Six putative homologs, Lmo1233, 1609, 1903, 2152, 2424, and 2830, were annotated as thioredoxin-like proteins (Gopal et al., 2009). Among these, only TrxA contained the canonical CGPC active motif of the thioredoxins (Figure 1A). Moreover, L. monocytogenes TrxA shows 40.0–64.1% amino acid sequence homology to the Trxs from other microorganisms (Figure 1A). Phylogenetic analysis further revealed TrxA and its homolog from B. subtilis form a sister branch, while the other five thioredoxin homologures of L. monocytogenes form a separate clade (Figure 1B), indicating that these Trxs within this clade are distinct from the TrxA found in other bacteria species. Recombinant TrxA efficiently catalyzed the thiol-disulfide oxidoreduction of insulin in the presence of DTT as an electron donor (Figures 1C,D). Both C28A and C31A single mutations completely abolished the oxidoreductase ability of the protein to catalyze the reduction of insulin (Figure 1D), clearly indicating that L. monocytogenes trxA encodes a functional thioredoxin and the residues Cys28 and Cys31 are required for its catalytic activity.

Previous work suggested a connection between the L. monocytogenes redox response and virulence (Reniere et al., 2016). These observations prompted us to investigate the role of the thioredoxin in L. monocytogenes. The trxA mutant strain exhibited similar growth to the parent strain in normal BHI medium (Figure 2A). To further determine whether the Trx system could contribute to oxidative tolerance, the bacteria were exposed to different oxidative agents. For oxidative stresses, paraquat was used as a superoxide generating compound, and H₂O₂ as a direct oxidant while diamide was used as a thiol-specific oxidizing agent. Notably, deletion of trxA led to a significantly increased lag phase in growth, starting from 4 h (P < 0.01), compared to its parental strain and two complement strains (CΔtrxA_P_trxA, expressing TrxA under its native promoter, and CΔtrxA_P_help, carrying the constitutive Listeria promoter, P_help) in BHI containing diamide, while H₂O₂ and paraquat did not exert marked effects on growth (Figures 2B–D). Immuno-blotting was used to determine expression of the Trx system under different oxidative stress conditions. As expected, diamide at a concentration of 4 mM induced expression of TrxA and TrxB (annotated as a thioredoxin reductase) to a significant extent, especially at the 30 and 60 min time-points (Figure 2E). However, no obvious induction was found in the presence of paraquat or H₂O₂ stress on TrxA expression, while TrxB was slightly induced by these two oxidants (Figures 2F,G). The results collectively suggest that diamide is a stronger inducer of TrxA expression than other oxidants, which enables L. monocytogenes to better survive in the thiol-specific oxidative conditions.

DsbA, a disulfide bond forming protein A comprising a Trx domain, catalyses the formation of disulfide bond that is involved in bacterial flagella synthesis (Heras et al., 2009). Thus, we examined whether TrxA is linked to motility in Listeria. Interestingly, motility of the ΔtrxA strain at 30°C was almost completely compromised, compared to its wild-type counterpart (Figure 3A). Impaired motility of the mutant was fully rescued in the complement strain, CΔtrxA_P_trxA, expressing TrxA under its native promoter, and partly restored in the other complement strain, CΔtrxA_P_help, carrying the constitutive Listeria promoter, P_help (Figure 3A), suggesting that L. monocytogenes TrxA contributes to bacterial motility. In contrast, no motility was observed at 37°C (Figure 3A), which inhibited flagellar formation in many Listeria strains (Grundling et al., 2004). Transmission electron microscopy further revealed that the trxA mutant is totally unable to produce visible flagella whereas morphologically normal-looking flagella were expressed by the wild-type and two trxA complement strains (Figure 3B). Our results demonstrate for the first time that L. monocytogenes TrxA participates in bacterial motility and flagellar production.

Temperature-dependent expression of Listeria flagellar motility genes is mediated by the opposing activities of MogR, a DNA-binding transcriptional repressor, and GmaR, an anti-repressor that functions as a direct antagonist of MogR (Shen and Higgins, 2006; Shen et al., 2006; Kamp and Higgins, 2011). Given that MogR and GmaR contain 2 and 4 cysteines, respectively, the CGPC domain of TrxA is predicted to play a critical role in the interactions between MogR and GmaR, and therefore contribute to Listeria flagella formation. Calorimetric response data on MogR during reaction with GmaR showed significant scatter with no discernible trends, which was more evident in the corresponding binding isotherm plot (Figure 3C), indicating that in vitro interactions between MogR and GmaR are not strong. However, upon titration of purified TrxA into MogR and GmaR, the binding event exhibited endothermic heat of reaction. A good fit to the data for calorimetric titration was achieved using a single-site binding model (Figure 3D), suggesting that TrxA is required to increase the binding affinity of MogR to GmaR. TrxA displayed endothermic reactions for strong binding to MogR in a manner similar to calorimetric data obtained in Figure 3D, while showing weak binding affinity to GmaR (Figures 3E,F). Notably, the endothermic response clearly observed during the whole titration process between TrxA and MogR is believed to arise due to breakage of intermolecular bonds among MogR molecules as a consequence of titration of TrxA (Figure 3F), suggesting that this protein contributes to breaking the intermolecular bonds of MogR that bind DNA of the flagella-related gene promoter region as a dimer (Shen et al., 2009). More importantly, mutation of either C28A or C31A located within the critical motif of thioredoxin significantly weakened the binding affinity of TrxA to MogR and GmaR, as determined from calorimetric data (Figures 3G–J).

As indicated from isothermal titration calorimetry (ITC) data (Figure 3), TrxA could contribute to catalyzing the reduction of misfolded MogR disulfides and maintaining its intracellular monomer status. To further characterize the reduction properties of recombinant TrxA, we assayed the ability of the thioredoxin to reduce MogR in vitro. The thioredoxin assay directly measures the redox state of the substrate by using a gel-based assay. In the presence of SDS, a protein with free cysteine residues is less compact than the same protein in which these residues form disulfide bonds. The reduced form of a protein therefore migrates more slowly upon polyacrylamide gel electrophoresis (PAGE) in the presence of SDS than its oxidized counterpart. As indicated in the Figure 4A, the recombinant MogR monomer exists as a mixture of the reduced and oxidized forms. However, the reduced form amount of MogR monomers gradually increased when an increasing amount of the recombinant TrxA added in the presence of DTT as an electron donor, while the amount of MogR dimers decreased (Figure 4A), strongly confirming that TrxA could act to reduce disulfides in MogR and maintain the monomer status.

In addition, the thioredoxin system is the major cellular disulfide reductase in cells, which can facilitate correct oxidative protein folding together with protein disulfide isomerases (PDI) and DSB proteins (Lu and Holmgren, 2014b). Searches for orthologs of B. subtilis DsbA and DsbB in L. monocytogenes revealed that this pathogen only encodes two putative DSB proteins, DsbA (Lmo0964), and DsbG (Lmo1059), but does not encode DsbB or any other DSB proteins. Therefore, the roles of Lmo0964 and Lmo1059 in correct dimerization of MogR were further investigated. To investigate mutant sensitivity to oxidative stress, Listeria were grown in the presence of the redox-active metal, copper, which rapidly and randomly oxidizes unpaired cysteines through a superoxide mechanism (Hiniker et al., 2005). WT and Δlmo1059 had a similar resistance to copper while Δlmo0964 was extremely copper-sensitive (Figure 4B). Cadmium sensitivity was further tested as an indicator of oxidase capacity due to the high affinity of Cd²⁺ for protein thiol groups (Vallee and Ulmer, 1972). The lmo0964-null strain was especially Cd²⁺ sensitive, while wild-type and Δlmo1059 strains were resistant to a concentration of 1 mM Cd²⁺ (Figure 4C). Our results suggest that Lmo0964 acts not only as an oxidase but also an isomerase, whereas bacteria in the absence of lmo1059 remain cadmium- and copper-resistant. Interestingly, in vitro enzymatic reactions indicate that Lmo1059 exhibits high disulfide interchange activity (Figure 4D) but fails to catalyze reduction of insulin in the presence of DTT (Figure 4E), suggesting that this protein mainly functions as an isomerase but not an oxidoreductase in vitro. Intriguingly, distinct from E. coli DsbA or BdbD in B. subtilis, Lmo0964 is unable to catalyze reduction of insulin, at least under standard assay conditions, suggesting that this protein is more stable in its oxidized form and mainly functions as an oxidase in vitro.

Whole-genome transcriptomic sequencing revealed that 160 genes show = 2.0-fold-higher transcript levels in parent EGD-e than the ΔtrxA mutant strain under oxidative conditions (expo0sure of stationary cells to 2 mM diamide for 1 h; Table S1, Figures 5A,B). Transcript levels of a total of 114 genes were =2.0-fold higher in the ΔtrxA mutant than the parent strain (Table S2, Figures 5A,B). Notably, the virulence genes, plcA, mpl, actA, and plcB, showed significantly higher expression in wild-type than the ΔtrxA mutant, with transcriptional ratios consistently >3.8 (Table S1). The internalin gene, inlC, also showed significant expression ratios of >2.0 (Table S1). In addition, three other virulence-associated genes, hly, inlA, and inlB, were slightly activated by TrxA (Table S1). Transcriptomic data were validated using quantitative RT-PCR, showing that the results of transcriptome were firmly reliable (Figures 5C,D). In addition, deletion of TrxA almost completely reduced cytoplasm LLO expression and significantly inhibited its secretion into culture supernatant, compared to the wild-type and complement strains (Figure 5E), which was further verified by the finding that loss of TrxA leads to thoroughly reduced LLO activity in lysis of sheep red blood cells (Figure 5F). These results collectively show for the first time that L. monocytogenes TrxA alters global expression profiles, especially those of genes encoding virulence-associated factors, and therefore most probably play a critical role in bacterial pathogenicity.

To investigate the involvement of TrxA in L. monocytogenes infection inside host cells, the epithelial cells Caco-2 was used to compare the wild-type and trxA mutant strains in terms of adhesion, invasion, and intracellular survival. The trxA mutant was significantly less adhesive and invasive, therefore resulting in the less efficient intracellular survival, compared to that of the wild-type strain (Figures 6A,B). Notably, EGD-e and the complement strains were able to associate with F-actin and formed long actin tails (Figure 6B), while a large proportion of the trxA mutant bacteria were not associated with polymerized actin, and the rare bacteria that were associated with polymerized actin displayed very short tails or clouds and were often located next to the plasma membrane (Figure 6B). However, unlike the circumstance in epithelial cells, the ΔtrxA mutant did not show a significantly decreased ability to grow intracellularly in RAW264.7 macrophages relative to the wild-type strain (data not shown), suggesting that TrxA mainly contributes to the early stages of L. monocytogenes infection within host. Moreover, the ΔtrxA mutant exhibited severely attenuated virulence, as infected mice exhibited bacterial burdens that were significantly lower relative to those infected with the wild-type strain at 48 h post infection (Figure 6C). Most importantly, the mutant strain containing the entire trxA CDS with its native promoter fully complemented virulence defects associated with loss of trxA, but not the P_help promoter (Figure 6C), suggesting no expression or overexpression of TrxA in vivo would disturb the equilibrium of the intracellular redox potential and thus bring harm to this pathogen for its ability to establish infections. Previous work suggested that the sigma factor H (SigH) is important for the regulation of mycobacterial response to oxidative stress by regulating the expression of thioredoxins (trxB1 and trxC) and thioredoxin reductase (Bhat et al., 2012), which therefore prompted us to investigate the role of SigH in regulating on the Trx expression in L. monocytogenes. As determined by EMSA assay, the purified recombinant SigH bound to the fragment containing the promoter region of trxA or trxB. Besides, upon incubation of the promoter DNA fragment with increasing amounts of SigH, shifted, supershifted, and super-supershifted DNA complexes were observed (Figure S1), supporting the theory that the thioredoxin system of L. monocytogenes is strictly regulated by an alternative sigma factor. Taken together, these data demonstrate that TrxA-mediated regulation of virulence-associated factors constitutes an important aspect of L. monocytogenes infection within host. Importantly, the ITC experiment indicates that TrxA exhibits exothermic reactions for strong binding to the virulence regulator, PrfA, a member of the Crp/Fnr family of bacterial transcription factors (Figure 6D), while the oxidized forms of TrxA (TrxA_ox) almost showed no binding affinity to PrfA (Figure 6E). Given the fact that PrfA activation is highly dependent on the reducing environment and only the reduced PrfA dimers can bind to DNA and then activate transcription of virulence genes (Reniere et al., 2015), we therefore suggest that TrxA plays a critical role in reducing the intermolecular disulfide bonds of PrfA and subsequently maintaining the intracellular monomer status for further dimerization.

L. monocytogenes EGD-e was used as the wild-type strain. E. coli DH5α was employed for cloning experiments and as the host strain for plasmids pET30a(+) (Merck, Darmstadt, Germany), pIMK2 and pKSV7. E. coli Rosetta (DE3) was used for prokaryotic protein expression. Listeria strains were cultured in brain heart infusion (BHI) medium (Oxoid, Hampshire, England). E. coli strains were grown at 37°C in Luria-Bertani broth (LB) (Oxoid). Stock solutions of ampicillin (50 mg/ml), erythromycin (50 mg/ml), kanamycin (50 mg/ml), or chloramphenicol (50 mg/ml) were added to media when necessary. All chemicals were obtained from Sangon Biotech (Shanghai, China), Merck or Sigma-Aldrich (St. Louis, USA) and were of the highest available purity. All primers used in this study are listed in Table S3.

The amino acid sequences of putative thioredoxin proteins from L. monocytogenes EGD-e and its homologs in other microbial species were obtained from the National Centre for Biotechnology Information database (NCBI). Trx amino acid sequences were aligned with MUSCLE using Geneious software (Edgar, 2004). The phylogenetic tree was constructed with the Neighbor-joining (NJ) method using 100 bootstrap replicates.

The recombinant proteins used in this study were expressed as an N-terminal His tag fusion using pET30a(+) as the expression vector and Rosetta (DE3) as the expression host. The full-length ORF of the gene of interest from the EGD-e genome was amplified with the primer pair, inserted into the pET30a(+) vector, and finally transformed into Rosetta competent cells. E. coli cells harboring recombinant plasmids were grown in 500 mL LB supplemented with 50 μg/mL kanamycin at 37°C until cultures reached 0.8–1.0 at OD_600nm. Isopropyl β-D-1-thiogalactopyranoside (IPTG) was added to a final concentration of 0.2 mM to induce expression of recombinant proteins for an additional 3 h under optimized conditions. His-tagged fusion proteins were purified using nickel-chelated affinity column chromatography.

Rabbits were initially immunized via subcutaneous injection of 500 μg protein with an equal volume of Freund's complete adjuvant (Sigma). After 2 weeks, rabbits were boosted by subcutaneous injection of 250 μg protein each in incomplete Freund's adjuvant (Sigma) three times at 10 day intervals. Rabbits were bled ~10 days after the last injection.

To identify the predicted active sites of TrxA, single (C28A and C31A) and a double mutant (C28AC31A) were generated using the original vector template, pET30a-TrxA, and the QuikChange Site-Directed Mutagenesis kit (Agilent Technologies, Palo Alto, USA) with the primer pairs described in Table S3. Template DNA was removed via digestion with DpnI (Toyobo Co., Osaka, Japan) for 2 h at 37°C. All mutant constructs were sequenced to ensure that only the desired single mutations had been incorporated correctly. The corresponding mutant proteins were designated TrxA_C28A, TrxA_C31A, and TrxA_C28AC31A, and expressed and purified as described above.

The ability of TrxA, DsbA, and DsbC/G to catalyze reduction of human insulin (Sigma) in the presence of DTT was measured as described previously by Holmgren (1979). Briefly, reaction mixtures were prepared using 0.1 M potassium phosphate buffer (pH 7.0), 150 μM insulin, 2 mM EDTA, and 1.0 μM purified proteins in a final volume of 200 μL. Reactions were initiated by adding DTT to a final concentration of 1 mM, and monitored as the increase in absorbance at 650 nm every 3 min (100 min in total) at 25°C using the Micro-plate reader Synergy H1 (BioTek Solutions, Inc., Santa Barbara, USA). The calibration curve for quantitative analysis of reduced insulin was determined by incubating 0.5 μM purified TrxA in PBS containing 1 mM DTT with various concentrations of insulin (0, 40, 80, 120, 160, and 200 μM).

Construction of L. monocytogenes gene deletion mutants were performed as described previously (Camilli et al., 1993). The temperature-sensitive pKSV7 shuttle vector was used for generating mutations in the L. monocytogenes strain EGD-e. A homologous recombination strategy with the splicing by overlap extension (SOE) PCR procedure was used for in-frame deletion to construct gene deletion mutants (Camilli et al., 1993; Monk et al., 2008). Specifically, DNA fragments containing homologous arms upstream and downstream of the gene of interest were obtained via amplification of EGD-e genomic DNA using SOE primer pairs. The obtained fragment was then cloned into the temperature-sensitive shuttle vector pKSV7 and transformed into E. coli DH5α. After confirmation by sequencing, the recombinant vector containing the target gene deletion cassette was then electroporated into the competent L. monocytogenes EGD-e cells. Transformants were selected on BHI agar plates containing chloramphenicol (10 μg/ml). A single transformant was serially passaged for ~30 generations at a non-permissive temperature (40°C) in BHI medium containing 10 μg/ml chloramphenicol to promote chromosomal integration. A single colony with chromosomal integration was successively passaged in BHI medium without antibiotic at a permissive temperature (30°C) for ~50 generations to enable plasmid excision and curing. The recombinants were identified as chloramphenicol-sensitive colonies, and the mutagenesis was further confirmed by PCR and DNA sequencing.

To complement the L. monocytogenes ΔtrxA strain, we constructed two complement strains using the integrative plasmid, pIMK2, which harbors a constitutive Listeria promoter P_help (Monk et al., 2008). For the first complement strain, we amplified the complete open reading frame (ORF) of trxA from EGD-e genomic DNA using the primer pair CΔtrxA-a/CΔtrxA-b and inserted the fragment downstream of P_help after digestion with the appropriate restriction enzymes. For the other complement strain, the complete ORF of trxA along with its promoter region was amplified using the primer pair CΔtrxA-c/CΔtrxA-d and cloned into pIMK2 following digestion with the appropriate restriction enzymes to remove the P_help region. The resulting plasmids were electroporated into the L. monocytogenes ΔtrxA strain. Regenerated cells were plated on BHI agar containing kanamycin (50 μg/ml). The two complement strains were designated CΔtrxA_P_help and CΔtrxA_P_trxA, respectively.

L. monocytogenes wild-type strain EGD-e, mutant ΔtrxA and two complement strains, CΔtrxA_P_help and CΔtrxA_P_trxA, were grown overnight at 37°C in BHI broth with shaking. Cultures were collected by centrifugation at 5,000 × g at 4°C, washed once in PBS (10 mM, pH 7.4) and initial OD_{600 nm} adjusted to 1.0. Bacteria were diluted (1:100) in fresh BHI broth or BHI containing 5 mM paraquat, 4 mM diamide, or 10 mM H₂O₂, and incubated at 37°C for 12 h. Kinetic growth at OD_{600 nm} was measured at 1 h intervals.

Bacteria were grown in BHI broth to the stationary growth phase and lysates prepared as described earlier (Ryan et al., 2009). Specifically, stationary bacteria were collected at the indicated times after treatment with 5 mM paraquat, 4 mM diamide, or 10 mM H₂O₂. Bacterial pellets were re-suspended in 1 mL of extraction solution (2% Triton X-100, 1% SDS, 100 mM NaCl, 10 mM Tris-HCl, 1 mM EDTA, pH 8.0). One gram of glass beads (G8772, Sigma) was added and samples lysed using the homogenizer Precellys 24 (Bertin, Provence, France) at 6,000 rpm for 30 s with intermittent cooling for 30 s (2 cycles in total), followed by centrifugation at 12,000 g for 15 min. Supernatants were retained as cell-free extracts. Proteins were separated via 12% SDS-PAGE and blotted onto 0.22 μm PVDF membrane (Merck). Membranes were blocked for 1 h with 5% skimmed milk, and incubated for 1 h with polyclonal antisera against recombinant TrxA or TrxB (prepared in this study) in 0.5% skimmed milk. Next, membranes probed with anti-TrxA or anti-TrxB were developed using HRP-conjugated goat anti-rabbit IgG (Santa Cruz, California, USA) as the secondary antibody. Chemiluminescence was detected via an UVP EC3 imaging system (UVP Inc., Upland, USA).

Motility assay was performed essentially on soft LB agar (0.25%) as described (Paxman et al., 2009). L. monocytogenes were grown overnight in BHI and the cultures adjusted at OD_{600 nm} to 0.40 (about 2 × 10⁸ CFU/ml). Bacterial samples (5 μL) were dropped onto soft TSA agar and incubated at 30° or 37°C for 48 h to allow growth. Motility was assessed by examining migration of bacteria through agar from the center toward the periphery of the colony. TEM experiments were performed as described previously (Dingle et al., 2011) in the Institute of Agrobiology and Environmental Sciences of Zhejiang University. Briefly, L. monocytogenes colonies grown overnight at 30°C from BHI agar plates were suspended in 50 μL monoethanolamine buffer (pH 10.0), and 10 μL of the suspension applied to carbon-coated copper grids and allowed to stand for 2 min at room temperature. Excess liquid was subsequently removed using filter paper, and bacteria stained with 10 μL of 0.5% phosphotungstic acid (pH 7.0) placed on the grids for 10 s at room temperature. Excess stain was gently wicked away using filter paper, and the dried grids examined under a Hitachi H-7650 transmission electron microscope (Hitachi High-Technologies Corporation, Tokyo, Japan).

The protein-protein interactions between TrxA (or its active site mutants), MogR, and GmaR were measured using isothermal titration calorimetry (ITC) with a VP-isothermal titration calorimeter from Microcal, Inc. (Northampton, USA). Calorimetric titrations of TrxA (80–200 μM in the syringe) and PrfA, MogR, or GmaR (8–20 μM in the cuvette) were carried out at 25°C in 25 mM Tris-HCl, pH 7.5, 100 mM NaCl. TrxA or its active site mutants (C28A, C31A, and C28AC31A) was titrated into PrfA, MogR, or GmaR in 15 μL injections with a spacing of 300 s between injections. Calorimetric data were analyzed by integrating heat effects normalized to the amount of injected protein and curve-fitting based on a 1:1 binding model using the Origin software package (Microcal). The dissociation constant was derived from data using standard procedures. ITC was carried out in the Life Sciences Institute of Zhejiang University.

Reactions with recombinant TrxA were carried out in 50 mM Tris-HCl buffer, pH 7.5 with the increasing amount (0–2 μg) of reduced TrxA and 2 μg MogR with or without 0.1 mM DTT as an electron donor. After 1 h incubation, an equal amount of SDS–PAGE sample buffer was added to quench the reactions. Reactions were terminated with 10% (v/v) trichloroacetic acid (TCA) and then separated on 12% polyacrylamide gels, and stained with Coomassie blue to visualize the redox state of MogR.

Cadmium and copper sensitivity assays are usually performed to test oxidase and isomerase activities, respectively (Ren et al., 2014). L. monocytogenes wild-type strain EGD-e, isogenic lmo1059 and lmo0964 deletion mutants and the corresponding complementation strains were grown overnight in BHI broth. Bacteria were diluted to OD_600nm of 1.0 (~10⁹ cfu ml⁻¹) in PBS. Bacteria were serially diluted 10-fold, and 10 μl of each dilution spotted onto BHI agar plates containing 1 mM cadmium chloride (Sangon) or 1 mM copper(II) chloride (Sangon). Following incubation at 37°C for 48–72 h, colony growth on each plate was assessed and imaged. All spot titer assays were performed in duplicate to verify the results.

Disulfide interchange reactions were used to investigate the ability of TrxA, DsbA, and DsbC/G to catalyze the reduction of oxidized glutathione (GSSG) by DTT. The DTT/glutathione equilibrium lies far on the side of oxidized DTT, and DTT_ox formation can be followed spectrophotometrically. GSSG (5 mM) was reduced by DTT (5 mM) at 25°C in 100 mM formic acid/NaOH (pH 4.0), 1 mM EDTA. Reactions were monitored based on the increase in absorbance at 287 nm.

To validate of the reliability of transcriptome sequencing data, real-time quantification of random mRNAs was performed using qRT-PCR with the same total cellular RNAs used for transcriptome experiments. We randomly selected 28 genes for qRT-PCR validation, i.e., 16 representative downregulated genes and 12 representative upregulated genes in the ΔtrxA mutant. Specific primers for these genes were designed and purchased from Sangon Biotech. cDNA was synthesized with reverse transcriptase (Toyobo). Quantitative PCR was performed in 20 μL reaction mixtures containing SYBR quantitative PCR mix (Toyobo) to measure the transcriptional levels of genes of interest using the Mx3000P PCR detection system (Agilent) with specific primer pairs. Real-time quantitative PCR was performed in a 20 μL reaction volume containing 200 ng cDNA, 10 μL SYBR quantitative PCR mix and 1 μL gene-specific primers (200 nM). The housekeeping gene, rpoB, was used as an internal control for normalization in each sample. Triplicate assays were performed for each gene.

To further establish, whether TrxA is involved in bacterial virulence, western blot was employed to analyze the changes in expression of the major virulence-associated factor, LLO. Bacterial overnight cultures of L. monocytogenes wild-type EGD-e, mutant ΔtrxA and two complement strains, CΔtrxA_P_help and CΔtrxA_P_trxA, were diluted into 100 mL fresh BHI broth and grown to the stationary phase. For isolation of secreted proteins, the fractionation procedure was performed as described before (Lenz and Portnoy, 2002), with minor modifications. Briefly, bacterial cells were pelleted via centrifugation at 13,000 g for 20 min at 4°C, and the resulting culture supernatant collected and filtered through a 0.22 μm polyethersulfone membrane filter (Merck). Trichloroacetic acid (TCA) was added to the supernatant to obtain a final concentration of 10%. Proteins were TCA-precipitated on ice overnight and washed with ice-cold acetone. Washed precipitates of supernatant proteins were re-suspended in SDS-PAGE sample buffer (5% SDS, 10% glycerol, and 50 mM Tris-HCl, pH 6.8). Samples were boiled for 5 min and stored at -20°C before electrophoresis. Total cell protein isolation was performed as described above. Protein samples were separated using 12% SDS-PAGE and immunoblotted with α-LLO or α-GAPDH antisera (prepared in this study). GAPDH was employed as an internal control.

Measurement of LLO-associated hemolytic activity was done as described previously (Xayarath et al., 2015). L. monocytogenes wild-type and mutant strains were grown for 16 h with shaking in BHI broth at 37°C. All cultures were adjusted to OD_600nm of 1.0 before supernatants were collected. Hemolytic activity was measured based on lysis of sheep red blood cells (SRBCs) using supernatants from cultures. The culture supernatant (50 μL) was diluted with an equal volume of PBS containing 6 mM cysteine (pH 5.6) and equilibrated to 37°C for 10 min. Next, 50 μL PBS-washed intact SRBCs were added to each sample and incubated at 37°C for 60 min. Samples were centrifuged and supernatants analyzed for hemoglobin absorption at 550 nm. PBS was employed as a negative control. The values corresponding to the reciprocal of the dilution of culture supernatant required to lyse 50% of HRBCs were used to compare the hemolytic activities in the different supernatants.

Bacterial survival in human intestinal epithelial Caco-2 cells was assessed as described previously (Reddy et al., 2016). Stationary L. monocytogenes wild-type strain EGD-e, mutant ΔtrxA and two complement strains, CΔtrxA_P_help and CΔtrxA_P_trxA, at 37°C in BHI were washed and re-suspended in PBS (10 mM, pH 7.4). Caco-2 cells cultured to 80% confluence in Dulbecco's modified eagle medium (DMEM) containing 10% fetal calf serum were infected with the above strains for 60 min at a multiplicity of infection (MOI) of 10:1. For adhesion, cells were lysed after being washed three times with PBS. For estimation of invasion, cells were washed with PBS after 1 h infection and incubated for an additional hour in DMEM containing gentamicin at 200 μg/mL. Caco-2 cells were lysed by adding 1 mL of ice-cold sterile distilled water. Lysates were 10-fold diluted for enumeration of viable bacteria on BHI agar plates that were considered the 0 h numbers invading cells. For intracellular proliferation, cells were further incubated for 6 or 12 h in 5% CO₂ at 37°C. Viable bacteria were enumerated as described above. Adhesion was expressed as the ratio of recovered colonies to colonies inoculated, while invasion was calculated as the ratio of colonies recovered after gentamicin treatment to colonies inoculated.

Cells were infected at MOI of 10:1 at 37°C with 5% CO₂ for 1 h. Extracellular bacteria were then killed with 200 μg/mL gentamicin for 1 h and incubated for an additional 6 h. Cells were washed gently with PBS (10 mM, pH 7.4), fixed with 4% paraformaldehyde and then permeabilized with 0.5% Triton X-100. The bacterial cells were stained with polyclonal antibodies to L. monocytogenes for 1 h at 37°C, washed twice with PBS and probed with Alexa Fluor 488-conjugated donkey anti-rabbit antibody (Santa Cruz) for 1 h at 37°C. F-actin was then stained with phalloidin-Alexa Fluor 568 (Thermo Fisher Scientific). DAPI (4′,6-diamidino-2-phenylindole; Thermo Fisher Scientific) was used to stain the nuclei. Actin tails recruited by the bacteria were visualized under a ZEISS LSM510 confocal microscope (Zeiss Germany, Oberkochen, Germany).

The L. monocytogenes wild-type strain EGD-e, mutant ΔtrxA and two complement strains, CΔtrxA_P_help and CΔtrxA_P_trxA, were tested for recovery in liver and spleen sections of ICR mice (female, 18–22 g, purchased from Zhejiang Academy of Medical Sciences, Hangzhou, China). ICR mice (eight per group) were injected intraperitoneally with ~10⁶ CFU of each strain. At 24 and 48 h post-infection, mice were sacrificed, and livers and spleens removed and individually homogenized in PBS (10 mM, pH 7.4). Surviving Listeria cells were enumerated by plating serial dilutions of homogenates on BHI agar plates. Results were expressed as means ± SD of the recovery rate per organ for each group.

To further demonstrate, whether the L. monocytogenes Trx system is regulated by SigH, we determined the binding abilities between SigH and promoter regions of trxA and trxB in vitro using the electrophoretic mobility shift assay (EMSA). DNA fragments of the promoter regions of trxA and trxB were generated by PCR with the specific primer pairs, and purified with the PCR Purification Kit (Sangon) according to the manufacturer's instructions. Next, 200 ng DNA was incubated with varying concentrations of purified recombinant SigH and incubated in binding buffer (50 mM Tris-HCl, pH 8.0, 250 mM NaCl, 5.0 mM MgCl₂, 2.5 mM DTT, 2.5 mM EDTA, and 20% glycerol) for 30 min at room temperature. Protein-DNA complexes were separated electrophoretically on a native 5% polyacrylamide gel at 80 V with 0.5 × Tris-acetate-EDTA (TAE) buffer and visualized using ethidium bromide staining.

All experiments were repeated at least three times. Data were analyzed using the two-tailed homoscedastic Student's t-test. Differences with P < 0.05 were considered statistically significant.

All animal care and use protocols were performed in accordance with the Regulations for the Administration of Affairs Concerning Experimental Animals approved by the State Council of People's Republic of China. The protocol was approved by the Institutional Animal Care and Use Committee of Zhejiang A&F University (Permit Number: ZJAFU/IACUC_2011-10-25-02).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.