Mechanism of nitric oxide release from S-nitrosothiols.

S-Nitrosothiols have many biological activities and have been suggested to be intermediates in signal transduction. The mechanism and products of S-nitrosothiol decomposition are of great significance to the understanding of nitric oxide (·NO) biochemistry. S-Nitrosothiols are stable compounds at 37°C and pH 7.4 in the presence of transition metal ion chelators. The presence of trace transition metal ions (present in all buffers) stimulates the catalytic breakdown of S-nitrosothiols to ·NO and disulfide. Thiyl radicals are not formed as intermediates in this process. Photolysis of S-nitrosothiols results in the formation of ·NO and disulfide via the intermediacy of thiyl radicals. Reduced metal ion (e.g. Cu+) decomposes S-nitrosothiols more rapidly than oxidized metal ion (e.g. Cu2+) indicating that reducing agents such as glutathione and ascorbate can stimulate decomposition of S-nitrosothiol by chemical reduction of contaminating transition metal ions. Transnitrosation can also stimulate S-nitrosothiol decomposition if the product S-nitrosothiol is more susceptible to transition metal ion-catalyzed decomposition than the parent S-nitrosothiol. Equilibrium constants for the transnitrosation reactions of reduced glutathione, either with S-nitroso-N-acetyl-DL-penicillamine or with S-nitroso-L-cysteine indicate that S-nitrosoglutathione formation is favored. The biological relevance of S-nitrosothiol decomposition is discussed.

ing to this proposal, S-nitrosothiols are synthesized chemically by reaction of ⅐ NO with thiol. Subsequently, these compounds are transported or diffuse to the site of action. Decomposition of the S-nitrosothiol then leads to ⅐ NO release and the corresponding biological effect. This hypothesis is mainly speculative and remains to be rigorously tested. Little is known about the reaction of ⅐ NO with glutathione (GSH) in vivo; however, the direct reaction of GSH with ⅐ NO does not generate GSNO but forms glutathione disulfide and nitroxyl anion (NO Ϫ ) (4,5). GSNO is formed only if ⅐ NO is oxidized, by reaction with oxygen, to form ⅐ NO 2 and N 2 O 3 (6). As intracellular oxygen concentrations at the tissue level are in the range of 10 -20 M (7) and as the rate of ⅐ NO oxidation is proportional to the squared power of the ⅐ NO concentration (8), it is likely that the oxidation of ⅐ NO by oxygen in vivo is a slow and insignificant process (4). Evidence for the formation of S-nitrosothiols from endogenous ⅐ NO remains scarce (9). Nevertheless, nitrosylation of protein thiols has been implicated in the ⅐ NO-dependent regulation of many enzymes, including protein kinase C (10) and glyceraldehyde-3-phosphate dehydrogenase (11). It has been reported that normal human serum contains S-nitroso-serum albumin (12,13) which has been proposed to act as an endogenous regulator of vessel tone (14).
Although the physiological relevance of S-nitrosothiols remains to be established, these compounds have been used as donors of ⅐ NO (1,15,16). The most commonly employed compounds are GSNO and S-nitroso-N-acetyl-DL-penicillamine (SNAP) (Fig. 1A). Such compounds have been shown to have diverse and remarkable biological effects. For example, SNAP is a potent vasodilator (1) and low concentrations of GSNO have been shown to afford significant protection to the ischemic myocardium (17). It is generally assumed that S-nitrosothiols decompose by homolytic cleavage of the S-N bond (Reaction 1). RSNO 3 RS ⅐ ϩ ⅐ NO REACTION 1 This process generates ⅐ NO and a thiyl radical, RS ⅐ (18). However, this assumption has not been effectively tested under physiologically relevant conditions. It has been established that S-nitrosothiols are sensitive to both photolytic (19,20) and transition metal ion-dependent breakdown (21) but are stable in the presence of transition metal ion chelators in the dark. The biological activity of S-nitrosothiols may not be exclusively dictated by the release of ⅐ NO as the chemistry of these compounds is complex. S-Nitrosothiols have also been shown to form NO Ϫ , which under appropriate conditions can lead to the formation of either nitrous oxide (4,5,22) or peroxynitrite (4,23). S-Nitrosothiols can also undergo nitrosonium (NO ϩ ) transfer to other cellular thiols by a process referred to as transnitrosation (24).
In this study we have investigated the mechanism of decomposition and transnitrosation reactions of S-nitrosothiols using electron spin resonance (ESR), optical spectroscopy, high performance liquid chromatography (HPLC), and electrochemical methods.
ESR Measurements-ESR spectra were recorded at room temperature on a Varian E109 spectrometer operating at 9.5 GHz and employing 100 kHz field modulation. Samples were prepared in the nitrogen glove box and taken up in a 100-l capillary (Corning), which was sealed at both ends with Miniseal (Baxter). The capillary was placed in a 4-mm quartz tube, which was then placed inside the ESR cavity. All samples were prepared in the dark and irradiated, when required, with visible light ( Ͼ 408 nm) inside the ESR cavity. Data were collected using the VIKING software, developed at the Medical College of Wisconsin, and simulated with the ESR software developed by David Duling from the Laboratory of Molecular Biophysics, NIEHS, National Institutes of Health, Research Triangle Park, NC. 2 Irradiation Procedures-Photolytic decomposition of GSNO was performed in 8-ml glass vials. Samples were irradiated while stirring with light from xenon arc lamp source (ILC Technology, Sunnyvale, CA) after passing through a copper(II) sulfate solution (100 g/liter) and a long pass filter of Ͼ 408 nm. Aliquots were removed for GSSG measurements. Samples for ESR studies were prepared in glass capillaries and were irradiated inside the ESR cavity.
HPLC Methods-Separation of GSNO and GSSG was performed on a Hewlett-Packard 1050 series HPLC system equipped with UV-vis detection. The mobile phase was 0.05% trifluoroacetic acid and methanol (94:6) with a flow rate of 0.75 ml/min. The stationary phase was an analytical C 18 reversed phase column (Partisil ODS-3, 5-m particle size, Whatman, Hillsboro, OR). UV absorption was used to detect GSNO and GSSG ( ϭ 220 nm).
⅐ NO Detection-⅐ NO was detected directly using a ⅐ NO electrode (World Precision Instruments Inc., Sarasota, FL) in a thermostated oxygen-electrode chamber (Yellow Spring Instruments, Yellow Springs, OH) modified by the addition of a glass window for irradiation purposes. The ⅐ NO electrode was stabilized at 37°C in PBS before use. The presence of ⅐ NO was confirmed using NNO, which rapidly scavenges ⅐ NO to form the corresponding imino nitroxide (29).
Kinetic Analysis of Transnitrosation Reactions-Transnitrosation reactions between SNAP and either GSH or CySH were monitored at 37°C in PBS containing DTPA (100 M) by following the change in absorbance at 234 nm. The initial rate of the transnitrosation reaction at a range of thiol concentrations was used to measure the rate constant (k 1 ) for this reaction according to Equation  [SNAP] 0 and [RSH] 0 represent the initial concentrations of SNAP and RSH. A similar series of experiments using NAP and either GSNO or CySNO was conducted to determine the rate constant for the reverse reaction (k Ϫ1 ). Equilibrium constants (K) were calculated from the ratio of the forward and the reverse rate constants.

Detection of Thiyl Radicals during the Decomposition of S-Nitrosothiols-
The chemical structures of the thiols and Snitrosothiols used in this study are shown in Fig. 1A. In order to measure the production of thiyl radicals from the decomposition of S-nitrosothiols, these compounds were mixed with DMPO at 37°C and ESR spectra were obtained. GSNO, SNAP, and CySNO, in the dark, gave no ESR-detectable DMPO spin adducts (Fig. 2, A, C, and E, respectively). Irradiation of a mixture of GSNO and DMPO with visible light ( Ͼ 408 nm) gave a four-line ESR spectrum (Fig. 2B), that could be simulated (Fig. 2B, dotted  The spectra in Fig. 2 were collected in the absence of transition metal ion chelators and under these conditions the Snitrosothiols decomposed in the dark with relative rates CySNO Ͼ SNAP Ͼ GSNO (Ref. 30 and data not shown). The observation that DMPO thiyl radical adducts were not formed in the dark suggests that transition metal ion-induced decomposition of S-nitrosothiols does not proceed through a thiyl radical intermediate (Fig. 2).

Effect of DMPO on GSSG Formation during Light-or Transition Metal Ion-induced Decomposition of S-Nitrosothiols-To
further investigate the intermediacy of thiyl radical, GSNO was allowed to decompose both in the dark and under irradiation, in the presence and absence of DMPO, and the extent of conversion of GSNO to GSSG was monitored. DMPO has been shown to react rapidly with thiyl radicals (k ϭ 10 7 M Ϫ1 s Ϫ1 ) (31,32). If thiyl radicals are formed, it is expected that they will react with DMPO and reduce the yield of GSSG.
Photolytic and transition metal ion-mediated decomposition of GSNO liberated ⅐ NO and resulted in the formation of GSSG (Reaction 2). Fig. 3 shows a typical HPLC trace for the separation of GSNO from GSSG. The retention times for GSNO and GSSG were 8.5 and 9.5 min, respectively. In the absence of DMPO, irradiation of GSNO resulted in the stoichiometric conversion of GSNO to GSSG (Fig. 3A) (20 nmol of GSNO produced 10 Ϯ 1.2 nmol of GSSG). In the presence of DMPO, irradiation resulted in an identical decrease in GSNO, indicating that DMPO did not affect the amount of light absorbed by GSNO; however, the concomitant formation of GSSG was inhibited (Fig. 3B). The kinetics of GSNO decay and GSSG formation are shown in Fig. 4. Decomposition of GSNO in the dark was a slow process (t 1 ⁄2ϭ 80 h) (Fig. 4A) that results in the formation of GSSG. 3 DMPO(100 mM) had no effect on the formation of GSSG. Decomposition of GSNO under irradiation was more rapid (t 1 ⁄2ϭ8 min) (Fig. 4B) and also resulted in the formation of GSSG. DMPO (100 mM) did not affect GSNO decomposition but resulted in a 50% reduction in GSSG formation. At higher concentrations of DMPO (200 mM), an 80% reduction in the yield of GSSG was observed during photolytic decomposition of GSNO (Table I). High concentrations of DMPO are required in order to compete with the diffusion-limited dimerization of GS ⅐ . Addition of Cu 2ϩ (2 M) resulted in a more rapid decomposition of GSNO. Under these conditions, DMPO/ ⅐ SG was not observed by ESR and DMPO (200 mM) did not affect the yield of GSSG (Table I). These results indicate that photolytic decomposition of GSNO occurs through a GS ⅐ intermediate, whereas dark, transition metal ion-mediated decomposition does not.
Effect of Cu 2ϩ and Cu ϩ on ⅐ NO Formation during GSNO Decomposition-Previous studies have indicated that the decomposition of S-nitrosothiols is catalyzed by Cu 2ϩ by a mechanism that does not involve Cu 2ϩ /Cu ϩ interconversion (33). Fig. 5 shows that addition of Cu 2ϩ to GSNO resulted in a rapid release of ⅐ NO. However, addition of Cu ϩ resulted in a much greater stimulation of ⅐ NO release, which was inhibited by DTPA, a transition metal ion chelator (data not shown). These results indicate that the binding of copper ions to GSNO is a prerequisite for enhanced decomposition, as has been described for other S-nitrosothiols (33). Both Cu 2ϩ -and Cu ϩ -induced ⅐ NO release were antagonized by NNO, and ESR analysis indicated that, under these conditions, NNO was partially converted to imino nitroxide confirming the production of ⅐ NO (data not shown). Interestingly, the addition of GSH to GSNO, in the absence of DTPA, stimulated the contaminating transition metal ion-dependent release of ⅐ NO (Fig. 5). In the presence of DTPA, GSH-stimulated GSNO decay does not result in ⅐ NO production (4).
These results are consistent with the proposal that Cu 2ϩ / Cu ϩ redox cycling mechanism is not responsible for GSNO decomposition. However, reduction of Cu 2ϩ to Cu ϩ by GSH, or other reducing agents, will accelerate GSNO decay. Decomposition of GSNO by Cu ϩ may occur by the mechanism shown in Reactions 3 and 4. This represents a catalytic redox cycle for Cu ϩ , which will not occur with Cu 2ϩ unless additional reducing equivalents are present. Under these conditions autoxidation of Cu ϩ to Cu 2ϩ may lead to superoxide and peroxynitrite formation.
Transnitrosation between S-Nitrosothiols and Thiols-Transnitrosation between thiols and S-nitrosothiols has been impli-cated in the biological activity of S-nitrosothiols (34,35). Previous determinations of the kinetic and thermodynamic parameters for these reactions have used extremely high concentrations of both thiol and nitrosothiol (0.5-10 mM) (24,34,35). The reasons for using high concentrations of S-nitrosothiols are: (i) most S-nitrosothiols have similar UV-vis spectra and (ii) the ⌬⑀ max for the transnitrosation reaction is less than 100 M Ϫ1 cm Ϫ1 . SNAP, on the other hand, has an additional absorbance peak at 228 nm, and transnitrosation between SNAP and other thiols can be monitored at 234 nm with a ⌬⑀ max of approximately 3000 M Ϫ1 cm Ϫ1 . Fig. 6A shows the spectral changes that occurred upon addition of GSH (100 M) to SNAP (100 M) in the presence of DTPA. As transnitrosation proceeded the decay of absorbance at 234 nm was accompanied by an increase in absorbance at 200 -220 nm. A sharp isosbestic point at 222 nm was observed, indicating that decomposition of S-nitrosothiol was not occurring. The ⌬⑀ for this change was measured to be 3040 M Ϫ1 cm Ϫ1 at 234 nm from the maximum change at a high concentration of GSH (500 M). The initial rate of SNAP decay, using a range of GSH concentrations (Fig. 6B), was used to calculate a rate constant for this reaction (Table II), assuming a reversible second-order reaction (Reaction 5).
SNAP ϩ GSH º GSNO ϩ NAP REACTION 5 A parallel set of experiments were performed using GSNO (100 M) and NAP (0, 100, 200, 300, 400, and 500 M) (Fig. 6B and data not shown) as the reactants to determine the rate constant of the reverse reaction (Table II). The ratio of these two numbers gave an equilibrium constant of 18.6, indicating that GSNO formation is favored. This is in contrast to Meyer et al. (34), who measured K eq to be 1.13 for this reaction. A similar set of reactions, using SNAP and CySH for the forward reaction and CySNO and NAP for the reverse reaction, gave the rate constants shown in Table II. The calculated equilibrium con- stant of 6.46 indicates that the formation of CySNO is favored in this reaction. This is larger than that measured by Meyer et al. (34), who reported an equilibrium constant of 2.2 for this reaction. The equilibrium constant (K GC ) for the reaction between GSNO and CySH was calculated to be 0.35 from the relationship K GC ϭ K SG /K SC , where K SG is the equilibrium constant for the reaction between SNAP and GSH and K SC is the equilibrium constant for the reaction between SNAP and CySH (Table II). This indicates that GSNO formation is thermodynamically favored in this reaction system.
The observation that photolysis of S-nitrosothiols generates thiyl radicals (Fig. 2) indicates another method by which transnitrosation reactions can be monitored. Photolysis of a mixture of a thiol and an S-nitrosothiol in the presence of DMPO, after a period of dark incubation, will give a "snapshot" of the transnitrosation reaction. Fig. 7 shows the transnitrosation reaction between SNAP and GSH as monitored by ESR. Computer simulations indicate that the presence of DMPO/ ⅐ SG and DMPO/ ⅐ S-NAP is sufficient to account for the spectra shown in Fig. 7. A solution of SNAP gave the ESR spectrum of DMPO/ ⅐ S-NAP (Fig. 7A) immediately after photolysis. However, if the solution was incubated with GSH in the dark for 20 min before photolysis, the ESR spectrum observed was a mixture of both the DMPO/ ⅐ SG and the DMPO/ ⅐ S-NAP spin adducts. After 60 min of dark incubation, the DMPO/ ⅐ SG spectrum dominated. These results are consistent with the optical data indicating that GSNO formation is favored during transnitrosation (Table II).
The transnitrosation reaction between GSH and CySNO (Reaction 6) was monitored by ESR as shown in Fig. 8.

TABLE II Kinetic and thermodynamic parameters of transnitrosation reactions
Reactions between thiols (0 -500 M) and S-nitrosothiols (100 M) were monitored at 234 nm in the presence of DTPA (100 M) at 37°C. Initial rates of the transnitrosation reactions were used to calculate rate constants for the forward (k 1 ) and reverse (k Ϫ1 ) reactions. Equilibrium constants were calculated as the ratio of the forward and reverse rate constants. ND, not determined.  it immediately (within the dead time of sample preparation and ESR analysis which is about 2 min) gave only the DMPO/ ⅐ SG (Fig. 8B). Incubation of GSNO with CySH in the presence of DMPO resulted in no ESR signal under anaerobic conditions (Fig. 8C); however, upon irradiation of the solution (after incubation in dark), it gave only the DMPO/ ⅐ SG (Fig. 8, D and E). This indicates that the transnitrosation reaction between CySNO and GSH is rapid and the equilibrium of Reaction 6 favors GSNO formation.
In the absence of transition metal ions or in the presence of metal ion chelators, the transnitrosation reaction is unlikely to be responsible for the release of ⅐ NO from S-nitrosothiols as both the reactant and product S-nitrosothiol are stable. However, in the presence of contaminating transition metal ions, transnitrosation may accelerate ⅐ NO release. This is shown for the reaction between GSNO and NAP (Fig. 9). In the presence of DTPA and in the dark, GSNO did not generate ⅐ NO and addition of NAP had no effect. In the absence of DTPA, however, GSNO generated a small amount of ⅐ NO, which was dramatically enhanced by the addition of NAP. Irradiation of GSNO, in the presence of DTPA, gave a much greater release of ⅐ NO, which was unaffected by NAP. However, in the absence of DTPA, the addition of NAP resulted in enhanced production of ⅐ NO. Although the equilibrium position of the transnitrosation reaction between GSNO and NAP favors GSNO formation (Reaction 5), SNAP is more susceptible to transition metal ioncatalyzed decomposition (21). SNAP decomposition will pull the equilibrium in Reaction 5 to the left, resulting in enhanced ⅐ NO release. It is also possible, as discussed earlier (cf. Fig. 5), that thiol-dependent metal ion reduction is responsible for the accelerated release of ⅐ NO.

Metal Ion-and Light-induced Decomposition of S-Nitrosothiols-
The decomposition chemistry of S-nitrosothiols has been discussed widely in the literature (36 -40). It has been shown that the biological effects of these compounds do not correlate with the rate of ⅐ NO release in solution (30), suggesting either selective metabolism or direct action of some of the S-nitrosothiols used in this study. Mathews and Kerr (30) highlighted the fact that the rate of decomposition of S-nitrosothiols in the experimental buffer cannot be used as an indicator of biological activity. The factors affecting S-nitrosothiol decomposition in solution include light, temperature, pH, and contaminating transition metal ions (19,20,37). In the absence of light and transition metal ion contamination, S-nitrosothiols are stable at physiological pH and temperature. In most reported measurements of S-nitrosothiol decomposition, the agent responsible for the breakdown of these compounds is the variable metal ion content of the buffer (21). This is exemplified by the halftime of SNAP, which has been variously measured as 1.15 h (30) and 4.6 h (1). The amount of free redox-active metal ion in each experimental system can vary between different batches and preparations of chemicals and buffers. Tissue preparations may also contribute to the total metal ion content.
The decomposition of S-nitrosothiols by photolysis is well understood. UV-visible light causes a homolytic cleavage of the sulfur-nitrogen bond, resulting in the release of ⅐ NO and a thiyl radical (19). It is often assumed that the decomposition of these compounds in the dark also occurs through homolysis, and a thiyl radical intermediate (21, 30, 36 -40). We show here that this is not the case. We observed no evidence of thiyl radical formation during decomposition of GSNO either by contaminating metal ions or by exogenously added Cu 2ϩ ions. Moreover DMPO, which is an efficient thiyl radical trap (31,32), did not affect the yield of GSSG during the dark decomposition of GSNO but inhibited GSSG formation during photolytic decomposition. The precise mechanism of transition metal ion-induced decomposition of S-nitrosothiols is unknown, and it is likely that the chelation of metal ions by thiol and thiol disulfides may complicate the kinetics of this process (39 -41). The observation that Cu ϩ is significantly more active than Cu 2ϩ implies that reducing agents such as glutathione and ascorbate will accelerate the transition metal ion-dependent decomposition of S-nitrosothiols by chemical reduction of contaminating transition metal ions. Additional studies are required to fully understand the transition metal ion-dependent decomposition chemistry of S-nitrosothiols.
Boese et al. (42) have shown that the S-transnitrosation reaction between serum albumin and dinitrosyl-iron complex exhibits a direct characteristic ESR signal. It is conceivable that a combination of both the photolytic spin-trapping methodology presented here and direct ESR of dinitrosyl-iron complexes can be used to assess the interaction between ⅐ NO and proteins.
Biological Relevance of Transnitrosation Reactions-Another important reaction of S-nitrosothiols, which has been implicated in the biological activity of GSNO, is transnitrosation (27). This reaction consists of NO ϩ transfer from an Snitrosothiol to a thiol. This may become biologically significant if the thiol is a protein cysteinyl residue and such modification leads to altered enzyme or receptor activity. In the case of the N-methyl-D-aspartate receptor, transnitrosation to a protein cysteinyl residue inhibits receptor activity and thus inhibits the excitotoxic response (43). This may be the only case where transnitrosation to protein is protective, as cysteinyl modification is more likely to result in cellular ion imbalance, via inhibition of plasma membrane channels (44), or enzyme inhibition. It has been suggested that glyceraldehyde-3-phosphate dehydrogenase, protein kinase C, and glutathione peroxidase can all be inhibited by transnitrosation (10,11,15). Previous measurements of the kinetics and thermodynamics of transnitrosation have resorted to high concentrations of thiol and S-nitrosothiol in order to achieve sufficient optical absorbance (24,34,35). This procedure is subject to error, as the reductive reaction between thiol and nitrosothiol is favored at such high concentrations (4,23,35). This reaction leads to the formation of nitrous oxide (from a NO Ϫ intermediate) and disulfide. Fortuitously, SNAP exhibits an additional absorbance shoulder at 234 nm, which can be used to monitor transnitrosation at lower concentrations of compounds. The equilibrium constants for transnitrosation between SNAP, GSH, and CySH favor GSNO formation. Because of the vast excess of GSH over every other thiol in the cell cytoplasm, it is likely that the transfer of the NO ϩ will be effectively unidirectional from S-nitrosothiols to glutathione. Thus GSH may "repair" S-nitrosylated proteins by a transnitrosation mechanism. It remains to be determined if this process is augmented by enzyme catalysis. Transnitrosation represents a possible pathway for the nitrosonium functional group to cross biological membranes and enter cells.
If, as is likely, GSH represents an intracellular thermodynamic sink for the transnitrosation processes, the fate of GSNO becomes an important consideration. The mechanism of GSNO decomposition in biological systems is not known, nor are the products of decomposition well characterized. In chemical systems, GSNO slowly decays in the presence of thiol to NO Ϫ and consequently to either peroxynitrite or nitrous oxide (depending upon oxygen availability) (4). If GSNO remains inside the cell in the presence of 5-10 mM GSH, the reaction between GSNO and GSH represents the most likely mode of decomposition. However, if, as with other S-substituted glutathione derivatives, GSNO is actively expelled from the cell its fate becomes dependent upon the extracellular milieu where metal ion-dependent breakdown and transnitrosation reactions are possible. It has been suggested that the enzyme ␥-glutamyl transpeptidase is responsible for GSNO decomposition (45). This enzyme enhances the decomposition of GSNO forming S-nitrosocysteinylglycine and glutamate. 4 ⅐ NO production is dramatically enhanced upon the addition of ␥-glutamyl transpeptidase to GSNO, as S-nitrosocysteinylglycine is more susceptible to transition metal ion-dependent decomposition than GSNO. However, DTPA completely inhibited the ⅐ NO release by preventing the decomposition S-nitrosocysteinylglycine.
Conclusions-In this study we conclude the following. (i) Biologically relevant S-nitrosothiols are stable in the dark in the presence of transition metal ion chelators. (ii) Photolytic decomposition of S-nitrosothiols generates ⅐ NO and the corresponding thiyl radical. (iii) Metal-ion catalyzed (dark) decomposition of S-nitrosothiols generates ⅐ NO and disulfides without the intermediacy of thiyl radicals. (iv) Transnitrosation from a stable thiol to an unstable thiol may facilitate ⅐ NO release in the presence of contaminating metal ions. (v) Reducing agents, such as ascorbate and thiols, may additionally accelerate metal ion-dependent decomposition as the reduced forms of the contaminating metal ions (e.g. Cu ϩ ) are more active at promoting the decomposition of S-nitrosothiols than their oxidized counterparts. (vi) In the absence of contaminating metal ions, thiols enhance S-nitrosothiol decomposition to yield NO Ϫ and not ⅐ NO.