, 23), which employs a n iterative process of nonlinear fitting, which calculates the equilibrium pH value a t each point in the potentiometric titration and refines the equilibrium constants in the selected model until the weighted discrepancies between the experimental and the calculated pH values are minimized (17). The titration curves for both meso- and ruc-DMSA with a 1igand:zincratio of 2:l and 1:l were analyzed using the BEST program because these ratios closely approximated the true stoichiometry of the zinc complexes formed in the system, based on the analysis of the formation function curves. The titration curves for the zinc-mesoDMSA system were analyzed by the BEST program; the complex species formed in the L:Zn system are Zn&OH5-, ZnzLz4-, ZnzLzH3-, and ZnLH-; and those in the 2L:Zn system are Zn2Lz4-, ZnzLzH3-, ZnLZ6-, and ZnLzH5-. The formation constants and the protonation constants of various zinc complexes with meso-DMSA are listed in Tables 1and 2, and the titration curves simulated by the BEST program for meso-DMSAzinc ratios of 2:l and 1 : l are shown in Figure 1(solid lines). The titration curves for rue-DMSA were analyzed by the BEST program, and

Fang and Fernando

774 Chem. Res. Toxicol., Vol. 7, No. 6, 1994 Table 1. Formation Constants for the Systems: H+-Zn2+-meso-DMSAat p = 0.10 and T = 25 "C"

3.80 -

species ratio of L:Zn

P

9

r

log BDR?

2:lC

1 1 2 2

2 2 2 2 1 2 2 2

0 1 0 1 1 0 1 -1

19.74 30.56 33.77 39.83 20.30 33.56 40.12 22.72

1:ld

1 2 2 2

3.60 n

ECI

5

c v1

I

n

L4- represents the completely deprotonated meso-DMSA. PpQr = [Znp4H,-(~-2p-r)y[Zn~+~[L4-]Q[H+~. a(pH)fit = 0.016 (17). a(pH)fit= 0.022 (17).

Table 2. Protonation Constants of the Zinc Complexes of meso-DMSAat p = 0.10 and T = 25 O c a log K protonation

+ +

-

ZnzLz4- H+ ZnzLzH3ZnLZ6- H+ * ZnLzH6-

6.06b 10.82b

-

I

Q

~~~

3.40

6.56c

.Y 3.20 -

E c) a u

3.00 -

2.80

a 0

2

4

8 1 0 1 2 1 4

6

L represents the completely deprotonated meso-DMSA. Values derived from the BpQrvalues obtained from the 2:lligand:zinc system. Value derived from the BpQrvalues obtained from the 1:l 1igand:zinc system.

Figure 4. Chemical shift of the methine protons of meso-DMSA

Table 3. Formation Constants for the Systems: H+-Zn2+-ruc-DMSAat fi = 0.10 and T = 25 "C"

Proton NMR Titration of the Disodium Salt of meso-DMSA The use of proton magnetic resonance

0

ratio of L:Zn

P

2:l'

1 1 1 2 2

1:ld

1 2 2 2 2

a

species 4 2 2 2 2 2 2 2 2 2 2

r

log Bp,rb

0 1 -1 1 2

25.28 32.87 14.68 42.26 46.52 43.13 34.79 42.26 46.22 24.55

2 0 1 2 -1

L4- represents the completely deprotonated rac-DMSA.

Bpqr

o(pH)fit = 0.0076 (17).

Table 4. Protonation Constants of the Zinc Complexes of ruc-DMSA at u = 0.10 and T = 25.0 "C" loe K

+ + +

-

7.47c 4.26b 7.59b 10.26d

(W) and ruc-DMSA ( 0 )versus pD.

spectroscopy in the study of conformations of rac-DMSA in aqueous solution has been described previously by Fang et al. (26). Similar to the proton NMR spectrum of rac-DMSA, the two methine proton resonances in mesoDMSA were not resolved in the proton NMR spectrum, under our experimental conditions, and were recorded as a singlet. Unlike rac-DMSA, addition of two deuterium ions to the fully deprotonated meso-DMSA does not result in a significant change in the chemical shift of the methine protons in the molecule (Figure 4). The chemical shift of the methine protons in the meso-DMSA is about 3.07 ppm in the pD range from 4.2 to 12.

= [Znp~Hr-(~-2P-r)y[Znz+~[L4-]~[H+]'. u(pH)fit = 0.0116 (17).

ZnzLz4- H' ZnzLzH3ZnzLzH3- + H' ZnzLzHZ2ZnLz6- H+ =-- ZnLzH5ZnLzH5- H+ ZnLzHz4-

PD

3.96c

L represents the completely deprotonated rac-DMSA. Values derived from the BpQr values obtained from the 2:l 1igand:zinc system. Value derived from the BPQrvalues obtained from the 1:l 1igand:zinc system. Values derived by combining the Bpqr values from the 2:l and 1:l 1igand:zinc systems. a

of the many models that were evaluated, the data for the statistically favorable models are summarized in Table 3; the protonation constants of various zinc complexes with rac-DMSA are listed in Table 4; and the titration curves simulated by the BEST program for rac-DMSA. zinc ratios of 2:l and 1:l are shown in Figure 2 (solid lines). In the potentiometric titration of both L:M and 2L:M systems, zinc interacted with rac-DMSA to form dimeric complexes a t different stages of protonation: ZnzLzOH5-,ZnzLz4-, Zn2L2H3-, and ZnzLzH22- in the 1:l system and Zn2LzH3- and ZnzLzH22- in the 2:l system.

Discussion Conformations of meso-DMSA in Aqueous Solution. The change in chemical shift of the methine proton in DMSA in the proton NMR spectrum, as a function of pD, has been used to probe the conformations of racDMSA in aqueous solution by Fang et al. (26). The downfield chemical shift due to the protonation of one thiol group in the ruc-DMSA molecule is ca. 0.139 ppm (26). This was not observed for the meso-DMSA, in which the chemical shift of methine protons remained almost constant (Figure 4) in the course of protonation of the two thiolate groups, when the pD was decreased from 12 to 4.2. meso-DMSA and rac-DMSA undergo similar conformational changes when the pD of the solution decreases. The proposed conformations of meso-DMSA, L4-, DL3-, and DzL2-, are shown in Figure 5. L4- is linear; however, upon the addition of one deuterium ion to one of the thiolate groups in the molecule, the mesoDMSA forms an internal six-membered ring structure via deuterium bonding between the sulfur of the deuterated thiol group and the oxygen of the carboxylic group which is a t a ,&position relative to the deuterated thiol group. The most energetically favorable conformation of this six-membered ring structure is that in which the two bulky carboxylic groups are in a staggered anti position, to minimize electrostatic repulsion and steric effects.

Chem. Res. Toxicol., Vol. 7,No.6,1994 775

Zinc-DMSA Complexes

0'

Figure 5. Conformationsof meso-DMSA in aqueous solution. Therefore, the methine proton, Ha in Figure 5, adjacent to the carboxylic group, which participates in the sixmembered ring formation, is always located above the carbonyl double bond plane of that carboxylic group, where it experiences a shielding effect (24,25). The deshielding effect caused by the deuteration of one thiol group is canceled out by the formation of one sixmembered ring resulting from the deuteration. The extent of the shielding effect caused by the long range interaction with an adjacent carbonyl double bond is close to the extent of the deshielding effect caused by the deuteration of an adjacent thiol group. Upon the addition of the second deuterium ion to DL3-, another internal six-membered ring is formed via the deuterium bond, which cancels out the deshielding effect of the deuteration of the second thiol group. This bis-six-membered ring conformation of D2L2- in Figure 5 is centrosymmetric. The meso-DMSA favors the formation of a n internal sixmembered ring when it is only monodeuterated, whereas the rac-DMSA forms a n internal six-membered ring structure after both thiol groups are deuterated. Speciation Models for Zinc-meso-DMSA and Zinc-rac-DMSk Ligands containing mercapto groups, such as meso-DMSA (221,the dimethyl ester of mesoDMSA (26),unithiol (21,and mercaptoacetic acid (271, have been reported to form dimeric complexes with zinc. The speciation model proposed for the zinc-meso-DMSA system by Harris e t al. (22)was used to refine the potentiometric titration data obtained for the zinc-mesoDMSA system, in which L:Zn = 1:1, and it was found to be the statistically favorable model for the system. The formation constants (Table 1)of the dimeric complexes Zn2L2*- and Zn2L2H3-, which are determined from both L:M and 2LM systems, agree very well with each other, within the margin of error of the computations with the BEST program (17,23). This strongly supports the models that define the zinc complex species that are present in solution in the course of the potentiometric titrations of the two systems in which the meso-DMSA zinc ratio is 1 : l and 2:l. The average values of the log formation constants determined by us for the Zn2h4-and Zn2L2H3- are 33.67 f 0.15 and 39.98 f 0.21, respectively, which are in good agreement with the values reported by Harris et al. (221,33.6 f 0.5 and 39.6 f 0.3, respectively; and by Jones et al. (21,34.08 f 0.03 and 40.07 f 0.02, respectively, despite the differences in ionic strength and temperature used by Jones et al. The log formation constant of Zn2L20H5- determined by us, 22.72, is also in good agreement with the value reported by Harris et al. (221,23.6 f 0.8;Jones e t al. (2)have postulated the formation of ZnLzs- on the basis of a statistically most favorable model obtained from MINI-

QUAD (28);however, Harris et al. (22)did not include this complex in their statistically most favorable model. Recently, the formation of ZnL$- has been clearly shown with proton NMR spectroscopy and infrared spectroscopy by Fang et a1.;2 therefore, the complexes of ZnL& and ZnL2HS- have been included in the 2L:M system. The log formation constant of ZnLz6- determined by us is 19.74, which is in good agreement with value reported by Jones et al., 19.46 f 0.06. Jones et al. also postulated the formation of protonated forms of ZnL31°-; however, we were unable to confirm the formation of ZnL310- by proton NMR spectroscopy and infrared spectroscopy.2 Comparison of results shown in Tables 1 and 3 shows that the complexation of zinc with ruc-DMSA is much more complicated than complexation with its meso isomer. Egorova et al. (13)suggested that the ruc-DMSA had a higher formation constant than meso-DMSA in the formation of ZnL26-. This qualitative observation was made on the basis of the suppression of the buffer region, which was observed after 3 equiv of base were added, in the potentiometric titration of a ruc-DMSA solution containing a 2:l 1igand:zinc ratio. No quantitative calculations were performed for zinc-ruc-DMSA systems, and no dimeric complexes were included in their interpretation of the potentiometric titrations of the zincmeso-DMSA systems. The shapes of the potentiometric titration curves obtained with a 2:l ratio of ligand:zinc, for zinc-meso-DMSA and zinc-ruc-DMSA systems, in our laboratory are exactly the same as those obtained by Egorova et al. (13).The speciation models for zincruc-DMSA proposed by us (Table 3) have been calculated with the aid of the computer program BEST (17,23) and shown to be the most favorable on a statistical basis. The log formation constants (Table 3) of the dimeric complexes Zn2LzH3- and Zn2L~H22-, which are determined from both L:M and 2L:M systems, agree very well with each other, within the margin of error of the computations with the BEST program. This strongly supports the models that define the zinc complex species that are present in solution in the course of the potentiometric titrations of the two systems in which the ruc-DMSA zinc ratios are 1 : l and 2:l. The ZnLz6- species of rucDMSA is extraordinarily stable with a log formation constant of 25.28, in comparison with the ZnL$- of mesoDMSA, which has a log formation constant of 19.74. The complexation of zinc with ruc-DMSA is more complicated than its complexation with the meso isomer because the solution of ruc-DMSA consists of two unresolved enantiomers, (R,R)-2,3-dimercaptosuccinicacid [(R,R)-DMSAll and (S,S)-2,3-dimercaptosuccinic acid [(S,S)-DMSA].' There are three types of 2:l ligand to metal complexes that may be formed in solution, i.e., Zn((R,R)-DMSA)((S,S)-DMSA),Zn((R,R)-DMSA)2,and Zn((S,S)-DMSA)2. The formation constants obtained potentiometrically for the zinc-ruc-DMSA may reflect the averaged behavior of all three types of complexes. The formation of various types of zinc complexes in the zincruc-DMSA system has been confirmed by proton NMR spectroscopy and infrared spectroscopy by Fang et a1.;2 however, the enantiomeric nature of the DMSA in the 2:l ligand to zinc complexes cannot be deduced unequivocally from the results of the NMR experiments. 2 Fang, X.,and Fernando, Q.(1994)Conformations of zinc chelates of meso- and rac-2,3-dimercaptosuccinicacid in aqueous solution (submitted for publication).

776 Chem. Res. Toxicol., Vol. 7, No. 6, 1994

Fang and Fernando I

100 1

-

80

80

I2n.LH.I

3

4

5

8

6 7 -log[H+J

9

10

Figure 6. Distribution diagram of zinc species for the H+M Zn2+-ruc-DMSA-meso-DMSA system containing 2.5 x Zn2+,5.0 x M ruc-DMSA, and 5.0 x M meso-DMSA (Le., 4:l 1igand:zinc system). The zinc species plotted are only those which contribute t o the total amount of zinc by more than 0.2% in the vicinity of pH 7.4.

Depletion of Zinc by meso- and rac-DMSA at Physiological pH. As indicated by the potentiometric titration curves and confirmed by the speciation models for meso- and rac-DMSA, the complex formation behavior of ruc-DMSA is dramatically different from that of mesoDMSA. Therefore, it is not valid to predict the extent of zinc complexation with the two isomers of DMSA on the basis of a naive comparison of formation constants. The distribution of zinc complexes in vivo varies as a result of the large pH variation in the various in vivo compartments. It is reasonable, however, to select a pH of 7.4 as a general physiological pH value. In order to compare the extent of formation of zinc complexes with meso- and rac-DMSA a t physiological pH, the program "SPE" (17) was used to calculate the distribution of various zinc complexes formed in the presence of both meso- and racDMSA under various conditions, as a function of pH by using the formation constants listed in Tables 1 and 3. It was assumed, in the calculation, that all the zinc complexes of meso- and ruc-DMSA, shown in Tables 1 and 3, were present. The concentrations of the ligands used in the calculation ranged from 0.1 to 1.0 mM, which covers the typical concentration range of a chelating agent used clinically. The concentration of zinc used in the calculation ranged from 0.001 to 1.0 mM in order to cover the biological levels of zinc. The concentrations of meso- and rac-DMSA were varied simultaneously and always kept the same in our computer calculations. It is shown in Figure 6 that in a solution containing (2mesoDMSA 2rac-DMSA):Zn, only two types of zinc complexes are present in the vicinity of pH 7.4, Zn(racDMSA)zHz4- and Znz(meso-DMSA)z4-,with about 88%of zinc bound in ruc-DMSA and about 12% of zinc bound in meso-DMSA. When the zinc concentration was increased by a factor of 2 while the DMSA concentrations were kept constant, a protonated zinc complex of meso-DMSA,Znz(meso-DMSA)zH3-,was formed in the vicinity of pH 7.4 along with an increase in the concentration of Znz(mesoDMSA)z4-and a decrease in the concentration of Zn(racDMSA)zHz4- (Figure 7). As a consequence, only about 47%of the zinc is bound in ruc-DMSA and about 53% of zinc in meso-DMSA. A further increase in the concentration of zinc to a ratio of (meso-DMSA rac-DMSA):2Zn did not result in any change in the concentrations of the individual zinc complexes with meso-DMSA, however, the concentration of the zinc complexes formed with racDMSA were drastically changed as a result of the

+

+

60

Figure 7. Distribution diagram of zinc species for the H+M Zn2+-ruc-DMSA-meso-DMSA system containing 5.0 x Zn2+,5.0 x M ruc-DMSA, and 5.0 x low4M meso-DMSA (i.e., 2:l 1igand:zinc system). "he zinc species plotted are only those which contribute to the total amount of zinc by more than 0.2% in the vicinity of pH 7.4.

-logJH+J

Figure 8. Distribution diagram of zinc species for the H+Zn2+-ruc-DMSA-meso-DMSA system containing 1.0 x M Zn2+,5.0 x M ruc-DMSA, and 5.0 x M meso-DMSA (Le., 1:l 1igand:zinc system). "he zinc species plotted are only those which contribute to the total amount of zinc by more than 0.2% in the vicinity of pH 7.4.

formation of two dimeric complexes, Znz(rac-DMSA)z4and Znz(meso-DMSA)zH3-,and the disappearance of Zn(rac-DMSA)zHz4-(Figure 8). The amount of zinc bound in ruc-DMSA remained unchanged a t about 47%. It is evident from Figures 6-8 that the ratio of zinc complexed by the two DMSA isomers depends on the concentration of zinc present in the solution and also depends on the concentrations of the DMSA isomers. The term plasma mobilizing index (PMI)' (2,291 has been used to assess the efficacy of a chelating agent in mobilizing a metal ion from a labile metal-protein complex in blood plasma. A similar term, the relative plasma mobilizing index (RPMI),' which is defined as: RPMI = the total concentration of zinc bound in ruc-DMSA the total concentration of zinc bound in meso-DMSA has been used to compare the relative extent of the depletion of zinc by rac-DMSA over meso-DMSA. RPMI values at pH 7.4 were calculated by using the program SPE over a wide range of zinc concentration and total DMSA concentration, but the individual concentrations of rac- and meso-DMSA were always kept the same. The results of the calculations were plotted in Figure 9 as log(RPM1)versus the log of the molar ratio of rac-DMSA to zinc. It is evident from Figure 9 that the RPMI value

Zinc-DMSA Complexes

Chem. Res. Toxicol., Vol. 7, No. 6, 1994 777 bound to ruc-DMSA than to meso-DMSA a t pH 7.4 while a low level of free zinc ion, about 10 pg/dL, is maintained in the solution. This ensures that there is practically no depletion of endogenous zinc by ruc-DMSA. It may be concluded, therefore, that Zn(ruc-DMSA)$+ is more efficacious and much less toxic than either meso-DMSA or ruc-DMSA in the treatment of acute lead poisoning. The ruc-DMSA is superior to meso-DMSA if the drug is administered to children and pregnant women as a prophylactic for the prevention of lead poisoning. In this case the required dosage of the drug is low. Lead blood levels as low as 10-15 ,&dL (30))i.e., 0.48-0.72 pM, and possibly lower, are linked with impaired neurobehavioral development in human fetuses and children. In order to maintain the lead blood level under 0.48 pM, a DMSA concentration of a few micromolar is required in the blood, which brings the DMSAzinc ratio to the vicinity of 1.0 or even lower, where the extent of zinc depletion by the two DMSA isomers is approximately equal. The administration of ruc-DMSA, therefore, is advantageous because of the higher lead mobilizing capability of rucDMSA over meso-DMSA.

4.5

3.5 c;'

E

2.5

&

u, 0

. I

1.5

0.5

-0.5 0.0

0.4

0.8

1.2

1.6

2.0

log(mmo1 rac-DMSNmmol Zn(I1)) Figure 9. Curves for log of zinc(I1) relative plasma mobilizing index (RPMI) of ruc-DMSA over meso-DMSA versus the molar ratio of ruc-DMSA to zinc(I1) at concentrations of rac-DMSA of 0.1 mM (01, 0.5 mM (H), and 1.0 mM (A).Note: The initial concentration of meso-DMSA is always equal to the initial concentration of rac-DMSA.

is affected by the DMSA to zinc ratio and the absolute concentration of the DMSA. The RPMI value is independent of the absolute concentration of DMSA (Figure 91, before the ratio of ruc-DMSAzinc approaches 1:l which corresponds to a total DMSAzinc ratio of 2. There is no significant difference in the depletion of zinc by the two isomers at a log(RPM1)value of around 0 despite the difference in the chemical nature of zinc-ruc-DMSA complexes (Figure 8). A significant increase by 3-4 powers of ten in the depletion of zinc by ruc-DMSA over meso-DMSA occurs when log(ruc-DMSAlzinc) increases from 0 to 1.0, which corresponds to a n increase in the ratio of total DMSA to zinc between 2 and 20. A further increase in the ratio of total DMSA to zinc from 20 to 200 does not cause a significant additional increase in the depletion of zinc by ruc-DMSA over meso-DMSA. A high dosage of DMSA is commonly used in the treatment of acute heavy metal poisoning. Hence, the ratio of DMSAzinc is between 20 and 200 or even higher, and the zinc depletion by ruc-DMSA is expected to be much greater than the zinc depletion by meso-DMSA. The RPMI value of ruc-DMSA over meso-DMSA for lead3 has been found to be about 10, and therefore, there is no significant advantage if the ligand ruc-DMSA, instead of meso-DMSA, is administered in the treatment of acute lead poisoning. The ruc-DMSA is superior to meso-DMSA in the treatment of acute heavy metal poisoning, e.g., lead poisoning, if the zinc complex, Zn(ruc-DMSA)26-, is administered instead of the pure ligand, ruc-DMSA. The lead3 is found to be completely bound to DMSA in the system initially containing 0.25 mM ruc-DMSA, 0.25 mM meso-DMSA, 0.25 mM zinc ion, and 0.25 mM lead ion, and about 10 times more lead is Fang, X., and Fernando, Q.(1994) A comparative study of lead complexes of meso-DMSA and rac-DMSA in aqueous solution (manuscript in preparation).

Acknowledgment. This work was supported in part by NIEHS Grant ES 03356. References (1) Lucis, 0. J., Lynk, M. E., and Lucis, R. (1969) Turnover of lo9Cd in rats. Arch. Environ. Health 18, 307-310. (2) Jones, D. C., Smith, G. L., May, P. M., and Williams, D. R. (1984) Assessment of pharmaceutical agents for removing cadmium from humans using chemical speciation models. Znorg. Chim. Acta 93, 93-100. (3) Cherian, M. G. (1980) Biliary excretion on cadmium in rat. 111. Effects of chelating agents and change in intracellular thiol content on biliary transport and tissue distribution of cadmium. J. Toxicol. Environ. Health 6, 379-391. (4) Cherian, M. G. (1980) Biliary excretion of cadmium in rat. IV. Mobilization of cadmium from metallothionein by 2,3-dimercaptosuccinic acid. J . Toxicol. Environ. Health 6, 393-401. (5) Aposhian, H. V., Hsu,C., and Hoover, T. D. (1983)DL- and mesodimercaptosuccinic acid: in vitro and in vivo studies with sodium arsenite. Toxicol. Appl. Pharmacol. 69,206-213. (6) Friedheim, E., Corvi, C., and Wakker, C. H. (1976) Mesodimercaptosuccinic acid a chelating agent for the treatment of mercury and lead poisoning. J.Pharm. Pharmacol. 28,711-712. (7) Magos, L. (1976) The effects of dimercaptosuccinic acid on the excretion and distribution of mercury in rats and mice treated with mercury chloride and methylmercury chloride. Br. J.Pharmacol. 56, 479-484. (8) Liebelt, E., Shannon, M., and Graef, J. W. (1994) Efficacy of oral meso-2,3-dimercaptosuccinic acid therapy for low-level childhood plumbism. J . Pediatr. 124 (2), 313. (9) Graziano, J. H., Siris, E. S., Lolacono, N. J., Silverberg, S. J., and Turgeon, L. (1985)2,3-Dimercaptosuccinicacid as an antidote for lead intoxication. Clin. Pharmacol. Ther. 37,431-438. (10)Cantilena, L. R., and Klaassen, C. D. (1981) Comparison of the effectiveness of several chelators after single administration on the toxicity, excretion, and distribution of cadmium. Toxicol. Appl. Pharmacol. 58, 452-460. (11) Mason, R. (1981) Effect of dimercaptosuccinate on the accumulation and distribution of cadmium in the liver and kidney of the rat. Biochem. Pharmacol. 30,2427-2433. (12) Rivera, M., Zheng, W., Aposhian, H. V., and Fernando, Q. (1989) Determination and metablism of dithiol-chelating agents: VIII. Metal complexes of meso-dimercaptosuccinic acid. Toxicol. Appl. Pharmacol. 100,96-106. (13) Egorova, L. G., Okonishnikova, I. E., Nirenburg, V. L., and Postovskii, I. Y. (1972) Complex formation of certain metals with dimercaptosuccinic acid stereoisomer. Khim. Farmat. Zhur. 6, 14-16. (14) Okonishnikova, I. E. (1970) The comparative antidotal potency of stereoisomeric dimercaptosuccinic acids in mercury poisoning. Gig. Tr. Prof. Zabol. 14,18-22. (15) Okonishnikova, I. E. (1971) The effect of steric isomers of dimercaptosuccinic acid on the elimination of some heavy metals from body. Gig. Tr. Prof. Zabol. 15, 50-52.

778 Chem. Res. Toxicol., Vol. 7, No. 6, 1994 (16) 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. Tozicol. 7, 148-156. (17) Martell, A. E., and Motekaitis, R. J. (1988) Determination and Use of Stability Constants, VCH Publishers, New York, NY. (18)Gran, G. (1952) Determination of the equivalence point in potentiometric titrations. Part 11. Analyst 77, 661-671. (19) Meites, L., Ed. (1963) Handbook ofAnalytical Chemistry, Tables 1-6 and 1-7, McGraw-Hill Book Co., New York, NY. (20) Sudmeier, J. L., and Reilley, C. N.(1964) Nuclear magnetic

resonance studies of protonation of polyamine aminocarboxylate compounds in aqueous solution. Anal. Chem. 36,1698-1706. (21) 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. (22) Harris, W. R., and Chen, Y. (1991) Stability constants for dimercaptosuccinic acid with bismuth(III), zinc(II), and lead(I1). J . Coord. Chem. 23,173-186. (23) Motekaitis, R. J., and Martell, A. E. (1982) BEST-A new program

Fang and Fernando for rigorous calculation of equilibrium parameters of complex multicomponent systems. Can. J. Chem. 60, 2403-2409. (24) Silverstein, R. M., Bassler, G. C . , and Morrill, T. C. (1991) Spectrometric Identification of Organic Compounds, 5th ed., p 175, John Wiley & Sons, New York, NY. (25) Jackman, L. M., and Sternhell, S. (1969)Applicationsof Nuclear Magnetic Resonance Spectroscopy in Organic Chemistry, 2nd ed., Pergamon Press, New York, NY. (26) 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 (11,142-147. (27) Perrin, D. D., and Sayce, I. G. (1967) Stability constants of Dolvnuclear mercaDtoacetate comDlexes of nickel and zinc. J . ‘Chin.SOC.A, 82-88. (28) Sabatini. A.. Vacca. A.. and Gans. P. (1974) MINIQUAD-A general computer pro&am for the’ computation of formation constants from potentiometric data. Talanta 21,53-77. (29) May, P. M., and Williams, D. R. (1977) Computer simulation of chelation therapy, plasma mobilizing index as a replacement for effective stability constant. FEBS Lett. 78, 134. (30) Davis, J. M., and Svensgaard, D. J. (1987) Low-level lead exposure and child development. Nature 329,297-300.