All strains and plasmids used in this study are listed in Supplementary Table S1. The B. thailandensis E264 derivatives were mutations of the B. thailandensis E264 strain (Zhao and Shao, 2015). The B. thailandensis strains were cultured on LB plates or in LB medium in a rotary shaker (220 rpm) at 37°C.

Genes encoding B. thailandensis E264 AhpC (BTH_I2092), dihydrolipoamide dehydrogenases Lpd (BTH_I2554), dihydrolipoamide acyltransferase SucB (BTH_I1865), TrxC (BTH_I2218), and thioredoxin reductase TrxR (BTH_I1560) were amplified using PCR with B. thailandensis E264 genomic DNA as the template. These DNA fragments were digested using corresponding restriction enzymes and cloned into pET-28a vectors to construct the plasmids pET-28a-ahpC, pET-28a-lpd, pET-28a-sucB, pET-28a-trxC, and pET-28a-trxR.

To prepare the ahpC (BTH_I2092) in-frame deletion mutant of B. thailandensis, pDM4-pheS-ΔahpC was constructed. The primer pairs ahpC-M1F SpeI/ahpC-M1R and ahpC-M2F/ahpC-M2R BglII were used to amplify the ahpC upstream and downstream fragments with B. thailandensis genomic DNA as the PCR template. The primer pair ahpC-M1F SpeI/ahpC-M2R BglII was used for overlap PCR to fuse the upstream and downstream fragments together. Finally, the PCR product was digested using SpeI and BglII and inserted into pDM4-phes to create the plasmid pDM4-pheS-ΔahpC. pME6032-ahpC was constructed to complement the ahpC mutant. The primers ahpC-F BamHI and ahpC-R XhoI were used to amplify the ahpC gene from B. thailandensis genomic DNA. The ahpC gene was digested by BamHI and XhoI and ligated into pME6032 and pGEX-6P-1, obtaining the plasmids pME6032-ahpC and pGEX-6P-1-ahpC, respectively. The point mutant ahpC^C57S was created using overlap PCR (Zhang et al., 2013) to substitute a Ser residue for the Cys residue at position 57 of ahpC. The primer pairs ahpC-F BamHI/ahpC^C57S R and ahpC^C57S F/ahpC-R XhoI were used to amplify the upstream and downstream fragments of ahpC. The full-length ahpC^C57S was obtained using overlap PCR. The other site-directed mutants such as ahpC^C171S, ahpC^C173S, ahpC^C57SC171S, ahpC^C57SC173S, and ahpC^C171SC173S were created the same way. The ahpC and its mutant fragments were digested using BamHI and XhoI, and ligated into pET-28a to obtain the expression vectors like pET-28a-ahpC. The point mutants ahpD^C131S and ahpD^C134S were obtained using overlap PCR in the same way. The primer pairs are listed in Supplementary Table S2. ahpD (BTH_I2091) and its mutant fragments were then digested using BamHI and XhoI, and ligated into pET-21a to obtain the expression vectors pET-21a-ahpD, pET-21a-ahpD^C131S, and pET-21a-ahpD^C134S.

The in-frame deletion and complementation mutant construction in B. thailandensis was performed as described previously (Si et al., 2017c). The pDM4-pheS derivatives were transformed into B. thailandensis using E. coli SM10(λ pir)-mediated conjugational mating to carry out a single crossover. The transconjugants were plated on LB agar with chloramphenicol and streptomycin. The single clones were assayed using PCR. For complementation, the pME6032 derivatives were transformed into the ΔahpC mutant using electroporation. The expression was induced using 1 mM IPTG.

A stationary phase strain was diluted 30-fold into M9 medium containing 0.25 mM CHP, 1 mM H₂O₂, 1 mM t-BOOH, or 20 mM 3-Morpholinosydnonimine (SIN-1) (Aladdin). After 30 min of treatment at 37°C, the culture was diluted to 10⁻⁴ and plated on LB agar. After 16 h of cultivation at 37°C, the colonies were counted and the survival percentage was calculated as described previously (Wang et al., 2015; Si et al., 2017b).

The fluorescent reporter dye 2′,7′-dichlorodihydrofluorescein diacetate (DCFHDA) (Invitrogen) was used to detect intracellular ROS as previously described (Si et al., 2014, 2017c). 1 mL of culture was collected, washed in PBS and resuspended in 1 mL PBS with 10 μM DCFHDA. The sample was incubated in the dark at 37°C for 40 min, pelleted and washed in PBS twice. 1 mM H₂O₂, 0.25 mM CHP, or 0.3 mM t-BOOH was added to the sample resuspended in LB and incubated for 30 min at 37°C. The cells were pelleted, washed in PBS and resuspended in 400 μL PBS. 200 μL of the resuspended sample was transferred into a dark 96-well plate. Fluorescence signals were measured using a SpectraMax M2 Plate Reader (Molecular Devices) with excitation/emission wavelengths of 495/520 nm.

To obtain the His₆-tagged and GST-tagged recombinant proteins, the pET-21a, pET-28a, and pGEX-6P-1 expression vectors were transformed into the E. coli BL21(DE3) host strain as described previously (Shen et al., 2005; Si et al., 2014, 2015a). The bacteria were cultured in 500 mL LB medium at 37°C to an OD of about 0.3, cooled to 26°C, and induced to express with 0.5 mM IPTG at 26°C for 8 h. The cell pellet was harvested, resuspended in 30 mL PBS and lysed using ultrasonication. After centrifugation, the supernatant was loaded onto 1 mL pre-equilibrated Ni-NTA (Qiagen) or the GST-Bind resin (Novagen). 5 column volume washing buffer was loaded onto Ni-NTA or GST-Bind resin to wash and elution buffer with 0.5 M imidazole or 10 mM glutathione (GSH) was used to elute the recombinant protein. The His₆ tag was cleaved by adding 10 units of Enterokinase-Max (Invitrogen, Karlruhe, Germany) and incubating at 4°C overnight to conduct subsequent enzyme activity analysis. Ni-NTA agarose was used to remove the cleaved tag and uncleaved protein from the tag-free protein. All enzymes were purchased from Sigma-Aldrich (St. Louis, MO, United States). The pooled protein was determined using SDS-PAGE to be over 90% homogeneous, and its concentration was determined using Bradford assay according to the manufacturer’s instructions (Bio-Rad) with BSA as the standard.

As reported previously, Lpd, SucB, AhpD, and AhpC together sustained peroxide-dependent oxidation of NADH. The enzymatic activity of AhpC was determined by monitoring the consumption of NADH at 340 nm (Bryk et al., 2002; Si et al., 2015a,b, 2017a). The Trx reaction system was the same as described previously (Wong et al., 2017). The reaction mixture contained 100 mM KPi (pH 7.0), 1 mM EDTA, 150 μM NADH, 1 μM AhpC, 40 μM AhpD, 2 μM SucB, 2 μM Lpd and peroxides (1 mM H₂O₂ or 0.5 mM t-BOOH or CHP); or contained 50 mM Tris–HCl buffer (pH 8.0), 2 mM EDTA, 250 μM NADPH, 1 μM AhpC, 1 mM peroxides, and the reduced Trx-generating system [5 μM thioredoxin reductase (TrxR) and 40 μM TrxC]. After 5 min reaction, peroxides were added into the reaction mixture and then the NADH or NADPH consumption was monitored by the absorbance at 340 nm. The catalytic kinetic parameters for one substrate were acquired by changing its concentration in the case of saturating concentrations of the other substrate (between 0 and 500 μM for AhpD or TrxC, and between 0 and 1 mM for peroxide). The activity was determined after subtracting the spontaneous reduction rate detected without AhpC, and the micromoles amount of NADPH or NADH oxidized per second per micromolar of enzyme (i.e., turnover number, s⁻¹) was calculated using the molar absorption coefficient of NADPH or NADH at 340 nm (ε₃₄₀) of 6220 M⁻¹ cm⁻¹. Three independent experiments were performed at each substrate concentration. The k_cat and K_m values of AhpC were acquired from a non-linear fit to the Michaelis–Menten equation using the program GraphPad Prism 5.

The reaction rate of AhpC with peroxynitrite was examined using a competition approach by measuring the inhibitor effect of increasing AhpC concentrations on peroxynitrite-mediated HRP oxidation to compound I in a stopped-flow spectrophotometer (Applied Photophysics Ltd., United Kingdom) under our experimental conditions (100 mM phosphate buffer containing 0.1 mM DTPA, pH 7.4, and 25 ± 0.2°C) as previously described (Hayashi and Yamazaki, 1979; Floris et al., 1993; Ogusucu et al., 2007; Trujillo et al., 2008; Hugo et al., 2009; Manta et al., 2009; Demicheli et al., 2016; Reyes et al., 2016). In brief, 1 μM peroxynitrite-mediated HRP (5 μM) oxidation to Compound I in the absence or presence of increasing reduced AhpC concentrations (from 0 to 10 μM) was evaluated at 398 nm (Δε₃₉₈ = 4.2 × 10⁴ M⁻¹ cm⁻¹). The rate constant of peroxynitrite-mediated AhpC oxidation was calculated from HRP-Compound I yield obtained at different AhpC concentrations (0–10 μM) as previously described (Ogusucu et al., 2007; Hugo et al., 2009; Manta et al., 2009).

To analyze the reaction intermediate Cys-SOH content, NBD chloride was used to label the site-directed mutant AhpC^C57SC171S, AhpC^C57SC173S, and AhpC^C171SC173S as previously described (Baker and Poole, 2003; Si et al., 2015c). Each mutant protein (50 μM) was incubated with 5 mM DTT for 30 min. The redundant DTT was removed by ultrafiltration. The protein sample was then made anaerobic by flushing with argon gas for 30 min and separated into two parts. One was treated with 50 μM anaerobic CHP and the other kept as a control. The sample was incubated with 5 mM anaerobic NBD chloride for 30 min at 26°C in the dark. After removing excess NBD chloride by ultrafiltration, the protein sample was analyzed spectrophotometrically at 200–600 nm.

Free sulfhydryl groups in AhpC variants were determined using 5, 5′-dithio-bis (2-nitrobenzoic acid) (DTNB) as previously described (Ellman, 1959; Si et al., 2014, 2015c, 2017c). Proteins (25 μM) were treated with 5 mM CHP or 50 mM DTT at 26°C for 30 min. DTT or CHP was removed by ultrafiltration. 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 sulfhydryl groups were determined using the molar absorption coefficient of TNB at 412 nm of 13,600 M⁻¹s⁻¹ (Eyer et al., 2003).

Electrophoretic mobility shift assay was performed using the method by Si et al. (2017c). To reduce non-specific binding in the EMSA assay, the ahpC promoter region (P_ahpC, 231 bp) containing the predicted OxyR binding site was amplified from the ahpC promoter region using primers P_ahpC-F/P_ahpC-R. Increasing concentrations of purified OxyR (0–16 μg) are incubated with 20 ng ahpC DNA promoter in EMSA buffer containing 20 mM Tris–HCl, pH7.4, 4 mM MgCl₂, 100 mM NaCl, 1 mM DTT, and 10% glycerol. The binding reaction mixture is incubated at room temperature for 30 min, subjected to electrophoresis on a 6% native polyacrylamide gel with 5% glycerol in 0.5 × TBE electrophoresis buffer, and detected using SYBR Green. A 230 bp fragment from the ahpC coding region amplified with primers Control-F and Control-R instead of the 231 bp ahpC promoter in the binding assays was used as a negative control.

The GST pull-down assay was performed as previously described with minor modifications (Xu et al., 2014; Si et al., 2017b). Briefly, purified GST-AhpC was mixed with His₆-AhpD in PBS on a rotator for 2 h at 4°C, and GST was used as a negative control. After adding 30 μL of prewashed glutathione beads slurry, binding proceeded for another 2 h at 4°C. The beads were then washed five times with TEN buffer (50 mM Tris–HCl, pH 8.0, 500 mM NaCl). Retained proteins were detected using immunoblot after SDS-PAGE with the anti-His antibody (Millipore).

Student’s t-test was used to test for statistical significance. Statistical analyses were performed using GraphPad Prism Software (GraphPad Software, San Diego, CA, United States).

To examine the antioxidant functions of AhpC in B. thailandensis, the sensitivities of the wild type, the ΔahpC mutant, and the complementary strain ΔahpC(ahpC) to CHP (0.25 mM), H₂O₂ (1 mM), and t-BOOH (0.3 mM) were tested (Figure 1A). Concentrations of peroxides applied could reduce the survival rate of the wild type and increase the mortality rate from 50 to 70% (Supplementary Figure S1A). Although deletion of ahpC did not affect bacterial growth under normal conditions (Supplementary Figure S1B), the survival rates of the ΔahpC mutant decreased by about 31.4–52.6% compared to wild type. However, the oxidative stress sensitive phenotype of the ΔahpC mutant recovered to the wild type level after complementation with the ahpC gene. In addition, deletion of ahpC was shown to affect the survival rate of B. thailandensis in response to peroxynitrite (Supplementary Figure S1C).

Excess intracellular ROS is harmful in bacteria. To investigate the function of AhpC in reducing ROS under oxidative stress, the intracellular ROS levels were determined using the membrane-permeable fluorogenic probe 2′,7′-dichlorodihydrofluorescein diacetate (DCFHDA). As shown in Figure 1B, the ROS levels of all three strains under oxidative stress were higher than those of the unstressed strains, indicating ROS induction by oxidative stress in the strains. Moreover, the ΔahpC mutant had a significantly higher ROS level than the wild type after oxidative stress treatment. However, the ROS level of the ΔahpC mutant was completely restored to that of the wild type by complementation (Figure 1B). These results indicate that AhpC plays important roles in counteracting oxidative stress by reducing ROS in B. thailandensis.

It was reported that AhpC metabolize peroxides and peroxynitrite via a conserved N terminal cysteine residue that is oxidized to form a disulfide bond. To complete the catalytic cycle, the Cys residue must be reduced. AhpC can be regenerated by the reducing agents Trxc and AhpD. Thus, to examine whether the in vitro peroxide reductase activity of B. thailandensis AhpC was supported by the AhpD system (Lpd, SucB, AhpD, and NADH) and Trx system (TrxC, TrxR, and NADPH), we measured the catalytic constants of B. thailandensis AhpC with TrxC or AhpD as recycling reductants under steady-state conditions at saturating concentrations of peroxides (1 mM) and varying concentrations of reductants (0–500 μM). As shown in Table 1, the apparent affinity of B. thailandensis AhpC toward AhpD obtained under steady-state conditions was, higher than the value determined with TrxC. The k_cat values of AhpC-mediated CHP reduction with the TrxC and AhpD systems were 11.6 ± 0.2 s⁻¹ and 29.7 ± 2 s⁻¹, respectively; and the respective K_m values were 147.3 ± 10 μM and 69.2 ± 4.1 μM, respectively. The catalytic efficiencies were 0.79 ± 0.2 × 10⁵ M⁻¹ s⁻¹ and 4.29 ± 1.2 × 10⁵M⁻¹ s⁻¹, respectively. So AhpD and TrxC were active with an apparent efficiency order AhpD > TrxC. Based on our biochemical studies of AhpC, we speculated that AhpC-dependent peroxide defense was supported primarily by NADH because of the preferred selectivity of AhpD for this cofactor. Moreover, AhpC showed a small preference for the larger hydroperoxides.

Next, enzymatic activities of B. thailandensis AhpC toward various oxidizing substrates were analyzed using the TrxC and AhpD system as the reducing power (Table 2). Results showed that B. thailandensis AhpC had strong reductase activity in treating peroxide H₂O₂, t-BOOH and CHP (the catalytic efficiencies k_cat/K_m in AhpD system were 0.31 ± 0.05 × 10⁵ M⁻¹s⁻¹, 4.24 ± 0.9 × 10⁵ M⁻¹s⁻¹, and 5.19 ± 1.5 × 10⁵ M⁻¹s⁻¹, respectively; 0.1 ± 0.03 × 10⁵ M⁻¹ s⁻¹, 0.91 ± 0.1 × 10⁵ M⁻¹ s⁻¹, and 1.42 ± 0.2 × 10⁵ M⁻¹ s⁻¹in TrxC system, respectively). The relatively high reactivity of AhpC with multiple substrates, at 10⁵∼10⁶ M⁻¹ s⁻¹, supports the view that it acts as a potent antioxidant defense across a range of peroxide substrates. Apart from these peroxides, peroxynitrite was also reported as the substrate of AhpC in M. tuberculosis, S. typhimurium, and H. pylori (Bryk et al., 2000). The capacity of AhpC to reduce peroxynitrite was analyzed using competition approaches as described (Ogusucu et al., 2007). 1 μM peroxynitrite in 10 mM NaOH was mixed with 5 μM HRP in the absence of or with varying concentrations of AhpC in 100 mM sodium phosphate buffer (pH 7.4) at 25°C. As shown in Figure 2A, mixing HRP (5 μM) with peroxynitrite (1 μM) without AhpC led to the formation of HRP-Compound I. However, addition of AhpC caused a decrease in HRP-Compound I formation yield. Kinetic analysis of the data indicated that the second order rate constant of peroxynitrite reduction by AhpC was (2.7 ± 0.53) × 10⁶ M⁻¹ s⁻¹ at pH 7.4 and 25°C (Figure 2B), which was similar to those reported for other AhpCs (Bryk et al., 2000) Nature, or Prxs (Floris et al., 1993; Ogusucu et al., 2007; Manta et al., 2009; Reyes et al., 2016) in stopped-flow experiments.

Typical peroxiredoxin protein family members possess two conserved Cys residues, and mycobacterial AhpCs employ three Cys residues involved in catalysis (Guimaraes et al., 2005). The amino acid sequence alignment of AhpC homologs indicated that B. thailandensis AhpC possesses three conserved Cys residues: Cys⁵⁷, Cys¹⁷¹, and Cys¹⁷³. To investigate the function of Cys in AhpC, the site-directed mutations AhpC^C57S, AhpC^C171S, AhpC^C173S, AhpC^C171SC173S were produced, which substituted Ser residue for a Cys residue. The peroxidase activities of these proteins were measured in the presence of H₂O₂, t-BOOH and CHP. As shown in Figure 3A, the single mutant of Cys⁵⁷ or double mutant of Cys¹⁷¹ and Cys¹⁷³ resulted in a loss of AhpC catalytic activity. However, single mutant of Cys¹⁷¹ or Cys¹⁷³ showed almost no obvious effect on the activity. Similar results were also observed when the Trx system was used as a reducing power. The result indicated that AhpC could be reduced by both AhpD and TrxC, and Cys¹⁷¹ and Cys¹⁷³ could substitute each other in function. To examine the intracellular function of conserved Cys residues in AhpC, the sensitivity assay of B. thailandensis strains with ahpC in-frame deletion or serial mutant complementation were performed using H₂O₂ treatment. The result confirmed that ahpC^C57S or ahpC^C57SC171S mutant gene complementation almost could not increase the survival rate of the respective strains, but the ahpC^C57S or ahpC^C57S complementation could nearly recover the survival rate (Figure 3B).

These results indicate that the conserved Cys⁵⁷ in the N-terminus domain plays a critical role in the catalytic activity of AhpC in B. thailandensis like in other bacteria (Wong et al., 2017). The C-terminal Cys¹⁷¹ and Cys¹⁷³ residues play important roles and they could functionally compensate for each other. For simplicity, we used the AhpD system only as the reducing power in the following experiments.

The peroxidate cysteine catalyzes the peroxidase reaction by temporarily producing sulfenic acids (Ellis and Poole, 1997). To investigate the functional difference between N-terminal cysteine and C-terminal cysteines in detail, AhpC^C171SC173S, AhpC^C57SC171S, and AhpC^C57SC173S variants were generated and treated with NBD-Cl (4-chloro-7-nitrobenzo-2-oxa-1,3-diazole) with or without CHP. NBD-Cl can react with the thiol group and sulfenic acids. The reaction between NBD-Cl and the thiol group can generate absorption at 420 nm, and the reaction with sulfenic acids increases absorption at 347 nm (Baker and Poole, 2003). After the reaction with CHP, the AhpC^C171SC173S variant exhibited Soret bands at 347 and 420 nm differing from the CHP untreated group. There was no peak change with or without CHP in the reaction of AhpC^C57SC171S and AhpC^C57SC173S variants (Figure 4A). The results suggest that Cys⁵⁷ works as a peroxidatic cysteine in the active center of AhpC. The quantification of free thiol levels of AhpC variants was also performed with or without H₂O₂ treatment (Figure 4B). AhpC^C57S showed two thiol groups per monomer with or without H₂O₂ treatment. But AhpC^C171S and AhpC^C173S lost two thiol groups, and AhpC^C171SC173S lost one. These data indicate that the N-terminal Cys⁵⁷ residue acts as a peroxidatic cysteine and the C-terminal Cys¹⁷¹ and Cys¹⁷³ residues act as resolving cysteines like in other AhpC homologs (Guimaraes et al., 2005).

In the AhpC/AhpD/SucB/Lpd peroxide reducing reaction system, AhpD is the electron donor for AhpC, which is important for the reaction. To test the direct interaction between AhpD and AhpC, the GST pulldown assay was performed as shown in Figure 5A. The results showed direct binding of AhpD to AhpC. AhpD possesses a signature motif at the C-terminus as Cys-X-X-Cys which is critical for the catalytic activity in M. tuberculosis (Bryk et al., 2002). To test the function of AhpD and the Cys-X-X-Cys motif (Figure 5B), Cys to Ser mutant proteins in the motif were designed and purified. AhpD and the variants were used to carry out the reductase assay with AhpC/SucB/Lpd against CHP, t-BOOH, H₂O₂ and peroxynitrite. The activities of AhpD^CXXS decreased to about 10% of wild type, and the activities of AhpD^SXXC nearly disappeared (Figure 5C). The results indicated that Cys¹³¹ is critical for AhpD the reductase activity, and Cys¹³⁴ also had an important influence on the enzyme activity that differs from AhpD in M. tuberculosis (Koshkin et al., 2004). The Cys-X-X-Cys motif of AhpD plays a critical role in the peroxide and peroxynitrite reducing reaction with AhpC. As such, AhpD functions as a dithiol mixed disulfide reductase with the active site cysteines.

The OxyR protein is a well-known regulator for oxidative stress in bacteria. In our previous study, the negative regulation of AhpC by OxyR was discovered using RNA sequencing (RNA-seq)-based transcriptomic analysis and quantitative reverse transcriptase PCR (qRT-PCR) (Si et al., 2017c). To explore the regulation between OxyR and the expression of ahpC, promoter sequence analysis was performed and an OxyR binding site in the promoter region of ahpC was identified using the online software Virtual Footprint ¹ (Figure 6A). To examine the direct interaction between OxyR and ahpC promoter DNA, a promoter region DNA fragment containing the putative OxyR binding site was amplified using PCR with P_ahpC-F/P_ahpC-R primer pair was incubated with varying concentrations of purified OxyR protein in vitro. The reaction mixture was analyzed using an EMSA. As shown in Figure 6B, OxyR directly bound to the ahpC promoter region DNA fragment, but the negative control unrelated DNA without the OxyR binding site showed no interaction with OxyR. These data provided direct evidence that OxyR binds to the promoter region of ahpC to regulate its expression.