The bacterial strains and plasmids used in this study are listed in Table 1.

ZY05719, a virulent strain of SS2, was isolated and identified from a dead pig during an outbreak in Sichuan, China in 2005. SS2 ZY05719 and its derivative mutants were cultured at 37°C in Todd-Hewitt (TH, BD, Cat.No.249240) broth or on TH agar. Escherichia coli strain DH5α was used as a host for plasmids and cultured in Luria-Bertani (LB, Sigma-Aldrich, Cat.No.L3147) broth or on LB agar at 37°C. Once the bacteria contained plasmids, 50 or 100 μg mL⁻¹ spectinomycin was added to the culture medium for E. coli or SS2, respectively. The murine microglia cell line BV2 was purchased from the Cell Resource Center, IBMS, CAMS/PUMC. As a valid substitute for primary microglia cells, a special type of phagocyte, this cell line exhibits morphological and functional characteristics of microglia (Blasi et al., 1990; Bocchini et al., 1992; Henn et al., 2009). It was cultured in Dulbecco's modified Eagle medium (DMEM) high glucose (GIBCO, Cat.No.11965) supplemented with 10% fetal bovine serum (GIBCO, Cat.No.10099-141) at 37°C under 5% CO₂.

The primers used in these experiments are listed in Table 2.

The upstream and downstream flanking sequences of the DNA fragment to be deleted were amplified utilizing two pairs of primers, UF/UR and DF/DR, by PCR. Then, they were integrated together by PCR using primer pair UF/DR. After digestion with restriction endonucleases (TaKaRa, EcoR I, Cat.No.1040S; BamH I, Cat.No.1010S; Sal I, Cat.No.1080S), the PCR products were ligated into pSET4s (Takamatsu et al., 2001a) with T4 DNA Ligase (TaKaRa, Cat.No.2011A). Recombinant vectors were electroporated into wild-type ZY05719. Transformants were cultured at 37°C on TH agar in the presence of spectinomycin. A single colony was subjected to serial passages at 28°C in TH broth free from spectinomycin. Subsequently, the bacterium solution was diluted and plated on TH agar. Spectinomycin-sensitive colonies were verified for deletion of a particular DNA fragment by PCR using primer pairs UF/DR and F/R as well as product sequencing.

The DNA fragment containing a whole ORF of a deleted gene and its promoter sequence was amplified by PCR using primer pair CF/CR and cloned into pSET2 (Takamatsu et al., 2001b). Verified recombinant vector was electroporated into the mutant to construct the complementation strain.

The wild-type SS2 strain ZY05719, its derived mutants, and complementation strains were cultured to the mid-log phase (at an OD₆₀₀ of 0.6). Total RNA was extracted using RNAiso Plus (TaKaRa, Cat.No.D9108A) following its manufacturer's instructions. The remaining genome was eliminated using DNase I (Fermentas, Cat.No.EN0521). cDNA synthesis was completed using the PrimeScript™ RT reagent kit (TaKaRa, Cat.No.RR037A) and mRNA levels were measured with the SYBR® Premix Ex Taq™ kit (TaKaRa, Cat.No.RR420A) according to the manufacturer's protocol. Primers used for the various qPCR assays are listed in Table 2. The 16s RNA gene was used as the reference gene (Wang et al., 2011). Relative changes in gene transcriptional levels were calculated utilizing the comparative CT method (Livak and Schmittgen, 2001). Each set of qPCR assay was repeated three times with independent RNA preparations.

Phagocytosis assays were performed as described before with some modifications (Meijerink et al., 2012; Redlich et al., 2012). BV2 cells grown to approximately 5.0 × 10⁵ cells/well in a 24-well plate were used for assays. Bacteria grown to the log-phase were harvested by centrifugation (5,000 g, 5 min), washed with and resuspended in DMEM, and added to cells at a multiplicity of infection (MOI) of 1. After 1 h of incubation at 37°C under 5% CO₂, cells were washed three times with phosphate buffered saline (PBS) and were added to DMEM containing 5 μg mL⁻¹ penicillin and 100 μg mL⁻¹ gentamicin for another 1 h of incubation to kill extracellular bacteria. Then, after another three washes with PBS, cells were lysed with double-distilled water to recover intracellular bacteria. Recovered bacteria were calculated using the plate count method. The phagocytic rate was calculated as the percentage of the CFU number of intercellular bacteria recovered in the CFU number of initial inoculum. Assays were repeated as three independent experiments.

Intracellular survival analysis in microglial cells was performed based on previous reports (Cybulski et al., 2008; Tang et al., 2012) with some modifications. Briefly, the protocols were the same as the phagocytosis assay except that the incubation time after the cells were added to DMEM containing antibiotics were 1 and 3 h, respectively. Intercellular bacteria were recovered after each time point of incubation and calculated. The survival rate was calculated as CFU_3h/CFU_1h × 100%.

For visual observation of intracellular SS2, samples were fixed with 2.5% glutaraldehyde and 2% formaldehyde in cacodylate buffer (0.01 M MgCl₂, 0.01 M CaCl₂, 0.09 M sucrose, 0.1 M cacodylate, pH 6.9) for 3 h at 4°C, washed with cacodylate buffer and treated with 1% osmium tetroxide in cacodylate buffer for 1 h at room temperature (Hammerschmidt et al., 2005). Then samples were dehydrated with a graded series of ethanol, embedded in an epoxy resin, sectioned, and counterstained with uranyl acetate. After air-drying, samples were observed on a Hitachi H7650 transmission electron microscope (Hitachi, Tokyo, Japan) at a magnification of 1,200 × and an acceleration voltage of 80 kV.

Survival assays in whole blood were performed as described earlier (Bonifait et al., 2010; Zhu et al., 2011). Blood samples were collected from healthy pigs and used immediately. Bacteria were cultured to an OD₆₀₀ of 0.6, collected by centrifugation (5000 g, 5 min), and washed and resuspended in PBS to an OD₆₀₀ of 0.2. Then, 1 mL of whole blood and 100 μL of bacteria in PBS were mixed together and incubated for 3 h at 37°C with occasional gentle shaking. After 0 or 3 h of infection with SS2, whole blood samples were diluted and plated on TH agar to determine the bacterial survival. The time point of 0 h was considered as 100 percent viability. Tests were performed three times.

The survival abilities of SS2 in hydrogen peroxide and acidic conditions were determined as described elsewhere with some modifications (Wen et al., 2011; Tang et al., 2012; Fulde et al., 2014). Briefly, bacteria grown to an OD₆₀₀ of 0.6 were harvested by centrifugation and washed three times with 0.1 M PBS (pH 7.4). Then, the pellets were resuspended in 0.1 M PBS at a pH of 5 for 1 h of incubation at 37°C for survival assays in the acidic condition, or in 0.1 M PBS (pH 7.4) containing 5 mmol/L H₂O₂ for 2 h of incubation at 37°C for survival assays in the peroxidation environment, respectively. The pellets were resuspended in 0.1 M PBS at a pH of 7.4 for 1 h or 2 h of incubation at 37°C was used as negative controls. Results were recorded as the percent survival compared with the initial inoculum. Each assay was performed three times.

BV2 cells were stimulated with killed bacteria as previously reported with some modifications (Segura et al., 2004; Houde et al., 2012). After overnight growth, bacteria were washed with PBS and heat killed at 60°C for 45 min. Killed bacteria from 1 × 10⁹ CFU were incubated with BV2 cells (5 × 10⁵ cells/well) that had previously been plated in 24-well plates. The culture supernatants were collected after 6, 12, 24, and 48 h, respectively. Production of nitric oxide (NO) by BV2 cells was measured with a Nitric Oxide Assay Kit (Beyotime, Cat.No.S0021) according to the manufacturer's instructions. Briefly, 50 μL of Griess Reagents I and II was added in turns to the collected 50-μL supernatant samples. Absorbance at OD₅₄₀ was spectrophotometrically determined. The NO concentration was calculated using the standard curve method with different concentrations of NO diluted in DMEM containing 10% FBS. Assays were repeated three times.

Cytokines were detected using an enzyme-linked immunosorbent assay (Kim and Weiser, 1998) as reported earlier with some modifications (Segura et al., 1999; Dominguez-Punaro et al., 2010). The SS2 wild-type strain ZY05719, mutant and complementation strains were added to BV2 at an MOI of 2 for 12 h at 37°C under 5% CO₂. Commercial ELISA kits (Boster, TNF-α (m) ELISA kit, Cat.No.EK0525, IFN-γ (m) ELISA kit, Cat.No.EK0375) were used to measure the production of TNF-α and IFN-γ in the harvested culture supernatant.

Each experiment was repeated three times, and every treatment group in each time of test was set up three biological repetitions. The data were collected for analyzing and plotting using GraphPad Prism (version 5.0; GraphPad Software, [http://www.graphpad.com]) as reported earlier (Weissgerber et al., 2015). Data are presented as means with standard deviations (SDs). Data statistical analyses for all pairwise comparisons were assessed using Student's t-test. Differences were considered to be significant for a P < 0.05.

The insertion mutant of SS2 ZY05719 hsdS gene mentioned in the introduction was named MhsdS. A complete genome sequencing of MhsdS was performed, and there is no DNA sequence change when compared with wild-type ZY05719 except the insertion mutation of hsdS gene (data not shown). To get a better insight of a series of trait changes, which were as results of the mutation of hsdS gene, the complementation strain that MhsdS containing pSET2::hsdS was successfully constructed and named CMhsdS. As the hsdS belongs to the Type I R–M system SsuZY05719II, we constructed a hsdS′ deletion mutant and its complementation strain, which were named ΔhsdS′ and CΔhsdS′, respectively, to analyze the difference in the functional impact of these two genes.

Meanwhile, RNA samples of wild-type ZY05719 and MhsdS grown to the mid-log phase were extracted and sent to Sinogenomax in China for RNA sequence and gene transcription analysis (Table S1). As a result, the transcriptional levels of both gene peptidoglycan-binding proteins lysM (locus_tag ZY05719_RS01210) and lysM′ (locus_tag ZY05719_RS10235) were more than two-fold decreased when MhsdS was compared with wild-type ZY05719. Then, the changes in the transcriptional levels of these two genes were shown to be significant, and the transcriptional regulation of these two genes was impacted by hsdS, not by hsdS′, according to qPCR assays (p < 0.001, Figures 2A,B). We constructed lysM, lysM′ and double gene deletion mutants, which were named ΔlysM, ΔlysM′ and ΔlysMΔlysM′, respectively. As both lysM and lysM′ were located downstream of the ORF of their upstream and downstream genes from the transcription direction (Figures 1B,C), the gene deletions did not theoretically affect the transcription of upstream and downstream genes; these observations were confirmed using qPCR assays, which showed there were no significant differences between the wild-type SS2 and the lysM or lysM′ deletion mutants on the transcript levels of their corresponding upstream and downstream genes (Figure 2C). As a result, the complementation strains of ΔlysM and ΔlysM′ did not need to be constructed.

There was no difference in the growth kinetics among wild-type ZY05719, its insertion or deletion mutations, and the complementation strains mentioned above (Figure S1).

Phagocytosis assays showed the phagocytic rates of BV2 cells to wild-type SS2, MhsdS, and CMhsdS were 2.46 ± 0.42%, 9.47 ± 0.58%, and 5.00 ± 0.29%, respectively (Figure 3). Compared with MhsdS, the phagocytic rates to wild-type SS2 and CMhsdS were reduced 74.0 and 47.2% (p < 0.001), which demonstrated hsdS was beneficial to SS2 against phagocytosis. The result of transmission electron microscopy also revealed that the mutation of hsdS made the bacteria more easily to be phagocytized by a phagocyte (Figure 4). Then, we analyzed the relationship between the function of hsdS′ and the anti-phagocytosis of SS2. As a result, there was no significant difference between wild-type ZY05719 and ΔhsdS′. Meanwhile, in comparison with the wild-type SS2, the phagocytic rates to ΔlysM, ΔlysM′ and ΔlysMΔlysM′ were significantly increased (p < 0.001). In addition, the phagocytic rate to ΔlysMΔlysM′ was significantly increased (p < 0.001) when compared with ΔlysM or ΔlysM′.

Next, we analyzed the influence of insertion mutation of hsdS on survival in phagocytes. The results showed that the survival rate of MhsdS in BV2 cells was significantly reduced when compared with wild-type ZY05719 or CMhsdS (p < 0.001, Figure 5A). The deletion of hsdS′ did not affect the survival of ZY05719 in BV2. Interestingly, the mean survival rates of ΔlysM, ΔlysM′ and ΔlysMΔlysM′ were less than 10%, and they were significantly reduced by 85.4, 82.2, and 90.9% (p < 0.001) when compared with wild-type ZY05719, respectively. Still, the survival ability of double gene deletion mutant ΔlysMΔlysM′ was significantly decreased (p < 0.001) when compared with ΔlysM or ΔlysM′.

The survival ability of wild-type ZY05719, its mutants and complementation strains in whole swine blood was determined (Figure 5B). The deletion of hsdS′ did not affect the survival ability of ZY05719 in whole blood. The insertion mutation of hsdS, as well as the deletion of lysM or/and lysM′, significantly reduced the survival rates of bacteria in whole blood compared with parent wild-type ZY05719 (p < 0.001). In comparison with MhsdS, the survival rate of the complementation strain, CMhsdS, was significantly increased (p < 0.001).

To gain a better insight into mechanisms against the host environment that were influenced by hsdS, lysM and lysM′, other survival assays in the special environment in vitro were evaluated. The survival rates of all mutants and complementation strains in 0.1 M PBS at a pH of 7.4 when compared with wild-type ZY05719 were not significantly different (Figure S2). The results of survival assays in PBS containing H₂O₂ showed the survival rates of MhsdS, ΔlysM, ΔlysM′ and ΔlysMΔlysM′ were significantly decreased when compared with wild-type ZY05719 (Figure 5C, p < 0.001). Meanwhile, the ΔlysMΔlysM′ significantly decreased the survival rate when compared with ΔlysM, or ΔlysM′ (p < 0.001). There were no significant differences in the survival rates among wild-type ZY05719, CMhsdS, ΔhsdS′ and CΔhsdS′ in PBS with H₂O₂, which indicated that the deletion of hsdS′ did not impact the survival ability of ZY05719 in the peroxidation environment.

Among all mutants, only the ΔhsdS′ significantly decreased the survival rate in acidic conditions (Figure 5D, p < 0.001) when compared with wild-type ZY05719. Additionally, the CΔhsdS′ significantly recovered this rate when compared with ΔhsdS′ (p < 0.001). This result revealed that bacterial resistance to the acid environment is regulated by hsdS′, not by hsdS, lysM, or lysM′, which was different from the aforementioned survival assays.

Nitric oxide, which is another antibacterial substance produced by phagocytes after incubation with SS2, was measured as time progressed. As a result, significantly increased NO production by bacteria stimulation appeared in the 12- to 24-h time period of incubation (Figure 6). Of all mutants and complementation strains, only the MhsdS and the CMhsdS significantly increased the NO production relative to wild-type ZY05719 after 24 and 48 h of incubation (p < 0.01, Figures 6C,D). The NO production, which was stimulated by CMhsdS, was significantly less than MhsdS after 24 or 48 h of incubation (p < 0.001, Figures 6C,D). These results indicated hsdS′, lysM, and lysM′ were not involved in the inhibition of NO production by phagocytes.

It was previously reported that tumor necrosis factor-alpha (TNF-α) and gamma-interferon (IFN-γ) act synergistically to induce the NO production by phagocytes (Ding et al., 1988). Therefore, the production of these two cytokines in BV2 cells was measured after 12 h of incubation with SS2. As a result, after they were infected with SS2, BV2 cells produced significantly higher levels of TNF-α (Figure 7A, p < 0.001) than un-infected cells. MhsdS significantly increased the induced levels of TNF-α compared to wild-type ZY05719 and CMhsdS (p < 0.001). On the other hand, the deletion of hsdS′, lysM, and lysM′ did not have a significant effect on the TNF-α production. BV2 cells did not produce INF-γ with or without SS2 infection (Figure 7B).

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.