Comparisons of pollen coat genes across Brassicaceae species reveal rapid evolution by repeat expansion and diversification

Reproductive genes and traits evolve rapidly in many organisms, including mollusks, algae, and primates. Previously we demonstrated that a family of glycine-rich pollen surface proteins (GRPs) from Arabidopsis thaliana and Brassica oleracea had diverged substantially, making identification of homologous genes impossible despite a separation of only 20 million years. Here we address the molecular genetic mechanisms behind these changes, sequencing the eight members of the GRP cluster, along with 11 neighboring genes in four related species, Arabidopsis arenosa, Olimarabidopsis pumila, Capsella rubella, and Sisymbrium irio. We found that GRP genes change more rapidly than their neighbors; they are more repetitive and have undergone substantially more insertion/deletion events while preserving repeat amino acid composition. Genes flanking the GRP cluster had an average K-a/K-s approximate to 0.2, indicating strong purifying selection. This ratio rose to approximate to0.5 in the first GRP exon, indicating relaxed selective constraints. The repetitive nature of the second GRP exon makes alignment difficullt; even so, K-a/K-s within the Arabidopsis genus demonstrated an increase that correlated with exon length. We conclude that rapid GRP evolution is primarily due to duplication, deletion, and divergence of repetitive sequences. GRPs may mediate pollen recognition and hydration by female cells, and divergence of these genes could correlate with or even promote speciation. We tested cross-species interactions, showing that the ability of A. arenosa stigmas to hydrate pollen correlated with GRP divergence and identifying A. arenosa as a model for future studies of pollen recognition.