national surveillance systems in Latin America
were hyperalert for cholera outbreaks, and as a
result, their sampling framework captured a
diverse collection of V. cholerae during both
epidemic and interepidemic periods (10, 11).
Coupling these precise epidemiological data,
which describe the beginning of the epidemic, to
increased sampling within Latin America, allows
studies within this region to offer unprecedented
opportunities to address the relationships between local populations and globally circulating
pandemic lineages of V. cholerae.
The relationships between these bacterial populations have been difficult to characterize until
now, primarily due to the molecular methods
used to assess the relatedness of V. cholerae
isolates to one another (10, 12–14). Inconsistency
in applying and interpreting results generated
using these methods, such as the use of different
restriction enzymes for pulsed-field gel electrophoresis and ribotyping, and the lack of standardized nomenclatures, has further complicated
comparisons between studies. However, it is possible now to unify these results by using whole-genome sequencing.
To examine the relationship between V. cholerae
lineages in Latin America, we sequenced a collection of 252 isolates. Phylogenetic analysis showed
that 164 strains were of the 7P El Tor (7PET)
cholera lineage and 88 were strains distinct from
7PET (collectively referred to as non-7PET) (15).
This collection is geographically and temporally
broad, and includes representative isolates from
14 countries spanning 1974 to 2014, including preepidemic, epidemic (isolated and typed during
the 1991 and Haitian epidemics), and interepidemic periods. Critically, this collection includes
serogroup O1 and non-O1 isolates, both clinical
and environmental isolates (figs. S1 to S3 tables
S1 to S3, and supplementary text note 1), isolates
typed by early molecular approaches (table S4)
(11, 12), and several key additional 7PET isolates from Africa [see companion analysis of
Weill et al. (16)].
We placed these isolates into a phylogenetic
framework, and determined the evolutionary
relationships between lineages in Latin America.
A global phylogeny, comprising a total of 665
isolates, revealed a marked diversity of V. cholerae
lineages present in this region (Fig. 1 and figs. S2
and S3). Representative isolates from both the
1991 and 2010 epidemics clustered within the
7PET lineage (table S4). The phylogeny also re-
vealed that isolates sampled in different years,
and in some cases across multiple countries, com-
prise 11 lineages in Latin America (Fig. 1 and
fig. S4). These lineages include Classical V. cholerae
isolated from Mexico during the mid-1990s, as well
as several V. cholerae O1 local lineages, such as the
Gulf Coast lineage (17), or those containing isolates
of Mx1 to Mx3 ribotypes (MX-1 to MX-3 lineages)
described in Mexico (11). The Tucumán variant
from Argentina (18) and the Amazonia variant
from Brazil (19) form a single lineage, named
Endemic Latin American 1 (ELA-1), in which these
isolates remain clearly separated phylogenetically
by their country of origin (Fig. 1A and fig. S2).
Although 7 of the 18 samples sequenced from the
Tucumán and other regions of Argentina belong
to ELA-1, the other samples are distributed among
five other lineages. More than 30 additional iso-
lates sampled across Latin America do not belong
to any previously known lineage and comprise at
least eight different serotypes (fig. S5 and table S2).
Local V. cholerae O1 lineages in Latin America
harbor a wide range of genetic determinants that
are associated with pandemic disease (figs. S6 and
S7 and tables S2 and S3). For instance, the genes
encoding the bipartite cholera toxin (ctxAB), the
primary virulence determinant of cholera borne
1 Wellcome Trust Sanger Institute, Wellcome Genome Campus,
Hinxton CB10 1SA, UK. 2Institut Pasteur, Unité des Bactéries
Pathogènes Entériques, Paris, 75015, France. 3Department of
Medicine, University of Cambridge, Addenbrooke’s Hospital,
Cambridge CB2 0SP, UK. 4Department of Veterinary Medicine,
University of Cambridge, Madingley Road, Cambridge, CB3 0ES,
UK. 5Department of Microbiology and Parasitology, Faculty of
Medicine, Universidad Nacional Autónoma de México, Mexico,
D.F., Mexico. 6Institut Pasteur, Unité Biodiversité des
Bactéries Pathogènes Emergentes, Paris, 75015, France.
7Centro de Investigación Científica y de Educación Superior
de Ensenada, Baja California, (CICESE), Ensenada, Baja
California, Mexico. 8Institut Pasteur, Plate-forme Génomique
(PF1), Paris, 75015, France. 9Centre for Genomic Pathogen
Surveillance, Wellcome Genome Campus, Hinxton, Cambridge
CB10 1SA, UK. 10Subsecretaría de Prevención y Promoción de
la Salud, Secretaría de Salud, Ciudad de México, Mexico.
11Enteric Diseases Laboratory Branch, Centers for Disease
Control and Prevention, Atlanta, GA, USA. 12Instituto Nacional
de Enfermedades Infecciosas, ANLIS, Buenos Aires,
Argentina. 13Faculty of Medicine, Universidad Nacional
Autónoma de México, Mexico, D.F., Mexico. 14London School
of Hygiene and Tropical Medicine, London WC1E 7HT, UK.
*Corresponding author. Email: email@example.com (D.D.); nrt@sanger.
ac.uk (N.R. T.) †These authors contributed equally to this work.
Year of isolation
1975 1980 1985 1990 1995 2000 2005 2010 2015
Fig. 1. Multiple lineages of V. cholerae are present in Latin America. (A) Maximum
likelihood phylogeny of 148 V. cholerae genomes. Local lineages present in Latin America are
highlighted. The 7PET lineage is shown as a collapsed triangle. Four 7PET genomes
(reference genome N16961, and examples of LAT-1 to -3) were used as representatives of the 518
7PETgenomes in this study. The scale bar denotes substitutions per variable site. The hash mark
denotes a branch that was artificially shortened; the full tree is shown in figs. S2 and S3.
(B) Geographical distribution of selected local V. cholerae lineages in Latin America. The size of
the circle denotes the number of genomes analyzed from that area. Only isolates annotated
with explicit geographic information are shown. (C) Temporal distribution of genomes sampled
from V. cholerae lineages present in Latin America. The size of the circle scales with the number
of genomes in our study for each lineage.