Melioidosis is caused by the environmental bacterium Burkholderia pseudomallei. Infections are acquired by direct contact with the pathogen, most commonly through traumatic inoculation with contaminated soil or water but also by ingestion or inhalation. Symptoms are nonspecific and can include pneumonia, skin lesions, abscess formation, and sepsis (1).

In Latin America, melioidosis is believed to be underdiagnosed because of the absence of reliable surveillance and the lack of available diagnostic tools and methods (2). Colombia has previously reported cases as sporadic, isolated events in a few geographic areas (2,3). The aim of this study was to genetically characterize isolates of B. pseudomallei recovered from clinical specimens in different departments of Colombia (4). (A department in Colombia is a geographic unit composed of municipalities led by a governor.) The goal was to better understand genetic relationships among the isolates from Colombia, as well as their relationships to isolates from other tropical and subtropical regions of the Americas. The study was internally reviewed at the US Centers for Disease Control and Prevention (Atlanta, GA, USA) and determined not to involve human subject research.

Melioidosis is not an officially reportable disease in Colombia, but when cases are identified, department public health laboratories are required to send isolates of B. pseudomallei to the Instituto Nacional de Salud. During 2016–2017, a total of 11 isolates of B. pseudomallei were recovered from 10 melioidosis patients in the departments of Cesar (n = 4 isolates), Antioquia (n = 4), Casanare (n = 2), and Santander (n = 1) (Appendix). The most common risk factor was diabetes mellitus (n = 6); 4 of the patients died (Table). Cesar, Antioquia, Casanare, and Santander vary in population from a few hundred thousand to >6 million (4).

We performed whole-genome sequencing of the 11 isolates and deposited sequences at the National Center for Biotechnology Information under BioProject PRJNA638548. Sequences were used for multilocus sequence typing and single-nucleotide polymorphism (SNP) analysis (Appendix[[ANCHOR###F1######Anchor]]). The multilocus sequence types (ST) we observed were ones previously described, such as ST92, ST349, ST518, and ST1459. Two novel STs from this study were designated ST463 and ST1701. Previous entries in the PubMLST database (http://pubmlst.org) indicate that ST92 has been identified in cases associated with Puerto Rico and Brazil and in 1 person in Switzerland who had travelled to Martinique. ST349 was represented in 2 examples, one from Martinique and the other in a person from Spain who had travelled to West Africa; ST518 is represented in 4 examples. The first was in a person from Arizona, USA, in whom melioidosis developed after sustaining an injury while swimming in Costa Rica (5). In addition, ST518 was identified in B. pseudomallei isolates from 3 pet green iguanas, 2 of them in California, USA, and 1 in Belgium, all of which were presumably imported from Central or South America (6,7). ST1459 was noted in 1 isolate from Brazil.

Figure

Figure. Dendrogram used for characterization of Burkholderia pseudomallei genomes from Colombia compared with reference genomes. Tree was generated in MEGA7 software (http://www.megasoftware.net) from results of maximum-parsimony phylogenetic...

SNP analysis determined from the whole genome sequences indicates that the Colombia isolates (N=11) are within the clade associated with Western Hemisphere B. pseudomallei based on a comparison with a panel of reference genomes (N=45) (Figure). Within this clade, a subgroup was resolved containing the Colombia genomes along with ones from Brazil and Guatemala. Also included is a genome from an isolate from a patient who had traveled to both Panama and Peru, as well as isolates from iguanas from California and Belgium, as noted, plus 1 from the Czech Republic that were presumably imported from Central or South America (Figure) (6–8).

The full panel (N = 56) was also used for quantifying SNP differences among the genomes. Patient isolates B107 and B108 had no SNPs between them, even though they were from different patients, suggesting a common source of infection or a clonal population of B. pseudomallei present in different sources. However, isolates B308 and B309 were from the same patient and had 1 SNP between them. The next closest relationship was for B199 (from Casanare), which diverged by 38 SNPs from B308 and by 39 SNPs from B309 (from Antioquia). The phylogenetic SNP tree indicates that isolates from Antioquia, Casanare, and Cesar for the most part do not uniformly group together by department. The largest divergence was seen between B109 and the genomes for B107 and B108, with >6,900 SNPs detected (all from Cesar). The amount of divergence plus the lack of grouping by department, even though we presume that patients’ main exposures would have been within a given department, suggests B. pseudomallei is well established in Colombia and has had time to diverge substantially since its introduction. In addition, the genomes from the 2 cases of melioidosis from pet iguanas from California and the 1 from Belgium cluster together with examples from Colombia, suggesting this region or a nearby region may have been the origin of the iguanas. Further studies, especially to recover and test environmental isolates, will improve our understanding of the population structure of B. pseudomallei in Colombia and improve the ability of public health stakeholders to respond to cases of melioidosis.

Top