The MarR family is unique to both bacteria and archaea. The members of this family, one of the most prevalent families of transcriptional regulators in bacteria, enable bacteria to adapt to changing environmental conditions, such as the presence of antibiotics, toxic chemicals, or reactive oxygen species (ROS), mainly by thiol-disulfide switches. Although the genome of Corynebacterium glutamicum encodes a large number of the putative MarR-type transcriptional regulators, their physiological and biochemical functions have so far been limited to only two proteins, regulator of oxidative stress response RosR and quinone oxidoreductase regulator QosR. Here, we report that the ncgl2617 gene (cosR) of C. glutamicum encoding an MarR-type transcriptional regulator plays an important role in oxidative stress resistance. The cosR null mutant is found to be more resistant to various oxidants and antibiotics, accompanied by a decrease in ROS production and protein carbonylation levels under various stresses. Protein biochemical function analysis shows that two Cys residues presenting at 49 and 62 sites in CosR are redox-active. They form intermolecular disulfide bonds in CosR under oxidative stress. This CosR oxidation leads to its dissociation from promoter DNA, depression of the target DNA, and increased oxidative stress resistance of C. glutamicum. Together, the results reveal that CosR is a redox-sensitive regulator that senses peroxide stress to mediate oxidative stress resistance in C. glutamicum.
Corynebacterium glutamicum is a high G+C-content aerobic Gram-positive bacterium. It is not only a workhorse in biotechnology for glutamate and lysine amino acid production, which are used as nutritive additives in food and feed, but also a model organism of systems biology for the investigation of its pathogenic relatives, Corynebacterium diphtheriae, Corynebacterium jeikeium, and mycobacterium [1–3]. C. glutamicum cells are frequently confronted with excessive reactive oxygen species (ROS) production, triggering sudden changes of fermentative conditions in temperature, pH, osmotic pressure, or toxic compounds [4–5]. The elevated ROS levels can deplete the cellular thiol pool and then lead to oxidative stress, potentially damaging DNA, lipids, and proteins . Thus, the ability of C. glutamicum to combat ROS challenges and maintain the cellular redox balance is crucial for cell survival and improves fermentation efficiency. As previously reported, C. glutamicum can robustly survive in the hostile environment of the fermenting process in that it possesses intrinsic resistance [7–11]. This intrinsic resistance results from thiol-based redox-sensitive transcriptional regulators, including the LysR-family regulator OxyR, the multiple antibiotic resistance (MarR)-family RosR and QorR, and stress-responsive ECF sigma factor SigH. These transcriptional regulators specifically sense and respond to changing conditions and then rapidly adjust their own biological activity to control the oxidant-prevailing regulon expression with thiol groups of Cys, mainly by the classic disulfide-based mechanism.
The MarR family, widely distributed among bacteria, is involved in the control of virulence factor, the response to environmental stress, or the regulation of aromatic compound catabolism, especially prominent in giving bacteria adaptation to environmental stress, such as antibiotics, toxic chemical, and ROS [12–16]. The classic characteristic of this family is a DNA-binding domain comprising a winged helix-turn-helix motif . Generally, the MarR-type regulators function as transcriptional repressors by binding to the promoter region of a target regulon under physiological conditions and then occluding RNA polymerase binding, whereas they form intermolecular disulfide bond-containing homodimers under ROS-inducing oxidative stress causing a release from the promoter . However, Hillion and Antelmann  report that there are redox-sensing model differences among the MarR-type subfamily from different bacteria. Bacillus subtilis OhrR and Staphylococcus aureus SarZ, belonging to the one Cys-type MarR/OhrR family, are oxidized by Cumene-OOH (CHP) to the sulfenic acid that still retains DNA-binding activity and reacts further to a mixed disulfide with BSH (S-bacillithiolated OhrR) or benzene thiol (S-thiolated SarZ) causing dissociation in response to oxidative stress [18,19]. In contrast, the one Cys-type OhrR homologs MgrA in the human pathogen S. aureus can be oxidized by CHP to Cys-SOH that leads to dissociation of MgrA from the operator DNA and induction of resistance in S. aureus in vivo . The quinone-sensing MarR/DUF24-family regulator QsrR in S. aureus shows that the conserved Cys5 of QsrR senses quinones by a thiol-S-alkylation model , while C. glutamicum QorR senses quinones by intersubunit disulfide bond formation [22,23]. More recently, Lee et al.  found that the prototypical regulator YodB of the MarR/DUF24 family from B. subtilis uses two distinct pathways to regulate transcription in response to two reactive electrophilic species (diamide or methyl-p-benzoquinone). Diamide stress induces obvious structural changes in the YodB dimer by promoting the formation of disulfide bonds between cysteine residues, whereas methyl-p-benzoquinone forms an adduct on a specific cysteine residue with little effect on the YodB dimer. Thus, the exact different molecular mechanism of multiple oxidants differing simultaneously by the MarR-type regulator or of the same oxidant sensing by the same MarR-type regulators in various bacteria is still a subject of active investigation.
The MarR-type family is one of the most prevalent families of transcriptional regulators in C. glutamicum, many members of which have been annotated as a putative transcriptional regulator. To the best of our knowledge, only two of the MarR-type regulators in C. glutamicum have been characterized as oxidative stress sensors: H2O2 (hydrogen peroxide)-sensing RosR (encoded by ncgl1124) by intramolecular disulfide model and quinone-sensing QorR (encoded by ncgl1317) by an intermolecular disulfide model. In fact, the regulation mechanisms and functions for most MarR family proteins are unknown in C. glutamicum. Thus, we here report the findings of the MarR-type regulator encoded by the ncgl2617 gene. Our results indicate that the NCgl2617 protein exerts a regulatory effect on oxidative stress response and antibiotic resistance in C. glutamicum. NCgl2617 protein is found to possess the winged helix DNA-binding motif and the two cysteine residues present in MarR family members.
Bacterial strains and culture conditions
Bacterial strains, plasmids and primers used in the present study are listed in Supplementary Tables S1 and S2. Escherichia coli and C. glutamicum are cultured at 37°C or 30°C in Luria-Bertani (LB) broth aerobically on a rotary shaker (220 rpm) or LB plates. Mineral salts medium (MM) containing 100 mM glucose is used for growth assay under different concentrations of various peroxides and for in vivo CHP-induced CosR structural analysis . For generating and maintaining C. glutamicum mutants, brain–heart broth containing 0.5 M sorbitol (BHIS) medium is used . To prepare ΔcosR in-frame deletion mutants, the suicide plasmid pK18mobsacB is used to construct the knockout plasmids pK18mobsacB-ΔcosR. The resulting pK18mobsacB-ΔcosR are transformed into C. glutamicum wild type (WT) through electroporation to carry out single crossover. The transconjugants are selected on LB agar medium containing 40 µg/ml nalidixic acid and 25 µg/ml kanamycin. Counter-selection for markerless in-frame deletion is performed on LB medium agar plates with 40 µg/ml nalidixic acid and 20% sucrose [26–28]. Strains growing on this plate are tested for kanamycin sensitivity (KmS) by parallel picking on 40 µg/ml nalidixic acid-containing LB plates added with either 25 µg/ml kanamycin or 20% sucrose. Sucrose-resistant and kanamycin-sensitive strains are tested for deletion by PCR using the DCosR-F1/DCosR-R2 primer pair (Supplementary Table S2). For complementation, the pXMJ19 derivatives are transformed into ΔcosR mutants by electroporation and the expression is induced by adding 0.5 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) into the medium. All chemicals were of analytical reagent grade purity or higher. Antibiotics are added at the following concentrations: kanamycin, 50 µg/ml for E. coli and 25 µg/ml for C. glutamicum; nalidixic acid, 40 µg/ml for C. glutamicum; chloramphenicol, 20 µg/ml for E. coli and 10 µg/ml for C. glutamicum.
To obtain the expression plasmid, OCosR-F/OCosR-R primer pair and C. glutamicum genomic DNA are used to amplify the gene coding for CosR (NCgl2617) by PCR. The amplified DNA fragments are digested and then cloned into digested pET28a plasmid, yielding the corresponding pET28a-cosR.
The suicide plasmid pK18mobsacB-ΔcosR is prepared using overlap PCR to construct the ΔcosR deletion mutants . Briefly, primer pairs DCosR-F1/DCosR-R1 and DCosR-F2/DCosR-R2 are used to amplify the 775-bp upstream fragment and the 818-bp downstream fragment of cosR, respectively. The primer pair DCosR-F1/DCosR-R2 is used to fuse the upstream and downstream fragments by overlap PCR . The obtained PCR products are digested with HindIII and EcoRI and clouded into similarly digested pK18mobsacB  to produce pK18mobsacB-ΔcosR.
To prepare pXMJ19-cosR or pXMJ19-His6-cosR, primer pair CCosR-F/CCosR-R is used to amplify the cosR gene fragment from C. glutamicum genomic DNA. The obtained DNA fragments are digested and cloned into similar digested pXMJ19 or pXMJ19-His6 vector.
To make the cysteine residue at position 49 of CosR into a serine residue (CosR:C49S), site-directed mutagenesis is performed . Briefly, to obtain the mutant cosR:C49S DNA segments, two rounds of PCR are used. In the first round of PCR, primer pairs DCosR-F1/CosR-C49S-R and CosR-C49S-F/DCosR-R2 are used to amplify segments 1 and 2, respectively. The second round of PCR is carried out using CosR-F/CosR-R as primers and fragment 1 and fragment 2 as templates to get the cosR:C49S fragment. The cosR:C49S DNA fragment is digested and cloned into digested pXMJ19, pXMJ19-His6, or pET28a plasmid, obtaining pXMJ19-cosR:C49S, pXMJ19-His6-cosR:C49S, or pET28a-cosR:C49S. The cosR:C62S fragments are gained using a similar procedure as described above and cloned into plasmid pXMJ19, pXMJ19-His6, or pET28a, obtaining pXMJ19-cosR:C62S, pXMJ19-His6-cosR:C62S, or pET28a-cosR:C62S.
The fidelity of all constructs was confirmed by DNA sequencing (Sangon Biotech, Shanghai, China).
To measure the response to various peroxides, stationary-phase C. glutamicum strains grown in LB broth medium at 30°C are diluted 100-fold into fresh LB broth medium and treated with 100 mM H2O2, 50 µM diamide, 1 mM hypochlorous acid (HClO), 11 mM CHP, and 20 mM tert-butyl hydroperoxide (t-BHP), respectively, at 30°C for 30 min at 100 rpm. After treatment, the cultures are diluted 1000-fold and plated onto LB agar plates, and colonies are counted after 36 h growth at 30°C. The survival percentage is calculated by dividing the number of CFU of stressed cells by the number of CFU of cells without stress, as recently reported [28,30]. All these assays are performed in triplicate.
Overexpression and purification of recombinant protein
To express and purify soluble glutathione S-transferase (GST) — and His6-tagged CosR and its variants proteins, the recombinant pGEX6p-1 and pET28a derivatives are transformed into BL21(DE3) host strains. Bacteria are cultured at 37°C in LB medium to an OD600 of 0.5, shifted to 22°C and induced with 0.5 mM IPTG, and then cultivated for an additional 12 h at 22°C. Harvested cells are sonicated and proteins are purified with the His · Bind Ni-NTA resin or the GST · Bind resin (Novagen, Madison, WI, U.S.A.) according to manufacturer's instructions. Eluted recombinant proteins are dialyzed against PBS at 4°C.
Analysis of sulfenic acid and disulfide bond formation
The formation of Cys-SOH in CosR variants is measured by the 4-chloro-7-nitrobenzofurazan (NBD-Cl) labeling assay [31–33]. After proteins are pretreated with 50 mM DTT (dithiothreitol) for 30 min, the remaining DTT is removed by ultrafiltration with Ultracel-30KD membrane-containing Amicon Ultra-4 (Millipore, MA, U.S.A.). Proteins in buffer (50 mM potassium phosphate with 1 mM EDTA at pH7.0) are prepared anaerobically by repeatedly flushing with argon gas and vacuum in alternating cycles for 20 min. An anaerobic solution of NBD-Cl (25 mM in DMSO) is prepared by bubbling argon for 10 min. Under anaerobic conditions, 50 µM proteins are treated with various concentrations of CHP or DTT (negative control). The resulting proteins are treated with 5 mM NBD-Cl at 25°C for 30 min in the dark. Remaining NBD-Cl is removed by ultrafiltration, and the absorbance of proteins are analyzed (200–600 nm) on a Beckman DU 7500 diode array spectrophotometer (Fullerton, CA, U.S.A.).
The formation of disulfide bond and sulfenic acid in CosR WT and its variants are performed with the thiol-reactive probe 4-acetamido-4′maleimidyldystilbene-2, 2′-disulfonic acid (AMS; Molecular Probes, Eugene, OR, U.S.A.) covalent modification method . Briefly, 25 µM proteins are treated with 10 mM DTT at room temperature for 30 min to obtain reduced CosR WT and its variants, and remaining DTT is removed by ultrafiltration. The resulting proteins are treated with 50 µM CHP at room temperature for 30 min. DTT-treated proteins are directly used as a negative control. The DTT- and CHP-treated proteins are precipitated with 10% (w/v) trichloroacetic acid (TCA) and then re-dissolved with 100 mM Tris–HCl (pH 8.0) including 1% SDS and 15 mM AMS. After reaction for 30 min in the dark, 2 µg proteins are subjected by non-reducing SDS–PAGE.
Quantitative analysis of sulfhydryl groups
Free sulfhydryl groups are measured using 5,5′-dithio-bis (2-nitrobenzoic acid) (DTNB) . After CosR WT and its variants (50 µM) are treated with 200 µM CHP and 50 mM DTT at room temperature for 30 min, respectively, a PD10 desalting column (GE Healthcare, Piscataway, NJ, U.S.A.) is used to remove residual CHP or DTT. The resulting proteins (10 µM) are cultured with 2 mM DTNB in 50 mM Tris–HCl buffer (pH 8.0) and the absorbance at 412 nm is measured against a 2 mM DTNB solution as the reference. The amounts of reactive sulfhydryl groups are measured using the molar absorption coefficient of TNB at 412 nm (ε412) of 13 600 M−1 cm−1 .
MALDI-TOF MS–MS analysis
DTT-treated CosR and CHP-treated CosR are incubated with 10 mM iodoacetamide for 15 min at room temperature. The alkylated proteins are subjected to non-reducing SDS–PAGE and stained with Coomassie Brilliant Blue (CBB). After staining, bands are excised, digested with trypsin, and analyzed by matrix-assisted laser desorption ionization time-of-flight tandem mass spectrometry (MALDI-TOF MS–MS) (Voyager-DE STR, Applied Biosystems).
Fluorescence dye-based intracellular ROS detection
ROS levels are detected using the fluorogenic probe 2′,7′-dichlorofluorescein diacetate (H2DCFDA), as described by Si et al. .
Oxidative stress-dependent structural change of CosR in vivo
The CHP-dependent structural change of CosR and its variants in vivo are determined by a recently reported method . ΔcosR(pXMJ19-His6-cosR), ΔcosR(pXMJ19-His6-cosR:C49S), and ΔcosR(pXMJ19-His6-cosR:C62S) strains are cultured in MM containing 100 mM glucose at 30°C. Bacterial cells are grown to mid-exponential phase and split into 100 ml aliquots for CHP treatment (1.5–10 mM, 10 min). The treated samples are harvested immediately by centrifugation. For recovery, the 10 mM CHP-treated samples are again inoculated into fresh MM containing 25 µg/ml streptomycin at 30°C for 2 h and then harvested. Crude cell lysates are subjected to non-reducing SDS–PAGE, and the structural properties of CosR and its variants are visualized by immunoblotting using the anti-CosR antibody.
β-Galactosidase activities are assayed with o-nitrophenyl-β-d-galactopyranoside (ONPG) as the substrate .
Electrophoretic mobility shift assay
Electrophoretic mobility shift assay (EMSA) was performed using the method of Si et al. . To reduce nonspecific binding in the EMSA assay, a DNA promoter (Pmpx; 300 bp) containing the predicted CosR-binding site is amplified from the mpx (mycothiol peroxidase) promoter region using primers PMPx-F2/PMPx-R. Increasing concentrations of purified CosR (0–4 µg) are incubated with 20 ng of mpx DNA promoter in EMSA buffer containing 20 mM Tris–HCl, pH 7.4, 4 mM MgCl2, 100 mM NaCl, and 10% glycerol. The binding reaction mixture is incubated at room temperature for 30 min, subjected to electrophoresis on a 6% native polyacrylamide gel with 5% glycerol in 0.5× TBE electrophoresis buffer, and then detected with SYBR Green. A 300 bp fragment from the mpx-coding region amplified with primers Control-F and Control-R instead of the 300 bp mpx promoter and BSA (bovine serum albumin) instead of CosR in the binding assays are used to act as negative controls.
Western blot analysis
Western blot analysis is performed as recently described . Samples subjected on SDS–PAGE are transferred onto polyvinylidene fluoride membranes. After blocking with 5% (w/v) nonfat milk powder for 2 h at room temperature, membranes are incubated with the primary antibody at 4°C overnight: anti-CosR (Ncgl2617) rabbit polyclonal antibody, 1 : 1000; anti-RNA polβ, 1 : 6000; anti-His (Millipore, MA, U.S.A.), 1 : 1000. The membrane is washed three times in TBST buffer [50 mM Tris, 150 mM NaCl, 0.05% (v/v) Tween 20, pH 7.4] and incubated with 1 : 5000 dilution of horseradish peroxidase-conjugated secondary antibodies (Shanghai Genomics) for 1 h. The protein bands are visualized with the ECL plus kit (GE Healthcare, Piscataway, NJ, U.S.A.). The RNA polβ antisera were made in our recent study . The purified His6-tagged CosR and MPx are used to generate rabbit anti-CosR and anti-MPx polyclonal antibodies, and the resulting antiserum is affinity-purified against the same proteins. The density of bands on Western blots is quantified by Image Lab (Bio-Rad, U.S.A.).
Quantitative RT-PCR analysis
Quantitative RT-PCR analysis (7500 Fast Real-Time PCR; Applied Biosystems, Foster City, CA, U.S.A.) is performed as recently described . To obtain standardization of results, the relative abundance of 16S rRNA is used as the internal standard.
Coupled Trx/TrxR and Mrx1/MSH/Mtr electron transfer assays
Oxidized CosR is reduced by the Trx (thioredoxin)/TrxR (thioredoxin reductase), Mrx1 (mycoredoxin 1)/MSH (mycothiol)/Mtr (mycothione reductase), and MSH coupled enzyme assays described by Van Laer et al.  with slight modification. Briefly, oxidized CosR is performed by 1 mM CHP for 30 min at room temperature and excess CHP is removed by ultrafiltration. The oxidized CosR/Trx/TrxR electron transfer pathway is reconstructed by incubating 500 µM NADPH, 15 µM TrxR, 5 µM Trx, and 50 µM CosR at 37°C in the assay buffer containing 50 mM HEPES (pH 8.0) and 500 mM NaCl. The oxidation of NADPH due to oxidized CosR reduction is monitored as a decrease in the absorption at 340 nm, measured as a function of time. The oxidized CosR/Mrx1/MSH/Mtr assay contains 500 µM NADPH, 5 µM Mtr, 400 µM MSH, 2 µM Mrx1, and 50 µM CosR. The oxidized CosR/MSH/Mtr assay contains 500 µM NADPH, 5 µM Mtr, 400 µM MSH, and 50 µM CosR. After 5 min pre-incubation of the other components in the reaction mixture, 50 µM oxidized CosR is added to start the reaction and the absorption at 340 nm is measured as a function of time.
MSH is purified from C. glutamicum with thiopropyl sepharose 6B followed by Sephadex LH-20 chromatography as recently described . The concentration of purified MSH is determined using the thiol-specific fluorescent-labeling HPLC method with commercial glutathione (GSH) as the thiol standard reference.
Bacterial two-hybrid assay
Bacterial two-hybrid complementation assays were carried out as recently described . Efficiencies of interactions between different hybrid proteins were quantified by the measurement of β-galactosidase activity of overnight cultures (OD600 = 1.5) grown in LB broth at 37°C. The overnight cultures (50 µl) were permeabilized with 420 µl of Z buffer containing 20 µl of chloroform and 10 µl of 0.1% SDS at 30°C for 1 h, and β-galactosidase activities were assayed with ONPG as the substrate as described . All the experiments were performed at least three times.
GST pull-down assay
The GST pull-down assay was performed as previously described . Retained proteins were detected by immunoblot after SDS–PAGE using the anti-His antibody (Millipore, MA, U.S.A.).
Statistical analyses of survival rate, ROS level, enzyme activity, and transcription level are determined with paired two-tailed Student's t-test. GraphPad Prism Software is used to carry out statistical analyses (GraphPad Software, San Diego, CA, U.S.A.).
Results and discussion
Involvement of cosR in peroxide resistance
The gene of C. glutamicum ATCC 13032 (ncgl2617) coding for a putative transcriptional regulator of the MarR family has been annotated, which composed of 447 bp and encoded a protein of 148 amino acids with a theoretical molecular mass of 16.8 kDa . Previously, Wilkinson and Grove  reported that many of the MarR (multiple antibiotic resistance regulator) family of prokaryotic transcriptional regulators were critical for control of bacterial response to peroxide-caused stresses. This prompted us to construct a mutant strain with deletions for the ncgl2617 gene and examine whether the Ncgl2617 plays a role in protection against peroxide-caused stress. Here, we found that the C. glutamicum ncgl2617 mutant showed increased resistance compared with the WT and the complement strain upon H2O2, diamide, CHP, and t-BHP challenge (Figure 1A), which was consistent with the role played by MarR-type transcriptional regulators MexR, AsrR, MosR, and OhrR in Pseudomonas aeruginosa, Enterococcus faecium, Mycobacterium tuberculosis, Mycobacterium smegmatis, and Streptomyces avermitilis under oxidative stress [43–48]. We then postulated that the ncgl2617 gene would influence growth of C. glutamicum under peroxide conditions. This prediction was confirmed by analyzing the growth of the C. glutamicum WT, ncgl2617 mutants and complement strains in MM medium supplemented with CHP and H2O2. The levels of growth of all tested strains were nearly identical in MM medium (Supplementary Figure S1A). The growth of the WT and complement strains was moderately impaired in MM medium under 30 mM H2O2 and 0.5 mM CHP, whereas the growth of the ncgl2617 mutants was significantly unaffected after a 1 or 3 h lag period (Supplementary Figure S1B,D). H2O2 (60 mM) and CHP (1.0 mM) severely inhibited the growth of the WT and complement strains in comparison with that of the ncgl2617 mutants (Supplementary Figure S1C,E). The physiological data indicated that Ncgl2617 exhibited peroxide-resistant phenotype. Thus, this protein is named CosR (C. glutamicumoxidant-sensing regulator).
CosR was involved in oxidative stress resistance.
Environmental stimuli induce the production of more harmful ROS. To examine the effect of CosR on ROS reduction upon peroxide stress, we assessed the intracellular ROS levels in C. glutamicum WT and mutant strains challenged with H2O2, CHP, and t-BHP, respectively, by using the fluorogenic probe H2DCFDA. As shown in Figure 1B, mutants lacking cosR gene (cosR mutant with empty pXMJ19, ΔcosR) had significantly lower ROS levels than the WT strains (WT C. glutamicum with empty plasmid pXMJ19, WT) and the complement strains (ΔcosR was complemented with plasmids carrying the WT cosR gene, ΔcosR+) after exposure to H2O2, CHP, and t-BHP, respectively, indicating that CosR was critical in reducing ROS accumulated in C. glutamicum under peroxide stress conditions (Figure 1B).
ROS escaping from the antioxidant defense system are more apt to react with the cysteine thiol groups of proteins, which results in reversible inter- or intra-protein disulfides (PrSSPr and PrSSPr), and mixed disulfides with LMW thiols, irreversible sulfoxidation products and carbonylation [49,50]. Since CosR is involved in reducing ROS levels, it was assumed that CosR might also play a role in protecting protein from carbonylation damage under oxidative stress conditions. To test this assumption, carbonyl groups in the total proteins isolated from the WT and ΔcosR mutant grown under oxidative stress were derivatized with 2,4-dinitrophenylhydrazine (DNPH) and detected by Western blotting with anti-DNPH antibody. As shown in Figure 1C, ΔcosR mutant showed significantly lower carbonyl contents than WT cells under H2O2, CHP and t-BHP treatment. This result demonstrates that CosR plays important roles in protecting C. glutamicum against oxidative stress via reducing intracellular ROS accumulation and protein carbonylation.
Dimer formation of CosR via intersubunit disulfide bonds upon peroxide oxidation in vitro
Many members of the MarR-type regulators, such as RosR, YodB, MexR, MosR, and 2-cys OhrR, exist as homodimers via intersubunit disulfide bonds upon oxidation [9,24,43,45,51–53]. The amino acid sequence showed that CosR contains two cysteine residues at positions 49 and 62 (Supplementary Figure S2). Thus, we thought it might share a similar oxidation-sensing mechanism in which CosR is oxidized to form intersubunit disulﬁde-containing dimer by a 2-Cys mechanism. Although non-reducing SDS–PAGE, native PAGE, and gel filtration chromatography showed that native CosR protein was monomeric with an apparent MW of ∼20 kDa (Supplementary Figure S3), corresponding well to the molecular mass of CosR deduced from its amino acid sequence, CosR incubated with CHP, t-BHP, and H2O2 migrated as a band of ∼40 kDa as judged by its behavior on 15% non-reducing SDS–PAGE, which corresponded to the dimeric form. The dimer formation was reversed by an excess of DTT (Figure 2A,B), suggesting that there was the formation of a disulfide bond between Cys49 and Cys62 in this protein. The formation of the disulfide bond in the t-BHP-treated CosR was further confirmed by MS analysis, with the identification of a mass of 3859.7 Da, lower by 2.01 Da than predicted for the sum of two peptides including the Cys49-containing 37–51peptide (calculated mass 1768.9 Da) and the Cys62-containing 62–80 peptide of CosR (calculated mass 2092.8 Da), indicating that the 37–51 and 62–80 peptides were cross-linked by a disulfide bond (Figure 2C,D), which was similar to the observation of Xanthomonas campestris OhrR that was a reversible disulfide bonding between the two subunits of the protein under oxidation treatment . This mass was not observed in the CosR with DTT and t-BHP treatment (Figure 2C).
Oxidative stress promoted reversible intersubunit disulﬁde-containing dimerization of CosR.
CHP-induced dimeric formation of CosR in vivo
To examine whether the dimeric formation of CosR was induced under CHP treatment in vivo, we treated cells of the ΔcosR+ (ΔcosR strains containing pXMJ19-His6 plasmids), ΔcosR+(C49S) (ΔcosR strains containing pXMJ19-His6-cosR:C49S plasmids), and ΔcosR+(C62S) (ΔcosR strains containing pXMJ19-His6-cosR:C62S plasmids) with various concentrations of CHP and probed the forms of CosR by immunoblotting with CosR antibody after non-reducing SDS–PAGE separation (Supplementary Figure S4A). Our results indicate that under normal conditions (no stress), CosR in ΔcosR+ strains existed as monomers and became intermolecular disulfide bond-containing dimers when exposed to certain CHP conditions (>7.5 mM CHP) (Figure 3A,B). For example, 10 mM CHP induced most of the proteins into dimer form (Figure 3A). However, whether under CHP treatment or not, CosR in ΔcosR+(C49S) and ΔcosR+(C62S) strains always existed in monomeric form (Figure 3B,C). Moreover, withdrawing CHP for 120 min led to a disappearance of CosR dimers formed in the ΔcosR+ cells treated with 10 mM CHP, which was accompanied with an increase in CosR monomers (Figure 3A, rightmost lane). These results indicate that CHP conferred reversible structural change of CosR, and Cys49 and Cys62 were responsible for the observed CosR reversible modification under CHP treatment.
Oxidative stress-dependent structural changes of relevant CosR in vivo.
Oxidation led to the generation of the sulfenic acid intermediate form of the sensing cysteine, C49
Previous studies have revealed a step-by-step molecular mechanism for peroxide-sensing transcription factor by which the conserved Cys is specifically oxidized by peroxide to form Cys-SOH, and Cys-SOH reacting with another Cys led to the disulfide bridge . Thus, we wondered whether oxidation could generate Cys-SOH in CosR. To reveal the state of Cys in CgohpsR under OHP treatment, the two mutants, C49S and C62S, with each Cys mutated to Ser, respectively, were constructed and used. Cys state was measured by using NBD-Cl sign, DTNB assay, and AMS modification.
NBD-Cl can exclusively react with thiol groups and sulfenic acids, but not with sulfinic or sulfonic forms. The covalent attachment of NBD-Cl generated an absorption peak at ∼420 nm upon reaction with thiol groups, whereas it peaked at ∼347 nm upon reaction with sulfenic acids [31–33]. By reaction with NBD-Cl, the cysteine thiol of the CosR:C49S under no exposure or exposure to different concentrations of CHP all formed adducts with NBD that absorb at 420 nm, (Figure 4A). DTNB assay showed the DTT-treated CosR:C49S contained one thiol per monomer, which was equal to thiol content of per CosR:C49S monomer after treating with CHP (Figure 4C). The redox state of thiol in CosR:C49S was further tested by detection of free thiol groups using AMS covalent modification. AMS covalently modified sulfhydryl groups of protein irreversibly. Since the molecular mass of AMS was 0.5 kDa, the apparent delay of electrophoretic mobility would occur in proportion to the number of free thiol groups in proteins . The CHP-treated, AMS-modified CosR:C49S migrated as same as their CHP-untreated, AMS-modified states (Figure 4D), indicating that CosR:C49S variants were not oxidized by CHP, of which Cys62 still existed in the thiol state after exposed to CHP.
The thiol content and form of DTT-treated or CHP-treated CosR.
The CosR:C62S variant (50 µM) with less than 125 µM CHP treatment showed soret band at 347 and 420 nm, indicating that the reaction of NBD-Cl with sulfenic acids and free thiol groups existed at the same time, and Cys49 was partly oxidized to a sulfenic acid form (Figure 4B). The 125 µM CHP-treated and NBD-labeled CosR:C62S variant had a specific absorbance maximum (λmax) of 347 nm, which clearly signified the detection and trapping of approximately stoichiometric amounts of SOH at Cys49. The decrease or the disappearance of this signal upon the treatment with 200 and 500 µM CHP, respectively, probably indicated that Cys49 was over-oxidized to sulfinic or sulfonic acid forms. Consistently, the DTT-treated CosR:C62S monomer showed one thiol, while close to 0 equivalent of free thiol groups was observed after treating with CHP, indicating that Cys49 was susceptible to oxidation (Figure 4B). AMS assay also showed that the electrophoretic migration state of the CHP-treated CosR:C62S protein modified with or without AMS was the same, and the DTT-treated CosR:C62S protein modified with AMS retarded mobility compared with the CHP-treated CosR:C62S protein modified with AMS (Figure 4D, lanes 3 and 4). These results suggest that the Cys49 residue is the peroxidative cysteine that can be oxidized to Cys-SOH.
Negative regulation of mpx by CosR
Previously, we showed that C. glutamicum mycothiol peroxidase MPx, similar to the glutathione peroxidase (Gpx), was resistant to and induced by organic and inorganic peroxides . Moreover, E. faecium gpx is regulated by MarR-type AsrR . Thus, we suggested C. glutamicum MPx was regulated by CosR. The lacZ activity of Pmpx::lacZ chromosomal promoter fusion reporter in relevant C. glutamicum strains and quantitative real-time PCR (qRT-PCR) profiling of mpx expression were quantitatively measured in bacterial cells either untreated or treated with different toxic agents of various concentrations (Figure 5A,B). Concentrations of CHP applied were able to reduce the growth rate but under sub-lethal concentrations (Supplementary Figure S5). As expected, high levels of the promoter lacZ activity of mpx were detected in the ΔcosR strain, regardless of whether or not CHP was present. Under normal conditions (without CHP treatment), the promoter lacZ activity of mpx in ΔcosR strain was 6.5 times higher than that of the WT strain, while ΔcosR+ had levels similar to WT, indicating that CosR acted as a transcriptional repressor of mpx expression. Under 1.2 and 2.0 mM CHP treatment, the promoter lacZ activity of mpx in WT went up by 5.4 and 6.2 times, respectively, almost reaching levels comparable with that of ΔcosR. The addition of CHP caused a mild decrease in promoter activity of mpx compared with that without CHP in ΔcosR strain, whereas it was still high when compared with WT without CHP treatment. A similar regulatory pattern of mpx by CosR was also observed at the mRNA transcriptional level by qRT-PCR analysis (Figure 5A,B). Survival rate assays also showed the ΔmpxΔcosR double mutants behaved identically to the mpx mutant that was less resistant than its parental strain when challenged with peroxides, and ΔmpxΔcosR double mutants complemented by the cosR WT gene were still sensitive to peroxides. In contrast, the introduction of mpx gene in ΔmpxΔcosR double mutants dramatically improved resistance to peroxides, reaching levels comparable with that of ΔcosR (Figure 5C). Further analysis at protein level indicated that similar regulation was observed for protein production of mpx, in which deletion of CosR increased its cellular level (Figure 5D and Supplementary Figure S4B). These data indicate that mpx was fully up-regulated in the ΔcosR mutant and that CosR negatively regulated the expression of mpx.
CosR-dependent expression of mpx.
Oxidation leads to dissociation of CosR from DNA
Conformational changes in the regulator were linked to its direct transcriptional activity. For example, most members of the MarR family in the DNA-binding activity were often reversibly inhibited by forming inter-monomer disulfide bonds under oxidation. Thus, we next tested whether the DNA-binding activity of oxidized CosR changed. To do so, oxidized CosR with a single disulfide between its active-site cysteines was prepared and was confirmed by DTNB assays (Figure 4C). EMSA was performed by the interaction of reduced (DTT-treated) or oxidized CosR with the mpx promoter region. As shown in Figure 6A,B, reduced CosR formed a tight complex with Pmpx, a 300 bp PCR fragment amplified from the mpx promoter, while incubation of oxidized CosR containing inter-monomer disulfide bonds with Pmpx did not cause retarded mobility of the probe, which was similar to the result of a 300 bp control DNA amplified from the mpx-coding region showing undetectable CosR binding. This important finding indicates that oxidation of CosR completely eliminated its DNA affinity, which may lead to depression of MPx
Reduced CosR bound to the mpx promoter DNA but intermolecular disulfide bond-containing dimerized CosR dissociated from the mpx promoter DNA.
C49 and C62 are required for CosR to dissociate from DNA under CHP condition
The above studies indicate well that the reduced form of CosR bound to target promoters and repressed transcription (Figure 6A,B). In the presence of peroxides, CosR was deactivated and released from the promoter. It was of interest to know whether C49 and C62 of CosR in the peroxide-sensing and inactivation mechanisms played a role. Thus, the capacity of the diverse CosR variants to depress and repress mpx expression in response to CHP was evaluated in ΔcosR background by the promoter lacZ activity and qRT-PCR analysis. Analysis of the lacZ activity revealed that ΔcosR+, ΔcosR+(C49S), and ΔcosR+(C49S) repressed mpx expression under untreated CHP conditions to equal degrees, similar to that of the WT strains (Figure 5). Otherwise, CHP treatment of ΔcosR strains harboring various mutant cosRs revealed different regulation patterns from ΔcosR+ strains and the WT strains. The ΔcosR+(C49S) and ΔcosR+(C62S) strains rescinded the ability of CHP to depress the expression of mpx, leading to constitutively low lacZ activity, in agreement with previous observations of X. campestris OhrR . In contrast, ΔcosR+ strains had an obvious positive effect on the ability of CHP to depress the expression of mpx, consistent with the observation that the promoter lacZ activity of mpx in the WT strains was significantly increased with the increase in CHP concentration (Figure 5A). These observations were consistent with the result from the CHP-induced dimeric formation of CosR in vivo. A similar regulator pattern of mpx expression in response to CHP was also observed in qRT-PCR analysis in ΔcosR+, ΔcosR+(C49S), and ΔcosR+(C49S) strains (Figure 5B). The results imply that Cys49 and Cys62 residues have a critical function in the CosR-sensing process.
To further probe whether Cys49 and Cys62 residues were responsible for the observed dissociation of CosR from DNA under oxidation, we pretreated CosR:C49S and CosR:C62S with 1 mM CHP treatment for 90 min to perform an EMSA experiment. As shown in Figure 6C,D, CHP-treated CosR:C49S and CosR:C62S also caused obvious retarded mobility of the probe, indicating that they completely failed to prevent the CHP-mediated dissociation of CosR from DNA. This result indicated that CosR:C49S and CosR:C62S proteins, expressed in ΔcosR+(C49S) and ΔcosR+(C62S) strains, respectively, always bind to the mpx promoter to block RNA polymerase binding regardless of CHP treatment. In addition, survival rate analysis also showed that increased bacterial susceptibility to all of these detected peroxides and antibiotics was still observed when the shuttle plasmid pXMJ19 encoding cosR:C49S or cosR:C62S was expressed in the cosR mutant strain compared with both WT and ΔcosR+ strain (Figure 1A). These results firmly established the thiol oxidation-based sensing and regulation mechanism for the CosR-mediated oxidative resistance.
CosR is specifically reduced by the thioredoxin system
C. glutamicum contained three alternative physiological reducing systems, MSH system (MSH/Mtr), Mrx1 system (Mrx1/Mtr/MSH), and Trx system [thioredoxin (Trx)/TrxR] . MSH averts disulfide stress, in which MSH acts in combination with thiol and mycothiol disulfide reductase as a biological relevant monothiol-reducing system. Mrx1 system reduces the mixed disulfide via monothiol, in which Mrx1 acts in combination with MSH. Trx system reduces disulfides via a dithiol mechanism, in which Trx reduces disulfide bond in other protein to form a transient inter-molecular disulfide bond between Trx and the reacted protein, and then inter-molecular disulfide bond is reduced by TrxR and NADPH. . Taking into account that the oxidized CosR was reduced in vitro by DTT treatment, we next monitored the ability of these reducing systems in providing reducing power for the oxidized CosR by the rate of NADPH consumption. As shown in Figure 7, only the addition of the oxidized CosR into the Trx system showed a consumption of NADPH. No reduction in the oxidized CosR was observed in the presence of MSH or Mrx1 system (Figure 7A–C). The data suggest that the Trx system could specifically reduce CosR.
Intermolecular disulfide bond-containing dimerized CosR was specifically reduced by the thioredoxin system in vitro.
To further confirm the results that CosR uses Trx as an electron donor, we investigated the direct interaction between Trx and CosR via bacterial two-hybrid and GST pull-down assays. Early reports suggested that sulfhydryl group regeneration at the active-site Cys of CosR was formed by the N-terminal Cys of Trx at the CXXC active sites (Trx_C32XXC35), and a transient intermolecular disulfide bond linking CosR with Trx can be stabilized by removing an internal cysteine of Trx at the CXXC active sites [56,57]. Therefore, to avoid attack by the C-terminal Cys of Trx at the CXXC active sites, a single mutant of Trx protein, namely Trx:C35S, was constructed. As shown in Figure 7D, the Trx:C35S variant showed a very strong interaction with CosR, similar to the positive control. To verify the Trx–CosR interaction seen in the bacterial two-hybrid assay, we produced recombinant His6-Trx:C35S and tested their interactions with GST-CosR using the GST pull-down assay in vitro. As shown in Figure 7E, we observed the specific interaction between GST-CosR and His6-Trx:C35S. The above in vitro results indicated that CosR can be regenerated by Trx.
Direct sensing and quick response by redox-sensing transcription regulators with cysteine residues were regarded as efficient ways to protect themselves against oxidative stress. Here, we investigated how the MarR-type CosR protein, a transcription repressor in C. glutamicum, monitored and responded to oxidative stresses. Oxidative stress led to the formation of disulfide bonds between cysteine residues. The chemical modification caused considerably different changes in the CosR structure, resulting in the release of CosR from the DNA of antioxidant genes. The redox-sensing transcription regulator CosR allowed C. glutamicum to respond to multiple oxidative signals of differing toxicity.
bovine serum albumin
electrophoretic mobility shift assay
reactive oxygen species
Can Chen and M.S. designed the research. M.S., T.S., Can Chen, Cheng.C, S.Y., Guang.L and Gui.L. performed the research and analyzed the data. G.Y. supervised the research. M.S. and G.Y. wrote the paper.
This work was supported by the National Natural Science Foundation of China (31500087), Science and Technology Plan Project of Shandong Higher Education Institutions (J16LE03), the international cooperation grant from Henan Province Science and Technology Agency (no. 172102410060) and Foundation of Henan Educational Committee (no. 17B180008).
The Authors declare that there are no competing interests associated with the manuscript.
These authors contributed equally to this work.