Molecular Insights into Hydrogen Peroxide-sensing Mechanism of the Metalloregulator MntR in Controlling Bacterial Resistance to Oxidative Stresses*

Manganese contributes to anti-oxidative stress particularly in catalase-devoid bacteria, and DtxR family metalloregulators, through sensing cellular Mn2+ content, regulate its homeostasis. Here, we show that metalloregulator MntR (So-MntR) functions dually as Mn2+ and H2O2 sensors in mediating H2O2 resistance by an oral streptococcus. H2O2 disrupted So-MntR binding to Mn2+ transporter mntABC promoter and induced disulfide-linked dimerization of the protein. Mass spectrometry identified Cys-11/Cys-156 and Cys-11/Cys-11 disulfide-linked peptides in H2O2-treated So-MntR. Site mutagenesis of Cys-11 and Cys-156 and particularly Cys-11 abolished H2O2-induced disulfide-linked dimers and weakened H2O2 damage on So-MntR binding, indicating that H2O2 inactivates So-MntR via disulfide-linked dimerization. So-MntR C123S mutant was extremely sensitive to H2O2 oxidization in dimerization/oligomerization, probably because the mutagenesis caused a conformational change that facilitates Cys-11/Cys-156 disulfide linkage. Intermolecular Cys-11/Cys-11 disulfide was detected in C123S/C156S double mutant. Redox Western blot detected So-MntR oligomers in air-exposed cells but remarkably decreased upon H2O2 pulsing, suggesting a proteolysis of the disulfide-linked So-MntR oligomers. Remarkably, elevated C11S and C156S but much lower C123S proteins were detected in H2O2-pulsed cells, confirming Cys-11 and Cys-156 contributed to H2O2-induced oligomerization and degradation. Accordingly, in the C11S and C156S mutants, expression of mntABC and cellular Mn2+ decreased, but H2O2 susceptibility increased. In the C123S mutant, increased mntABC expression, cellular Mn2+ content, and manganese-mediated H2O2 survival were determined. Given the wide distribution of Cys-11 in streptococcal DtxR-like metalloregulators, the disclosed redox regulatory function and mechanism of So-MntR can be employed by the DtxR family proteins in bacterial resistance to oxidative stress.

Manganese is essential for organisms in the resistance of oxidative stresses (1)(2)(3) and required by many pathogenic bacteria for virulence and persistence in infected hosts (4 -7). Manganese ion not only serves as a cofactor of superoxide dismutase but also forms non-proteinaceous manganese antioxidants by complexing with orthophosphate, lactate, or other small molecules (8 -10). It is even found that cellular Mn 2ϩ assists bacteria in ␥-radiation resistance, which is virtually a tolerance to reactive oxygen species generated by ionizing radiation (11). Manganese antioxidants play particularly important roles in catalase void Lactobacillus and streptococci (4,8,12,13). Acquisition of manganese is also crucial for virulence of Streptococcus pathogens like Streptococcus pneumoniae and Streptococcus pyogenes, most likely because manganese-antioxidants equip the bacterial pathogens for survival against the oxidants generated by the infected hosts (4,6).
Manganese transportation is tightly controlled by metalloregulators because of the high cellular toxicity of the manganese ions when at excess concentration. Bacteria employ metal ion-dependent transcriptional regulators to control metal ion homeostasis, mainly through control of metal ion transport in or out of cells (14,15). Through activating or suppressing the expression of divalent metal ion transporter genes, metalloregulators are responsible for the regulatory mechanisms of metal ion uptake and efflux that ensure sufficient cellular metal ions for bacteria metabolism and oxidative resistance while avoiding the toxic effect to cells (16 -20). Recently, metalloregulators are found to regulate virulence gene expression in pathogenic bacteria (21,22), making them the promising therapeutic targets in controlling pathogen infection. The streptococcal DtxR/MntR family metalloregulator homologues, including SloR from Streptococcus mutans (23), ScaR from Streptococcus gordonii (24), and PsaR from S. pneumoniae (25) are affiliated with the manganese/iron type of metalloregulators. They all repress Mn 2ϩ transport by sensing cellular Mn 2ϩ sufficiency to prevent excess metal toxicity, e.g. PsaR acts as the Mn 2ϩ -dependent repressor of the Mn 2ϩ importer psaABC and the virulent genes prtA and pcpA (26), and ScaR suppresses the virulence-related Mn 2ϩ permease scaCBA operon expression by sensing cellular Mn 2ϩ concentration (24). The structural studies of ScaR, SloR, and PsaR in the presence of Cd 2ϩ or Zn 2ϩ reveal that Cys-123 is one of the key residues for metal ion binding (27)(28)(29).
Given the essentiality of Mn 2ϩ for the protection of bacteria from oxidative harm, it is crucial to understand how these DtxR-like metalloregulators regulate Mn 2ϩ import by sensing cellular redox status (H 2 O 2 level). Streptococcus oligofermentans is a beneficial oral commensal and generates a high concentration of H 2 O 2 to inhibit the growth of dental caries pathogen S. mutans and also tolerates a high amount of H 2 O 2 (30 -32). Although previously we found that in response to H 2 O 2 , the peroxide-responsive repressor PerR de-represses the expression of mntABC gene that encodes a Mn 2ϩ uptake transporter in S. oligofermentans (13), a direct connection between PerR and mntABC has not been found.
In the present study we identified a ScaR/PsaR homolog, tentatively named mntR So , in the S. oligofermentans genome. By sensing intracellular Mn 2ϩ concentration, So-MntR directly repressed the expression of the Mn 2ϩ transporter operon mnt-ABC. In the following redox proteome study, to screen the redox-sensitive cysteines in H 2 O 2 -pulsed S. oligofermentans, it was found that all three cysteines of the So-MntR protein were oxidized to form disulfide bond. 3 This suggests that So-MntR can be a metalloregulator in control Mn 2ϩ uptake by sensing oxidative stress. In the present study we, through redox biochemistry, genetic, and physiological studies, demonstrated that in response to H 2 O 2 , mntR So de-represses the expression of mntABC and facilitates Mn 2ϩ import; we also found that H 2 O 2 inactivates So-MntR. The cellular So-MntR oligomers appear to be readily degraded and so cause de-repression of mntABC and an increase of Mn 2ϩ import. Thus, this work provides a new mechanism of a metalloregulator contributing to oxidative stress resistance in response to H 2 O 2 .

S. oligofermentans MntR Suppressed mntABC Expression and Mn 2ϩ Import by Sensing Both the Cellular Mn 2ϩ and Redox
Status-To search the ScaR/PsaR ortholog in S. oligofermentans, scaR from S. gordonii (SGO_1816) was used as a probe to query the completed S. oligofermentans genome. A gene (I872_01020) annotated as "DtxR family manganese-dependent transcriptional regulator" was hit at a 73.8% identity with scaR. Similar to ScaR, this 25-kDa protein possessed three domains: an N-terminal DNA binding domain, a central metal binding and dimerization domain, and a C-terminal FoeA domain. Moreover, it contained the conserved metal ion binding amino acids that present in the ScaR-like DtxR homologues from streptococci (27), namely Asp-7, Glu-99, Glu-102, and His-103 (the primary site) and Glu-80, Cys-123, His-125, and Asp-160 (the secondary site) (Fig. 1A). Thus this gene was tentatively assigned as mntR So and So-MntR of the encoded protein. A protein superimposition of the So-MntR homology model and the dimeric crystal structure of ScaR (3hrs.1.pdb, X-ray diffraction at 2.70 Å) also showed good matching (Fig.  1B).
To confirm the regulatory role of So-MntR in transport of Mn 2ϩ and other metal ions, a mntR So deletion strain was con-structed by means of double-crossover recombination. Inductively coupled plasma mass spectrometry (ICP-MS) 4 was used to determine the cellular content of metal ions in the wild strain and mntR So mutant that grows in brain heart infusion (BHI) broth supplemented with a variety of concentrations of MnCl 2 . As shown in supplemental Table S1, deletion of mntR So caused an ϳ1.5-fold increased cellular Mn 2ϩ content in the presence of Յ3 M MnCl 2 and a 3-fold increase with a 100 M and 5 mM MnCl 2 supply. Correspondingly, mntR So deletion decreased S. oligofermentans Mn 2ϩ tolerance, and growth was significantly suppressed by 5 mM Mn 2ϩ , whereas the wild type tolerated up to 5 mM Mn 2ϩ (supplemental Fig. S1). Noticeably, the cellular iron level was also increased in the mntR So mutant. Cellular iron content decreased as the manganese supply increased (supplemental Table S1), implying that the Mn 2ϩ transporter prefers manganese uptake.
Previously, we determined that MntABC is the major Mn 2ϩ transporter of S. oligofermentans (13). To link So-MntR-controlled Mn 2ϩ uptake with the immediate regulation of Mn 2ϩ transporters, mntABC expression was measured in luciferase reporter strains WT-PmntABC-luc (13) and ⌬mntR So -Pmnt-ABC-luc. As expected, in comparison with the wild strain, mntABC expression increased 3.2-and 4.6-fold in ⌬mntR So -PmntABC-luc strain that cultured statically or anaerobically in BHI broth (supplemental Table S2), respectively. When growing the two reporter strains, a given concentration of MnCl 2 or FeSO 4 in a metal ion-deprived BHI (B-BHI) broth, luciferase activity assay did not detect a significant repression of mntR So on mntABC unless the addition was Ն0.5 M Mn 2ϩ . However, even up to 200 M, ferrous ion did not activate mntR So repression of mntABC (supplemental Table S2). This indicated that manganese is the cognate ligand of So-MntR. Surprisingly, only half the level of mntABC expression was determined in an anaerobic culture (luciferase activity, 1.05 Ϯ 0.17 ϫ 10 6 relative light units (RLU)/A 600 ) compared with the statically cultured cells (2.01 Ϯ 0.35 ϫ 10 6 RLU/A 600 ). This implies that So-MntR may also sense redox potential in controlling mntABC expression. Taken together, So-MntR suppression of mntABC requires Mn 2ϩ as the ligand and responds to oxidation status.
H 2 O 2 Decreased Mn 2ϩ -dependent Binding of So-MntR to mntABC Promoter-To characterize So-MntR protein, it was overexpressed in Escherichia coli and purified. Size exclusion chromatography showed that So-MntR molecular mass was ϳ50 kDa, indicating that it presents a dimer in the solution (supplemental Fig. S2). Next, electrophoretic mobility shift assay (EMSA) was performed to determine the direct association of So-MntR and mntABC. Fig. 2A showed that in the presence of 100 M Mn 2ϩ , as low as 5 nM So-MntR protein bound to the promoter of mntABC, and a dose-dependent protein-DNA complex formation was observed. The shifted DNA bands were abolished by the addition of increased unlabeled mntABC promoter into the binding mixture. However, no DNA shift was found in the absence of Mn 2ϩ , indicating that So-MntR specifically binds to mntABC promoter by using Mn 2ϩ as the ligand.
Given that a higher mntABC expression occurred in static than in anaerobic culture (supplemental  (15 nM) to the mntABC promoter, it was found that 0.1 and 1 mM H 2 O 2 , respectively, caused partial and complete dissociation of So-MntR from the target DNA (Fig. 2C). Interestingly, the H 2 O 2 -disrupted binding complex was mostly recovered by 10 mM 1,4-dithiothreitol (DTT) treatment, indicating an involvement of reversible disulfide formation upon H 2 O 2 oxidation (Fig. 2C). To avoid possible Fenton chemistry damage, the So-MntR-His 6 protein was purified and dialyzed in the presence of 1 mM EDTA, and treated with 10 g/liter Chelex 100 to remove metal ions at the last dialysis. Metal ion content -dimer homology model of S. oligofermentans MntR was generated via the SWISS-MODEL web server by automatically matching 3hrs.1.pdb, a dimeric crystal structure of ScaR from S. gordonii, as template. The dimeric So-MntR (two monomers shown in blue and green, respectively) was overlaid with ScaR (gray). The three cysteine residues were shown as sticks, and the distance (Å) between each pair was predicted with PyMOL.
in the So-MntR-His 6 was then measured using ICP-MS, which detected only trace amounts of transition metal ions in 15 nM protein with iron (0.09 nM), cobalt (0.0003 nM), copper (0.018 nM), and manganese (0.0015 nM).
H 2 O 2 Facilitated Disulfide-linked So-MntR Dimers-So-MntR contains three cysteine residues (Fig. 1A), and the thiol groups can be oxidized by H 2 O 2 to form a disulfide bond. Dimers were indeed observed for the purified So-MntR on denaturing non-reducing SDS-PAGE gel. Therefore, H 2 O 2 oxidation on So-MntR disulfide-linked dimer formation was further analyzed. Fig. 3A showed that after 90 min of treatment with 0.1 mM H 2 O 2 , an ϳ4-fold increase in dimer formation (lane 1 versus lane 4) as well as slightly increased disulfidelinked monomers were noted. The dimer bands were all abolished by DTT reduction (Fig. 3B) Fig. 3A). To avoid incorrect disulfide linkages induced during protein denaturation, So-MntR was pretreated with N-ethylmaleimide (NEM), a thiol alkylating agent, to block the free thiol groups before being subjected to nonreducing SDS-PAGE gel. Notably, disulfide-linked dimers/olig-omers frequently occurred in So-MntR without H 2 O 2 treatment (lane 1, Fig. 3A), suggesting that this protein is extremely liable to air oxidation.
To determine the disulfide bond linkages in H 2 O 2 -treated So-MntR, the dimer and disulfide-linked monomer bands (Fig.  3A, lanes 1 and 4) were excised and subjected to LC-MS/MS analysis. Iodoacetamide was used to block the free thiol groups before trypsin digestion of the gel. Cys-11/Cys-11, Cys-11/Cys-156, and Cys-156/Cys-156 fragments were all identified in the dimers, whereas Cys-11/Cys-156 and Cys-11/Cys-11 were overrepresented in H 2 O 2 -treated protein according to the increased disulfide-linked peptide spectral matches numbers compared with the untreated protein (supplemental Table S3). Fig. 3C shows the representative MS/MS spectrometric maps in which Cys-11/Cys-11 and Cys-11/Cys-156 disulfide linkages were each identified as a 3-charged peptide fragment of CIYEIGTRQEK/CIYEIGTR (observed m/z of 764.378) and CIYEIGTR/YHQSLASAQEPGNYLICR (observed m/z of 1001.151), respectively. In disulfide-linked monomer only Cys-11/Cys-156 peptide was identified (supplemental Table S3). Noticeably, the intramolecular Cys-11/Cys-156 linkage appeared to change the protein conformation by migrating faster than the common monomer (band im, Fig. 3A). Cys-123 seemed not to involve either the inter-or the intramolecular disulfide bond; likely it is involved in Mn 2ϩ binding as in other DtxR-like metalloregulators (27).

Cys-11 and Cys-156 Contribute to H 2 O 2 -induced Disulfidelinked Dimerization and So-MntR Inactivation, Whereas Cys-123
Linked to DNA Binding-To verify the contribution of each cysteine to the disulfide linked dimers, we mutated each of the three residues to serine and examined dimerization in H 2 O 2treated mutant proteins. A non-reducing SDS-PAGE ( Fig. 4A) assay showed that mutation of Cys-11 completely eliminated the dimerization, whereas the Cys-156 mutation reduced the dimers but increased oligomers. This indicates that both Cys-11 and Cys-156 are involved in disulfide-linked dimerization. In contrast, in response to H 2 O 2 treatment, the So-MntR C123S mutant substantially increased both intra-and intermolecular disulfide bond linkages. To further determine the relative level of same cysteine disulfide-linked dimers, C11S and C156S mutagenesis was each introduced into So-MntR C123S mutant to generate double cysteine mutagenesis. As shown in Fig. 4A, either the C11S/C123S or C156S/C123S double mutants (dimer:monomer ϭ 0.10) produced ϳ8.7-fold less disulfide-linked oligomers than that the C123S mutant (dimer: monomer ϭ 0.87) upon H 2 O 2 treatment. This indicates that same cysteine disulfide linkage between Cys-11/Cys-11 and Cys-156/Cys-156 would contribute 20% of the S-S dimers, and Cys-11/Cys-156 contributes the remaining 80%. Noticeably, there were more disulfide-linked di-and oligomers in C123S mutant, implying that C123S mutagenesis might change the So-MntR conformation, thereby facilitating the interaction of Cys-11 and Cys-156. Circular dichroism (CD) spectra analysis did show a dramatic secondary structure change for the C123S mutant that reduced the ␣ helix (from 39% in the wild protein to 24%) but increased the ␤ strand content (from 17% in the wild protein to 28%). There was only a slight conformation change for the C11S and C156S mutants (supplemental Fig. S3).  Next, impact of cysteine mutation on So-MntR binding to DNA was detected by EMSA. Cys-11 mutation remarkably increased So-MntR affinity to DNA by showing binding at as low as 1 nM protein (Fig. 4B), most likely because of its position in the DNA binding domain. By comparing the DNA binding ability change by H 2 O 2 treatment, it was found that either C11S or C156S mutation reduced H 2 O 2 damage on So-MntR affinity to DNA, indicating that disulfide-linked dimer formation causes So-MntR inactivation (Fig. 4B). In addition, C123S mutation reduced about half of the DNA affinity regardless of H 2 O 2 oxidation, indicating that this residue contributes DNA binding.
H 2 O 2 -induced Disulfide-linked Oligomerization Caused Cellular So-MntR Decrease-To test the intracellular So-MntR status, we constructed the S. oligofermentans mntR So -His 6 strain by inserting the C-terminal-His 6 -tagged mntR So into shuttle vector pDL278 (34) and transformed it into the mntR So deletion strain. The anaerobically cultured ⌬mntR So -mntR cells at the middle exponential phase were harvested and divided into four aliquots. Two aliquots were exposed to air for 30 or 60 min, 1 was air-exposed but with 200 g/ml catalase addition, and another aliquot was retained anaerobically. Cell samples were then sonicated in RIPA buffer containing NEM, and the supernatant was subjected to redox-Western hybridization using anti-His-tag antibody. As shown in Fig. 5A, compared with the anaerobically stand sample, 30 min of air exposure caused more So-MntR oligomer, which became much less in the catalase-treated sample. This indicates that H 2 O 2 oxidation contributes to the oligomer formation. Next, anaerobically cultured mntR So -His 6 strain was treated with various concentrations of H 2 O 2 , and then the in vivo So-MntR level was determined as described above. It turned out that either 30 min of pulsing of 40 M H 2 O 2 or 5 min of pulsing of 120 M H 2 O 2 significantly increased both monomer and oligomer formation, whereas extending the 120 M H 2 O 2 pulsing to 10 or 30 min A, culture aliquots of the ⌬mntR So -mntR strain were exposed to air for 30 or 60 min. ϩcat, catalase (200 g/ml) was added to the culture. Cells were collected and sonicated in RIPA buffer containing the thiol alkylating agent NEM. After centrifugation, 50 g of cellular protein were resolved by non-reducing SDS-PAGE, and a 4000-fold dilution of the anti-His-tag antiserum was used to probe the in vivo  caused the oligomer to be diminished and the monomer to be decreased significantly as well (Fig. 5B). The redox Western hybridization data strongly suggest that under H 2 O 2 stress cellular MntR forms inactive oligomers and are then subjected to a proteolysis. The monomer increase can be due to H 2 O 2 induction of mntR So expression, found in an unpublished Microarray data. 3 H 2 O 2 induced cellular So-MntR dimerization/oligomerization, and vanishing was further tested on the cysteine-mutated mntR So mutants, which are insensitive to H 2 O 2 oxidation in formation of dimers and oligomers. The mntR So deletion mutant was complemented with each of the three cysteinemutated (C11S, C123S, and C156S) mntR So -His 6 s, designated as ⌬mntR So -mntRC11S, ⌬mntR So -mntRC123S, and ⌬mntR So -mntRC156S, respectively. These strains were grown anaerobically until mid-log phase and then subjected to 30 min of pulsing with 120 M H 2 O 2 . Redox Western blot was performed to examine the So-MntR proteins in both the supernatant and precipitate of the cell extract. It was observed that upon H 2 O 2 pulsing, So-MntR protein reduced by 33 and 65% in the ⌬mntR So -mntR and ⌬mntR So -mntRC123S strains, respectively. However, only a 14 and 25% decrease of So-MntR protein was noted in the ⌬mntR So -mntRC11S and ⌬mntR So -mntRC156S strains, respectively (Fig. 5C, upper panel). No So-MntR protein was detected in the cell extract precipitation of any tested strains (Fig. 5C, lower panel). This suggests that diminished So-MntR protein in the cell can be attributed to a proteolytic action. The H 2 O 2 sensitivity of each So-MntR protein is in line with its tendency in dimerization/oligomerization (Fig. 4A).
So  Fig. 6 showed that upon H 2 O 2 treatment ϳ2.8and ϳ2.5-fold elevated mntA transcript was detected in ⌬mntR So -mntR and ⌬mntR So -C123S strains, respectively, whereas no significantly increase was detected in the ⌬mntR So -mntRC11S and C156S strains, indicating that the cysteine residues contribute to So-MntR repression of mntABC expression when encountering H 2 O 2 .
As MntABC functions in Mn 2ϩ transport (13), ICP-MS was then used to determine the cellular Mn 2ϩ content in ⌬mntR So -pDL278, ⌬mntR So -mntR, and ⌬mntR So -mntRC11S, -C123S, and -C156S strains. Strains were all grown statically in either BHI broth or BHI supplemented with 0.1 mM MnCl 2 . Under either culture conditions, a significantly reduced intracellular Mn 2ϩ was detected in the ⌬mntR So -mntRC11S and C156S strains (Table 1). This is consistent with the reduced mntABC expression in the ⌬mntR So -mntRC11S and -C156S mutants that retained more So-MntR protein in the cell (Figs. 6 and 5C).
So-MntR Mediated Manganese-dependent Anti-oxidative Stress Based on the H 2 O 2 -reactive Cys-11 and Cys-156 -To examine the effect of mntR So deletion and So-MntR cysteine mutation on manganese-mediated H 2 O 2 survival of S. oligofermentans, the above strains were grown in either BHI broth or BHI supplemented with 0.1 mM MnCl 2 . When the growth reached A 600 ϳ 0.4, 1 aliquot of each culture was exposed to 10 mM H 2 O 2 for 10 min while leaving another aliquot untreated. H 2 O 2 survival rate was determined by the percentage of colony forming units (cfu) in H 2 O 2 -exposed aliquots against those in non-exposed aliquots. As shown in  (Fig. 1A) to survey the distribution of H 2 O 2 -reactive cysteine residues. Cys-11, positioned at the DNA binding domain, is present in all the streptococcal DtxR homologues except for S. mutans SloR. Cys-123, a predicted metal ion binding residue situated in the dimerization domain, is widely distributed in all the DtxR-like proteins from streptococci, whereas Cys-156 presents only in the MntR of S. oligofermentans and Streptococcus lutetiensis (35). Of note, a group of streptococcal DtxR homologues, including those derived from pathogen S. pneumoniae, Streptococcus pseudopneumoniae and Streptococcus equi, contain the 3rd cysteine at position 57 associated to the DNA binding domain. This implies that the DtxR family proteins from streptococci might all be H 2 O 2 -reactive.

Discussion
Metal homeostasis plays a central role in the anti-oxidative stress of Gram-positive bacteria particularly of the catalase void lactic acid bacteria (1,2,4,12). Metalloregulators are the key controllers in maintaining bacterial metal homeostasis by sensing the cellular metal redundancy or deficiency (14,15,36). This work presents that a manganese homeostasis regulator MntR, by sensing the cellular oxidant, de-represses the Mn 2ϩ transporter mntABC expression and enables an oral commensal S. oligofermentans to resist oxidative stress. Fig. 7 shows the working mechanism of the S. oligofermentans MntR. By sensing the cellular Mn 2ϩ redundancy, So-MntR represses the expression of mntABC, the major Mn 2ϩ importer. When encountering H 2 O 2 stress either from the metabolic accumulation or from external source, So-MntR is oxidized at the redox-reactive thiol groups of Cys-11 and/or Cys-156 to form disulfide-linked dimers or oligomers, which can be subjected to a proteolysis. Depletion of the cellular MntR protein leads to mntABC de-repression and increased Mn 2ϩ uptake and consequently facilitates the bacterium manganese-mediated resistance of oxidative stress.
Streptococci are well known for accumulating high cellular H 2 O 2 due to a lack of catalase. Therefore, they have evolved unusual protective approaches from oxidative stress. Maintaining metal ion homeostasis is one of the important strategies (2,13). Streptococci usually use Mn 2ϩ instead of Fe 2ϩ centralized metabolism, likely to avoid Fenton chemistry, which converts H 2 O 2 to a more deleterious oxidant hydroxyl radical (7,37). Although manganese, a first row transition metal, catalyzes Fenton-like reaction at acidic pH as well (38,39), at physiological pH values it exhibits inorganic superoxide dismutase and catalase activities by complex with low molecular weight molecules, such as orthophosphate and bicarbonate (9,10,40). Manganese has been reported to help streptococci cope with oxidative stresses (4). Thus, intensive studies have been focused on how the streptococcal and other Firmicutes metalloregulator switch functions by sensing and binding the ligand metals (41,42), e.g. the S. pyogenes MtsR and S. mutans SloR sense and regulate both cellular manganese and iron levels (29,43), whereas the S. gordonii ScaR and S. pneumoniae PsaR function as the Mn 2ϩ -sensitive regulators (24,27,28). Similar to ScaR and PsaR, the S. oligofermentans MntR also senses intracellular Mn 2ϩ and then inhibits the expression of Mn 2ϩ transporter operon mntABC (supplemental Table S2). Notably, this work demonstrates that So-MntR can be inactivated by H 2 O 2 via oxidation of its redox sensitive cysteines (Figs. 2 and 3). H 2 O 2 induces MntR to form Cys-11/Cys-11 and Cys-11/Cys-156 disulfide-linked dimers and oligomers, leading to a de-repression of mntABC expression and Mn 2ϩ uptake for coping with the oxidative stress. This is supported by the fact that the mntR So mutant was more resistant to H 2 O 2 injury than the wild type strain ( Table 1). Deletion of mntR So also led to elevated intracellular iron (supplemental Table S1) but not decreased the H 2 O 2 sensitivity of the bacterium. This can be attributed to the role of Dpr, a Fe 2ϩ chelating protein, as an unpublished microarray analysis showed ϳ3-fold up-regulation of dpr in the mntR So inactivation mutant. 3 To insight into the disulfide bond formation on protein conformation distortion, we, using SWISS-MODEL protein structure homology modeling server, performed a homology model of So-MntR protein and overlaid it with the dimeric crystal structure of the S. gordonii ScaR (3hrs.1.pdb, X-ray diffraction at 2.70 Å, Fig. 1B). The two proteins show an overall structural similarity. Based on this modeling structure of So-MntR, a distance of 25.67 Å was predicted between Cys-11 and Cys-156 within each monomer, whereas between monomers, distances of 55.69 Å and 41.83 Å for Cys-11/Cys-156, and Cys-11/Cys-11 were measured with PyMOL, respectively. Apparently, the predicted cysteine distances are much larger than 2.8 Å to allow for a disulfide-bond formation. Therefore, the identified Cys-11/ Cys-11 or Cys-11/Cys-156 peptides in H 2 O 2 -treated So-MntR (Fig. 3C) strongly indicate a structure distortion of the H 2 O 2oxidized protein. According to the protein domain analysis (Fig. 1A) and structural study on S. gordonii ScaR (27), Cys-11 situates on the ␣1 helix associated to the DNA binding domain. This work shows that the Cys-11 thiol group exhibits high H 2 O 2 reactivity to form intermolecular disulfide linkage with Cys-156 or Cys-11. Thus oxidation must change the protein into inactive form, which is supported by the in vitro EMSA data (Fig. 2, B and C).

TABLE 1 Effect of each cysteine mutation of mntR So on the cellular manganese content and Mn 2؉ -mediated H 2 O 2 survival of S. oligofermentans
Data are the means Ϯ S.D. of three independent experiments. Suffixes of the ⌬mntR So represent the complemented gene types of the mntR So , and pDL278 indicates an empty vector only.

Measured items
⌬mntR Consistently, redox Western blot also detected oligomerization and degradation of the cellular So-MntR upon air exposure and H 2 O 2 pulse (Fig. 5). This indicates that a bacterial surveillance system can sense the aberrant protein and initiate proteolysis to eliminate the dysfunctional So-MntR dimers and oligomers. Similar finding is reported for H 2 O 2 -mediated degradation of the uracil-DNA glycosylase (UNG1) in human mitochondria (44). Oxidized MntR degradation can also be similar to that oxidative carbonylation causes protein aggregation and degradation (45).
Therefore, a conclusion can be drawn that the metalloregulator MntR plays a dual role in controlling cellular manganese homeostasis. In the absence of oxidative stress, MntR, by sensing the cellular Mn 2ϩ level, modulates Mn 2ϩ uptake to meet the bacterium physiological requirement. When encountering oxidative stress, it "sacrifices" itself through redox-sensitive thiol group-mediated oligomerization to release suppression of mnt-ABC for uptake Mn 2ϩ to cope with the oxidants. The H 2 O 2reactive cysteine residue Cys-11, identified in S. oligofermentans MntR, is widely distributed in the DtxR family metalloregulators from streptococci (Fig. 1A). Although Cys-156 presents only in the MntR of S. oligofermentans and S. lutetiensis, a group of streptococcal DtxR homologues, including those derived from pathogen S. pneumoniae, S. equi, and Streptococcus uberis, contain a third cysteine at position 57 that associated to DNA binding domain. A preliminary study implemented in our laboratory found that upon H 2 O 2 oxidation, the recombinant purified PsaR (spr1480) from S. pneumoniae formed disulfide-linked dimers as well. Further study is needed to extensively investigate the possible role of cysteines of these metalloregulators in H 2 O 2 sensing in pathogenic streptococci. This mechanism would benefit the oral commensals and pathogenic streptococci, enabling them adapted to the oxidative tension in human ecological niches and, therefore, pathogenic.
Redox switches based on the redox-sensitive cysteine thiol groups have been widely utilized by H 2 O 2 sensors in conducting the regulatory functions in response to reactive oxygen species that derived from the extracellular environments or respiratory metabolism. Those include the E. coli OxyR (46), the Bacillus subtilis OhrR (47) and the Staphylococcus aureus MgrA (48). In addition to the specialized redox regulators, recently quorum-sensing regulators, such as the Pseudomonas aeruginosa LasR (49), the S. aureus AgrA (50), and the E. coli SdiA (51) have been found to respond to oxidants via cysteine oxidation as well. Based on these findings and on that of the metalloregulator in this study, it is conceivable that any protein, including regulatory proteins, carrying redox-reactive thiol groups in cysteine or methionine residues can be involved in the regulation of redox reactions. Further studies on redoxsensitive cysteines at proteomics level are being implemented in our laboratory. These works would, therefore, much improve our understandings on how bacteria cope with oxidative stress via oxidizing protein post-translation modification.

Experimental Procedures
Bacterial Strains and Culture Condition-S. oligofermentans AS 1.3089 (52) and its derivative strains were grown in BHI broth (BD Difco, Franklin Lakes, NJ) statically or anaerobically under 100% N 2 at 37°C. To test the effect of Mn 2ϩ or Fe 2ϩ on gene expression, MnCl 2 or FeSO 4 was supplemented to 5% Chelex 100 (C7901, Sigma)-treated BHI plus 0.1 mM CaCl 2 and 2 mM MgCl 2 (B-BHI). BHI plates supplemented with spectinomycin (1 mg ml Ϫ1 ) or kanamycin (1 mg ml Ϫ1 ) were used to select transformants. E. coli strains used for cloning and plasmid construction were grown in Luria-Bertani medium. Spectinomycin (250 g ml Ϫ1 ) or kanamycin (50 g ml Ϫ1 ) was used when necessary. The metalloregulator MntR suppresses expression of the mntABC operon, which encodes the major Mn 2ϩ ABC transporter MntABC, and so restricts manganese uptake by sensing the cellular manganese sufficiency. When encountering oxidative stress, MntR sacrifices itself, via the redox-reactive thiol group-mediated oligomerization and degradation, and releases mntABC expression to initiate the uptake of Mn 2ϩ to cope with the oxidants. Solid lines specify the proved reactions, and broken lines represent predicted ones. Yellow oval, MntR protein; red dot, manganese ion; purple and blue lines, cysteine residue; red lines, disulfide bond.
Construction of Genetic Strains-Genomic DNA of S. oligofermentans was extracted as described previously (53,54). All primers (supplemental Table S4) were designed according to the genome sequence (33) and synthesized by Sangon Co. (Shanghai, China). mntR So deletion strain was constructed by the PCR ligation method (55). The upstream and downstream DNA fragments of mntR So were amplified using primer pairs listed in supplemental Table S4. The BamHI-digested PCR products were ligated with non-polar kanamycin resistance gene cassette released from plasmid pALH124 (56) and transformed into S. oligofermentans as described previously (54). Various mntR So gene ectopically expressed strains were constructed using the shuttle plasmid pDL278. A DNA fragment containing the promoter and six histidines tag-fused mntR So gene was PCR-amplified using the primer pair mntRcomF/ mntRHiscomR, and the PCR product was inserted into the BamHI and HindIII sites of pDL278 to produce pDL278-mntR-His 6 . Meanwhile, pDL278-mntRC11S-His 6 , C123S-His 6 , or C156S-His 6 was constructed using site-directed gene mutagenesis kit (Beyotime Biotechnology Co., Shanghai, China) and the mutagenesis primer pairs listed in supplemental Table S4. The correct recombinants were transformed into S. oligofermentans mntR So mutant. The PmntABC luciferase reporter in the mntR So mutant was constructed by deleting mntR So in the wild type PmntABC-luc fusion strain previously constructed (13). All the correct S. oligofermentans transformants were selected on BHI plates containing kanamycin or spectinomycin and identified by PCR and sequencing.
Overexpression and Purification of So-MntR Protein-A C-terminal fusion of His 6 to the So-MntR was constructed as follows. A 648-bp region containing the entire mntR So gene was PCR-amplified with primer pair mntR28aFNcoI/ mntR28aRXhoI (supplemental Table S4). The resultant product was digested with NcoI/XhoI and ligated into the compatible sites on pET-28a (Novagen, Madison, WI) to produce pET-28a-mntR. The construct was verified by DNA sequencing. So-MntR C11S, C123S, and C156S mutants were constructed using the site mutagenesis kit and primers described above by using pET-28a-mntR as the template. The substitution of serine for cysteine was verified by sequencing. pET-28a carrying wild type or cysteine-mutated mntR So was transformed into E. coli BL21 (DE3) (Novagen) cells. Correct transformants were cultured in LB medium supplemented with 50 g/ml kanamycin. Cells were grown at 37°C to A 600 of 0.4 -0.6, and 0.5 mM isopropyl-␤-D-thiogalactopyranoside (Sigma) was added. After an additional 3-4 h incubation, cells were collected by centrifugation at 8000 rpm for 10 min, resuspended in a 1 ⁄ 10 volume of binding buffer (20 mM sodium phosphate, 500 mM NaCl, 30 mM imidazole, 1 mM EDTA, 1 mM DTT, pH 7.4), and then lysed by sonication for 30 min. The cell lysate was centrifuged at 12,000 rpm for 15 min, and the supernatant was filtered through a 0.45-nm polyvinylidene difluoride membrane (Millipore, Billerica, MA) and then applied to a Ni 2ϩcharged chelating column (GE Healthcare) previously equilibrated with binding buffer. Proteins were eluted by elution buffer (20 mM sodium phosphate, 500 mM NaCl, 500 mM imidazole, 1 mM DTT, pH 7.4), and the elution fractions were analyzed by electrophoresis on a 12% sodium dodecyl sulfate-poly-acrylamide gel. The fractions with desired protein were pooled and dialyzed against phosphate-buffered saline (PBS: 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 , 137 mM NaCl, and 2.7 mM KCl, pH 7.4) containing 1 mM DTT and 1 mM EDTA 3 times, with the last dialysis in PBS containing 1 mM DTT and 10 g/liter Chelex 100. The purified proteins were stored in aliquots in 10% glycerol at Ϫ80°C until use.
EMSA-The mntABC promoter fragment was generated by PCR amplification using a biotin-labeled primer pair of PmntEMSAF/PmntEMSAR (supplemental Table S4). EMSA was performed using Light Shift Chemiluminescent EMSA Kit (Pierce). Briefly, 0.2 nM biotin-labeled dsDNA probe and increasing amounts of So-MntRs (1-50 nM) were mixed in the binding buffer (10 mM Tris-HCl, pH 8.0, 5% glycerol, 50 mM NaCl, 10 g/ml BSA, 2 ng/l poly(dI-dC), 0.5 mM DTT, and 0.1 mM MnCl 2 ). The reaction remained at 30°C for 30 min and then was electrophoresed on 8% polyacrylamide gel on ice. The DNA-protein complex was transferred onto a nylon membrane and detected by Chemiluminescent Nucleic Acid Detection Module kit (Thermo Scientific TM ).
Non-reducing SDS-PAGE-Five micrograms of So-MntR and its cysteine-mutated proteins were incubated with 0.1 or 1 mM H 2 O 2 for 90 min at room temperature, respectively. Before electrophoresis, NEM (final concentration 40 mM) was added and put in the dark for 30 min. 4-Fold diluted non-reducing SDS loading buffer (0.2 M Tris-HCl, pH 6.8, 40% glycerol, 8% SDS, 0.4% bromphenol blue) was added, and then proteins were separated on 12% SDS-PAGE gel.
LC-MS/MS Identification of Disulfide-linked Peptide-Nonreducing SDS-PAGE was used to separate H 2 O 2 -untreated and -treated So-MntR protein (total amount 5 g/lane). After staining with Coomassie Blue G-250, the gel bands were cut into pieces. Gel pieces were washed twice with MS-grade water and directly alkylated with 55 mM iodoacetamide in the dark for 1 h at 37°C and then digested by sequencing grade modified trypsin (Promega, Fitchburg, WI) in 50 mM NH 4 HCO 3 , pH 8.0, at 37°C overnight. The digested products were extracted twice with 1% formic acid in 50% acetonitrile aqueous solution and dried to reduce volume by SpeedVac. For LC-MS/MS analysis, the peptides were separated by a 65-min gradient elution at a flow rate of 0.250 ml/min with the EASY-nLC integrated nano-HPLC system (Thermo-Fisher), which was directly interfaced with the Thermo Q-Extractive mass spectrometer. The analytical column was a homemade fused silica capillary column (75-mm internal diameter, 150-mm length; Upchurch) packed with C-18 resin (300 Å, 5 m; Varian). Mobile phase A consisted of 0.1% formic acid, and mobile phase B consisted of 100% acetonitrile and 0.1% formic acid. The Q-Extractive mass spectrometer was operated in the data-dependent acquisition mode using the Xcalibur 3.0 software. A single full-scan mass spectrum in the Orbitrap (300 -1800 m/z, 70,000 resolution) was followed by 20 data-dependent MS/MS scans in the ion trap at 27% normalized collision energy. Each mass spectrum was searched against So-MntR protein sequence (Uniprot accession number N0C1Z2) using the SEQUEST searching engine of Proteome Discoverer software (v1.4). Peptide fragments with disulfide linkage were analyzed by PMi-Byonic software (Protein Matrix Inc.).
Cellular Metal Content Measurement-Concentrations of metal ions in static cultures of various S. oligofermentans strains were measured using ICP-MS. Overnight BHI cultures of tested strains were diluted 1:50 into fresh BHI broth and incubated at 37°C under static conditions. Mid-log phase cells were harvested by centrifugation at 13,400 ϫ g for 10 min. The cell pellets were washed twice in PBS with 1 mM EDTA and once in PBS without EDTA and then resuspended in 1 ml of PBS. 100 l of suspension was used to measure protein concentration with a BCA protein analysis kit per the manufacturer's recommendations. The remaining 900 l of suspension was collected by centrifugation at 13,400 ϫ g for 10 min. Pelleted bacteria were resuspended in 500 l of nitric acid (ultra pure). After incubation at room temperature overnight, the volume was brought to 1.5 ml with deionized distilled water. The samples were then analyzed for metal ions content with ICP-MS (DRCII, Perkin-Elmer Life Sciences) at Peking University Health Science Center. Beryllium, indium, and uranium were used as metal ions standard to calibrate ICP-MS. Experiments were conducted in triplicate, and each was repeated at least three times. Metal content was expressed in nmol/mg of protein.
Assay of Hydrogen Peroxide Sensitivity-Overnight cultures were 1:100 diluted into fresh BHI broth and incubated anaerobically. Until the growth reached A 600 ϳ 0.5-0.6, 0.2 ml of cells were harvested by centrifugation and then resuspended in 0.2 ml of fresh BHI broth. One aliquot (200 l) was pulsed with 10 mM H 2 O 2 , leaving another aliquot not treated. After incubation at 37°C for 10 min, cells from both samples were collected and washed twice with PBS buffer and resuspended in 200 l of BHI broth. Cell chains were separated by sonication for 30 s with a XC-3200D ultrasonic cleaner (Xinchen Co., Nanjing, China) and then 10-fold series dilutions were performed. Appropriate dilutions were plated on BHI agar, and cfus were counted after 24 h incubation in a candle jar at 37°C. The survival percentage was calculated by the cfus of the H 2 O 2 -challenged sample over those in controls. Experiments were executed in triplicate, and each was repeated at least three times independently.
Luciferase Activity Assay-Twenty-five microliters of 1 mM D-luciferin (Sigma) solution (in 1 mM citrate buffer, pH 6.0) was added to 100-l sample, and luciferase assays were performed using a TD 20/20 luminometer (Turner Biosystems, Sunnyvale, CA). The optical density of the samples (A 600 ) was measured with a 2100 visible spectrophotometer (Unico, Shanghai, China) and used to normalize the luciferase activity. All the measurements were done for triplicate samples and repeated at least three times.
Quantitative PCR-Total RNA was extracted from mid-log phase (A 600 ϳ 0.4 -0.5) cells using TRIzol reagent (Invitrogen) as recommended by the suppliers. After quality confirmation with a 1% agarose gel, the RNA was treated with RNase-free DNase (Promega, Madison, WI) and analyzed by PCR for possible chromosomal DNA contamination. cDNA was generated from 2 g of total RNA with random primers using Moloney murine leukemia virus reverse transcriptase (Promega) according to the supplier's instructions and used for qPCR amplification with the corresponding primers (supplemental Table S4). Amplifications were performed with a Mastercycler ep realplex 2 (Eppendorf, Germany). To estimate copy numbers for a given mRNA, a standard curve of the tested gene was generated by quantitative PCR using 10-fold serially diluted PCR product as the template. The 16S rRNA gene was used as the biomass reference. The copy number of mntA gene was normalized to the number of 16S rRNA copies. The number of copies of the transcript of each gene per 1000 16S rRNA copies is shown.
Author Contributions-Z. C. expressed MntR, constructed the mutants, assayed redox status of MntR protein, and performed the H 2 O 2 survivals of the strains. X. W., F. Y., and Q. H. prepared and characterized the MntR disulfide-linked peptides. H. T. and X. D. supervised the project and wrote the paper with input and approval from all authors.