Genome Sequences Reveal Cryptic Speciation in the Human Pathogen Histoplasma capsulatum

Abstract

Histoplasma capsulatum is a pathogenic fungus that causes life-threatening lung infections. About 500,000 people are exposed to H. capsulatum each year in the United States, and over 60% of the U.S. population has been exposed to the fungus at some point in their life. We performed genome-wide population genetics and phylogenetic analyses with 30 Histoplasma isolates representing four recognized areas where histoplasmosis is endemic and show that the Histoplasma genus is composed of at least four species that are genetically isolated and rarely interbreed. Therefore, we propose a taxonomic rearrangement of the genus.IMPORTANCE The evolutionary processes that give rise to new pathogen lineages are critical to our understanding of how they adapt to new environments and how frequently they exchange genes with each other. The fungal pathogen Histoplasma capsulatum provides opportunities to precisely test hypotheses about the origin of new genetic variation. We find that H. capsulatum is composed of at least four different cryptic species that differ genetically and also in virulence. These results have implications for the epidemiology of histoplasmosis because not all Histoplasma species are equivalent in their geographic range and ability to cause disease.

Keywords: Histoplasma; cryptic speciation; genomic alignment; phylogenetic species concept; taxonomy.

FIG 1

Principal-component analysis for Histoplasma using whole-genome allelic frequencies. The genetic variance in the genus Histoplasma indicates the existence of discrete genetic clusters. Red, NAm 1; blue, NAm 2; yellow, LAm A; gray, Africa; light blue, Panama. (Left panel) Principal components (PC) 1 and 2 are shown, and the percentage of variance explained by each eigenvalue is shown within parentheses on each axis. (Right panel) PC3 and PC4.

FIG 2

Rooted phylogram for the species of the Histoplasma genus using whole-genome data. The leftmost tree (A) shows a tree constructed using the variation from the whole genome. The trees to the right (B) were built with data from each of the 14 supercontigs. (The length of each supercontig is not shown). The length of each branch indicates the amount of genetic divergence. Arrows in the trees to the right show isolates with conflicting positioning for a given supercontig. All topologies are rooted with Emmonsia sp.

FIG 3

Species tree of the genus Histoplasma shows that all species have high levels of genomic concordance. Shown is a genome-wide primary concordance tree produced by BUCKy (α = 5), based on 100-kb windows. Values above each branch show the concordance factor (CF) for each node, and values below branches show the 95% Bayesian credibility intervals for this statistic.

FIG 4

Pairwise genetic distance indicates that Histoplasma species are genetically differentiated across the whole genome. Values in the diagonal are the estimate of π_intra (intraspecific variability). All other values are π_inter (genetic distance between species). As expected Emmonsia crescens is more distantly related to Histoplasma species than they are to themselves. In all cases, π_inter is higher than π_intra, indicating strong genetic differentiation (Table 1).

FIG 5

TreeMix results show genetic differentiation between species from the Histoplasma genus. (A) Best-fitting genealogy without admixture for Histoplasma species calculated from the variance-covariance matrix of genome-wide allele frequencies. We estimated the likelihood of three demographic scenarios with 1 to 3 migration events (m). The caption for each panel shows the number of migrations (m), the LS of the demographic scenario, and the P value (p) of the comparison between that model and the model with m⁻¹ migration events (also shown in Table 3). (B) m = 1. (C) m = 2. (D) m = 3. To determine the best-fitting demographic scenario, we used LRTs to compare nested models (one migration versus two migrations and two migrations versus three migrations). We also used wAIC to find the best-supported migration scenario (see text). The caption for each panel contains the number of migrations, the likelihood, and the associated P value when comparing that demographic scenario with the previous one. The best-fitting demographic history involved two migration events: one from NAm 1 to Africa and a second one from Panama to NAm 1 (B).

FIG 6

ADMIXTURE results for NAm 1 and NAm 2. ADMIXTURE proportions among NAm 1 and NAm 2 isolates (K = 2). Each of the two clusters is represented by a different color, and each isolate is represented by a vertical line divided into two colored segments with heights proportional to genotype memberships in the clusters. Thin black lines separate isolates. 1: 505; 2, CI_19; 3, CI_22; 4, CI_24; 5, CI_42; 6, CI_43; 7, CI_7; 8, Downs; 9, UCLA_531; 10, WU24; 11, 1986; 12, CI_10; 13, CI_17; 14, CI_18; 15, CI_30; 16, CI_35; 17, CI_4; 18, CI_6; 19, CI_9; 20, G217B; 21, G222B. We ran independent analyses for the largest 14 supercontigs, indicated by the number on the right. Results from the whole-genome analyses are concordant with the results from supercontigs 2.1 to 2.10.