Evolutionary plasticity of restorer-of-fertility-like proteins in rice

Hybrid seed production in rice relies on cytoplasmic male sterility (CMS) induced by specific mitochondrial proteins, whose deleterious effects are suppressed by nuclear Restorer of Fertility (RF) genes. The majority of RF proteins belong to a specific clade of the RNA-binding pentatricopeptide repeat protein family. We have characterised 'restorer-of-fertility-like' (RFL) sequences from 13 Oryza genomes and the Brachypodium distachyon genome. The majority of the RFL sequences are found in genomic clusters located at two or three chromosomal loci with only a minor proportion being present as isolated genes. The RFL genomic cluster located on Oryza chromosome 10, the location of almost all known active rice RF genes, shows extreme variation in structure and gene content between species. We show evidence for homologous recombination events as an efficient mechanism for generating the huge repertoire of RNA sequence recognition motifs within RFL proteins and a major driver of RFL sequence evolution. The RFL sequences identified here will improve our understanding of the molecular basis of CMS and fertility restoration in plants and will accelerate the development of new breeding strategies.