S-Cysteinylation Is a General Mechanism for Thiol Protection of Bacillus subtilis Proteins after Oxidative Stress*

S-Thiolation is crucial for protection and regulation of thiol-containing proteins during oxidative stress and is frequently achieved by the formation of mixed disulfides with glutathione. However, many Gram-positive bacteria including Bacillus subtilis lack the low molecular weight (LMW) thiol glutathione. Here we provide evidence that S-thiolation by the LMW thiol cysteine represents a general mechanism in B. subtilis. In vivo labeling of proteins with [35S]cysteine and nonreducing two-dimensional PAGE analyses revealed that a large subset of proteins previously identified as having redox-sensitive thiols are modified by cysteine in response to treatment with the thiol-specific oxidant diamide. By means of multidimensional shotgun proteomics, the sites of S-cysteinylation for six proteins could be identified, three of which are known to be S-glutathionylated in other organisms.

S-Thiolation is crucial for protection and regulation of thiolcontaining proteins during oxidative stress and is frequently achieved by the formation of mixed disulfides with glutathione. However, many Gram-positive bacteria including Bacillus subtilis lack the low molecular weight (LMW) thiol glutathione. Here we provide evidence that S-thiolation by the LMW thiol cysteine represents a general mechanism in B. subtilis. In vivo labeling of proteins with [ 35 S]cysteine and nonreducing twodimensional PAGE analyses revealed that a large subset of proteins previously identified as having redox-sensitive thiols are modified by cysteine in response to treatment with the thiolspecific oxidant diamide. By means of multidimensional shotgun proteomics, the sites of S-cysteinylation for six proteins could be identified, three of which are known to be S-glutathionylated in other organisms.
Protein S-thiolation by low molecular weight (LMW) 3 thiols prevents the irreversible oxidation of cysteine residues during oxidative stress and plays a pivotal role in the redox regulation of thiol-containing proteins. For example, the formation of mixed disulfides between target proteins and the LMW thiol glutathione is a key event in the regulation of the eukaryotic proteins, including c-Jun, thioredoxin, and glyceraldehyde-3phosphate dehydrogenase (1)(2)(3)(4)(5). S-Glutathionylation has also been reported to modulate the activity of Escherichia coli methionine synthase (MetE), phosphoadenosine phosphosul-fate reductase (CysH), and the oxidative stress-specific transcription factor OxyR (6 -8).
Because many Gram-positive bacteria lack glutathione, the nature of S-thiolation in these organisms remains elusive (9). In the Gram-positive model organism Bacillus subtilis, cysteine represents the most abundant LMW thiol (9). One of the most obvious responses of B. subtilis to disulfide stress is the strong induction of cysteine biosynthesis genes (10). Although the origins of this effect are unclear, it might reflect a consumption of free cysteine by oxidation to cystine and the formation of mixed disulfides with proteins.
In a previous proteome analysis, we reported that reversible thiol oxidation occurs in a number of B. subtilis proteins after oxidative stress indicating a general thiol-protection mechanism in this organism (11). We also showed that overoxidation of cysteine residues to sulfonic acid caused by high level peroxide stress results in protein damage and in the irreversible loss of protein activity in B. subtilis and Staphylococcus aureus (11,12). However, the nature of the reversible thiol modifications was not apparent and likely reflected a mixture of intra-or intermolecular protein disulfides as well as mixed disulfides with LMW thiols. Recently, Lee et al. (13) demonstrated that the organic peroxide-sensing repressor OhrR of B. subtilis is reversibly inactivated by S-thiolation by cysteine, an unknown thiol of 398 Da, or coenzyme A (CoASH) under conditions of oxidative stress. S-Thiolation by CoASH also has been shown to occur during spore formation in Bacillus megaterium and 45% of the CoASH in spores was found as mixed disulfides with protein (14).
In the present study, we conducted a global survey to detect proteins modified by S-cysteinylation under conditions of disulfide stress. By in vivo [ 35 S]cysteine labeling and two-dimensional PAGE, a large subset of proteins was observed, which showed S-thiolation in response to diamide treatment. Multidimensional "shotgun" proteomics revealed S-cysteinylation and identified the sites of this modification in six different B. subtilis proteins. No evidence was obtained under these conditions for S-thiolation by either CoASH or the recently described 398-Da LMW thiol (13,15). Our results suggest that S-cysteinylation represents a general mechanism of thiol protection and functions as an actual mode of redox regulation under disulfide stress conditions in B. subtilis.

EXPERIMENTAL PROCEDURES
Strain and Growth Medium-The B. subtilis wild type strain 168 (16) was grown aerobically at 37°C in a synthetic medium (17).
[ 35 S]Cysteine Monitoring of S-Thiolations-To prevent protein synthesis, 500 g/ml chloramphenicol was added to exponentially growing cells (A 500 nm 0.5) 30 min before cell treatment. The cells were incubated with 10 Ci/ml of L-[ 35 S]cysteine (specific activity 1,000 Ci/mmol and radioactive concentration 10 mCi/ml at reference date, GE Healthcare) for 20 min to provide labeled cysteine. Conditions of oxidative stress were achieved by the addition of a final concentration of 1 mM diamide (11,18) for 30 min. For control purposes, cultivation was carried out without the addition of diamide.
Cell Sampling and Blocking of Thiol Groups-Cells were harvested by centrifugation with 8,900 ϫ g for 5 min at 4°C. Subsequently, the resulting pellet was resuspended in denaturing buffer (8 M urea, 1% CHAPS, 1 mM EDTA, 200 mM Tris-HCl, pH 8.0) containing 100 mM iodoacetamide. Crude cell extracts were obtained by sonication and were then incubated for 20 min at room temperature to alkylate accessible free thiol groups. To terminate the blocking reaction, proteins were precipitated by addition of 4 parts ice-cold acetone and stored at Ϫ20°C for 1 h. Samples were then centrifuged at 21,900 ϫ g for 15 min at 20°C and washed with 80% acetone and pure acetone. The resulting protein pellet was dried in a vacuum centrifuge. Samples were dissolved in 150 l of denaturing buffer without iodoacetamide and the protein concentration was determined. To prove reversible S-thiolations, aliquots of the labeled protein extracts were treated with the disulfide reducing agent at final concentrations of 10 mM dithiothreitol or 1 mM Tris-(2carboxyethyl)-phosphine for 30 min at room temperature.
Measurement of Protein-associated Radioactivity with and without Reducing Agents-Similar protein amounts (150 and 300 g) of the untreated or reduced samples were loaded on filter disks. The protein bound radioactivity was measured in a scintillation counter (19). A dilution series of L-[ 35 S]cysteine from 2 to 1000 nCi was used to calculate the radioactivity from the measured counts per minute.
Nonreducing Two-dimensional PAGE-Similar protein amounts (100 g) with 35 S-labeled S-thiolations were separated by two-dimensional PAGE according to Büttner et al. (20). To ensure nonreducing conditions dithiothreitol and thiourea were omitted from the procedure. Gels were stained with silver nitrate, and total protein images were obtained using a light scanner with a transparency unit. For autoradiograms, gels were dried on filter paper, exposed to storage phosphor screens, and scanned with the Typhoon 9400 variable mode imager (Amersham Biosciences).
Shotgun Proteomics-Diamide was added to exponentially growing cells (A 500 nm 0.5) to a final concentration of 1 mM for 20 min. Cultivation was carried out without the addition of diamide for control purposes. Cells were harvested, and thiol groups were blocked as described above. The protein mixture was mixed with nonreducing PAGE loading buffer and separated by electrophoresis in a one-dimensional SDS-PAGE mini-gel. In-gel digestion of equal gel pieces was carried out with trypsin as described recently (21). The peptide mixture was subjected to a reversed phase separation using the Ettan MDLC System (GE Healthcare) and analyzed using a LTQ FTICR mass spectrometer (Thermo Electron Corp., San Jose, CA) (22). MS data sets were analyzed as described in the legend for supplemental Fig. S2. All data were obtained from two independently performed MS analyses of at least two independent cell cultivation experiments.

RESULTS AND DISCUSSION
To find out whether B. subtilis uses free cysteine to form mixed protein disulfides under conditions of disulfide stress, we analyzed cells treated with the disulfide-generating agent dia-mide by means of the [ 35 S]cysteine labeling approach (23). In brief, growing cells were treated with chloramphenicol to prevent protein synthesis followed by the addition of L-[ 35 S]cysteine to provide free labeled cysteine. To provoke thiol oxidation, the cells were stressed with diamide (11,18). After extraction of intracellular proteins, the protein-associated radioactivity was finally measured. Compared with untreated control cells, intracellular proteins of cells treated with diamide showed an approximately 6-fold increase in the amount of 35 S radioactivity (Fig. 1).
To evaluate whether the detected increase in 35 S-protein labeling is a reversible thiol modification, we incubated aliquots of the labeled protein extracts with disulfide reducing agents. The reduction step resulted in about an 81% loss of radioactivity in the diamide-treated sample, indicating that the vast majority of 35 S labeling accounts for mixed protein disulfides initiated by cysteine (Fig. 1).
On the basis of these results, we considered that calculating the amount of protein bound [ 35 S]cysteine could give us a lower limit of the number of S-thiolations per cell. We estimated about a 30-fold increase in cysteine-based S-thiolation events per cell after 30 min of diamide treatment. This corresponds to an absolute number of at least 1000 cysteine-based S-thiolation events in one single cell after disulfide stress. Because of putative S-thiolation using unlabeled cysteine and significant losses of thiol modifications during the experimental procedure itself, we have provided this number as the lower limit.
To visualize the overall pattern of S-thiolated proteins, [ 35 S]cysteine-labeled protein extracts were separated by two-dimensional gel electrophoresis (20) under nonreducing conditions. The autoradiograms of the two-dimensional gels revealed a strong increase in the number of protein spots with associated 35 S radioactivity after diamide stress, again indicating S-thiolation after oxidative stress (see supplemental Fig. S1). These results suggest that many intracellular proteins of B. subtilis are readily S-thiolated by free cysteine (or a derivative) after diamide treatment.
To characterize the precise molecular nature of S-thiolations and identify the specific sites of the S-thiolation events, we surveyed cytoplasmic proteins from diamidetreated cells by multidimensional shotgun proteomics using a LTQ FTICR mass spectrometer. In summary, growing cells were treated with diamide to provoke severe reversible thiol oxidation. The protein fraction was separated by nonreducing one-dimensional protein gel electrophoresis, and in-gel digestion by trypsin using gel pieces was carried out. Peptide mixtures were then subjected to a reversed phase separation using a nanoscale liquid chromatography system and were measured with the LTQ FTICR mass spectrometer. The data were analyzed using the SEQUEST algorithm and a data base of B. subtilis proteins as described in the legend for supplemental Fig. S2. In brief, the data set was scanned for peptides with mixed disulfide bonds to cysteine, homocysteine, N-acetylcysteine, CoASH, ␥-glutamylcysteine, glutathione, or the novel 398-Da LMW thiol. The filtering was achieved by using Bio-worksBrowser software to identify peptides with a mass deviation below 2 ppm between the measured and the calculated mass (see supplemental Fig. S2 for details).
The LTQ FTICR analysis revealed sites of S-cysteinylation in six different peptides from six different cysteine-containing proteins after disulfide stress ( Table 1). The modified proteins included inosine-monophosphate dehydrogenase (GuaB), cobalamin-independent methionine synthase (MetE), and  inorganic pyrophosphatase (PpaC). All six proteins showed an additional mass of 119 Da at a cysteine-containing peptide, which is specific for S-cysteinylation ( Fig. 2 and supplemental  Fig. S2). The fragment ion spectra of the target peptides were used to confirm the exact position of the S-cysteinylation at the cysteine within the amino acid sequence of the corresponding peptide ( Fig. 2 and Table 1). In peptide mixtures derived from control cells (grown without diamide) no S-thiolated proteins were detected, indicating that protein S-cysteinylation accumulates during oxidative stress. Four of the six identified S-cysteinylated proteins, namely GuaB, MetE, PpaC, and the protein YwaA with similarity to branched-chain amino acid aminotransferases, were already reported to show increased amounts of reversible thiol oxidation after oxidative stress using a fluorescence-based thiol modification assay (11). Three homologous proteins (GuaB, MetE, and PpaC) are known to be S-glutathionylated in other organisms (6,24,25) (Table 1).
The inosine-monophosphate dehydrogenase GuaB was found to be S-cysteinylated at its active site cysteine (Cys-308). A regulatory significance of this GuaB S-thiolation is yet unknown, but our results point to a S-cysteinylation-dependent inhibition of GuaB, which in turn may cause an arrest of GTP biosynthesis and therefore may help to prevent the accumulation of the highly mutagenic 7,8-dihydro-8-oxoguanine (26) under conditions of oxidative stress.
For E. coli, it has been shown that the inhibition of methionine biosynthesis is caused by the reversible thiol oxidation of methionine synthase MetE after oxidative stress (6). Likewise, we have demonstrated thiol oxidation of MetE and methionine auxotrophy in B. subtilis cells after diamide or peroxide treatment (11). In contrast to the S-glutathionylation that was detected for MetE of E. coli, our present study shows that MetE of B. subtilis is S-cysteinylated on Cys-719. We therefore suggest that B. subtilis MetE is protected and reversibly inactivated by S-cysteinylation at Cys-719 under conditions of oxidative stress. In B. subtilis, the cysteine biosynthesis is tightly controlled by different regulators, including CymR, CysL, and Spx (27)(28)(29)(30). In response to disulfide stress, B. subtilis cells initiate a strong induction of cysteine biosynthesis genes including the methionine to cysteine conversion pathway (10,31). The activation of cysteine biosynthesis after thiol-specific oxidative stress is supposed to be caused by depletion of the intracellular cysteine pool following highly elevated S-cysteinylation to protect redox-sensitive thiols from irreversible oxidation. Therefore, it can be presumed that the reported inhibitory S-cysteinylation of MetE at Cys-719 contributes in concert with the strong induction of cysteine biosynthesis genes to redirecting the metabolism from methionine toward cysteine biosynthesis for LMW thiol generation.
A high number of proteins were detected to be S-thiolated by the [ 35 S]cysteine assay. The six proteins identified with S-cysteinylation by mass spectrometry appear to be just the tip of the iceberg, because in this initial study we obtained only proteins with S-cysteinylation that were reported to be among the 115 most abundant proteins in growing B. subtilis cells (21). Apart from the S-cysteinylation, no other type of S-thiolation could be detected in B. subtilis cells after diamide treatment. The lack of S-glutathionylation and S-␥-glutamylcysteinylation may be explained by the lack of the genes encoding the corresponding enzymes for biosynthesis of glutathione or ␥-glutamylcysteine in B. subtilis. The failure to detect mixed disulfides with CoASH and the recently described 398-Da LMW thiol (13,32) may be due to a comparatively low abundance of these thiols, a preferential reaction of diamide with cysteine versus other LMW thiols, or a differential sensitivity of detection of the various modified peptides by LTQ FTICR mass spectrometry. It is known, for example, that CoASH-modified peptides are very poorly detected in positive ion mode mass spectrometry studies, likely because of the poor ionization of the phosphate-containing CoASH moiety (32). Further analyses will be required to develop methods to accurately quantify the various LMW thiols in B. subtilis, and their relative contributions to thiol modification under various oxidative stress conditions.
In the present work, we have demonstrated that the S-thiolation of proteins of B. subtilis can be detected by the addition of labeled free cysteine during oxidative stress treatment. Importantly, we were able to identify six different proteins by multidimensional shotgun proteomics that showed in vivo S-cysteinylation after disulfide stress. Taken together, our results suggest that S-cysteinylation is a general mechanism to protect protein thiols from further oxidation and irreversible damage in B. subtilis and may play a role in redox regulation of several thiol-containing proteins.