A novel adhesive factor contributing to the virulence of Vibrio parahaemolyticus

Bacterial adhesins play a pivotal role in the tight bacteria-host cells attachment to initiate the downstream processes and bacterial infection of hosts. In this study, we identified a novel adhesin, VpadF in V. parahaemolyticus. Deletion of VpadF in V. parahaemolyticus markedly impaired its attachment and cytotoxicity to epithelial cells, as well as attenuated the virulence in murine model. Biochemical studies revealed that VpadF recognized both fibronectin and fibrinogen. The binding of VpadF to these two host receptors was mainly dependent on the its fifth bacterial immunoglobulin-like group domain and its C-terminal tail. Our finding suggested that VpadF is a major virulence factor of V. parahaemolyticus and a potential good candidate for V. parahaemolyticus infection control for both vaccine development and drug target.

domains that mediate its adherence to different types of cells by binding to fibronectin (Fn) and phosphatidic acid in the initial stage of infection and subsequent cytotoxicity against 3T3 fibroblasts and RAW264.7 macrophages, but not to HeLa or Caco-2 epithelial cells. Sequence homology search showed that MAM7 is not only present in V. parahaemolyticus, but also a conservative protein possessed by several Gram-negative pathogens, including enteropathogenic Escherichia coli, Vibrio cholerae and Yersinia pseudotuberculosis 22 . Given that each Gram-negative pathogen usually produces more than one adhesive factor to initiate infection [23][24][25] and V. parahaemolyticus frequently acquires virulence factors, it is reasonable to speculate that V. parahaemolyticus has evolved or acquired specific, yet poorly understood mechanism to strengthen host-pathogen interactions. In this study, we identified a novel adhesin gene vp1767, which was referred to as VpadF (V. parahaemolyticus adhesive Factor), contributing to the attachment to and cytotoxicity against epithelial cells. This gene is essential for the lethal effect of V. parahaemolyticus on mice. Most importantly, VpadF bound to cell surface receptors, fibronectin (Fn) and fibrinogen (Fg), which may contribute to its host colonization and pathogenesis.

Methods
Bacterial strains, plasmids and growth conditions. V. parahaemolyticus strains, E. coli strains and plasmids used in this study were listed in ST 1. V. parahaemolyticus strains were cultured in LB medium supplemented with 2.5% sodium chloride (LBS) at 37 °C. Thiosulfate-citrate-bile salts-sucrose agar (TCBS) was used to select V. parahaemolyticus strains.

Construction of gene deletion and complementary strains.
The vp1767 gene was deleted from clinical V. parahaemolyticus strain VP3218 by homologous recombination as described previously 14 . Similar procedures were used to obtain Δ vcrD1 and Δ vcrD1Δ VpadF using specific primers (ST 2). To construct VpadF complementary strain, DNA fragment encoding vp1767 gene with a C-terminal Flag tag and N-terminal ribosome binding site was amplified using primers vp1767com-F and vp1767com-R (ST 2). PCR product was digested and cloned into the same digested pMMB207 to create pMMB207:VpadF. The recombinant plasmid was transformed to E. coli SY327 λpir and then conjugated to V. parahaemolyticus strains with a helper partner E. coli SY327 λpir carrying pPK2013. Transconjugants were selected on TCBS containing 5 μ g/ml chloramphenicol to obtain the complementary strains.
Bioinformatics analysis. Protein domain search was performed using PFAM (http://pfam.sanger. ac.uk/search), InterPro (http://www.ebi.ac.uk/interpro) and SMART (http://smart.embl-heidelberg. de). The subcellular location and the transmembrane region were predicated using TMHMM Server v2.0 (http://www.cbs.dtu.dk/services/TMHMM/) and TMpred (http://www.ch.embnet.org/software/ TMPRED_form.html). Multiple sequence alignments were achieved using Clustal W2 (http://www.ebi. ac.uk/Tools/msa/clustalw2/). Phylogenetic analysis was performed using MEGA version 5 after multiple alignment of the data via CLUSTAL_X. Distances were obtained using options according to Kimura's two-parameter model and clustering was performed by using the neighbor-joining method. The topology of the neighbor-joining phylogenetic tree was evaluated by using bootstrap resampling with 1000 replications 26 . Fractionation of bacterial cells. Outer membrane proteins of V. parahaemolyticus were obtained as previously described 27 . Briefly, exponential phase cultures (4 ml) were pelleted, lysed by sonication and then centrifuged at 20,000 g for 2 min. Supernatants were then transferred to fresh tubes and centrifuged again (20,000 g, 30 min, 4 °C). The pellets were re-suspended in 500 μ l 1% sodium lauryl sulfate in 10mM HEPES (pH7.4) and incubated at room temperature for 30 min. Outer membrane proteins were obtained after centrifugation at 20,000 g for 30 min (4 °C). The detergent soluble and insoluble fractions of the outer membrane proteins were separated by SDS-PAGE on 11% polyacrylamide gel, transferred to PVDF membrane, and subjected to Western Blotting using rabbit α -VpadV polyclonal antibody (1:2000, Pierce).
Expression and purification of recombinant proteins. Different fragments of VpadF were amplified by PCR using the primers listed in ST 2 and cloned into pET28 vector. Recombinant proteins were induced by IPTG and then purified using affinity chromatograph methods as previously described 28 , dialyzed into PBS buffer and examined by SDS-PAGE. Protein concentrations were measured by comparison with the BSA standards (Amresco).
Solid phase binding assay. One hundred μ l each of recombinant fibrinogen (F3879, sigma), full-length fibronectin (F2006, Sigma), HBD (F9911, Sigma) and CBD (F0162, 45-kDa, Sigma) domains of fibronectin at a concentration of 5 μ g/ml in 50mM Na 2 CO 3 , pH 9.6 was coated onto ELISA plates (IWAKI) at 4 °C overnight. After three-time washes with PBS, the plate was blocked with 1% (w/v) BSA in 50 mM Na 2 CO 3 (pH 9.6) and incubated at room temperature for 1h. The ELISA plate was then washed three times and then incubated with 3xFlag-tagged, recombinant VpadF truncation fragments at room temperature for 1 h. After triple washes, 100 μ l mouse α -3xFlag-HRP antibody (diluted 1:1000 0) was added to each well and incubated at room temperature for 1 h. 100 μ l TMB (T4444, sigma) was added to Scientific RepoRts | 5:14449 | DOi: 10.1038/srep14449 each well for color development (5 min, room temperature). After quenching with 100 μ l H 2 SO 4 (1 M), absorbance was read at 450 nm. BSA was coated simultaneously as negative control.

Cell attachment and cytotoxicity assays. HeLa and HT-29 cells were maintained in Dulbecco's
Modified Eagle Medium (DMEM, Invitrogen) supplemented with 10% fetal bovine serum (FBS, Invitrogen) at 37 °C in 5% CO2. Attachment assay was carried out as previously described 29 . Briefly,70-90% confluence cells were infected with freshly prepared V. parahaemolyticus at a multiplicity of infection (MOI) of ~10 CFU/cell. To consider the factor that V. parahaemolyticus may replicate during the experiment, bacterial cell were added to the empty wells of the cell culture plates and incubate for the same time as the binding experiment to determine the final total number of V. parahaemolyticus for the binding experiment. After 1 h infection, cells were washed three times with PBS, lysed with 1% Triton X-100 at 37 °C for 10 min to get the successful attached bacteria. The cell lysates and control bacteria were serially diluted and plated on LBS agar. Attachment rate was calculated by dividing bound bacteria to the total bacterial load.
For cytotoxicity assay, similar conditions were used as mentioned above. The supernatants from infected epithelial cell cultures (50-80% confluence) were collected at specific time points. The amounts of LDH release were determined using CytoTox 96 Non-Radioactive Cytotoxicity kit (Promega) following the manufacturer's instructions. The LDH activity in the supernatant of uninfected cells was also measured to obtain a spontaneous LDH value. Percent cytotoxicity was calculated with the following formula: (test LDH release-spontaneous release)/(maximal release-spontaneous release).

RT-PCR.
Overnight V. parahaemolyticus culture was firstly diluted 1:100 in fresh LBS broth and grown to exponential growth phase (OD600 ≈ 0.5-0.7). Cells were collected and re-suspended in the same volume of DMEM with different concentration of FBS (0, 5% and 10% FBS), respectively. After incubation for 30 min at 37 °C, bacteria were collected. Exponential growth phase culture that grown in LBS was directly collected. RNA was extracted using Trizol (Invitrogen) following the manufacturer's instructions. Residual DNA was removed from the sample with DNase (Turbo DNase, Ambion). RT-PCR reactions were performed using Superscript one-step RT-PCR system (Invitrogen) following the manufacturer's instructions. Primer pairs, rtVpadF-F/rtVpadF-R and rtrpoA-F/rtrpoA-R, were used to amplify the target genes, respectively (ST 2).
Murine Infection assay. V. parahaemolyticus strains (10 8 CFU) were intraperitoneally injected into 6to 10-week-old C57BL/6 mice (n = 10) as described previously [30][31][32] and the survival of mice was measured at the indicated time points. Three independent experiments were performed. The animal experiments were conducted in the National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention (CDC) following the guidelines and policies approved by China CDC.
Statistical analysis. Statistical analysis was performed using Prism software (version 5.0, GraphPad software). The data were the averages of three independent experiments. A P value of 0.05 or lower was considered significant. For analysis of the murine survival ratio, Kaplan-Meier and log rank tests were performed and P values of < 0.01 was considered statistically significant.

Results
VpadF is probably a adhesin gene in V. parahaemolyticus. To identify potential adhesins, we searched the genome sequence of a pandemic strain RIMD2210633. An open-reading frame, vp1767, designated VpadF, which encodes a putative protein with 736 amino acid residues, attracted our attention. VpadF consists of a putative transmembrane domain at the N-terminal, followed by five bacterial immunoglobulin-like tandem repeats (Bigs). Sequence alignment analysis showed that Bigs in VpadF possesses ~30% amino acid identity to the last Big of LigA and LigB, which possess 13 and 12 Bigs, respectively (SF3). LigA and LigB were described as adhesive molecules in Leptospira interrogans [33][34][35][36] . Bigs are widely present in numerous proteins, especially in surface proteins that involved in pathogen-host cells interactions 36,37 . It is likely that VpadF is an adhesive factor.
Proteins that possess similar domain configuration to that of VpadF also exist in other bacteria, such as Butyrivibrio fibrisolvens, Gemmatimonas aurantiaca, V. campbellii, V. harveyi, etc (Fig. 1A), suggesting that they share a common ancestor and represent a protein family. Phylogenetic analysis based on the core regions of VpadF and its homologs revealed that VpadF formed a distinct branch related to its homolog from V. cholerae, while it was separated from that of V. harveyi (Fig. 1B). PCR screening of VpadF on environmental and clinical V. parahaemolyticus strains showed that almost all isolates possessed this gene with a few exceptional cases ( Table 1). BLAST of the 293 available draft genome sequences of V. parahaemolyticus in GenBank confirmed that about 256 out of 293 (87%) whole genome sequences contained VpadF gene. BLASTN of GenBank non-redundant (NR) database failed to find any high homologues in any bacteria other than V. parahaemolyticus. In addition, the codon adaption index (CAI) and expression level analysis of VpadF were found to be normal compared to other genes in V. parahaemolyticus (Data not shown). These data suggested that VpadF is likely to be an intrinsic gene in V. parahaemolyticus and gets lost in some cases rather than a horizontal gene transferred (HGT) gene. VpadF localizes to the outer membrane. Firstly, we tried to find the appropriate condition for endogenous VpadF expression because some adhesin genes are not able to be expressed in vitro, or just transcribed under specific condition 38,39 . As shown in Supplementary Figure 1, the transcription of VpadF was stimulated in DMEM media in a FBS-independent manner, while its expression was inhibited in LBS broth. We then analyzed the localization of VpadF by heterogeneous expression in E. coli BL21. After Coomassie blue staining of the SDS-PAGE gel, a ~110 KDa protein band was clearly visible from the outer membrane fractions of E. coli BL21 expression VpadF ( Fig. 2A). To further determine the cellular localization of VpadF in V. parahaemolyticus, the outer membrane fractionation of three clinical V. parahaemolyticus strains, VPATCC17802, VP3218 and VP1074, were probed with VpadF antibody generated in our lab. It showed that VpadF was localized at the detergent insoluble fraction of the outer membrane of V. parahaemolyticus (Fig. 2B).

VpadF mediates V. parahaemolyticus infection to epithelial cells. Having confirmed that VpadF
is a surface protein, we then assessed the role of VpadF on V. parahaemolyticus attachment to HeLa and HT-29 epithelial cells. A Δ vcrD1 background strain (T3SS1 negative), which showed a much slower cell lysis rate, was constructed as wild type (WT) V. parahaemolyticus caused cell lysis quickly after infection, making it unfeasibl for the attachment assay 40 . The deletion of VpadF in the Δ vcrD1 strain dramatically decreased its attachment to HeLa and HT-29 epithelial cells (Fig. 3B). The complementary strain, Δ VpadFΔ vcrD1:pVpadF, showed ~3 fold higher cell adherence ratio than that of the Δ VpadFΔ vcrD1 strain (Fig. 3B), suggesting that VpadF is required for bacterial cell attachment and in trans expression of VpadF increases its cell attachment ratio.  To test whether the host cell adherence mediated by VpadF affects cytotoxicity of V. parahaemolyticus against the epithelium, the lysis of HeLa and HT-29 epithelial cells were measured by monitoring the release of lactate dehydrogenase (LDH) after infection with WT V. parahaemolyticus, Δ VpadF strain (vcrD1 positive) and the complementary strain Δ VpadF:pVpadF. At 3.5 h, Δ VpadF caused a ~50% decreased cytotoxicity compared to that caused by WT and Δ VpadF::pVpadF. After 5.5 h infection, HeLa and HT-29 epithelial cells were nearly completely lysed by WT and the VpadF complementary strains, while Δ VpadF infected cells showed ~30% less lysis (Fig. 3C). This suggested that VpadF plays an important role in epithelial cell infection.

VpadF is responsible for lethality in mice.
To explore whether VpadF affects V. parahaemolyticus pathogenesis, mice were infected with WT V. parahaemolyticus VP3218 and its VpadF deletion mutant, respectively. As shown in Fig. 4A-C, mice infected with WT strain displayed only 20% survival, while deletion of VpadF nearly completely abolished the lethality in mice after 48 h infection. After 96 h infection, no further lethal effect was observed in the mice infected by WT V. parahaemolyticus VP3218 or its VpadF deletion mutant (data not shown). This indicated that VpadF mediating V. parahaemolyticus-host interaction is required for its pathogenesis in mice.

VpadF binds both fibronectin (Fn) and fibrinogen (Fg).
To gain further insight into the mechanism by which VpadF mediates host cells attachment and thus contributes to the pathogenesis, we set out to identify potential host receptors of VpadF. It was suggested that besides mam7, V. parahaemolyticus produces other adhesins to bind Fn 29 . Initially, we assessed the Fn binding ability of VpadF. VpadFA, the VpadF truncated fragment that lack of 36 N-terminal amino acid residues was first purified (Fig. 5A and  SF2). The specific interaction between VpadFA and Fn was examined using the solid-phase binding assay. As shown in Fig. 5B, VpadFA was able to interact with Fn in a pronounced dose-dependent manner and reached a saturated status, with a Kd (the concentration able to saturate 50% of substrate) of ~200 nM. To evaluate which region in VpadF contributes to its binding property, all five Big repeats, VpadFB1 (the first Big domain), VpadFB2 (the second Big domain), VpadFB3 (the third Big domain), VpadFB4 (the fourth Big domain), VpadFB5 (the fifth Big domain) and the C-terminal end, VpadFC, were individually purified, respectively. Our results demonstrated that VpadFB2, VpadFB5 and VpadFC strongly bound to Fn and VpadFC displayed highest binding affinity. In contrast, VpadFB1, VpadFB3 and VpadFB4 showed weak binding activities to Fn (Fig. 5C).  To further demonstrate which parts of Fn are bound by VpadF, we tested the binding activities of VpadFA, B2, B5 and C to the immobilized N-terminal 30-KDa (heparin binding domain, HBD) and 45-KDa fragments (gelatin and collagen binding domain, CBD) of Fn. As shown in Fig. 5D,E, both HBD and CBD were targeted by VpadFA, B2, B5 and C. VpadFB5 and VpadFC bound HBD and CBD with similar avidities, while higher than those of VpadFB2.
Septicemia caused by V. parahaemolyticus is potential fatal to immunocompromised and liver failure individuals 41 . In Gram-positive pathogens, septicemia is related to their binding abilities to Fg since Fg is not only rich in human blood but also present in extracellular matrix 42,43 . Thus, we tested the binding ability of VpadF to Fg. Our results displayed that Fg was bound by VpadFA, with a Kd of ~400 nM (Fig. 5F). Similar to VpadF binding Fn, this protein also interacted with Fg mainly dependent on its C-terminal region, VpadFB5 and VpadFC (Fig. 5G).

Discussion
Initial host cell adhesion is the first important step in bacterial pathogenesis. In the case of V. parahaemolyticus, its infectious processes require the adherence to intestinal epithelial cells, resulting in epithelial cell extrusion, villus disintegration and formation V. parahaemolyticus-filled cavities 44 . Adhesins play a pivotal role in the tight bacteria-host cells attachment to initiate the downstream processes. In the present Figure 4. Survival rates of murine models infected with V. parahaemolyticus strains. C57BL/6 mice (n = 10) were infected intraperitoneally with WT or Δ VpadF strains (10 8 CFU) and the dead mice were recorded at specific time points during a 96 period of time. Only the data for the first 48 h was shown since no further change was observed in these mice for the next 48 h. Data from three independent experiments, (A-C), were shown. Kaplan-Meier and log rank tests were used to analyze the data (P < 0.001).
Scientific RepoRts | 5:14449 | DOi: 10.1038/srep14449 study, we discover a novel adhesin gene, VpadF, that plays a significant role on the pathogenesis of V. parahaemolyticus.
VpadF has several features that distinguish it from other previously described adhesive factors, such as MSHA pilus, enolase, capsular polysaccharide, T6SSs and MAM7 in V. parahaemolyticus. Firstly, this protein shows a unique domain organization. Secondly, VpadF is required for the lethal effects of V. parahaemolyticus on mice. Moreover, VpadF is a multifunctional adhesin that is capable of interact with both Fn and Fg. To the best of our knowledge, VpadF is the first reported adhesin that be able to bind Fg in Vibrio species. It was shown that Fg binding proteins have essential roles in pathogenesis by enabling bacteria to penetrate host barriers and spread in tissues 43 . VpadF is likely to both strengthen the attachment of V. parahaemolyticus to epithelial cells and modulate its spreading in the infected tissues. The molecular mechanism of the interactions between VpadF and host receptors is also different from those of other adhesins in Vibrio. For instance, OmpU from V. vulnificus recognizes RGD tripeptide that localizes the middle part of Fn and MAM7 interact with HBD of Fn 29,45 . In contrast, VpadF can bind the whole N-terminal fragments, HBD and CBD of Fn. VpadF also differs from Bigs possessing adhesins in Leptospira interrogans. Biochemical analysis revealed that the fifth Big repeat in VpadF significantly contributes to Fn and Fg binding, while in LigA and LigB, their binding avidities are mainly dependent on the 13 th and 12 th Big domain, respectively [46][47][48] . Moreover, unlike in LigA and LigB, only the last Bigs can recognize host components, the 2 nd Big repeat in VpadF also exhibited Fn/Fg binding property to some extent. These results implicates the number of Bigs is not the primary determinant of function and individual Big repeat in a single protein could be functional diverged even each Big fold shares some sequence similarity with each other (SF3).
In conclusion, we identified and characterized a novel and essential adhesion factor from V. parahaemolyticus and demonstrated its significant role on host cell attachment, cytotoxicity and pathogenicity. VpadF is a potential good candidate for V. parahaemolyticus infection control for both vaccine development and drug target.