OxyR2 Functions as a Three-state Redox Switch to Tightly Regulate Production of Prx2, a Peroxiredoxin of Vibrio vulnificus*

The bacterial transcriptional regulator OxyR is known to function as a two-state redox switch. OxyR senses cellular levels of H2O2 via a “sensing cysteine” that switches from the reduced to a disulfide state upon H2O2 exposure, inducing the expression of antioxidant genes. The reduced and disulfide states of OxyR, respectively, bind to extended and compact regions of DNA, where the reduced state blocks and the oxidized state allows transcription and further induces target gene expression by interacting with RNA polymerase. Vibrio vulnificus OxyR2 senses H2O2 with high sensitivity and induces the gene encoding the antioxidant Prx2. In this study, we used mass spectrometry to identify a third redox state of OxyR2, in which the sensing cysteine was overoxidized to S-sulfonated cysteine (Cys-SO3H) by high H2O2 in vitro and in vivo, where the modification deterred the transcription of prx2. The DNA binding preferences of OxyR25CA-C206D, which mimics overoxidized OxyR2, suggested that overoxidized OxyR2 binds to the extended DNA site, masking the −35 region of the prx2 promoter. These combined results demonstrate that OxyR2 functions as a three-state redox switch to tightly regulate the expression of prx2, preventing futile production of Prx2 in cells exposed to high levels of H2O2 sufficient to inactivate Prx2. We further provide evidence that another OxyR homolog, OxyR1, displays similar three-state behavior, inviting further exploration of this phenomenon as a potentially general regulatory mechanism.

Reactive oxygen species build up as metabolic by-products from oxygen during aerobic respiration processes (1,2). Of the reactive oxygen species, hydrogen peroxide (H 2 O 2 ) is toxic to cellular components, especially when it is converted to the hydroxyl radical via the Fenton reaction in the presence of iron or copper ions (3,4). Pathogenic bacteria inevitably encounter elevated levels of H 2 O 2 imposed by the host immune response system during infection. Therefore, pathogenic bacteria have to cope with H 2 O 2 to survive host environments and in turn to ensure developing illness. Therefore, the bacteria's mechanisms to defense against H 2 O 2 are closely linked to their virulence (5). Bacterial defense against H 2 O 2 relies on a variety of antioxidant defense enzymes, such as peroxiredoxins (Prxs) 3 and catalases (6).
OxyR is a central regulator of the antioxidant genes in many bacteria and forms a homotetramer (7). OxyR is a modular structure consisting of an N-terminal DNA binding domain and C-terminal regulatory domain that contains two conserved redox-sensitive cysteines. Typical OxyR, such as Escherichia coli OxyR, has been considered as a "two-state redox switch" that precisely modulates expression of many antioxidant genes in a H 2 O 2 -dependent manner (7)(8)(9). When H 2 O 2 is low and within safe limits, the redox-sensitive cysteines are present as free thiols separated by an ␣-helix, resulting in the "reduced state" of OxyR. In the reduced state, the OxyR tetramer adopts an extended conformation (10) and binds to the extended DNA-binding sites to mask the Ϫ35 region of the target antioxidant genes (11,12). Thereby, the reduced OxyR turns off the expressions of target antioxidant genes under conditions of low H 2 O 2 (12)(13)(14). When H 2 O 2 increases and exceeds safe limits, one of the redox-sensitive cysteines, named sensing cysteine, senses H 2 O 2 and is rapidly oxidized to Cys-SOH (15). Given the high reactivity, the Cys-SOH readily forms an intramolecular disulfide bond (Cys-S-S-Cys) with the other redox-sensitive cysteine. This "oxidized state or disulfide state" of OxyR adopts compact conformation, binds to more compact DNA-binding sequences, and unmasks the Ϫ35 region (11,12). The oxidized OxyR interacts with the C-terminal domain of RNA polymerase ␣-subunits and turns on the target antioxidant genes under conditions of high H 2 O 2 (8,12,16).
The facultative aerobic pathogen Vibrio vulnificus has two OxyRs, OxyR1 and OxyR2 (11). Amino acid sequences of both OxyR proteins have the conserved residues that are functionally important in OxyR, indicating that OxyR1 and OxyR2 share common functional and structural characteristics with the OxyR homologs of many bacteria. OxyR1 and OxyR2 activate prx1 and prx2, respectively (11). Each of the Prxs scavenges distinct ranges of H 2 O 2 (i.e. working ranges of H 2 O 2 for each Prx). Prx2 has a higher affinity for H 2 O 2 than Prx1 and thus scavenges lower levels of H 2 O 2 more effectively, whereas Prx1 detoxifies higher levels of H 2 O 2 more effectively. It is notewor-thy that Prx2 is sensitive, whereas Prx1 is relatively robust, to H 2 O 2 . Prx2 is irreversibly inactivated by overoxidation at one of the catalytic cysteines, peroxidatic cysteine, to S-sulfonated cysteine (Cys-SO 3 H) in cells exposed to 30 M or higher levels of H 2 O 2 that exceed its working range (17).
This observation led us to raise a question on the redox state and transcriptional activity of OxyR2 in the cells in which cellular H 2 O 2 level is high enough to irreversibly inactivate Prx2. Accordingly, the present study examined the redox state of OxyR2 exposed to various levels of H 2 O 2 in vivo as well as in vitro. MS analysis found that the sensing cysteine is overoxidized to Cys-SO 3 H by high H 2 O 2 in vitro. The "overoxidized state" is a third redox state in addition to the reduced and disulfide states of OxyR2. We confirmed the presence of the overoxidized state of OxyR2 in V. vulnificus cells exposed to H 2 O 2 exceeding 30 M. Biochemical and genetic studies demonstrated that the overoxidized OxyR2 deters the production of Prx2 that is no longer functional. The sequences for binding of the overoxidized OxyR2 were determined to explain the mechanism of overoxidized OxyR2 to turn off the prx2 expression. Finally, the possibility that the OxyR1 is also overoxidized to play a role in the regulation of its antioxidant genes was investigated.

Results
Sensing Cysteine of OxyR2 Is Overoxidized at High Levels of H 2 O 2 -The purified OxyR2 protein was reacted with 500 M H 2 O 2 in vitro, and the redox state of the sensing cysteine Cys 206 was subjected to MALDI-TOF MS analysis in positive and negative ion reflector modes (Fig. 1). Peptides can be detected in positive and/or negative ion reflector modes, depending on the intrinsic properties of the peptides that influence the ionization behavior (18). Mass spectra for the tryptic digests of the OxyR2 protein were analyzed after the free cysteine residues in the sample were alkylated with iodoacetamide to prevent further oxidation during the analysis. As shown in Fig. 1A, the peptide fragment containing both redox-sensitive cysteine residues (EHC 206 LTEHAVSAC 215 K) was detectable in the negative ion reflector mode. It remained reduced in the absence of H 2 O 2 and thus produced the doubly alkylated version (Cys-S-CH 2 CONH 2 ) with a monoisotopic mass [M Ϫ H] Ϫ of 1539.6 ( Fig. 1A, top)  indicating that Cys 215 did not undergo the overoxidation. The EHS 206 LTEHAVSAC 215 K peptide fragment was not detectable in the negative ion mode (Fig. 1B, left panels). Therefore, the combined results demonstrated that the sensing cysteine, Cys 206 , but not the other cysteine, Cys 215 , of OxyR2 is overoxidized to Cys-SO 3 H in the presence of high levels of H 2 O 2 in vitro.
Overoxidation of OxyR2 in the V. vulnificus Cells Exposed to H 2 O 2 -To ascertain whether the overoxidation of OxyR2 at the sensing cysteine indeed occurs by cellular H 2 O 2 in V. vulnificus, the redox state of cellular OxyR2 was analyzed. The oxyR2 mutant OH0703 with pDY1025 expressing oxyR2 was grown anaerobically, exposed to various concentrations of H 2 O 2 , and then treated with 0.5-kDa 4-acetamido-4Ј-maleimidylstilbene-2, 2Ј-disulfonic acid (AMS; Invitrogen) to alkylate free thiols in the proteins. The total cellular proteins (3.5 g/well) were resolved on non-reducing SDS-PAGE and immunoblotted with anti-OxyR2 antibody. Alkylation of a free thiol in the proteins with AMS adds 0.5 kDa of molecular mass ( Fig. 2A,  top). OxyR2 in the cells without exposure to H 2 O 2 existed in the reduced state in which all seven cysteine residues present in OxyR2 were alkylated with AMS (7AMS control). An OxyR2 band containing pentuply alkylated cysteine residues coexisted with the heptuply alkylated OxyR2 band when the cells were exposed to 10 M H 2 O 2 , indicating that a part of OxyR2 was oxidized to form a disulfide bond, as observed in the previous report (11). More importantly, a portion of OxyR2 in the cells exposed to H 2 O 2 exceeding 30 M appeared to have sextuply alkylated cysteine residues, indicating that one of the seven cysteine residues of OxyR2 had a single modification, such as an oxidation, that could prevent alkylation by AMS. When determined based on the band intensities, relative amounts of the OxyR2 with the single cysteine modification (or sextuply alkylated) were increased gradually by exposure to higher levels of H 2 O 2 ( Fig. 2A, top).
To examine whether the OxyR2 with a single cysteine modification observed in the cells exposed to H 2 O 2 exceeding 30 M is the OxyR2 with overoxidized Cys 206 , the total cellular proteins (7.0 g/well) were resolved on non-reducing SDS-PAGE and immunoblotted with the anti-OxyR2-Cys 206 -SO 3 H antibody ( Fig. 2A, bottom). The OxyR2 protein with overoxidized Cys 206 was specifically detected in cells exposed to H 2 O 2 exceeding 30 M and increased gradually in the cells exposed to higher levels of H 2 O 2 . The results indicated that the sensing cysteine of OxyR2 becomes overoxidized in vivo when the V. vulnificus cells are exposed to high levels of H 2 O 2 , as was demonstrated in vitro ( Fig. 1).
Concurrently, to determine the specificity of the anti-OxyR2-Cys 206 -SO 3 H antibody by ELISA, the antibody was tested to react with the Cys 206 -SO 3 H or Cys 206 -SH peptides that were synthesized and attached to the microtiter 96-well plates. The antibody specifically bound to the Cys 206 -SO 3 H

FIGURE 2. Overoxidation of OxyR2 Cys 206 in V. vulnificus cells exposed to various levels of H 2 O 2 .
A, OH0703 (pDY1025) was grown anaerobically to an A 600 of 0.3 and exposed to various H 2 O 2 concentrations for 3 min. Cellular proteins were precipitated with TCA and alkylated with fresh AMS buffer for 1 h at 37°C. Proteins (3.5 g for the top panel and 7 g for the bottom panel) were resolved by non-reducing SDS-PAGE and immunoblotted using anti-OxyR2 antibody (top) or anti-OxyR2-Cys 206 -SO 3 H antibody (bottom). The predicted numbers of AMS that alkylated each OxyR2 molecule and their redox states are indicated at the ends of the gel. Negative control (NC) was OH0703 (pJH0311), 6AMS control was OH0703 (pBANG1416), and 7AMS control was OH0703 (pDY1025). B, the specificity of anti-OxyR2-Cys 206 -SO 3 H antibody to overoxidized and reduced OxyR2 was determined using ELISA. The microtiter 96-well plates were coated with 0.1 g of synthetic peptides corresponding to either OxyR2 active site with overoxidized (S-sulfonated) Cys 206 (EKEHC 206 (SO 3 H)LTEHA) (E) or with reduced Cys 206 (EKEHC 206 (SH)LTEHA) (•), and the peptides were reacted with various concentrations of the antibody as indicated. C, the S-sulfonated OxyR2 peptide (0.1 g) was attached to the microtiter 96-well plates and then reacted with 4 g of anti-OxyR2-Cys 206 -SO 3 H antibody. As a binding competitor, either S-sulfonated or reduced OxyR2 peptides (12.5 ng) were added to the reaction with the antibody as indicated. Black bar, control where no competitor was added. Relative binding of the antibody to the specific peptides is presented as A 450 . All data in B and C represent mean Ϯ S.D. (error bars). peptide (Fig. 2B). Furthermore, the binding of the antibody to the attached Cys 206 -SO 3 H peptide was effectively inhibited by the Cys 206 -SO 3 H peptide added to the reaction as a competitor (Fig. 2C). The antibody slightly bound to the Cys 206 -SH peptide (Fig. 2B), and the Cys 206 -SH peptide competitor could marginally inhibit the antibody reaction to the Cys 206 -SO 3 H peptide (Fig. 2C), reflecting that Cys 206 -SH peptide could be oxidized to Cys 206 -SO 3 H peptide in the aerobic conditions we tested. The results indicated that the anti-OxyR2-Cys 206 -SO 3 H antibody we used is specific to the OxyR2 with Cys 206 -SO 3 H. However, it was not possible to examine the antibody's specificity to Cys 206 -SO 2 H. Therefore, the OxyR2 with Cys 206 -SO 2/3 H is hereafter designated as the overoxidized state of OxyR2.
Overoxidized State of OxyR2 Turns Off the prx2 Promoter-To examine the transcriptional activity of the overoxidized state of OxyR2, the expression of prx2 was monitored in the wild-type V. vulnificus cells. The expression of prx2 was significantly induced when the anaerobically grown cells were exposed to 10 M H 2 O 2 (Fig. 3A), indicating that OxyR2 was oxidized to the disulfide state under this condition to activate the prx2 promoter (P prx2 ) ( Fig. 2A). However, the level of prx2 transcript gradually decreased along with increasing concentrations of H 2 O 2 exceeding 30 M (Fig. 3A), which was in accordance with the increase of the overoxidized state of OxyR2 ( Fig. 2A). The changes in prx2 expression were dependent on OxyR2 activity, but not on oxyR2 expression level, because the expression level of oxyR2 was not significantly affected by increasing concentrations of H 2 O 2 below 1 mM (Fig.  3A). The combined results suggested that the overoxidized state of OxyR2 deters further expression of prx2. Because it has been reported that Prx2 is inactivated by H 2 O 2 exceeding 30 M (17), this deterrence of the prx2 expression may prevent the worthless production of Prx2 in the environments of high H 2 O 2 , where Prx2 is no longer functional. However, it was still possible that H 2 O 2 over 30 M can also detrimentally affect other transcription factors, such as components of RNA polymerase, to deter the prx2 expression. To rule out this possibility, OH0703 containing pBANG1416 expressing OxyR2-C206D was grown aerobically without exogenously added H 2 O 2 , and the prx2 expression was compared. It has been reported that the size and the polar properties of aspartic acid are comparable with those of the overoxidized cysteine residues (Cys-SO 2 H and Cys-SO 3 H), and therefore OxyR2-C206D is anticipated to mimic the overoxidation state of OxyR2 (10). As shown in Fig. 3B, OH0703 producing wild-type OxyR2 expressed almost 13-fold greater prx2 than the negative control strain lacking OxyR2. This supported our previous observation that wild-type OxyR2 activates the prx2 expression under aerobic conditions (11). In contrast, the prx2 expression level of the strain producing OxyR2-C206D was almost the same as that of the negative control strain (Fig. 3B) (Figs. 1 and 2). So far in the widely accepted mechanism of OxyR proteins, OxyR2

Three-state Redox Switch OxyR2
JULY 29, 2016 • VOLUME 291 • NUMBER 31 shifts its redox state from the reduced state to the disulfide state upon reaction with H 2 O 2 and activates the expression of prx2. However, when exposed to H 2 O 2 exceeding the working range for Prx2, OxyR2 moves to an overoxidized state, a third redox state, and turns off the expression of prx2. It is obvious that OxyR2 is able to prevent production of useless Prx2 and thus save valuable cellular resources by working as a three-state redox switch (Fig. 3C).
Mechanism of the Overoxidized OxyR2 to Turn Off P prx2 -To gain insight into the mechanism by which the overoxidized state of OxyR2 deters the expression of prx2, the sequences for binding of the overoxidized OxyR2 were determined. Purified OxyR2 5CA -C206D, which mimics overoxidized OxyR2 and is soluble under aerobic conditions, was used in vitro for the DNA footprinting assay. OxyR2 5CA , in which all five non-catalytic cysteine residues in OxyR2 were replaced with alanine residues, was constructed previously (11). The purified OxyR2 5CA did not form oligomers in vitro under nonreducing conditions (11). As shown in Fig. 4, the sequences for binding of OxyR2 5CA -C206D were extended from Ϫ80 to Ϫ26 relative to the transcription start site of P prx2 . Therefore, one possible mechanism is that the overoxidized OxyR2 presumably turns off the expression of prx2 by masking the Ϫ35 region and thereby preventing RNA polymerase binding to P prx2 .
Furthermore, the binding sequences of OxyR2 5CA -C206D are identical to those of the reduced OxyR2. Enhanced cleavage in several nucleotides (Ϫ53 to Ϫ55), which was observed in the binding sequences of the reduced OxyR2, was also found in the binding sequences of OxyR2 5CA -C206D (Fig. 4B). It has been proposed that OxyR tetramer in the reduced state adopts an extended conformation (10) and binds to the elongated DNAbinding sites (11,12). The combined results suggest that the overoxidized state of OxyR2 turns off the activation of prx2 by adopting a conformation similar to that of the reduced state, binding to the elongated sequences, and preventing RNA polymerase binding to the promoter sequences.
Possible Overoxidation of OxyR1-To probe whether other OxyR proteins could function as a three-state redox switch, the expression of prx1 and katG in the V. vulnificus cells exposed to various levels of H 2 O 2 was determined (Fig. 5, A and B). It has been reported that OxyR1, a homolog of E. coli OxyR, directly activates the expression of prx1 and katG (11). 4 Expression of both antioxidant genes in V. vulnificus increased along with increasing levels of H 2 O 2 but decreased when H 2 O 2 levels exceeded certain levels; the expression levels of prx1 and katG showed the highest peaks in cells exposed to 30 and 10 M H 2 O 2 , respectively, and then decreased in cells exposed to higher levels of H 2 O 2 (Fig. 5, A and B). This variation of the prx1 and katG expression in response to increasing levels of H 2 O 2 was similar to that of the prx2 expression (Fig. 3A), indicating that OxyR1 was also probably overoxidized to turn off expression of prx1 and katG when the H 2 O 2 levels exceeded the working ranges for Prx1 and KatG. This result suggested that OxyR homologs of other bacteria could function as a three-state redox switch to tightly regulate their target antioxidant genes.

Discussion
In this study, we present evidence that V. vulnificus OxyR2 regulates the expression of prx2 as a three-state redox switch. When OxyR2 encountered H 2 O 2 exceeding the Prx2 working range, the sensing cysteine of OxyR2 was overoxidized to Cys 206 -SO 3 H in vivo as well as in vitro (Figs. 1 and 2). Although MS analysis was unable to detect the peptide fragment of OxyR2 with Cys 206 -SO 2 H (Fig. 1A), the S-sulfinated form of

Three-state Redox Switch OxyR2
In addition to the reduced and oxidized (disulfide) states, the overoxidized state is the third redox state of OxyR2, whose function was primarily characterized in this study. The overoxidized OxyR2 actively turns off the transcription of prx2 by binding to elongated sequences masking the Ϫ35 region of P prx2 in the environments of high H 2 O 2 , where Prx2 is no longer functional (Figs. 3 and 4) (17). Obviously, the overoxidized OxyR2 can prevent production of useless Prx2 and thus save valuable cellular resources. Therefore, the overoxidized OxyR2 is not simply an expired or perished form of OxyR2 but rather plays an important role in the survival of V. vulnificus during pathogenesis.
It was noteworthy that expression of prx1 and katG was increased, reached a maximum, and then was decreased by exposure of cells to increasing levels of H 2 O 2 in a pattern similar to that of prx2 (Fig. 5, A and B). Both prx1 and katG are regulated by OxyR1, a homolog of E. coli OxyR (11), implying that OxyR2 is not the only OxyR regulating antioxidant genes as a three-state redox switch. OxyR has been primarily thought of as a two-state redox switch that turns the antioxidant genes off and on in an H 2 O 2 -dependent manner (7)(8)(9). When bacteria encounter manageable ranges of H 2 O 2 , OxyR turns on its target genes encoding antioxidants, such as Prxs and catalases. Scavenging the H 2 O 2 using the antioxidants and maintaining metabolic activities to keep growing is beneficial for bacteria. Accordingly, expression of most of the OxyR regulon is governed by the housekeeping factor RpoD ( 70 ) (20). However, when H 2 O 2 exceeds the manageable ranges, maintaining expressions of the antioxidant genes would be worthless because growth and proliferation of bacterial cells under this condition could be reckless. Instead, bacteria rather express many stress tolerance genes governed by the stress-responsive factor RpoS ( 38 ) and shift their global physiology to stationary (or dormant) phase (21,22). Bacteria can turn off the antioxidant genes using OxyR, as a three-state redox switch, under conditions in which expression of the antioxidants is useless. It is obviously advantageous that bacteria utilize more cellular resources to express stress tolerance genes encoding many additional protective systems (20,21).
The mechanism for the overoxidation of the sensing cysteine remains unclear. However, structural analysis of the Pseudomonas aeruginosa OxyR-C199D mutant, in which the sensing cysteine residue is replaced with aspartic acid, gives us a hint as to this mechanism (10). The crystal structure OxyR-C199D holds a H 2 O 2 molecule near aspartic acid that is even bulkier than a cysteine residue, implying that H 2 O 2 can also bind to the same site even after the sensing cysteine is oxidized to Cys-SOH or Cys-SO 2 H. The bound H 2 O 2 can eventually oxidize the cysteine to Cys-SO 3 H by the same H 2 O 2 -driven oxidation mechanism proposed previously (10). We modeled Cys-SOH, Cys-SO 2 H, and Cys-SO 3 H forms based on the P. aeruginosa OxyR-C199D structure that contains H 2 O 2 molecules (Fig. 6A). Steric clash was not observed around the modeled cysteine residues with modification. Moreover, the lone pair electrons of the sulfur atoms in the cysteine residues could be facing H 2 O 2 within a reasonable distance for nucleophilic attack on the H 2 O 2 molecule when the -SOH or -SO 2 H moiety is simply rotated in the model. These findings support the idea that overoxidation at the sensing cysteine residue by H 2 O 2 can occur faster than at other cysteine residues, presumably by the successive H 2 O 2 -driven oxidation mechanism.
Three different redox states of OxyR are determined, depending on the cellular levels of H 2 O 2 as depicted in Fig. 6B. At low levels of H 2 O 2 below the sensing level of OxyR, OxyR is maintained in the reduced state. Under the Prx2 working range of H 2 O 2 , a kinetic path renders the sensing cysteine in OxyR oxidized to the Cys-SOH intermediate that forms a disulfide bond with the other redox-sensitive cysteine, leading to activation of the target antioxidant genes. However, when H 2 O 2 levels exceed the Prx2 working range, the Cys-SOH intermediate would be rapidly overoxidized to the Cys-SO 3 H via Cys-SO 2 H by H 2 O 2 before making the disulfide bond. The resulting overoxidized OxyR2 turns off the production of useless Prx2. We do not yet know what happens to the overoxidized OxyR when the cellular H 2 O 2 level returns to the working range. One plausible mechanism is that the overoxidized OxyR is degraded and new OxyR is synthesized, so that the cell can constantly respond to the changing environment. Given that the cellular amount of OxyR2 is very low (11), the energy needed for this process is not high.
In summary, V. vulnificus OxyR2 is primarily characterized as a three-state redox switch. A thiol-based sensor protein OxyR2 senses H 2 O 2 and shifts to a third redox state, the overoxidized state, as a result of overoxidation of sensing cysteine residue by H 2 O 2 . Overoxidized OxyR2 prevents production of Prx2 where Prx2 is no longer functional and thereby saves val- Wild-type V. vulnificus was grown anaerobically to an A 600 of 0.3 and exposed to various concentrations of H 2 O 2 as indicated for 3 min. Total RNAs were isolated, and the relative levels of prx1 (A) and katG (B) transcripts were determined by quantitative real-time PCR analyses. The level of each transcript from wild type unexposed to H 2 O 2 is presented as 1. *, statistically significant difference (p Ͻ 0.05) between groups. All data represent mean Ϯ S.D. (error bars).
uable cellular resources. The DNA sequences for binding of OxyR2 5CA -C206D suggested that overoxidized OxyR2 adopts an extended conformation and turns off P prx2 by binding to elongated sequences masking the Ϫ35 region. Expression patterns of prx1 and katG in response to various H 2 O 2 levels suggested that OxyR1, a homolog of E. coli OxyR, also functions as a three-state redox switch, as does OxyR2.

Experimental Procedures
Bacterial Strains, Plasmids, and Culture Media-The strains and plasmids used in this study are listed in Table 1. The V. vulnificus strains were grown in LB medium supplemented with 2.0% (w/v) NaCl (LBS) at 30°C. Anaerobic conditions were obtained using an anaerobic chamber with an atmosphere of 90% N 2 , 5% CO 2 , and 5% H 2 (Coy Laboratory Products, Grass Lake, MI). For anaerobic culture, the media were preincubated to remove dissolved O 2 in the anaerobic chamber, which was verified by adding 0.00001% (w/v) resazurin salt (Sigma) to the media as described previously (11).
Site-directed Mutagenesis of oxyR2 and Purification of Mutant OxyR2 Proteins-A mutant OxyR2 in which Cys 206 was replaced with aspartic acid was constructed using the QuikChange site-directed mutagenesis kit (Agilent Technologies, Loveland, CO) as described previously (11). The complementary mutagenic primers listed in Table 2 were used in conjunction with the plasmid pDY1025 (oxyR2 cloned into a broad host-range vector pJH0311 under the lac promoter) to create pBANG1416 (oxyR2-C206D on pJH0311) ( Table 1). The mutation was confirmed by DNA sequencing. E. coli SM10 pir, tra (23) harboring pJH0311, pDY1025, or pBANG1416 was used as a conjugal donor to the oxyR2 mutant (OH0703). The conjugation was conducted as described previously (11).
MALDI-TOF MS Analysis-To investigate the redox state of Cys 206 in OxyR2 in vitro, the full-length His 6 -tagged OxyR2 and OxyR2-C206S proteins were analyzed by MALDI-TOF MS, as described previously (11,17). Briefly, to reduce the proteins, 20 g of proteins in a reaction buffer (20 mM Tris (pH 7.4), 0.3 M KCl, 5 mM MgCl 2 , 0.5 mM EDTA, and 10% (v/v) glycerol) were treated with 100 mM DTT for 1 h, and then the DTT was removed by gel filtration chromatography under anaerobic conditions. To prepare oxidized proteins, reduced OxyR2 proteins were reacted with 500 M H 2 O 2 for 10 min after the DTT removal. Subsequently, the reduced cysteine residues of each reduced and overoxidized OxyR2 protein were alkylated with 50 mM iodoacetamide for 1.5 h in the dark under anaerobic conditions. Alkylated OxyR2 proteins were resolved on nonreducing SDS-PAGE, and the protein bands were excised and in-gel digested with trypsin (Sigma). Peptides were extracted from the gel pieces with 0.1% trifluoroacetic acid in 50% aceto-nitrile, concentrated to volumes of 10 l using a SpeedVac concentrator (Savant Instruments Inc., Farmingdale, NY), and desalted using Zip-Tip C 18 reverse phase peptide separation matrix (Millipore, Billerica, MA).
MALDI-TOF MS analyses were carried out on a Voyager-DE TM STR biospectrometry work station (Applied Biosystems Inc., Foster City, CA) operating in positive and negative ion reflector modes. ␣-Cyano-4-hydroxycinnamic acid (10 mg/ml) in 50% acetonitrile and 0.1% trifluoroacetic acid was used as a matrix. The theoretical monoisotopic masses ([M Ϫ H] Ϫ for the deprotonated form in the negative ion reflector mode or [M ϩ H] ϩ for the protonated form in the positive ion reflector mode) of the cleavage peptides were determined using Peptide-Mass software from the ExPASy proteomics server. The masses from the positive ion reflector mode were calibrated internally with known masses of autolytic trypsin peptides. However, the masses from the negative ion reflector mode were uncalibrated because the trypsin peaks were not detectable in the negative ion mode.
In Vivo Alkylation of OxyR2 and Western Blotting Analysis-The oxyR2 mutant OH0703 with pDY1025 expressing oxyR2 was used for Western blotting analysis of OxyR2 as described previously (11). OH0703 (pDY1025) was grown anaerobically to an A 600 of 0.3, aliquoted to the same volume, and exposed to various concentrations of H 2 O 2 . To alkylate free thiols in the proteins with AMS, the cells were immediately precipitated with ice-cold TCA, and then the resulting pellets were dissolved in 50 l of the fresh AMS buffer (15 mM AMS, 1 M Tris, 1 mM EDTA, 0.1% (w/v) SDS, pH 8.0) (11).
After incubation at 37°C for 1 h, the same amounts of pelleted total protein were resolved on SDS-PAGE under nonreducing conditions and immunoblotted with either anti-OxyR2 or anti-OxyR2-Cys 206 -SO 3 H antibody as described previously (11). The anti-OxyR2 polyclonal antibody was prepared previously in a rabbit by using the purified His 6 -tagged OxyR2 as an antigen (11). An S-sulfonated peptide corresponding to the active site of OxyR2, EKEHC 206 (SO 3 H)LTEHA, was synthesized and conjugated with keyhole limpet hemocyanin and then used to raise rabbit anti-OxyR2-Cys 206 -SO 3 H polyclonal antibody (AbFrontier, Seoul, South Korea).