Vibrio parahaemolyticus can be found in multiple cell types, including free-swimming forms and colony formations attached to inert surfaces such as shellfish (Letchumanan et al., 2014). A previous study utilized whole transcriptome next-generation sequencing (RNA-Seq) for analyzing the OpaR regulon in V. parahaemolyticus, which was cultured on a solid growth medium. RNA-Seq data were further validated through qRT-PCR for a select subset of transcription factor genes that were highly regulated by OpaR (Burke et al., 2015). In this study, we investigated the regulation of the protease genes of V. parahaemolyticus, which was cultured in a liquid growth medium.

We first inserted a FLAG-tag in front of the opaR gene on the VP93 chromosome. Western blot results demonstrated that FLAG-OpaR is steadily expressed from the early-log to late-stationary phases (Supplementary Figure 1). We collected bacterial cells from the late-log (OD₆₀₀ 2.8) and stationary (OD₆₀₀ 4) phases to perform ChIP-seq assays. FLAG-OpaR was bound to the genome and cross-linked by anti-FLAG beads; the genome was sheared in 300-bp fragments, and immunoprecipitated DNA was purified and sequenced. Reads were obtained from independent ChIP-seq assays in late-log and stationary phases. The input control of ChIP-seq was the VP93 genome without anti-FLAG antibody enrichment. All reads were mapped to the VP93 genome. Our ChIP-seq data (NCBI Gene Expression Omnibus database, GEO Accession GSE122479) revealed the presence of 435 and 835 OpaR-binding sites in late-log and stationary phases, respectively; there were 432 overlapping OpaR-binding sites between the two phases (Figure 1A).

OpaR regulates nine transcription factors in V. parahaemolyticus: CpsR, the Crl family, the Ars family, the CsgD/VpsT family, the FhlA family, the AsnC family, ExsA, AphA, and LafK (Burke et al., 2015). Moreover, transcriptional regulation of opaR, qrr2–4 (encoding QS regulatory small RNAs), and aphA by OpaR has been reported in V. parahaemolyticus (Zhang et al., 2012). In this study, we used Multiple Expectation Maximization for Motif Elicitation (MEME)¹ to analyze 838 OpaR-binding sequences, which showed a conserved OpaR-binding motif (TATTGATAAAATTATCAATA) (Figure 1B). From genes situated near to OpaR-binding sites, we identified seven protease genes that might be regulated by OpaR in V. parahaemolyticus. The ChIP-seq-defined regions of the seven protease genes are listed in Table 1. In addition, we used TOMTOM² to align the conserved OpaR-binding motif with other homologs of QS regulators in Vibrio spp., namely LuxR (Figure 1C) and HapR (Figure 1D). We observed that the OpaR-binding motif showed a high similarity with other QS regulator motifs. Comparison of the three QS regulator motifs—the conserved OpaR-binding motif, LuxR motif, and HapR motif (Figure 1)—indicated that TATTGATAAAATTATCAATA is the most conserved binding region of QS regulators in Vibrio spp.

Table 2 reveals the functions of seven putative OpaR-regulated protease genes classified using Gene Ontology (GO) and Clusters of Orthologous Groups (COG) annotations. These seven genes could be classified into two categories: metalloendopeptidases (LytM, Mcp02, M6 protease) and serine-type endopeptidases (serine protease, DegS, periplasmic protease, protease II). Moreover, they could be classified based on their cellular locations: inner membrane protease, DegS (VP0432) (Gottesman, 2017); periplasmic protease (VP2032) (Baird et al., 1991); and extracellular proteases, LytM (VPA1649) (Gode-Potratz et al., 2011), Mcp02 (VPA0755), M6 protease (VP0907) (Rompikuntal et al., 2015), serine protease (VPA0449) (Chimalapati et al., 2018) and protease II (VPA1467). Three of these—namely Mcp02, serine protease and protease II—are secreted by the type II secretion system (Chimalapati et al., 2018).

The distribution of OpaR-binding peaks relative to the seven protease genes are shown in Figure 2. These peaks overlaid the translational start site of serine protease and degS. OpaR-binding peaks appeared on the upstream of lytM, mcp02, M6 protease, and protease II; however, one OpaR-binding peak presented at the terminator region and overlaid the translational stop site of the periplasmic protease. Predicted OpaR-binding sites in the promoters of lytM, mcp02, M6 protease, serine protease, and degS were found not only in the late-log phase but also in the stationary phase, whereas the OpaR occupancy in the promoter of protease II and terminator of periplasmic protease was only found in the stationary phase. Normalized reads per million (RPM) values [total reads/mapped reads (millions) x region length (bp)] represent the binding strength of OpaR-binding peaks. In addition, the binding length of OpaR ranged from 177 bp (degS) to 372 bp (lytM) (Table 1).

Each protease gene expression levels at different growth phases were examined using qRT-PCR. Genes encoding LytM (Figure 3A), Mcp02 (Figure 3B), DegS (Figure 3E), and protease II (Figure 3F) were expressed from late-log to stationary phases, whereas those encoding M6 protease (Figure 3C), serine protease (Figure 3D), and periplasmic protease (Figure 3G) were primarily expressed during the late-log phase. Growth curves of VP93 were monitored by measuring OD₆₀₀ at every growth stage (Figure 3H). Therefore, we used late-log and stationary phase cultures to study the correlation between OpaR and the seven protease genes.

We used qRT-PCR to measure the relative transcript levels of seven protease genes in wild-type VP93 and mutant ΔopaR strains at late-log and stationary-phase growth. Our results revealed that OpaR represses the expression of genes encoding LytM (1.9-fold, p = 0.0069), Mcp02 (1.5-fold, p = 0.0058), M6 protease (2.0-fold, p = 0.0074), Serine protease (4.8-fold, p = 0.0054), DegS (2.1-fold, p = 0.0085), Protease II (25.7-fold, p = 0.0001), and Periplasmic protease (2.6-fold, p = 0.0072) in the late-log phase (Figure 4A). Moreover, this differential gene expression was also observed in the stationary phase for both Serine protease (repressed 12.7-fold, p = 0.0060) and DegS (repressed 1.5-fold, p = 0.0018) (Figure 4B).

To determine the role of these OpaR-occupied positions of protease genes from Figure 2, the promoter-fusion luxAB reporter assay was conducted. We amplified the aforementioned putative OpaR-binding sequences of each protease gene and cloned them into the promoter-less vector pSAluxAB. Bioluminescence was measured for the wild-type VP93 and the mutant ΔopaR strains at late-log and stationary phases. The results revealed that OpaR represses the bioluminescence expression for protease genes through OpaR-binding sites in the late-log phase, including lytM (1.2-fold, p = 0.0002), mcp02 (1.8-fold, p = 0.0089), m6 protease (1.1-fold, p = 0.0017), serine protease (1.4-fold, p = 0.00001), degS (6.0-fold, p = 0.0106), protease II (1.1-fold, p = 0.0001), and periplasmic protease (1.2-fold, p = 0.0094) (Figure 4C). However, the expressions of degS (1.6-fold, p = 0.0030), protease II (1.1-fold, p = 0.0003), and periplasmic protease (1.1-fold, p = 0.0076) were regulated by OpaR in the stationary phase (Figure 4D).

The electrophoretic mobility shift assay (EMSA) was performed to evaluate the potential of the direct binding of OpaR to the promoters of the protease genes. We amplified the DNA fragments as the probes (P1 and P2) for each target gene, and the positions of EMSA probes are listed in Table 1. EMSA demonstrated that OpaR bound six target protease gene promoters (Figure 5). Two OpaR-binding sites (P1 and P2) were present within lytM, m6 protease, and protease II promoter regions (Figures 5A–C), and only one (P1) was present on degS, mcp02, and serine protease promoter regions (Figures 5D,F,G). Indeed, in OpaR EMSA, two shifted bands were observed in the P1 region of lytM, protease II, and degS promoters (Figures 5A,C,D). These bands are attributable to the presence of more than one individual binding sites in the promoter P1 region; they may also represent two or more different states of DNA binding. Notably, we also observed a periplasmic protease 3’ P1 region that shifted in the presence of 2.8–11.2 pmol of OpaR (Figure 5E). OpaR did not bind to 16S rDNA as the negative control of EMSA (data not shown).

We amplified the promoter region containing all the EMSA probe sequences of each protease gene to design footprinting probes. The 5′-end of the sense and antisense footprinting probes were labeled with 6-carboxyfluorescein. Labeled probes were incubated with purified His-OpaR and digested by DNase I. Capillary electrophoresis was used to analyze the digested DNA fragments. The OpaR-binding sequences on the seven protease genes are listed in Table 3, and the specific OpaR-binding positions are presented in Figure 6. The OpaR-binding position was defined by the difference of peak observed between the control without OpaR and samples with OpaR. We found two OpaR-binding sequences for lytM, m6 protease, protease II, degS (Figures 6A–D). Capillary electrophoresis analysis identified one OpaR-binding sequence for periplasmic protease, mcp02 and serine protease (Figures 6E–G). Notably, upon aligning the OpaR-binding sequences obtained from the DNase I footprinting assay with the EMSA probe sequences, degS EMSA revealed a double band shift, and we could locate and identify two OpaR-binding sequences in the DNase I footprinting assay result (Figures 5D, 6D).

To further verify whether the OpaR-binding site appeared at the periplasmic protease 3′region, we performed sequence analysis using ARNold program (Naville et al., 2011). The result revealed that a Rho-independent transcriptional terminator (intrinsic terminator) was present and located 29 to 50-nt downstream of the periplasmic protease stop codon. DNase I footprinting assay result showed that one OpaR-binding sequence located 33 to 52-nt upstream of the stop codon (Figure 6E). OpaR binding close to the stop coden of the ORF may block transcriptional elongation and result in the target gene repression. These results conclude that OpaR binds to the ORF near 3′ region to repress periplasmic protease expression.

At a high cell density, OpaR repressed three groups of proteases: extracellular metalloprotease, extracellular serine protease, and periplasmic serine protease. However, PrtA (VPA0227) is an extracellular serine protease that was promoted by OpaR, as reported in a previous study (Chang and Lee, 2018). As shown in Figure 7, OpaR directly regulated the gene expression of the eight protease genes: three environmental survival genes—periplasmic protease (Baird et al., 1991), serine protease, and protease II (Kanatani et al., 1991, 1992)—and five bacterial virulence genes—lytM (Gode-Potratz et al., 2010), mcp02 (Gao et al., 2010; Miyoshi, 2013), m6 protease (Vaitkevicius et al., 2008; Rawlings et al., 2018), prtA (Lee et al., 2002), and degS (Mathur et al., 2007).

All bacterial strains, plasmids, and primers are listed in Supplementary Tables 1, 2. V. parahaemolyticus strains and their derivatives were cultured in tryptic soy bean broth containing 3% NaCl (TSB3) (Difco) at 35°C. Escherichia coli strains were grown in Luria–Bertani (Difco) medium at 35°C. XL1-Blue (Bullock et al., 1987) and S17-1 λpir (Delorenzo et al., 1993) were used for cloning and conjugation, respectively. Mutant and reporter strains were derived from VP93. When necessary, antibiotics were added to the culture medium: ampicillin, 100 μg ml^–1; kanamycin, 25 μg ml^–1; and chloramphenicol, 20 μg ml^–1 for E. coli and 5 μg ml^–1 for V. parahaemolyticus.

We constructed a FLAG-tag insertion strain using homologous recombination. First, a FLAG-tag (DYKDDDDK) was inserted after the translational start site of opaR using overlapping PCR. The FLAG-opaR fragment was then ligated to pDS132, a suicide vector (Philippe et al., 2004), to yield the pDS132–FLAG-opaR plasmid. This mobilizable plasmid was transferred from E. coli S17-1 λpir to VP93 via conjugation (Schafer et al., 1990). The conjugants were selected from a TSB3 Cm plate for the first step of the homologous recombination. Next, thiosulfate–citrate–bile salts–sucrose plates containing 6% sucrose were used to select the second homologous recombination conjugants. PCR amplification and sequencing were used to confirm the generation of the FLAG-tag insertion strain.

FLAG-OpaR/VP93 strain culture was collected at every growth stage by centrifugation. Total proteins of the FLAG-OpaR/VP93 strain were separated using 12.5% SDS-PAGE gels and transferred to PVDF membranes (Merck Millipore, Germany). The membrane was blotted with anti-FLAG antibody as the primary antibody. The secondary antibody is an anti-rabbit IgG HRP-linked antibody. Immobilon Western Chemiluminescent Horseradish Peroxidase Substrate (Merck Millipore, Germany) was added, and signals were detected using a chemiluminescence image system, MutiGel-21 (Bio Pioneer Tech, Taipei, Taiwan).

FLAG-OpaR/VP93 was grown to the late-log (OD₆₀₀ 3) and stationary (OD₆₀₀ 4.2) phases. Next, the bacterial culture was incubated with 1% formaldehyde for 20 min at 25°C, and cross-linking was ceased using 0.5 M glycine. Bacterial cells were then harvested by centrifugation, washed three times with Tris-buffered saline buffer, and resuspended in a lysis buffer [10 mM Tris, 100 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 0.1% deoxycholate, and 0.5% N-lauroylsarcosine (pH 8.0)] containing a protease inhibitor cocktail (Sigma). The lysed cells were sonicated using Bioruptor Pico^® (Diagenode) for 20 cycles. Cell debris was removed by centrifugation; supernatant DNA fragments ranged from 300 bp to 500 bp. Anti-FLAG antibody beads (Sigma) were incubated with supernatant DNA fragments overnight at 4°C. After the ChIP reaction, samples were washed five times with a radioimmunoprecipitation assay buffer [50 mM hydroxyethyl piperazineethanesulfonic acid buffer, 500 mM LiCl, 1 mM EDTA, 1% Non-idet P-40, and 0.7% deoxycholate (pH 7.5)] and twice with Tris-buffered saline buffer. ChIP DNA was eluted in an elution buffer (50 mM Tris-EDTA buffer, 1% EDTA, pH 7.5) at 65°C for 30 min to de-cross-link. Samples were treated with RNase A and proteinase K (Dong and Mekalanos, 2012), and ChIP DNA was then purified using the Agencourt AMPure XP PCR Purification Kit (Beckman Coulter). Libraries for ChIP-seq were prepared using the Ovation Ultralow Library System V2 1–96 (NuGEN). Sequencing was performed using the HiSeq 2000 sequencing system (Illumina). ChIP-quantitative PCR was performed using SYBR green (BioRad) to determine the quality (S/N ratio) of ChIP-DNA. AphA and VPA0606 were used as positive and negative controls, respectively (Burke et al., 2015).

We used ”Map Reads to Reference” in CLC Genomics Workbench v9.5 platform conducting Reference mapping. We also used ”Transcription factor ChIP-Seq Analysis” conducting the peak finding. Maximum P-value for peak calling is < 0.05. The OpaR-binding motif was generated using MEME¹ (Bailey et al., 2009). TOMTOM, a web-based tool² (Bailey, 2011), was used to align the conserved OpaR-binding motif with other homologs. The ChIP-seq data could be obtained with the GEO accession number GSE122479.

His₆-OpaR protein was expressed and purified as previously described (Chang and Lee, 2018). Briefly, to generate the pET28a–opaR plasmid, opaR was cloned into the six-His-tag expression vector pET28a (Novagen) (Supplementary Table 1). The clone was transformed into E. coli BL21 (DE3) to express OpaR with an N-terminal fusion tag. E. coli BL21 (DE3) with pET28a–opaR was grown in 200 mL Luria–Bertani broth supplemented with 25 μg mL^–1 kanamycin at 37°C. When OD₆₀₀ reached 0.4–0.6, 1 mM isopropyl β-D-1-thiogalactopyranoside was used to induce His₆-OpaR protein expression at 30°C for 4 h. The cells were then collected and disrupted using the Constant Cell Disrupter System (Constant Systems Ltd., Daventry, United Kingdom), and nickel–nitrilotriacetic acid chromatography (GE Healthcare, Chicago, IL, United States) was used to purify His₆-OpaR. His₆-OpaR purity was confirmed using sodium dodecyl sulfate polyacrylamide gel electrophoresis. Proteins were quantified using the method described by Bradford (Bradford, 1976).

We collected the bacterial cultures from different growth phases of VP93 and ΔopaR. Cells were mixed with RNAprotect Bacteria Reagent (Qiagen) and centrifuged at 5000 g at 4°C for 15 min. Total RNA was isolated using TRIzol. Reverse transcription–PCR (RT-PCR) was performed using HiScript I Reverse Transcriptase (Bionovas Biotechnology), according to the manufacturer’s instructions, and real-time quantitative reverse-transcriptase (qRT-PCR) was performed using SYBR Green Master Mix (iQTM SYBR Green^® Supermix, BioRad) on a BioRad CFX96 system. Relative expression values were determined using 2^{–(Δ Ct Target – Δ Ct Reference)}. Ct is the fractional threshold cycle, and the reference was 16S ribosomal RNA gene. The specific primer sets used for qRT-PCR are listed in Supplementary Table 2.

The promoter regions of the six OpaR-regulated protease genes and one 3’ end region of periplasmic protease were amplified using specific primer sets (Supplementary Table 2). The promoter fragments were cloned into a promoter-less vector pSAluxAB (Supplementray Table 1), with SmaI and XbaI restriction sites. These plasmids were transformed into VP93 and ΔopaR by using the Gene Pulser Xcell Electroporation System (BioRad). Luciferase activity was measured by adding n-decanol [final concentration 0.001% (vol/vol)] as the substrate to 2 mL bacterial cultures. Luminescence was detected on a Spectrofluorometry F-2500 device (Hitachi, Tokyo, Japan) and is reported as specific light units (SLU; relative luminescent light units per second per milliliter per OD₆₀₀ unit).

The specific primer sets used to amplify EMSA probes are listed in Supplementary Table 2. These probes were labeled with digoxigenin-11-ddUTP at their 3′-ends using the second generation DIG Gel Shift Kit (Roche Applied Sciences, Mannheim, Germany). In the binding reaction, 0.4 ng labeled fragment was incubated with various quantities of purified His-OpaR protein in 20 μL binding buffer [100 mM hydroxyethyl piperazineethanesulfonic acid (pH 7.6), 5 mM EDTA, 50 mM (NH₄)₂SO₄, 5 mM dithiothreitol, 1% (wt/vol) Tween 20, and 150 mM KCl] at 25°C for 30 min. In competition analyses, the protein–DNA complex was mixed with 100 ng unlabeled fragment. These samples were separated using 6% native polyacrylamide gel electrophoresis in 0.5 × Tris-borate-EDTA buffer and transferred by electroblotting to a positively charged nylon membrane. Finally, immunological treatment and chemiluminescent signal detection were performed according to the instructions for the second generation DIG Gel Shift Kit (Roche Applied Sciences, Mannheim, Germany).

The DNase I footprinting assay was performed using the method of Zianni et al. (2006). The probes of the DNase I footprinting assay were amplified using specific primer sets (Supplementary Table 2). These probes contained the promoter regions or the 3′ end region of the protease genes. The 5′-ends of the sense and antisense probes were labeled with 6-carboxyfluorescein. In the binding reaction, 75 ng labeled probe and various quantities of His-OpaR protein were mixed in 20 μL binding buffer (second generation DIG Gel Shift Kit) and incubated at 25°C for 30 min. We then added 0.015 U DNase I (Thermo scientific, Germany) at 25°C for 1 min to digest the DNA–protein complex and ceased this reaction using 2.77 μL stop solution (5 mM EDTA) at 65°C for 10 min. The Agencourt AMPure XP PCR Purification Kit (Beckman Coulter) was used to purify the digested DNA fragments. Subsequently, 1 μL digested DNA fragments were added to a mixture of 8.5 μL highly deionized formamide and 0.5 μL GeneScan^TM LIZ 600 Size Standard (Applied Biosystems). The samples were analyzed using the ABI 3730xl DNA Analyzer (Applied Biosystems), and the results were analyzed using Peak Scanner Software v1.0 (Applied Biosystems).

Statistical differences between the three samples were measured using Student’s t-test with a two-tailed distribution. P < 0.05 or P < 0.01 indicated statistical significance.