Dynamic nucleotide-binding site and leucine-rich repeat-encoding genes in the grass family

Abstract

The proper use of resistance genes (R genes) requires a comprehensive understanding of their genomics and evolution. We analyzed genes encoding nucleotide-binding sites and leucine-rich repeats in the genomes of rice (Oryza sativa), maize (Zea mays), sorghum (Sorghum bicolor), and Brachypodium distachyon. Frequent deletions and translocations of R genes generated prevalent presence/absence polymorphism between different accessions/species. The deletions were caused by unequal crossover, homologous repair, nonhomologous repair, or other unknown mechanisms. R gene loci identified from different genomes were mapped onto the chromosomes of rice cv Nipponbare using comparative genomics, resulting in an integrated map of 495 R loci. Sequence analysis of R genes from the partially sequenced genomes of an African rice cultivar and 10 wild accessions suggested that there are many additional R gene lineages in the AA genome of Oryza. The R genes with chimeric structures (termed type I R genes) are diverse in different rice accessions but only account for 5.8% of all R genes in the Nipponbare genome. In contrast, the vast majority of R genes in the rice genome are type II R genes, which are highly conserved in different accessions. Surprisingly, pseudogene-causing mutations in some type II lineages are often conserved, indicating that their conservations were not due to their functions. Functional R genes cloned from rice so far have more type II R genes than type I R genes, but type I R genes are predicted to contribute considerable diversity in wild species. Type I R genes tend to reduce the microsynteny of their flanking regions significantly more than type II R genes, and their flanking regions have slightly but significantly lower G/C content than those of type II R genes.

Figure 1.

Comparison of the presence and absence haplotypes of the R gene. The black boxes represent R genes, the gray boxes represent homologous sequences that are connected by shaded parallelograms between two haplotypes, and the white boxes represent unrelated sequences between two haplotypes. A, Deletion of the R gene in the absence haplotype was caused by unequal crossovers between two transposons (MuDR-5), resulting in a chimeric transposon. B, The absence haplotype was generated by a deletion followed by homologous repair using a homologous sequence from another chromosome as a template. C, The deletion end points were likely ligated through nonhomologous repair. D, Poor synteny between the presence and absence haplotypes. [See online article for color version of this figure.]

Figure 2.

An integrated map of R loci in Poaceae. R genes from different genomes were mapped onto the 12 chromosomes of rice cv Nipponbare. Prevalent P/A polymorphism between different genomes is evident.

Figure 3.

Chimeric structure of type I R genes. Sequence-exchange tracts are shaded. Only part of each exchange tract is shown. The consensus sequence is shown at the bottom. Dots represent nucleotides identical to the consensus sequence. Numbers at the top show nucleotide positions in gene BGIOSGA021935.

Figure 4.

Contrasting topology of neighbor-joining distance trees for type I and type II R genes. The numbers on nodes are bootstrap values, and values lower than 60 are not shown. A, Neighbor-joining distance tree for three type I R genes from Nipponbare and their close homologs from other cultivars. B, Neighbor-joining distance tree for three type II R genes from Nipponbare and their close homologs from other cultivars. Genes from Nipponbare are marked with black circles.

Figure 5.

Nonsense mutations conserved in a type II lineage. The left column shows the names of the cultivars and wild accessions from which the sequences were obtained. The numbers at the top show nucleotide positions in gene Os06g33360. The R gene Pi36 is included for comparison, and its amino acid sequence is shown at the bottom. Nineteen cultivars with sequences identical to Os06g33360/BGIOSGA023023 are not shown.