Strong positive selection drives rapid diversification of R-genes in Arabidopsis relatives.

Although plant resistance (R) genes are extremely diverse and evolve rapidly, little is known about the mechanisms that generate this sequence divergence. To investigate these forces, we compared all nucleotide binding sites and leucine-rich repeat R-genes between two closely related species, Arabidopsis thaliana and Arabidopsis lyrata. Our analyses revealed two distinct evolutionary patterns driven by either positive or stabilizing selection. Most R-genes (>50%) were evolving under strong positive selection characterized by high Ka/Ks ratios (>1), frequent recombination, copy number variation, and extremely high sequence divergence between the two species. The stably selected R-genes (<30%) have exactly the opposite four characters as the positively selected genes. The remaining R-genes (about 20%) are present in only one genome and absent from the other. A higher proportion of such genes were found to be part of TNL class (23.5%) compared to the non-TNL class (5.6%), suggesting different evolutionary patterns between these two groups. A significant correlation between Ka and divergence was revealed, indicating that the rapid evolution and diversification of R-genes were initiated by selectively generated, frequently shuffled and selectively maintained non-synonymous substitutions. Our genome-wide analyses confirmed an amazing mechanism by which plants to selectively accumulate and efficiently exploit these non-synonymous substitutions for their resistance to various pathogens.