Bacterial strains and plasmids used in this study are shown in Table 1. The V. cholerae O1 El Tor biotype, Ogawa serotype strain C7258 was used as wild-type (WT) precursor (Peru isolate, 1991). The mutant Δcrp (WL7258) was generated previously (Li et al., 2019). E. coli DH5αλpir and S17-1λpir were, respectively, used for cloning and conjugation purposes, ER2566 is used as host for the expression and purification of hns cloned into the pTXB1 vector. All strains were grown in Luria–Bertani (LB) broth (Oxoid, Basingstoke, United Kingdom) containing 1% NaCl (170 mM) at 37°C. When necessary, culture media were supplemented with ampicillin (Amp, 100 μg/ml), chloramphenicol (Cm, 10 μg/ml for E. coli, 2.5 μg/ml for V. cholerae) or polymyxin B (100 units/ml). Isopropyl-β-D-thiogalactopyranoside (IPTG) was used at a concentration of 0.5 mM for induction purposes.

Mutants Δhns, Δfis were constructed by homologous recombination mediated by suicide plasmid using C7258 as a precursor, while ΔcrpΔhns and ΔcrpΔfis used Δcrp (WL7258) as a precursor. Upstream and downstream chromosomal DNA fragments flanking the hns or fis were amplified from C7258 genomic DNA using the corresponding primers listed in Table 2. The amplicons were stitched together by overlapping PCR. Δhns or Δfis fragments were cloned into pWM91 and constructed in DH5αλpir. The resulting pWM91-Δhns or pWM91-Δfis was introduced into C7258 or Δcrp (WL7258) by conjugation from S17-1λpir. Exconjugants and mutants were selected as described previously (Wu et al., 2015; Li et al., 2019).

Chromosomal complementation strain Δhns::hns was constructed using a temperature-sensitive transposable plasmid PGRG25. pGRG25 contains a Tn7 transposon that can carry the target fragment for specific recombination with the chromosome of the host bacterium, directionally inserted into the downstream of the glms in the chromosome (McKenzie and Craig, 2006). For this purpose, hns promotor region and open reading frame (ORF) was amplified and cloned into pGRG25 to generate pGRG25-hns which was mobilized into Δhns by conjugation. The chromosomal insertion of the transposon was induced by 0.1% arabinose in LB broth at 30°C for 16 h, then screened on LB agar by a temperature at 42°C. Grown colonies were tested for Amp sensitivity, and proper insertion of hns downstream of glms was verified by PCR with primers targeting the glms and hns sequences. Primer sequences used here are shown in Table 2.

Vibrio cholerae strains were cultured to OD₆₀₀ 1.0. Total RNA extraction, removal of chromosomal DNA contamination and cDNA synthesis were performed as previously described (Wu et al., 2015). Equation R = 2^{− (ΔCq rfbT-ΔCq thyA)} was used to calculate the relative expression values (R) of rfbT, where Cq is the threshold cycle fraction and thyA was used as an internal reference. A control reaction with total RNA as a template was performed for each sample to exclude contamination from chromosomal DNA. Primers used were listed in Table 2.

Two different length fragments of rfbT promoter region were, respectively, amplified and cloned into pBBRlux upstream of the promoterless luxCDABE operon. The resultant fusion plasmids pBBRlux-rfbT1 and pBBRlux-rfbT2 were constructed in DH5αλpir and then mobilized into V. cholerae strains C7258 and ∆hns by conjugation from S17-1λpir. pBBRlux-rfbT1-M1, pBBRlux-rfbT1-M3-1, and pBBRlux-rfbT1-M3-2 reporter fusions containing the specific mutations in the predicted H-NS binding sites were generated by PCR-based site-directed mutagenesis with pBBRlux-rfbT1 as a template. Overnight cultures of V. cholerae strains containing lux reporter fusion plasmid were diluted at 1:100 in fresh LB and incubated at 37°C with shaking to grow to exponential phase. 200 μl of the broth was transferred into 96 well microtiter plates (Costar 3,917) every 1 h, and luminescence and OD₆₀₀ were measured using a microplate reader (Infinite M200 Pro, Tecan, Austria). Luminescence activity was calculated as light unit/OD₆₀₀ as previously described (Pan et al., 2018).

E. coli strain ER2566 containing the recombinant expression plasmid pXTB1-HNS (Wang H. et al., 2012) was cultured to OD₆₀₀ of 0.5 with shaking at 37°C and then protein expression was induced with 0.4 mM IPTG for 4 h at 28°C. The cells were collected by centrifugation, resuspended in ice-cold Column buffer (20 mM Tris–HCl, pH 8.0, 0.5 M NaCl, and 1 mM EDTA), and lysed by sonication. The cell debris was removed by centrifugation, and H-NS-intein fusion protein with chitin binding domain (CBD) was purified using IMPACT™ Kit (New England Biolabs, United Kingdom) according to the manufacturer’s instructions. The clarified lysate was slowly loaded onto the equilibrated chitin column, and then the chitin column was washed with 20 bed volumes of Column Buffer. Subsequently, the column was quickly washed with 5 bed volumes of the Cleavage Buffer (Column Buffer containing 80 mM DTT), and then incubated at 4°C overnight for full cleavage reaction on-column. Finally, the H-NS was eluted with Column Buffer. H-NS-containing fractions were combined and dialyzed against Column Buffer at 4°C to remove DTT. The purity of the recombinant H-NS was analyzed by SDS-PAGE (Figure 1A), and the protein concentration was determined by a Pierce BCA protein assay kit (Thermo Fisher Scientific, United States). The protein was stored in 20% glycerol at - 80°C.

Probe-hns was a 256 bp fragment of rfbT promoter region containing all of the predicted H-NS binding sequences, which was amplified with 5′ biotin-labeled primers. Competing cold probe was amplified using the same primers without a biotin label. Probe-N7 (Li et al., 2019) was a 140 bp fragment of rfbT promoter region containing the previously determined CRP binding site and used as a nonspecific control probe for H-NS binding (Li et al., 2019). The reaction mixture of 15 ng biotin-labeled probe with increasing amounts of purified H-NS protein in reaction buffer (10 mM Hepes, 150 mM KCl, 1 mM EDTA, 1 mM DTT, 10 mM (NH₄)₂SO₄, 0.2% Tween 20) together with 100 ng BSA and 100 ng CF-DNA in each reaction (20 μl) was incubated at 28°C for 30 min and then separated on a 6% native polyacrylamide gel. The free and H-NS-bound probes were visualized with the Chemiluminescent Nucleic Acid Detection Module (Thermo Fisher Scientific, United States) according to the manufacturer’s instruction after transferring them onto nylon membranes.

For preparation of fluorescent FAM-labeled probes, the promoter region was PCR amplified with 2 × HIFI DNA polymerase premix from the plasmid pBBRlux-rfbT1 using primers of rfbT-up (FAM)and rfbT-dn. The FAM-labeled probes were purified by the Wizard® SV Gel and PCR Clean-Up System (Promega, United States) and were quantified with NanoDrop 2000C (Thermo, United States).

DNase I footprinting assay was performed as previously described (Wang Y. et al., 2012). For each assay, 250 ng probes were incubated with different amounts of protein in a total volume of 40 μl. After incubation for 30 min at 25°C, 10 μl solution containing about 0.015 unit DNase I (Promega, United States) and 100 nmol freshly prepared CaCl₂ was added and further incubation was performed at 37°C for 1 min. The reaction was stopped by adding 140 μl DNase I stop solution (200 mM unbuffered sodium acetate, 30 mM EDTA, and 0.15% SDS). Samples were first extracted with phenol/chloroform, and then precipitated with ethanol. Pellets were dissolved in 30 μl MiniQ water. The preparation of the DNA ladder, electrophoresis, and data analysis were the same as described before (Wang Y. et al., 2012), except that the GeneScan-LIZ600 size standard (Applied Biosystems) was used.

GraphPad Prism 9 software was used for statistical analysis and graphical representation of data. Statistical significance was determined by an unpaired two-tailed Student’s t-test.

Though the gene rfbT has been recognized as the genetic determinant of Ogawa serotype of V. cholerae O1 serogroup for more than 30 years (Stroeher et al., 1992) and various kinds of mutations were revealed in the rfbT coding sequence of isolates from different space–time sources (Sharifnia et al., 2012; Liang et al., 2013; Karlsson et al., 2016), its regulation and the molecular structural features except the transcriptional start site, putative −35 and − 10 elements of the promoter-regulatory region remain unclear. Previously, we identified a cis-regulatory element, i.e., a non-canonical CBS in the promoter region, through which global regulator CRP exerts an activational effect (Li et al., 2019). Further sequence analysis revealed that the G + C content of promoter-intergenic region of rfbT (41.2%) is quite low compared with the V. cholerae genome in general (47.7% for chromosome I and 46.9% for chromosome II). In other words, the promoter-intergenic region of rfbT is AT rich and probably prone to be regulated by small nucleoid associated proteins such as H-NS and Fis which tend to bind AT rich sequences. Indeed, subsequent Virtual Footprint and PRODORIC analysis using the 10-bp H-NS consensus (Bouffartigues et al., 2007; Figure 2A) revealed 3 potential H-NS binding elements, site 1 (5’-CCTATTAAAG-3′), site 2 (5’-TATCAAACGT-3′), and site 3 (5’-TCAATAAAAT-3′) in the rfbT promoter-intergenic region (Figure 2B). The three binding sites are, respectively, located at nucleotides −283 to −273, −250 to −240, and − 120 to −110 relative to the rfbT start codon, and are all downstream of nonclassical CBS (Figure 2B). Of these, 8 of the 10 bp at site 3 is consistent with the consensus, followed by site 1 with 5 bp, and finally site 2 with only 4 bp (Figure 2A). It’s worth noting that the site 3 overlaps the predicted −10 promoter element (Figure 2B). These findings strongly indicated the possibility that H-NS regulates serotype-shifting gene rfbT expression.

To determine whether H-NS is involved in the regulation of rfbT, we constructed an hns deletion mutant using C7258 as a precursor and detected the rfbT mRNA level in WT C7258, and Δhns mutant. As shown in Figure 3A, compared to the C7258, Δhns mutant statistically produced more rfbT mRNA. To further confirm the result, we constructed an hns complementation strain C7258Δhns::hns where a single copy of hns gene with its native promoter region was integrated downstream of chromosomal glms. As displayed, the rfbT mRNA abundance was reduced in C7258Δhns::hns compared to Δhns mutant, though the expression level was not restored to the WT level (Figure 3A). Together, these results showed that H-NS negatively regulates rfbT expression.

In our previous study, we demonstrated that CRP could positively regulate rfbT transcription, and then we wondered whether H-NS regulates rfbT expression in a CRP-dependent manner (Li et al., 2019). For this purpose, we constructed a ΔcrpΔhns double mutant and compared its rfbT mRNA expression with those of Δhns and Δcrp mutants. We found that, contrary to the Δcrp mutant where rfbT expression was obviously lower than its WT and Δhns, the rfbT mRNA level was significantly increased in ΔcrpΔhns (Figure 3A), and additionally, the Δhns and ΔcrpΔhns mutants displayed roughly similar rfbT mRNA level. These results proved that H-NS negatively regulates the expression of rfbT in a CRP-independent manner. From another perspective, these results also indicated that CRP is not required for rfbT activation in H-NS negative background.

Like H-NS, Fis belongs to the small family of nucleoid-associated proteins and is widely implicated in the control of gene expression through binding to the A−/AT-tracts-constituted binding site (Cho et al., 2008). To find out whether Fis takes part in the regulation of rfbT, we constructed Δfis and ΔcrpΔfis deletion mutants and measured the rfbT mRNA levels. As displayed in Figure 3B, the rfbT mRNA level in Δfis is similar to the WT, and deletion of crp significantly reduced rfbT expression regardless of the presence or absence of fis, implying that Fis is not involved in the regulation of rfbT.

To determine whether the H-NS-mediated repression of rfbT occurs at transcription level, we constructed two transcriptional reporter plasmids by fusing the different length fragments of promoter region of rfbT to the promoterless bioluminescence reporter genes luxCDABE. The 552-bp promoter fragment in pBBRlux-rfbT1 contains both the previously identified CBS and the predicted H-NS binding sites, while the 403-bp promoter region in pBBRlux-rfbT2 lacks the CBS. Consistent with the rfbT mRNA expression, bioluminescence activities of both pBBRlux-rfbT1 and pBBRlux-rfbT2 in Δhns were significantly higher than in its WT (Figure 4). These results demonstrated that H-NS negatively regulates rfbT at the promoter level. H-NS represses the transcription of rfbT probably through binding to the predicted binding sites and therefore the direct interaction still needs to be clarified.

To verify the direct binding of H-NS to the rfbT promoter region, we performed EMSA with purified H-NS protein. A 256-bp DNA fragment of rfbT promoter region encompassing the three predicted H-NS binding sites were labeled with biotin at 5′ end and used as specific probe. The same fragment without biotin label was employed as competing cold probe. The 135 amino acids V. cholerae H-NS protein was induced and purified from pTXB1-HNS recombinant plasmid (Wang H. et al., 2012) in E. coli host ER2566 and reached a purity of higher than 90% (Figure 1A). As displayed in Figure 1B, inclusion of H-NS (0.66 μM) generated two shifted bands with slower mobility. With the increment of H-NS (>1.32 μM), the intensity of the higher-shifted band was increased, concomitantly with the decrease of the amount of free probe. Adding the same, but unlabeled 256-bp DNA fragment greatly competed with the labeled probe in a dose-dependent manner. Both of the 10-fold and 50-fold addition of the competing cold probe completely abolished the shifted-band, and concurrently the labeled probe was released as the free one. As expected, the amount of the released free probe is more in 50-fold cold probe reaction mixture than the 10-fold one. To further validate the specificity of H-NS binding, we adopted Probe-N7 which is a 140-bp DNA fragment of rfbT promoter region containing the intact CBS but without the predicted H-NS binding element (Li et al., 2019). No apparent shifted band was observed even when Probe-N7 was incubated with higher amount of H-NS (1.32 μM or 1.98 μM) under the same reaction condition, indicating that H-NS could not bind to the rfbT promoter fragment lack of specific binding site.

To further clarify the real H-NS binding site in the rfbT promoter and dissect the binding sequences, we conducted DNase I footprinting analysis. As shown in Figure 1C, the assay revealed one clearly protected region against DNase I digestion which is composed of TCCTTCTAAATGTCAATAA extending from −132 to −113 relative to the start codon of rfbT, i.e., this region completely overlaps the predicted binding site 3 extending from −120 to −110. Though EMSA displayed two H-NS retarded bands that seems somehow in accordance with the predicted existence of two more conserved binding sites 1 and 3, unexpectedly, the less conserved binding site 1 was not confirmed in the DNase I footprinting assay, which suggested that the more conserved site 3 is intrinsically the real binding site with high-affinity. Of course, we cannot exclude the possibility that the current assay condition did not favor binding to the low-conservation site 1.

To further validate that site 3 is the real functional H-NS binding site and the less conserved site 1 has no function, we set out to introduce mutations at sites 1 and 3 through PCR-based site-directed mutagenesis with pBBRlux-rfbT1 as the template (Figure 5A). We introduced 4-bp changes in the site 1 by replacing the ATT and the antepenultimate A with CAC and C, respectively, to generate a new construct pBBRlux-rfbT1-M1. Considering that site 3 overlaps the predicted −10 promoter element, we introduced two sets of mutations to construct pBBRlux-rfbT1-M3-1 and pBBRlux-rfbT1-M3-2. In pBBRlux-rfbT1-M3-1, the conserved ATA and the antepenultimate A were changed to CAC and C, where TA and A are involved in the −10 motif of rfbT promoter. In pBBRlux-rfbT1-M3-2, the TC and second A were changed to GA and C. These new constructs were mobilized into V. cholerae WT and Δhns to measure the corresponding bioluminescence activities. As depicted in Figure 5B, pBBRlux-rfbT1-M1 and pBBRlux-rfbT1 had similar bioluminescence activities in the WT background which indicated that the mutation of sites 1 has no effect on the rfbT promoter, i.e., site 1 is indeed not a H-NS binding site. However, compared to the pBBRlux-rfbT1, pBBRlux-rfbT1-M3-2 displayed significantly high bioluminescence activity in WT as same as in the Δhns, implying that the introduced mutation affects H-NS binding and thus relieves its repression. Not surprisingly, pBBRlux-rfbT1-M3-1 almost completely lost the bioluminescence signal due to the mutations that destroyed the −10 motif of rfbT promoter by changing TAAAAT to ACACAT. Altogether, these results further experimentally validated that site 3 partially overlaps the −10 motif of rfbT promoter and is the intrinsically functional H-NS binding site.