Genome-wide comparison of nucleotide-binding site-leucine-rich repeat-encoding genes in Arabidopsis

Abstract

Plants, like animals, use several lines of defense against pathogen attack. Prominent among genes that confer disease resistance are those encoding nucleotide-binding site-leucine-rich repeat (NB-LRR) proteins. Likely due to selection pressures caused by pathogens, NB-LRR genes are the most variable gene family in plants, but there appear to be species-specific limits to the number of NB-LRR genes in a genome. Allelic diversity within an individual is also increased by obligatory outcrossing, which leads to genome-wide heterozygosity. In this study, we compared the NB-LRR gene complement of the selfer Arabidopsis thaliana and its outcrossing close relative Arabidopsis lyrata. We then complemented and contrasted the interspecific patterns with studies of NB-LRR diversity within A. thaliana. Three important insights are as follows: (1) that both species have similar numbers of NB-LRR genes; (2) that loci with single NB-LRR genes are less variable than tandem arrays; and (3) that presence-absence polymorphisms within A. thaliana are not strongly correlated with the presence or absence of orthologs in A. lyrata. Although A. thaliana individuals are mostly homozygous and thus potentially less likely to suffer from aberrant interaction of NB-LRR proteins with newly introduced alleles, the number of NB-LRR genes is similar to that in A. lyrata. In intraspecific and interspecific comparisons, NB-LRR genes are also more variable than receptor-like protein genes. Finally, in contrast to Drosophila, there is a clearly positive relationship between interspecific divergence and intraspecific polymorphisms.

Figure 1.

Phylogeny of CNL and CN proteins based on the NB domain. Vertical lines on the right side indicate several examples of proteins encoded in tandem (black) and mixed (blue) clusters in A. thaliana (for cluster assignments, see Supplemental Tables S3 and S4). Branches in red indicate A. lyrata sequences, and those in black indicate A. thaliana sequences. Asterisks indicate proteins without LRR domains. For complete names at each branch, see Supplemental Figure S1.

Figure 2.

Phylogeny of TNL and TN proteins based on the NB domain. Annotation is as for Figure 1. For complete names at each branch, see Supplemental Figure S2.

Figure 3.

Lengths of gene conversion tracts. Whiskers above boxes indicate the greatest value after excluding outliers; those below boxes indicate the least value after excluding outliers.

Figure 4.

Divergence of NB-LRR genes. A, Divergence of different classes. B, Divergence of singletons and cluster members.

Figure 5.

Potential deletions in 80 A. thaliana genomes. A, Missing calls. Dashed lines indicate the median of all NB-LRR genes. B, Allele frequencies of potential deletions in 80 genomes (Cao et al., 2011).

Figure 6.

Comparison of the RPP7 region in three haplotypes from A. thaliana and A. lyrata. A, Dot plots comparing two A. thaliana alleles. B, Annotation. The closest Col-0 homologs of two Cvi-0 genes in the center of the cluster are indicated in gray. NB-LRR genes are in red. Numbers indicate the sizes of gaps in kb (not to scale).

Figure 7.

Inferred evolutionary history of the RPP7 cluster. A, Phylogeny of NB-LRR genes in the RPP7 region based on the NB domain. Red, CNL; blue, CN; black, NB. B, Proposed history of the RPP7 cluster. A. lyrata constitutes the ancestral state. Color change indicates tandem duplications (black connectors), and gray indicates loss of gene (red connectors).

Figure 8.

Divergence of RLPs compared with NB-LRRs. A, Divergence of singletons (S) and cluster members (C). B, Missing calls in 80 resequenced A. thaliana genomes (Cao et al., 2011). NB-LRR data are the same as in Figures 4 and 5.