A super PPR cluster for restoring fertility revealed by genetic mapping, homocap-seq and de novo assembly in cotton.

Rf candidate genes were related to the super D05_PPR-cluster and verified to be individually nonfunctional. Restorer of fertility (Rf) genes of cytoplasmic male sterility (CMS) is commonly found to be PPR (pentatricopeptide repeat) genes, which are mostly located in a cluster of PPR genes with high similarity. Here, Homocap-seq was applied to analyze PPR clusters in 'three lines,' and we found broad variations within the D05_PPR-cluster in a restorer line and deduced that the D05_PPR-cluster was associated with fertility restoration. Genetic mapping of Rf and Homocap-seq analysis of three genotypes in the F2 population validated that the D05_PPR-cluster was the origin of Rf. Three Rf candidates were cloned that were the most actively expressed genes in the D05_PPR-cluster in the restorer line as revealed by their high-depth amplicons. However, further transgenic experiments showed that none of the candidates could restore fertility of the CMS line independently. Then, the members of the brand-new super D05_PPR-cluster in the restorer line, containing 14 full-length PPRs and at least 13 PPR homologous sequences, were identified by long-read resequencing, which validated the effectiveness of variation and expression prediction of Homocap-seq. Additionally, we found that several PPR duplications, including 2 of the 3 Rf candidates, had undergone site-specific selection as potentially important anther development-associated genes. Finally, we proposed that multiple PPRs were coordinately responsible for the fertility restoration of the CMS line.