Low effective dispersal of asexual genotypes in heterogeneous landscapes by the endemic pathogen Penicillium marneffei

Abstract

Long-distance dispersal in microbial eukaryotes has been shown to result in the establishment of populations on continental and global scales. Such "ubiquitous dispersal" has been claimed to be a general feature of microbial eukaryotes, homogenising populations over large scales. However, the unprecedented sampling of opportunistic infectious pathogens created by the global AIDS pandemic has revealed that a number of important species exhibit geographic endemicity despite long-distance migration via aerially dispersed spores. One mechanism that might tend to drive such endemicity in the face of aerial dispersal is the evolution of niche-adapted genotypes when sexual reproduction is rare. Dispersal of such asexual physiological "species" will be restricted when natural habitats are heterogeneous, as a consequence of reduced adaptive variation. Using the HIV-associated endemic fungus Penicillium marneffei as our model, we measured the distribution of genetic variation over a variety of spatial scales in two host species, humans and bamboo rats. Our results show that, despite widespread aerial dispersal, isolates of P. marneffei show extensive spatial genetic structure in both host species at local and country-wide scales. We show that the evolution of the P. marneffei genome is overwhelmingly clonal, and that this is perhaps the most asexual fungus yet found. We show that clusters of genotypes are specific to discrete ecological zones and argue that asexuality has led to the evolution of niche-adapted genotypes, and is driving endemicity, by reducing this pathogen's potential to diversify in nature.

Figure 1. Co-Emergence of AIDS and Human Penicilliosis in Northern Thailand

Temporal emergence of HIV (antenatal data, 1990–2000 [38]) and human P. marneffei–associated penicilliosis (1985–2001; Maharaj Hospital, Chiang Mai) for the Chiang Mai region, northern Thailand.

Figure 2. Spatial Distribution of Isolates and Genetic Diversity in Thailand

(A) Spatial distribution of sampled isolates in Thailand. Isolates that are found to co-infect both humans and bamboo rats are shown in red, and the clone range of each of these co-infecting MTs is shown as coloured regions. Inset shows the geographic variation of the PC1, which accounts for 18.8% of the total genetic variation. Spatial interpolation of the PC1 values (with colours representing the predicted degree of difference between sample points) illustrates gradients in allele frequencies between northern, eastern, and south-central Thailand. (B) eBURST analysis of the Thailand P. marneffei dataset, with isolates coloured as noted in (A). The single green point refers to an isolate that was recovered from soil. Each clone is represented by a point, the area of which is representative of the frequency of the clone. Clones that differ from one another at a single locus, and are thus inferred to be directly related (single locus variants), are joined by lines.

Figure 3. Correlograms Showing the Correlation Coefficient r as a Function of Distance for Human Clinical Isolates and Bamboo Rat Isolates

Data for (A) human clinical isolates and (B) bamboo rat isolates. Values for r are shown as solid lines with 95% CIs calculated by bootstrapping the dataset. Also shown are the 95% CIs around the null hypothesis of randomly distributed genotypes of P. marneffei, calculated over 1,000 permutations of the data, shown as dotted lines.

Figure 4. Predicted Distributions for Three Classes of Genotype: Northern, Eastern, and Southern Thailand

The “overlap index” represents the synthesised output from 20 optimal GARP models. See Figure S2 for data sources.