Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Apr 7;117(14):7897-7904.
doi: 10.1073/pnas.1918763117. Epub 2020 Mar 30.

Toxigenic Vibrio cholerae Evolution and Establishment of Reservoirs in Aquatic Ecosystems

Affiliations
Free PMC article

Toxigenic Vibrio cholerae Evolution and Establishment of Reservoirs in Aquatic Ecosystems

Carla Mavian et al. Proc Natl Acad Sci U S A. .
Free PMC article

Abstract

The spread of cholera in the midst of an epidemic is largely driven by direct transmission from person to person, although it is well-recognized that Vibrio cholerae is also capable of growth and long-term survival in aquatic ecosystems. While prior studies have shown that aquatic reservoirs are important in the persistence of the disease on the Indian subcontinent, an epidemiological view postulating that locally evolving environmental V. cholerae contributes to outbreaks outside Asia remains debated. The single-source introduction of toxigenic V. cholerae O1 in Haiti, one of the largest outbreaks occurring this century, with 812,586 suspected cases and 9,606 deaths reported through July 2018, provided a unique opportunity to evaluate the role of aquatic reservoirs and assess bacterial transmission dynamics across environmental boundaries. To this end, we investigated the phylogeography of both clinical and aquatic toxigenic V. cholerae O1 isolates and show robust evidence of the establishment of aquatic reservoirs as well as ongoing evolution of V. cholerae isolates from aquatic sites. Novel environmental lineages emerged from sequential population bottlenecks, carrying mutations potentially involved in adaptation to the aquatic ecosystem. Based on such empirical data, we developed a mixed-transmission dynamic model of V. cholerae, where aquatic reservoirs actively contribute to genetic diversification and epidemic emergence, which underscores the complexity of transmission pathways in epidemics and endemic settings and the need for long-term investments in cholera control at both human and environmental levels.

Keywords: cholera; evolution; phylodynamics; reservoir.

Conflict of interest statement

The authors declare no competing interest.

Figures

Fig. 1.
Fig. 1.
Cholera epidemic in the Ouest Department, Haiti, 2010 to 2015. (A) Geographical distribution of environmental (n = 27) and clinical (n = 89) isolates sampled from the Haitian toxigenic V. cholerae O1 lineage in the Ouest Department between 2010 and 2015. (B, Upper) Monthly case counts of cholera infections in the Ouest Department and mean monthly precipitation (mm/mo) between 2010 and 2015. (B, Lower) Temporal distribution of V. cholerae genomes sampled in Haiti and of the number of cases reported monthly to the World Health Organization from October 2010, the beginning of the epidemic, until December 2015 (ticks correspond to the month of January for each year indicated on the x axis). The size of the circle is proportional to the number of environmental (green) and clinical (violet) genomes sampled in our study for each year. (C) Ancestral-state reconstruction using maximum likelihood inferred with PastML from a rooted ML phylogenetic tree inferred with IQ-TREE (SI Appendix, Fig. S1E). ML tree clades are merged vertically (clade circles contain the number of tips of the initial tree contained in them) and horizontally (branch size corresponds to the number of times its subtree is found in the initial tree) to cluster independent events of the same kind.
Fig. 2.
Fig. 2.
Contribution of toxigenic V. cholerae O1 environmental isolates to the evolution of the cholera epidemic in Haiti between 2012 and 2014. (A) MCC phylogeny for 116 environmental and clinical toxigenic V. cholerae O1 isolates collected between October 2010 and December 2015 inferred from genome-wide hqSNP data using the Bayesian phylogeography framework implemented in BEAST package version 1.8.4. Branch lengths are scaled in time by enforcing a relaxed molecular clock. Environmental and clinical states are indicated in green and violet, respectively. Circles at internal nodes indicate high posterior probability (PP) support (PP > 0.9). (B) Trunk reward proportion (TRP) at each ancestral location state estimated over time inferred using the continuous-time Markov chain model. Green- and purple-shaded areas represent the trunk proportions over time for environmental and clinical transitions, respectively. (C) Number of jumps (%) of all possible intrastate (clinical-to-clinical, environmental-to-environmental) and interstate transitions (clinical-to-environmental, environmental-to-clinical) normalized by total numbers of transitions obtained from the MCC phylogeny. Internal refers to all internal branches including the backbone path, and external to terminal branches of the tree. (D) Weighted averages of synonymous substitution rate estimates for environmental toxigenic V. cholerae O1 isolates were based on 200 randomly sampled trees from the posterior distribution of molecular clock-calibrated Bayesian phylogenies. Internal refers to estimates based on all internal branches of the tree, while external refers to estimates based on terminal branches. An asterisk indicates a significant (P < 0.001) difference between rate estimates. Error bars indicate standard deviation.
Fig. 3.
Fig. 3.
Evolution and adaptation of V. cholerae in Haitian aquatic reservoirs. Phylogenetic relationship and geographical distribution of 27 environmental isolates collected between 2012 and 2015 in Haiti. Nonsynonymous mutations acquired during the evolution of the environmental population, reconstructed by Bayesian inference of ancestral states, are indicated along the backbone of the tree. The SNPs detected along the trunk (surviving lineage) of the tree were sequentially numbered from 1 to 7 to facilitate comparison with the additional information reported in SI Appendix, Table S7. Maps show the sampling sites, grouped by aquatic source, labeled with yellow, red, orange, and cyan dots in the Gressier region, purple for Carrefour, and blue in the Jacmel region.
Fig. 4.
Fig. 4.
Mixed-transmission model of cholera and vaccination prediction. (A) Ne (effective population size) observed from data (gray) and in simulation with environmental replication (green) and Ne in simulation without environmental replication (violet). (B) Distinct scenarios of vaccination and control for the model simulation with environmental replication, in particular vaccination coverages of 64% (rate of 0.01 d−1), 88% (rate 0.04 d−1), and 64% with a 10% increase in environmental decay rate. Here we define vaccination coverage as the percent reduction in susceptible individuals (upon reaching an approximate steady state after 2 to 3 mo of vaccination). As opposed to Kirpich et al. (27) and simulations without environmental replication (SI Appendix, Fig. S10) in which 64% vaccination coverage eradicates the pathogen within a year, here either more vaccination (88% coverage) or increased environmental decay (10% increase) is needed to control cholera in the presence of environmental replication.

Similar articles

See all similar articles

References

    1. Cvjetanovic B., Barua D., The seventh pandemic of cholera. Nature 239, 137–138 (1972). - PubMed
    1. Colwell R. R., Huq A., Environmental reservoir of Vibrio cholerae. The causative agent of cholera. Ann. N. Y. Acad. Sci. 740, 44–54 (1994). - PubMed
    1. Stine O. C., et al. , Seasonal cholera from multiple small outbreaks, rural Bangladesh. Emerg. Infect. Dis. 14, 831–833 (2008). - PMC - PubMed
    1. Alam M., et al. , Seasonal cholera caused by Vibrio cholerae serogroups O1 and O139 in the coastal aquatic environment of Bangladesh. Appl. Environ. Microbiol. 72, 4096–4104 (2006). - PMC - PubMed
    1. Weill F. X., et al. , Genomic history of the seventh pandemic of cholera in Africa. Science 358, 785–789 (2017). - PubMed

LinkOut - more resources

Feedback