| Gene duplication is the major source of new genes and gene functions. Gene families are produced by gene duplication and comprise a large proportion of many recently sequenced genomes. Plant disease resistance is an intriguing system to study the evolution of duplicated genes, as many proteins involved in disease resistance belong to multigene families (R gene families). Here we present a genome wide analysis of the evolutionary diversification of three putative R gene families. The proteins in these families contain a Leucine Rich Repeat region (LRR) that may determine resistance specificity and is subject to positive selection in experimentally characterized R genes. Most of the members of the families in our study are A. thaliana putative genes. An indirect way to determine their involvement in disease resistance is by determining whether they are subject to positive selection. We performed bioinformatic analyses of positive selection in members of the nucleotide binding site (NBS)-LRR gene superfamily. We found substantial evidence of positive selection in 32% of the superfamily. Positively selected positions were disproportionately located in the LRR domain (p < 0.001) particularly a nine amino acid beta-strand submotif that in the corresponding proteins is likely to be solvent exposed. However, a substantial proportion (30%) of positively selected sites were located outside LRRs, suggesting regions other than the LRR may determine resistance specificity. Positive selection on these genes acts on variation introduced by point mutation and, perhaps, gene conversion. The diversifying role of this second mechanism is controversial, as NBS-LRR genes are deemed to evolve independently. We employed bioinformatic analysis to characterize the extent, determining factors and diversifying role of gene conversion in NBS-LRR, Receptor-Like Kinases (RLKs) and Receptor-Like Proteins (RLPs). We found a substantial amount of gene conversion in highly identical and physically clustered genes from the three families. The occurrence of positive selection in both NBS-LRR and RLP superfamilies suggest their members are potentially diversified by positive selection acting on diversity introduced by gene conversion. In contrast the RLK superfamily evolves mostly by purifying selection, although a proportion of its members are involved in gene conversion. |
