The two-component system expression patterns and immune regulatory mechanism of Vibrio parahaemolyticus with different genotypes at the early stage of infection in THP-1 cells

ABSTRACT Vibrio parahaemolyticus must endure various challenging circumstances while being swallowed by phagocytes of the innate immune system. Moreover, bacteria should recognize and react to environmental signals quickly in host cells. Two-component system (TCS) is an important way for bacteria to perceive external environmental signals and transmit them to the interior to trigger the associated regulatory mechanism. However, the regulatory function of V. parahaemolyticus TCS in innate immune cells is unclear. Here, the expression patterns of TCS in V. parahaemolyticus-infected THP-1 cell-derived macrophages at the early stage were studied for the first time. Based on protein-protein interaction network analysis, we mined and analyzed seven critical TCS genes with excellent research value in the V. parahaemolyticus regulating macrophages, as shown below. VP1503, VP1502, VPA0021, and VPA0182 could regulate the ATP-binding-cassette (ABC) transport system. VP1735, uvrY, and peuR might interact with thermostable hemolysin proteins, DNA cleavage-related proteins, and TonB-dependent siderophore enterobactin receptor, respectively, which may assist V. parahaemolyticus in infected macrophages. Subsequently, the potential immune escape pathways of V. parahaemolyticus regulating macrophages were explored by RNA-seq. The results showed that V. parahaemolyticus might infect macrophages by controlling apoptosis, actin cytoskeleton, and cytokines. In addition, we found that the TCS (peuS/R) could enhance the toxicity of V. parahaemolyticus to macrophages and might contribute to the activation of macrophage apoptosis. IMPORTANCE This study could offer crucial new insights into the pathogenicity of V. parahaemolyticus without tdh and trh genes. In addition, we also provided a novel direction of inquiry into the pathogenic mechanism of V. parahaemolyticus and suggested several TCS key genes that may assist V. parahaemolyticus in innate immune regulation and interaction.

V ibrio parahaemolyticus is a foodborne pathogen with a multi-host range that can infect marine animals and humans (1). The most widely studied virulence factors of V. parahaemolyticus have thermostable hemolysin (tdh/trh) and Type III secretory system (T3SS1 and T3SS2) (2)(3)(4). It is generally agreed upon that possessing more virulence factors will have stronger pathogenicity. However, among the 42 samples of V. parahaemolyticus with different genotypes that we isolated from fecal samples of patients treated in Shanghai hospital, most of the pathogens carried just one thermosta ble hemolysin (tdh or trh), a small number of pathogens had both tdh and trh, and there is also a pathogen without tdh and trh (5). Furthermore, VopZ, a T3SS2 effector protein, was discovered to be important for intestinal colonization and diarrheagenic, which is essential for the pathogenicity of V. parahaemolyticus (6). However, we found that it was not detectable in some strains (7). Zhang et al. found that VopC was critical for V. parahaemolyticus T3SS2-mediated invasion (8), but another research reveals that VopC is not necessary for pathogenicity in an animal infection model (9). And several strains of 42 samples were undetectable VopC (5). Since we cannot ignore the diversity of pathogenic V. parahaemolyticus in clinical practice, we chose V. parahaemolyticus of different genotypes as research objects to investigate new pathogenic factors.
Generally, V. parahaemolyticus infects humans through the skin, gastrointestinal wounds, and diet. It must overcome several challenges to cause illness, and the first hurdle is the innate immune system (10). When the immune phagocytes ingest the bacteria, they will be under various extreme conditions, including an acidic pH, hypoxia, reactive oxygen species, reactive nitrogen species, and nutritional deprivation (11). Therefore, recognizing and reacting quickly to external environmental signals is crucial for bacterial survival in the cell. Two-component system (TCS) is a critical process for bacteria to sense environmental signals and transmit them to the interiors to activate regulatory mechanisms (12). It is known that TCS is composed of a transmembrane sensor protein known as histidine kinase (HK) on the bacterial membrane and a response regulator (RR) in the cytoplasm (12). According to the P2CS (http://www.p2cs.org) prokaryote database, V. parahaemolyticus had 50 HKs and 55 RRs. The HK transfers its phosphate group to RR, and RR changes downstream target gene expression to control intercellular communication, environmental adaptation, growth, adhesion, and other processes. Moreover, TCS can also influence bacterial pathogenicity by regulating itself or interfering with the host immune system (13,14). For example, Salmonella typhimu rium recognizes the host by a surface-located HK PhoP sensor and phosphorylates the intracellular effector PhoQ, thus promoting the bacterial expression of lipid A-modifying enzymes. Modification of lipid A attenuates toll-like receptor 4 (TLR4)-mediated nuclear factor kappa-B (NF-κB), thereby inhibiting the generation of inflammatory cytokines and preventing the immune system from responding normally (15).
Studies have shown that V. parahaemolyticus effector proteins, VopQ and VopS, can inhibit the activation of the nucleotide-binding oligomerization domain (NOD), leucine-rich repeat and caspase recruitment domain-containing 4 (NLRC4) by inducing autophagy and cell division cycle 42 (Cdc42) inactivation in macrophages, respectively (2). And VopQ can also prevent phagocytosis (16). However, it is fully unclear how V. parahaemolyticus interacts with human innate immune cells. The RNA-seq has greatly helped researchers fully understand pathogenic mechanisms and key pathogenic factors of pathogens. Nevertheless, persistent infection worsens intracellular pathogen status, particularly non-parasitic pathogens, which is difficult for RNA-seq technology with strict requirements for sample RNA. Moreover, the content of pathogens in the cell is too small, resulting in a low proportion of pathogens in the total RNA extracted, which also increases difficulty to fully comprehend the pathogenic mechanisms of bacteria in the interaction with host immune cells (17). Therefore, in this study, we chose the early infection stage of immune cells and utilized quantitative real-time PCR (qPCR) to examine the transcriptional expression levels of TCS in V. parahaemolyticus. qPCR is a reliable and rapid method to assess the level of target gene expression, which is particularly sensitive to the detection of some low-copy mRNAs (18). This method aids in understanding the changes of TCS in V. parahaemolyticus-infected immune cells. In addition, we chose THP-1 cell-derived macrophages as an innate immune cell model because numerous studies have demonstrated that THP-1 cells are a suitable model for early infection (19).
This study aimed to identify the expression patterns of TCS in V. parahaemolyticus with various genotypes under macrophage stress and to explore the potential critical genes and mechanisms of V. parahaemolyticus regulating host innate immunity. First, we determined that the optimal time for V. parahaemolyticus to infect cells was 3.5 h (infection for 1.5 h and gentamicin treatment for 2 h) based on intracellular survival bacterial counts. And the TCS expression profiles of V. parahaemolyticus were detected by quantitative real-time PCR (qRT-PCR). On this basis, differently expressed TCS genes (DETGs) and their interacted proteins were analyzed by protein-protein interaction (PPI) network analysis. Subsequently, the RNA-seq was used to comprehensively understand the changes at the transcription level of macrophages at 3.5 h after V. parahaemolyticus infection. Furthermore, we explored the potential role of TCS in regulating macrophages by the knockout of critical TCS genes.
In this study, the interaction of V. parahaemolyticus TCS in regulating innate immune cells was studied for the first time. Moreover, the object of the study was comprehensive, including V. parahaemolyticus of various genotypes. This study is beneficial to a thorough understanding of the potential function of TCS during host cell infection and provides new directions for studying V. parahaemolyticus in innate immune regulation.

Intracellular survival of V. parahaemolyticus
To evaluate the number of V. parahaemolyticus in macrophages, V. parahaemolyticus with four genotypes [VPC17 (tdh+/trh−), VPC44 (tdh−/trh−), VPC49 (tdh+/trh+), and VPC85 (tdh−/trh+)] was used to infect cells in the antibiotic-free medium at 0-3 h. The results showed that the number of V. parahaemolyticus increased within 1.5 h but diminished after more than 2 h (Fig. 1A). During infection, not all bacteria were in direct touch with cells, and extracellular V. parahaemolyticus proliferated quickly in the antibiotic-free environment. The drop in intracellular bacteria may be due to a considerable increase in external bacteria, intensifying the attack on cells and continuing cell lysis.

Research Article mSystems
We examined that four strains could not grow in the medium containing 100 ng/mL gentamicin. Then, the cells were infected with four strains for 1.5 h; the extracellular bacteria were removed and treated with the gentamicin-containing medium for 0-6 h. The results showed that the cell's bacterial population peaked at 3 h and decreased between 4 h and 5 h (Fig. 1B). This may be due to the proliferation of intracellular bacteria leading to cell lysis overflow (8). The number of VPC85 was much lower at 4-6 h than the other three strains, showing that the intracellular bacteria proliferated rapidly and aggravated macrophage lysis. Based on the above research results, we determined the V. parahaemolyticus infection time to be 1.5 h and the gentamicin treatment time to be 2 h.

Expression patterns of TCS in V. parahaemolyticus-infected macrophages
The expression patterns of TCS in V. parahaemolyticus with four genotypes were shown by cluster heatmap (Fig. 1C). The red square in the heatmap indicates high expression, and the blue one indicates low expression. After infection with macrophages, 63, 40, 43, and 32 TCS genes could not be detected in VPC17, VPC44, VPC49, and VPC85, respectively. These V. parahaemolyticus TCS genes did not express or had extremely low expression abundance, suggesting that most TCS genes had changed considerably under macrophage stress. The expression patterns of TCS were separated into four categories; groups A and B were similar. Moreover, VPC85 in group C differed most from the other three strains. The common genes among the four strains were located in groups A and B, indicating that this part of TCS genes may contain essential genes that function as regulators when V. parahaemolyticus infects macrophages.

TCS in V. parahaemolyticus before and after infected macrophages
We compared expressed levels of V. parahaemolyticus TCS genes between uninfected and infected ( Fig. 2A). The red part in the map indicates up-regulated genes, and the blue one indicates down-regulated genes. In VPC17, VPC44, VPC49, and VPC85, the number of genes with fold change >1 was 4, 20, 8, and 15, respectively (P < 0.05), and the number of genes with fold change <1 was 100, 83, 91, and 85, respectively (P < 0.05). We selected fold change >1.5, <0.5, and P < 0.05 as DETGs, of which 1, 1, 3, and 8 genes were significantly up-regulated, and 84, 77, 60, and 66 genes were significantly down-regulated (Fig. 2B). We found that most TCS genes of strains were down-regulated, indicating that TCS was inhibited under macrophage stress. Notably, a few genes were still up-regulated in all four strains, such as VP_ RS15995 (VPA0148). According to the Venn diagram, there were 39 identical DETGs in the four strains (P < 0.05) (Fig. 2C). And 21 of the 39 DETGs had similar expression patterns combined with the expression pattern of TCS after infection. We selected these 21 DETGs as potential regulatory genes for the study of V. parahaemolyticus regulating macrophages; the information and Kyoto Encyclopedia of Genes and Genomes (KEGG) description of these 21 DETGs were shown in Table 1.
VPA0021 can also interact with VPA0018-encoded outer membrane lipoprotein (Fig.  3D). VPA0018 protein contains a Lol A domain that moves lipoprotein from the inner membrane to the outer membrane. Then, the lipoprotein is received and located inside the cell's outer membrane through the membrane receptor protein Lol B. This process plays a vital role in assembling bacterial outer membrane structure and regulating bacterial life activities (21). Research Article mSystems Figure 3E shows that the VP1735 protein might interact with the thermostable hemolysin protein encoded by VP1729. Thermostable hemolysin protein has hemolytic activity, enterotoxin activity, cardiotoxicity, and cytotoxicity (22)(23)(24)(25). In addition, VP1735 may also interact with the iron-containing redox enzyme family protein (VP1731encoded), which may play a role in iron uptake.
The uvrC is a noteworthy protein in the proteins interacting with uvrY (Fig. 3F). The uvrC contains the GIY-YIG domain, which aids in repairing DNA damage and preserving genomic stability (26). In addition, uvrY can also interact with csrA (Fig. 3F). The posttranscriptional regulator csrA plays a central role in adapting pathogens to infect animal hosts (27,28).
VPA0148 might interact with the protein encoded by VPA0156 that contains the tetratricopeptide repeat (TPR) domain. Proteins containing TPR can mediate PPIs, assemble multiprotein complexes, and participate in several biological activities,  (29). In addition, it is currently known that VPA0148 also interacts with TonB-dependent siderophore enterobactin receptor peuA (Fig. 3G) (30). Iron carrier is one of the ways for microorganisms to obtain iron, which gives bacteria more robust adaptability and competitiveness (31). The above analysis shows that these seven TCS genes have the potential in studying V. parahaemolyticus regulating macrophages.

RNA-seq of THP-1 cell-derived macrophages after infection
To fully understand the regulation of V. parahaemolyticus regulating macrophages at the early stage of infection, RNA-seq was performed on host cells infected with VPC17, VPC44, VPC49, and VPC85 at 3.5 h. Principal component analysis (PCA) analysis of the four groups was shown in Fig. 4A. The sequencing data of VPC17_THP1, VPC44_THP1, VPC49_THP1, and VPC85_THP1 were compared with the reference genome by 96.63%, 96.74%, 96.97%, and 96.93%, respectively. We selected eight genes from each group of transcription samples, a total of 32 genes for qPCR verification. The correlation between qPCR and RNA-seq showed that the expression trends of the two groups were consistent (R 2 >0.8), and the transcriptome data were credible (Fig. 4B). The sequencing data of uninfected THP-1 cells were used as the control group for downstream analy sis. The cluster heatmap of significant genes (P adj <0.05) revealed that VPC17_THP1, VPC44_THP1, VPC49_THP1, and VPC85_THP1 had similar expression patterns (Fig. 4C). The screening conditions for differentially expressed genes (DEGs) were |log 2 FC| ≥1 and P adj < 0.05; the volcanic map showed that there were 2,410, 1,715, 2,369, and 1,576 DEGs in VPC17_THP1, VPC44_THP1, VPC49_THP1, and VPC85_THP1, respectively (Fig. 4D).
Gene ontology (GO) enrichment analysis of DEGs showed that the four groups had similar enrichment results in biological processes, cell components, and molecular functions (false discovery rate [FDR] <0.05) (Fig. 5A). The same biological processes were regulation of transcription by RNA polymerase II, inflammatory response, positive regulation of I-κB kinase/NF-κB signaling, cellular response to lipopolysaccharide, etc. The same cell components were chromatin, nucleoplasm, cytoplasm, and the same molecular functions were DNA-binding transcription factor activity, protein binding, and sequence-specific double-stranded DNA binding. The macrophages exhibited similar immunological responses even when infected with V. parahaemolyticus of different Research Article mSystems genotypes, suggesting that the four strains may share a similar regulatory mechanism. This might be one of the reasons why they were clinically pathogenic, even if they carry different virulence genes. The results indicated that the shared metabolic pathways or DEGs might be the key to V. parahaemolyticus regulating macrophages. Therefore, we used gene set enrichment analysis (GSEA) to evaluate the complete transcriptome of four groups. The transcriptome data of four groups were summarized as the treatment group, named VPC_THP1, and the uninfected macrophages' transcriptome data were used as the control group. The GSEA results revealed that among the 178 gene sets of VPC_THP1, 87 gene sets were up-regulated, and 14 gene sets were considerably enriched (FDR <25%). The pathways were listed according to FDR (Fig. 5B), including pathogenic Escherichia coli infection, NOD-like receptor signaling pathway, cytokinecytokine receptor interaction, T cell receptor signaling pathway, apoptosis, and other pathways. These pathways pertain to immunological reaction, apoptosis, immune evasion, etc. We pay more attention to the pathways and particular regulated genes associated with immune evasion.

Research Article mSystems
In the pathogenic Escherichia coli infection (Fig. 5C) and T cell receptor signaling pathway (Fig. 5D), we found that Cdc42 was up-regulated in all four groups. Cdc42 can affect alterations in cell structure by regulating the actin cytoskeleton that facilitates pathogen invasion or impedes cell function (8). This result indicates that V. parahaemoly ticus may modify the cytoskeleton actin of macrophages during infection.
In the Jak-STAT signaling pathway (Fig. 5E), interleukin 10 (IL-10) was up-regulated in VPC17_THP1, VPC44_THP1, VPC49_THP1, and VPC85_THP1, respectively (P < 0.05). The highly expressed IL-10 can inhibit the inflammatory response and dendritic macrophage activity, lowering the body's capacity to fight infections and generate immune responses (32). V. parahaemolyticus might inhibit the inflammatory response and macrophage activity by stimulating the expression of the inflammatory factor IL-10.
Moreover, all four strains induced apoptosis pathway activation of macrophage (Fig.  5F). We summarized the DEG expression patterns of four groups in the apoptotic pathway (|log 2 FC| ≥ 1 and P adj < 0.05) (Fig. 5G). The red square represents up-regula ted genes, and the green square represents down-regulated genes. The three classical pathways mediating apoptosis are the mitochondrial, endoplasmic reticulum, and death receptor pathways (33). Although tumor necrosis factor-alpha (TNF-α) was triggered in the death receptor and endoplasmic reticulum pathways, caspase-8 (CASP8) was not activated (Fig. 5G). Meanwhile, caspase-10 (CASP10) was inhibited, and endoplas mic reticulum stress-mediated caspase-12 (CASP12) was not activated (Fig. 5G). In the mitochondrial pathway, the expression levels of Cyt-C and caspase-3 (CASP3) were noticeably up-regulated in four groups (Fig. 5G). Cyt-C can be used as a carrier to transfer electrons in the mitochondrial respiratory chain and establish a transmembrane potential. When it is released into the cytoplasm, it can bind to apoptotic protease activating factor-1 to form an apoptosome and then activate CASP3 to initiate a series of apoptotic reactions (34). The findings implied that V. parahaemolyticus might activate the macrophage apoptosis pathway through Cyt-C-mediated mitochondrial pathway.

The effect of TCS peuS/R in V. parahaemolyticus
To preliminarily explore the potential role of TCS in V. parahaemolyticus regulating macrophages, we selected VPA0148 to construct a mutant strain to infect macrophages. VPA0148 was up-regulated in all four strains, and Tanabe et al. named VPA0149/ VPA0148 TCS as peuS/R (30). We selected the model strain ATCC17802 as representative wild-type (WT) to construct peuS/R mutant and complement strains. Firstly, we used WT to infect macrophages for 1.5 h and treated with gentamicin for 2 h. The gene expression trends of peuS and peuR verified by qPCR were consistent with the four strains in this study (Fig. 6A). The expression levels of IL-10 and CASP3 in macrophages were up-regulated by 5.78-and 7.37-folds, respectively (P < 0.05). Subsequently, the mutants of peuR and peuS were successfully constructed by homologous recombination (Fig. 6B), named △peuR and △peuS. The growth curve of the △peuR and △peuS showed that the deletion of peuR and peuS did not affect the growth of the strain (Fig. 6C), indicating that peuS/R was not involved in the growth of V. parahaemolyticus. Based on the mutant, the complementary strains were successfully constructed and named △peuR: C and △peuS: C, respectively.
The results of the CCK8 assay showed that the survival rate of macrophages infected with △peuR was higher than that of WT (Fig. 6D), indicating that peuR had an inhibitory effect on macrophage growth. The number of intracellular bacteria demonstrated that peuS/R did not impact bacterial invasion and proliferation (Fig. 6E).
After macrophages were infected with mutant and complement strains, the expres sion level of CASP3 in △peuR_THP1 was significantly lower than that in WT_THP1. When peuR was restored, the gene expression of CASP3 was restored significantly up-regulated compared with WT_THP1, suggesting that peuR may assist V. parahaemolyticus in inducing macrophage apoptosis. In addition, △peuS_THP1 also affected the expression level of CASP3, but the one in △peuS: C_THP1 did not recover significantly, suggesting that there may be other HKs that can activate peuR. The expression level of IL-10 in △peuS_THP1 and △peuR_THP1 was significantly higher than that in WT_THP1, indicating that peuS/R may inhibit the expression of IL-10 (Fig. 6F).

DISCUSSION
In this study, we found that V. parahaemolyticus can grow and proliferate in macrophages at the early stage of infection. Bacteria are forced to adapt to huge environmental changes after being taken up by macrophages as foreign bodies. V. parahaemolyticus should quickly give response strategies to maintain its survival, such as activating the corresponding signal transduction system and regulating the expression and transcrip tion of related genes. According to the analysis of expression profiles of V. parahaemoly ticus TCS, the majority of the TCS genes were inhibited or not expressed, and only a small number of genes were up-regulated. The significant change of the TCS genes after infection may result from its interaction with macrophages. We investigated seven crucial TCS genes that might assist V. parahaemolyticus control cells, and V. parahaemolyticus can change the cell's actin cytoskeleton, induce inflammatory cytokines, and activate the apoptosis pathway during the infection of macrophages.
Interestingly, the Cdc42 was significantly up-regulated in macrophages infected with V. parahaemolyticus. This result resembled Orth's report that Cdc42 was activated when V. parahaemolyticus entered non-phagocytic cells. Moreover, the study showed that V. parahaemolyticus activates Rac-GTP and CDC42 through the effector protein VopC, which changes the actin cytoskeleton and promotes bacteria to enter non-phagocytic cells (8). However, VPC17 and VPC44 did not contain VopC (5). But some studies have shown that V. parahaemolyticus effector protein, VopS, can mediate the inactivation of Cdc42 in mouse bone marrow-derived macrophages. In this paper, all four strains of V. parahaemolyticus had VopS (7), but the relationship between VopS and Cdc42 in THP-1 cell-derived macrophages remains to be further studied.

Research Article mSystems
In this study, V. parahaemolyticus may induce macrophage apoptosis through Cyt-C-mediated mitochondrial pathway at the early stage of infection (3.5 h). Apoptosis has two sides to bacterial infection. On the one hand, apoptosis is one of the means by which host cells eliminate bacteria. On the other hand, the induction of apoptosis cannot always protect host cells in infection because bacteria can use the host's apoptosis mechanism to eliminate the cells required for immune response (35). As a result, apoptosis serves a variety of intricate roles in bacterial infection regulation. However, combined with the amounts of intracellular bacteria (Fig. 1B), we discovered that V. parahaemolyticus was still undergoing multiplication at 3.5 h. The number of intracellular bacteria started to decline after 4.5 h. We speculated that V. parahaemolyticus might trigger apoptosis to aggravate cell rupture, which helps V. parahaemolyticus detach from cells.
To investigate the potential role of V. parahaemolyticus TCS in the regulation process, we infected macrophages after knocking out the TCS genes HK (peuS) and RR (peuR). The results revealed that △peuR had lesser toxicity than WT, suggesting that peuR had some involvement in the process of infection. In the TCS, the peuS is HK which detects external signals and phosphorylates the downstream peuR. And the peuR is an RR that primarily controls the transcription of particular downstream genes. However, the results suggested that there may be other unknown genes that can activate peuR. TCS is a complex regulatory mode with classical HK and RR modes but also has cross-regulation modes, which was also reflected in the PPI of peuR. Our research indicated that TCS might be involved in the V. parahaemolyticus infecting macrophages.

Conclusion
In summary, our work understood the interaction between V. parahaemolyticus and macrophages from two aspects of bacteria and cells. From the perspective of bacteria, the expression patterns of TCS in V. parahaemolyticus with different genotypes infected macrophages were characterized and the multiple vital genes for V. parahaemolyticus immune escape were provided. From the perspective of cells, the analysis of the cell gene alterations has revealed that V. parahaemolyticus might infect macrophages by controlling apoptosis, actin cytoskeleton, and cytokines. In addition, we also identified that peuS/R in TCS might affect the toxicity of V. parahaemolyticus to macrophages and might play a role in the induction of apoptosis, which supported the study of V. parahaemolyticus TCS in innate immunity regulation. This study provided several new insights into the pathogenicity of V. parahaemolyticus.

Bacterial strain and cells culture
Four V. parahaemolyticus strains were stored in our laboratory (Table 2) (5). The tryptic soy broth with 3% NaCl (wt/vol) (TSB + N) was used for culture bacteria. The THP-1 cell line was maintained in RMID1640 (Gibco, New Zealand) supplemented with 5% fetal bovine serum (FBS) (Gibco, New Zealand) in a humidified 5% carbon dioxide atmosphere. All strains and cells were cultured at 37℃.

Infection and intracellular survival bacterial counts
The THP-1 were seeded at 2 × 10 5 cells in 12-well plates , and added phorbol 12-myristate 13-acetate (PMA) (Sigma, USA) with a final concentration of 100 nmol/L to induce cell transformation into macrophages for 48-72 h. Overnight-grown bacteria were induced in TSB + N medium containing 0.05% (wt/vol) bile salt for 2 h and diluted with RMID1640. The cells were infected with bacteria at a multiplicity of infection (MOI) of ~10 per cell and then centrifuged at 200 × g for 5 min to synchronize the infection (17). At the specified times, the culture medium was removed, rinsed with 1× phosphate buffered saline (PBS), and treated with an infection media containing 100 ng/mL gentamicin. The number of intracellular bacteria is counted by CFU/mL at the indicated time points. Host cells were washed with 1× PBS to remove extracellular dead bacteria and lysed with 1% (vol/vol) TX-100 (8).

qRT-PCR
The gene expression level was identified by qRT-PCR (Roche Light Cycler 480, USA). RNA extraction and cDNA transcription were performed according to the kit steps (Vazyme, Nanjing, China). And the primers for TCS genes were shown in Table 3. The housekeeping gene of V. parahaemolyticus was 16S rRNA. The fold changes of relative gene expression were determined by the 2 -ΔΔCT method (7).

RNA sequencing, GO, and GSEA
After 1.5 h of infection at 37°C, cells were treated for 2 h with infection media containing 100 ng/mL gentamicin, and host cells were washed with 1× PBS in preparation for total RNA extraction. RNA sequencing was performed in Allwegene Gene Technology Co., Ltd. (Nanjing, China), and the platform is Illumina NovaSeq 6000. The genes were identified by qRT-PCR, the primers were shown in Table 4, and the housekeeping gene of the cell was GADPH. The bioinformatics tool Database for Annotation, Visualization, and Integrated Discovery (https://david.ncifcrf.gov) performed GO enrichment analysis on the DEGs (36). GSEA was applied to the whole transcriptome by the GSEA tool, and the KEGG database from MSigDB (v7.4) was used for the GSEA of the cell. The number of random combinations was 1,000. The gene set under the pathway of |NES| >1, NOM P-value <0.05, and FDR q-value <0.25 is significant (17,37).

Construction of deletion mutations
The allelic exchange was used to create the deletion mutations using pDS132, which contains the upstream and downstream sequences of the target genes (38). Briefly, using certain primers with restriction enzyme sites, the target gene's upstream and downstream homology arms were amplified and combined into the suicide plasmid pDS132. After being inserted into E. coli S17pir, the recombinant plasmid was conjugated into WT V. parahaemolyticus. A sacB counter-selectable marker and a chloromycetin resistance cassette are present on the plasmid, which could exchange genetic fragments twice with the genomes of V. parahaemolyticus by intermolecular recombination. PCR was used to identify the potential deletion mutant, and sequencing was used to confirm it. Based on the mutant strains, pBAD33 was used to complement the mutant gene to construct the complemented strain. The primer sequence, strains, and plasmids were shown in Tables 5 and 6.

Cell viability
To determine if the experimental treatment was harmful to cell survival, the Trans Detect Cell Counting Kit-8 (Sangon Biotech, Shanghai, China) was employed. 1 × 10 4 cells were put into each well of a 96-well plate. The cells were then incubated at 37℃, 5% CO 2 . After that, 10 µL of CCK-8 was added to the wells, and the cells continued to incubate for another hour. Finally, optical density values were calculated at 450 nm wavelength (39).