Local Population Structure and Patterns of Western Hemisphere Dispersal for Coccidioides spp., the Fungal Cause of Valley Fever

Abstract

Coccidioidomycosis (or valley fever) is a fungal disease with high morbidity and mortality that affects tens of thousands of people each year. This infection is caused by two sibling species, Coccidioides immitis and C. posadasii, which are endemic to specific arid locales throughout the Western Hemisphere, particularly the desert southwest of the United States. Recent epidemiological and population genetic data suggest that the geographic range of coccidioidomycosis is expanding, as new endemic clusters have been identified in the state of Washington, well outside the established endemic range. The genetic mechanisms and epidemiological consequences of this expansion are unknown and require better understanding of the population structure and evolutionary history of these pathogens. Here we performed multiple phylogenetic inference and population genomics analyses of 68 new and 18 previously published genomes. The results provide evidence of substantial population structure in C. posadasii and demonstrate the presence of distinct geographic clades in central and southern Arizona as well as dispersed populations in Texas, Mexico, South America, and Central America. Although a smaller number of C. immitis strains were included in the analyses, some evidence of phylogeographic structure was also detected in this species, which has been historically limited to California and Baja, Mexico. Bayesian analyses indicated that C. posadasii is the more ancient of the two species and that Arizona contains the most diverse subpopulations. We propose a southern Arizona-northern Mexico origin for C. posadasii and describe a pathway for dispersal and distribution out of this region.

Importance: Coccidioidomycosis, or valley fever, is caused by the pathogenic fungi Coccidioides posadasii and C. immitis The fungal species and disease are primarily found in the American desert southwest, with spotted distribution throughout the Western Hemisphere. Initial molecular studies suggested a likely anthropogenic movement of C. posadasii from North America to South America. Here we comparatively analyze eighty-six genomes of the two Coccidioides species and establish local and species-wide population structures to not only clarify the earlier dispersal hypothesis but also provide evidence of likely ancestral populations and patterns of dispersal for the known subpopulations of C. posadasii.

FIG 1

Bayesian phylogenetic analysis of C. immitis and C. posadasii isolates from all known regions of endemicity. The Bayesian statistical framework incorporated in BEAST 1.8.1 was used to integrate prior information, in the form of internal node timing estimates (from a fossil record) with a rooted tree, to produce a calibrated phylogeny. The analysis was performed on WGS data from 69 Coccidioides genomes (18 C. immitis and 51 C. posadasii). Clades of interest are highlighted as follows: green, Arizona; blue, Phoenix subclade of Arizona; orange, Guatemala; pink, Texas-Mexico-South America; and yellow, C. immitis. Stars highlight strains of interest. Posterior probabilities are indicated by node size. Purple node bars are shown for each node and are informative for the 95% confidence interval for the timing estimate. The timeline represents millions of years before the present.

FIG 2

Bayesian phylogenetic analysis of C. posadasii isolates. Beast 1.8.1 was used to produce a calibrated phylogeny, where the time to most recent common ancestor estimated for C. posadasii from the dual-species analysis whose results are presented in Fig. 1 was used for calibration here (TMRCA for C. posadasii, 818,100 years). The analysis was performed on WGS data from 51 C. posadasii genomes. Clades of interest are highlighted as follows: green, Arizona; blue, Phoenix subclade of Arizona; orange, Guatemala; and pink, Texas-Mexico-South America. Stars highlight strains of interest, while dotted circles represent TMRCA of interest. Posterior probabilities are indicated by node size. Purple node bars are shown for each node and are informative for the 95% confidence interval for the timing estimate. The timeline represents years before the present.

FIG 3

Bayesian phylogenetic analysis of C. immitis isolates. Beast 1.8.1 was used to produce a calibrated phylogeny, where the time to most recent common ancestor estimated for C. immitis from the dual-species analysis whose results are presented in Fig. 1 was used for calibration here (TMRCA for C. posadasii, 365,700 years). The analysis was performed on WGS data from 18 C. posadasii genomes. Posterior probabilities are indicated by node size. Purple node bars are shown for each node and are informative for the 95% confidence interval for the timing estimate. The timeline represents years before the present.

FIG 4

Genome-sharing analysis of Coccidioides species complex. fineStructure analysis was performed using the SNP matrix developed for Fig. S1 in the supplemental material. The SNPs from 66 Coccidioides genomes were reduced to a pairwise similarity matrix, which was used to identify population structure based on shared haplotype regions of genome. The x axis analysis represents the strain as a “recipient” of genomic regions, and the y axis represents the strain as a “donor” of genomic regions. The scale bar represents the number of shared genome regions, with blue representing the largest amount of sharing and yellow representing the smallest. The shading of isolates on the y axis correlates with clades in Fig. 1.

FIG 5

Population structure analysis of C. posadasii. Bayesian analysis of C. posadasii population structure was carried using BAPS 6.0, 3 fixed genetically diverged groups previously established by phylogenetic inferences (see Fig. S2 in the supplemental material), and 10 replicates. Admixture graphs of three identified C. posadasii mixtures (populations) were plotted using 200 simulations, and the percentage of genetic composition from each isolate was plotted. The shading of isolates in left column correlates with clades shown in Fig. 1.

FIG 6

Model for Coccidioides dispersal in the Western Hemisphere. A proposed dispersal model for C. posadasii from a hypothetical founder population in southern Arizona (South AZ)-northern Mexico and C. immitis from a hypothetical founder population in California’s Central Valley (Central CA Valley) is shown in conjunction with a timeline (labeled in thousands of years ago [kya]) that is color coded to match dispersal paths and annotated with the timing of great American biotic interchange (GABI) events 3 and 4, as well as of the final draining of the Central California basin. Central Am., Central America; South Am., South America; SE, southeastern; Mx, Mexico.