Sequence Polymorphisms in Vibrio cholerae HapR Affect Biofilm Formation under Aerobic and Anaerobic Conditions

ABSTRACT

We investigated the influence of hapR sequence mutations on the biofilm formation of Vibrio cholerae. In this study, hapR sequences from 85 V. cholerae strains belonging to both pandemic and nonpandemic serogroup were investigated through phylogenetic and sequence analyses. Biofilm formation assays under aerobic and anaerobic conditions were also performed. Sequence variations include single point mutations and insertions/deletions (indels) leading to either truncated or frameshifted HapR. Population structure analysis revealed two major hapR haplogroups, hapR1 and hapR2. Phylogenetic reconstruction displayed a hypothetical ancestral hapR sequence located within the hapR1 haplogroup. Higher numbers of single nucleotide polymorphisms and genetic diversity indices were observed in hapR1, while indels occurred dominantly in hapR2. Aerobic conditions supported more robust biofilms compared to anaerobic conditions. Strains with frameshifted HapR produced the largest amount of biofilm under both oxygen conditions. Quantitative real-time PCR assay confirmed that strains with truncated and frameshifted HapR resulted in a nonfunctional regulator as exhibited by the significantly low hapA gene expression. The present study shows that HapR mutations had a strong influence on biofilm formation and that sequence polymorphisms leading to the disruption of DNA-binding sites or dimerization of the HapR will result in more-robust V. cholerae biofilms.

IMPORTANCE Our study revealed an ancestral hapR sequence from a phylogenetic reconstruction that displayed the evolutionary lineage of the nonpandemic to the pandemic strains. Here, we established hapR1 and hapR2 as major hapR haplogroups. The association of the O1 and O139 serogroups with the hapR2 haplogroup demonstrated the distinction of hapR2 in causing cholera infection. Moreover, mutations in this regulator that could lead to the disruption of transcription factor-binding sites or dimerization of the HapR can significantly affect the biofilm formation of V. cholerae. These observations on the relationship of the hapR polymorphism and V. cholerae biofilm formation will provide additional considerations for future biofilm studies and insights into the epidemiology of the pathogen that could ultimately help in the surveillance and mitigation of future cholera disease outbreaks.

INTRODUCTION

Vibrio cholerae lives and survives in aquatic environments in both freshwater and seawater ecosystems (1). To facilitate its survival in the natural environment and persist in different niches, V. cholerae forms biofilm to provide protection against different environmental conditions (2–6). Biofilm is a complex association of self-aggregating bacterial cells and extracellular matrix attached to a substrate or as floating aggregates (7, 8). In the natural environment, V. cholerae biofilms are frequently associated with the chitin exoskeletons of zooplankton (9, 10). V. cholerae biofilms on chitin surfaces also induce competence for natural transformation, which allows serogroup conversion in the aquatic environment (11). This could result to V. cholerae variants that are well suited to dynamic water environments and more pathogenic to humans. Moreover, biofilm-associated cells are more virulent than the planktonic phenotype, resulting in lower infectious doses to cause the infection (12). As such, better understanding of the biofilm formation in V. cholerae could assist in ongoing epidemiological studies of this pathogen.

Biofilm formation is influenced by a variety of environmental factors, including oxygen availability, salinity, temperature, and pH (13–16). For example, reduced biofilm formation under anoxia compared to atmospheric conditions was observed in uropathogenic Escherichia coli, Vibrio vulnificus, and Listeria monocytogenes (17–19). In contrast, anaerobic conditions supported higher biofilm formation for Vibrio parahaemolyticus (19). For Pseudomonas aeruginosa, the oxygen condition showed no significant influence on its biofilm formation (20). Throughout its life cycle, V. cholerae experiences fluctuations in its extracellular oxygen level during its transition from the gastrointestinal tract to the environment and vice versa (21). Since there are only limited biofilm studies on V. cholerae administered under anaerobic conditions, we investigated its biofilm formation under both aerobic and anerobic conditions.

Biofilm formation is facilitated by a process called quorum sensing system. The quorum-sensing system depends on the production and detection of extracellular signaling molecules known as autoinducers (AIs). In V. cholerae, four AI-receptor pairs are involved in overlooking the quorum-sensing system and consequently biofilm regulation (22). The well-studied AI-receptor pairs CAI-1/CqsS and AI-2/LuxPQ facilitate the phosphorylation of a response regulator, LuxO (23). The phosphorylated LuxO drives the expression of regulatory RNAs, which activates a low-cell-density master regulator, AphA. The activation of AphA initiates biofilm formation and virulence-related gene expression such as type IV pilus, accessory colonization factors, and cholera toxin (24–26). At high cell density (HCD), the hemagglutinin (HA) protease regulatory protein (HapR) is activated, AphA is repressed, and biofilm-associated cells shift into motile cells. Additional QS receptors, VpsS (27) and CsqR (28), operate similarly to the above-mentioned QS receptors; however, their cognate AIs and AI synthases are yet to be determined. Along the biofilm regulatory pathway, HapR is directly involved in the regulation of the secreted HA/protease, which acts as a “detachase” for effective colonization (29).

The HapR is a member of the TetR-family transcriptional regulators known to repress both virulence gene expression and biofilm formation of V. cholerae under HCD (30–32). HapR is a dimeric protein with 203 amino acid residues. The first three helices of the HapR monomer form a helix-turn-helix DNA-binding motif, while the six remaining helices are involved in the dimerization of the repressor (33). Sequence polymorphisms in the hapR gene have been reported to alter the functionality of the repressor, especially on biofilm regulation. For example, several toxigenic biofilm-proficient V. cholerae strains have either truncated or frameshifted HapR, which resulted in a nonfunctional HapR protein (31, 34). V. cholerae serovar O37 strain V2 has a point mutation that also yielded a nonfunctional HapR regulator (35). Due to the diversity of sequence polymorphisms occurring within the hapR coding region, other mutations could also provide additional information on the biofilm formation of the pathogen.

Phylogenetic analysis of V. cholerae hapR revealed two governing haplotypes: hapR1, which is mostly comprised of environmental non-O1/non-O139 serovars, and hapR2, which is associated with the pandemic O1 and O139 serogroups (36). The evolution of hapR2 from hapR1 was likely the result of strong natural selection, since it provides an advantage in the human host (37, 38). Pandemic V. cholerae strains can regulate the expression of virulence factors, such as cholera toxin and toxin coregulated pilus, during infection, and their prevalence in the aquatic environment was correlated with their resistance against bacteriophages (39). Whereas majority of the environmental non-O1/non-O139 strains are unable to express such genes (40). This hapR genetic structure and its association with the pathogenicity of V. cholerae have led us to investigate further the significant relationship of HapR and biofilm formation. In the present study, we focused on the relationship of HapR to the biofilm formation of V. cholerae, as well as the influence of hapR sequence polymorphisms and oxygen conditions on the biofilm formation of the pathogen.

RESULTS

Diverse sequence mutations occurred in the hapR of V. cholerae.

To identify the sequence variations in the hapR, we amplified and sequenced the target gene from 50 different V. cholerae strains. Sequence alignment revealed high similarity to a TetR regulator encoding a HA protein regulator, with 612-bp nucleotide length and 203 amino acid residues. We then identified the mutational events that occurred along the hapR gene sequence (Table 1). Transition mutations were widely distributed along the non-O1/non-O139 hapR sequences and frequently resulted in synonymous mutations. Relative to the hapR2 reference sequence (accession no. EU523710), 20 hapR2 strains, comprised mostly of O1 and O139 strains, did not exhibit any sequence polymorphism. Transversion mutations lead to either intact (569B, BO306) or truncated (25-6/9, SG21) HapR protein. Indel events resulted in frameshift mutations on the hapR sequences of P1418, CO840, BO302, 4-11/9, 15-9/9, 17-9/9, BO311, BO214, B4202-64, and 211-72. In special cases, a deletion of A¹⁷⁸CTCGTGAAGACT in BO320 has led to a premature termination of its hapR, which produced a truncated protein with 58 amino acid residues. In V. cholerae O139 MO14, a 12-bp interstitial sequence deletion starting from position 507 resulted in the deletion of the amino acid residues Ile¹⁷⁰FYS. This deletion did not produce any truncation or frameshift mutation when the nucleotide sequence was translated in silico.

TABLE 1.

Sequence polymorphisms in the V. cholerae hapR coding region^a

^aGray areas indicate complete linkage disequilibrium sites between hapR1 and hapR2 haplogroup. Numbers above the sequence corresponds to the nucleotide positions in the complete hapR coding region. Dots indicate sequence similarity with the reference sequence, while dashes indicate nucleotide deletion.

Next, we evaluated the degree of sequence variations that occurred on the hapR sequences of V. cholerae used in this study (Table 2). Single nucleotide polymorphism (SNP) analysis showed higher numbers of polymorphic sites in the non-O1/non-O139 serogroup, while indel events were found mostly in the O1 serogroup with 7 indel haplotypes (Indel_Hapn). Diversity indices were also analyzed, and the nonpandemic serogroup showed higher haplotype diversity (Hd; 0.912 ± 0.05) and nucleotide diversity (π; 0.0049).

TABLE 2.

Molecular diversity indices of the hapR coding region among the V. cholerae serogroups

Serogroup	SNP parametera
	n	Hapn	M	P		Hd (means ± the SD)	π (means ± the SD)	IndelHd	IndelHapn
				S1	Par
O1	26	6	510	4	3	0.450 ± 0.11	0.0016 ± 0.00	0.462	7
O139	8	3	598	2	0	0.417 ± 0.19	0.0007 ± 0.00	0.250	2
Non-O1/non-O139	14	8	586	3	7	0.912 ± 0.05	0.0049 ± 0.00	0.546	2
hapR	50	13	540	7	9	0.652 ± 0.07	0.0040 ± 0.01	0.425	12

^an, number of sequences; Hap_n, number of haplotypes; M, number of monomorphic sites; P, number of polymorphic sites; S1, number of singletons; Par, number of parsimony informative sites; Hd, haplotype diversity; π, nucleotide diversity; Indel_Hd, Indel haplotype diversity; Indel_Hapn, number of Indel haplotypes.

Phylogenetic and network analyses revealed a hypothetical ancestral hapR sequence of V. cholerae.

To better understand the population structure and phylogeny of V. cholerae hapR, we constructed a maximum-parsimony phylogenetic tree (Fig. 1A) and then performed a Bayesian analysis of population structure (BAPS) (Fig. 1B) using the sequences from our present study (n = 48) and sequences retrieved from NCBI GenBank (n = 37). In reference to the previous report (36), we then regarded hapR1 and hapR2 as major hapR haplogroups and not as haplotypes, as confirmed by BAPS. Among the 50 strains that were used in the present study, we identified 17 strains in the hapR1 haplogroup; these include three environmental O1 (KV8, KV28, and KV41) and 14 non-O1/non-O139 serovars, whereas 33 O1 and O139 strains belonged to the hapR2 haplogroup. Consequently, we performed a similar SNP analysis with the sequences and have found 29 unique haplotypes (Table 3). The diversity indices, haplotype diversity (Hd; 0.906 ± 0.03) and nucleotide diversity (π; 0.0046 ± 0.00), were highest in hapR1 compared to hapR2. Tajima’s D and Fu_FST values were all negative for the two haplogroups, indicating an excess of unique nucleotide sequence variants compared to the expected neutral model of evolution. As such, these values suggest that the hypothesis of neutral evolution was significantly rejected for the hapR2 haplogroup (Tajima’s D = –2.4506, P = 0.0000; FU_FST = –22.8094, P = 0.000) and not for the hapR1 haplogroup (Tajima’s D = –1.0571, P = 0.1480; FU_FST = –2.3601, P = 0.1881).

FIG 1.

(A) Maximum parsimony phylogeny of the hapR protein-coding region. Sequences were aligned by ClustalW. Bootstrap values were calculated from 1,000 replicates and are indicated next to the branches. Phylogenetic analysis showed two major haplogroups from 85 hapR sequences: hapR1 (shown in the blue region) and hapR2 (shown in the red region). The sequences of the strains used in this study are indicated in boldface. (B) Bayesian analysis of population structure (BAPS) based on the coding region of the hapR gene. The colors represent the different hapR haplogroups.

TABLE 3.

Molecular diversity indices of the hapR coding region between the V. cholerae hapR haplogroups

Haplogroup	SNP parametera
	n	Hapn	M	P		Hd (means ± the SD)	π (means ± the SD)	IndelHd	IndelHapn	Tajima’s D	FuFST
				S1	Par
hapR1	30	14	596	6	9	0.906 ± 0.03	0.0046 ± 0.00	0.251	4	–1.057ns	–2.451ns
hapR2	55	16	548	12	3	0.525 ± 0.08	0.0012 ± 0.00	0.420	13	–2.360s	–22.809s
hapR	85	29	518	14	15	0.780 ± 0.05	0.0046 ± 0.00	0.361	16	–1.897s	–25.715s

^an, number of sequences; Hap_n, number of haplotypes; M, number of monomorphic sites; P, number of polymorphic sites; S1, number of singletons; Par, number of parsimony informative sites; Hd, haplotype diversity; π, nucleotide diversity; Indel_Hd, Indel haplotype diversity; Indel_Hapn, number of Indel haplotypes; ns, not significant; s, significant (Tajima: ns if P > 0.10; Fu_FST: ns if P > 0.02).

To further assess the relationship among the hapR haplotypes, we performed a genetic reconstruction of the sequences through a haplotype network employing a median-joining (MJ) criterion (Fig. 2). The clustering of the haplogroups were replicated in both the MP tree and the haplotype network (Fig. 1). The two occurring median vectors (small red circles) were considered unsampled hapR sequences. The hapR1 haplogroup showed a reticulated network, while a star-like pattern was observed for hapR2. Interestingly, we identified a possible hapR ancestral haplotype found in haplotype 20 (H20). The associated haplotypes within the H20 cluster contained the conserved T²⁴⁰ and T²⁴³ nucleotide positions (Table 1). Next, an intermediate haplotype H24 between H20 cluster and H16 was also observed with a transition mutation of T²⁴⁰ to C. A further transition mutation occurred from H24 to H16 via T²⁴³ to C. The strains belonging to this cluster could be the more distant relatives of the hapR2 haplogroup since environmental strains of O1 serogroup (KV1, KV28, and KV41) were in H4. Between H16 and H1 (hapR2 core haplotype), the transition of T⁵⁷² to C and the transition mutation of A⁴⁵⁰ to G were regarded as the linkage disequilibrium between the two haplogroups (36), giving rise to an independently diverged hapR2 haplogroup from hapR1.

FIG 2.

Haplotype network of 85 V. cholerae hapR sequences constructed using median-joining algorithm. The colors of the nodes represent the different serogroups, the sizes of the nodes represent the numbers of samples, and the dashes on the branches show the numbers of mutations between nodes. The two occurring median vectors (small red circles) were considered unsampled hapR sequences.

Oxygen limitation affects biofilm formation of V. cholerae.

We then investigated the influence of environmental oxygen on the biofilm formation of V. cholerae by performing a static biofilm assay under aerobic and anaerobic conditions. When biofilms were developed under aerobic condition, biofilm formation was higher compared in anaerobic condition (P = 1.670E–17, r = –0.695) (Fig. 3). The largest amount of biofilm under an aerobic condition was observed among strains O1 El Tor P1418 and 19-8/9, O139 BO312, and non-O1/non-O139 strains 211-72, BO320, and 7007-62. In contrast, the least biofilms were exhibited by the O1 strains 569B, 124, 35A3, KV8, KV41, and BO301 and O139 strains MO2, MDO6, and VO10. When the strains were cultured under anoxia, biofilm formation was only evident in some strains. Particularly, high biofilm producers were the non-O1/non-O139 strains SG11 and 7007-62. Several O1 strains, such as P1418, BO302, and CO840 and the O139 strains BO311 and MO20 also formed relatively high levels of biofilm. Biofilm formation under anaerobic condition was not evident in approximately 50% of the biofilm-forming strains under normoxia.

FIG 3.

Box plot representation of V. cholerae biofilm formation under aerobic and anaerobic conditions. The P value was considered significant by the Wilcoxon signed rank test. The black dots indicate the average biofilm formation of V. cholerae strains (N = 50). Data means are indicated by red asterisks.

The reduced biofilms under anoxia were speculated to be influenced by oxygen limitation and not by growth changes. To verify this assumption, we performed a growth kinetic analysis of each strain under both oxygen conditions, and this was followed by a correlation analysis between biofilm formation and the growth kinetic parameters, lag time (t_lag), growth rate (μ), and final optical density (y_max) (Table 4). Based on the correlation analysis under both oxygen conditions, no significant relationship with the biofilm formation and growth kinetic parameters was observed. As such, the reduced biofilm of V. cholerae strains under anoxia is likely caused by the anoxic condition more than the growth of the pathogen.

TABLE 4.

Spearman correlation analysis between V. cholerae biofilm formation and growth kinetic parameters

Oxygen condition	Parametera	Growth kinetic parameter
		Lag time (tlag)	Growth rate (μ)	Final OD (ymax)
Aerobic	rs	0.127	0.182	–0.164
	P	0.375	0.207	0.256
Anaerobic	rs	0.291	–0.216	0.191
	P	0.040b	0.131	0.184

^ar_s, Spearman correlation coefficient (P).

^bP < 0.05.

hapR sequence mutations significantly influence V. cholerae biofilm formation.

In V. cholerae under HCD, hapR negatively regulates biofilm formation by repression of vpsT expression, reduction of intracellular c-di-GMP concentration, and activation of cell dispersal from the biofilms (41). Initially, we compared the biofilm formation of V. cholerae between hapR haplogroups and among serogroups under both oxygen conditions. Between hapR haplogroups, no significant difference was found. Among serogroups, V. cholerae non-O1/non-O139 exhibited higher biofilms under normoxia (H = 8.631, P = 0.013) and anoxia (H = 15.857, P = 1.44E–4) compared to the pandemic serogroups, O1 and O139. A medium effect of the V. cholerae serogroup on biofilm formation under aerobic (η² = 0.063) and anaerobic (η² = 0.096) conditions was observed. Previously, sequence mutations in hapR were previously reported to significantly influence biofilm formation (27, 30, 31). Hence, we further analyzed the relationship of these mutations on the degree of biofilm formation under aerobic and anaerobic conditions. Frameshifted HapR mutants, except CO840, regardless of the hapR haplogroup produced the highest levels of biofilm under aerobic (H = 29.281, P = 2.00E–6, η² = 0.232) and anaerobic (H = 34.369, P = 1.66E–7, η² = 0.195) conditions. Terminal deletion mutants also formed robust biofilms under aerobic conditions, but with negligible to minimal biofilm under anaerobic conditions. Strain MO14, which had an interstitial deletion on its hapR sequence, also produced a degree of biofilm similar the the frameshift and terminal deletion mutants. Some strains, such as 14-9/9, 19-8/9, MO20, and BO312, which have an intact hapR2 sequence also produced robust biofilms. Frameshift mutations led to more biofilm formation under both oxygen conditions than unaltered hapR2 and point mutations (Fig. 4). Although high biofilm formation was exhibited by terminal deletion mutants under aerobic conditions, their low biofilm formation under anoxia could be a result of an intermediate phenotype. However, our data are not sufficient to confirm this speculation. Based on the effect size, we observed that hapR mutations have a stronger influence on the biofilm formation of V. cholerae than the strain’s hapR haplogroup or serogroup. We also examined the interaction between the hapR mutation and oxygen conditions on the biofilm formation of V. cholerae. There was an observed significant interaction between the effects of hapR mutation and oxygen condition (F = 2.794, P = 0.041); however, the effect was small (η² = 0.029) compared to the significant effects of either hapR mutation (F = 23.697, P = 1.07E–13, η² = 0.202) or oxygen condition (F = 49.815, P = 1.33E–11, η² = 0.151). As such, hapR mutations have a more significant effect under aerobic than under anaerobic conditions.

FIG 4.

Box plot representation of the biofilm formation of V. cholerae HapR variants under aerobic (A) and anaerobic (B) conditions. A Kruskal-Wallis test with pairwise comparison was used to determine significant differences between groups. The P values of the pairwise comparison are indicated above the box plots. The black dots indicate the average biofilm formation of each V. cholerae strain. Data means are designated by red asterisks.

To verify whether these sequence mutations affect the functionality of the HapR regulator, we then performed a real-time PCR assay on the hapA gene, a gene encoding the HA protease, which is directly regulated by HapR (31). Based on qRT-PCR analyses, we observed that HapR variants that produced robust biofilms exhibited low hapA expression. Higher hapA expression were exhibited by unaltered hapR2 and point mutations with intact HapR strains (H = 59.476, P = 7.61E–13, η² = 0.146) and significantly low hapA expression in terminal deletion and frameshift mutants (Fig. 5). Results from this experiment confirmed that HapR variants with a terminal deletion, interstitial deletion, or frameshift mutation, along with some single point mutants, will possess a nonfunctional HapR regulator.

FIG 5.

(A) Box plot representation of hapA gene expression of V. cholerae HapR variants. (B) Scatterplot between hapA gene expression and biofilm formation. The Kruskal-Wallis test with pairwise comparison was used to determine significant differences between groups. The P values of the pairwise comparisons are indicated above the box plots. The black dots indicate the average biofilm formation of each V. cholerae strain. Data means are indicated by red asterisks.

DISCUSSION

The hapR is a known repressor of biofilm formation in V. cholerae, particularly in the negative regulation of the vps (Vibrio polysaccharides) operon (31). Here, we investigated the hapR sequence mutation and its affect on the biofilm formation of 50 V. cholerae strains. As previously reported, we confirmed two major hapR haplogroups in V. cholerae that distinguish between the pandemic strains, O1 and O139 serogroups, and the nonpandemic strains (36). The pandemic hapR2 haplogroup regulates genes for cholera toxin, a colonization factor toxin-coregulated pilus, and a regulatory protein ToxR, which are not found in most non-O1/non-O139 V. cholerae strains (42).

We further evaluated the genetic phylogeny of hapR through a haplotype network. The hapR1 haplogroup showed a network with reticulation, while a star-like pattern was observed in hapR2. Within the hapR2 network, a star-like pattern accompanied by high haplotype diversity and low nucleotide diversity was observed. This structure indicates low levels of sequence divergence and a high frequency of unique mutations, such as indels, because of a rapid population expansion by the pandemic haplogroup.

The evolution of hapR2 haplogroup from hapR1 was a result of a recent selective sweep that resulted in the emergence of the pandemic strains (36). Based on our phylogenetic reconstruction, we have identified a hypothetical ancestral hapR sequence within the hapR1 cluster. The associated haplotypes of the ancestral H20 cluster contain the conserved T²⁴⁰ and T²⁴³ nucleotide positions (Table 1). The initial sequence mutation occurred at T²⁴⁰ to C, followed by T²⁴³ to C. These nucleotide positions were previously identified as part of a putative ligand-bonding socket in the HapR dimer (33). Although the mechanism for how the succeeding mutations in these positions could affect the function of the regulator has not been elucidated, we have speculated that the mutations in the ligand-binding socket was an initial step in the transition of hapR1 to hapR2. Next, the adjacent hapR1 haplotype H4 was regarded as the more distant relatives of the hapR2 haplogroup since environmental strains of O1 serovars (KV1, KV28, and KV41) clustered in this haplotype. Between H16 and hapR2 dominating haplotype H1, the transitions of A⁴⁵⁶ to G (Glu¹⁵⁰) and T⁵⁷⁸ to C (Val¹⁹¹ to Ala¹⁹¹) were the linkage disequilibrium between the two haplogroups (36), giving rise to an independently diverged hapR2 haplogroup. The hapR1 nucleotide sequences in these positions are similar to V. harveyii LuxR, a homolog of HapR. There is more sequence similarity with hapR1 and LuxR than with hapR2. Glu¹⁵⁰ was conserved among the TetR regulators; meanwhile, the mutation to Ala¹⁹¹ was only observed in hapR2. This residue functions as a dimer interface, and mutations in such positions could have affected the HapR dimer formation or enzyme activity.

In V. cholerae, sequence variations that include SNPs and indels were located in the hapR coding region. Sequence variations are frequently observed in growth-associated genes in response to environmental conditions (43); as in this study, polymorphisms in the quorum-sensing regulator HapR could be result of both environmental and pathogen-host interactions. We then categorized the V. cholerae into strains with (i) unaltered hapR2; (ii) point mutations, which includes both transition and transversion, with intact hapR; (iii) terminal deletions; and (iv) frameshift mutations. Strains with intact HapR, regardless of mutation, showed the least biofilm under aerobic and anerobic condition, whereas terminal deletion and frameshift mutants produced robust biofilms under both oxygen conditions. Moreover, some strains that showing rugose colony phenotype were robust biofilm formers. Phase variation was previously reported in V. cholerae, wherein rugose variants produce more biofilm than their smooth counterpart due to the overproduction of extracellular polysaccharides during biofilm formation (44). A transition variant V. cholerae O137 strain V2, isolated from Kolkata, India, has a similar transition mutation with MO45 (31). This mutation caused the carboxylate side chain of Asp³⁹ to electrostatically repel the phosphate backbone of DNA, thus preventing the DNA binding of the HapR (45). The interstitial deletion of the amino acid residues I¹⁷⁰FYS in MO14, located in the α8 helix, could also reveal another type of mutation that affects the functionality of the HapR. This deleted region contains (i) the Phe¹⁷¹ residue acting as a putative ligand-binding pocket and (ii) Tyr¹⁷² forming a stacking interaction in the helix α8, stabilizing the hydrophobic contacts in helices α8 and α9, and thus affecting the dimerization of the HapR (33). Both MO45 and MO14 exhibited robust biofilms and low hapA gene expression, indicating that such mutations lead to a nonfunctional biofilm repressor. Indels in the hapR sequence have resulted in frameshift mutations producing nonfunctional HapR protein (34, 46). Strains with such mutations in HapR showed a high degree of biofilm formation under both aerobic and anaerobic conditions. These mutations were frequently observed in the hapR2 haplogroup and are speculated to provide an advantage for environmental persistence and host pathogenicity.

Our study has confirmed and established two major hapR haplogroups, hapR1 and hapR2. hapR1 is dominated by environmental strains, while hapR2 is closely associated with the pandemic cholera strains. Furthermore, an ancestral hapR sequence was identified within the hapR1 haplogroup. From this ancestral haplotype, four mutational events occurred and resulted in the emergence of the pandemic hapR2 haplogroup. However, the specific functions of these SNP sites on HapR function and their significance for hapR2 evolution are yet to be elucidated. We also observed that HapR mutations have a stronger influence on biofilm formation compared to the V. cholerae serogroup or the hapR haplogroup. Mutations resulting in either disruption of the transcription factor-binding sites or dimerization of the HapR will result in a nonfunctional biofilm repressor. These mutations will allow V. cholerae to regulate robust biofilms under both aerobic and anaerobic conditions. As such, future V. cholerae biofilm research should consider the HapR polymorphism of the strain prior to experimentation.

MATERIALS AND METHODS

Bacterial strains, growth conditions, and growth kinetic analysis.

We investigated a total of 50 V. cholerae strains belonging to O1, O139, and non-O1/non-O139 serogroups (Table 5). The strains were retrieved from −80°C glycerol stock cultures and subsequently grown in Luria-Bertani (LB) medium (LB broth/agar; Nacalai Tesque, Japan) at 37°C.

TABLE 5.

V. cholerae strains used in the study

Strain	Location, yr of isolation	Biotype and serogroup	Reference(s)	Accession no.
569B	India, 1948	Classical O1, clinical	57	EU523721
124		Classical O1	58	MW506237
35A3	India, early 1940s	Classical O1, Inaba	59, 60	MW506238
P1418		El tor, O1, Ogawa	61, 62	MW670444
CO840	India, 1995	El tor o1, Ogawa	63, 64	MW506239
KV8	Japan, 2002	O1, environmental	This study	MW506240
KV28	Japan, 2002	O1, environmental	This study	MW506241
KV41	Japan, 2002	O1, environmental	This study	MW506242
BO301	Bangladesh, 2002	O1, environmental	This study	MW506243
BO302	Bangladesh, 2003	O1, environmental	This study	MW506244
BO303	Bangladesh, 2003	O1, environmental	This study	MW506245
BO304	Bangladesh, 2003	O1, environmental	This study	MW506246
BO305	Bangladesh, 2003	O1, environmental	This study	MW506247
BO306	Bangladesh, 2003	O1, environmental	This study	MW506248
BO307	Bangladesh, 2003	O1, environmental	This study	MW506249
BO308	Bangladesh, 2003	O1, environmental	This study	MW506250
BO309	Bangladesh, 2003	O1, environmental	This study	MW506251
BO310	Bangladesh, 2003	O1, environmental	This study	MW506252
1-11/9	India, 2003	O1, clinical	This study	MW506253
4-11/9	India, 2003	O1, clinical	This study	MW506254
14-9/9	India, 2003	O1, clinical	This study	MW506255
15-9/9	India, 2003	O1, clinical	This study	MW506256
17-9/9	India, 2003	O1, clinical	This study	MW506257
19-8/9	India, 2003	O1, clinical	This study	MW506258
25-6/9	India, 2003	O1, clinical	This study	MW506259
26-6/9	India, 2003	O1, clinical	This study	MW506260
33-6/9	India, 2003	O1, clinical	This study	MW506261
MO2	India, 1992	O139, clinical	65, 66	MW506262
MO14	India, 1992	O139, clinical	67	MW506263
MO20	India, 1992	O139, clinical	67	MW506264
MO45	India, 1992	O139, clinical	65, 66	MW506265
SG21	India, 1992	O139	64, 67	EU523725
MDO6	India, 1992	O139, clinical	67, 68	MW506266
VO10	India, 1992	O139, clinical	67	MW506267
BO311	Bangladesh, 2002	O139, environmental	This study	MW670445
BO312	Bangladesh, 2003	O139, environmental	This study	MW506268
BO214	Bangladesh, 2002	O5, clinical	This study	MW506269
B4202-64	Philippines, 1964	O5, clinical	69, 70	MW506270
7007-62	India, 1962	O6, clinical	70 – 72	MW506271
8394-62	Philippines, 1962	O7, clinical	70, 71	MW506272
112-68	Philippines, 1968	O9, clinical	70, 71	MW506273
10843-62	India, 1962	O11, clinical	70, 71	MW506274
211-72	India, 1972	O12, clinical	70	MW506275
VTE1235	India, 2002	O18, clinical	This study	MW506276
BO319	Bangladesh, 2003	O37, environmental	This study	MW506277
BO320	Bangladesh, 2003	O39, environmental	This study	MW506278
H-08942	India, 2002	O47, environmental	This study	MW506279
SG11		O69, environmental	73	MW506280
BO313	Bangladesh, 2003	O93, environmental	This study	MW506281
BO209	Bangladesh, 2002	O126, environmental	This study	MW506282

The individual growth curves of the strains under both aerobic and anaerobic conditions were also investigated. Overnight broth cultures were inoculated (1:100) in fresh LB broth. For the anaerobic growth condition, 50 mM filter-sterilized (20-μm cellulose acetate filter; Advantec, Japan) trimethylamine N-oxide (TMAO; 95%; Sigma-Aldrich, USA) were supplemented with LB medium (LB_TMAO). Next, 150-μL aliquots were transferred to a flat-bottom polystyrene microplate (Falcon 96-well plate; Corning, Japan). To facilitate the anaerobic growth, 100 μL of mineral oil was overlaid on the LB_TMAO. Th optical density at 600 nm (OD₆₀₀) was then measured every 20 min for at least 20 h (until mid-stationary phase). The growth curve was then plotted using a Dynamic Modeling Fit (DMFit 3.5) MS Excel add-in, which was based on a previously developed growth model (47). The growth kinetic parameters—lag time (t_lag), growth rate (μ), and final OD (y_max)—were also calculated from this model. Late-stationary-phase time points were excluded from the analysis since they distorted the model fit.

hapR gene amplification.

Genomic bacterial DNA was extracted using the boiled lysate method (48). A 0.8-kbp region containing the hapR gene was amplified (Table 6). The PCR conditions include an initial denaturation at 98°C in 1 min, followed by 30 cycles of DNA denaturation at 98°C for 10 s, annealing at 56°C for 15 s, and primer extension at 68°C for 30 s, with a final elongation at 68°C for 15 s. The PCR products were then cleaned (ExoSAP-IT; Thermo Fisher Scientific, Japan) and processed for nucleotide sequencing by Eurofins Genomics (Japan). The DNA sequences were viewed and trimmed using 4Peaks software v1.8 (Mekentosj, Amsterdam, Netherlands) and submitted to GenBank/DDBJ/ENA for the accession numbers.

TABLE 6.

Primers used in this study^a

Target gene	Primer	Sequence (5′–3′)
hapR	HapR-F1	ACC ATT ACA CTC ATA GGG CT
	HapR-R1	CAA CAC CAA GTC GTT TAG GT
hapA	hapA-qF	CAC CTT ACC ATT CGG CAA CTG
	hapA-qR	CGC AGC AAC AGT ACA GTA TG
gyrB	gyrB-F1	GGA TTG GCT GAT CAA AGA GTC G
	gyrB-R1	TCC ATC GTA GTT TCC CAC AGC

^aAll primers were from the present study.

Nucleotide sequence polymorphism analysis.

Next, we analyzed the level of hapR sequence polymorphisms among the different V. cholerae serogroups. The hapR sequences of the strains (n = 50) investigated in the study (n = 50) were aligned by ClustalW in MEGA software v10.1.8 (49). SNP parameters, which include singletons and parsimony informative sites, and insertion-deletion events were enumerated along the hapR coding region. Estimates of diversity indices—nucleotide diversity (π), haplotype diversity (Hd), and Indel haplotype diversity (Indel_Hd)—were also determined. The software DnaSP (v6.12.03) was used to calculate the SNP parameters and diversity indices from the sequence data (50).

Phylogenetic and network analysis.

To determine the genetic structure and phylogeny of the V. cholerae hapR, we aligned 85 hapR sequences retrieved from this study (n = 48) and from NCBI GenBank (n = 37), as mentioned above. The major haplogroup of the strains was initially identified by constructing a maximum parsimony (MP) phylogenetic tree with 1,000 bootstrap replicates (49). The population genetic structure was also investigated using the BAPS 7.13 program employing an “analysis of genetic mixture with linked loci or sequences” (51, 52), which further confirmed the haplogroup designations of the investigated strains. SNPs and genetic diversity indices were also calculated using DnaSP software (50). The same software was then used to determine the haplotype profile of the aligned sequences. Furthermore, the neutrality tests for the hapR haplogroups, Tajima’s D and Fu’s FS, were calculated by using Arlequin software (v3.5.2.2) (53). Thereafter, the relationships among the hapR haplotypes were further characterized by constructing a haplotype network using the median-joining algorithm in the PopArt 1.7 software (54).

Biofilm formation assay.

We implemented a biofilm formation assay for the 50 V. cholerae strains using the crystal violet staining method (55). Overnight bacterial cultures were diluted (1:100) in LB broth medium. Then, 200-μL aliquots were then transferred into a V-bottom 96-well polystyrene microplate (Micro test plate 96-well; Nerbe Plus, Germany), covered with a lid, and sealed with parafilm to prevent evaporation. The plates were then incubated under static conditions at 30°C for 24 h. Planktonic cells were removed, and wells were washed with distilled water and dried for 10 min. Biofilm staining was administered using 0.1% aqueous crystal violet solution for 15 min. After washing, the bound CV stain was solubilized with 95% ethanol. Thereafter, 150 μL of the solubilized stain was transferred to a flat-bottom polystyrene microplate, and the relative biofilm was quantified at 570 nm using a microplate reader (Multiskan Sky; Thermo Scientific, Finland).

For anaerobic biofilm formation, LB broth was supplemented with filter-sterilized TMAO to a final concentration of 50 mM (16). A biofilm assay was performed as described above with microplates placed in an anaerobic growth chamber with a gas pack (Anaero Pack; Mitsubishi Gas Chemical, Japan) to provide an anaerobic culture environment.

RNA extraction and RT-qPCR.

To investigate whether the observed mutations that occurred in hapR resulted in inactivation of the regulator, we performed RT-qPCR on the hapA gene that is directly regulated by HapR. Strains were cultured in LB broth until early stationary phase. Total RNA was extracted using the TRIzol protocol (TRIzol reagent; Invitrogen, USA) according to the manufacturer’s instruction. RNA extracts were then subjected for DNase treatment (DNase I recombinant; Roche, Sigma-Aldrich, Germany). RNA concentration and quality were checked by NanoDrop (BioSpec-nano; Shimadzu, Japan). cDNA was then prepared from 2 ng of DNase-treated RNA using a ReverTra Ace cDNA synthesis kit (Toyobo, Japan). A relative standard quantitation was then performed using SYBR green PCR kit (TB Green Premix Ex Taq II [Tli RNase H Plus]; TaKaRa, Japan) in a StepOne Plus real-time PCR system (Applied Biosystems, Japan) with cDNA (1:10 dilution) as the template. The gyrB gene was used as the endogenous control, and the hapA expression was determined by relative standard quantification (Table 6).

Statistical analysis.

Statistical analysis for each assay was performed using the averages of three independent runs. A Wilcoxon signed rank test was employed to compare the biofilm formation between hapR haplogroups. Mann-Whitney U test was used for comparing the biofilm formation of haplogroups between aerobic and anaerobic conditions. The effects of sizes for the signed rank test and Mann-Whitney test were determined based on the r value (56). Significant differences in biofilm formation among serogroups and HapR mutants or hapA gene expression among HapR mutants were determined using a Kruskal-Wallis test set at a confidence interval of 95%. In case of significant differences, a pairwise comparison was performed. Eta-squared (η²) was used to determine the effect size of this test. To determine the interactions between hapR mutations and oxygen conditions on the biofilm formation, a two-way analysis of variance was employed. For biofilm and growth correlation statistics, bivariate correlation analysis using the Spearman correlation statistic was utilized. All statistical computations were done in SPSS v.27 (IBM, USA). Biofilm formation and growth curve experiment assays were performed at least three times with three biological replicates. For RT-qPCR, three independent experiments with two biological replicates were done.

Data availability.

All hapR sequence data in this study were registered at GenBank/DDBJ/ENA, as shown in Table 1. Source data for Fig. 3 to 5 and Table 4 are provided.