Regulatory role of the RstB‐RstA system in adhesion, biofilm production, motility, and hemolysis

Abstract For infection, initial invasion of the host is of great importance, with adhesion playing a critical role. We previously demonstrated rstA and rstB are remarkably downregulated in Vibrio alginolyticus cultured under heavy metal and acidic stresses, with impaired adhesion, suggesting that rstA and rstB might be involved in adhesion regulation. The present study showed that rstA and rstB silencing resulted in impaired adhesion, biofilm production, motility, hemolysis, and virulence. Meanwhile, changes of temperature, starvation, and pH remarkably affected rstA and rstB expression. These findings indicated that (1) rstA and rstB are critical regulators of adhesion in V. alginolyticus; (2) rstA and rstB have remarkable effects on biofilm production, motility, hemolysis, and virulence in V. alginolyticus; (3) rstA and rstB modulate adhesion in response to environmental changes of temperature, pH, and starvation.

changes (Yamamoto et al., 2005). Every two-component system includes a response regulator (RR) and a sensor protein-histidine kinase (HK). Through histidyl-aspartyl phospho-relay, HK and RR form a signal transduction pathway. RstB and RstA are HK and RR, respectively.
The function of the RstB-RstA system in bacterial virulence is not well characterized. Tran, Han, Shi, and Guo (2016) found that invasion, motility, pyrimidine metabolism, and iron acquisition in Salmonella typhimurium are controlled by the RstB-RstA system. Menanteau-Ledouble and Lawrence (2013) demonstrated that mutation of the RstB-RstA system results in decreased adhesion and virulence in Edwardsiella ictaluri. Terceti et al. (2017) reported that RstB plays key roles in the hemolytic activity and pathogenicity of Photobacterium damselae. As V. alginolyticus adhesion is likewise impaired under stress, the RstB-RstA system was hypothesized to participate in this process.
This study aimed to examine whether the RstB-RstA system contributes to V. alginolyticus attachment to the host mucus, by assessing (1) the association of rstA and rstB with V. alginolyticus attachment, and (2) whether rstA and rstB modulate adhesion in response to environmental stimuli.

| Bacterial strains and culture conditions
Vibrio alginolyticu strain ND-01 is a clinical isolate from naturally infected P. crocea . It was cultured at 28°C in Luria-Bertani (LB) broth supplemented with 2% NaCl, with shaking (220 rpm).
To assess the effects of temperature on rstA and rstB expression levels, V. alginolyticus was incubated overnight in LB at 4°C, 15°C, 28°C, 37°C, and 44°C, respectively. The bacteria were harvested, resuspended, and incubated for 30 min at their respective culture temperatures , in triplicate. V. alginolyticus cultured at 4°C was used as a control group here, which means the expression levels of rstA and rstB at 15°C-44°C were compared to their counterparts at 4°C.
To assess the effects of pH on rstA and rstB expression levels, V. alginolyticus cells were cultured at pH5, pH6, pH7, pH8, and pH9, respectively, overnight at 28°C, and washed with PBS at the same pH used for culture, in triplicate (Yan, Chen, Ma, Zhuang, & Wang, 2007).
V. alginolyticus cultured at pH5 was used as a control group here, which means the expression levels of rstA and rstB at pH6-9 were compared to their counterparts at pH5.
To evaluate the effects of starvation on rstA and rstB expression levels, V. alginolyticus was incubated in PBS and adjusted to OD 600 nm = 0.3, and starved for 1, 3, 5, and 7 days, respectively, at 28°C. The plate counting method was used to quantify culturable cells Jiang et al., 2017;Lin et al., 2017). Three replicates were set up. V. alginolyticus starved for 1 day was used as a control group here, which means the expression levels of rstA and rstB in V. alginolyticus starved for 3-7 days were compared to their counterparts in V. alginolyticus starved for 1 day.

| Stable gene silencing
The rstA and rstB genes were silenced with vectors containing short hairpin RNA (shRNA) sequences targeting the rstA and rstB coding regions as previously described . The shRNA sequences were obtained from Shanghai Generay Biotech Co., Ltd.
Recombinant plasmids were transformed into Escherichia coli SM10 (Dongsheng, Guangzhou, China) by heat shock. Recombinant plasmids were then transferred from strain SM10 into V. alginolyticus by conjugation. The empty pACYC184 vector was used as a control.
Chloramphenicol was employed to screen clones with stable silencing at a concentration of 34 μg/ml.

| RNA extraction and reverse transcription
TRAzol (Dongsheng, Guangzhou, China) was used for total RNA extraction from V. alginolyticus. A Revert Aid Mu-MLV cDNA synthesis kit (Dongsheng, Guangzhou, China) was employed to synthesize firststrand cDNA from total RNA. These experiments were conducted according to the manufacturer's instructions.

| Quantitative RT-PCR (qRT-PCR)
Quantitative RT-PCR was carried out on a QuantStudio ™ 6 Flex realtime PCR system (ABI, USA) with the SYBR green I fluorescent dye (Dongsheng, Guangzhou, China). The mRNA expression levels were determined as previously described (Liu et al., 2017). Primers were designed with Primer Premier 5.0 ( Table 2). The expression levels of rstA and rstB were normalized to that of 16S RNA. Relative Expression Software Tool (REST 2008.-version 2) was used to assess the relative

| Mucus preparation
Based on a previously described method , skin mucus was collected from healthy P. crocea in Ningde, Fujian Province, China.
Briefly, the fish was washed with sterile PBS. For skin mucus collection, the surface gel layer of the skin was scrapped with a plastic spatula. The collected mucus was homogenized in PBS, and centrifuged twice at 4°C for 30 min (20,000g) to remove particulate materials. Then, 0.45-and 0.22μm pore filters were successively used for filtration. The mucus sample was adjusted to a final concentration of 1 mg protein/ml as previously proposed (Bradford, 1976).

| In vitro adhesion assay
According to Huang, Hu, et al. (2016), V. alginolyticus adhesion was analyzed. Briefly, 50 μl of mucus was spread onto a glass slide (22 × 22 mm) evenly, and fixed with methanol. Twenty minutes later, 1 ml of bacterial suspension (10 8 CFU/ml) was gently applied to mucus coated glass slides, and incubated for 2 hr in a humidified chamber at 25°C. The slides were then washed with PBS (5 times) to remove nonadherent bacteria.
The specimens were then fixed with 4% methanol for 30 min, followed by crystal violet staining for 3 min. Finally, microscopic observation (×1,000) was carried out, and the average number of adherent bacteria was determined. In each assay, 20 fields of view were randomly selected.

| Soft agar plate motility assay
By the soft agar approach, the flagellar motility of V. alginolyticus was evaluated. Overnight cultured V. alginolyticus was diluted to OD 600 = 0.03. Then, 1 μl of the suspension was gently dropped onto the center of LB agar plates, followed by incubation at 28°C. After 20 hr of culture, colony diameters were measured.

| Biofilm assay
As previously described (Luo et al., 2016), the biofilm assay was carried out for V. alginolyticus. Overnight V. alginolyticus cultures were adjusted to OD 600 nm = 0.2. Then, 150 μl of LB was mixed with 50 μl of bacterial culture per well in 96-well plates. After incubation at 28°C for 24 hr, sterile PBS was used for 3 washes. The samples were stained for 15 min with 200 μl of 1% crystal violet, rinsed with sterile PBS, and air dried. Finally, 200 μl acetic acid (33%) was used for solubilizing the stained biofilm, which was quantitated by measuring OD 590 nm . The experiment was performed in triplicate.

| Hemolysis assay
As previously described (Tsou & Zhu, 2010), hemolysis assays were performed. First, commercial rabbit blood (PingRui, Beijing, China) was rinsed three times with PBS. After incubation of 245 μl of culture supernatant with 5 μl of washed rabbit blood at 37°C for 1 hr with shaking (220 rpm), detection of released hemoglobin was carried out by measuring OD 540 nm . The percentage of total hemolysis was determined by comparison with negative and positive control samples (100% lysis by 1% Triton X-100). The experiment was performed in triplicate.

| Artificial infection
Epinephelus coioides was used for artificial infection as previously described (Liu et al., 2017). Sixty healthy E. coioides were randomly divided into 3 groups. Each fish was intraperitoneally administered 0.1 ml of V. alginolyticus suspension (10 7 CFU/ml) of wild type and silenced strains, respectively. Instead of V. alginolyticus suspension, sterile PBS was used in the negative control group. Mortality was observed every day for 10 days.

| Data processing
Results were reported as mean ± standard deviation (SD). Statistical analysis was conducted with the SPSS 13.0 software (SPSS, Chicago, IL, USA). Differences were assessed by one-way analysis of variance (ANOVA) followed by Dunnett's multiple comparison test. p < .05 was considered statistically significant.
These results supported the reliability of previous RNA-Seq findings.

| Effects of environmental changes on rstA and rstB expression levels
To evaluate rstA and rstB responses to temperature changes, their expression levels were detected at different temperatures (Figure 2A).

Gene
Primers for qRT-PCR The expression levels of both rstA and rstB showed an inverted Ushaped trend. However, these genes showed highest levels at 28°C.
These findings suggested that both rstA and rstB were sensitive to temperature changes.
To evaluate the responses of these genes to pH changes, their expression levels were assessed at different pH levels. As shown in Figure 2B, an inverted U-shaped trend was also obtained. Highest expression levels were found at pH 7.0 for both rstA and rstB, indicating that they were sensitive to pH changes.
To evaluate the responses of these genes to starvation, their expression levels were assessed under starving conditions. Starvation resulted in remarkably decreased gene expression levels, in a timedependent manner ( Figure 2C). These findings suggested that rstA and rstB were both sensitive to starvation.

| Effects of rstA and rstB silencing on adhesion
As shown by qRT-PCR, rstA and rstB were remarkably reduced in stably silenced clones, by 6.67-and 5.26-fold, respectively ( Figure 3A).
The adhesion capabilities of rstA-and rstB-silenced clones were compared with that of the wild type. A total of 1499 ± 115 adherent bacteria were obtained per field of view in the control group, for only 322 ± 21 and 108 ± 7 in the rstA-and rstB-RNAi groups, respectively ( Figure 3B). This finding indicated that the adhesion capability of V. alginolyticus was remarkably reduced after rstA and rstB silencing, respectively.

| Effects of rstA and rstB silencing on motility
The motility abilities of rstA-and rstB-silenced clones were also analyzed. As shown in Figure 3C, motility in rstA-and rstB-RNAi cells was significantly reduced, by 3.50-and 7.00-fold, respectively.

| Effects of rstA and rstB silencing on biofilm production
The biofilm formation ability of V. alginolyticus was remarkably reduced after rstA and rstB silencing compared with the control group, by 3.44-and 7.17 -fold, respectively ( Figure 3D).

| Effects of rstA and rstB silencing on hemolytic activity
Hemolytic activities were remarkably impaired in the rstA-and rstB-RNAi groups compared with the control group; rstB-RNAi displayed a stronger suppression of hemolysis than rstA-RNAi ( Figure 3E).

| Effects of rstA and rstB silencing on virulence
After artificial infection, mortality was dramatically lower in fish administered rstA-and rstB-RNAi strains, respectively, compared with the control group ( Figure 3F). Survival rates were 0%, 65%, and 80% in groups infected with wild-type, rstA-RNAi and rstB-RNAi strains, respectively. In the wild-type, rstA-RNAi, and rstB-RNAi groups, death occurred at days 2, 3 and 4, respectively.

| DISCUSSION
This study showed that rstA and rstB are involved in the regulation of adhesion, consistent with RNA-Seq data. These findings supported the hypothesis that rstA and rstB sensitivity to environmental stresses might constitute a mechanism by which environmental conditions affect adhesion. Meanwhile, decreased motility was observed in rstAand rstB-RNAi strains, indicating that rstA and rstB might influence adhesion by controlling motility. This is consistent with previous findings that RstB-RstA system is necessary for motility in Salmonella (Tran et al., 2016).
Environmental factors can markedly influence the bacterial adhesion capacity. Indeed, pH, as an important environmental factor, significantly affects bacterial adhesion (Balebona et al., 1995;Yan et al., 2007). Heavy metals occurring in the environment also affect microorganisms (Haferburg & Kothe, 2007;Xiao, Zong, & Lu, 2015). Kong et al. (2015) reported that Cu 2+ , Pb 2+ , and Hg 2+ significantly reduce V. alginolyticus adhesion to the skin mucus of large yellow croakers. However, the mechanisms underlying the effects of environmental factors on bacterial adhesion remain unclear. In this study, the effects of temperature, pH, and starvation on rstA and rstB expression levels were assessed.
The adhesion capability V. alginolyticus at various temperatures showed an inverted U-shaped trend (Huang, Hu, et al., 2016;. V. alginolyticus showed remarkably stronger adhesion at 28°C compared with the other temperatures, which is  Urtaza, 2010;Reilly, Reilly, Smith, & Baker-Austin, 2011;Sterk, Schets, de Roda Husman, de Nijs, & Schijven, 2015). The trends of rstA and rstB gene expression levels and in vitro adhesion at various temperatures were similar; indicating that V. alginolyticus attachment was affected by temperature, with rstA and rstB likely contributing to adhesion control at different temperatures.
V. alginolyticus adhesion under diverse pH values also displayed an inverted U-shaped trend, peaking at pH 7.0 (Huang, Hu, et al., 2016;. The trends of rstA and rstB gene expression levels and in vitro adhesion under diverse pH levels were similar, suggesting that pH affects V. alginolyticus adhesion, with rstA and rstB likely involved in adhesion control at different pH levels. Huang, Hu, et al. (2016) and   Biofilm production is a way by which bacteria protect themselves from the host immune system (She et al., 2016). However, whether rstA and rstB take part in biofilm formation remains undefined. The above results revealed that rstA and rstB silencing remarkably decreased biofilm formation. These findings indicated that rstA and rstB play key roles in V. alginolyticus biofilm formation.
Hemolysin is a chief virulence factor of many Vibrio (Syed et al., 2009). Terceti et al. (2017) found that rstB controls the production of damselysin, phobalysin C, and phobalysin P in Photobacterium damselae subsp. damselae. However, whether rstA and rstB regulate the hemolytic activity of V. alginolyticus remains unclear. In the present study, rstA and rstB silencing, respectively, in V. alginolyticus resulted in remarkably decreased hemolytic capability.
Several studies demonstrated that rstA and rstB are closely as- subsp. damselae virulence for fish was also reported (Terceti et al., 2017). In the present study, the effects of rstA and rstB on V. alginolyticus virulence were also demonstrated. Monitoring fish postchallenge revealed that mortality was remarkably lower in groups treated with rstA-and rstB-RNAi strains compared with the control group.
Meanwhile, death was delayed in animals injected with rstA-RNAi and rstB-RNAi strains compared with controls. These findings revealed that rstA and rstB knockdown has remarkable effects on V. alginolyticus virulence.
In conclusion, these findings suggest that: (1) rstA and rstB are tightly associated with V. alginolyticus adhesion; (2) rstA and rstB contribute to motility, biofilm production, hemolysis, and virulence in V. alginolyticus; (3) rstA and rstB modulate adhesion in response to environmental changes of temperature, pH, and starvation.