Previous studies have identified a putative mycothiol peroxidase (MPx) in Corynebacterium glutamicum that shared high sequence similarity to sulfur-containing Gpx (glutathione peroxidase; CysGPx). In the present study, we investigated the MPx function by examining its potential peroxidase activity using different proton donors. The MPx degrades hydrogen peroxide and alkyl hydroperoxides in the presence of either the thioredoxin/Trx reductase (Trx/TrxR) or the mycoredoxin 1/mycothione reductase/mycothiol (Mrx1/Mtr/MSH) regeneration system. Mrx1 and Trx employ different mechanisms in reducing MPx. For the Mrx1 system, the catalytic cycle of MPx involves mycothiolation/demycothiolation on the Cys36 sulfenic acid via the monothiol reaction mechanism. For the Trx system, the catalytic cycle of MPx involves formation of an intramolecular disulfide bond between Cys36 and Cys79 that is pivotal to the interaction with Trx. Both the Mrx1 pathway and the Trx pathway are operative in reducing MPx under stress conditions. Expression of mpx markedly enhanced the resistance to various peroxides and decreased protein carbonylation and intracellular reactive oxygen species (ROS) accumulation. The expression of mpx was directly activated by the stress-responsive extracytoplasmic function-σ (ECF-σ) factor [SigH]. Based on these findings, we propose that the C. glutamicum MPx represents a new type of GPx that uses both mycoredoxin and Trx systems for oxidative stress response.
Reactive oxygen species (ROS) are produced continually as a natural by-product of aerobic metabolism in cells and their production can be enhanced by various environmental insults, such as oxidants, antibiotics, acids and heavy metals. Consequently, aerobic organisms have developed multiple defences via antioxidant enzymatic and non-enzymatic mechanisms to withstand ROS-induced oxidative damage . Low-molecular-mass thiols (LMWTs) belonging to non-enzymatic systems act as redox buffers essential for cell defence against ROS, maintaining the reducing state of the cyto-plasm. Eukaryotes and Gram-negative bacteria produce the tripeptide glutathione (GSH; γ-L-glutamyl-L-cysteinylglycine) as the LMWT redox buffer [2,3], whereas some Gram-positive bacteria, such as members of Corynebacterium, Mycobacterium, Rhodococcus and Streptomyces genera, produce the related redox buffer mycothiol (MSH; AcCys-GlcN-Ins) [4,5]. As the dominant LMWT restricted to the high-(G+C)-content Gram-positive Actinobacteria, MSH has been regarded as a functional equivalent of GSH in these species and plays important roles in cytosolic redox homoeostasis and in adapting to ROS [6,7]. In addition, MSH has been reported to be involved in detoxification of a broad range of poisonous chemicals, such as oxidants, electrophiles, antibiotics, aromatic compounds, heavy metals and ethanol [8,9]. In Mycobacterium tuberculosis, MSH acts in combination with mycothiol disulfide reductase (MR) and mycoredoxin-1 [Mrx1, a glutaredoxin (Grx)-like mycothiol-dependent oxidoreductase] as a biologically relevant reducing system for MtAhpE . Moreover, protein S-mycothiolation was recently discovered as an important thiol protection and redox switch mechanism in response to oxidative stress in Corynebacterium glutamicum .
The cellular enzymatic antioxidant defence system, including superoxide dismutase (SOD), GSH reductase (GR), thioredoxin (Trx), Grx and peroxidase, has been identified. Of these, peroxidases, such as catalase and GSH peroxidase (GPx), play important roles in detoxification of peroxides, including the simplest H2O2 and larger cumene peroxides (Cumene-OOH). Many catalases do not react with larger compounds such as organic t-butyl hydroperoxides (t-butyl-OOH), whereas GPxs can act on both H2O2 and larger hydroperoxides using various reducing powers. Owing to their significantly high affinity for the peroxide, GPxs rather than catalases are generally considered to be the principal antioxidant enzymes for detoxifying H2O2 [12,13]. According to their amino acid sequences and catalytic redox centres, GPxs are divided into two principal subfamilies: selenoperoxidases (SecGPxs) and sulfur-containing peroxidases (CysGPxs) . SecGPxs are typical of vertebrates and sporadic in protists and bacteria, which include soluble tetrameric proteins commonly using GSH as electron donor. CysGPxs are widely distributed in Nature and have been found in terrestrial plants, insects, bacteria, fungi and protozoa, which encompass almost exclusively monomeric proteins using some redoxins, i.e. Trx or Trx-related proteins, as electron source . Although GPxs have been extensively studied biochemically and physiologically in GSH-containing organisms, their MSH-dependent paralogues in MSH-containing Gram-positive bacteria remain to be identified.
C. glutamicum, a well-known amino acid producer in industry, is also widely used as a model organism for systems biology . As reported, C. glutamicum employs multiple strategies to ensure survival under oxidative stress conditions, including a variety of antioxidant enzymes and MSH [6,16,17]. But surprising observations emerging in the previous reports showed that, although MSH-deficient C. glutamicum mutants were highly sensitive to peroxides, MSH did not directly react with peroxides to protect the cells. Thus the existence of MSH-dependent peroxidases in C. glutamicum has been suggested [6,16,17]. A recent study has identified an S-mycothiolated GPx (NCgl2502, annotated as GPx) under sodium hypochlorite (NaOCl) stress using thiol-redox proteomics and MS in C. glutamicum . This study suggested that this putative GPx might be the MSH-dependent peroxidase that utilizes MSH as electron donor for peroxide detoxification. This study contributes to a deeper characterization of the C. glutamicum mycothiol peroxidase (MPx). Evidence that the MPx can protect against the damaging effects of ROS induced by oxidative stresses was presented and biochemical studies to investigate the capacity and mechanisms of MSH, Mrx1 and Trx to serve as reducing power for MPx were conducted. Based on these results, a catalytic model for the reduction in MPx by Mrx1 and Trx is proposed.
Bacterial strains and culture conditions
Bacterial strains and plasmids used in the present study are listed in Supplementary Table S1. C. glutamicum and Escherichia coli strains were cultured in Luria–Bertani (LB) broth aerobically on a rotary shaker (220 rev/min) or on LB plates at 30°C or 37°C respectively. For generation of mutants and maintenance of C. glutamicum, BHIS (brain-heart infusion supplemented with 0.5 M sorbitol) medium was used . In this medium, sorbitol was used to create an incomplete cell wall and to improve the transformation efficiency of C. glutamicum competent cells and brain heart broth and glycerol were used to compensate for the growth defects induced by sorbitol. To construct the Δmpx in-frame deletion mutant, the pK18mobsacB-Δmpx plasmid was transformed into C. glutamicum wild-type (WT) by electroporation [19,20]. Integration of the introduced plasmid into C. glutamicum chromosome by single cross-over was selected on BH (brain-heart broth) plates containing 25 μg/ml kanamycin and 40 μg/ml nalidixic acid. The kanamycin-resistant (KmR) colonies were grown overnight in LB broth allowing for a second cross-over to occur. Selection for loss of the genome integrated sacB-containing plasmid was performed on LB plates containing 20% sucrose and 40 μg/ml nalidixic acid. Strains growing on this plate were tested for kanamycin sensitivity (KmS) by parallel picking on LB plates containing either kanamycin or sucrose. Kanamycin-sensitive and sucrose-resistant strains were tested for deletion by PCR using the DMPx-F1/DMPx-R2 primer pair (Supplementary Table S2) and confirmed by DNA sequencing. The Δmrx1 deletion mutant was similarly constructed by using the pK18mobsacB-Δmrx1 plasmid. To test the susceptibility of C. glutamicum strains to peroxides, overnight cell cultures were diluted 100-fold with fresh LB medium and exposed to 100 mM hydrogen peroxide (H2O2), 11 mM Cumene-OOH and 120 mM linoleic acid hydroperoxides (LA-OOHs) for 30 min at 30°C with shaking respectively. The cultures were serially diluted and plated on to LB agar plates, and then the survival percentage was calculated as reported in .
Cloning, heterologous expression and recombinant protein purification
The genes coding for MPx (NCgl2502) were amplified by PCR using C. glutamicum genomic DNA as template with indicated primers listed in Supplementary Table S2. The amplified DNA fragments were digested and then subcloned into a similarly digested pET28a plasmid, obtaining plasmid pET28a-mpx. To construct the deletion plasmid pK18mobsacB-Δmpx, a 986-bp upstream fragment and a 935-bp downstream fragment of mpx were amplified using primer pairs DMPx-F1/DMPx-R1 and DMPx-F2/DMPx-R2 respectively (Supplementary Table S2). In the next step, the upstream and downstream PCR fragments were fused together with the primer pair DMPx-F1/DMPx-R2 by overlap PCR . The resulting DNA fragments were digested with HindIII/EcoRI and inserted into a similarly digested suicide plasmid pK18mobsacB  to create pK18mobsacB-Δmpx. Plasmid pK18mobsacB-Δmrx1 was constructed by a similar approach using primer pairs DMrx1-F1/DMrx1-R1 and DMrx1-F2/DMrx1-R2 (Supplementary Table S2). The lacZ fusion reporter vector pK18mobsacB-Pmpx::lacZ was made by fusion of the mpx promoter to the lacZY reporter gene via overlap PCR . To make sure the whole mpx promoter was included, a 1000-bp sequence upstream of the mpx start codon was amplified with primer pair PMPx-F1/PMPx-R. In the second round of overlap PCR, the promoter fragment was fused with the lacZY fragment (amplified with lacZY-F/lacZY-R) by using primers PMPx-F1 and lacZY-R. The resulting PCR fragments were digested with SmalI/PstI and inserted into the SmalI/PstI site of the suicide vector pK18mobsacB to get the pK18mobsacB-Pmpx::lacZ fusion construct. Site-directed mutagenesis was carried out by overlap PCR to make the active-site cysteine residue at position 36 of MPx into a serine residue (MPx:C36S). The mutant mpx:C36S DNA segment was amplified by two rounds of PCR. Primer pairs DMPx-F1/MPx-C36S-R and MPx-C36S-F/DMPx-R1 were used to amplify segments 1 and 2, containing upstream and downstream flanking sequences of the mpx ORF respectively. The second round of PCR was carried out using the primer pair MPx-F/MPx-R with fragments 1 and 2 as templates, to generate the mpx:C36S segment containing a mutation in the Cys36 site of MPx. The mpx:C36S DNA fragment was digested and cloned into similar digested plasmid pET28a, obtaining plasmid pET28a-mpx:C36S. The mpx:C64S, mpx:C79S, mpx:C64SC79S, mpx:C36SC64S, mpx:C36SC79S, mrx1:C12S, mrx1:C15S, trx:C32S and trx:C35S DNA fragments were obtained following a similar procedure as described above with primers listed in Supplementary Table S2. For overexpression or complementation in C. glutamicum strains, mpx and mpx:C64SC79S DNA fragments were digested and cloned into similarly digested pXMJ19 vector  to yield corresponding pXMJ19 derivatives and transformed into C. glutamicum strains by electroporation . Expression in C. glutamicum was induced by addition of 0.5 mM IPTG into cultures. For overexpression of His6-tagged proteins in E. coli, the obtained DNA fragments were cloned into plasmid pET28a, obtaining corresponding expression plasmids. These pET28a derivatives were transformed into E. coli BL21(DE3) host strain, and recombinant proteins were prepared essentially as described previously [19,24]. Protein concentration was determined using the Bradford assay (Bio-Rad Laboratories) according to the manufacturer's instructions, with BSA as standard. The fidelity of all constructs was confirmed by DNA sequencing (Sangon Biotech).
Peroxidase activity assays were performed by monitoring the decrease in absorbance at 340 nm arising from NADPH oxidation [25,26]. The catalytic properties of MPx were determined using a reduced Trx-generating system [4 μM Trx reductase (TrxR) and 40 μM Trx], Mrx1 system [4 μM mycothione reductase (Mtr), 240 μM MSH and 40 μM Mrx1] or MSH system (4 μM Mtr and 500 μM MSH) as the possible electron donors. These assays were carried out in a total volume of 500 μl containing 50 mM Tris/HCl buffer (pH 7.5), 1 mM EDTA, 250 μM NADPH, 0.5 μM MPx (WT or its variants) and the electron donor (Trx system, Mrx1 system or MSH system). The reactions were started by the addition of 200 or 500 μM peroxide substrates following 5 min of pre-incubation. The catalytic parameters for one substrate have been obtained by varying its concentration at saturating concentrations of the other substrate (between 0 and 150 μM for Trx or Mrx1, between 0 and 3.0 mM for MSH, and between 0 and 1000 μM for peroxide substrates). NADPH oxidation was monitored as A340. The activity was determined after subtracting the spontaneous reduction rate observed in the absence of MPx and the number of micromoles of NADPH oxidized per second per micromole of enzyme (i.e. turnover number, s−1) was calculated using the molar absorption coefficient of NADPH at 340 nm (ε340) of 6220 M−1·cm−1. Three independent experiments were performed at each substrate concentration. The kcat and Km values of MPx for Trx, Mrx1, MSH or peroxide substrates were obtained from a non-linear fit with the Michaelis–Menten equation using the program GraphPad Prism 5.
Peroxidase activity was also detected by using the ferrous Xylenol Orange (FOX) assay . The consumption of peroxide was measured kinetically at different time points during the initial linear rate of catalysis. The reaction mixture contained the same components as described above. After the reactions were initiated by the addition of peroxide, 16.7 μl of HCl (1 M) was added into 83.3 μl of reaction mixture to stop the reaction at different intervals. Then, the resulting mixture (100 μl) was mixed with 900 μl of FOX reagent (88 mg of butylated hydroxytoluene, 7.6 mg of Xylenol Orange and 9.8 mg of (NH4)2Fe(SO4)2·6H2O added to 90 μl of methanol and 10 μl of 250 mM sulfuric acid) and incubated at 37°C for 30 min. The colour developed was read colorimetrically at 560 nm. All assays were performed three times.
Examination of redox and oligomerization states
The oligomerization state of native MPx was determined by 15% native PAGE . The purified native MPx was mixed with the loading buffer [250 mM Tris/HCl (pH 6.8), 0.5% Bromophenol Blue and 50% (v/v) glycerol], separated by 15% PAGE without SDS buffer (pH 8.8) and then stained with Coomassie Brilliant Blue R-250.
The oligomerization state of the native MPx was also determined by gel filtration using HPLC equipped with a Superdex 75 10/300 GL column (GE Healthcare). Proteins (100 μl) at a concentration of 5.62 mg/ml were loaded on to the column equilibrated in 50 mM potassium phosphate buffer (pH 7.2) including 0.15 M NaCl. The flow rate was fixed at 0.5 ml/min and the detection was recorded at 280 nm. The molecular mass standards used were aprotinin (6.5 kDa), ribonuclease A (13.7 kDa), carbonic anhydrase (29 kDa), ovalbumin (44 kDa) and conalbumin (75 kDa) (GE Healthcare). The Kav values for each of the standard protein were calculated using the equation: Kav=(Ve−V0)/(Vc−V0), where V0=column void volume, Ve=elution volume and Vc=geometric column volume (24 ml). The elution volume of 8.61 ml for Blue Dextran 2000 is equal to the V0. The plots of Kav values obtained from the above known standard protein were against the log of the molecular mass of the standard protein to form the calibration curve. These data were suited with a linear equation. The molecular mass of an unknown protein could be determined from the calibration curve once its Kav value was calculated from its measured elution volume.
The redox state of MPx WT (50 μM) was analysed by incubating the proteins with 50 mM DTT or 200 μM H2O2 for 30 min before separating on non-reducing SDS/PAGE (15% gel). For non-reducing conditions, DTT and β-mercaptoethanol were omitted from the above loading buffer and 50 mM iodoacetamide (IAM) was included. All samples were boiled for 5 min prior to electrophoresis.
Analysis of sulfenic acid and disulfide bond formation
The formation of Cys-SOH as a reaction intermediate was examined by the assays of MPx:C64SC79S, MPx:C36SC64S and MPx:C36SC79S proteins labelled with 4-chloro-7-nitrobenzofurazan (NBD-Cl) [26,29,30]. Protein was pre-incubated with 50 mM DTT for 30 min and then excess DTT was removed by ultrafiltration. MPx:C64SC79S, MPx:C36SC64S and MPx:C36SC79S proteins in buffer (50 mM potassium phosphate with 1 mM EDTA at pH 7.0) were prepared anaerobically by repeatedly flushing with argon gas and vacuum in alternating cycles for 20 min. Anaerobic solution of NBD-Cl (25 mM in DMSO) was prepared by bubbling argon through the preparation for 10 min. Under anaerobic conditions, MPx:C64SC79S, MPx:C36SC64S and MPx:C36SC79S proteins were divided into seven portions (final concentration 25 μM), five for oxidation treatment with various concentrations of H2O2, from 25 to 500 μM respectively and the other two for direct use as the reduced samples (negative control). Treated or untreated proteins were incubated with 5 mM NBD-Cl for 30 min at 25°C in the dark. Excess NBD-Cl was removed by ultrafiltration, and the absorbance of protein samples was analysed (200–600 nm) on a Beckman DU 7500 diode array spectrophotometer (Fullerton).
The formation of disulfide bond and sulfenic acid was measured by covalent modification with the thiol-reactive probe 4-acetamido-4′-maleimidyldystilbene-2,2′-disulfonic acid (AMS; Molecular Probes) . Briefly, MPx WT and its variants (50 μM) were reduced by incubation with 50 mM DTT at room temperature for 30 min and excess DTT was removed by ultrafiltration. The resulting protein was treated with 200 μM H2O2 at room temperature for 30 min or directly used as negative control. The resulting proteins were precipitated in 10% (w/v) trichloroacetic acid (TCA) and then redissolved in 100 mM Tris/HCl (pH 8.0) containing 1% SDS and 15 mM AMS. After incubation for 30 min in the dark, 2 μg aliquots of proteins were separated by non-reducing SDS/PAGE.
Formation and separation of heterodimers
The assays of heterodimers were performed based on the method described by Rouhier et al.  with minor modifications. MPx (WT and its variants; final concentration 25 μM) and Trx or Mrx1 (WT and its variants; final concentration 15 μM) were mixed in TE buffer (30 mM Tris/HCl, pH 8.0, and 1 mM EDTA) to a final volume of 20 μl. This reaction mixture was incubated at room temperature for 15 min before the addition of 50 μM H2O2. The mixture was incubated at room temperature for another 30 min and then subjected to non-reducing SDS/PAGE (15% gel).
Quantitative analysis of thiol groups
Free thiol groups in MPx WT and its variants were determined using 5,5′-dithio bis-(2-nitrobenzoic acid) (DTNB) . Proteins (50 μM) were treated with 500 μM H2O2 and 50 mM DTT at room temperature for 30 min respectively, followed by removing DTT or H2O2 by a PD10 desalting column (GE Healthcare). The resulting proteins (10 μM) were added to DTNB (2 mM) in 50 mM Tris/HCl buffer (pH8.0) and the absorbance at 412 nm was measured against a 2 mM DTNB solution as the reference. The amounts of reactive thiol groups were determined using the molar absorption coefficient of 5-thio-2-nitrobenzoic acid (TNB) at 412 nm (ε412) of 13600 M−1·cm−1 .
Reduced Mrx1:C15S, Trx:C35S, MPx (WT and its variants), oxidized MPx variants and the dead-end intermediate between MPx and Trx variants were incubated with 10 mM IAM for 15 min at room temperature. The alkylated proteins were subjected to non-reducing SDS/PAGE and Coomassie Brilliant Blue-stained bands were excised, in-gel digested with trypsin and analysed by MALDI–TOF–MS/MS (Voyager-DE STR, Applied Biosystems).
Detection of S-mycothiolated MPx in vitro
For the in vitro test, S-mycothiolated MPx variants were produced according to the method of Chi et al. . MSH was purified from C. glutamicum RES167 as described previously [5,35]. MPx:C64S (50 μM) was incubated with an excess of reduced MSH (1 mM) prior to the addition of 200 μM H2O2. After 30 min of incubation, the sample was loaded on Ni2+-nitrilotriacetate (Ni-NTA) His·Bind resin-containing column (Novagen), and MSH and H2O2 were removed by washing (50 mM HEPES, pH 8.0, 500 mM NaCl). S-mycothiolated MPx:C64S was eluted in the same buffer containing 300 mM imidazole. The resulting proteins were analysed by MALDI–TOF–MS/MS.
Measurement of intracellular ROS levels and determination of cellular protein carbonylation
In vivo ROS levels were measured using the fluorogenic probe 2′,7′-dichlorofluorescein diacetate (DCFH-DA) as described by Schurig-Briccio et al. . Protein carbonylation assays were performed based on the method described by Vinckx et al. . For Western blots, samples resolved by SDS/PAGE were transferred on to PVDF membranes and detected with anti-dinitrophenyl antibody (1:500 dilution; Millipore).
Construction of chromosomal fusion reporter strains and β-galactosidase assay
The lacZ fusion reporter plasmid pK18mobsacB-Pmpx::lacZ was transformed into relevant C. glutamicum strains by electroporation. The chromosomal pK18mobsacB-Pmpx::lacZ fusion reporter strain was selected by plating on LB agar plates supplemented with kanamycin. β-Galactosidase activities were assayed after exposure to different peroxides at indicated concentrations for 30 min, with ONPG (o-Nitrophenyl β-D-galactopyranoside) as the substrate .
EMSA was performed using the method of Si et al. . To reduce non-specific binding in the EMSA assay, a shorter DNA promoter probe (Pmpx; 400 bp) containing the predicted SigH [the stress-responsive extracytoplasmic function-σ (ECF-σ) factor]-binding site was amplified from the 1000-bp mpx promoter region of the pK18mobsacB-Pmpx::lacZ reporter vector using primers PMPx-F2/MPx-R (Supplementary Table S2). Increasing concentrations of purified His6–SigH (0–4 μg) were incubated with 20 ng of DNA probes in EMSA buffer (20 mM Tris/HCl, pH 7.4, 4 mM MgCl2, 100 mM NaCl, 1 mM DTT and 10% glycerol). The binding reaction mixture incubated for 30 min at room temperature was subjected to electrophoresis on a 6% native polyacrylamide gel containing 5% glycerol in 0.5×Tris/borate/EDTA (TBE) electrophoresis buffer and the DNA probe was detected with SYBR Green. As negative controls, a 240-bp fragment from the mpx coding region amplified with primers Control-F and Control-R instead of the 400-bp mpx promoter and BSA instead of His6–SigH were included in the binding assays.
RESULTS AND DISCUSSION
Roles of MPx as an antioxidant in C. glutamicum
The physiological role of MPx in detoxifying peroxides was assessed in the Δmpx mutant by cell viability assays. Concentrations of peroxides applied were able to reduce the survival rate of the WT and cause the mortality rate from 30% to 50% (Supplementary Figure S1). Compared with the WT, the survival rates of the Δmpx mutant decreased ~20.3–35.6% to H2O2 (100 mM), Cumene-OOH (11 mM) and LA-OOH (120 mM) treatment (Figure 1A). However, the hypersensitive phenotype was completely reversed in the complementary strain. In addition, mpx overexpression markedly increased the resistance of the WT strain to H2O2, Cumene-OOH and LA-OOH challenge (Figure 1A). Next, we investigated the role of MPx in ROS reduction upon oxidative stress. ROS levels were examined in various strains using DCFH-DA, a membrane-permeant dye which passively diffuses into cells . After 30 min of exposure to H2O2, Cumene-OOH and LA-OOH, the Δmpx mutant showed obviously higher ROS levels as compared with the WT (P≤0.05; Figure 1B). Moreover, the ROS level in the Δmpx mutant was completely restored to that of the WT by introducing a plasmid overexpressing mpx (Figure 1B). These data demonstrate that MPx can perform an antioxidant function by reducing intracellular ROS accumulation induced by various oxidants.
MPx was required for cellular resistance to oxidant-induced stresses in C. glutamicum
To test whether MPx protects cells by reducing oxidative damage to proteins, the carbonylation assays were performed on total proteins isolated from Δmpx and WT cells grown under oxidative stresses. As expected, the carbonylation level of protein extracts was significantly lower in the WT than in the Δmpx mutant after 30 min of exposure to H2O2, Cumene-OOH and LA-OOH (Figure 1C).
Since MPx was shown to promote the survival of C. glutamicum upon oxidative stress, mpx expression in response to various oxidants was investigated by chromosomal Pmpx::lacZ fusion reporter analysis. Compared with the untreated control, the promoter activity of Pmpx in the WT increased by 38.7%, 44.5% and 32.3% upon H2O2 (35 mM), Cumene-OOH (5.5 mM) and LA-OOH (60 mM) treatment respectively (Figure 1D). Furthermore, the expression of the Pmpx::lacZ fusion exhibited a dose-dependent increase in response to the oxidants tested (Figure 1D). The strong induction of mpx by various oxidants suggests that a stress sensor/transcriptional regulator might be involved in the regulation of its expression. As SigH, the stress-responsive ECF-σ factor, was reported to increase the resistance of C. glutamicum to multiple stresses and regulate the expression of many oxidative stress-resistance genes [39,40], we further examined the role of SigH in mpx gene expression. After 30 min of exposure to oxidants, a marked decrease in the level of mpx promoter activity was observed in the ΔsigH mutant compared with the WT strain (Figure 1D). Indeed, the decreased mpx expression in the ΔsigH mutant was even observed in the absence of H2O2, Cumene-OOH and LA-OOH treatment. Moreover, the mpx expression in the ΔsigH mutant was fully recovered when the regulatory protein was complemented (Figure 1D). The SigH-dependent mpx expression was further confirmed by determining the direct interaction between His6–SigH and mpx promoter with EMSA assay (Figure 1E). Incubation of His6–SigH with a 400-bp PCR fragment (Pmpx) led to retarded mobility of the probe and the DNA–protein complexes increased in response to more His6–SigH used in the reactions (Figure 1E), indicating direct binding of this protein to the mpx promoter. A 240-bp control DNA fragment amplified from the mpx coding ORF region showed no detectable His6–SigH binding (Figure 1E). Collectively, these results indicate that multiple peroxides induce the expression of mpx, which, in turn, directly contributes to cell tolerance to these adverse stresses.
Cys36 is the peroxidative cysteine that is oxidized to sulfenic acid
Amino acid sequence comparison showed that MPx had the same catalytic triad with all reference GPxs, comprising the cysteine/selenocysteine residue co-ordinated to the glutamine, tryptophan and asparagine. The residues of the catalytic triad were distributed in distant parts of the sequence and extremely conserved in the whole GPx superfamily, with typical characteristics being NVA(S/T)X(U/C36)GXT, FPC64NQFXXQEP and WNFXKFLV (Supplementary Figure S2) [14,41]. This suggests that the conserved Cys36 of MPx may be the peroxidatic cysteine (CP).
In addition, MPx was strongly homologous to CysGPxs (showing 31%, 29% and 23% amino acid identities to CysGPxs from Trypanosoma brucei, Plasmodium falciparum and Drosophila melanogaster respectively), but not to SecGPxs (the highest similarity to human SecGPx was only 11%; Supplementary Figure S2). This is mainly because MPx had another conserved cysteine residue of a designated ‘cysteine block’ [also called the resolving residue (CR)] at position 79, which was found in many Trx-dependent monomeric CysGPxs but not generally possessed by SecGPxs. Moreover, MPx sequence manifested a major deletion, which was the subunit interface (designated ‘tetramer interface’ and ‘dimer interface’) of SecGPxs and similarly seen in monomeric CysGPxs (Supplementary Figure S2) [14,41]. The above phenomena suggested that MPx not only was monomeric, as verified by native PAGE and gel filtration chromatography of the native protein (Supplementary Figures S3A–S3C), but also possessed Trx-dependent peroxidase activity. In addition, H2O2-treated MPx still appeared as monomer on non-reducing SDS/PAGE gel, suggesting that MPx was functionally monomeric and did not form disulfide-bonded dimer in clearing peroxides (Supplementary Figure S3D).
The CP residue has been reported to be first oxidized by peroxide to a cysteine sulfenic acid (CP-SOH) intermediate . During catalysis, the labile peroxidatic Cys-SOH is easily attacked by the other cysteine of GPx to form the redox-active disulfide. To trap Cys-SOH, the C-terminal cysteine of the active-site disulfide pair must be removed. So the double variants of MPx (i.e. MPx:C64SC79S, MPx:C36SC64S and MPx:C36SC79S) were constructed and treated with NBD-Cl under anaerobic conditions with and without previous exposure to different concentrations of H2O2 to identify whether Cys36 is the most sensitive cysteine residue and oxidized to sulfenic acid.
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 . Following the reaction with NBD-Cl, the absorption spectra of the MPx:C36SC64S and MPx:C36SC79S proteins were unchanged before or after exposure to different concentrations of H2O2 (Supplementary Figure S4), exhibiting only the 420 nm peak. DTNB assay for free thiol contents also showed that the DTT-treated MPx:C36SC64S and the MPx:C36SC79S proteins contained 0.89±0.2 and 0.99±0.4 thiol groups per monomer (Figure 2B) respectively, equal to the thiol contents measured in the H2O2-treated MPx:C36SC64S and MPx:C36SC79S proteins. These results indicate that no sulfenic acid is formed in Cys64 or Cys79 and that Cys64/Cys79 still exist in thiol state in the H2O2-treated MPx:C36SC64S or MPx:C36SC79S proteins.
The thiol content and form of DTT-treated or H2O2-treated MPx in C. glutamicum
The MPx:C64SC79S variant (25 μM) with less than 50 μM H2O2 treatment showed Soret bands 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 Cys36 was partly oxidized to a sulfenic acid form (Figure 2A). The 100 μM H2O2-treated and NBD-labelled MPx:C64SC79S variant had a specific absorbance maximum (λmax) of 347 nm, which clearly signified the detection and trapping of approximately stoichiometric amounts of SOH at Cys36. The decrease or the disappearance of this signal upon the treatment with 250 and 500 μM H2O2 respectively, probably indicated that Cys36 was over-oxidized to sulfinic or sulfonic acid forms. Consistently, the thiol content of the DTT-treated MPx:C64SC79S monomer was significantly higher than the H2O2-treated MPx:C64SC79S monomer and the difference of 0.70 thiol content between the H2O2-treated and the DTT-treated MPx:C64SC79S revealed that the H2O2-treated MPx:C64SC79S contained no free thiol groups (Figure 2B). The redox state of thiol was further tested by detection of free thiol groups using AMS covalent modification. AMS covalently modifies thiol groups of protein irreversibly. Since the molecular mass of AMS is 0.5 kDa, the apparent delay of electrophoretic mobility would occur in proportion to the number of free thiol groups in proteins . The H2O2-treated AMS-modified MPx:C36SC64S and MPx:C36SC79S migrated the same as their H2O2-untreated AMS-modified states (Figure 2C), indicating that MPx:C36SC64S and MPx:C36SC79S variants were not oxidized by H2O2, of which Cys64 or Cys79 still existed in the thiol state after exposure to H2O2. However, the electrophoretic migration state of the H2O2-treated MPx:C64SC79S protein modified with or without AMS was the same, indicating that Cys36 did not exist in the thiol state in the H2O2-treated MPx:C64SC79S variant (Figure 2C, lanes 3 and 4). These results suggest that the Cys36 residue is the peroxidative cysteine that can be oxidized to Cys-SOH.
Cys36 and Cys79 form an intramolecular disulfide bond under oxidative stress
The disulfide bond formation in the MPx WT and its single cysteine mutants incubated with and without H2O2 was analysed by reducing or non-reducing SDS/PAGE, AMS modification and the free thiol content assay. As shown in Figure 2(D), the MPx:C64S variant treated with H2O2 and modified with AMS, migrated faster than its non-H2O2-treated AMS-modified form and the same as the H2O2-treated non-AMS-modified form, suggesting that there were a lack of thiols and the formation of disulfide bond between Cys36 and Cys79 in this protein (Figure 2D). Both the H2O2-treated and the H2O2-untreated MPx:C36S proteins were modified by AMS and had the same electrophoretic mobility (Figure 2D, lanes 2 and 4). This observation indicates that MPx:C36S was not oxidized by H2O2, of which Cys64 and Cys79 still existed in the thiol state after exposed to H2O2. Unexpectedly, besides the monomeric form, MPx:C79S showed the presence of covalent dimer upon high concentration of H2O2 (200 μM) treatment. Both the dimeric and the monomeric states of MPx:C79S treated with H2O2 and modified with AMS migrated slower than their non-AMS-modified forms, implying that the MPx:C79S variant is not fully oxidized, of which Cys64 remains in the thiol state and Cys36 is oxidized. Formation of the dimer was also observed in the H2O2-treated C64SC79S variant, though to a much lesser extent. These results suggest that Cys36 was involved in formation of non-specific dimers through intermolecular disulfide bonds. Although the physiological roles remain unrevealed, the formation of the non-specific dimers upon strong oxidative stress treatment has also been reported for the thiol-dependent peroxidase Ohr  and AhpC .
Consistent with the above observations, the difference of free thiol content in MPx:C64S proteins before and after H2O2 treatment was 1.73 (Figure 2B). However, the MPx:C36S and MPx:C79S proteins under H2O2 treatment lost zero and one thiol group respectively when compared with the thiol content of DTT-treated states (Figure 2B).
The formation of the disulfide bond in the H2O2-treated MPx:C64S was further confirmed by MS analysis, with the identification of a mass of 5687.3 Da lower by 2.01 Da than predicted for the sum of two peptides including the Cys36-containing 36–47 peptide (calculated mass 1336.7 Da) and the Cys36-containing 56–95 peptide of MPx:C64S (calculated mass 4351.8 Da), indicating that the 56–95 and 36–47 peptides were cross-linked by a disulfide bond (Supplementary Figure S5D). This mass was not observed in the DTT-treated MPx:C64S (Supplementary Figures S5B and S5C). Altogether, these results demonstrate that MPx is capable of forming a disulfide bond between the functionally important cysteine residues, i.e. Cys36 and Cys79.
MPx exhibits peroxidase activity using both Trx and Mrx1 reducing systems as electron donors
Since C. glutamicum contains three alternative physiological reducing systems, i.e. the MSH system (MSH/Mtr), the Mrx1 system (Mrx1/Mtr/MSH) and the Trx system (Trx/TrxR), we next compared their efficiency in providing reducing power for MPx. MPx activity was first measured using the MSH system as the electron donor coupled to NADPH oxidation. No activity was found with the MSH system at the concentration lower than 120 μM and the electron-providing role of the MSH system for MPx was only at a low rate and quite variable with values of (1.7–25.2)×102 M−1·s−1 using different concentrations of various peroxides (0–1000 μM) and 500 μM MSH (500 μM MSH was nearly the highest concentration in C. glutamicum; Supplementary Figures S6A—S6D; Supplementary Table S3) [5,6,44]. Under these conditions, peroxidase assays were also performed with a fixed concentration of oxidants (500 μM) and different concentrations of MSH (0–3.0 mM). The Km MSH values, Kcat MSH values and the catalytic efficiencies of MPx were calculated to be 1488±206 μM, 0.4±0.01 s−1 and 0.3×103 M−1·s−1 for Cumene-OOH; 1813±282 μM, 0.2±0.01 s−1 and 0.1×103 M−1·s−1 for H2O2; 2346±655 μM, 0.1±0.01 s−1 and 0.04×103 M−1·s−1 for LA-OOH (Supplementary Figure S6E; Supplementary Table S4). These data indicated that MPx had a very low affinity for MSH and presented apparent maximum velocities with MSH at the concentrations far above the estimated range of evaluated physiological values (500 μM) [5,6,44]. Therefore, the MSH system cannot act as the electron donor for MPx in vivo and the reactions were not evaluated in further detail.
Next we tested whether Mrx1 could reduce MPx in the presence of MSH and Mtr. When H2O2 and Cumene-OOH were served as the substrates, addition of Mrx1 to the MSH (240 μM)/Mtr/MPx system (Supplementary Figures S7A and S7B) resulted in an obvious reduction in NADPH compared with only the MSH/Mtr/MPx system, indicating that the Mrx1/MSH/Mtr system could effectively act as the electron donor for MPx during decomposition of H2O2 and Cumene-OOH. However, the role of the Mrx1/MSH/Mtr system in reducing MPx was undetectable when LA-OOH was used as the substrate, indicating that the Mrx1/MSH/Mtr system cannot act as the electron donor for MPx when LA-OOH was used as the substrate (Supplementary Table S4). As for the negative control, no reduction in NADPH was observed when MSH was omitted from the reaction mixture, indicating that Mrx1 alone was not sufficient to aid MPx in reducing peroxide without MSH. Similarly, the Mrx1/MSH/Mtr system alone could not effectively reduce peroxides in the absence of MPx (Supplementary Figures S7A and S7B). The kinetic parameters of MPx have been determined with a fixed concentration of Mrx1 (40 μM) and different concentrations of H2O2 and Cumene-OOH (0–1000 μM). The Km values, Kcat values and catalytic efficiencies of MPx were calculated to be 16.9±2.1 μM, 6.4±0.2 s−1 and 3.8×105 M−1·s−1 for Cumene-OOH; 21.3±3.0 μM, 4.9±0.1 s−1 and 2.3×105 M−1·s−1 for H2O2 (Figure 3A; Table 1).
Mrx1- and Trx-dependent peroxidase activities of MPx
|Substrate||Km (μM)||kcat (s−1)||kcat/Km × 105 (M−1·s−1)||Km (μM)||kcat (s−1)||kcat/Km × 105 (M−1·s−1)|
|Substrate||Km (μM)||kcat (s−1)||kcat/Km × 105 (M−1·s−1)||Km (μM)||kcat (s−1)||kcat/Km × 105 (M−1·s−1)|
As MPx has high sequence similarity to CysGPxs, we speculated that MPx could also catalyse the reduction in oxidants using Trx as the electron donor. As expected, a Trx-dependent peroxidase activity was found for MPx under oxidative stresses induced by H2O2, Cumene-OOH and LA-OOH (Supplementary Figure S7C). The catalytic constants of MPx with the Trx system as the recycling reductant were determined under steady-state conditions at saturating concentration of reductants (40 μM) and different concentrations of various peroxides (0–1000 μM). The Km values, Kcat values and catalytic efficiencies of MPx were 10.6±1.3 μM, 8.5±0.2 s−1 and 8.0×105 M−1·s−1 for Cumene-OOH; 14.8±2.8 μM, 5.9±0.2 s−1 and 4.0×105 M−1·s−1 for H2O2; 31.0±5.7 μM, 3.1±0.1 s−1 and 1.0×105 M−1·s−1 for LA-OOH (Figure 3B and Table 1). The role of Mrx1 and Trx in facilitating MPx activity was further corroborated by monitoring the consumption of peroxides with FOX assay (Supplementary Figures S7D–S7F).
Consistent with the prediction that Cys36 acts as the peroxidatic cysteine residue essential for MPx activity, no activity was observed for the MPx:C36S variant using either the Trx/TrxR or the Mrx1/Mtr/MSH system as reducing power (Figure 4). The resolving cysteine residue Cys79 was also sensitive to inactivation as the MPx:C79S variant showed significantly decreased activity in the Trx/TrxR system (Figure 4). However, the mutation of Cys79 had no effect on MPx activity in the Mrx1/Mtr/MSH system. Indeed, Cys36 alone was sufficient to MPx activity in the Mrx1/Mtr/MSH system, with the observation that the MPx:C64SC79S variant, although completely inactivated in the Trx/TrxR system, showed higher activity than that of MPx in the Mrx1/Mtr/MSH system. In addition, both the MPx:C36SC64S and the MPx:C36SC79S variants showed no detectable activity using either the Trx/TrxR or the Mrx1/Mtr/MSH system as the reducing power. Thus, it was clear that Mrx1 and Trx employed different mechanisms in reducing MPx. The role of Cys64 in MPx activity was negligible as the MPx:C64S variant showed comparable activity with MPx WT using either the Trx/TrxR or the Mrx1/Mtr/MSH system as reducing power.
Effect of cysteine mutation on catalytic activity of MPx
We then analysed the kinetic parameters of MPx WT and its variants (i.e. MPx:C64S, MPx:C79S and MPx:C64SC79S) (Supplementary Table S5). The Kcat and Km values of MPx:C64S variant were similar to those of MPx WT in both the Trx/TrxR and the Mrx1/MSH/Mtr systems, indicating that Cys64 was non-essential for the catalytic activity of MPx. The MPx:C79S and MPx:C64SC79S variants exhibited a decreased Kcat/Km ratio in the Trx/TrxR system (~100-fold lower than MPx WT), but an obviously increased Kcat/Km ratio in the Mrx1/MSH/Mtr system (~56- and 78-fold higher than MPx WT respectively.).
Next, enzymatic activities of MPx towards various substrates were analysed using the Mrx/MSH/Mtr and the Trx/TrxR system as the reducing power respectively (Table 1). Results showed that MPx was active with H2O2, Cumene-OOH, t-butyl-OOH, as well as LA-OOH and L-α-phosphatidylcholine dilinoleoyl (PCdili-OOH) in the Trx/TrxR system. Whereas MPx showed comparable activities towards H2O2, Cumene-OOH and t-butyl-OOH in the Mrx1/MSH/Mtr system, its activity towards LA-OOH and PCdili-OOH was undetectable, indicating that the Mrx1 system cannot act as the reducing power when MPx reduces these two lipid hydroperoxides. In both systems, the rate constant of the reaction of MPx was the highest with Cumene-OOH as the substrate. Therefore, MPx displayed the kinetic pattern characteristic of typical GPxs with Cumene-OOH as the substrate. From the above results, we supposed that both Mrx1 and Trx could act as the electron donor for MPx with H2O2, Cumene-OOH and t-butyl-OOH as the substrates and MPx exhibited different affinities for various substrates when using different reducing powers. In addition, MPx showed almost equal affinity for Mrx1 and Trx in vitro, as indicated by Km values when using H2O2 and Cumene-OOH as the substrates (Supplementary Table S4).
MPx forms a mixed disulfide with Trx
The formation of a transient intermolecular disulfide bond is the prerequisite for an efficient reaction with Trx . To explore the catalytic mechanism for Trx-dependent MPx reduction, MPx:C64S was incubated with three versions of Trx (WT, Trx:C32S and Trx:C35S) in the presence of H2O2. These mixtures were separated by non-reducing SDS/PAGE and evaluated based on the molecular mass determinations (Figure 5A). In all assays, the reduced MPx:C64S polypeptide was present with an apparent molecular mass of 22 kDa (band labelled +), the oxidized MPx:C64S was present with an apparent molecular mass of 20 kDa (band labelled #) and Trx was present with an apparent molecular mass of 16 kDa (band labelled -). An additional 38-kDa polypeptide occurred in the mixture of Trx:C35S, MPx:C64S and H2O2 (band labelled *; Figure 5A), which did not exist in H2O2-containing Trx:C35S or MPx:C64S (Supplementary Figure S8) and may correspond to the MPx–Trx heterodimer. Consistent with this assumption, a peak with a mass of 6946.7 Da from band labelled * and analysed by MALDI–TOF–MS/MS after in-gel digestion occurred, corresponding to the Cys79-containing the 56–95 peptide of MPx:C64S [calculated mass 4351.8 Da and observed mass 4546.8 Da formed by adding IAM (185 Da) into Cys79; Supplementary Figure S5B] and the Cys35-containing the 14–36 peptide of Trx:C35S (calculated mass 2595.8 Da and observed mass 2773.2 Da formed by adding IAM (185 Da) into Cys32; Supplementary Figure S6A) connected by a disulfide bridge was observed (Figure 5B). This peak was not observed in reduced Trx:C35S and MPx:C64S (Supplementary Figures S5A–S5C). These data demonstrate that Trx can reduce the disulfide bond in oxidized MPx and Cys79 is the favourable cysteine residue in the formation of the transient disulfide intermediate with Cys32 of Trx.
Validation of the interaction between MPx and Trx in vitro
Reduction in mycothiolated MPx by Mrx1 coupled with Mtr/MSH
To clarify the mechanism used by Mrx1 for regenerating MPx activity, three single variants (MPx:C36S, MPx:C64S and MPx:C79S), a double variant (MPx:C64SC79S) and MPx WT were incubated with three versions of Mrx1 (WT, Mrx1:C12S and Mrx1:C15S) respectively, in the presence of H2O2. The control samples with just one type of protein were also included in the assay (Supplementary Figure S8). However, no heterodimer was detected between any one of the Mrx1 and MPx (WT and its variants). Moreover, NADPH consumption was not observed in the oxidized MPx:C64S (containing disulfide bonds) and the oxidized MPx:C64SC79S (containing sulfenic acid) coupled to the Mrx1/MSH/Mtr system (results not shown). These results indicate that Mrx1 directly reduces neither MPx-containing disulfide bond nor MPx-containing sulfenic acid (Figure 6A).
MPx was reduced by the Mrx1 system via a monothiolic mechanism
Recently, Chi et al.  identified the S-mycothiolation (mixed disulfides between protein thiol groups and MSH) modification of MPx under hypochlorous acid stress in a thiol-redox proteomics analysis . As shown in Figure 6(B), S-mycothiolation formed on the catalytic Cys36 was also detected by peptide MS after tryptic digestion of H2O2/MSH-treated MPx:C64S in vitro, with the identification of a mass of 1820.0 Da that was 484 Da higher than the Cys36-containing 36–47 peptide of the reduced MPx:C64S (calculated and observed mass 1336.6 Da; corresponding to CGLTPQYEGLQK), consistent with the results from the addition of MSH to MPx on Cys36 (Figure 6B, upper panel). However, this modification did not occur after the treatment with MSH alone (Figure 6B, lower panel). This finding promoted us to test whether Mrx1 regenerates MPx by the mycothiolation/demycothiolation monothiol mechanism.
The monothiol mechanism of action required only the N-terminal cysteine of the C12XXC active site in Mrx1 and was employed for reducing mixed disulfides between MSH and the functional cysteine residue of the target protein [6,45]. Thus, studies of the interaction between MPx:C64SC79S-SSM and Mrx1/Mrx1:C15S and reduction activity of S-mycothiolated MPx:C64SC79S with the Mrx1/MSH/Mtr and Mrx1:C15S/MSH/Mtr electron transfer pathway as reducing substrate in vitro was performed. The enzyme activity confirmed that the S-mycothiolated but not the non-mycothiolated MPx:C64SC79S could be reduced with the Mrx1/MSH/Mtr and Mrx1:C15S/MSH/Mtr electron transfer pathway (Figure 6C). As shown in Figure 6(C), only the sample with MPx:C64SC79S-SSM coupled to the Mrx1/MSH/Mtr and Mrx1:C15S/MSH/Mtr electron transfer pathway showed consumption of NADPH, but NADPH consumption was not observed when Mrx1/Mrx1:C15S was omitted from the assay, or when non-mycothiolated MPx:C64SC79S was used instead of S-mycothiolated MPx:C64SC79S. Interestingly, the complex between Mrx1:C15S and MPx:C64SC79S-SSM was not observed (Figure 6D), indicating that the cysteine residue in Mrx1:C15S reacts with the sulfur of MSH in MPx:C64SC79S-SSM to produce reduced MPx through the reaction for a monothiolic mechanism of reduction, in line with the result reported for M. tuberculosis AphE recently .
To further explore the role of S-mycothiolation in MPx reduction, we performed the NBD-Cl labelling assay. 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 . As expected, the reduced MPx:C64SC79S (MPx:C64SC79S-SH) reacted with NBD-Cl by forming species with an absorbance maximum at 420 nm. Whereas MPx:C64SC79S-SOH reacted with NBD-Cl by forming species with an absorbance maximum at 347 nm in the absence of MSH, no NBD-Cl labelling occurred in the presence of MSH (Figure 6E). These data suggest that MSH cannot directly reduce MPx:C64SC79S-SOH to reduced MPx:C64SC79S. Instead, MSH reacts with MPx:C64SC79S-SOH to form MPx:C64SC79S-SSM. These data, combining with the sulfenic acid formation of Cys36 and the disulfide linkage formation between Cys36 and Cys79 in H2O2-treated MPx:C64SC79S and MPx:C64S (Figure 2) respectively, demonstrate that Cys36 is first oxidized to its sulfenic acid derivative and then either reacts with MSH to undergo S-mycothiolation or forms a disulfide bond with Cys79.
These results were further confirmed in vivo by comparing the ability of MPx:C64SC79S to complement the oxidative-sensitive phenotype of the Δmpx mutant (contained both the Trx/TrxR and the Mrx1/Mtr/MSH regeneration pathways) and the ΔmpxΔmshCΔmrx1 mutant (contained the Trx/TrxR regeneration pathway only) (Figure 6F). The activity of MPx:C64SC79S was exclusively dependent on the Mrx1/Mtr/MSH regeneration pathway (Figure 4). Whereas the WT mpx greatly recovered the survival rates to H2O2 and Cumene-OOH stress in the Δmpx mutant, its complementation efficiency was only slightly (17.8% and 13.3%) abolished in the ΔmpxΔmshCΔmrx1 mutant that contained the Trx/TrxR regeneration pathway only. Similar complementation efficiency was obtained by expression of MPx:C64SC79S in the Δmpx mutant. However, the Mrx1-dependent MPx:C64SC79S variant completely lost the complementation ability in the ΔmpxΔmshCΔmrx1 mutant lacking the Mrx1/Mtr/MSH regeneration pathway. Consistent with the results of enzyme assay that MPx was inactive to LA-OOH decomposition by using the Mrx/MSH/Mtr regeneration system, the expression of MPx:C64SC79S in either the Δmpx or the ΔmpxΔmshCΔmrx1 mutant has no effect on recovering the oxidative-tolerant phenotype to LA-OOH treatment as the WT mpx did. These data indicate that in vivo Mrx1 and Trx employ different mechanisms to reduce MPx and the full recovery of the antioxidative activity of MPx towards H2O2 and alkyl hydroperoxides can be obtained in the presence of either the Trx/TrxR or the Mrx1/MSH/Mtr pathway.
Proposed catalytic mechanism of MPx
Based on our experimental data, a catalytic model for the alternatively reduction in MPx by Mrx1 and Trx could be hypothesized (Figure 7). The first step consists of peroxides reduction with the concomitant formation of a sulfenic acid intermediate on catalytic Cys36. The sulfenic acid intermediate was attacked by MSH, leading to the liberation of one molecule of water and the formation of a mycothione adduct. Next, the mycothione adduct was solved by Mrx1. Alternatively, a nucleophilic attack by Cys79 on the sulfenic acid intermediate leads to the release of one molecule of H2O and the formation of a transient disulfide bond between Cys36 and Cys79. This disulfide bond was directly reduced via the Trx/TrxR system to complete the regeneration cycle. MPx showed almost equal affinity for Mrx1 and Trx in vitro (Supplementary Table S4). By comparing the survival rates of the Δmpx and ΔmpxΔmshCΔmrx1 mutants that complemented with the mpx WT or variants in trans, we further confirmed that both the Trx- and the Mrx1-reducing pathways are active in the regeneration of MPx under stress conditions (Figure 6F). Moreover, almost complete recovery of the antioxidative activity of MPx towards H2O2 and alkyl hydroperoxides could be obtained in the presence of either Trx/TrxR or Mrx1/MSH/Mtr pathways, indicating that Trx and Mrx1 pathways reduced MPx in an alternative way. However, MPx was inactive to LA-OOH decomposition by using the Mrx/MSH/Mtr regeneration system, which was opposite to the Trx/TrxR system (Table 1). This property may provide it with a unique advantage to decompose a broad variety of peroxides efficiently.
Proposed reaction mechanism for MPx in C. glutamicum
We have shown that C. glutamicum MPx is more related to CysGPx than to SecGPx, the two principal groups of GPx. MPx is composed of two special regions: the conserved cysteine residue of a ‘cysteine block’ that widely exists in many Trx-dependent monomeric CysGPxs and a major deletion that was the subunit interface (designated ‘tetramer interface’ and ‘dimer interface’) of SecGPxs. Indeed, our data revealed that MPx employed the Trx/TrxR pathway for regeneration and oxidative stress resistance. In addition, we provided evidence that Mpx could also carry out the H2O2 and alkyl hydroperoxides antioxidant functions via the Mrx1/Mtr/MSH pathway but not the MSH/Mtr pathway. The catalytic mechanism of the Trx system involves the formation of an internal disulfide between Cys36 and Cys79 that is pivotal to the interaction with Trx, whereas for the Mrx1 system, the catalytic cycle involves mycothiolation/demycothiolation on the Cys36 sulfenic acid via the monothiol reaction mechanism. It is worth noting that, since almost complete recovery of the antioxidative activity of MPx towards H2O2 and alkyl hydroperoxides can be obtained in the presence of either the Trx/TrxR or the Mrx1/MSH/Mtr pathway and MPx showed almost equal affinity for Mrx1 and Trx in vitro, the Trx and Mrx1 pathways may reduce MPx in an alternative way.
Xihui Shen and Meiru Si designed the research. Meiru Si, Yixiang Xu, Tietao Wang, Mingxiu Long, Xinmeng Guan, Yingbao Liu and Can Chen performed the research and analysed the data. Yao Wang, Xihui Shen and Shuang-Jiang Liu supervised the research. Meiru Si and Xihui Shen wrote the paper. Xihui Shen, Meiru Si and Shuang-Jiang Liu revised the paper.
We thank Dr Shutao Sun in the Institute of Microbiology, Chinese Academy of Sciences for MS analysis.
This work was supported by the National High Technology Research and Development Program of China [grant number 2013AA102802]; the National Natural Science Foundation of China [grant numbers 31270078, 31170100 and 31170121]; the Key Science and Technology R&D Program of Shaanxi Province, China [grant number 2014K02-12-01]; and the Opening Project of State Key Laboratory of Microbial Resource, Institute of Microbiology, Chinese Academy of Sciences [grant number SKLMR-20120601].
5,5′-dithio bis-(2-nitrobenzoic acid)
ferrous Xylenol Orange
linoleic acid hydroperoxide
L-α-phosphatidylcholine dilinoleoyl hydroperoxide
reactive oxygen species
stress-responsive ECF-σ factor