Regulation of Protein S-Thiolation by Glutaredoxin 5 in the Yeast Saccharomyces cerevisiae *

The irreversible oxidation of cysteine residues can be prevented by protein S-thiolation, a process by which protein -SH groups form mixed disulfides with low molecular weight thiols such as glutathione. We report here that this protein modification is not a simple response to the cellular redox state, since different oxidants lead to different patterns of proteinS-thiolation. SDS-polyacrylamide gel electrophoresis shows that glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is the major target for modification following treatment with hydroperoxides (hydrogen peroxide or tert-butylhydroperoxide), whereas this enzyme is unaffected following cellular exposure to the thiol oxidant diamide. Further evidence that protein S-thiolation is tightly regulated in response to oxidative stress is provided by the finding that the Tdh3 GAPDH isoenzyme, and not the Tdh2 isoenzyme, isS-thiolated following exposure to H2O2 in vivo, whereas both GAPDH isoenzymes are S-thiolated when H2O2 is added to cell-free extracts. This indicates that cellular factors are likely to be responsible for the difference in GAPDH S-thiolation observed in vivo rather than intrinsic structural differences between the GAPDH isoenzymes. To begin to search for factors that can regulate theS-thiolation process, we investigated the role of the glutaredoxin family of oxidoreductases. We provide the first evidence that protein dethiolation in vivo is regulated by a monothiol-glutaredoxin rather than the classical glutaredoxins, which contain two active site cysteine residues. In particular, glutaredoxin 5 is required for efficient dethiolation of the Tdh3 GAPDH isoenzyme.

Sulfydryl (-SH) groups play a remarkably broad range of roles in the cell, with their redox state affecting the activity and structure of many enzymes, receptors, and transcription factors. As a result, all organisms contain complex regulatory machinery to maintain the redox states of -SH groups in both proteins and low molecular weight sulfydryls (reviewed in Refs. [1][2][3]. However, cysteine residues are among the most easily oxidized residues in proteins, and oxidation can result in intermolecular protein cross-linking and enzyme inactivation, eventually leading to cell death (4). Such irreversible oxidation events can be prevented by protein S-thiolation, in which pro-tein -SH groups form mixed disulfides with low molecular weight thiols such as GSH (5,6).
A variety of proteins that become S-thiolated in response to cellular stress have been detected in mammalian cells. These include key metabolic enzymes such as carbonic anhydrase, glycogen phosphorylase, creatine kinase, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), 1 glutathione transferase, and superoxide dismutase as well as structural and transport proteins such as hemoglobin, actin, and crystallin (reviewed in Ref. 6). Studies have also suggested that this modification may be involved in the regulation of the function and activity of proteins, including the human immunodeficiency virus type 1 protease (7), ubiquitin-conjugating enzymes in bovine retina cells (8), and DNA binding by the transcription factor c-Jun (9). There does not appear to be any one unifying feature of the proteins that become S-thiolated apart from their relative abundance in mammalian cells. Not all -SH-containing proteins are thiolated in response to an oxidative stress, and the specificity of this protein modification indicates that it must be a tightly regulated process. However, the exact role of thiolation in protection against reactive oxygen species-mediated damage is unknown, as are the consequences of this process on cell growth and metabolism.
Recent studies in yeast have shown that protein S-thiolation is important for survival during exposure to oxidative stress conditions. The basal levels of S-thiolation are maintained at low levels, but they are increased following treatment with hydrogen peroxide (10). GAPDH was identified as the major target of protein S-thiolation following exposure to H 2 O 2 (11). This protein modification to GAPDH is tightly regulated, since, despite a high degree of sequence homology (98% similarity, 96% identity), the Tdh3, but not the Tdh2 GAPDH isoenzyme was S-thiolated. Furthermore, activity of both the Tdh2 and Tdh3 GAPDH isoenzymes was reduced following exposure to H 2 O 2 , but only Tdh3 activity was restored within a 1-h recovery period, indicating that S-thiolation of the Tdh3 polypeptide was readily reversible. This protein modification was also shown to be physiologically important, since mutants lacking TDH3 showed increased sensitivity to a challenge with H 2 O 2 , indicating that the S-thiolated Tdh3 polypeptide is required for survival during conditions of oxidative stress. In contrast, the nonthiolated Tdh2 polypeptide was required for survival during exposure to continuous low levels of oxidants, conditions in which the Tdh3 polypeptide was S-thiolated and hence inactive. These findings supported a model in which both enzymes are required for survival during conditions of oxidative stress, playing complementary roles depending on their ability to undergo S-thiolation (11).
Taken together, these previous studies indicate that protein S-thiolation is a novel form of post-translational modification regulating protein activity in response to growth or other cellular signals. Protein S-thiolation must therefore be reversible, and in vitro studies have shown that dethiolation can occur via direct reduction by GSH as well as enzymatically via both glutaredoxins and thioredoxins (5,12). Glutaredoxin appears to be the most active dethiolase, and in addition, a recent study has shown a correlation between protein-SSG reduction and glutaredoxin activity in mammalian cells (13). However, little is known about the factors regulating this protein modification in vivo. In this study, we provide the first in vivo evidence that protein dethiolation can be catalyzed by a monothiol-glutaredoxin rather than the classical glutaredoxins, which contain two active site cysteine residues.

EXPERIMENTAL PROCEDURES
Yeast Strains and Growth Conditions-The Saccharomyces cerevisiae strains used in this study are described in Table I. The double grx3::kanMX4 grx5::kanMX4 and grx4::kanMX4 grx5::kanMX4 mutant strains were constructed using standard yeast genetic methods. Strains were grown in rich YEPD medium (2% (w/v) glucose, 2% (w/v) bactopeptone, 1% (w/v) yeast extract) or minimal SD medium (0.17% (w/v) yeast nitrogen base without amino acids, 5% (w/v) ammonium sulfate, 2% (w/v) glucose) supplemented with appropriate amino acids and bases: 2 mM leucine, 4 mM isoleucine, 1 mM valine, 0.3 mM histidine, 0.4 mM tryptophan, 1 mM lysine, 0.15 mM adenine, 0.2 mM uracil. Media were solidified by the addition of 2% (w/v) agar. Sensitivity to oxidants was determined by growing cells to exponential phase in SD medium at 30°C and treating with the required oxidant. Aliquots of cells were diluted into fresh YEPD medium and plated in triplicate on YEPD plates to obtain viable counts after 3 days of growth.
Determination of Glutathione Levels-Glutathione levels were determined as described previously (10). Briefly, cells were grown in minimal SD medium to an A 600 of 1 (1-2 ϫ 10 7 cells/ml) and harvested by centrifugation. Cells were washed twice with phosphate-buffered saline (pH 7.4) to remove any traces of growth medium and resuspended in ice-cold 8 mM HCl, 1.3% (w/v) 5-sulfosalicyclic acid. Cells were broken with glass beads using a Minibead beater (Biospec Scientific, Bartlesville, OK) for 30 s at 4°C before incubating on ice for 15 min to precipitate proteins. Cell debris and proteins were pelleted in a microcentrifuge for 15 min (13,000 rpm, 4°C), and the supernatant was used for the determination of free glutathione.
Analysis of Protein S-Thiolation-The analysis of protein S-thiolation in yeast has been described previously (11). Briefly, yeast cells were grown to exponential phase in minimal SD medium and treated with 50 g/ml cycloheximide for 15 min to inhibit cytoplasmic protein synthesis. The intracellular pool of low molecular weight sulfydryls was radiolabeled by incubating cells with ϳ0.5 nM L-[ 35 S]cysteine/methionine for 1 h. Cells were then washed and resuspended in fresh SD medium with oxidants to induce S-thiolation. Cell-free extracts were prepared in 20 mM sodium phosphate buffer, pH 7.4, containing 100 mM phenylmethylsulfonyl fluoride, by breaking cells with glass beads using a Minibead Beater for 30 s at 4°C. Parallel extracts were prepared in buffer containing either 50 mM N-ethylmaleimide (NEM) to prevent thiolation during the sample preparation, or 25 mM dithiothreitol (DTT) to reduce any S-thiolated proteins. To follow dethiolation, cells were resuspended in fresh SD medium containing 50 g/ml cycloheximide to inhibit protein synthesis. To quantify protein S-thiolation, aliquots of cell extracts were precipitated on Whatman GF/C glass microfiber filters using 10% (w/v) trichloroacetic acid. Radioactive incorporation was measured by scintillation counting and S-thiolation expressed as the difference between the NEM-and DTT-treated extracts (cpm/g of protein). Proteins were separated by SDS-polyacrylamide gel electrophoresis and visualized using a Typhoon 8600 Imaging system (Amersham Biosciences). Quantification on gels was performed using ImageQuant Solutions software (Amersham Biosciences), and S-thiolation was again expressed as the difference between the NEM-and DTT-treated extracts.
GAPDH Enzyme Assays-GAPDH activity was measured according to McAlister and Holland (14) and expressed as mol of NADH formed per min per g of protein.
Western Blot Analysis-Protein extracts were electrophoresed under reducing conditions on 10% SDS-PAGE minigels and electroblotted onto polyvinylidene difluoride membrane (Amersham Biosciences). The blot was incubated in 2 g/ml anti-GAPDH monoclonal antibody (MAB374; Chemicon International Inc., Temecula, CA). Bound antibody was visualized by chemiluminescence (ECL; Amersham Biosciences) following incubation of the blot in rabbit anti-mouse immunoglobulin-horseradish peroxidase conjugate (Santa Cruz Biotechnology, Inc., Santa Cruz, CA).

Effect of Different Oxidants on Protein S-Thiolation-To
characterize the effects of oxidative stress induced by different oxidants on protein S-thiolation, the intracellular pool of low molecular weight sulfydryls was radiolabeled as described under "Experimental Procedures." Cells were challenged with 2 mM H 2 O 2 , tert-butylhydroperoxide, or diamide, and cell extracts were prepared in the presence of NEM to prevent Sthiolation during the extraction procedure or with DTT to reduce any S-thiolated proteins. The levels of protein S-thiolation were quantified as the difference in counts incorporated between the NEM-and DTT-prepared extracts (Fig. 1A). The basal level of S-thiolation in unstressed cells was very low but increased following treatment with all three oxidants. H 2 O 2 and tert-butylhydroperoxide caused a similar increase in protein S-thiolation, whereas that from diamide was much lower.
To determine the number and range of proteins that are S-thiolated following treatment with the different oxidants, total cell extracts were examined by SDS-PAGE (Fig. 1B). As expected from the radioactive counts, few S-thiolated proteins were detected in unstressed cells, whereas, several proteins were detected following exposure to oxidants. Treatment with 2 mM H 2 O 2 or tert-butylhydroperoxide induced a similar pattern of protein S-thiolation. In both cases, a prominent protein of ϳ38 kDa was detected, which we have previously identified as GAPDH (11). In addition, several unknown proteins were Sthiolated in response to both oxidants (e.g. of 45, 70, and 100 kDa). These unknown proteins were also detected following treatment with diamide, but surprisingly, diamide did not induce protein S-thiolation of GAPDH. Radioactivity incorpora- As in strain S173-6B but tdh1::URA3 tdh2::URA3 As in strain S173-6B but tdh1::URA3 tdh3::URA3 Research Genetics, Inc. 14033 As in strain BY4742 but grx3::kanMX4 Research Genetics, Inc. 16170 As in strain BY4742 but grx4::kanMX4 Research Genetics, Inc. 12769 As in strain BY4742 but grx5::kanMX4 Research Genetics, Inc. Y681 As in strain BY4742 but grx3::kanMX4 grx5::kanMX4 This study Y682 As in strain BY4742 but grx4::kanMX4 grx5::kanMX4 This study tion was confirmed to occur as a result of S-thiolation, since it was reversed by treatment with DTT. These results confirm that protein S-thiolation is tightly regulated in the cell and indicate that different oxidant treatments result in different patterns of modification. Protein S-Thiolation Does Not Correlate with Cellular Redox State or Viability-To determine whether this difference in S-thiolation arose as a result of the oxidants having different effects on the cellular oxidation state, we determined whether the increases in protein S-thiolation levels correlated with changes in the cellular redox state. Glutathione is the most abundant low molecular weight sulfydryl, and differences in the ratio of reduced to oxidized glutathione (GSH/GSSG) can be used as a sensitive indicator of the overall redox balance (15). Exponential phase yeast cells grown under normal aerobic conditions have a high glutathione redox ratio, indicating that most glutathione is maintained in a reduced (GSH) form (16,17). Treatment with both 2 mM H 2 O 2 and tert-butylhydroperoxide reduced the GSH/GSSG ratio, but to different extents (Fig. 2B). H 2 O 2 treatment at this concentration did not increase the levels of GSSG, but it decreased GSH by ϳ40% ( Fig.  2A), reducing the GSH/GSSG ratio from 188 in the control culture to 137. In contrast, tert-butylhydroperoxide elevated GSSG levels by approximately 6-fold, reducing the GSH/GSSG ratio to 32. Diamide, which is a membrane-permeable thiolspecific oxidant (18), readily oxidized glutathione, resulting in a greater than 40-fold increase compared with the control culture ( Fig. 2A). This caused a much reduced GSH/GSSG ratio (4.8), indicating that ϳ15% of the total free glutathione was present in an oxidized form following treatment with diamide (Fig. 2B). Thus, protein S-thiolation did not correlate with changes in the cellular redox state, since the diamide treat-ment caused a large reduction in the cellular redox balance but only a small increase in the levels of S-thiolation compared with the other oxidants. In addition, there was no correlation between the levels of protein S-thiolation and cell viability. Treatments with 2 mM tert-butylhydroperoxide or diamide did not affect viability (Fig. 2C), whereas 2 mM H 2 O 2 resulted in greater than 90% loss of cell viability within 1 h. Thus, protein S-thiolation is a regulated process that occurs in response to particular stress conditions. In order to identify the processes responsible for this regulation, we next examined protein Sthiolation of GAPDH as a model protein.
Both Tdh2 and Tdh3 Can Be S-Thiolated in Vitro-Three unlinked genes (TDH1, TDH2, and TDH3) encode isoenzymes of GAPDH in yeast (19). We have previously shown that the Tdh3 isoenzyme, but not the Tdh2 isoenzyme, is S-thiolated in response to oxidative stress (11). This difference in S-thiolation between the Tdh2 and Tdh3 polypeptides might arise due to differences in their structures that affect the accessibility of sulfydryl groups to S-thiolation, or alternatively, the difference might be due to trans-acting factors that regulate the process in the cell. To discriminate between these two possible mechanisms, we examined the ability of Tdh2 and Tdh3 to be Sthiolated in vitro. Specifically, the pool of low molecular weight sulfydryls was radiolabeled in the wild-type strain and in the tdh1 tdh2 and tdh1 tdh3 mutants. No S-thiolation was detected in the absence of oxidative stress (Fig. 3A, -H 2 O 2 lanes). Treatment of cells with H 2 O 2 resulted in S-thiolation of GAPDH in strains containing Tdh3 (wild type and tdh1 tdh2) but not in strains that lacked Tdh3 (tdh1 tdh3), confirming that Tdh3 is the S-thiolated isoenzyme of GAPDH (Fig. 3A, ϩH 2 O 2 lanes).

FIG. 1. Protein S-thiolation in response to oxidant treatments.
A, following 35 S radiolabeling of the intracellular pool of low molecular weight sulfydryls, cells were treated with 2 mM H 2 O 2 , tert-butylhydroperoxide (t-BH), or diamide. Cell extracts were prepared in the presence of NEM to prevent S-thiolation during the extraction procedure or in the presence of DTT to reduce any S-thiolated proteins. The levels of protein S-thiolation are quantified as the difference in counts incorporated between the NEM-and DTT-prepared extracts and are expressed as cpm incorporated/g of protein. B, following the radiolabeling and oxidant treatments described above, proteins were separated on a 10% SDS-PAGE gel and detected by phosphor imager analysis. Radioactivity incorporation was confirmed to occur as a result of S-thiolation, since it was reversed by treatment with DTT. GAPDH is indicated with an arrow. Molecular mass markers are indicated in kDa.

FIG. 2. Effect of oxidant treatments on glutathione metabolism and cell viability.
Following the same oxidant treatments as described for Fig. 1, the levels of reduced (GSH) and oxidized (GSSG) glutathione (A) as well as the glutathione redox ratios (B) were determined. C, the wild-type strains (CY4) was grown to exponential phase in SD medium and treated with 2 mM H 2 O 2 , tert-butylhydroperoxide (t-BT), or diamide for 1 h. Cells were diluted and plated in triplicate onto YEPD medium to monitor cell viability at 20-min intervals. Percentage survival is expressed relative to the untreated control cultures (100%).
Cell-free extracts were prepared from the nonstressed strains and treated with 1 mM H 2 O 2 for 30 min at 30°C in vitro (Fig.  3B). No S-thiolation was detected in the unstressed extracts, but GAPDH was S-thiolated in response to treatment with H 2 O 2 in extracts from all three strains. Radiolabeling was confirmed to occur as a result of S-thiolation, since it was reversed by treatment with DTT. Thus, unlike the situation found in cells (Fig. 3A), Tdh2 as well as Tdh3 can be S-thiolated in vitro (Fig. 3B). This indicates that cellular factors are likely to be responsible for the differences in GAPDH S-thiolation observed in vivo rather than differences in structure between the GAPDH isoenzymes. To begin to search for possible transacting factors that might regulate the S-thiolation process, we next examined protein S-thiolation in mutants lacking glutaredoxins.
Regulation of GAPDH Activity in Glutaredoxin Mutants-We have previously identified two genes, designated GRX1 and GRX2, in yeast that encode glutaredoxins (20). Mutants lacking GRX1 and GRX2 were unaffected in mixed disulfide formation (20), and similarly, S-thiolation of GAPDH is unaffected in grx1 and grx2 mutants (data not shown). Recently, a new family of glutaredoxin-related proteins has been identified in yeast (GRX3-5) that is conserved throughout evolution from bacterial to mammalian species (21). In addition, Grx5 was found to play an important role in protection against oxidative damage (21). These glutaredoxin-like proteins differ from classical glutaredoxins in that they contain a single cysteine residue at their putative active sites. Thus, they would be able to reduce protein-mixed disulfides in a reaction that proceeds via a monothiol mechanism (21,22). We therefore investigated the role of Grx5 in protein S-thiolation by determining its effect on GAPDH enzyme activity. The wild-type and grx5 mutant strains were grown to exponential phase and treated with 2 mM H 2 O 2 to induce protein S-thiolation (Fig. 4). Prior to the peroxide treatment, the basal level of GAPDH activity in the grx5 mutant was ϳ10-fold higher than in the wild-type strain. Following treatment with H 2 O 2 , GAPDH activity was inhibited by ϳ60 and 89% in the wild-type and grx5 mutant strain, respectively (Fig. 4A). This reduction in GAPDH activity may have arisen due to irreversible oxidation or S-thiolation of the enzyme active site, both of which would inhibit GAPDH activity. To distinguish between these two possibilities, cells were transferred into fresh SD medium to follow the recovery of GAPDH activity once the H 2 O 2 stress was removed. In the wild-type strain, GAPDH activity recovered within 20 min to ϳ90% of the basal level, before the oxidant treatment. In contrast, GAPDH activity in the grx5 mutant was only restored to 30% of the control activity during a 60-min recovery period, indicating that Grx5 is required for efficient dethiolation of S-thiolated GAPDH. The differences in GAPDH activity did not arise due to changes in protein levels. In agreement with the measured enzyme activity, GAPDH protein was elevated in the grx5 mutant compared with the wild-type strain. In addition, Western blot analysis showed that differences in protein levels between the untreated, H 2 O 2 -treated, and recovery extracts could not account for the changes in GAPDH activity detected (Fig. 4B).
Protein S-Thiolation in the grx5 Mutant-To further investigate the role of Grx5 in protein S-thiolation, cells in which the pools of low molecular weight thiols had been radiolabeled were challenged with H 2 O 2 at concentrations between 1 and 4 mM for 30 min (Fig. 5). Protein synthesis in the BY4742 strain background appears to be somewhat more resistant to cycloheximide than the CY4 strain. This was seen as the incorporation of label that could not be reversed by DTT and was therefore due to protein synthesis rather than protein S-thiolation. A peak of GAPDH S-thiolation was seen in the wild-type strain at a concentration of 1 mM H 2 O 2 . In contrast, significantly higher levels of protein S-thiolation were detected in the grx5 mutant at all concentrations of H 2 O 2 tested. In particular, S-thiolation of GAPDH was very prominent, and there was elevated S-thiolation of several unknown proteins with approximate sizes of 30, 45, 60, 70, and 100 kDa.
Loss of GRX5 Results in Delayed Dethiolation of Tdh3-To examine the role of Grx3-5 in dethiolation, wild-type and glutaredoxin mutant cells were radiolabeled and treated with 2 mM H 2 O 2 for 30 min to induce S-thiolation. Cells were washed to remove the H 2 O 2 and any unincorporated radiolabel and resuspended in fresh minimal SD medium to follow the dethiolation process (Fig. 6). In the wild-type strain, dethiolation of

FIG. 3. Protein S-thiolation of GAPDH in vivo and in vitro.
A, protein S-thiolation was analyzed following treatment with 2 mM H 2 O 2 for 30 min in the wild-type (wt) strain (6B) and in strains deleted for both TDH1 and TDH2 and for both TDH1 and TDH3, all encoding GAPDH isoenzymes. S-Thiolation of GAPDH was only detected in strains containing Tdh3 (wild type and tdh1 tdh2) but not in strains that lacked Tdh3 (tdh1 tdh3), confirming that Tdh3 is the S-thiolated isoenzyme of GAPDH. Radiolabeling was confirmed to occur as a result of S-thiolation, since it was reversed by treatment with DTT. B, cell-free extracts were prepared from the nonstressed strains and treated with 1 mM H 2 O 2 for 30 min at 30°C in vitro. No S-thiolation was detected in the unstressed extracts, but GAPDH was S-thiolated in response to treatment with H 2 O 2 in extracts from all three strains. Radiolabeling was again confirmed to occur as a result of S-thiolation, since it was reversed by treatment with DTT GAPDH was ϳ80% complete within 5 min. In contrast, dethiolation of GAPDH was significantly slower in the grx5 mutant compared with the wild-type strain and was only 50% complete after a 10-min recovery period (p Ͻ 0.001). However, dethiolation did occur in the grx5 mutant and was complete within 20 min. These results indicate that Grx5 is required for the efficient dethiolation of GAPDH, but other factors can also catalyze the reaction. These other factors do not appear to be Grx3 or Grx4, since dethiolation of GAPDH occurred with similar kinetics to the wild-type strain in a grx3 and a grx4 mutant (data not shown). To further examine the roles of Grx3 and Grx4, dethiolation was examined in grx3 grx5 and grx4 grx5 double mutant cells (Fig. 6B). Loss of GRX3 or GRX4 in the grx5 mutant had little or no effect on dethiolation, which was still complete within the 20-min recovery period.
Loss of GRX5 Lowers the Glutathione Redox Ratio-To determine whether glutaredoxins 3-5 are required to maintain the cellular redox state, the levels and redox ratio of glutathi- FIG. 4. Inhibition and recovery of GAPDH enzyme activity in a grx5 mutant. A, GAPDH enzyme activity (mol/min/g) was determined in the wild-type (wt) strain (BY4742) and grx5 mutant grown to exponential phase in SD medium (control). Cells were treated with 2 mM H 2 O 2 for 30 min to induce S-thiolation, and aliquots were taken to determine GAPDH activity (ϩH 2 O 2 ). Cells were washed and resuspended in fresh SD medium to follow the recovery of GAPDH activity after 20 or 60 min, respectively. Results are expressed as percentages relative to the untreated control cultures (basal levels: wild-type strain ϭ 4.4 mol/min/g and grx5 mutant ϭ 54.0 mol/min/g). B, Western blot analysis of extracts from the wild type and grx5 mutant probed with anti-GAPDH antibody.
FIG . 5. Protein S-thiolation is elevated in a grx5 mutant. Following radiolabeling of the intracellular pool of low molecular weight sulfydryls, wild-type (wt) (A) and grx5 mutant cells (B) were treated with various concentrations of H 2 O 2 for 30 min. Proteins were separated on 10% SDS-PAGE gels and analyzed by phosphorimaging. Radioactivity incorporation was confirmed to occur as a result of S-thiolation, since it was reversed by treatment with DTT. GAPDH is indicated with an arrow, and unknown proteins that show higher levels of S-thiolation in the grx5 mutant are indicated with an asterisk. Molecular mass markers are indicated in kDa. one were measured in grx3-5 mutants (Table II). Previous work has shown that glutathione is unaffected in grx1 or grx2 mutants (20). Similarly, strains lacking GRX3 or GRX4 contained levels of GSH and GSSG that were comparable with the wild type, resulting in a redox ratio of ϳ140 (Table II). In contrast, the grx5 mutant contained approximately 4-fold higher levels of GSSG, which resulted in a lowered GSH/GSSG redox ratio of 46. The treatment with 2 mM H 2 O 2 for 30 min, which was used to induce protein S-thiolation in previous experiments (Figs. 4 -6), did not affect the wild-type strain but caused a small increase in GSSG (27%) and decrease in the GSH redox ratio (17%). DISCUSSION It has long been recognized that low molecular weight thiols, such as GSH, can interact in a reversible manner with the cysteine -SH groups in many cellular proteins (6,23). In particular, protein S-thiolation/dethiolation is a dynamic process that occurs in cells under physiological conditions as well as following exposure to an oxidative stress (5,24,25). Models have been proposed in which the modification of proteins by S-thiolation does not require an enzymatic activity but pro-ceeds via the reaction of partially oxidized protein sulfydryls (thiyl radical or sulfenic acid intermediates) with thiols such as cysteine or GSH or by thiol/disulfide exchange reactions with the oxidized disulfide form of glutathione (GSSG) (5). However, the tight control of GAPDH S-thiolation described in this present study indicates that cellular factors must regulate the modification of Tdh3. The Tdh2 and Tdh3 GAPDH isoenzymes share extensive sequence homology (98% similarity, 96% identity), particularly in the regions surrounding their active site cysteine residues (14,19), but the Tdh3, and not the Tdh2 isoenzyme, is reversibly S-thiolated following exposure to H 2 O 2 . In contrast, both GAPDH isoenzymes could be S-thiolated in vitro using cell-free extracts. Thus, there do not appear to be any intrinsic structural differences to account for the difference in protein-mixed disulfide formation between the two GAPDH isoenzymes. Our studies also show that S-thiolation of GAPDH is a specific response to particular oxidants, since it is not increased in response to oxidative stress induced by treatment with diamide. This was despite the fact that diamide readily shifts the cellular redox balance to a more oxidized state and induces the S-thiolation of other unknown proteins. Therefore, S-thiolation in yeast is a controlled process that is regulated by specific cellular factors. The factors that are responsible for the differential S-thiolation of Tdh2 and Tdh3 remain to be identified.
To protect protein -SH groups against irreversible oxidation or to serve an enzyme-regulatory function, protein S-thiolation must be readily reversible. Many studies have demonstrated that S-thiolated proteins formed during oxidative stress are readily dethiolated once the stress is removed (25)(26)(27). In addition, the reduction of mixed disulfides is stimulated by both NADH-and NADPH-dependent systems (27,28), but the physiological electron donors have not been identified. Most studies investigating the factors controlling dethiolation have been performed in vitro using radiolabeled protein-mixed disulfides as substrates. These studies have implicated various oxidoreductases including those of the GSH/glutaredoxin and thioredoxin systems (12,29). Glutaredoxin appears to be the most efficient dethiolase enzyme in these in vitro experiments. In addition, a correlation has been demonstrated between protein-SSG reduction and glutaredoxin activity in mammalian cells (13), and the reversible S-glutathiolation of human immunodeficiency virus type 1 protease can be catalyzed by a glutaredoxin in vitro (7). We therefore investigated the role of yeast glutaredoxins in protein S-thiolation.
Yeast contains two classical glutaredoxin genes, designated GRX1 and GRX2, whose protein products share 40 -52% identity and 61-76% similarity with those from bacterial and mammalian species (20). Additionally, a family of glutaredoxinrelated proteins has been identified in yeast (encoded by GRX3-5), which is conserved throughout evolution (21). These glutaredoxin-like proteins differ from Grx1 and Grx2 in that they contain a single cysteine residue at their putative active FIG. 6. Dethiolation of Tdh3 is slower in a grx5 mutant. Following radiolabeling of the intracellular pool of low molecular weight sulfydryls, wild-type (wt), grx5, grx3 grx5, and grx4 grx5 mutant cells were treated with 2 mM H 2 O 2 for 30 min to induce S-thiolation. Cells were washed, resuspended in fresh SD medium, and incubated for 20 min to follow the dethiolation process. A, a representative experiment showing dethiolation in the wild-type and grx5 mutant following a 10and 20-min recovery period. Radioactivity incorporation was confirmed to occur as a result of S-thiolation, since it was reversed by treatment with DTT. GAPDH is indicated with an arrow, and molecular mass markers are indicated in kDa. B, Tdh3-S-thiolation was quantified in the wild-type and glutaredoxin mutants treated with the same conditions as described above. Values shown are percentages relative to the levels of S-thiolation induced by the H 2 O 2 treatment (100%). All data points represent the mean of at least three independent determinations, with A showing a single representative experiment. sites. Thus, they would be unable to substitute for glutaredoxins or thioredoxins as disulfide reductases with substrates like ribonucleotide reductase that require a dithiol mechanism (22). However, they could reduce protein-mixed disulfides in a reaction that proceeds via a monothiol mechanism (21,22). In agreement with this, our studies show that Grx5 is required for dethiolation in vivo and can reduce mixed disulfides formed by GAPDH and other S-thiolated proteins. In contrast, the dithiol glutaredoxins (Grx1, Grx2) are not affected in the formation of mixed disulfides (20) as well as in the S-thiolation/dethiolation of GAPDH (data not shown). Similarly, strains lacking GRX3 or GRX4 are unaffected in the protein S-thiolation/dethiolation process.
This study shows that the grx5 mutant contains higher levels of S-thiolated Tdh3 compared with the wild-type strain following exposure to 1 mM H 2 O 2 . Western blot analysis shows that the grx5 mutant contains higher levels of GAPDH protein, so it is possible that the amount of thiolation per molecule of GAPDH may be the same between the grx5 mutant and the wild-type strain. However, it seems unlikely that elevated levels of GAPDH protein can account for the increased GAPDH thiolation observed in the grx5 mutant in response to higher levels of H 2 O 2 (2 and 4 mM). The level of Tdh3 S-thiolation peaks at 1 mM H 2 O 2 in the wild-type strain, whereas thiolation is elevated at concentrations of 2 and 4 mM in the grx5 mutant. In addition, the levels of thiolation for several unknown proteins are elevated in the grx5 mutant compared with the wildtype strain, consistent with a role for Grx5 in dethiolation.
Loss of GRX5 results in a delayed recovery of GAPDH enzyme activity following oxidant treatment, with only a 2-fold recovery detected in the grx5 mutant following a 1-h growth period. In contrast, although the grx5 mutant shows a slower rate of Tdh3 dethiolation, it is still complete within a 20-min recovery period. This difference presumably reflects the presence of three GAPDH isoenzymes in yeast, of which only Tdh3 can be protected against irreversible oxidation by protein Sthiolation. Thus, dethiolation of the Tdh3 isoenzyme in the grx5 mutant would account for the limited recovery of GAPDH activity observed. Oxidation of the Tdh1 and Tdh2 isoenzymes will result in irreversible oxidation. Given that the grx5 mutant is extremely sensitive to oxidative stress (21), these nonthiolated isoenzymes are likely to be readily inactivated. Thus, simply dethiolating Tdh3 in the grx5 mutant does not restore GAPDH activity to the basal levels found prior to the oxidative stress within a 1-h time period. It is also interesting to note that GAPDH protein levels are elevated in the grx5 mutant, and this may arise in a mechanism where GAPDH synthesis is increased to compensate for any oxidized, and hence inactivated, GAPDH that may be formed.
Loss of GRX5 results in extremely slow growth and elevated levels of protein oxidative damage compared with strains lacking GRX1-GRX4 (21). Here, we have shown that the cellular redox balance is also shifted toward a more oxidized form (measured as GSH/GSSG) in a grx5 mutant, whereas it is unaffected by the loss of GRX1-GRX4. Thus, GRX5 appears to be the most important glutaredoxin for protection against oxidative stress. However, the synthetic lethality described for a grx2 grx5 and grx3 grx4 grx5 mutant indicates that there must be some overlapping functions shared by the different glutaredoxins (21). Although this study has indicated a role for Grx5 in the dethiolation of protein-mixed disulfides, other activities must be required, since the reduction of S-thiolated GAPDH can still proceed. In addition, there may also be enzymes that regulate the forward S-thiolation reaction. Mammalian glutaredoxins have been shown to catalyze both the formation and reduction of mixed disulfides (30). However, yeast glutaredoxins do not appear to be required for S-thiolation, since mixed disulfides are formed in grx1-grx5 mutants in response to oxidative stress. Other enzymes that regulate protein S-thiolation may represent already known factors involved in the maintenance of cellular redox homeostasis or as yet unidentified activities. Yeast will provide an ideal model system to search for these factors, given the genetic and biochemical tractability of the organism as well as the ready availability of null mutants.