Population Genetics of Fusarium graminearum at the Interface of Wheat and Wild Grass Communities in New York

is primarily understood as an agricultural pathogen affecting cereal crops, but its host range also includes diverse, noncultivated grasses ubiquitous across agricultural and natural environments. Wild grasses may select for the production of diverse toxin variants (chemotypes) and serve as reservoirs of genetic diversity or sources of disease-inciting inoculum. Populations at the intersection of wheat and wild grass communities were described using 909 isolates collected from wheat spikes, wild grass spikes, and overwintered wild grass stems found at natural and agricultural sites in regions of high and low crop production. Trichothecene (TRI) genotypes correlated to pathogen chemotype were predicted from two loci, and multilocus genotypes (MLGs) were determined using eight microsatellite loci. The genetic diversity of wild grass and wheat-derived populations was comparable, and their differentiation was low. Duplicate MLGs were rare even in samples collected from a single square meter, although they could be found in multiple hosts, environments, regions, and years. TRI genotype frequencies differed between region and land use. Admixture between TRI genotype-defined populations, which correspond to three previously described sympatric North American populations, was detected and was highest in a region with remote host communities and little agricultural production. Nonagricultural environments may maintain different pathogen TRI genotypes than wheat fields and provide an opportunity for recombination between isolates from different populations. A lack of structural barriers suggests that pathogen gene flow is uninhibited between wheat and wild grass communities, and the recovery of putative clones from multiple hosts and environments provides initial evidence that noncultivated grasses are a source of local and regional inoculum.