Population genomic footprints of host adaptation, introgression and recombination in coffee leaf rust

Abstract

Coffee leaf rust, caused by Hemileia vastatrix (Hv), represents the biggest threat to coffee production worldwide and ranks amongst the most serious fungal diseases in history. Despite a recent series of outbreaks and emergence of hypervirulent strains, the population evolutionary history and potential of this pathogen remain poorly understood. To address this issue, we used restriction site-associated DNA sequencing (RADseq) to generate ∼19 000 single nucleotide polymorphisms (SNPs) across a worldwide collection of 37 Hv samples. Contrary to the long-standing idea that Hv represents a genetically unstructured and cosmopolitan species, our results reveal the existence of a cryptic species complex with marked host tropism. Using phylogenetic and pathological data, we show that one of these lineages (C3) infects almost exclusively the most economically valuable coffee species (tetraploids that include Coffea arabica and interspecific hybrids), whereas the other lineages (C1 and C2) are severely maladapted to these hosts, but successfully infect diploid coffee species. Population dynamic analyses suggest that the C3 group may be a recent ‘domesticated’ lineage that emerged via host shift from diploid coffee hosts. We also found evidence of recombination occurring within this group, which could explain the high pace of pathotype emergence despite the low genetic variation. Moreover, genomic footprints of introgression between the C3 and C2 groups were discovered and raise the possibility that virulence factors may be quickly exchanged between groups with different pathogenic abilities. This work advances our understanding of the evolutionary strategies used by plant pathogens in agro-ecosystems with direct and far-reaching implications for disease control.