Biological Thiols and Carbon Disulfide: The Formation and Decay of Trithiocarbonates under Physiologically Relevant Conditions.

Carbon disulfide is an environmental toxin, but there are suggestions in the literature that it may also have regulatory and/or therapeutic roles in mammalian physiology. Thiols or thiolates would be likely biological targets for an electrophile, such as CS2, and in this context, the present study examines the dynamics of CS2 reactions with various thiols (RSH) in physiologically relevant near-neutral aqueous media to form the respective trithiocarbonate anions (TTC-, also known as "thioxanthate anions"). The rates of TTC- formation are markedly pH-dependent, indicating that the reactive form of RSH is the conjugate base RS-. The rates of the reverse reaction, that is, decay of TTC- anions to release CS2, is pH-independent, with rates roughly antiparallel to the basicities of the RS- conjugate base. These observations indicate that the rate-limiting step of decay is simple CS2 dissociation from RS-, and according to microscopic reversibility, the transition state of TTC- formation would be simple addition of the RS- nucleophile to the CS2 electrophile. At pH 7.4 and 37 °C, cysteine and glutathione react with CS2 at a similar rate but the trithiocarbonate product undergoes a slow cyclization to give 2-thiothiazolidine-4-carboxylic acid. The potential biological relevance of these observations is briefly discussed.

Introduction

This laboratory has recently compiled literature information on the known and suggested physiological properties of carbon disulfide (n class="Gene">CS2) and identified certain analogies to the small-molecule bioregulators (SMBs) nitric oxide, carbon monoxide, and hydrogen sulfide (sometimes called “gasotransmitters”).[1] The common properties of these SMBs include partial solubility in aqueous and lipid systems, the ability to diffuse readily in physiological structures, and known toxicity at higher concentrations.[2] Similarly, CS2 is a nonpolar, readily diffusible molecule considered to be an environmental toxin.[3] In addition, there are indications that CS2 is formed endogenously or in the associated gut microbiome of mammals.[4] Biological sulfhydryls (R–SH) would be likely targets, and the dynamics of the “on” and “off” reactions of this electrophile with such nucleophiles should be crucial to any physiological roles. Thus, the present study is focused on exploring the reactivity of CS2 with thiols, such as cysteine (CysSH) and glutathione (GSH), as well as with several model thiols to form trithiocarbonate anions (TTC–, also known as “thioxanthate anions”) in near-neutral aqueous media (eq ). From the biomedical perspective, the trithiocarbonate anion (PhCH2SCS2–) has been studied as an inhibitor to carbonic anhydrases and as a possible therapeutic in glaucoma treatment.[5] However, to our knowledge, the dynamics of the formation and decay of TTC salts under physiologically relevant conditions have not been previously reported.

Elucidating prospective biological roles of carbon disulfide will depend on having vehicles for controlled n class="Gene">CS2 release under experimental biological conditions. Certain TTCs are unstable toward the slow release of carbon disulfide,[6] and such reactivity may be relevant to the biological activity of CS2 as well as a desirable property for CS2 delivery. In this context, we describe the kinetics of CS2 dissociation from several prepared TTC– salts in aqueous solution. The latter studies complement earlier investigations of CS2 generation by photosensitized oxidation of 1,1-dithiooxalate[7] and by the thermal decay of dithiocarbamate anions.[8] The CS2 release rates from the latter precursors vary considerably, thereby providing a wide range of activities for physiological experiments. The TTC derivatives of CysSH and GSH also undergo a slow cyclization reaction to give 2-thiothiazolidine-4-carboxylic acid (TTCA), a product that, when found in the urine, is considered diagnostic of exposure to carbon disulfide.[9,10] Notably, this cyclization also releases an equivalent of hydrogen sulfide.

Results and Discussion

Trithiocarbonate Decay

As noted above, we have recently described the kinetics for a set of n class="Chemical">dithiocarbamate salts that decay by releasing CS2 with lifetimes ranging from seconds to days in near-neutral, aerobic aqueous media at 37 °C.[8] In this section, we describe analogous decays of several RSCS2– anions under similar conditions.

The TTC– n class="Chemical">salts used here were prepared by the reaction of the corresponding thiol precursor with CS2 in strongly alkaline solution. The electronic spectrum of each displays intense absorption bands at approximately 310 and 330–350 nm with extinction coefficient of ∼8 × 103 M–1 s–1 (e.g., Figure ) that we assign to π → π* transitions largely localized on the −SCS2– functional group. Time-dependent density functional theory calculations (Supporting Information (SI) Figure S1) support this assignment. Similar but somewhat higher-energy absorptions are seen in the spectra of analogous dithiocarbamate (R2NCS2–) and xanthate (ROCS2–) anions.[8,11]

Figure 1

Temporal solution spectra at 49 s intervals tracking the decay of Na2[PTTC] (initially about 0.11 mM) in pH 7.4, 37 °C aqueous solution (phosphate buffer at 100 mM). The increasing absorbance at 206 nm is consistent with the strong absorbance of CS2 at this wavelength.

Temporal solution spectra at 49 s intervals tracking the decay of Na2[n class="Chemical">PTTC] (initially about 0.11 mM) in pH 7.4, 37 °C aqueous solution (phosphate buffer at 100 mM). The increasing absorbance at 206 nm is consistent with the strong absorbance of CS2 at this wavelength.

Decays of the TTC anions in aqueous media are readily followed by the temporal den class="Chemical">creases of these two UV bands (Figure ). These spectral changes were accompanied by the release of at least 90% of the CS2 predicted, for example, by the stoichiometry of eq , using the CS2 analysis method described in detail elsewhere[7,8] and briefly in the Experimental Section.

As discussed below, the decays of these TTC– anions are reversible; therefore, it is necessary to take the n class="Disease">back-reaction into account as indicated in eq when evaluating the kinetics of the temporal spectral changes. These reactions were conducted in buffered solutions at specific pH values, and the effects of pH, buffer, and other factors are incorporated into the apparent rate constants koff and kon for the forward reaction and back-reaction. Although [CS2] is not constant during an individual experiment, the temporal spectral data can be fit numerically using a nonlinear least-squares regression program (see Experimental Section and SI Figure S2); however, the kon values so obtained are inherently less accurate than are the corresponding koff values.Figure illustrates temporal spectral changes observed when a sample of the sodium salt of 3-trithiocarbonatopropionate (Na2[PTTC]) undergoes decay in pH 7.4 aqueous solution at 37 °C (eq ). Table summarizes the rate constants koff and kon calculated as described. There was no significant effect of pH on koff over the range 6.5–7.8; thus, the decay of PTTC– is not catalyzed by acid, an observation that is in direct contrast to the decay under analogous conditions of the dithiocarbamate ions. The kon values for the reverse reaction appear to increase with pH, as would be expected, if this step involves the reaction of the electrophile CS2 with the thiolate group of the product. This type of reactivity is discussed in greater detail below. No significant effects were seen for increasing ionic strength from 0.154 to 0.308 M or buffer concentration from 50 to 100 mM.

Table 1

koff and kon Values Determined for the Decay of PTTC– in Aqueous Phosphate Buffer Solution at 37 °C

pH	koff (in 10–3 s–1)	kon (M–1 s–1)
6.5a	2.43 ± 0.13	0.06 ± 0.03
7.4b	2.36 ± 0.05	0.08 ± 0.04
7.8	2.36 ± 0.02	0.22 ± 0.01

50 mM phosphate buffer and μ = 0.154 M.

Average of values determined at 50 and 100 mM phosphate and μ = 0.154 and 0.308 M.

A linear Eyring plot of the koff values determined for the decay of PTTC– in pH 7.4 aqueous buffer solution over the temperature range of 5–55 °C (SI Figure S3) gave the activation parameter values ΔH‡ = 70.3 ± 0.7 kJ mol–1 and ΔS‡ = −69 ± 2 J K–1 mol–1.

MPTTC–, the n class="Chemical">methyl ester of PTTC–, shows similar behavior (eq , SI Figure S4). At 37 °C, the rate constant for decay was (2.40 ± 0.01) × 10–3 s–1. Kinetics data collected at 27.1, 37.0, and 46.4 °C at pH 7.4 gave the activation parameters ΔH⧧ = 75.7 ± 0.1 kJ mol–1 and ΔS⧧ = −51.4 ± 0.4 J K–1 mol–1.

N-Acetylcysteine trithion class="Chemical">carbonato anion (NacTTC–, eq ) behaved similarly but gave larger values for both koff (0.0118 ± 0.004 s–1) and kon (0.7 ± 0.2 M–1 s–1) in pH 7.4 aqueous phosphate buffer solution (100 mM) at 37 °C. An Eyring plot of koff values measured over the temperature range of 5–55 °C (SI Figure S3) gave the activation parameter values ΔH‡ = 66.8 ± 0.9 kJ mol–1 and ΔS‡ = −67 ± 3 J K–1 mol–1 quite similar to those determined for the decay of PTTC–.

Analogous kinetics studies for decay of the benzyl trithiocarbonate anion (Bnn class="Chemical">TTC–, eq ) displayed decreases in the 302 and 332 nm absorption band characteristic of such TTC– anions (SI Figure S5). The reactions are somewhat faster than seen for PTTC–, but again the rate constants koff for the forward reaction showed essentially no sensitivity to the solution pH over the range of 6.5–10.1 (Table ). In contrast, the kon value calculated was markedly larger at the highest pH, and this was reflected by the reaction not going to completion but reaching an equilibrium or steady state (Figure ).

Table 2

koff and kon Values Determined for the Decay of BnTTC– in Aqueous Phosphate Buffer Solution at 37 °C

pHa	koff (in 10–3 s–1)	kon (M–1 s–1)
6.5	8.8 ± 0.1
7.4	8.5 ± 0.1	0.7 ± 0.1
7.8	8.5 ± 0.1	1.8 ± 0.1
10.1	7.3 ± 0.1	58 ± 1

100 mM phosphate buffer and μ = 0.18–0.29 M.

Figure 2

Effect of pH on the decay kinetics of BnTTC– at 37 °C and in 0.1 M phosphate buffer, except for pH 10.1, which is in 0.1 M carbonate buffer (all experiments done in duplicate).

The absence of a pH effect on the koff values in 37 °C aqueous media for the decays of PTTC– and Bnn class="Chemical">TTC– is consistent with the rate-limiting step dissociation of CS2 from the conjugate thiolate anion, RS–, followed by protonation of the latter, as illustrated in Scheme . In accord with this sequence where the first step is rate limiting, that is koff = k1, the reactivity order NacTTC– > BnTTC– > PTTC– ∼ MPTTC at pH 7.4 is roughly antiparallel to the increasing basicity of the thiolate anion RS– as reflected in pKa values of the respective thiols (RSH) (SI Table S1).[12−16]

Scheme 1

Proposed Sequence of Steps Leading to TTC– Decay in Aqueous Media

Effect of pH on the decay kinetics of BnTTC– at 37 °C and in 0.1 M n class="Chemical">phosphate buffer, except for pH 10.1, which is in 0.1 M carbonate buffer (all experiments done in duplicate).

In other words, the more basic n class="Chemical">thiolate ions are slower to dissociate from the CS2 electrophile. However, one might expect a positive value of ΔS‡ for unimolecular dissociation of the RS–CS2 bond illustrated in the first step of Scheme , but that was not the case for PTTC–, MPTTC–, or NacTTC–. One possible explanation for the observed negative ΔS‡ values would be a pathway involving concerted protonation of the exiting thiolate group by a general acid (solvent or buffer conjugate acid). However, no general acid catalysis was observed and protonation by H2O itself would generate hydroxide ion, which would be unfavorable. Thus, a more likely explanation for a negative entropy of activation draws from solvent reorganization as the negative charge delocalized over the −CS2– functional group becomes localized on the thiolate ion at the transition state (Scheme ).

Scheme 2

Solvation Reorganization upon CS2 Dissociation from a TTC– Anion

Formation of Trithiocarbonates

Physiologically, thiols are likely targets in the action of n class="Gene">CS2 either as a toxin or in potential bioregulatory or therapeutic roles.[1] For example, modification of a key protein thiol by the formation of trithiocarbonate would be expected to have profound effects on that protein’s activity. The “on” reaction noted above is, of course, the formation of the TTC– adduct from the parent thiol plus CS2. The goal in this section is to examine the dynamics of trithiocarbonate formation with cysteine and several cysteine derivatives, including glutathione, in greater detail.

Figure illustrates the temporal spectrum changes that occur rapidly after stopped-flow mixing of a solution containing excess cysteine [n class="Chemical">CysSH] with a second solution containing CS2 in pH 7.4 aqueous buffer at 37 °C. Very similar spectral changes were shown to result from the reactions of CS2 with (respectively) N-acetylcysteine (NacSH), glutathione (GSH, SI Figure S6), and cysteine methyl ester (MecSH, SI Figure S7). In each case, the appearance of the spectrum characteristic of a TTC– anion (eq ) followed an exponential rise as seen in the figure inset.

Figure 3

Temporal absorption spectra recorded over a period of 60 s of a solution prepared by 1:1 stopped-flow mixing of aqueous solutions of cysteine and CS2 with spectra recorded every 0.6 s. Concentrations after mixing: [CysSH] = 5 mM, [CS2] = 0.2 mM. Conditions: T = 37 °C, pH 7.4, [phosphate buffer] = 50 mM, μ = 154 mM. Inset: absorbance change at 332 nm.

Figure illustrates the absorption changes at 332 nm characteristic of formation of CysTTC– (eq ) when a pH 7.4 buffered solution of excess n class="Chemical">CysSH was mixed with a solution containing CS2 in the stopped-flow spectrometer. Under such “pseudo-first-order” conditions where [CysSH] ≫ [CS2], the temporal absorption changes can be fit to a simple exponential function from which the observed first-order rate constant kobs was obtained. Under these conditions where the system is relaxing to equilibrium and CS2 is the limiting reactant, the relationship described by eq holds true.[17]Thus, when RSH = CysSH, a plot of kobs versus [CysSH] should be linear with a slope equal to kon and an intercept equal to koff. The inset of Figure is such a plot for the reaction of CS2 with excess CysSH in 37 °C, pH 7.4 buffered aqueous solutions from which the values kon = 2.9 ± 0.1 M–1 s–1 and koff = 0.103 ± 0.002 s–1 were determined. Plots similar to Figure were generated for the reactions of glutathione (GSH), cysteine methyl ester (MecSH), and N-acetylcysteine (NacSH) (SI Figures S6–S8), and the kon and koff values so determined are summarized in Table . Notably, the values of kon and koff determined in 37 °C, pH 7.4 solution for NacSH (0.606 ± 0.008 M–1 s–1 and 0.0138 ± 0.001 s–1, respectively) can be compared to those (0.7 ± 0.2 M–1 s–1 and 0.018 ± 0.004 s–1) obtained above by following the decay of NacTTC– under similar conditions. Given the different procedures and apparatus used in these two experiments, the agreement is quite good.

Figure 4

Exponential fit (red curve) of absorption changes at 332 nm (black dots) upon stopped-flow mixing of cysteine (1 mM) with CS2 (0.1 mM) in 37 °C, pH 7.4, aq phosphate buffer (50 mM, μtot =154 mM). Inset: plot of kobs values calculated from similar fits at various [CysSH]. Conditions after mixing: pH 7.4, [buffer] = 100 mM, T = 37 °C, [cysteine] = 5–30 mM, [CS2] = 0.5 mM. kon = 2.9 M–1 s–1 (slope), koff = 0.103 s–1 (intercept).

Table 3

Values of kon and koff Determined Using a Stopped-Flow Kinetics Spectrophotometer for the Reactions of CysSH, NacSH, MecSH, and GSH with CS2 in 37 °C, pH 7.4 Buffered Aqueous Solutiona

RSH	pKa (SH)	pKa (NH3+)	kon (M–1 s–1)	koff (s–1)	kon/koff (M–1)
MecSH	6.56b	8.99b	4.3 ± 0.1	0.116 ± 0.002	37
CysSH	8.33c	10.78c	2.9 ± 0.1	0.103 ± 0.002	28
GSH	8.66b	9.12b	2.5 ± 0.1	0.036 ± 0.001	69
NacSH	9.52b		0.606 ± 0.004	0.0138 ± 0.001	44

[CS2] = 0.5 mM, 100 mM phosphate buffer solution, μ = 0.308 M.

Ref (18).

Ref (19).

Temporal absorption spectra recorded over a period of 60 s of a solution prepared by 1:1 stopped-flow mixing of aqueous solutions of cysteine and n class="Gene">CS2 with spectra recorded every 0.6 s. Concentrations after mixing: [CysSH] = 5 mM, [CS2] = 0.2 mM. Conditions: T = 37 °C, pH 7.4, [phosphate buffer] = 50 mM, μ = 154 mM. Inset: absorbance change at 332 nm.

Exponential fit (red curve) of absorption changes at 332 nm (black dots) upon stopped-flow mixing of cysteine (1 mM) with n class="Gene">CS2 (0.1 mM) in 37 °C, pH 7.4, aq phosphate buffer (50 mM, μtot =154 mM). Inset: plot of kobs values calculated from similar fits at various [CysSH]. Conditions after mixing: pH 7.4, [buffer] = 100 mM, T = 37 °C, [cysteine] = 5–30 mM, [CS2] = 0.5 mM. kon = 2.9 M–1 s–1 (slope), koff = 0.103 s–1 (intercept).

[CS2] = 0.5 mM, 100 mM n class="Chemical">phosphate buffer solution, μ = 0.308 M.

As noted above for several other TTC anions, the koff values reported in Table for the n class="Chemical">cysteine derivatives decrease as the basicity of the RS– increases owing presumably to the higher RS–CS2– bond strength through this series. The values of kon at pH 7.4 show a similar decrease (hence, the ratio kon/koff varies only modestly over this series). The behavior of kon would be consistent with the mechanism suggested by the microscopic reverse of Scheme . If koff = k1, then kon = k2 f(H+), where f(H+) = Ka/(Ka + [H+]). As the acidity of RSH (Ka) increases, more of the conjugate base thiolate RS– is available at pH 7.4. Thus, one would expect kon to increase. However, the nucleophilicity of RS– is likewise also decreasing over this series, so one might expect the rate constant k2 for the nucleophilic attack of RS– on CS2 to correspondingly decrease in going from NacS– to MeCysS–, the two trends therefore countering each other. This may help account for a reactivity difference on only an order of magnitude between NacSH and MeCysSH at pH 7.4 despite the much greater difference in Ka values.

SI Figure S9 displays Eyring plots of the kon and koff values determined in a similar manner for the reaction of n class="Chemical">CysSH with CS2 in pH 7.4 buffered aqueous solution over the temperature range of 25–45 °C. The apparent activation parameters for kon are ΔH⧧ = +84.3 kJ mol–1 and ΔS⧧ = +37.4 J K–1 mol–1, and those for koff are ΔHoff⧧ = +75.1 kJ mol–1 and ΔSoff⧧ = −21.7 J K–1 mol–1. Notably, ΔH⧧ for koff in this case are quite similar to that recorded above at pH 7.4 for the decays of PTTC–, MPTTC–, and NacTTC–, but because ΔS⧧ remains negative in this case, it is less so than for the other three (SI Table S2). According to the proposed mechanism described in Scheme , koff = k1, so the apparent ΔH⧧ and ΔS⧧ for this step equal ΔH1⧧ and ΔS1⧧, respectively. The relationship is more complex for kon because it equals k2 f(H+). If one were to make the rough approximation that, at pH 7.4, [H+] ≫ Ka, then kon ≅ k2Ka[H+]−1 and then ΔHon⧧ ≅ ΔH2⧧ + ΔHa0 and ΔSon⧧ ≅ ΔS2⧧ + ΔSa0. The values of ΔHa0 and ΔSa0 can be calculated as +30.9 kJ/mol and −52 J K–1 mol–1, respectively, from the pH dependence of the pKa of cysteine (SI Figure S10).[19] Thus, on the basis of this model, ΔH2⧧ ≅ 53 kJ mol–1 and ΔS2⧧ ≅ +89 J K–1 mol–1 for the reaction of CysSH with CS2.

Table reports the results of analogous stopped-flow kinetics studies of the reactions of CysSH and n class="Chemical">GSH with CS2 to form CysTTC– and GTTC– (eqs and 9, respectively) at different pH values over the range of 7.4–9.6. The kobs values were determined for a range of different initial CysSH concentrations at each pH, and linear plots of kobs versus [CysSH] similar to those shown in Figure and SI Figures S6–S8 gave the pH-dependent values of kon and koff. Both systems show dramatic increases in kon at higher pH as anticipated from the simple mechanism proposed in Scheme . Consistent with the experiments for PTTC– and BnTTC– reported in Tables and 2, the koff values for GTTC– are pH-independent; however, this was not the case for CysTTC–, for which koff decreases by about a factor of four from pH 7.4 to 9.6. We attribute this difference to the relative proximity of the protonated amine group (NH3+) to the trithiocarbonate functionality in CysTTC–. The resulting change in the inductive effect upon deprotonation of this group at higher pH should enhance the basicity of the thiolate, thus leading to slower dissociation of CS2. Although GTTC– has a similar amine, it is positioned further from the TTC functionality and its deprotonation would have a much smaller impact.

Table 4

Rate Constants Measured for Reaction between CS2 (0.1 mM) and CysSH (1–8 mM) or GSH (1–8 mM) as a Function of pH at 37 °C in Buffered Aqueous Solution

	CysSH			GSH
pH	kon (M–1 s–1)	koff (s–1)	kon/koff (M–1)	kon (M–1 s–1)	koff (s–1)	kon/koff (M–1)
7.4	2.9	0.103	28	2.5	0.036	69
7.6	6.4	0.099	65
8.0	11.7	0.086	136	8.7	0.039	224
8.4	14.9	0.075	198	13.7	0.036	380
8.8	18.4	0.054	340	23.5	0.034	692
9.2	22.0	0.039	563	32.5	0.034	956
9.6	26.0	0.023	1129	46.9	0.034	1340

Subsequent Reactions of CysTTC– and GTTC–

Although the reaction of CS2 with n class="Chemical">cysteine initially showed the rapid appearance of new absorption bands at ∼295 and 332 nm consistent with the formation of the trithiocarbonate CysTTC– (eq ), a much slower subsequent reaction was evidenced by further spectral changes (Figure ) involving the disappearance of the bands at ∼295 and 332 nm and the appearance of a new band at 270 nm. The latter band can be assigned to the cyclized compound thiazolidine-2-thione-carboxylate (TTCA, eq ),[9] which has been identified[20,21] as a urinary excretion product from humans who have been exposed to CS2. The temporal decay at 332 nm and rise at 270 nm could be fit to exponential functions to give the respective first-order rate constants 5.2 × 10–5 and 4.9 × 10–5 s–1 in pH 7.4 aq phosphate buffer at 37 °C (SI Figure S11). However, inspection of the spectral changes shows the absence of isosbestic points, as well as an absorbance increase followed by a decrease of a shoulder at ∼230 nm, which are clear indications that a transient intermediate species is formed. When the reaction was run with a very large excess of CysSH ([CS2] = 0.2 mM; [CysSH] = 10 mM), the kobs value measured was nearly the same (5.8 × 10–5 s–1). Thus, it appears that this slow cyclization process is unimolecular and does not involve the reaction of free CysSH.

Figure 5

Spectral changes upon mixing CS2 (1 mM) with CysSH (1 mM) in pH 7.4 aq phosphate buffer (100 mM) at 37 °C in a sealed cell indicating the slow transformation of CysTTC– (red spectrum) to TTCA (blue spectrum). Total time = 17 h. Spectra recorded at 600 s intervals.

The reaction of cysteine (10 mM) and n class="Gene">CS2 (10 mM) was further studied in a buffered phosphate deuterium oxide solution (prepared with anhydrous Na3PO4 plus D2O, pD 7.5) to characterize more thoroughly the formation of TTCA. DCl solution (35 wt % in D2O) was used to correct the pD. The reaction was run for 24 h at 37 °C. A small aliquot of the product solution was used to check the electronic spectrum, which showed the presence of a single and intense absorption band at 271 nm (SI Figure S12), indicating the formation of the cyclic compound because unreacted cysteine and oxidized cysteine products, such as cystine, have no strong absorption bands in this wavelength range. The 1H NMR spectrum of the product solution (SI Figure S13) showed not only resonances belonging to unreacted cysteine but also two groups of proton resonances at 3.67–3.63 ppm (S–CH2, doublet of doublets) and 3.92–3.89 (CH, broad doublet of doublets) attributed by van Doorn et al. to the hydrogens of TTCA.[22]

Spectral changes upon mixing CS2 (1 mM) with n class="Chemical">CysSH (1 mM) in pH 7.4 aq phosphate buffer (100 mM) at 37 °C in a sealed cell indicating the slow transformation of CysTTC– (red spectrum) to TTCA (blue spectrum). Total time = 17 h. Spectra recorded at 600 s intervals.

GTTC– formed by the relatively rapid reaction of n class="Chemical">GSH with CS2 also goes through a very slow transformation to a new species reported to be TTCA (SI Figure S14). However, similar spectral changes/secondary reaction was not seen with 3-trithiocarbonatopropionate (PTTC–) but was observed with 3-trithiocarbonatopropylamine and thus an amine functionality appears to be required. This is not simply the transfer of CS2 from the thiolate to the amine to form a dithiocarbamate (DTC–) analogue because the spectrum formed by the reaction of aqueous base and glycine displays bands at 253 and 284 nm (data not shown) typical of DTC anions and very different from the characteristic of TTCA. The mechanism(s) of these transformations are the subject of continuing investigation.

Summary

We have described the kinetics of the formation and decay of n class="Chemical">trithiocarbonate derivatives formed by the reactions of carbon disulfide and various thiols (RSH) under physiologically relevant conditions (near-neutral pH, 37 °C). Rates of TTC– formation are strongly dependent on solution pH, indicating that the rate-limiting step involves the reaction of the thiolate anion RS– with the electrophilic CS2 substrate. Rates of TTC– decay are pH-independent, consistent with microscopic reversibility. The relative decay rates have a reverse correlation to the basicity of the RS– anions as expected if simple dissociation of CS2 is rate determining, as illustrated by the qualitative reaction coordinate diagram shown in Figure . At pH 7.4, glutathione and cysteine react with CS2 at similar rates, and in both cases, the resulting TTC– anions undergo slow cyclization reactions to a cyclized species identified previously as 2-thiothiazolidine-4-carboxylic acid. As noted above, the latter has been found to be urinary excretion products from individuals exposed to toxic levels of carbon disulfide. Similar reactions with protein thiols could lead to irreversible modifications of such targets as well as releasing an equivalent of H2S.

Figure 6

Qualitative reaction coordinate diagram for the formation (kon) and decay (koff) of trithiocarbonate anions (RSCS2–) from the respective thiols plus carbon disulfide. According to Scheme , TTC– decay is independent of pH (koff = k1), whereas the forward reaction is strongly pH-dependent owing to the RSH ⇌ RS– + H+ equilibrium.

Qualitative reaction coordinate diagram for the formation (kon) and decay (koff) of n class="Chemical">trithiocarbonate anions (RSCS2–) from the respective thiols plus carbon disulfide. According to Scheme , TTC– decay is independent of pH (koff = k1), whereas the forward reaction is strongly pH-dependent owing to the RSH ⇌ RS– + H+ equilibrium.

The present data do not resolve the question whether CS2 might serve some type of bioregulatory role. Although n class="Chemical">thiols would appear to be logical targets for this electrophile, it is clear that the rates and equilibrium constants for TTC– adduct formation are many orders of magnitude smaller than those for reactions of NO with heme proteins, the best characterized targets of that SMB.[23,24] It appears likely that any such bioregulatory relevant sites for CS2 would have higher binding constants, for example, an activated thiolate protein, where the TTC– anion is stabilized by other intramolecular interactions. Another could be a metal center because it is well established that transition-metal ions activate CS2 toward reaction with coordinated nucleophiles and insertion into metal–ligand bonds, owing to the strong bonding of the 1,1-dithiolate (−CS2–) functionality as a ligand. Such reactivity has been reported for metal–sulfur bonds[25,26] as well as with alkoxide,[27] amine,[28,29] and (even) phosphine[30] ligands. Given the relatively high cytosolic concentrations of GSH[31] and plasma-based protein sulfhydryls,[32] the reversible formation of unstabilized TTCs may play a role in CS2 transport.

Experimental Section

Materials

Except where otherwise noted, all materials were of analytical or reagent grade and were used without further purification. N-Acetyl-l-cysteine (≥99%, TLC), n class="Chemical">3-mercaptopropionic acid (≥99%), l-cysteine (97%), glycine (≥98.5%), and reduced l-glutathione (98%) were purchased from Sigma-Aldrich. Carbon disulfide (ACS Reagent Grade, ≥99.9%), benzyl mercaptan (Alfa Aesar, 99%), potassium hydroxide, sodium hydroxide, and mono- and dibasic sodium phosphate and sodium chloride used to prepare buffers were purchased from Fisher Scientific.

Synthesis of TTCs

Sodium 3-(Trithiocarbonato)propionate (Na2[PTTC])

This species had been generated as an intermediate in the preparation of a corresponding trithiocarbonate ester.[33] A round-bottom flask was charged with n class="Chemical">pentane and CS2 (30 mL each) plus a stir bar and then was cooled to approximately 0–5 °C with an ice bath and purged with inert gas. During rapid stirring, finely ground NaOH (3.09 g, 77.4 mmol) was added to the flask under inert gas. Sufficient water was added to solubilize the hydroxide (1–3 mL), and to the resulting base solution, 3-mercaptopropionic acid (4.11 g, 38.7 mmol) in diethyl ether (10 mL) was added dropwise over the course of a few minutes. A deep yellow powder immediately began to precipitate. The reaction was allowed to warm to room temperature, and stirring was continued for 6–12 h. The solid was collected via filtration and rinsed with cold Et2O. The yield was 3.68 g (37%). The product proved to be soluble only in methanol or water; however, attempts to recrystallize from these solvents resulted in decomposition. The optical spectrum of the initial product in pH 7.4 water displayed bands at λmax = 303 and 332 nm with an extinction coefficient at 332 nm (8650 M–1 cm–1) close to those previously observed for alkyl TTCs.[21] The infrared spectrum and CHN analysis are consistent with a hydrated sample. The solid is rather odorous. The compounds were kept in a freezer for long-term storage. Infrared spectrum: (attenuated total reflection in cm–1) selected bands 989, 879 (ν, C–S), 1148 (C=S), 1395 (δ, C–O), 1555 (ν, C–O), 1655 (δ, O–H), 3300 (ν, O–H). Anal.: (calculated values for C4H4Na2S3·2H2O in parentheses): C, 18.1 (18.32); H, 2.8 (3.07).

Triethylammonium (N-Acetylcysteine)trithiocarbonate ([HNEt3][NacTTC])

To an ice-cold (0–5 °C) flask containing diethyl ether (30 mL) and n class="Chemical">triethylamine (3.72 g, 36.8 mmol) under inert atmosphere, a 3 g portion of N-acetylcysteine (18.4 mmol) was added. Dimethylformamide (2–3 drops) was added to solubilize the N-acetylcysteine, after which CS2 (1.90 g, 25 mM) was added dropwise, resulting in rapid separation of an insoluble, viscous, translucent, golden oil. This oil was separated from the solvent by decanting and was dried in vacuo for several days at elevated temperature (∼40 °C). The oil was then dissolved in ethanol, heated gently, and the resulting solution stirred vigorously. Vacuum removal of the ethanol resulted again in a golden oil. This product was stored in a freezer. Optical spectrum: λmax = 302 and 333 nm in pH 7.0 water with a molar extinction coefficient of ∼8400 M–1 cm–1 at the latter wavelength. ESI-MS (neg. mode) (75% ACN, 25% H2O): Anion predicted (+H+), 237.96 m/z; found, 237.94 m/z, 259.89 m/z (+Na+), 339.04 m/z (+TEAH+). Also observed, 162.17 m/z (N-acetylcysteine, impurity/decomposition). Anal. (calculated values for C18H39N3O3S3 in parentheses): C, 48.0 (48.95); H, 8.22 (8.90); N, 9.39 (9.51).

Potassium Benzyl Trithiocarbonate (K[BnTTC])

The preparation was adapted from a known procedure.[5] A round-bottom flask was charged with a magnetic stirring bar, diethyl ether (30 mL), and n class="Chemical">carbon disulfide (10 mL). The solution was cooled to 0–5 °C with an ice bath and then purged with nitrogen. Finely ground KOH (2.00 g, 35.6 mmol) was added and then benzyl mercaptan (5 mL, 5.29 g, 42.6 mmol) was added to this cloudy mixture over a period of 2 min. The resulting solution rapidly became yellow and cloudy to give a chalky precipitate. This mixture was stirred for an additional 3 h. The solid product was collected by filtration and washed with diethyl ether and cold ethanol (4 × 10 mL). The solid was then added to a flask containing 150 mL of Et2O and collected by filtration again to remove excess dimethylformamide. Recrystallization from a 50/50 mixture of ethanol and acetonitrile gave a deep yellow, mildly odorous powder that was dried in vacuo and stored in a freezer. The yield was 4.58 g (54%). K[BnTTC] decomposes when dissolved in water or water/alcohol mixtures to give highly odorous product(s). CHN anal. (calculated values for C8H7KS3 in parentheses): C, 39.8 (40.3); H, 2.84 (2.96). Optical spectrum: λmax = 302 and 332 nm in pH 7.4 water (extinction coefficient of ∼9300 M–1 cm–1 at the latter wavelength). (Note: Benzyl mercaptan has a particularly unpleasant odor and is toxic. Proper environmental controls must be observed when handling this liquid.)

Kinetics Methods

Phosphate buffers were prepared using n class="Chemical">mono- and dibasic sodium phosphate and sodium chloride (to maintain ionic strength). Nanopure water (≥18 megohm) was obtained from a Barnstead Nanopure II system and used in solution preparations.

The rates of the TTC decompositions were determined by monitoring temporal spectral changes on a Shimadzu UV-2401 spectrophotometer with UVProbe kinetics software. The reaction cell was a septum-capped, sealable quartz cuvette with a 1.0 cm path length and an approximate volume of 4.80 mL. The cells were thermostated (±0.2 °C) at the desired reaction temperature, and the solutions were stirred with a Starna “Spinnette” stirrer. Septum-sealed cells containing the buffer solution of interest temperature equilibrated (7–10 min) under the desired conditions (typically 37 °C). A small amount of the solid n class="Chemical">TTC salt (5–10 mg) was then added to a vial containing 3–5 mL of buffer to prepare a stock solution. A 100 μL aliquot of this stock solution was then syringed into the thermostated cuvette. The volumes were chosen to minimize the headspace in the cuvette. Data acquisition was initiated after an estimated dead time of 60–90 s.

Kinetics of the reactions of CS2 with various n class="Chemical">thiols to examine the rates of TTC formation were generally performed using an Applied Photophysics SX.19MV stopped-flow UV–visible spectrophotometer with photodiode array (temporal spectra) or a photomultiplier tube (single wavelength) detector. Aqueous buffered solutions of the thiol (RSH) and CS2 were loaded into Hamilton salt syringes (#1725 AD-SL), which were attached to the mixing block of the spectrophotometer and sealed from the outside atmosphere by a three-way valve. These syringes were maintained at the desired temperature (generally 37 °C unless otherwise noted) by a circulating water bath. The connector tubing and observation cell were purged with the sample solution by filling and emptying the drive syringes 3× prior to data acquisition. Stopped-flow mixing of the two solutions initiated the reaction. The entire unit was controlled with “Applied Photophysics Pro-Data SX” software. Only simultaneous symmetric mixing was used (1:1 dilution).

Kinetics data were processed using Igor Pro 6.37 software by WaveMetrics, Inc., and Prism 7 by GraphPad Software, Inc. For global fittings of equilibrium analyses and determination of n class="Disease">back-reaction values, a DynaFit (BioKin Ltd) nonlinear least-squares regression program was used.

Computational Studies

All density functional theory (DFT) computations were performed using Gaussian’09 software packages. Optimizations were performed using Kohn–Sham DFT with the hybrid M06-2X exchange-correlation functional and the 6-311+G(d,p) n class="Chemical">basis set. No symmetry restriction was imposed, and implicit solvent effects were included using PCM (solvent = water) methods, as implemented in Gaussian 09. Vibrational frequency calculations were performed at the same level of theory to verify that no imaginary frequencies were present and to ensure the true local minima energies. Single-point energy and time-dependent DFT calculations were performed using the M062X/6-311+G(3df,3pd) level of theory.

Analysis of CS2 Release

This procedure was similar to that described by Schwach and Nyanzi[34] and to that used in this laboratory to detect n class="Gene">CS2 released by the photolysis of CdSe quantum dots surface decorated with 1,1-dithiooxalate[7] and for the CS2 released in the thermal reactions of dithiocarbamate salts.[8]