Evolution of the Genomic Region Controlling Self-Incompatibility in Arabidopsis: Junk DNA or Remarkable Regulatory Complexity?

Self-incompatibility is a genetic system in the flowering plants preventing self-fertilization as well as mating between genetically related individuals. In the Brassicaceae, it is controlled by a single highly polymorphic genomic region that contains two tightly linked genes encoding the pollen and pistil recognition proteins. Using a set of dedicated BAC libraries constructed from individuals containing specific S-haplotypes, we fully sequenced this challenging genomic region in 20 different accessions from two outcrossing species of Arabidopsis. We found a very high level of sequence divergence among haplotypes, to the point that orthologous sequences could not be aligned. The region was in particular characterized by a very strong enrichment in transposable elements, in line with the absence of recombination among distinct haplotypes and the reduced efficacy of purifying selection among gene copies within functional haplotypes. Beside this accumulation of junk DNA, we also uncovered in this region an exquisitely complex regulatory system controlling dominance/recessivity interactions between S-haplotypes. This system comprises at least eight new distinct miRNA families collectively targeting various positions in or near the gene encoding pollen specificity and determining the complex network of dominance observed in the species.