Vibrio alginolyticus HN08155 and its derivative strains were cultured at 30°C and 180 rpm in 2216E medium consisting of 5 g/L tryptone, 1 g/L yeast extract, and 0.01 g/L FePO₄. Escherichia coli β2163 was grown at 37°C and 180 rpm in standard Luria-Bertani (LB) medium containing 0.3 mM diaminopimelic acid (DAP) (Luo et al., 2015). Antibiotics were used at the following concentrations: 100 μg/mL ampicillin (Amp) was used for the growth of Vibrio spp., 10 μg/mL chloramphenicol (Cm) was additionally added to complemented strains for integrated plasmids. All reagents were purchased from Solarbio (Beijing, China).

All strains and plasmids used are described in Supplementary Table 1. The (p)ppGpp synthetase gene mutants of V. alginolyticus HN08155 were generated by allelic exchange (Milton et al., 1996). This method has been described in detail in our previous work (Yin et al., 2021), and the oligonucleotides used in these procedures are listed in Supplementary Table 2. In brief, the sucrose-sensitive suicide plasmid pDM4-relA containing the fusion product of a 542 bp upstream fragment and a 542 bp downstream fragment of relA coding sequence was introduced into E. coli β2163 cells by heat shock. Subsequently, the recombinant plasmid pDM4-relA was integrated into wild type through conjugation, and the relA internal coding sequence was deleted after homologous recombination occurred on both homologous arms. Single crossover mutants with plasmids integrated into specific chromosomal loci were obtained on 2216E agar plates supplemented with Amp and Cm. The second crossover mutants were screened on 2216E agar plates supplemented with Amp and 10% sucrose. The in-frame deletion mutant ΔrelA was confirmed by PCR and DNA sequencing. For generation of the double deletion strain ΔrelAΔspoT, the suicide plasmid pDM4-spoT containing the fusion product of a 563 bp upstream fragment and a 500 bp downstream fragment of spoT coding sequence was transferred into the single deletion mutant ΔrelA, and the spoT internal coding sequence was deleted after double cross-over recombination. High cellular (p)ppGpp level in strains would inhibit cell growth, due to the lack of hydrolysis activity mediated by SpoT to degrade (p)ppGpp synthesized by RelA, so we failed to construct a spoT single mutant (Xiao et al., 1991). In order to construct the complemented plasmids, the open reading frame of relA and the linear segment pACYC184 (without tetracycline resistance but with chloramphenicol resistance gene) were amplified, respectively. The open reading frame of relA was cloned into the linear segment of pACYC184 by using ClonExpress II One Step Cloning Kit (Vazyme, Nanjing, China). The opening reading frame of tetracycline resistance gene was replaced by that of relA gene but with its native promoter to produce the pACYC184-relA plasmids expressing relA gene. And then, the resulting pACYC184-relA plasmid was transformed into the ΔrelA mutant to generate the complemented strain with Amp^R and Cm^R, designated ΔrelA-pRelA. Similarly, the pACYC184-spoT plasmid was transformed into the ΔrelAΔspoT mutant to generate the complemented strain with Amp^R and Cm^R, designated ΔrelAΔspoT-pSpoT.

The biofilm assay was performed following a previously described protocol with modification (O’Toole and Kolter, 1998). After overnight incubation, the cultures were adjusted to an OD₆₀₀ of 0.6, diluted 1:100 into 50 mL fresh LB medium without antibiotics. Each 200 μL sample was taken from the diluted medium into the 96-well polystyrene microplate and cultivated statically at 30°C. After culturing for 2, 4, 6, 8, 10, 12, 14, 18, 24, and 36 h respectively, bacterial cells were removed, rinsed once with phosphate buffer saline (PBS) and 220 μL of 0.1% crystal violet was added to stain biofilm for 30 min. The residual stains were then removed, and the stained biofilms were washed twice with PBS. Finally, the amount of biofilm was determined at 570 nm using a microplate reader (Epoch2, BioTeK) after dissolving crystal violet with 200 μL of 95% ethanol for 1 h. Three separate experiments were performed with biological triplicates each. To exclude growth effects on biofilm formation, the cell density was measured at 600 nm, and biofilm formation was normalized by dividing total biofilm by the cell density.

In spotty pellicle assay, 5 mL above diluted medium was added in each glass tube and incubated without shaking for 16 h at 30°C. For colonial morphology assay, the cultures were adjusted to an OD₆₀₀ of 0.6, and collected by centrifuging at 12,000 rpm for 2 min. After removing the supernatant, bacteria are resuspended in 10 μL LB medium. Subsequently, 10 μL mixed medium is spotted on a trypticase soy sheep blood agar plate (Huankai Microbial, Guangdong, China) and cultivated statically at 30°C for 24 h, after which the colonial morphology was observed and photographed.

The modified phenol-sulfuric acid method was used to determine the polysaccharides production (Liu et al., 2021). In the beginning, 8 mg glucose was dissolved into the 100 mL volumetric flask, and then 0, 0.2, 0.4, 0.6, 0.8, and 1 mL glucose solution were added to each centrifuge tube, respectively. Double-distilled water was added to the centrifuge tube containing less than 1 mL glucose solution. Subsequently, 1 mL phenol solution and 5 mL concentrated sulfuric acid were added into each tube, after 30 min of reaction, the absorbance at 490 nm was respectively measured to draw the standard glucose curve. Subsequently, 1 mL phenol solution and 5 mL concentrated sulfuric acid were added to the tube, after 30 min of reaction, the absorbance at 490 nm was respectively measured to draw the standard glucose curve. Each overnight culture was adjusted to an OD₆₀₀ of 0.6, diluted 1:100 into 50 mL LB medium and 6 mL broth was incubated to the 6-well polystyrene microplate statically at room temperature for 6 and 24 h, respectively. 5 mL bacterial culture was collected and centrifuged at 5,000 rpm and 4°C for 5 min, and 1 mL supernatant was collected and transferred into a new tube. Afterward, the supernatant, phenol solution (9%, v/v) and concentrated sulfuric acid were added to the tube sequentially in the ratio of 1:1:5 (v:v:v). After 30 min of reaction, the absorbance at 490 nm was measured, and the amount of polysaccharides was further calculated according to the above standard glucose curve.

Vibrio containing Amp antibiotic (OD₆₀₀ = 0.6) were fixed in 2.5% glutaric dialdehyde, and then bacterial flagella were observed by using the transmission electron microscope (TEM, JEM-2100, Japan). To evaluate swarming ability, fresh cultures of V. alginolyticus strains containing antibiotics (OD₆₀₀ = 0.6) were centrifuged, precipitated, and resuspended in 10 μL 2216E medium, and then 1 μL culture was spotted on 0.9% (w/v) 2216E agar plates at room temperature.

The WT and mutants were cultured in LB medium containing Amp overnight, and then diluted 1:100 into fresh medium and grown to the exponential phase (OD₆₀₀ = 0.6) and stationary phase (OD₆₀₀ = 1.2). RNA was extracted with a Bacteria Total RNA Extraction Kit (Promega, Madison, WI, United States), and then HiScript II Q RT SuperMix (+gDNA wiper) (Vazyme, Nanjing, China) was used for reverse transcription after determining the RNA quality by Nanophotometer NP80 (IMPLEN, München, Germany). The final cDNA samples were analyzed by Quantitative Reverse Transcription PCR (qRT-PCR) using ChamQ Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China). Primers used in the qRT-PCR analysis are listed in Table 1. Expression of the genes encoding helix-turn-helix transcriptional regulator VpsT, polysaccharides biosynthesis protein VpsM and VpsH, chemotaxis protein CheR, flagellar hook protein FlgE, flagellar basal body L-ring protein FliD, and alternative sigma transcription factor RpoN was determined in triplicate. The 16S rRNA gene was used as an endogenous control for normalization of expression values. The 2^–ΔΔCt method was used to quantify and compare each gene expression (Livak and Schmittgen, 2001).

Three separate experiments were performed with biological triplicates each, and results were presented as mean ± standard deviation (SD; n = 3). Analysis of statistical differences was conducted on GraphPad Prism version 7.01 for Windows (GraphPad Software, SanDiego, CA, United States) through one-way analysis of variance (ANOVA) with Dunnett’s post-test. Differences were considered significant by calculated p value (*p ≤ 0.05, ^**p ≤ 0.01), and p value ≥ 0.05 was regarded as no significant.

The amount of biofilm formation of WT, ΔrelA, ΔrelAΔspoT, ΔrelA-pRelA, and ΔrelAΔspoT-pSpoT was evaluated by staining with crystal violet. Afterward, biofilm formation was normalized by dividing total biofilm by the cell density (Supplementary Figure 1) to exclude growth effects (Figure 1A). Overall, the trend of biofilm formation in WT was to continuously increase at low cell density and then reduce at high cell density. The increasing amount of biofilm formation reached a peak at 6 h (OD₅₇₀/OD₆₀₀ = 3.61), and kept a high level for a period of time until 10 h. Interestingly, the second small peak of biofilm formation appeared at 12 h (OD₅₇₀/OD₆₀₀ = 1.04) and decreased again to a low level. The amount of biofilm formation in ΔrelA was significantly lower than that of the WT at low cell density (Figure 1B, p ≤ 0.01). However, it maintained a high level (OD₅₇₀/OD₆₀₀ = 2.61–3.36) during 8–14 h and then began to decompose slowly until close to the level of WT (Figure 1C). In terms of the biofilm formation in ΔrelAΔspoT, its absorbance value was significantly lower than that of WT at low cell density (Figure 1B, p ≤ 0.05). Surprisingly, the amount of biofilm formation reached its peak at 8 h (OD₅₇₀/OD₆₀₀ = 4.66), but the biofilm did not decompose within 36 h, even maintained at a high level (Figure 1C). The biofilm formation of ΔrelA-pRelA and ΔrelAΔspoT-pSpoT were similar to WT and ΔrelA, respectively. An obvious crystal violet-stained circular ring appeared at the air-liquid interface (data not shown), suggesting that bacteria might tend to accumulate in oxygen-rich areas. However, there was additional biofilm staining at the bottom of the microtiter plates in ΔrelAΔspoT at high cell density (8–36 h), exhibiting robust bacterial self-aggregation ability. These results revealed that the deletion of two (p)ppGpp synthetase genes led to postponed biofilm formation at low cell density, and the absence of relA and spoT resulted in delayed and failed biofilm disassembly, respectively.

It has been reported that the formation of spotty pellicle strongly depends on oxygen level (Hare et al., 1981). After static incubation at 30°C for 16 h in the glass tube, the liquid solutions of WT, ΔrelA, and ΔrelAΔspoT-pSpoT (OD₆₀₀ = 0.8–1.0) were more turbid than that of ΔrelAΔspoT (OD₆₀₀ = 0.6–0.8), without apparent bacterial flocculation and transparent air-liquid surface layer (Figure 2A). However, a spotty, unattached floating and thicker pellicle of ΔrelAΔspoT was observed at the air-liquid surface layer (Figure 2A). This result implied that spoT might participate in sensing oxygen and regulation of spotty pellicle.

A surprising but expected phenomenon was discovered in the hemolysis test when we determined whether the hemolytic ability of V. alginolyticus HN08155 could be affected by the absence of (p)ppGpp synthetase genes. In the beginning, there was no obvious difference in colony morphology among all strains (data not shown), whereas the colony morphology of ΔrelAΔspoT began to wrinkle gradually during cultivation. As shown in Figure 2B, WT and ΔrelA grew into gray-white, smooth/translucent, and neat-edged colonies after 24 h of cultivation on trypticase soy sheep blood agar plate. In contrast, the colony morphology of ΔrelAΔspoT showed lackluster and surface rugose, and ΔrelAΔspoT-pSpoT exhibited slightly wrinkled colony morphology. This finding indicated that spoT rather than relA played a variable role in controlling colony phase variation of V. alginolyticus: the lack of spoT resulted in the transformation of gray-white and smooth/translucent colonies into surface wrinkled colonies, while the presence of spoT inhibited the formation of opaque/rugose colonies.

Exopolysaccharides are an indispensable component of biofilm formation in V. alginolyticus (Chen et al., 2009). Based on our previous experiments, 6 and 24 h were chosen as the representatives of the low cell density and high cell density, respectively (Zhang et al., 2021). Our results indicated that the EPSs content of WT was the highest at low cell density (Figure 3A), but it gradually decreased with the disruption of (p)ppGpp synthetase genes, which was almost 3- and 5-fold of that in ΔrelA and ΔrelAΔspoT, respectively (Figure 3B). Complementation of the relA or spoT gene into mutant strains ΔrelA-pRelA and ΔrelAΔspoT-pSpoT respectively increased the synthesis of EPSs, although these contents were relatively lower in comparison with that in WT (p ≤ 0.01). At high cell density, WT had the lowest EPSs content, followed by ΔrelA, while ΔrelAΔspoT had the highest EPSs content. Notably, the content of EPSs in ΔrelAΔspoT was almost 1.5 times that of WT and ΔrelA. The EPSs content in ΔrelA was almost the same as that in WT (p ≥ 0.05), and supplementation of the spoT gene in ΔrelAΔspoT could reduce the secretion of EPSs to some extent. Overall, both relA and spoT favored the synthesis of EPSs at low cell density, and only spoT helped the degradation of it at high cell density.

Transmission electron microscope observation showed that V. alginolyticus HN08155 was short and rod-shaped, with a single curved polar flagellum that was 2–3 times longer than that of the bacterium (Figure 4A). The flagellum of ΔrelA and ΔrelAΔspoT-pSpoT was identical to WT (Figures 4B,D), but the lack of spoT resulted in the loss of flagellum (Figure 4C). There was no doubt that spoT played a critical role in synthesizing flagellum. Previous study revealed that the swimming ability of ΔrelAΔspoT decreased sharply (Yin et al., 2021). Here we further investigated whether the absence of spoT would affect its swarming ability. After 12 h of continuous culture on 0.9% 2216E agar plates at room temperature, the swarming halo diameter of the WT, ΔrelA, and ΔrelAΔspoT-pSpoT averaged 6.80 ± 0.3, 7.65 ± 0.5, and 5.75 ± 0.4 mm, respectively, while those of ΔrelAΔspoT averaged 2.10 ± 0.8 mm only (Figures 5A,B). The diameters of the swarming rings of ΔrelAΔspoT were significantly smaller than that of the WT, ΔrelA and complemented strains (Figure 5B, p ≤ 0.01), demonstrating that the swarming ability of ΔrelAΔspoT was sharply diminished. Taken together, the disruption of spoT in V. alginolyticus not only negatively affected the production of flagellum, but also significantly impaired its swarming ability.

Although the content of extracellular polysaccharides can be determined by the phenol-sulfuric acid method by which the product concentration is qualified by observing the absorbance value (Liu et al., 2021). However, we could not rule out the interference of the residual nucleic acid and other sugars in the samples. Considering that vps-related genes affect the synthesis and transport of bacterial polysaccharides (Casper-Lindley and Yildiz, 2004; Chang et al., 2010), and cheR, flgH, and fliD are closely related to bacterial flagella synthesis and motility (Echazarreta and Klose, 2019; Liu et al., 2020), the expression levels of some functional genes were further determined by qRT-PCR. As shown in Figures 6A–C, at low cell density, the expression level of vpsH in WT was 1.92- and 6.14-fold higher than that of ΔrelA and ΔrelAΔspoT, respectively. Conversely, at high cell density, the transcription of vpsH in ΔrelAΔspoT was 5- and 3.5-fold higher than that of WT and ΔrelA, respectively. Similarly, the expression levels of the other EPSs-related genes, vpsT and vpsM, in WT and ΔrelA showed a trend of first increasing and then decreasing, which was similar to the trend of biofilm formation and EPSs production in the culture medium under the same conditions. As for motility-related genes, the expression level of cheR and flgE in ΔrelA was significantly lower than that of WT, whereas the expression levels of flgH in ΔrelA were not significantly different from that of WT (Figure 6D). The transcriptional profiles of these motility-related genes in ΔrelAΔspoT exhibited lower levels compared with WT, which was consistent with the results of the swarming ability test. These results will further help us understand the effects of (p)ppGpp synthetase gene-mediated alteration in EPSs and motility on biofilms.

Previous study has shown that the deletion of rpoN leads to failure of biofilm detachment, loss of flagellum, p roduction of spotty pellicle, transparent surface rugose colonies, and decreased swimming motility (Zhang et al., 2021). To explore the relationship between (p)ppGpp synthetase genes and rpoN, we measured the expression level of rpoN at low and high cell densities. At low cell density, the expression levels of rpoN in ΔrelA was significantly higher than that in WT, but that in ΔrelAΔspoT was significantly reduced by almost 3-fold compared with WT (Figure 7). At high cell density, the transcription levels of rpoN in ΔrelA and ΔrelAΔspoT was lower than that of WT even if the difference was not significant. It was likely that (p)ppGpp synthetase genes had no obvious effect on rpoN at high cell density. These results implied that the deletion of (p)ppGpp synthetase genes had negative effects on the expression of rpoN at low cell density.