All chemicals were of analytical grade and obtained from Sigma-Aldrich Chemicals (St. Louis, MO, United States). Thiosulfate-Citrate Bile salt Sucrose agar (TCBS) was from Eiken (Japan) and Luria Bertani (LB) medium was from Difco (United States).

Representative Vibrio cholerae O1 strains El Tor (V100 and V114; isolated during 1990) and Haitian variants (IDH00990 and IDH02003; isolated during 2008) from the strain repository of ‘National Institute of Cholera and Enteric Diseases’ (NICED), Kolkata, India, were used in this study. These strains were isolated from the cholera patients admitted to the Infectious Diseases and Beliaghata General (ID&BG) Hospital, Kolkata, India. Detailed phenotypic and genetic characteristics of these strains are presented in Table 1. V. cholerae strains were grown in Luria-Bertani (LB) medium containing 1% NaCl at 37°C with shaking. Strains were stored at −80°C in LB containing 20% glycerol (V/V).

Medial growth curve analysis was performed with canonical El Tor and Haitian variant strains. Single colony was inoculated in 5 mL of LB broth and grown overnight at 37°C under shaking condition. Overnight cultures were sub-cultured in fresh 25 mL LB broth in a sterile glass conical flask at a dilution of 10³. The cultures were further incubated at 37°C and the O.D₆₀₀ was measured at every 1 h time interval up to 14 h (Medrano et al., 1999; Chatterjee et al., 2017).

New Zealand white rabbits of either sex weighing 2.0–2.5 kg, 4 to 5 days-old suckling BALB/c mice and 4–5 weeks adult BALB/c mice were used in this study (Table 2). All animals were maintained under specific pathogen free condition in ventilated cages in the Animal House facility of NICED and provided with sterilized food and water.

Animal experiments were conducted following the standard operating procedure framed by Committee for the purpose of Control and Supervision of Experiments on Animal (CPCSEA) (Approval No. PRO/111/Nov2014-Dec2016), Ministry of environment and forest, Government of India. The animal experimental protocol was approved by the Institutional Animal Ethics Committee of NICED.

Overnight cultures of V. cholerae strains were diluted 1:1000 in fresh LB medium supplemented with 100 μg of streptomycin/ml (LBS) and grown at 37°C with shaking to mid-log phase. Bacterial cultures were centrifuged, the pellets washed twice with ice cold phosphate-buffered saline (PBS), and then suspended in cold PBS (1 × 10⁹ CFU/50 μl).

Colonization assay was performed in suckling mouse model with the modified protocol described previously (Camilli and Mekalanos, 1995; Angelichio et al., 1999). Briefly, 4 to 5 day-old suckling BALB/c mice were separated from their mothers 1 h prior to inoculation with V. cholerae. Each mouse was then intragastrically inoculated with 50 μl of the diluted bacterial culture. Negative control mice were fed with 50 μl of PBS only. The bacterial titer in each inoculum was determined by plating serial dilutions on LBS agar plates. Infected mice were kept at 26°C separated from their mother. These were sacrificed 18 h after infection by cervical dislocations under isoflurane. Their intestines were then removed and mechanically homogenized in cold PBS (100 μl) using a tissue homogenizer at 4°C (POLYTRON PT1600E, Kinematica AG). Serial dilutions of the homogenates were then plated onto LBS agar to enumerate viable V. cholerae cells and expressed in log of Colony Forming Unit/gram (CFU/gram) of the intestine. Values were calculated as mean ± SD. Nine mice per group were used in this study.

The rabbit ileal loop test was performed as described previously (Koley et al., 1999). New Zealand white rabbits were fasted for 48 h before the surgery and only fed water ad libitum. Rabbits were anesthetised by intramuscular injection of ketamine (35 mg/kg body weight) and xylazine (5 mg/kg body weight). A laparotomy was performed, and the ileum was washed and ligated into discrete loops of about 10 cm. Each segment was separated by uninoculated segments of 1 to 2 cm. Test loops were inoculated with 10⁹ CFU of the challenge strain (V100, V114, IDH00990, IDH02003) in PBS. Negative-control loops were infused with PBS. The intestine was put back into the peritoneal cavity, animals were sutured and returned to their cages. After about 18 h, the animals were sacrificed with sodium pentobarbital (150 mg/kg) and the abdomens were re-opened. Loops were taken out. The volume of the accumulated fluid and the length of the loops were measured. The extent of the fluid accumulation (FA) was expressed as loop fluid volume (ml)/length (cm) ratio. Values were expressed as mean ± SD (three rabbits per group). Using the same model the bacterial colonization and histopathology were also performed. Bacterial colonization was quantified as mentioned earlier.

Pathogen free, 4 to 5-week-old BALB/c mice (five groups consisting three mice/group) were orally treated with 500 μl of sterile water containing 20 mg of streptomycin for 7 days as described previously (Sinha et al., 2016). Water and food were withdrawn 4 h before the streptomycin treatment as well as before infecting the rabbits with V. cholerae. The mice were then infected orally with 50 μl of 8.5% (wt/vol) sodium bicarbonate immediately followed by 50 μl of PBS containing 10⁹ CFU test strains of V. cholerae (V100, V114, IDH00990, and IDH02003). Negative control mice were fed with 50 μl of PBS only. Food was then immediately offered to all animals. To analyze bacterial shedding pattern, stool samples from three mice from each group were collected at a particular time up to 7 days.

To analyze intestinal colonization a similar but separate experiments like above (five groups; n = 21) were performed. Mice were sacrificed by cervical dislocations under anesthesia (isoflurane inhalation). The small intestines were removed aseptically from three animals of each group and homogenized in 4°C cold PBS. This procedure was repeated for 7 days. After dilution, homogenized stool samples and intestinal soups were diluted and plated on LBS agar. V. cholerae was enumerated in log of Colony Forming Unit/gram (CFU/gm) of stool and intestine. Values were expressed as mean ± SD. Intestinal tissues of these infected mice were used also for hematoxilin and eosin (H/E) staining and transmission electron microscopic analysis.

Histopathology was performed with the intestinal tissues of V. cholerae infected and uninfected adult mice and also of the rabbit models that were used for colonization assays. Briefly, tissue sections were embedded in paraffin block and then, 2–3 μm thick sections from paraffin-embedded intestinal segments were cut, placed on glass slides, and stained with H/E staining. All samples were examined by light microscopy (Leica DMLB), and digital photographs of stained tissues were taken at a magnification of 40x. H/E strained intestinal samples were then analyzed through Leica Quin Digital Image processing software and also by a pathologist blinded to the experimental sample set (D.R. Saha, National Institute of Cholera and Enteric Diseases, India). Scoring of the histological sections was also done (Barthel et al., 2003).

For TEM, small pieces of tissues were fixed in 3% cacodylate buffered glutaraldehyde (pH 7.4) for 4 h which were then rinsed in 0.1 M cacodylate buffer overnight. Post-fix specimens were incubated for 1 h in cacodylate buffer containing 1% Osmium tetroxide and then rinsed twice in 0.1 M cacodylate buffer (10 min each). Tissue samples were dehydrated with the gradient alcohol concentration. After that, infiltration with resin and embedding were done using BEEM capsule following standardized laboratory method. Tissue blocks were prepared and images were taken at 100 kV (FEI, Tecnai 126 12, BIOTWIN, Netherlands). TEM analysis was also performed by a pathologist who was blinded to the experimental protocol.

Data were expressed as mean ± standard deviation (SD). Statistical analysis was evaluated by Kruskal–Wallis one way ANOVA or two way ANOVA using GraphPad Prism software, version 5 (GraphPad Software, Inc., San Diego, CA, United States); ^∗p < 0.05 and ^∗∗p < 0.005 were considered as statistically significant.

The growth kinetics of both El Tor representatives (V100, V114) and Haitian variants (IDH00990 and IDH02003) were found to be predominantly sigmoidal. This result indicates the consistency in their growth patterns. Although the V. cholerae strains had different genetic backgrounds, their growth pattern did not differ significantly (Figure 1).

Intestinal colonization in suckling mouse was examined to find the colonization efficiency of El Tor and Haitian variant strains. In suckling mouse model, we found that Haitian variants (IDH00990 and IDH02003) had significantly higher colonization than the canonical El Tor strains (V100 and V114) after 18 h of infection. Mean Log₁₀CFU/gm of intestine count of V100 (7 ± 1) and V114 (7.4 ± 0.9) yielded significantly lower colonization efficacy compared to IDH00990 (11.64 ± 1.27) and IDH02003 (11.42 ± 1, p = 0.0001) (Figure 2).

Less fluid accumulation (FA) was observed with El Tor strains compared to the Haitian variant. This test depicted FA ratio of 0.58 ± 0.06 and 0.56 ± 0.05 with strains V100 and V114, respectively, while the strains IDH00990 and IDH02003 displayed a significantly higher FA ratio of 1.16 ± 0.15 and 1.3 ± 0.2 (p = 0.0141), respectively (Figure 3). In the colonization assay, Log₁₀CFU/gm of intestine count in V100 (6.7 ± 0.68) and V114 (6.72 ± 1.59) showed significant decrease in the colonization ability as compared to and IDH00990 (10.37 ± 1.25) and IDH02003 (11.51 ± 0.5, p = 0.0143), respectively (Figure 4). Intestinal tissues from the same rabbits were used for H/E analysis and TEM. Widely dilated villi were observed in the H/E stained rabbit intestine infected with V100. Congestion and scattered hemorrhage were also found. On the other hand, hemorrhagic changes in both mucous and sub-mucous layer were seen in the case of IDH02003 infected tissue. Structural alteration and disruption of the villous surface was clear (Figure 5A). TEM analysis showed the damaged microvillus structure with IDH02003 but no such changes were seen with V100 treated sections (Figure 5B).

Altered genomic features, higher colonization ability and FA effects of Haitian variant strains driven us to compare the disease persistency of El Tor and Haitian variants. Bacterial shedding prevailed throughout the 7 days. V100 displayed a peak fecal shedding on day 1 with ≈9.3 log₁₀CFU/gm of stool. This was followed by a gradual declining trend. No bacterial shedding of IDH02003 was observed on days 3 and 4. Bacterial shedding was again observed from day 5 with a rise on day 6. This shedding profile was almost identical with the other strain IDH00990. The shedding profiles of V100 and V114 showed a gradual declining phase (Figure 6). Our results depicted persistent diarrhea by Haitian variants which resembles the infection pattern of non-O1 and non-O139.

Though bacterial shedding was observed for 7-days, but there was no shedding on days 3 and 4 in Haitian variant infected mice. To confirm this persistent diarrheal pattern, we performed a 7-day colonization assay. Colonization also occurred throughout the 7 days in both El Tor and Haitian variant strains. Log₁₀CFU/gm of intestine count in IDH02003 was significantly higher up to day 5 in comparison to V100 (Figure 7).

Results of bacterial shedding and colonization motivated us to analyze the intestinal tissue damaging ability of the Haitian variant strains. Histopathological staining of the mouse tissue infected with V100 showed widely dilated villi with congestion, hemorrhage and increased cellularity with bacterial colonization on day 1. Inflammatory changes were observed to some extent up to day 2 and turned to normal from day 3 onward. On the other hand, IDH02003 showed increased mucosal damaging, inflammation and colonization ability until days 4–5. Gross structural alteration with distorted villus epithelium could be seen even on day 5. Day 6 onward, severity of the lesion gradually declined and became almost normal on day 7 (Figure 8A–C). TEM analysis with the same intestinal tissues has shown the greater damaging potential of IDH02003 and IDH00990 with higher colonization and the ability to disrupt the microvillus structure compared to V100 and V114 (Figure 9A,B).