Two Spx Regulators Modulate Stress Tolerance and Virulence in Streptococcus suis Serotype 2

Streptococcus suis serotype 2 is an important zoonotic pathogen causing severe infections in pigs and humans. The pathogenesis of S. suis 2 infections, however, is still poorly understood. Spx proteins are a group of global regulators involved in stress tolerance and virulence. In this study, we characterized two orthologs of the Spx regulator, SpxA1 and SpxA2 in S. suis 2. Two mutant strains (ΔspxA1 and ΔspxA2) lacking the spx genes were constructed. The ΔspxA1 and ΔspxA2 mutants displayed different phenotypes. ΔspxA1 exhibited impaired growth in the presence of hydrogen peroxide, while ΔspxA2 exhibited impaired growth in the presence of SDS and NaCl. Both mutants were defective in medium lacking newborn bovine serum. Using a murine infection model, we demonstrated that the abilities of the mutant strains to colonize the tissues were significantly reduced compared to that of the wild-type strain. The mutant strains also showed a decreased level of survival in pig blood. Microarray analysis revealed a global regulatory role for SpxA1 and SpxA2. Furthermore, we demonstrated for the first time that Spx is involved in triggering the host inflammatory response. Collectively, our data suggest that SpxA1 and SpxA2 are global regulators that are implicated in stress tolerance and virulence in S. suis 2.


Introduction
Streptococcus suis (S. suis) is an important zoonotic pathogen causing significant economic losses to the swine industry worldwide each year and is responsible for a variety of diseases in pigs, including meningitis, septicemia, arthritis, pneumonia, and even acute death [1]. Among the 33 serotypes (types 1 to 31, 33, and 1/ 2) identified on the basis of capsular polysaccharides, S. suis serotype 2 (S. suis 2) is considered to be the most virulent and the most frequently isolated serotype in association with diseases in most countries [2,3]. S. suis, especially serotype 2, is also the causative agent of serious infections in humans in contact with infected pigs or pork-derived products and causes meningitis, septic shock, and permanent hearing loss [4][5][6]. Since the first reported case of S. suis infection in human in Denmark in 1968, more than 700 human cases worldwide have been recorded [7]. In 1998 and 2005, two large outbreaks of human S. suis 2 infection in China raised enormous public concerns because of the high pathogenicity of this microbe [4]. Recently, cases of human S. suis infection have been reported in Canada, the United States, Australia, New Zealand and Korea [8][9][10]. In addition, S. suis was identified as the leading cause of adult meningitis in Vietnam, the second most common cause in Thailand and the third most common cause of community-acquired bacterial meningitis in Hong Kong [9,10].
Spx proteins are a group of global transcriptional regulators that are highly conserved among low-GC-content Gram-positive bacteria [28]. Unlike most regulators that interact with promoter regions, Spx binds to the C-terminal domain (CTD) of the asubunit of the RNA polymerase (RNAP) and alters the holoenzyme architecture [29]. By modifying the interactions between the RNA polymerase and specific promoter regions, Spx can activate or repress transcription [30]. To date, Spx regulators have been described in many bacterial species, including Lactococcus lactis [31], Bacillus subtilis [32][33][34], Staphylococcus aureus [35,36], Streptococcus pneumoniae [30], Streptococcus mutans [37], Streptococcus sanguinis [38], and Enterococcus faecalis [39]. Regulation by Spx has been characterized in extensive detail in B. subtilis. Under conditions of disulfide stress, Spx regulates the expression of different subsets of genes, negatively or positively, by interacting with the CTD of the RNAP a subunit [32]. A crystal structure of the B. subtilis Spx protein, in complex with the a-CTD of the RNAP reveals interactions between Spx and the a-CTD, and suggests that subtle conformational changes may be important for the role of Spx in regulating organosulfur metabolism [33,40]. A recent study identified 144 transcription units comprising 275 genes that were potentially under regulation by Spx in B. subtilis [34]. Using B. subtilis as a model organism, it was demonstrated that Spx is critical for the prevention of protein aggregate formation during thermotolerance [41]. Due to its important role in regulation, much attention has been paid to the role of Spx in organism pathogenesis. In S. aureus, Spx was shown to be a global effector impacting stress tolerance, biofilm formation and cell wall antibiotic resistance [35,36]. In S. mutans, two Spx proteins were shown to modulate stress tolerance, survival and virulence [37]. SpxA1 was shown to be involved in X-state (competence) development in S. pneumonia [30], and involved in hydrogen peroxide production, stress tolerance and endocarditis virulence in S. sanguinis [38]. In addition, the Spx regulator was also shown to modulate stress responses and virulence in E. faecalis [39].
In this study, the roles of Spx regulators in S. suis 2 were explored. Two Spx proteins were identified: SpxA1 and SpxA2. Functional studies revealed that SpxA1 and SpxA2 play important roles in stress tolerance and virulence in S. suis 2. Global gene transcription profiles indicated that the two Spx proteins are global transcriptional regulators in S. suis 2.

Ethics statement
This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of Hubei Province, China. The protocol was approved by the Laboratory Animal Monitoring Committee of Huazhong Agricultural University. All efforts were made to minimize suffering.

Bacterial strains, plasmids, and growth conditions
The bacterial strains and plasmids used in this study are listed in Table 1. S. suis 2 strains were grown in Tryptic Soy Broth (TSB) or plated on Tryptic Soy Agar (TSA; Difco Laboratories, Detroit, MI, USA) with 10% (vol/vol) newborn bovine serum at 37uC. Escherichia coli strain DH5a was grown in Luria broth (LB) liquid medium or on LB agar and used for plasmid construction and propagation. If required, spectinomycin (Sigma) was added to the growth media at the following concentrations, 100 mg/mL for S. suis 2 and 50 mg/mL for E. coli.

Construction of mutant strains
The spxA1 and spxA2 genes were deleted in the SC19 background by allelic exchange using the thermosensitive suicide vectors pSET4s as previously described [42]. Two flanking fragments (LA and RA) of an internal fragment of the spxA1 gene (bases 1 to 342) were amplified from the S. suis 2 genome by PCR using primers listed in Table 2. After digestion with the appropriate restriction enzymes, the two PCR products were simultaneously cloned into pSET4s to generate a spxA1 knockout vector, pSET4s::spxA1. The plasmid was introduced into SC19 competent cells by electroporation. After two steps of allelic exchange, spectinomycin-sensitive clones were selected and the presence of spxA1 in the genome was detected by PCR using specific primers listed in Table 2. Reverse transcription PCR (RT-PCR) and DNA sequencing was performed to confirm the deletion. D spxA2 was created in a similar manner.

Stress challenges
To investigate the role of the two Spx regulators in stress responses, the wild-type (WT) and mutant strains were subjected to a variety of stress challenges (0.5 mM H 2 O 2 , 0.01% SDS, 1.5% NaCl, 0.3 mM diamide, 42uC, 30uC, pH 5.5, and medium lacking newborn bovine serum). Overnight cultures of the WT and mutant strains were diluted in fresh medium adjusted to each specific condition, and growth was evaluated by measuring the optical density at 600 nm (OD 600 ) every hour.

Mouse infections
100 female BALB/c mice (4-6 weeks old) were randomly divided into 10 groups with 10 mice per group. Groups I, II and III, which served as the high dose groups, were inoculated intraperitoneally with 7610 8 CFU in 200 mL PBS of the WT, DspxA1 or DspxA2 strain. Groups IV, V and VI, which served as the lower dose groups, were inoculated intraperitoneally with 3.5610 8 CFU in 200 mL PBS of each strain. Groups VII, VIII and IX were inoculated following the same protocol as Groups I, II and III, except that bacteria were heat-inactivated (10 min at 80uC) prior to inoculation. Mice in group X were injected with 200 mL PBS as the control group. Mice were monitored daily for 14 days to determine survival rates. Surviving animals were sacrificed on day 14 post-infection. After euthanization, brain samples were collected and prepared for histological examination. To examine bacterial burden, mice were injected intraperitoneally with 1610 8 CFU of each strain. At 24 h following injection, mice were euthanized for collection of blood, brain, and spleen samples. Brains and spleens were weighed and homogenized in 1 mL PBS. Homogenates were serially diluted and plated on TSA to determine the number of viable bacteria. Blood samples were directly diluted for plating.

Histopathological studies
To examine differences in pathological changes, brain samples were collected from mice infected with the WT, DspxA1 and DspxA2 strains. Samples from the brain were fixed in 4% formaldehyde. After paraffin embedding, sections 4 mm thick were cut and stained with hematoxylin and eosin. Histology micrographs were obtained by light microscopy (Nikon, Tokyo, Japan).

Competitive-infection assay
For competitive-infection assay, 200 mL of a 1:1 D spxA1:WT or DspxA2:WT (5610 CFU) mixture was inoculated intraperitone-7 ally into mice. The ratio in the inoculum was determined by plating a suspension of each strain prior to mixing. Mice were sacrificed to collect blood samples 18 h after inoculation, and blood samples were diluted for plating. The DspxA1:WT ratio in blood samples was determined by analyzing 70 colonies from each sample with colony PCR using primers A1out1 and A1out2, which yielded 462-bp and 804-bp PCR products for DspxA1 and WT strains, respectively. The D spxA2:WT ratio was determined by the same method using primers A2out1 and A2out2, giving PCR products of 609-bp and 1008-bp for DspxA2 and WT strains, respectively. The competitive index (CI) was calculated as the mutant:WT ratio in blood samples divided by the ratio in the inoculum.

Measurement of Inflammatory Cytokines
To assess the differences in cytokine release trigged by the WT and mutant strains, a total of 36 infected and four uninfected mice were included for measurement of inflammatory cytokines. 36 mice were assigned randomly to three groups, and inoculated intraperitoneally with 2610 CFU of the WT, 8 DspxA1 or DspxA2 strains. At 6, 9, 12, 15 h post-infection, three mice per group and one uninfected mouse were sacrificed for collection of blood samples. Serum samples were isolated and preserved at 280uC until analysis. Levels of IL-6 and TNF-a in serum were determined using commercially available enzyme-linked immu-

Bactericidal assays
Bactericidal assays were performed as previously described [43][44][45], with slight modifications. WT, DspxA1 and DspxA2 strains were harvested at the mid-exponential growth phase, washed three times, and diluted in PBS to yield 1610 6 CFU/mL. Subsequently, 900 mL heparinized blood from healthy pig (free of S. suis) was mixed with 100 mL bacterial suspension and incubated at 37uC for 3 h while rotating. The number of viable bacteria in each sample after 1 or 3 h and original inocula were determined by plating. The growth factor was defined as the ratio of CFU in each sample after 1 or 3 h incubation over the CFU in the corresponding inoculum.

RNA isolation
S. suis strains were grown in TSB with 10% (vol/vol) newborn bovine serum to mid-exponential phase (OD 600 = 0.6). Total RNA was isolated using an SV total RNA isolation system (Promega), according to the manufacturer's recommended protocol. RNA concentrations and integrity were determined using an Agilent 2100 Bioanalyzer. The qualified RNA was then used for microarray analysis and qRT-PCR.

Microarray analysis
DNA microarray analysis was performed using an Agilent custom-designed oligonucleotide microarray. Based upon the whole genome sequence of SC84 [46], specific 60-mer oligonucleotide probes were designed using eArray (https://earray.chem. agilent.com/earray/), to cover all annotated genes, with the exception of 10 genes for which it was not possible to design specific probes. The final coverage was 99.47% (1888/1898 genes). Probes were printed seven times on microarray slides. Two biological replicates of total RNA from WT strain and three from each mutant strain were amplified and labeled with Cy3-CTP using Low Input Quick Amp Labeling Kit, one-color (Agilent technologies, US), following the manufacturer's instructions. Labeled cRNA was purified using the RNeasy mini kit (Qiagen). After fragmentation, microarray slides were hybridized with 600 ng Cy3-labeled cRNA. Hybridization was performed at 65uC for 17 h with rotation at 10 rpm. Microarray slides were washed and scanned by an Agilent Microarray Scanner (G2565CA). Those genes with greater than two-fold change ratios were regarded as differentially expressed genes. Microarray data has been deposited into the NCBI Gene Expression Omnibus (GEO) with accession number GSE56760.

Quantitative Real-time PCR evaluation
A subset of genes was selected to confirm the accuracy of the microarray data by qRT-PCR with SYBR Green detection. The primers (Table S1) were designed according to the genomic sequence of SC84 [46]. First-strand cDNA was generated from total RNA using an Reverse Transcriptase kit (Toyobo, Japan). Quantitative PCR was conducted using the THUNDERBIRD SYBR qPCR Mix (Toyobo, Japan) according to the manufacturer's instructions. Quantitative analysis was performed in triplicate with an ABI 7500 Fast Real-Time PCR system. 16S rRNA was used as a housekeeping control gene. The relative expression level was calculated using the comparative cycle threshold (2 2DDCt ) formula normalized to the 16S rRNA level [47]. Student's t test was performed to verify the significance of the real-time PCR quantifications.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 5 (San Diego, USA). Survival data were analyzed with the log-rank (Mantel-Cox) test. Differences in bacterial burdens were analyzed using the two-tailed Mann-Whitney test. The data in competitiveinfection assay were analyzed using the two-tailed paired t test. The two-tailed unpaired t test was used to analyze the production of inflammatory cytokines in mice and bacterial survival in pig blood. P-values,0.05 were considered statistically significant.

Identification of Spx homologues in S. suis
A BlastP search against the proteins annotated in the genome of S. suis strain SC84 [46] was performed using the SpxA1 (locus tag spr1262) and SpxA2 (locus tag spr0173) proteins of S. pneumoniae strain R6 [48]. Two significant hits were identified, designated SpxA1 (locus tag SSUSC84_0997), and SpxA2 (locus tag SSUSC84_0059). The spxA1 and spxA2 genes of S. suis are located at opposite positions on the chromosome (Fig. 1A). SpxA1 consists of 133 amino acids with a predicted pI value of 6.74 while SpxA2 consists of 132 amino acids with a predicted pI value of 8.89. Pfam searches (http://pfam.janelia.org/) placed both proteins in the ArsC family. BlastP searches of protein database available from the National Center for Biotechnology Information confirmed that the two proteins are conserved across all sequenced S. suis strains. A multiple alignment was performed of S. suis SpxA1, SpxA2 and other Spx proteins, from either streptococci species (SpxA and SpxB from S. mutans, SpxA1 and SpxA2 from S. pneumoniae and S. sanguinis), or other species (Spx from B. subtilis, S. aureus, and E. faecalis). The results (Fig. 1B) showed that S. suis SpxA1 and SpxA2 possess two conserved residues/ motifs: the amino terminal CXXC motif involved in redox state stress sensing by disulfide bond formation in B. subtilis [33,49], and the Gly52 residue responsible for the interaction of B. Subtilis Spx with the RNA polymerase a-CTD [29,33]. A carboxyl terminus RPI motif implicated in both modulating the reactivity of the CXXC motif and binding sulfate in vivo [33] is present in SpxA1, but is found as SPI in SpxA2. The high level of homology exhibited by these proteins suggests that SpxA1 and SpxA2 of S. suis may also share important functions similar to other Spx proteins.

Microbiological characterization of Dspx strains
To investigate the role of the Spx regulators in S. suis 2, inframe deletion mutants of the two spx genes, designated DspxA1 and DspxA2 were constructed through homologous recombination ( Fig. 2A). The mutation was confirmed by PCR detection (Fig. 2B), RT-PCR (Fig. 2C), and direct DNA sequencing of the mutation sites (data not shown). In both S. mutans UA159 and S. sanguinis SK36, a double mutant of the two spx genes is viable [37,38]. In contrast, simultaneous inactivation of spxA1 and spxA2 in S. pneumoniae R6 is lethal [30]. We also failed to obtain the double mutant of the two spx genes in S. suis SC19.
The effects of deletion of the spx genes on the basic biological properties of S. suis were examined. In solid medium, DspxA1 formed colonies of reduced size, while DspxA2 showed no major difference from the WT strain. Deletion of the spx genes had no obvious effects on the haemolytic activity (unpublished observations). Cells grown to the exponential phase were observed by TEM (Fig. S1). Measurement of capsule thickness showed no obvious differences between the WT and mutant strains. The growth kinetics of Dspx strains were compared to those of the WT strain by measuring OD 600 values. When grown in liquid culture at 37uC under static growth conditions, DspxA2 produced fewer CFU (data not shown), but exhibited a higher OD 600 value during the stationary phase than WT (Fig. 3A). In comparison, DspxA1 demonstrated moderate reductions in OD 600 value and CFU numbers (data not shown) during the stationary phase. In liquid culture with shaking at 180 rpm, DspxA1 showed a significant defect in growth as compared with the WT, while the growth kinetics of DspxA2 was almost identical to that of the WT strain (Fig. 3B).

Role of SpxA1 and SpxA2 in stress tolerance
Spx regulators were found to play important roles in stress response in many species [35][36][37][38][39]. We, therefore, examined whether S. suis SpxA1 and SpxA2 were also involved in stress tolerance. The WT and mutant strains were cultured under a variety of stress conditions, and growth curves were compared with those of strains cultured under normal conditions (TSB with 10% newborn bovine serum at 37uC under static conditions). The results indicated that DspxA1 is highly sensitive to hydrogen peroxide (Fig. 3C), while DspxA2 is highly sensitive to SDS (Fig. 3D) and NaCl (Fig. 3E). Both mutants were defective when cultured in TSB in the absence of newborn bovine serum (Fig. 3F). However, no obvious growth difference was observed between strains cultured at 37uC and those cultured at a higher temperature (42uC, Fig. S2A), or a lower temperature (30uC, Fig. S2B). Furthermore, Spx proteins in S. suis were not associated with either acid tolerance (Fig. S2C) or resistance to diamide, a thiol-specific oxidant (Fig. S2D). These results strongly suggest that Spx proteins play a general role in stress response. Deletion of the spx genes affects S. suis virulence in mice Spx had been shown to modulate virulence in several Grampositive bacterial species [37][38][39]. We investigated the role of S. suis SpxA1 and SpxA2 proteins in virulence using a murine infection model. In the higher inoculum dose groups, mice infected with the WT and DspxA2 strains developed typical clinical symptoms of S. suis 2 infection, including rough coat hair, limping, lethargy, and swollen eyes. Mice infected with D spxA1 mutant showed only mild symptoms. As shown in Figure 4A, 80% of the mice in the WT group and 100% of the mice in the D spxA2 group died within 24 h. In contrast, only two mice in the DspxA1 group died within 48 h. The survival rates were significantly lower in mice infected with the WT strain than in those infected with the DspxA1 strain (P = 0.0059). No significant difference was observed between the WT-infected group and DspxA2-infected group (P = 0.1462). In the lower dose inoculum groups (Fig. 4B), mice infected with the WT and DspxA2 strains also displayed typical clinical symptoms of S. suis 2 infection, with survival rates of 40% and 10%, respectively (P = 0.1003 for DspxA2). No mice in the DspxA1 group developed clinical symptoms or died (P = 0.0040 for DspxA1). All mice inoculated with heat-inactivated bacteria or PBS remained healthy (data not shown). In addition, pathological examination showed that the meninges of the mice infected with the WT strain were severely thickened and a mass of macrophages and neutrophils could be observed. Similar pathological alterations occurred in the meninges of DspxA2-infected mice, but to a lesser degree, while the meninges of DspxA1-infected mice were similar to those of normal mice (Fig. 5). These findings indicate that the deletion of spxA1 decreases the lethality of S. suis 2 in mice, and that the mortality was due to infection and not the toxic effects of bacterial components.
To compare the abilities of the WT and mutant strains to establish infection, live bacterial cells from the three strains were examined in tissues of mice infected with sublethal doses of approximately 10 8 CFU at 24 h post-infection. The number of bacterial cells of the WT strain recovered from the blood (Fig. 6A), brain (Fig. 6B), and spleen (Fig. 6C) was significantly higher than those from the DspxA1 and DspxA2 strains. Subsequently, competitive infection assays were performed to further evaluate the abilities of the WT and mutants strains to colonize the blood. Groups of six mice were inoculated intraperitoneally with a 1:1 mixture of DspxA1:WT or DspxA2:WT. Bacterial cells recovered  3) and DspxA2 strains (lanes 5 and 7), were used as templates. (C) RT-PCR identification of the mutant strains. Total RNAs were extracted from the WT, DspxA1 and DspxA2 strains. cDNAs generated from these RNA samples were subjected to RT-PCR analysis with primer pairs A1in1/A1in2 (for detection of spxA1 gene transcripts) or A2in1/A2in2 (for detection of spxA2 gene transcripts). The RT-PCR products were analyzed by electrophoresis on a 1% agarose gel (lanes 1 and 3, the WT strain; lanes 2, the DspxA1 strain; lanes 4, the DspxA2 strain). doi:10.1371/journal.pone.0108197.g002 from the blood were analyzed by colony PCR to determine the CI. The results show that the CI values for D spxA1 and DspxA2 in vivo were significantly less than 1 (Fig. 7), suggesting that both mutant strains had reduced abilities to colonize the blood.

Induction of inflammatory cytokines in mice
To compare the capacity of S. suis strains to induce inflammatory cytokines, production of TNF-a and IL-6 was measured in the serum of infected mice. As shown in Figure 8A, the production of TNF-a, an important host mediator in the pathogenesis of septic shock [50], induced by either DspxA1 or DspxA2, is significantly lower than that induced by the WT strain at 6 h post-infection. However, the production of IL-6, an important inducer of acute phase proteins [51], is clearly higher in DspxA2 infected mice, and lower in DspxA1 infected mice (Fig. 8B). To evaluate production of inflammatory cytokines over time after infection with S. suis strains, production of TNF-a and IL-6 was measured in serum of infected mice at four time points (6, 9, 12, 15 h after infection). As shown in Figure 8C and 8D, serum levels of TNF-a and IL-6 from the WT-infected mice decreased at 9 h post-infection, remained at high levels, and returned to basal levels at 12 h post-infection. The mutant strains DspxA1 and DspxA2 triggered a very low production of TNF-a. Additionally, serum levels of IL-6 from DspxA1 and DspxA2-infected mice returned to basal levels at 9 h post-infection, approximately three hours earlier than the WT-infected mice.

Deletion of the spx genes significantly attenuates survival of S. suis in pig blood
To determine whether deletion of the spx genes affects survival of S. suis in whole blood, we measured the ability of S. suis strains to grow in healthy nonimmune pig whole blood. After 1 h of incubation, the mean growth factors (ratio of S. suis CFU in cultures over inocula), of WT, DspxA1 and DspxA2 were 2.03060.879, 0.93760.306, and 0.37360.117, respectively. After 3 h of incubation, the mean growth factors were 48.605610.945, 0.30360.287, and 0.71060.030, respectively (Fig. 9). These results suggest the WT strain can evade immune components in blood and proliferate, while the mutants both showed decreased survival in blood.

Microarray analysis reveals the global regulatory roles of SpxA1 and SpxA2
To reveal the scope of the Spx regulation and explore the differences in the regulatory roles of SpxA1 and SpxA2, the global  gene transcription profiles of the WT, DspxA1 and DspxA2 strains grown to mid-exponential phase were determined by DNA microarray analysis. As expected, the expression levels of flanking genes of spx were unaltered in the mutant strains, confirming that spx deletions don't have a polar effect on the genes residing downstream of spx. A large number of genes were differentially expressed in each mutant strain compared to the WT strain (165 genes in DspxA1 and 404 genes in DspxA2), revealing the global regulatory role of SpxA1 and SpxA2. Interestingly, there was minimal overlap between genes that were differentially expressed in DspxA1 and DspxA2 (only 15 genes expressed with the same trends in both strains), suggesting each Spx protein exerts its regulatory functions in an independent manner. These differentially expressed genes can be classified into several functional categories (Table S2), including information storage and processing, cellular processes and signaling, metabolism, and poorly characterized. A subset of 10 genes with varying expression levels was selected to confirm the accuracy of the microarray data by qRT-PCR. There was a strong positive correlation between the data obtained by the two methods (Fig. 10).
Several genes involved in the oxidative stress response are positively regulated by SpxA1 and/or by SpxA2 The expression of several genes known to be involved in the oxidative stress response (gor, glutathione reductase; nox, NADH oxidase; tpx, putative thiol peroxidase; sodA, superoxide dismutase; dpr, Dps-like peroxide resistance protein) [37] were downregulated in DspxA1 and/or in DspxA2 (Table 3), suggesting these genes were positively regulated by SpxA1 and/or SpxA2. Downregulation of nox, tpx and sodA in DspxA1 might be responsible at least in part for the impaired growth of DspxA1 in the presence of hydrogen peroxide. In addition, gor, tpx and dpr were downregulated in DspxA2, suggesting that SpxA2 might also be involved in the oxidative stress response under other conditions.

Expression of several proven or putative virulenceassociated factors is altered in the mutant strains
Analysis of the microarray data revealed a change in expression of several genes involved in virulence (Table 3). Compared to the WT strain, the expression of genes encoding superoxide dismutase (sodA), accessory pilus subunit (SSUSC84_1907 and SSUSC84_1908) and Ectonucleotidase NudP (SSUSC84_1385) was downregulated in DspxA1. In addition, the microarray analyses also revealed downregulation of vicR (encoding response regulator protein of the VicRK TCS), endA (encoding competence associated endonuclease) and gene SSUSC84_1908 (encoding accessory pilus subunit) in D spxA2. These genes have been studied in S. suis or other streptococci species and shown to be implicated in the infection process [52][53][54][55][56][57]. Downregulation of virulenceassociated factors might help to explain the phenotypes of the mutant strains in relation to the reduced abilities to colonize the tissues, and the decreased level of survival in pig blood.

Discussion
Bacteria use global regulatory networks to sense and modify gene expression in response to changing environments. In many low-GC Gram-positive bacteria, Spx proteins are global transcriptional regulators that play a pivotal role in the regulation of stress tolerance and virulence [28,[37][38][39]. Although it has been described in a number of species, including the major streptococcal pathogens S. pneumoniae [30], S. mutans [37] and S. sanguinis [38], the role of Spx in S. suis has, so far, received little attention. Studies in S. pneumoniae [30], S. mutans [37] and S. sanguinis [38] uncovered two Spx proteins. In S. pneumoniae,  simultaneous inactivation of spxA1 and spxA2 was lethal, but single-gene inactivation suggested that SpxA1 negatively regulates the X-state development by repressing transcription of the early competence operon comCDE [30]. In S. mutans, SpxA and SpxB were shown to modulate stress tolerance and were, for the first   time, demonstrated to be required for virulence in Gram-positive pathogen [37]. In S. sanguinis, SpxA1 was shown to be involved in hydrogen peroxide production, stress tolerance and endocarditis virulence, while SpxA2 affected growth under normal conditions [38].
In the present study, we investigated the functional role of Spx in S. suis 2. Like some other streptococci [30,[37][38], two spx genes were discovered in the genome of S. suis 2, which were renamed spxA1 and spxA2. Bioinformatics analysis showed that the Spx proteins were highly conserved not only among streptococci species, but also among other low-GC Gram-positive bacteria, suggesting that Spx regulators of S. suis 2 very likely function similarly to their orthologs in other bacteria, especially streptococci. To evaluate the role of Spx in S. suis 2, we constructed two mutants, DspxA1 and DspxA2. Interestingly, we failed to obtain the double mutant of spxA1 and spxA2 in S. suis 2, which has similarly been noted in S. pneumoniae [30], suggesting their simultaneous inactivation is lethal. Although SpxA1 and SpxA2 share 45% identity at the amino acid level, independent deletion of spxA1 and spxA2 leads to different phenotypes, such as defective growth for DspxA1 and normal growth for D spxA2 in liquid culture with shaking at 180 rpm.
To investigate the involvement of SpxA1 and SpxA2 in stress tolerance, the WT, DspxA1 and DspxA2 strains were subjected to a variety of stress challenges. Our data show that DspxA1 is highly sensitive to hydrogen peroxide, while DspxA2 is highly sensitive to SDS and NaCl, and that both mutants are defective in medium lacking newborn bovine serum. The impaired growth of DspxA1 under hydrogen peroxide stress, coupled with microarray analysis, indicates that SpxA1 plays an important role in the oxidative stress responses in S. suis. Similar result has also been reported for SpxA in S. mutans [37] and Spx in E. faecalis [39]. Although no oxidation-sensitive phenotype was observed in DspxA2, downregulation of several genes involved in oxidative stress suggests that SpxA2 might play a secondary role in control of oxidative stress or in regulation of the oxidative stress response under other conditions, such as anaerobic conditions. Unlike the staphylococcal spx mutant that was hypersensitive to high and low temperatures [35], no phenotype is associated with spxA1 or spxA2 deletion for heat or cold tolerance in S. suis, which is in agreement with spxA1 inactivation in S. pneumoniae [30]. In addition, deletion of spxA1 or spxA2 did not affect the stress response to reduced pH. These findings are not surprising, as neither of the spx genes is included in the genes of S. suis 2 induced by temperature [58] or acidic stress [59]. Although Spx regulators have been reported to be involved in thiol-specific oxidative stress responses in many species [32,35,37,39], mutant strains showed no significant differences in growth compared to the WT strain in the presence of diamide, a thiol oxidizing agent, similarly to results with DspxA1 in S. pneumoniae [30].
To evaluate the role of SpxA1 and SpxA2 in the pathogenesis of S. suis 2, a murine infection model was used. Deletion of spxA1 results in decreased lethality of S. suis 2 in mice while deletion of spxA2 has no effect on lethality. In order to confirm that death of mice was due to infection and not to toxic effects of bacterial components, we performed infection of mice with heat-killed bacteria and observed no clinical symptoms. Pathological examination revealed that no obvious change was observed in the meninges of DspxA1-infected mice, while the meninges of DspxA2infected mice suffered less damage than that of the WT-infected mice. This observation could be explained by subsequent Figure 10. Correlation between DNA microarray data and qRT-PCR results. The relative transcriptional level of 10 selected genes determined by DNA microarray and qRT-PCR analyses were log 2 transformed, and the values were plotted against each other to evaluate their correlation. The genes analysed by qRT-PCR are listed in Table S1. doi:10.1371/journal.pone.0108197.g010 Table 3. Expression ratios of genes involved in oxidative stress response and virulence in the mutant strains relative to the WT strain by microarray analysis a . colonization experiments, which showed that the number of bacterial cells of the mutant strains recovered from the brain was much less than that of the WT strain. Previous studies have shown that Spx contributes to colonization during infections, such as SpxA of S. mutans [37] and Spx of E. faecalis [39]. Our colonization analysis showed that recovered bacterial number of the mutant strains from the blood, brain, and spleen was significantly reduced. Competitive infection assay further revealed that the CI values for DspxA1 and DspxA2 in blood were significantly less than 1. These results suggest that both mutant strains reduce their abilities to colonize the tissues. It is surprising that DspxA2 exhibits high lethality and reduced ability to colonize the tissues at the same time. Considering that the inflammatory response plays an important role in the pathogenesis of S. suis infection in mice [50], the effect of Spx regulators on the inflammatory response was investigated. The results show that serum levels of TNF-a in mice infected with mutant strains, either DspxA1 or DspxA2, are significantly lower than in WT-infected mice. Although the difference is not significant, serum levels of IL-6 in mice infected with DspxA2 are obviously higher than in WTinfected mice at 6 h post infection. Besides, Serum levels of TNF-a and IL-6 in mice infected with mutant strains return to basal levels at 9 h post infection, more quickly than in WT-infected mice. Previous study of group A streptococci showed that high levels of both TNF-a and IL-6 were inversely correlated with survival time in patients with sepsis [60]. We speculated that a higher dose of DspxA2 induced excessive production of IL-6, which would damage the mice and lead to the high fatality rate. While mice were infected with a lower dose of DspxA2 in colonization and competitive infection assays, low levels of IL-6 production would be beneficial for the mice to clear infection, thus lead to reduced DspxA2 recovered from the tissues. To the best of our knowledge, this is the first description of Spx involvement in triggering of the host inflammatory response. Dissemination via the bloodstream is considered to be an important step in the pathogenesis of S. suis 2 infection [61]. Bactericidal assays were further performed to examine survival of WT and mutant strains in pig blood. Both mutant strains show a significantly decreased survival ability in pig whole blood, indicating that SpxA1 and SpxA2 are involved in bacterial resistance to phagocytosis. The lower survival ability of mutant strains in blood might be partly responsible for their decreased abilities to colonize the tissues.
Considering the global regulatory role of Spx in B. subtilis [32], S. aureus [35] and S. mutans [37], microarray analysis of the WT and mutant strains was performed to identify genes under Spx regulation. The analysis revealed that transcription of a large number of genes is affected by SpxA1 and SpxA2, directly or indirectly. Of note, few genes are expressed with the same trends in both strains, suggesting that each Spx protein regulates gene expression independently, similarly to SpxA and SpxB in S. mutans. This might be the reason that we failed to obtain a mutant containing spxA1 and spxA2 deletions simultaneously, and that different phenotypes are observed in DspxA1 and DspxA2. Downregulation of several genes involved in the oxidative stress response in mutants might be partly responsible for the phenotypes of defective growth under conditions of oxidative stress. It has been demonstrated that vicR inactivation could increase susceptibility to osmotic pressure in S. pyogenes [62]. The impaired growth under NaCl stress of DspxA2 might be involved in the downregulation of vicR (Table 3). During the infection process, S. suis 2 may transform its metabolic mode to adapt to the new environment within the host [11]. Our data show that a large number of genes involved in metabolism are differently expressed in mutants, suggesting that Spx regulators in S. suis 2 play a critical role in regulation of genes involved in metabolism, which might not only be responsible for the defective growth of the mutant strains in culture without newborn bovine serum, but also facilitate the survival of S. suis 2 within the host. In addition, several proven or putative virulence-associated factors were downregulated in mutants, either DspxA1 or DspxA2, helping to explain the reduced ability to colonize the tissues, and the decreased level of survival in pig blood. It has been shown that S. suis, especially its cell wall components, could induce both TNF-a and IL-6 production by murine macrophages [63]. Some genes involved in cell wall/ membrane biogenesis were expressed differently in mutant strains (Table S2), which might be the reason that the two spx genes have an effect on induction of inflammatory response. Also, it is not surprising that DspxA2 induced a higher production of IL-6, as the gene pgdA, which has been demonstrated to promote the production of IL-6 [64], was 2.16-fold upregulated in DspxA2.
In summary, bioinformatics, mutational and microarray analysis were used to identify and characterize two Spx regulators of S. suis 2. The present study clearly demonstrates that Spx regulators modulate stress tolerance in S. suis 2. Specifically, we show that Spx regulators globally modulate gene expression, especially genes involved in metabolism. We also demonstrate that the mutant strains show reduced abilities to colonize the tissues and decreased survival in pig blood. More importantly, this study demonstrates, for the first time, that Spx regulators are involved in triggering the host inflammatory response.