In the past decades, the grasses of the Paspalum genus have emerged as a versatile model allowing evolutionary, genetic, molecular, and developmental studies on apomixis as well as successful breeding applications. The rise of such an archetypal system progressed through integrative phases, which were essential to draw conclusions based on solid standards. Here, we review the steps adopted in Paspalum to establish the current body of knowledge on apomixis and provide model breeding programs for other agronomically important apomictic crops. In particular, we discuss the need for previous detailed cytoembryological and cytogenetic germplasm characterization; the establishment of sexual and apomictic materials of identical ploidy level; the development of segregating populations useful for inheritance analysis, positional mapping, and epigenetic control studies; the development of omics data resources; the identification of key molecular pathways via comparative gene expression studies; the accurate molecular characterization of genomic loci governing apomixis; the in-depth functional analysis of selected candidate genes in apomictic and model species; the successful building of a sexual/apomictic combined breeding scheme.

Restorer-of-fertility (Rf) is a suppressor of cytoplasmic male sterility (CMS), a mitochondrion-encoded trait that has been reported in many plant species. The occurrence of CMS is considered to be independent in each lineage; hence, the question of how Rf evolved was raised. Sugar beet Rf resembles Oma1, a gene for quality control of the mitochondrial inner membrane. Oma1 homologues comprise a small gene family in the sugar beet genome, unlike Arabidopsis and other eukaryotes. The sugar beet sequence that best matched Arabidopsis atOma1 was named bvOma1; sugar beet Rf (RF1-Oma1) was another member. During anther development, atOma1 mRNA was detected from the tetrad to the microspore stages, whereas bvOma1 mRNA was detected at the microspore stage and RF1-Oma1 mRNA was detected during the meiosis and tetrad stages. A transgenic study revealed that, whereas RF1-Oma1 can bind to a CMS-specific protein and alter the higher-order structure of the CMS-specific protein complex, neither bvOma1 nor atOma1 show such activity. We favour the hypothesis that an ancestral Oma1 gene duplicated to form a small gene family, and that one of the copies evolved and acquired a novel expression pattern and protein function as an Rf, i.e. RF1-Oma1 evolved via neofunctionalization.