Background
The plant trait of cytoplasmically-inherited male sterility (CMS) and its suppression by nuclear restorer-of-fertility (Rf) genes can be viewed as a genetic arms race between the mitochondrial and nuclear genomes. Most nuclear Rf genes have been shown to encode P-type pentatricopeptide repeat proteins (PPRs). Phylogenetic analysis of P-class PPRs from sequenced plants genomes has shown that Rf-proteins cluster in a distinct clade of P-class PPRs, RFL-PPRs, that display hallmarks of positive evolutionary selection. Genes encoding RFL-PPRs (RFLs) within a given plant genome tend to be closely related both in sequence and position, but a detailed understanding of how such species-specific expansion occurs is lacking. In the canola, (oilseed rape) species Brassica napus, previous work has indicated the nuclear restorer genes for the two native forms of CMS, Rfn (for nap CMS) and Rfp (pol CMS), represent alternate haplotypes, or alleles, of a single nuclear locus.
Conclusions
Unlike other nuclear restorer loci containing multiple RFL genes, the RFL genes in the Rf region of B. napus are not present in tandem arrays but rather are dispersed in genomic location. The genes do not share similar flanking non-coding regions and do not contain introns, indicating that they have duplicated primarily through a retrotransposition-mediated process. In contrast, segmental duplication has been responsible for the distribution of the 10 sequences we annotated as RFL genes in the corresponding region of the A. lyrata genome. Our observations define the Brassica Rf locus and indicate that different mechanisms may be responsible for the proliferation of RFL genes even among closely related genomes.
Results
Fine genetic mapping indicates that Rfn does indeed localize to the same genomic region as Rfp. We find this region is enriched in RFL genes, three of which, based on their position and expression, represent potential candidates for Rfn; one of these genes, designated PPR4, is a preferred candidate in that it is not expressed in the nap CMS line. Comparison of the corresponding regions of the genomes of B. rapa, B. oleracea, Arabidopsis thaliana and A. lyrata provides insight into the expansion of this group of RFL genes in different lines of evolutionary descent. Conclusions: Unlike other nuclear restorer loci containing multiple RFL genes, the RFL genes in the Rf region of B. napus are not present in tandem arrays but rather are dispersed in genomic location. The genes do not share similar flanking non-coding regions and do not contain introns, indicating that they have duplicated primarily through a retrotransposition-mediated process. In contrast, segmental duplication has been responsible for the distribution of the 10 sequences we annotated as RFL genes in the corresponding region of the A. lyrata genome. Our observations define the Brassica Rf locus and indicate that different mechanisms may be responsible for the proliferation of RFL genes even among closely related genomes.