Genetic differentiation of Magnaporthe oryzae populations from scouting plots and commercial rice fields in Korea

Abstract

A previous study of the diversity and population structure of the rice blast fungus, Magnaporthe oryzae, over a 20-year period in Korea, found novel fingerprint haplotypes each year, and the authors hypothesized that populations might experience annual bottlenecks. Based on this model, we predicted that M. oryzae populations would have little or no genetic differentiation among geographic regions because rice blast is commonly found throughout Korea each year and M. oryzae would have to disperse from small populations surviving annually between rice crops. To test this hypothesis, we sampled M. oryzae from rice fields in eight provinces in Korea in a single year (1999). In four provinces, we sampled from a set of rice cultivars commonly grown in commercial fields (group I); because of low disease incidence in four other provinces, we could not sample from commercial fields and instead sampled from scouting plots of different cultivars set up for detecting new pathotypes of M. oryzae (group II). All isolates were genotyped with DNA fingerprint probes MGR586 and MAGGY, a telomere-linked gene family member TLH1, the PWL2 host specificity gene and mating type. Fingerprint haplotypes clustered into two distinct lineages corresponding to the two sets of cultivars (groups I and II), with haplotype similarities of 71% between lineages and >76% within lineages. Isolates from the same cultivar within group I were genetically differentiated among locations, and isolates within the same location were differentiated among cultivars. Differentiation for TLH1 and PWL2 was significant (P < 0.03), but not as strong as for fingerprint markers. Similar analyses were not possible among group II isolates because too few isolates were available from any one cultivar. All isolates were in the same mating type, Mat1-1, ruling out sexual reproduction as a source of novel haplotypes. When the 1999 samples were compared with the historical samples from the previous study, haplotypes of group I formed a separate cluster, while those of group II clustered with haplotypes from the historical sample. Altogether, geographic subdivision, monomorphism of mating type, and correlation of haplotypes to sets of cultivars are not consistent with the hypothesis of repeated turnover of haplotypes. Instead, the previous correlations of haplotypes to year might have been caused by inadequate sampling of haplotypes each year, highlighting the need for studies of population genetics to be conducted with systematic samples collected to address specific questions.