Search

S-glutathionylation, the post-translational modification forming mixed disulfides between protein reactive thiols and glutathione, regulates redox-based signaling events in the cell and serves as a protective mechanism against oxidative damage. S-glutathionylation alters protein function, interactions, and localization across physiological processes, and its aberrant function is implicated in various human diseases. In this review, we discuss the current understanding of the molecular mechanisms of S-glutathionylation and describe the changing levels of expression of S-glutathionylation in the context of aging, cancer, cardiovascular, and liver diseases.

Oxidative stress is known to contribute to the progression of apoptosis. Staurosporine is a broad-spectrum inducer of apoptosis, but its mechanism of action is not well understood. The goal of the present work was to elucidate the role of glutathione and reactive oxygen species (ROS) in the execution of staurosporine-induced apoptosis. HeLa cells were treated with staurosporine at 1 μM for up to 4 h. The concentration of glutathione, generation of ROS, and activation of caspase-3 were measured. The introduction of staurosporine significantly decreased the concentration of cellular glutathione and increased the presence of ROS after 3 h. These findings were concurrent with the activation of caspase-3. Interestingly, pre-treatment of cells with N-acetylcysteine, a precursor of glutathione, and a thiol antioxidant failed to block the depletion of glutathione, generation of ROS, and activation of caspase-3. Collectively, these results suggest that the cellular redox status may be one of the critical factors of the apoptotic pathway leading to caspase-3 activation by staurosporine.

Mycobacterium tuberculosis is the causative agent of many strains of tuberculosis, as it is composed of an impenetrable, complex cell wall. The proteins active in the synthesis of the cell wall are mycolyl transferase antigens 85A, 85B, and 85C, encoded by genes fbpA, fbpB, and fbpC. Antigen 85C contains one cysteine residue. S-Glutathionylation is the formation of a mixed disulfide between a protein cysteine residue and glutathione (GSH), an abundant antioxidant molecule. It is a post-translational modification of cysteine residues which can occur under oxidative stress or physiological conditions. It is a known mechanism to regulate enzyme activity, signaling pathways, and the progression of diseases. By S-glutathionylation, the lone cysteine residue in antigen 85C is modified by biotinylated GSH ethyl ester to form a mixed disulfide. This modification results in a decrease in enzyme activity by 90%, representing a decrease in ability of the protein to synthesize the bacterial cell wall. Both the modification and the enzymatic activity of the protein are concentration dependent and can be reversed upon addition of a thiol reducing agent. The results provide a potential strategy for inhibiting the synthesis of the cell wall of M. tuberculosis by promoting oxidation of the lone cysteine residue. To our knowledge, this is a novel finding to demonstrate the modification of antigen 85C and the regulation of its activity by a physiological molecule. © 2018 IUBMB Life, 70(11):1111-1114, 2018.

S-glutathionylation, the formation of mixed disulfides of glutathione with cysteine residues of proteins, is a broadly observed physiological modification that occurs in response to oxidative stress. Since cysteine residues are particularly susceptible to oxidative modification by reactive oxygen species, S-glutathionylation can protect proteins from irreversible oxidation. In this study, we show that the kinase activity of the non-receptor tyrosine kinase c-Abl is inhibited by in vitro thiol modification; specifically, the cysteine residues of c-Abl are modified by S-glutathionylation and by thiol alkylating agents such as 4-acetamido-4'-maleimidylstilbene-2,2'-disulfonic acid and N-ethylmaleimide. Modification of cysteine residues of c-Abl tyrosine kinase using glutathione disulfide and thiol alkylating agents corresponds to a concomitant loss of kinase activity. We also demonstrate that S-glutathionylation of c-Abl can be reversed using a physiological system involving glutaredoxin and this reversal restores c-Abl kinase activity. To our knowledge, these are the first data to show S-glutathionylation of c-Abl, and this modification may represent a mechanism of regulation of c-Abl kinase activity in cells under oxidative stress.

Oxidoreductases such as glutaredoxin are a major class of enzymes that reversibly catalyze thiol-disulfide exchange reactions. Transfection experiments using geneticin (G418) selection to identify the specific protein S-thiolated substrates of glutaredoxin-1 (Grx-1) noted the curious phenomenon that nontransfected control cells treated with G418 had increased levels of Grx-1 expression. Varied concentrations of gentamicin, kanamycin, and hygromycin increased Grx-1 expression in a time- and dose-dependent fashion in human cultured retinal pigment epithelial cells. Reactive oxygen species formation after aminoglycoside exposure correlated directly to aminoglycoside treatment. Further indication that oxidation regulates Grx-1 expression was noted by the positive effect of phorbol 12-myristate 13-acetate, a known inducer of redox-sensitive AP-1 transcription factor. In agreement with this hypothesis was the finding that the physiologic reductant N-acetylcysteine decreased Grx-1 expression whereas tert-butyl hydroperoxide increased Grx-1 expression. Our data suggest that aminoglycosides increased Grx-1 expression in response to oxidative stress.

Oxidized low-density lipoprotein (oxLDL), which accumulates in vascular lesions, alters vascular cell function in ways that can be construed as atherogenic. Among these is the observation that oxLDL and its lipids promote smooth muscle cell (SMC) proliferation. A number of schemes have been proposed to explain this phenomenon. Our published data support the concept that part of the proliferation is mediated by lysophosphatidylcholine (lysoPC) and structurally related phospholipids borne by oxLDL, which cause FGF-2 release via an oxidant-dependent mechanism. Since FGF-2 can bind extracellular matrices, we wanted to determine whether the FGF-2 released came from an intracellular or an extracellular matrix-bound pool. We tested whether lysoPC was capable of releasing FGF-2 from SMC matrices, whether agents that release FGF-2 from matrices could cause proliferation, and whether lysoPC-mediated proliferation could occur by stimulating metalloproteinase (MMP)-induced matrix degradation, which released matrix-bound FGF-2. Our results indicate that the source of FGF-2 released by lysoPC and related lipids is a preexisting cellular pool and not from matrix, and that the mechanism likely involves transient, sublethal cell permeabilization. These results enhance understanding of a mechanism by which oxLDL could contribute to SMC proliferation in arterial lesions.

We previously reported that oxidized low-density lipoprotein and one of its constituents, lysophosphatidylcholine (lysoPC), caused smooth muscle cell proliferation that was inhibitable by vitamin E and by a neutralizing antibody against basic fibroblast growth factor-2 (FGF-2). We now show that the mitogenic activity of lysolipids is highly dependent on structure. Phospholipids with palmitoyl fatty acid and phosphocholine induced DNA synthesis optimally. Shorter and longer fatty acids were significantly less potent, as were phosphoserine and phosphoethanolamine head groups. Structurally related phospholipids [platelet-activating factor (PAF) and lysoPAF] were also mitogens and acted via an analogous FGF-2-dependent, vitamin E-inhibitable mechanism. The mechanism of lysoPC stimulation was distinct from that of another phospholipid mitogen, lysophosphatidic acid (lysoPA), in that lysoPC stimulation was not pertussis toxin inhibitable. Furthermore, lysoPA stimulation was not inhibitable by vitamin E. Despite its distinct cellular pathway for stimulation, lysoPA also ultimately led to FGF-2 release. Our data show that specific structural attributes of lysoPC, PAF, and lysoPAF enable these agents to mediate smooth muscle cell release of FGF-2, which in turn stimulates proliferation.

The first reports of the influences of oxidized LDL (oxLDL) on cell function pertained to negative effects on cell growth—growth arrest, injury, and toxicity. Since these studies, it has become apparent that sublethal levels of oxLDL cause some, but not all, cells to proliferate. This review highlights the growth-promoting effects of oxLDL rather than its inhibitory or injurious effects. Smooth muscle cells (SMCs) and monocyte-macrophages proliferate after exposure to oxLDL; endothelial cells do not. Scavenger receptors are involved in the proliferative effects on monocyte-macrophages, whereas the effects of oxLDL on SMCs appear to be receptor independent. Lysophosphatidylcholine (lysoPC), and structurally related lipids are among the growth-promoting constituents of oxLDL. OxLDL exerts at least a part of its effects by inducing expression or causing the release of growth factors. OxLDL (or lysoPC) can cause the release of basic fibroblast growth factor (bFGF) from SMCs; oxLDL (or lysoPC) can induce heparin binding EGF-like growth factor (HB-EGF) synthesis and release from macrophages. An imposing array of changes in cytokine and growth factor expression and/or release can be imposed by oxLDL on a wide variety of cell types. These effects and the studies probing the cell signaling events leading to them are described.

Increased nitric oxide (NO) production by inducible nitric oxide synthase (iNOS) has been associated with intestinal inflammation, including human inflammatory bowel disease. However, NO can downregulate endothelial activation and leukocyte adhesion, critical steps in the inflammatory response. Using primary cultures of human intestinal microvascular endothelial cells (HIMEC), we determined the role of NO in the regulation of HIMEC activation and interaction with leukocytes. Both nonselective ( NG-monomethyl-l-arginine) and specific ( N-iminoethyl-l-lysine) competitive inhibitors of iNOS significantly increased binding of leukocytes by HIMEC activated with cytokines and lipopolysaccharide. Increased adhesion was reversible with the NOS substratel-arginine and was not observed in human umbilical vein endothelial cells (HUVEC). Activation of HIMEC significantly upregulated HIMEC iNOS expression and NO production. NOS inhibitors did not augment cell adhesion molecule levels in activated HIMEC but did result in sustained increases in intracellular reactive oxygen species. In addition, antioxidant compounds reversed the effect of NOS inhibitors on HIMEC-leukocyte interaction. Taken together, these data suggest that after HIMEC activation, iNOS-derived NO is an endogenous antioxidant, downregulating leukocyte binding and potentially downregulating intestinal inflammation.

We have previously shown that oxidized low density lipoprotein (LDL) but not native LDL stimulated DNA synthesis in cultured smooth muscle cells (SMC) and that α-tocopherol (vitamin E) inhibited this proliferative response (Lafont, A., Chai, Y. C., Cornhill, J. F., Whitlow, P. L., Howe, P. H., and Chisolm, G. M. (1995) J. Clin. Invest. 95, 1018-1025). The moiety of oxidized LDL that stimulates DNA synthesis and the cellular mechanism for this potentially mitogenic effect are not known. We now report that lipid fractions containing lysophospholipids from oxidized LDL or phospholipase A2-treated native LDL stimulated SMC DNA synthesis as did palmitoyl lysophosphatidylcholine (lysoPC). Protein kinase C inhibitors and down-regulation of protein kinase C activity by phorbol ester inhibited oxidized LDL- and lysoPC-induced DNA synthesis. A neutralizing monoclonal antibody against fibroblast growth factor-2 significantly inhibited oxidized LDL and lysoPC-induced DNA synthesis in SMC; irrelevant antibodies were ineffective. Vitamin E inhibited the DNA synthesis stimulated by lysoPC, an observation that distinguished this effect from DNA synthesis induced by another detergent, digitonin. These results suggest that oxidized LDL and its lysoPC moiety stimulate SMC to enter the cell cycle via an oxidative mechanism that causes the release of fibroblast growth factor-2 and a subsequent autocrine or paracrine response.

S-thiolation of carbonic anhydrase III (CA III) in cultured rat hepatocytes under oxidative stress was studied by immunodetection on nitrocellulose blots of isoelectrofocusing gels. In cells treated with menadione, three S-thiolated forms of CA III were detected, whereas only two forms were observed in hepatocytes treated with t-butyl hydroperoxide. Two "nonreducible" oxidized forms of CA III were also detected on nitrocellulose blots. These forms increased with the amount of stress and were the only modified forms of CA III in buthionine sulfoxide-treated hepatocytes containing 10-fold less glutathione than control hepatocytes. These experiments support the concept that S-thiolation protects CA III from irreversible oxidation during oxidative stress. Partly and fully S-thiolated forms of CA III were easily detected in both male and female hepatocytes by the immunoblotting method, although female cells contained 15-fold less CA III than did male liver. S-thiolated forms of CA III were also detected in rat skeletal muscle and heart showing the utility of this method for determining the effect of oxidative stress on specific S-thiolatable protein in several tissues in vivo.

Cellular molecules possess various mechanisms in responding to oxidant stress. In terms of protein responses, protein S-glutathionylation is a unique post-translational modification of protein reactive cysteines forming disulfides with glutathione molecules. This modification has been proposed to play roles in antioxidant, regulatory and signaling in cells under oxidant stress. Recently, the increased level of protein S-glutathionylation has been linked with the development of diseases. In this report, specific S-glutathionylated proteins were demonstrated in human embryonic kidney 293 cells treated with two different oxidative reagents: diamide and hydrogen peroxide. Diamide is a chemical oxidizing agent whereas hydrogen peroxide is a physiological oxidant. Under the experimental conditions, these two oxidants decreased glutathione concentration without toxicity. S-glutathionylated proteins were detected by immunoblotting and glutathione concentrations were determined by high performance liquid chromatography. We further show the effect of alteration of the cellular thiol pool on the amount of protein S-glutathionylation in oxidant-treated cells. Cellular thiol concentrations were altered either by a specific way using buthionine sulfoximine, a specific inhibitor of glutathione biosynthesis or by a non-specific way, incubating cells in cystine-methionine deficient media. Cells only treated with either buthionine sulfoximine or cystine-methionine deficient media did not induce protein S-glutathionylation, even though both conditions decreased 65% of cellular glutathione. Moreover, the amount of protein S-glutathionylation under both conditions in the presence of oxidants was not altered when compared to the amount observed in regular media with oxidants present. Protein S-glutathionylation is a dynamic reaction which depends on the rate of adding and removing glutathione. Phenylarsine oxide, which specifically forms a covalent adduct with vicinal thiols, was used to determine the possible role of vicinal thiols in the amount of glutathionylation. Our data shows phenylarsine oxide did not change glutathione concentrations, but it did enhance the amount of glutathionylation in oxidant-treated cells.

A disulfide bond between key redox-sensitive cysteine residues and glutathione is one mechanism by which redox related allosteric effectors can regulate protein structure and function. Here we test the hypothesis that glutaredoxin-1 (Grx-1), a member of the oxidoreductase family of enzymes, may be a critical component of redox-sensitive molecular switches by mediating reversible protein S-glutathionylation and enzymatic catalysis of thiol/disulfide exchange. Deglutathionylation of a 70 kDa protein by Grx-1 was detected using a monoclonal antibody specific to protein S-glutathionylation. Heat shock cognate protein 70 (Hsc70) was identified as a substrate of Grx-1 through mass spectrometry. Recombinant Hsc70 was glutathionylated in vitro, and protein S-glutathionylation reversed by Grx-1. Glutathionylated Hsc70 was more effective in preventing luciferase aggregation at 43 degrees C than reduced Hsc70 in a dose dependent fashion. ATP did not effect the chaperone activity of Hsc70-SG but did increase the activity of reduced Hsc70-SG. Reversible glutathionylation of Hsc70 may provide a mechanism for post-translation regulation of chaperone activity.

Mitochondria play central roles in cellular metabolism and apoptosis and are a major source of reactive oxygen species (ROS). We investigated the role of ROS and mitochondria in radiation-induced apoptosis in multiple myeloma cells. Two distinct levels of ROS were generated following irradiation: a small increase observed early, and a pronounced late increase, associated with depletion of reduced glutathione (GSH) and collapse of mitochondrial membrane potential (deltapsi(m)). Exogenous ROS and caspase-3 induced deltapsi(m) drop and cytochrome c release from mitochondria, which could be prevented by molecular (dominant-negative caspase-9) and pharmacologic (zVAD-fmk) caspase inhibitors and overexpression of Bcl-2. Exogenous ROS also induced mitochondrial permeability transition (PT) pore opening and cytochrome c release in isolated mitochondria, which could be blocked by inhibition of PT with cyclosporin A. These results indicate that the late ROS production is associated with increased PT pore opening and decreased deltapsi(m), and GSH, events associated with caspase activation and cytochrome c release.

Protein cysteines can serve both sensory and activation roles in the regulation of protein function. The modulation of mixed disulfides with glutathione may promise to be a broad mechanism of redox signalling. Using both protein extract and intact RPE cells, we have generated covalent adduction of glutathione to protein cysteines and further show that glutaredoxin (Grx-1) is able to remove glutathione from protein S-glutathiolated substrates. Our data demonstrate that glutathione can modify a wide range of RPE proteins in intact cells, but that the reversal of this process–deglutathiolation and thiol bond restoration–may require a specific catalytic reaction with glutaredoxin. More generally, our experiments support the hypothesis that glutathione can non-specifically become adducted to protein cysteines during oxidative stress, but that the specific, functional reconstitution of protein thiols depends on recognition by an oxidoreductase such as glutaredoxin. This concept offers the idea that redox signalling involves both adduction of a non-specific non-protein reducing equivalent such as glutathione and specific protein based removal by glutaredoxin.

Oxidized low-density lipoprotein (oxLDL), which accumulates in vascular lesions, alters vascular cell function in ways that can be construed as atherogenic. Among these is the observation that oxLDL and its lipids promote smooth muscle cell (SMC) proliferation. A number of schemes have been proposed to explain this phenomenon. Our published data support the concept that part of the proliferation is mediated by lysophosphatidylcholine (lysoPC) and structurally related phospholipids borne by oxLDL, which cause FGF-2 release via an oxidant-dependent mechanism. Since FGF-2 can bind extracellular matrices, we wanted to determine whether the FGF-2 released came from an intracellular or an extracellular matrix-bound pool. We tested whether lysoPC was capable of releasing FGF-2 from SMC matrices, whether agents that release FGF-2 from matrices could cause proliferation, and whether lysoPC-mediated proliferation could occur by stimulating metalloproteinase (MMP)-induced matrix degradation, which released matrix-bound FGF-2. Our results indicate that the source of FGF-2 released by lysoPC and related lipids is a preexisting cellular pool and not from matrix, and that the mechanism likely involves transient, sublethal cell permeabilization. These results enhance understanding of a mechanism by which oxLDL could contribute to SMC proliferation in arterial lesions.

Studies of restenosis after angioplasty, and more generally studies of arterial responses to injury, have revealed that the variety of potential outcomes includes proliferation of vascular smooth muscle cells, what appears to be a chronic constrictive response, or vascular remodeling. These are not mutually exclusive of one another and they are detailed more fully in other chapters of this book. They are mentioned here to point out that a mechanical arterial injury such as is inflicted by angioplasty, the inflammatory response to that injury and the stimuli for remodeling emanating from flow and pressure alterations caused by the injury, may compete in mediating changes in vascular tone and in the cell migration, proliferation and death that govern the healing process. Thus, basic mechanisms controlling cell death, mitogenesis or vascular relaxation that are mediated by factors already present in atherosclerotic lesions, factors induced by vascular injury or the injury response, and factors elicited by changes in flow or pressure are important to understand. Our purpose in the following pages is to describe a subcategory of the influences on cell growth and vascular reactivity that have been identified in vitro, that is, those resulting from the actions of lipid and lipoprotein oxidation products known to reside in atherosclerotic plaques or enter the angioplasty injury site.

The ability of alpha-tocopherol to reduce restenosis after angioplasty was tested in a rabbit model in which angioplasty was performed on established atherosclerotic lesions. Lesions induced by 4 wk of cholesterol feeding after focal desiccation of femoral arteries were balloon dilated. 3 wk after angioplasty, angiographically determined minimum luminal diameters were less in the untreated group (0.80 +/- 0.51 mm) than in the group treated with oral alpha-tocopherol beginning 19 d before angioplasty (1.38 +/- 0.29 mm; P < 0.01). The cross-sectional area of the intima-media was greater in the untreated group (1.18 +/- 0.48 mm2) than in the alpha-tocopherol group (0.62 +/- 0.25 mm2, P < 0.0001). These differences were not due to vasoconstriction or altered plasma cholesterol. Alpha-tocopherol thus reduced restenosis after angioplasty in this model. In rabbit vascular smooth muscle cells, oxidized low density lipoprotein stimulated DNA synthesis. Alpha-tocopherol treatment inhibited DNA synthesis stimulated by oxidized low density lipoprotein, but not by serum. The findings are consistent with the hypothesis that oxidized lipids can stimulate hyperplasia and that antioxidants may limit hyperplasia by inhibiting either the oxidation or the proliferative effects of oxidants on cells.

In order to examine potentially important S-thiolated proteins, 35 S-labeled hepatocytes were exposed to oxidative stress. A similar group of S-thiolated proteins including carbonic anhydrase III was observed in cells treated with t -butyl hydroperoxide, menadione, or stimulated neutrophils. The radioactive thiols bound to hepatocyte proteins were identified by HPLC and more than 85% was glutathione. In menadione-treated hepatocytes, proteins were gradually S-thiolated over 30 min and 25% of the cellular glutathione pool became protein-bound. In t -butyl hydroperoxide-treated cells, S-thiolation was more transient and 11% of the glutathione was protein-bound. Neutrophil-treated hepatocytes had nearly the same amount of protein S-thiolation (8% after 25 min). Two major proteins that were S-thiolated in untreated hepatocytes did not increase during any form of oxidative stress. In neutrophil-treated hepatocytes protein S-thiolation was not accompanied by either formation of glutathione disulfide or a measureable change in the total amount of glutathione. In both t -butyl hydroperoxide- and menadione-treated cells there was extensive formation of glutathione disulfide and in menadione-treated cells a significant increase in the total hepatocyte glutathione pool was observed. This result suggests that protein S-thiolation may occur by mechanisms that do not result from thiol/disulfide exchange between glutathione disulfide and protein sulfhydryls. It is suggested that a thiyl radical intermediate is important in neutrophil-mediated protein S-thiolation.

Protein S-thiolation, a reversible modification of protein sulfhydryls resulting in formation of mixed-disulfides, was studied in human neutrophils stimulated with phorbol diester to produce superoxide anion. Rapid S-thiolation of several proteins was examined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Glutathione was identified as the primary protein-bound thio by HPLC chromatography, contributing considerably more than 85% of the total. Minor amounts of homocysteine and/or cysteine were also detected as protein-bound thiols. During the first 30 min after stimulation, 10% of the cellular glutathione became protein bound (2 nmol/mg of protein). There was no increase in glutathione disulfide suggesting that S-thiolation of the proteins did not occur by thiol/disulfide exchange. Approximately 10 mol% of one heavily modified band (29 kDa) was S-thiolated after 30 min. A second major band of 42 kDa was identified as actin. It contained 1/10th of the total protein-bound glutathione and approximately 5 mol% was S-thiolated after 30 min. These experiments identify a subset of S-thiolated neutrophil proteins, including actin, whose modification is related to the phorbol diester stimulation of superoxide anion production in human neutrophils. Ten percent of the total glutathione pool became protein-bound without an appreciable change in non-bound concentration of glutathione or glutathione disulfide. These results suggest that glutathione was synthesized during initial phases of the respiratory burst, compensating for the amount of glutathione that became protein-bound. Since there was no significant increase in glutathione disulfide, it was probably not important in the observed protein S-thiolation.

Publisher Summary S-Thiolated proteins—mixed disulfides of proteins and low molecular weight thiols—are very early products of protein oxidation during oxidative stress, occurring within seconds after the generation of oxygen radicals. As a result, assessment of the extent and specificity of this process during oxidative stress is one of the best measures of the primary effects of oxygen radical generation on intact cells. The development of methods for measuring the S-thiolation status of individual proteins, eventually studying organs of intact animals and even humans, is essential for a more complete understanding of the role of oxidative stress in human disease. This chapter introduces a method to quantify the S-thiolation status of mixed cellular proteins in intact cells and second method to determine the activity of the cellular reductive processes—namely, dethiolases. This method tests the effects of toxic materials on protein S-thiolation. Protein dethiolation can be measured with S-thiolated substrate proteins containing a radioactive low molecular weight thiol.

Many cellular proteins have highly reactive sulfhydryls that are especially prone to modification during oxidative stress. Proteins with at least one “reactive” sulfhydryl are prevelant in cytoplasm, membranes, and nuclei, and the concentration of these reactive sites is impressive, reaching or exceeding the concentration of glutathione in many cells. Most proteins contain a single reactive sulfhydryl and are unlikely to form protein-protein disulfides under oxidizing conditions. Therefore, the most likely modification of these single reactive sulfhydryls by oxidative stress is formation of a mixed-disulfide with a low molecular weight cellular thiol such as glutathione, a modification we have termed protein S-thiolation (Ziegler, D.M., 1985 and Grimm, et.al., 1985). (If the modification is known to involve glutathione, the term S-glutathiolation is more descriptive.) This modification is metabolically labile and the rapid “dethiolation” of these proteins by several reductive processes (Park and Thomas, 1989) prevents lasting damage to the protein. Thus, the dynamic modification of proteins by S-thiolation/dethiolation processes represents one of the more important cellular mechanisms for protection against oxidative stress.

An S-thiolated 30-kDa protein has been purified from rat liver by two steps of ion-exchange chromatography. This monomeric protein has two “reactive” sulfhydryls that can be S-thiolated by glutathione (form a mixed disulfide with glutathione) in intact liver. The protein has been identified as carbonic anhydrase III by sequence analysis of tryptic peptides from the pure protein. The two “reactive” sulfhydryls on this protein can produce three different S-thiolated forms of the protein that can be separated by isoelectric focusing. Using this technique it was possible to study the S-thiolation and dethiolation reactions of the pure protein. The reduced form of this protein was S-thiolated both by thiol-disulfide exchange with glutathione disulfide and by oxyradical-initiated S-thiolation with reduced glutathione. The S-thiolation rate of this 30-kDa protein was somewhat slower than that of glycogen phosphorylase b by both S-thiolation mechanisms. The S-thiolated form of this protein was poorly dethiolated (i.e., reduced) by glutathione, cysteine, cysteamine, or coenzyme A alone. Enzymatic catalysis by two different enzymes (glutaredoxin and thioredoxinlike) greatly enhanced the dethiolation rate. These experiments suggest that carbonic anhydrase III is a major participant in the liver response to oxidative stress, and that the protein may be S-thiolated by two different nonenzymatic mechanisms and dethiolated by enzymatic reactions in intact cells. Thus, the S-thiolation/dethiolation of carbonic anhydrase III resembles glycogen phosphorylase and not creatine kinase.