Environment predicts Batrachochytrium dendrobatidis lineage distribution and zones of recombination in South Africa.

ABSTRACT

The amphibian-infecting chytrid fungus, Batrachochytrium dendrobatidis (Bd), is widespread throughout Africa and is linked to declines of populations and species across the continent. While it is well established that the lineage of Bd encodes traits which determine disease severity, knowledge around how lineages are distributed according to environmental envelope is unclear. We here studied the distribution of Bd in South Africa based on the two lineages found, BdGPL and BdCAPE, in terms of their genome and environmental envelope statistically associated with their distribution. We used Bd surveillance data from published studies, as well as data collected during fieldwork from across South Africa, Lesotho, and eSwatini with samples collected along a transect spanning most of South Africa from Lesotho to the west coast. We utilized lineage-typing qPCR to resolve the spatial distribution of BdGPL and BdCAPE across South Africa and used the resulting surveillance data to create a predictive ecological niche model for Bd lineages in South Africa. Phylogenomic analyses were performed on isolates sourced from across the transect. We show that BdGPL demonstrates a strong isolation by distance suggestive of stepping-stone dispersal, while BdCAPE showed two distinct clusters within their genomic structure that appear geographically and temporally clustered, indicating two separate invasions. Our predictive niche model revealed that the two lineages tended to occur in different ecotypes; BdGPL was associated with lower altitude, arid regions while BdCAPE occurred across cooler, higher altitude environs. Niche predictions identified a zone of lineage contact, where genomics identified inter-lineage recombinants. We argue that this zone of recombination should be prioritized for disease surveillance as it is a potential hotspot for the evolution of variants of amphibian chytrid with novel traits that may be epidemiologically relevant.