Population Analysis of Vibrio cholerae in Aquatic Reservoirs Reveals a Novel Sister Species (Vibrio paracholerae sp. nov.) with a History of Association with Humans

Abstract

Most efforts to understand the biology of Vibrio cholerae have focused on a single group, the pandemic-generating lineage harboring the strains responsible for all known cholera pandemics. Consequently, little is known about the diversity of this species in its native aquatic environment. To understand the differences in the V. cholerae populations inhabiting regions with a history of cholera cases and those lacking such a history, a comparative analysis of population composition was performed. Little overlap was found in lineage compositions between those in Dhaka, Bangladesh (where cholera is endemic), located in the Ganges Delta, and those in Falmouth, MA (no known history of cholera), a small coastal town on the United States east coast. The most striking difference was the presence of a group of related lineages at high abundance in Dhaka, which was completely absent from Falmouth. Phylogenomic analysis revealed that these lineages form a cluster at the base of the phylogeny for the V. cholerae species and were sufficiently differentiated genetically and phenotypically to form a novel species. A retrospective search revealed that strains from this species have been anecdotally found from around the world and were isolated as early as 1916 from a British soldier in Egypt suffering from choleraic diarrhea. In 1935, Gardner and Venkatraman unofficially referred to a member of this group as Vibrio paracholerae. In recognition of this earlier designation, we propose the name Vibrio paracholerae sp. nov. for this bacterium. Genomic analysis suggests a link with human populations for this novel species and substantial interaction with its better-known sister species. IMPORTANCE Cholera continues to remain a major public health threat around the globe. Understanding the ecology, evolution, and environmental adaptation of the causative agent (Vibrio cholerae) and tracking the emergence of novel lineages with pathogenic potential are essential to combat the problem. In this study, we investigated the population dynamics of Vibrio cholerae in an inland locality, which is known as endemic for cholera, and compared them with those of a cholera-free coastal location. We found the consistent presence of the pandemic-generating lineage of V. cholerae in Dhaka, where cholera is endemic, and an exclusive presence of a lineage phylogenetically distinct from other V. cholerae lineages. Our study suggests that this lineage represents a novel species that has pathogenic potential and a human link to its environmental abundance. The possible association with human populations and coexistence and interaction with toxigenic V. cholerae in the natural environment make this potential human pathogen an important subject for future studies.

Keywords: Vibrio cholerae; cholera; novel species; paracholera; population genomics.

FIG 1

Abundance and diversity of Vibrio cholerae populations in two geographic locations, Dhaka and Oyster Pond. (A) Absolute average abundance of V. cholerae in two locations. viuB gene copy numbers were quantified from qPCR data; the average of viuB gene copies for all the samples in two locations was calculated and used as a proxy for the V. cholerae abundance. Total height of the bar represents total V. cholerae (viuB), the black segment represents viuB-73, and the clear segment represents other viuB alleles. (B) Evenness and diversity of the two V. cholerae populations measured by Peilou’s evenness and Shannon diversity indices based on analysis of viuB alleles. Statistical significance was measured by Kruskal-Wallis test; ***, statistically significant differences (Kruskal-Wallis, P < 0.01).

FIG 2

Nonmetric multidimensional scaling (NMDS) plot comparing beta diversities of Vibrio cholerae populations from two aquatic environments. Population compositions were compared using a Bray-Curtis dissimilarity matrix, with ellipses representing 95% confidence intervals. The data set was composed of viuB gene amplicon sequences normalized by qPCR copy numbers. The NMDS plot (stress 0.16) shows distinct clustering of samples from the two locations shown along the first two axes labeled as NMDS 1 and NMDS 2. Analyses of similarity (ANOSIM) results are displayed in the box inside the plot describing dissimilarity between pairs of samples from the two locations.

FIG 3

Abundance of the most prevalent viuB alleles at two locations: Dhaka (Bangladesh) (A) and Oyster Pond (USA) (B). Total viuB gene copy numbers were obtained by qPCR. Relative abundance of each allele was determined by amplicon sequencing. Specific colors were used for individual alleles to be consistent with a scheme described elsewhere (9). The 10 most abundant alleles for each location were selected for comparison between the two locations.

FIG 4

Whole-genome phylogeny of Vibrio cholerae strains found in Dhaka and Oyster Pond populations. The phylogenetic tree was inferred using Parsnp v1.2 (69) based on the reference genome of V. cholerae O1 El Tor N16961 and includes representative strains from other environments. Leaves of the tree were colored according to the viuB allele found in that particular genome. Statistical support of relevant nodes was estimated by bootstrap analysis (1,000 replicates, indicated as a percentage). The scale bar represents nucleotide substitutions per site.

FIG 5

Whole-genome phylogenetic tree of Vibrio paracholerae along with its closest relatives. The maximum likelihood phylogenetic tree was constructed from the core genome alignment of ≈2.1 Mbp using the GTR gamma substitution model. Corresponding nodes with relevant bootstrap support over 70% from the 100 replicates were indicated with an asterisk (*). The scale bar represents nucleotide substitutions per site.

FIG 6

Phylogenetic tree of O antigen cluster genes found in Vibrio paracholerae and Vibrio cholerae. Maximum likelihood trees were constructed using a 705-bp nucleotide alignment of the gene encoding UDP-glucose 4-epimerase (A) and a 564-bp alignment of the gene encoding UDP-N acetylgalactosaminyltransferase (B). Nodes with relevant bootstrap support over 70% of 100 replicates are indicated with an asterisk (*). The scale bar represents nucleotide substitutions per site.