Self-incompatibility in the grasses: evolutionary relationship of the S gene from Phalaris coerulescens to homologous sequences in other grasses.

Self-incompatibility is widespread in the grasses and it is proposed that the grasses share a common incompatibility mechanism that is distinct from those operating in the dicotyledonous species studied in great detail. Where good genetic data are available, all grass species appear to have an incompatibility mechanism controlled by two unlinked loci, S and Z. A putative S gene has been cloned from Phalaris coerulescens. This gene is characterized by two major domains: an allele specificity domain and a thioredoxin catalytic domain. A family of sequences with varying degrees of homology to this gene has been identified among 15 grass species covering all subfamilies of the Poaceae. These S-related sequences appear to be present in the grass family regardless of self-compatibility. Evidence is presented to show that at least one of the sequences is transcribed, suggesting a functional gene. In contrast to the high expression of the S gene in Phalaris pollen, expression of the related gene in the pollen (or anthers) of the grass species examined was so low that RNA gel blot analysis failed to display a significant signal. However, reverse transcription-based polymerase chain reaction (RT-PCR) successfully amplified the region corresponding to the S thioredoxin domain from 10 of the grass species. With grasses other than Phalaris, RT-PCR showed limited success in amplifying the region corresponding to the S variable portion at the 5' end of the Phalaris S gene. Sequencing of the PCR-amplified S thioredoxin region from wheat, barley, rye and Dactylis revealed that this is a highly conserved gene with 94-97% sequence similarity with the corresponding Phalaris S gene. The conservation of sequence and ubiquitous expression of the gene across the grass family strongly suggest that the S-related gene is carrying out a significant biological function in the Poaceae. On the basis of these findings, a model for the evolution of the S self-incompatibility gene in the grasses is proposed.