Vibrio cholerae N16961, O1 biovar El Tor (NCBI, Taxonomy ID: 243277), was used as standard strain for V. cholerae O1 in this study (Heidelberg et al., 2000). Vibrio cholerae isolated in this study were grown in Luria-Bertani medium (LB) at 37°C with shaking and stored at −80°C as a glycerol stock solution.

During May 2012 to May 2017, environmental water sample was collected regularly a couple of times a month throughout the year from 3 ponds in the vicinity of Salt Lake of Kolkata, India. Contamination of filth into these ponds often occurred. The water in these ponds was never used for drinking, but was used for fish farming and washing. The environmental water was collected in sterilized glass sampling bottles (Shibata Scientific Technology LTD., Japan). A portion of the environmental water collected was centrifuged at 5,000 × g for 15 min at 25°C and pellet was recovered and suspended in 1/10 volume of sterilized phosphate-buffered saline (PBS). One hundred μL of bacterial suspension and the original environmental water were spread onto thiosulphate citrate bile-salt sucrose (TCBS) agar plates (Eiken Chemical Co., Japan). As it has been shown that some of V. cholerae taking vegetative form in environmental water enter into viable but non-culturable state (VBNC state) under certain conditions, though the conditions have not yet been elucidated (Senoh et al., 2014). The bacteria entered into VBNC state cannot grow on agar medium with normal composition, but can grow on the agar medium supplemented with catalase (Mizunoe et al., 2000; Imamura et al., 2015). So, we used TCBS agar medium supplemented with catalase (final concentration 100 U/mL) to detect V. cholerae entered into VBNC state.

We inoculated our water sample onto TCBS agar medium with and without catalase and cultured for 18 h at 37°C, all sucrose-fermenting yellow colonies yielded on these plates were picked up and inoculated onto enhanced selectivity (ES) vibrio agar plate (Eiken Chemical Co., Japan). After cultivation for 18 h at 37°C, typical V. cholerae-like bacteria grown in the plate were tested the possession of the gene of outer membrane protein W (ompW) of V. cholerae by PCR using specific primers listed in Table 1 (Nandi et al., 2000). The strains that exhibited DNA bands for ompW were regarded as V. cholerae. Then, the presence of the gene for A subunit of CT (ctxA) and the toxin-coregulated pilus A gene (tcpA) encoding subunit A of V. cholerae toxin-coregulated pili (TCP) were examined by PCR using the specific primers listed in Table 1. The nucleotide sequence of the 16S ribosomal RNA (rRNA) gene of the isolates used for further analyses in this experiment was determined to confirm the species of these bacteria.

The serogroup of the isolates was determined by immunoagglutination with specific antisera. For the determination of serogroups O1 and O139, we used commercially available antisera (Denkaseiken, Tokyo, Japan). Serogroups other than O1 and O139 were determined using 206 O group-specific serum supplied by National Institute of Infectious Diseases of Japan (Yamai et al., 1997).

Genomic DNA of the objective NAG strains was extracted from cells, cultivated overnight in LB medium at 37°C with shaking, using a DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany) in accordance with the manufacturer’s protocol. Illumina libraries of these DNA sequences were prepared using an Illumina DNA Prep Kit (Illumina, Inc. San Diego, CA, United States) and sequencing with HiSeq 2500 (Illumina) or MiSeq (Illumina) sequencers. De novo genome assemblies were constructed using CLC Genomics Workbench (Qiagen) and a multicontig draft genome was prepared. Nucleotide sequence data were submitted to the DDBJ Sequenced Read Archive (DRA), and each accession number is listed in Table 2.

The genes corresponding to each factor in the strains were obtained from their genomic sequences using Genetyx Mac Ver. 19 (Genetyx Corp, Tokyo, Japan). The nucleotide sequence of the reference gene of each factor was cited from GenBank. The accession number and locus tag used are shown in Table 3. Most of reference genes were cited from the genome of V. cholerae N16961 strain (Accession No. NC_002505 for chromosome I and NC_002506 for chromosome II) and other genes that were absent in the genome of V. cholerae N16961 strain were cited from each reference listed in Table 3.

To examine the ability of V. cholerae isolates to produce CT, we cultivated the strains in AKI medium (Iwanaga and Yamamoto, 1985). The strain was pre-cultured in LB medium overnight at 37°C with shaking, and 0.2 mL of the pre-culture was inoculated into 10 mL of AKI medium in a 250 mL flask. Two culture suspensions were prepared and cultured statically for 24 h at 25°C and at 37°C, respectively. After cultivation, the culture supernatant was recovered by centrifugation at 20,000 × g for 5 min at 4°C. The concentration of CT in culture supernatant was quantified using the GM1-ganglioside ELISA method reported by Kanjilal et al. with a slight modification (Kanjilal et al., 2010). Briefly, a 96-well microplate was coated with monosialoganglioside-GM1 (Sigma-Aldrich, St. Louis, MO, United States) (1 μg/mL in PBS containing 60 mM Na₂CO₃) by incubating at 37°C for 4 h. Then, the plate was washed three times with PBS containing 0.05% Tween-20 (PBST). For blocking, the wells were treated with 2% bovine serum albumin in PBS containing 60 mM Na₂CO₃ at 4°C overnight. After washing the wells three times with PBST, 100 μL of sample adequately diluted with PBS was applied to the wells and incubated for 2 h at room temperature. After washing the wells three times with PBST, the well was treated with a 1:20,000 dilution of rabbit anti-CT antibody (Sigma-Aldrich) for 1 h at room temperature. Then, the wells were washed three times with PBST and reacted with 1:5,000 dilution of anti-rabbit horseradish peroxidase-conjugated antibody (Abcam, Cambridge, United Kingdom) for 1 h at room temperature, followed by washing the wells three times with PBST. Subsequently, 100 μL of chromogenic tetramethylbenzidine substrate (Becton Dickinson, Franklin Lakes, NJ, United Kingdom) was applied to wells and the plate was kept at room temperature for a few minutes until color appeared. Then, the reaction was stopped by the addition of an equal volume of 3 M H₂SO₄. The optical density was measured at 450 nm. Commercially available CT (Sigma-Aldrich) was serially diluted (0.625–10.0 ng/mL) and used for making a calibration curve for CT in this assay.

Rabbit ileal loop assays were performed as described previously, using New Zealand White rabbits (weighing about 2 kg) (Koley et al., 1999). Animals were maintained in specific pathogen-free conditions in ventilated cages in the animal house facility of National Institute of Cholera and Enteric Diseases (NICED).

Rabbits were fasted for 24 h before treatment and were anesthetized with sodium pentobarbital. Then, the abdomen of the rabbits was incised and the intestines were exteriorized through a midline incision. After the intestinal lumen was rinsed three times with saline, the intestine was ligated in order to make ileal loops, into which the sample was injected.

Bacteria were cultivated in LB medium at 37°C for 18 h with shaking. The cells were harvested by centrifugation and suspended in an adequate volume of sterilized PBS to make a bacterial suspension adjusting to 2 × 10⁹ colony forming units (CFUs)/mL. The bacterial samples were mixed with an equal volume of either PBS or anti-CT antiserum or normal serum. One mL of these mixtures containing 1 × 10⁹ CFU was injected to the intestinal loop. After injection, the rabbits were kept for 18 h. Then, the rabbits were sacrificed under anesthesia and the intestinal ileal loops were recovered. The volume of accumulated fluid in the loop (mL) and the length of the ileal loop (cm) were measured, and the fluid accumulation ratio (FA ratio), which was the ratio of the volume of fluid accumulated in the intestinal loop to the length of the loop (mL/cm), was calculated.

Histopathological analysis of the intestinal tissues of the loops that were exposed to bacteria was performed. Each ileal intestinal loop separated from the body was immediately incised and put into 10% formalin neutral buffer solution. After incubating at room temperature for few days, the tissue sections were embedded in paraffin, and 2–3-μm thick sections made from paraffin-embedded intestinal segments were stained with hematoxylin and eosin. Stained sections were examined using light microscopy, and digital photographs of the stained tissues were taken.

The hemolytic activity of fluid accumulated in the ileal loop was examined using the lysis of sheep blood cells. The fluid accumulated in these loops was kept at -20°C until use. Just before the examination of hemolytic activity, the frozen samples were thawed and centrifuged 20, 000 × g at 4°C for 5 min to remove contaminating material. Fifty μL of the supernatant was transferred to a hole with a diameter of 5.0 mm that had been made in a sheep blood agar plate. After incubation at 37°C for 18 h, formation of a transparent zone around the well was examined.

Experiments except for animal experiments were carried out in triplicate, independently. The data are presented as arithmetic means ± standard deviation.

We collected water samples from ponds in Kolkata and cultured these on TCBS agar plates at 37°C for 18 h. Formation of yellow colony was observed in every samples collected. All yellow colonies formed on TCBS agar plate were inoculated onto ES vibrio agar plates, and the presence of ompW of V. cholerae in all colonies grown on ES vibrio agar was examined by PCR using the primers listed in Table 1. The number of colonies examined was more than 100,000. The band from ompW was detected in samples from more than 10,000 colonies. Subsequent searches for ctxA and tcpA by PCR showed that 5 strains possessed both ctxA and tcpA, and 32 strains possessed tcpA but not ctxA. Homology searches from nucleotide sequences of 16S rRNA gene showed that all 37 strains were V. cholerae.

The serogroup of O-antigen of the 5 strains possessing both ctxA and tcpA was determined. Three of the 5 strains did not react with O1 and O139 antisera, whereas the other two strains reacted with O1 antiserum. The former 3 strains were each isolated from separate samples. On the other hand, the latter two isolates were isolated from the same point on the same day. The two isolates appear to have come from the same patient. However, the patient from which these originated has not been identified.

Our interest was to examine the properties of NAG vibrio possessing genes related to the virulence of vibrios. We selected the 3 ctx-positive NAG vibrio strains that did not react with O1 and O139 antisera. We designated these 3 NAG vibrio strains as strains No. 1, 2, and 3, respectively. Then we determined the O-serogrop of 3 NAG vibrios using an immunoagglutination test. The serogroup of strains No. 1 and 2 were O124 and O152, respectively. However, it was impossible to determine the O-serogroup of strain No. 3 due to self-agglutination of the bacteria in the test (Table 2).

To examine the enterotoxicity of the NAG strains isolated, we carried out rabbit ileal loop assays. The bacteria that were examined in this assay were V. cholerae N16961 and four strains of NAG vibrios, strains No. 1, 2, 3, and A. Strain No. A, which we isolated from environmental water, does not possess both ctxA and tcpA (data not shown). One mL of bacterial suspension containing 1 × 10⁹ CFU/mL in PBS was injected into each loop of the intestine, and the accumulation of fluid in the loop was examined 18 h after injection of the bacterial suspension. PBS and strain N16961 were used as a negative control and a positive control in the assay, respectively. Strain A was used to get the information about the toxicity of NAG Vibrio which did not possess ctxA and tcpA.

As shown in Figure 1A, the fluid accumulated in loops administered with N16961. In loops administered with strain No. A and PBS, fluid accumulation was not observed. In the loop administered with strain No. 1, marked fluid accumulation was observed. In the loop administered with strain No. 2, a small but significant volume of fluid accumulated. On the other hand, in the loop administered with strain No. 3, no fluid accumulated.

Then, we collected the fluid accumulated and calculated the FA ratio. As shown in Figure 1B, the FA ratios obtained supported the results of the visual observations.

Subsequently, we investigated the contribution of CT to the enterotoxicity of strain No. 1. A bacterial suspension of strain No. 1 containing 2 × 10⁹CFU/mL in PBS was prepared. The bacterial suspension was mixed with an equal volume of serum and PBS. The composition of the mixtures prepared is described in the table in Figure 2, and 1 mL of each mixture was administered into the intestinal loop. The fluid accumulated in the loops was measured 18 h after administration. Fluid accumulated in every loop in which strain No. 1 was administered, although there was difference in the volume accumulated. The FA ratio of these loops was calculated (Figure 2). As shown in the results from samples 3 and 4, fluid accumulation by strain No. 1 was not reduced by the addition of normal serum. However, the FA ratio was reduced by the addition of anti-CT serum. The reduction was dose dependent on the volume of anti-CT serum added to the sample (samples 1 and 2 in Figure 2). When the bacterial suspension of strain No. 1 was mixed with an equal volume of anti-CT antiserum (sample No. 1), the FA ratio caused by the sample was 0.63, which meant that anti-CT serum sufficiently suppressed the enterotoxicity of strain No. 1. This result showed that CT is mainly responsible for the accumulation of fluid caused by strain No. 1.

To determine the productivity of CT from the isolated NAG strains, CT content in the culture medium was examined.

The three NAG strains isolated and V. cholerae N16961 were cultured in AKI medium for 24 h at 25°C and 37°C, and the concentration of CT in the culture supernatants was measured by GM1-ganglioside ELISA method using 96-well microplate coated with monosialoganglioside-GM1 as describe in the Materials and Methods. As shown in Figure 3, CT was detected in the culture supernatant of strain N16961, which was used as a CT-positive strain. The amounts of CT in the sample prepared by culturing at 25°C and 37°C was 48.4 ng/mL and 35.6 ng/mL, respectively. Similarly, CT was detected in the culture of strain No. 1 at 25°C. However, the CT level from strain No. 1 cultured at 37°C was very low (Figure 3). However, it is certain that CT is produced even at 37°C. Therefore, it is highly possible that CT is produced in the patient’s body and the CT produced causes diarrhea.

On the other hand, the amounts of CT produced by strains No. 2 and No. 3 were extremely low. Although a small amount of CT was detected in the supernatant of these strains cultured at 25°C, it was not detected in the sample cultured at 37°C (Figure 3).

We considered whether these NAG strains would be able to secrete CT from the cells, although they were capable of synthesizing CT. Therefore, we prepared cell extracts from the cells and measured the CT content in the cell extracts. The cells were collected from the cultures prepared for the experiment in Figure 3 by centrifugation at 15,000 × g for 5 min at 4°C and they were suspended in one tenth of the original amount of PBS, and the cell suspensions were disrupted by sonication, and the insoluble material was removed by centrifugation at 20,000 × g for 5 min. The recovered supernatant was used as the cell extract, and we measured the concentration of CT. No significant amount of CT was detected in the cell extracts prepared from the 4 strains (N16961 and strains No. 1, 2, and 3) (data not shown).

To elucidate the virulence properties of the isolated NAG vibrio strains, we determined genome sequences of these strains using next-generation sequencing methods. Genome sequences of the strains analyzed in this study were deposited in a public database. Accession numbers are listed in Table 2.

The sequences of ctx gene of these NAG vibrios were highly homologous with that of the reference strain, with an identity level of 99.7% or more (Table 3).

It was shown that the B subunit of CT (CTB) has been divided into 12 groups by differences in its sequence (Kim et al., 2015). The sequences determined showed that the type of CTB of strain No. 1 was ctxB1, and those of strains No. 2 and 3 were ctxB8 (Table 2).

The identity levels of the sequences of tcpA in these NAG strains were low, with a 76.7% homology level or less of with that of reference strain (Table 3).

Judging from the degree of homology of DNA sequence, it is considered that CT produced from these NAG vibrios has sufficient activity. However, since the homology of tcpA of isolated NAG vibrios with that of reference strain is in the 70% range, more studies are necessary to clear the activity of TcpA of these NAG vibrios isolated.

Several virulence factors are thought to be related to the toxicity of V. cholerae (Figueroa-Arredondo et al., 2001; Ramamurthy et al., 2020), therefore, we searched these genes in 3 NAG strains isolated (Table 3). These virulence factors (such as zot encoding zonula occludens toxin, ace encoding accessory cholera enterotoxin, hapA encoding hemagglutinin, and hlyA encoding pore forming cytotoxin) were well conserved in all NAG strains examined (Table 3). It has been revealed from comparative examinations of the sequences of these genes with their reference sequences that nonsense mutations and frameshift mutations were not present in these genes among the three strains examined.

On the contrary, the preservation rate of the genes for rtxA (multifunctional autoprocessing repeats-in-toxin), chxA (cholix toxin gene), and stn (heat-stable toxin (NAG-ST) gene) was low. chxA and stn were not conserved in the NAG strains examined. rtxA was conserved in strain No. 2 but not in the other two strains (Table 3).

The contribution of the type three secretion system (TTSS) to the pathogenicity of NAG vibrios has been reported (Luo et al., 2013; Zeb et al., 2019), although V. cholerae strain N16961 does not contain this system (Heidelberg et al., 2000). As shown in Table 3, genes encoding each protein involved in the TTSS were conserved in the genome of strains No. 1 and 2 but not in strain No. 3.

The presence of several regulator genes of ctx in the isolated NAG strains was examined. The regulatory genes examined were toxT, toxRS, tcpPH, and hapR. These gene products were reported to be involved in the expression of virulence genes, especially in the expression of ctx genes (Skorupski and Taylor, 1997; Ramamurthy et al., 2020). As shown in Table 3, all genes were highly conserved (98.0–100% homology) in the genome of all strains examined, and there were no nonsense mutation and no frameshift mutation in these genes.

The strain responsible for the 7th pandemic of cholera is V. cholerae O1 biotype El Tor. It is characteristic that the chromosome of V. cholerae O1 biotype El Tor consists of core and acquired genomes. The typical acquired genomes have four pathogenicity islands, (i) Vibrio pathogenicity island -1 (VPI-1), (ii) VPI-2, (iii) Vibrio seventh pandemic island-I (VSP-I), and (iv) VSP-II (Pant et al., 2020). We examined the presence of these pathogenicity islands in our isolated NAG vibrios.

The genes constituting VSP-I were not detected in any of the NAG vibrios examined (Table 4).

VSP-II consisted of the regions from VC0490 to VC0516 in the genome of V. cholerae N16961 chromosome I (Heidelberg et al., 2000). Homology analysis revealed that strains No. 2 and 3 do not contain any genes homologous with VSP-II. On the contrary, some regions that are homologous with VC0490–VC0498, VC0501, and VC0504–VC0510, are reserved in the genome of strain No. 1 (Table 4). However, the genes homologous to VC0499, VC0502–VC0503, and VC0512–VC0516, were not detected in the genome of strain No. 1 (Table 4). This showed that strain No. 1 possessed a VSP-II like element in which some loci constituting the canonical VSP-II (VSP-II of V. cholerae N16961) were deleted.

VPI-1 contains a gene for a crucial virulence factor of cholerae pathogens, toxin-coregulated pilus (TCP), in the loci from VC0824 to VC0837. These loci were maintained in all NAG strains examined (Table 4). All other loci constituting VPI-1 were present in strain No. 3, but the DNA sequences of VC0818 were disturbed in the genome of strain No. 1, and the DNA sequences of VC0818 and VC0846 in the genome of strain No. 2 were disturbed.

All loci constituting VPI-2 were conserved in strain No. 3, but more than half of the loci that constituted VPI-2, 29 loci of 48 loci of VPI-2, were not detected in strains No. 1 or 2 (Table 4).

We noticed that fluid caused by the administration of No. 1 strain in rabbit ileal loop assays was hemorrhagic red. Therefore, we assumed that strain No. 1 produced toxic factors to induce damage to intestinal tissue in vivo. It is known that hemolysin produced by V. cholerae often produces cytotoxic action (Mitra et al., 2000). Then, we investigated the hemolytic activity of the accumulated fluid using sheep blood agar plates. As shown in Figure 4, supernatant of fluid accumulated by strain No. 1 exhibited hemolysis, whereas those with N16961 and strain No. 2 did not.

Subsequently, we examined the intestinal tissues histopathologically from loops exposed to V. cholerae N16961 and strain No. 1 in the experiment for Figure 1. The examination showed that many villi were shortened or disappeared in the intestine that was administered with strain No. 1, whereas villi remained without incurring severe damage in intestinal tissue administered with N16961 (Figure 5).

Judging from these results, it was thought that cytotoxic hemolysin was produced by strain No. 1 in the intestinal lumen and the hemolysin produced might damage intestinal tissue, which resulted in the disappearance of villi.