Worldwide Phylogenetic Distributions and Population Dynamics of the Genus Histoplasma

Abstract

Background: Histoplasma capsulatum comprises a worldwide complex of saprobiotic fungi mainly found in nitrogen/phosphate (often bird guano) enriched soils. The microconidia of Histoplasma species may be inhaled by mammalian hosts, and is followed by a rapid conversion to yeast that can persist in host tissues causing histoplasmosis, a deep pulmonary/systemic mycosis. Histoplasma capsulatum sensu lato is a complex of at least eight clades geographically distributed as follows: Australia, Netherlands, Eurasia, North American classes 1 and 2 (NAm 1 and NAm 2), Latin American groups A and B (LAm A and LAm B) and Africa. With the exception of the Eurasian cluster, those clades are considered phylogenetic species.

Methodology/principal findings: Increased Histoplasma sampling (n = 234) resulted in the revision of the phylogenetic distribution and population structure using 1,563 aligned nucleotides from four protein-coding regions. The LAm B clade appears to be divided into at least two highly supported clades, which are geographically restricted to either Colombia/Argentina or Brazil respectively. Moreover, a complex population genetic structure was identified within LAm A clade supporting multiple monophylogenetic species, which could be driven by rapid host or environmental adaptation (~0.5 MYA). We found two divergent clades, which include Latin American isolates (newly named as LAm A1 and LAm A2), harboring a cryptic cluster in association with bats.

Conclusions/significance: At least six new phylogenetic species are proposed in the Histoplasma species complex supported by different phylogenetic and population genetics methods, comprising LAm A1, LAm A2, LAm B1, LAm B2, RJ and BAC-1 phylogenetic species. The genetic isolation of Histoplasma could be a result of differential dispersion potential of naturally infected bats and other mammals. In addition, the present study guides isolate selection for future population genomics and genome wide association studies in this important pathogen complex.

Fig 1. Maximum Likelihood (ML) tree of Histoplasma capsulatum generated by IQ-TREE software for 232 taxa through 4 different loci (arf, ole1, tub and anti-H loci) reveals at least monophyletic braches as following: NAm 1, NAm 2, RJ, LAm B, NAm LAm A1, LAm A2, BR1-4, and Cluster 6 containing Netherlands, Panama, Africa, Australia and BAC1.

Dual branch support, inferred by non-parametric bootstrap for ML analysis, combined with posterior probabilities obtained for the BI analysis, was added to the branches. Monophyletic branches that were supported by two methods (Bootstrap≥70/Posterior Probabilities≥0.95) were designated high confidence clades. We also identified possible in-group variation that may be associated with specific niches. Low supported clades such as Eurasia, Unknown 1 and Unknown 2 were detected but do not follow our monophyletic branches supporting criteria.

Fig 2. Population structure of Histoplasma capsulatum deduced by Bayesian Analysis of Population Structure (BAPS).

A) Structure plots of 205 isolates revealing 6 different major populations (Clusters 1–6). Phylogenetic species were assigned to each of the six deduced populations as follows: Cluster 1 representing the phylogenetic species NAm 1, Cluster 2 representing the phylogenetic species RJ, Cluster 3 containing LAm B, Cluster 4 constituted by phylogenetic species NAm 2, Cluster 5 constituted by LAm A1, LAm A2, BR1-4, and the paraphyletic low supported clades Eurasia, Unknown 1 and Unknown 2 and Cluster 6 containing Netherlands, Panama, Africa, Australia and BAC1. B) Bayesian population tree based on substructures of the 6 initial clusters deduced by BAPS. Gene flow is represented by mixture isolates that are annotated with brackets along the tree.

Fig 3. Extended Bayesian Skyline Plot (EBSP) of Histoplasma capsulatum sensu lato.

EBSPs represents population size changes over time and divergence dating was inferred using BEAST v1.8.2 based on conservative intervals of nucleotide substitutions rates and dates (0.00043–0.00656 subst/site/lineage/My; 0.0–15 Ma) that encompass values obtained by Kasuga et al. [40]. Y-axes are effective population size divided by generation time. X-axes are in millions of years. Confidence intervals of each dated phylogenetic species were added to the nodes.