Rapid generation of new powdery mildew resistance genes after wheat domestication

Plant defence against pathogens is controlled by disease resistance (R) gene products that directly or indirectly detect specific pathogen effectors. Plant-pathogen interactions have been proposed to follow a co-evolutionary arms-race model where R genes are recent and evolve rapidly in response to structural changes in matching pathogen effectors. However, the longevity and extensive polymorphism of R genes studied were more consistent with balancing selection maintaining ancient and diverse R genes or alleles. In bread wheat (Triticum aestivum), the Pm3 locus confers race-specific resistance to wheat powdery mildew (Blumeria graminis f.sp. triticii). Here we describe recently generated Pm3 resistance alleles that all derive from one susceptible allele, Pm3CS, which is widespread among hexaploid bread-wheat lines. One group of four Pm3 resistance alleles shows few, clearly delimited, polymorphic sequence blocks of ancient origin, embedded in sequences identical to Pm3CS and possibly derived from gene conversion. A second group of three alleles differs from Pm3CS by only two to five mutations, all non-synonymous, and all in the leucine-rich repeat-encoding region. Transient transformation experiments confirmed that Pm3 resistance specificities are based on one or few amino acid changes. The Pm3CS allele was found in wild tetraploid wheat, the ancestor of hexaploid bread wheat, specifically from southern Turkey, a region proposed to be the site of wheat domestication. Based on these data, we propose that the Pm3 resistance alleles were generated in agricultural ecosystems after domestication of wheat 10 000 years ago. The evolution of Pm3 alleles in wheat is best described by the model of evolved recycling, where novel genetic variation is integrated in plant populations together with recycling of old variation.