A novel reaction mechanism for the formation of S-nitrosothiol in vivo.

The objective of this study was to investigate the mechanism of S-nitrosothiol formation under physiological conditions. A mechanism is proposed by which nitric oxide (·NO) reacts directly with reduced thiol to produce a radical intermediate, R-S-N·-O-H. This intermediate reduces an electron acceptor to produce S-nitrosothiol. Under aerobic conditions O2 acts as the electron acceptor and is reduced to produce superoxide (O2). The following experimental evidence is provided in support of this mechanism. Cysteine accelerates the consumption of ·NO by 2.5-fold under physiological conditions. The consumption of O2 in the presence of ·NO and cysteine is increased by 2.4-fold. The reaction orders of ·NO and cysteine are second and first order, respectively. The second order of reaction for ·NO may result from interaction between ·NO and O2 to form peroxynitrite. In the presence of Cu,Zn-superoxide dismutase, the reaction of ·NO with cysteine generates hydrogen peroxide, indicating that the reaction generates O2. Finally, the formation of S-nitrosothiol is demonstrated in an anaerobic environment and, as predicted by the mechanism, is dependent on the presence of an electron acceptor. These results demonstrate that under physiological conditions ·NO reacts directly with thiols to form S-nitrosothiol in the presence of an electron acceptor.

strated in an anaerobic environment and, as predicted by the mechanism, is dependent on the presence of an electron acceptor. These results demonstrate that under physiological conditions ⅐ NO reacts directly with thiols to form S-nitrosothiol in the presence of an electron acceptor.
S-Nitrosothiols are important physiological regulators capable of producing vasodilation and inhibition of platelet aggregation (1)(2)(3)(4). An increasing number of proteins such as albumin, glyceraldehyde-3-phosphate dehydrogenase, hemoglobin, and p21 ras have been found to be S-nitrosylated in vivo (5)(6)(7)(8)(9). With the discovery of new S-nitrosylated proteins, it becomes evident that the formation of S-nitrosothiols may be important in such diverse processes as signal transduction, DNA repair, and blood-pressure regulation. However, at present the mechanism of the biosynthetic pathway for the formation of Snitrosothiols is unclear (10).
It has been shown previously that the reaction of ⅐ NO 1 with sulfhydryl groups under anaerobic conditions at neutral pH does not produce S-nitrosothiol (11)(12)(13). This has led to the conclusion that S-nitrosothiols are formed by the autoxidation of ⅐ NO, a second order reaction with respect to ⅐ NO, to higher oxides of nitrogen (NO x ), by metal catalysis (12,14,15), or by the action of dinitrosyl-iron complexes (16). However, the reaction of ⅐ NO and oxygen is slow, approximately 3 to 300 pmol/s, at physiological concentrations of ⅐ NO (0.1-1.0 M), and the availability of redox metal is unclear (17,18). Although dinitrosyl-iron complexes represent one possible mechanism for the formation of S-nitrosothiols, the reaction mechanism under physiological conditions also remains unclear (10).
Here we propose a novel mechanism for S-nitrosothiol formation that would operate at physiological concentrations of ⅐ NO. In this mechanism ⅐ NO reacts directly with a reduced thiol to produce a radical intermediate, In the presence of an electron acceptor, such as oxygen, this intermediate can be converted to S-nitrosothiol by the reduction of the acceptor.
Therefore, the reaction of ⅐ NO and cysteine in buffer under aerobic conditions will form superoxide via the reduction of O 2 by R-S-N ⅐ -O-H (Equation 2). Superoxide reacts at a nearly diffusion-limited rate with ⅐ NO to form peroxynitrite (19) (Equation 3). The overall reaction mechanism is shown below (Equation 4).
From this reaction mechanism the following testable predictions can be made: first, that free thiol will accelerate the decomposition of ⅐ NO and will result in the formation of a ⅐ NO donor; second, that thiol will accelerate the consumption of O 2 by ⅐ NO and generate H 2 O 2 in the presence of Cu,Zn-superoxide dismutase; and third, that the reaction will proceed under anaerobic conditions in the presence of an electron acceptor. Here experimental evidence is provided for each of the above predictions in support of the proposed mechanism.

EXPERIMENTAL PROCEDURES
Materials-Cu,Zn-superoxide dismutase was obtained from Fluka (Switzerland), and DEANO was obtained from Cayman Chemical (Ann Arbor, MI). All other chemicals were obtained from Sigma. Spectrophotometric measurements were made using a UV diode array spectrophotometer (Hewlett-Packard), and ⅐ NO in solution was measured using a * This work was supported by National Institutes of Health Service Award HL07748 (to A. J. G.) and Grant EY09269 (to D. G. B.), a grant from the Center for Neurogenerative Diseases, University of Pennsylvania, and by an Established Investigator Award from the American Heart Association (to H. I.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
⅐ NO-specific electrode (World Precision Instruments, Sarasota, FL). Synthesis and Measurement of ⅐ NO-Nitric oxide was synthesized by bubbling nitrogen through KNO 2 that was acidified with HCl. Nitric oxide and other nitrogen oxides formed in the reaction flask were forced through a gas-washing column containing 1 M NaOH. Nitric oxide was then collected in a nitrogen-purged, sealed vessel containing double distilled H 2 O that was passed through a Chelex-100 column to remove metal contaminants.
Nitric oxide concentrations were continuously monitored by means of a specific electrode in a 1-ml solution of PBS, pH 7.4, 100 M DTPA at a constant stirring velocity and at constant temperature (22°C) in a 46-well plate. Electrode output was recorded on a chart recorder, and initial rates of decay were measured as tangents within 10 s of reaching peak ⅐ NO concentration. Cysteine was added from a freshly prepared 10 mM stock solution upon reaching peak ⅐ NO concentration. In order to measure release of ⅐ NO from reaction products, ⅐ NO was added to PBS, 100 M DTPA containing cysteine. The reaction was allowed to proceed for 1.5 min, and then the solution was vortexed vigorously in three 10-s bursts to remove any free ⅐ NO remaining from the reaction. ⅐ NO release was then monitored as described above.
Measurement of O 2 and H 2 O 2 in Solution-Hydrogen peroxide and oxygen measurements were made by adding authentic ⅐ NO to PBS, 100 M DTPA, 2,400 units of Cu,Zn superoxide dismutase in the presence or absence of cysteine. All measurements were made in an open system at a constant stirring speed and temperature (22°C). Oxygen was continuously monitored by electrochemical reduction using a recessed microelectrode with a gold cathode polarized at Ϫ0.65 V relative to an Ag/AgCl reference electrode (20). The current sensitivity of the microelectrode was 1.2 pA/M. Hydrogen peroxide was measured by electrochemical oxidation using a 25-m diameter platinum wire sealed in glass that was polarized at 0.65 V relative to an Ag/AgCl reference electrode (21). The electrode was covered with a thin membrane by dip coating in a solution of 5% Nafion dissolved in aliphatic alcohols (Aldrich). The current sensitivity of the electrode was 0.32 nA/M. Hydrogen peroxide was also measured by the horseradish peroxidase-mediated oxidation of o-phenylenediamine; 25 g of horseradish peroxidase and 1 mM o-phenylenediamine were added to a 2-ml reaction mixture, and the mixture was vortexed and incubated at room temperature for 30 min. The mixture was acidified by adding 60 l of 5 N HCl, and the absorbance at 420 nm was measured (22).
Reaction of ⅐ NO and Cysteine in an Anaerobic Environment-2 ml of PBS, 100 M DTPA was placed in a 4-ml sealed vial. The solution was degassed by bubbling through N 2 for approximately 30 min. 750 M cysteine and/or 10 mM NAD ϩ were added to the solution prior to degassing. After degassing, 2 l of 37 mM DEANO, stored under N 2 , was injected into the vial via a gas-tight syringe with a Teflon seal. The solution was incubated at room temperature for 30 min. Under these conditions DEANO decomposes to release ⅐ NO at a rate of approximately 6 M/min (23). After incubation the vial was unstoppered and 1 ml of the solution was monitored for the release of ⅐ NO by electrode as described above. Another 1 ml was placed immediately in a quartz cuvette, and absorbance at 336 nm was measured.

RESULTS AND DISCUSSION
Cysteine Accelerates the Decomposition of ⅐ NO in Solution-In order to test the prediction that thiol will accelerate the decomposition of ⅐ NO, we examined the loss of ⅐ NO from an open aerobic system in the presence and absence of cysteine. Fig. 1 shows that cysteine accelerates the decomposition of ⅐ NO from buffer under aerobic conditions. In the absence of cysteine the initial rate of ⅐ NO loss was approximately 1.6 nM/s, whereas in the presence of cysteine the rate of ⅐ NO loss was approximately 3.8 nM/s. These experiments were performed at a constant stirring speed and temperature (22°C) and in the presence of a metal chelator. We have previously found that these factors are critical in determining the rate of loss of ⅐ NO in an open system. 2 The acceleration of ⅐ NO decomposition is dependent on the relative concentrations of ⅐ NO and cysteine. It was necessary for the thiol concentration to exceed the ⅐ NO concentration by approximately 100-fold. In addition, the effect of thiol on ⅐ NO decay was only relevant at ⅐ NO concentrations of less than 50 M. At higher concentrations of ⅐ NO the addition of thiol did not alter the decomposition of ⅐ NO, as has been reported previously (12). The reaction of ⅐ NO with O 2 , which is second order with respect to ⅐ NO, becomes increasingly relevant at higher concentrations of ⅐ NO (17). Therefore, at higher concentrations of ⅐ NO a NO x -based mechanism for S-nitrosothiol formation and other nitrosation reactions becomes increasingly relevant (12,14,24).
Orders of Reaction for Cysteine and ⅐ NO-The apparent orders of reaction for the loss of ⅐ NO were calculated from the initial rate of decay measured at varying concentrations of ⅐ NO and cysteine. Fig. 2 shows the plots of log(d[ ⅐ NO]/dt) versus log[substrate], where (d[ ⅐ NO]/dt) is expressed as the initial rate of decay in the presence of cysteine minus the initial rate in the absence of cysteine. As stirring velocity, ionic strength, and temperature are kept constant the difference between the initial rates in the presence and absence of cysteine represents the rate of reaction of ⅐ NO with cysteine. The slopes of the linear regression lines shown in Fig. 2 represent the apparent orders of reaction for ⅐ NO ( Fig. 2A) and cysteine (Fig. 2B). The reaction orders for ⅐ NO and cysteine approximate to second and first order, respectively. The apparent second order for ⅐ NO can be explained by the production of O 2 . upon decomposition of the radical intermediate. Superoxide will react with a second molecule of ⅐ NO in an almost instantaneous fashion. The Product of the Reaction Is S-Nitrosocysteine-Spectrophotometric examination of the reaction mixture containing cysteine and ⅐ NO reveals an absorbance peak at 336 nm, which is typical of S-nitrosothiol (5). The absorbance at 336 nm is dependent on the quantity of ⅐ NO added (Fig. 3A). However, the product yield expected from these experiments is very low, on the order of 1 M or less, because of the low concentrations of ⅐ NO in the reaction mixture. In addition, the molar absorptivity of S-nitrosocysteine is low, 3869 M Ϫ1 cm Ϫ1 . As a result the absorbances measured are at the limit of spectrophotometric detection and are hence quite variable. In order to confirm that the product of the reaction was S-nitrosocysteine, we assayed its ability to act as a ⅐ NO donor, a recognized property of S-nitrosocysteine.
Cysteine-based nitrosothiols release ⅐ NO in the presence of reduced thiol (25,26). In our experiments, cysteine was in approximately 100-fold excess and, therefore, any S-nitrosothiol formed will readily decompose to release free ⅐ NO. Fig. 3B shows that the reaction of ⅐ NO and cysteine results in the production of a rapidly decaying ⅐ NO donor. In this reaction ⅐ NO is added to PBS either in the presence or absence of cysteine. After 1.5 min the reaction mixture is vortexed to remove any residual ⅐ NO and then monitored for release of ⅐ NO by electrode. In the absence of cysteine no release is measured, whereas in the presence of cysteine a peak of ⅐ NO is observed. The presence of the characteristic S-nitrosothiol absorbance and the ability of the product to release ⅐ NO indicate that the reaction of physiological concentrations of ⅐ NO and cysteine forms S-nitrosocysteine.
Oxygen Consumption and Superoxide Production-In order to test the second prediction of the model, the consumption of O 2 by the decomposition of ⅐ NO was measured by means of an oxygen electrode. Fig. 4 shows that when ⅐ NO was added to buffer containing Cu,Zn superoxide dismutase alone, there was a small consumption of O 2 . However, when cysteine was present there was a 2.4-fold increase in O 2 consumption. This observation is confirmed by previous work that shows that thiol will increase O 2 consumption in the presence of a ⅐ NO donor (17). When varying concentrations of ⅐ NO were added to buffer in the presence and absence of cysteine, there was a direct correlation between ⅐ NO added and O 2 consumed (Fig. 4, inset). The slope of the regression line was 0.59 M O 2 /M ⅐ NO in the presence of cysteine. As Cu,Zn superoxide dismutase removes superoxide and, therefore, there is no possibility of superoxide and ⅐ NO interaction, a 1:1 stoichiometry between O 2 and ⅐ NO would be expected. However, Cu,Zn superoxide dismutase regenerates one O 2 molecule for every two superoxide molecules dismutated; therefore, these data are consistent with a 1:1 stoichiometry. The observation that the predicted stoichiometry holds true would seem to indicate that in the presence of superoxide dismutase all the ⅐ NO present reacts with thiol, producing S-nitrosothiol and superoxide anion.
To further confirm that superoxide was generated by the reaction of thiol and ⅐ NO, the production of H 2 O 2 in the presence of Cu,Zn superoxide dismutase was monitored by means of a specific electrode (21). The addition of approximately 6 M ⅐ NO had no effect on the current within the electrode in the absence of cysteine. However, in the presence of 750 M cysteine there was an increase in the current that corresponded to an increase in H 2 O 2 concentration of 4.9 Ϯ 1.9 M (n ϭ 5).
The formation of H 2 O 2 was confirmed independently by adding horseradish peroxidase and o-phenylenediamine 10 min after the addition of ⅐ NO. In the presence of cysteine, ⅐ NO, and superoxide dismutase this led to the generation of an absorbance characteristic of the oxidation of o-phenylenediamine, A 420 ϭ 0.155 Ϯ 0.029 (n ϭ 4). In the absence of either cysteine (A 420 ϭ 0.018) or superoxide dismutase (A 420 ϭ 0.020), oxidation of o-phenylenediamine was not observed. These results indicate that the reaction of ⅐ NO and cysteine in the presence of O 2 produces superoxide.
Formation of S-Nitrosocysteine under Anaerobic Conditions-In order to confirm the final prediction of the model, ⅐ NO was reacted with cysteine in an anaerobic environment in the presence of an electron acceptor, NAD ϩ . Fig. 5 shows that in  3. A, varying quantities of synthesized ⅐ NO were added to buffer as above. After peak electrode output was reached, 7.5 l of 10 mM cysteine was added. Two minutes after cysteine addition the solution was removed from the electrode and measured for absorbance at 336 nm in a UV-visible spectrophotometer. There is a significant correlation between ⅐ NO added and absorbance at 336 nm (p Ͻ 0.05). B, 2.5 l of 10 mM cysteine was added to 1 ml of PBS, 100 M DTPA, and 6 M ⅐ NO. The mixture was incubated for 1.5 min at 22°C and vortexed vigorously. The mixture was then monitored for ⅐ NO by electrode. Solid line represents no addition of cysteine, and dashed line represents addition of cysteine. the presence of 10 mM NAD ϩ the reaction of ⅐ NO and cysteine produced a compound that releases ⅐ NO at a steady rate. The decay of DEANO over the course of the anaerobic period will release approximately 18 M ⅐ NO. However, the vial contains both a gaseous and an aqueous phase, so a considerable portion (approximately 80%) of the ⅐ NO released by DEANO may be in the gaseous phase. The ⅐ NO measured in solution after the anaerobic period is most likely the residual dissolved material left from the decomposition of DEANO. It is possible that some of the DEANO remains intact and is, therefore, releasing ⅐ NO after the anaerobic period. However, the half-life of DEANO under these conditions is 2 min and thus it is unlikely that a significant quantity of intact DEANO remains after 30 min (23).
In the absence of either cysteine or NAD ϩ the rate of loss of ⅐ NO from the buffer immediately after the anaerobic incubation period was 6.7 and 6.0 nM/s, respectively. This rate probably represents the loss of dissolved ⅐ NO from the solution. In contrast, the rate of loss of ⅐ NO was only 3.1 nM/s in the presence of cysteine and NAD ϩ , indicating the presence of a ⅐ NO donor. The addition of excess reduced thiols confirmed the presence of a ⅐ NO donor, as it elicited an increase in the concentration of ⅐ NO within the reaction mixture only when both cysteine and NAD ϩ were present. Spectrophotometric analysis immediately after anaerobic incubation reveals that the product of the reaction possesses an absorbance peak of 0.05 unit at 336 nm, equivalent to 11 M S-nitrosothiol. Previous work has shown that the anaerobic interaction of glutathione and ⅐ NO in the absence of an electron acceptor proceeds through the same radical intermediate but that the products are disulfide and N 2 O (13). These results indicate that S-nitrosothiol can be formed from the anaerobic reaction of cysteine and ⅐ NO only in the presence of an electron acceptor.
Previous work has shown that it is possible to form S-nitrosothiol via mechanisms involving either ⅐ NO reaction with molecular O 2 or metal catalysis (12,14,15). However, at the concentrations of ⅐ NO used in this study, there is little or no production of higher oxides of nitrogen, and interaction with metals was avoided by using a metal chelator. Therefore, the formation of S-nitrosothiols under the conditions used in this study shows that S-nitrosothiols can be formed by a novel mechanism. An alternative to the mechanism proposed within this paper is that ⅐ NO reacts with oxygen to form a nitrosyldioxyl radical (Equation 5). The nitrosyldioxyl radical then reacts with a reduced thiol to produce S-nitrosothiol and superoxide (Equation 6).
This superoxide could then react with another molecule of ⅐ NO to produce peroxynitrite (Equation 3). This reaction mechanism would result in the same overall reaction equation described above (Equation 4). Therefore, this mechanism could explain all the data shown here except for the formation of S-nitrosothiol under anaerobic conditions, as oxygen is a requirement.
The reactions described here are significant because they operate at low concentrations of ⅐ NO and thus allow for the formation of S-nitrosothiols under physiological conditions. Our results and the already proven mechanisms of S-nitrosothiol formation indicate that S-nitrosothiols can be formed in vivo under a wide variety of pathophysiological conditions.