Population genomic analysis of the rice blast fungus reveals specific events associated with expansion of three main clades

Abstract

We examined the genomes of 100 isolates of Magnaporthe oryzae (Pyricularia oryzae), the causal agent of rice blast disease. We grouped current field populations of M. oryzae into three major globally distributed groups. A genetically diverse group, clade 1, which may represent a group of closely related lineages, contains isolates of both mating types. Two well-separated clades, clades 2 and 3, appear to have arisen as clonal lineages distinct from the genetically diverse clade. Examination of genes involved in mating pathways identified clade-specific diversification of several genes with orthologs involved in mating behavior in other fungi. All isolates within each clonal lineage are of the same mating type. Clade 2 is distinguished by a unique deletion allele of a gene encoding a small cysteine-rich protein that we determined to be a virulence factor. Clade 3 isolates have a small deletion within the MFA2 pheromone precursor gene, and this allele is shared with an unusual group of isolates we placed within clade 1 that contain AVR1-CO39 alleles. These markers could be used for rapid screening of isolates and suggest specific events in evolution that shaped these populations. Our findings are consistent with the view that M. oryzae populations in Asia generate diversity through recombination and may have served as the source of the clades 2 and 3 isolates that comprise a large fraction of the global population.

Fig. 1

Phylogenetic relationship of M. oryzae isolates. a Phylogenomic tree of M. oryzae isolates based on whole-genome SNP data. b STRUCTURE analysis of M. oryzae isolates. Each color in the plots represents the cluster membership coefficients. The presence of several colors in the same strain suggests admixture. c Principal component analysis (PCA) of M. oryzae isolates. d Origin of samples in this study. Red, blue, and green dots represent isolates of clade 1, clade 2 and clade 3. Some dots represent more than one isolate with same collection site. Suriname, French Guiana (left), and Ghana (right) are presented separately in two windows, for more information see Table S1

Fig. 2

Estimation of divergence time of M. oryzae populations. Chronogram from Bayesian phylogenetic analysis on whole-genome SNP dataset. Whole-genome SNP sequences were calibrated by collection dates. Blue bars indicate 95% highest posterior density intervals of node-age estimates. SV represent isolates collected from S. viridis. The red dashed line shows the break in the time axis of SV to rice population branch. Red stars marked divergence (star 2) or expansion (stars 1 and 3) time of clades

Fig. 3

Genomic divergence of M. oryzae field populations. a Distribution of Pi values against the corresponding density of genes for three clades. b Distribution of Ka/Ks values against the corresponding density of genes for three clades. The red, blue, and green colored lines represent the number of genes of clade 1, clade 2, and clade 3, respectively

Fig. 4

Recombination rate (ρ), nucleotide diversity (π), and Tajima’s D for the three clades of M. oryzae. a–d Whole-genome distribution of recombination rate (ρ), nucleotide diversity (π), Tajima’s D, and F_st on chromosome I ~ VII of M. oryzae with 50 kb windows. e–h Boxplot of recombination rate, Pi, Tajima’s D, and F_st values of three clades. The Tukey whiskers indicate 1.5 times the interquartile range from the 25th and 75th percentiles. n.s. represents no significance, * represents p-value smaller than 0.05, ** represents p-value smaller than 1e-5, *** represents p-value smaller than 1e-10

Fig. 5

Venn diagram of genes with F_st > 0.8 in the three clades. Venn diagram depicting the pair-wise comparison of genes with F_st > 0.8 between the three clades. Numbers in brackets represent the number of genes with non-synonymous substitutions

Fig. 6

Deletion of a putative effector gene defines clade 2 isolates. a Schematic of MGG_17227 in all of clade 2 isolates indicating MGG_17227 was replaced by a 5977 bp LINE retrotransposon MGL from 3757 bp upstream to 758 bp downstream of the coding region. b MGG_17227 suppresses BAX-induced cell death following Agrobacterium-mediated transfection of tobacco leaves. Both MGG_17227 full-length (1) and secretion signal-deleted construct (3) can suppress BAX-induced cell death (2, 4, 6, 8). AvrPiz-t (5) and empty vector (7) have been used as positive and negative controls, respectively