Skip to main content
PMC home page
Plant Physiology logoLink to Plant Physiology
. 2006 Mar;140(3):998–1008. doi: 10.1104/pp.105.073080

Point Mutations with Positive Selection Were a Major Force during the Evolution of a Receptor-Kinase Resistance Gene Family of Rice1,[W]

Xinli Sun 1, Yinglong Cao 1, Shiping Wang 1,*
PMCID: PMC1400573  PMID: 16461382

Abstract

The rice (Oryza sativa) Xa26 gene, which confers resistance to bacterial blight disease and encodes a leucine-rich repeat (LRR) receptor kinase, resides at a locus clustered with tandem homologous genes. To investigate the evolution of this family, four haplotypes from the two subspecies of rice, indica and japonica, were analyzed. Comparative sequence analysis of 34 genes of 10 types of paralogs of the family revealed haplotype polymorphisms and pronounced paralog diversity. The orthologs in different haplotypes were more similar than the paralogs in the same haplotype. At least five types of paralogs were formed before the separation of indica and japonica subspecies. Only 7% of amino acid sites were detected to be under positive selection, which occurred in the extracytoplasmic domain. Approximately 74% of the positively selected sites were solvent-exposed amino acid residues of the LRR domain that have been proposed to be involved in pathogen recognition, and 73% of the hypervariable sites detected in the LRR domain were subject to positive selection. The family is formed by tandem duplication followed by diversification through recombination, deletion, and point mutation. Most variation among genes in the family is caused by point mutations and positive selection.

Plants are constantly being subjected to pathogen attack and depend on the presence of specific genetic systems against the rapid evolution of pathogens. More than 40 resistance (R) genes against various plant diseases caused by bacteria, fungi, viruses, and nematodes have been characterized, providing insight into their function and evolution (Martin et al., 2003). The most prevalent domain encoded by functionally defined R genes is the Leu-rich repeat (LRR), which is considered to contribute to specificity in pathogen recognition (Dangl and Jones, 2001). Three classes of R genes encode proteins with an LRR domain (Hammond-Kosack and Parker, 2003). The largest class encodes intracellular nucleotide-binding site (NBS)-LRR proteins. The next largest class of LRR-containing R genes encodes extracellular LRR (eLRR)-transmembrane (TM) proteins. The third class encodes eLRR-TM-kinase or LRR receptor-kinase proteins. Only three genes—two, Xa21 and Xa26, from rice (Oryza sativa) and one, FLS2, from Arabidopsis (Arabidopsis thaliana)—have been identified as belonging to the third class (Song et al., 1995; Gomez-Gomez and Boller, 2000; Sun et al., 2004).

Increasing evidence indicates that R genes tend to be clustered in the genome (for review, see Michelmore and Meyers, 1998; Hulbert et al., 2001). The diversity of members in a complex R-gene family may provide variations in resistance specificity. The evolutionary mechanisms for the generation of novel resistance specificity of several complex R-gene families have been studied, and most results have indicated that the solvent-exposed amino acid residues of the xxLxLxx motif (“x” indicates any amino acid residue) that form a solvent-exposed surface for pathogen recognition in the LRR domain are subject to positive selection (Parniske et al., 1997; Botella et al., 1998; Dixon et al., 1998; Meyers et al., 1998; Wang et al., 1998; Noël et al., 1999; Dodds et al., 2001a, 2001b; Mondragón-Palomino et al., 2002; Caicedo and Schaal, 2004; Strain and Muse, 2005). These results suggest that positive selection of the predicted ligand-binding domain plays an important role in the evolution of these R-gene families. However, most of these studies have not reported the proportions and exact locations of the sites for positive selection, although some of them have demonstrated hypervariable sites in the LRR domain. Only two analyses of the NBS-LRR gene family and one analysis of the LRR receptor-kinase gene family in Arabidopsis and lettuce (Lactuca sativa) revealed the position of positively selected sites (Mondragón-Palomino et al., 2002; Kuang et al., 2004; Strain and Muse, 2005). The study of the complete NBS-LRR gene family in the Arabidopsis genome showed that only approximately 4% to 20% of the amino acid sites in different phylogenetic groups are subject to positive selection and that approximately 70% of the positively selected sites are located in the LRR domain, whereas the remaining 30% are located outside the LRR domain (Mondragón-Palomino et al., 2002). Only 43 of approximately 493 amino acid sites were identified under positive selection in the RGC2 gene family of lettuce (Kuang et al., 2004). Five of the 24 LRR receptor-kinase family groups in the Arabidopsis are subject to positive selection (Strain and Muse, 2005). Thus, it appears that the positive-selection pressure only applies to a small number of amino acid sites in the LRR-containing gene families of Arabidopsis and lettuce.

R-gene polymorphism is a significant source of resistance-specificity diversity. The tandem repeated homologous members (or paralogs) of a haplotype and their alleles (or orthologs) in different haplotypes provide a structural reservoir for evolutionary forces to generate rapid variation in resistance specificity in a plant species (Michelmore and Meyers, 1998). The intraspecific haplotype divergence versus interspecific divergence in the evolution of R genes has been studied in two of three complex R-gene families that encode the LRR domain. Studies of tomato (Lycopersicon esculentum) Cf-4/9 and Cf-2/5 complex loci encoding eLRR-TM proteins have revealed that the degree of haplotype diversity can determine the rate of generation of novel R gene and that haplotype polymorphism permits rapid exchange to create sequence diversity in these families (Parniske et al., 1997; Dixon et al., 1998). Analysis of the RPP5 complex locus encoding NBS-LRR proteins of Arabidopsis also showed pronounced intraspecific haplotype divergence (Noël et al., 1999). The rice bacterial blight R gene Xa21 belongs to a multigene family that encodes LRR receptor-kinase proteins (Song et al., 1995). The evolution of the Xa21 family has been studied (Wang et al., 1998). However, it is unknown whether interallelic polymorphism is an important component of variation for resistance to pathogens in this type of R protein.

The Xa26 gene, which confers resistance to Xanthomonas oryzae pv oryzae, the cause of bacterial blight disease, belongs to a multigene family that is clustered in the long arm of rice chromosome 11 (Yang et al., 2003; Sun et al., 2004). The encoding product of Xa26 shares 53% sequence similarity with that of Xa21 (Song et al., 1995; Sun et al., 2004). The resistance spectrum mediated by the two genes is different (Wang et al., 1996; Yang et al., 2003): Xa26 confers resistance at both the seedling and adult stages, whereas Xa21 is developmentally regulated and confers full resistance only at the adult stage (Century et al., 1999; Sun et al., 2004). Several other rice R genes for resistance against both bacterial blight disease and fungal blast disease have also been mapped to the chromosomal region that corresponds to the location of the Xa26 family, which suggests that these uncharacterized rice R genes might also be members of this family (Yoshimura et al., 1995; Ahn et al., 1996; Lin et al., 1996; Yu et al., 1996; Chen et al., 1999). Thus, studies of the sequence diversification of the Xa26 family members will not only provide insight into the evolution of pathogen perception by plants conferred by R genes but will also facilitate the identification of new R genes located in this region.

In this study, we analyzed the Xa26 family in four rice haplotypes from three indica varieties—Minghui 63, which carries the Xa26 gene; Teqing, which carries another bacterial blight R gene, Xa4; and 93-11, which carries at least one uncharacterized bacterial blight R gene—and one japonica variety, Nipponbare. Comparative sequence analysis of this genomic region from the four varieties revealed that haplotype polymorphism occurs in numbers, order of arrangement, and the level of diversity among the paralogs. The orthologs are more similar than the paralogs, which indicates a birth-and-death model of evolution. Positive selection on point mutation for diversification is also observed in limited amino acid sites located on the extracellular domain of the encoding products of the Xa26 family.

RESULTS

Organization of the Xa26 Gene Family in Four Rice Varieties

A bacterial artificial chromosome (BAC) clone, 3H8, from indica rice variety Minghui 63, carries the Xa26 gene (Sun et al., 2004). Analysis of approximately 95-kb sequences of 3H8 identified four genes: MRKa (representing receptor-kinase gene a in Minghui 63), Xa26 or MRKb, MRKc, and MRKd, which belong to the Xa26 gene family. The four paralogs covered an approximately 40-kb region (Fig. 1). In addition to Xa26, MRKa and MRKc were predicted to be intact genes with similar size and structure to the Xa26 (Supplemental Fig. 1). MRKd was a pseudogene with two large inserts of 5,881 and 2,092 bp, two frame-shift sites, and one in-frame stop codon (Supplemental Fig. 1). The predicted coding regions of the four paralogs shared 60% to 80% sequence identity.

Figure 1.

Figure 1.

Open in a new tab

The organization of the Xa26 gene family in four rice varieties, Minghui 63 (A), Nipponbare (B), Teqing (C), and 93-11 (D). The members of the Xa26 family and their transcription orientation are indicated by red (phylogenetic group I) and green (phylogenetic group II) arrows. The figure between the members of the family indicates nucleotide sequence identity (%) of predicted coding regions of adjacent members. The dotted slant indicates the border between two sequence contigs. The question mark indicates that the neighborhood of two adjacent contigs is deduced according to sequence similarity and the Xa26 family organization in the other three varieties. The double slants in B indicate that these regions are not on scale. The two-arrowhead line indicates repeat regions. The intergenic regions showing the same color among the four rice varieties indicate that the corresponding regions, flanked by the same orthologs, share more than 95% (purple color), 97% (brown color), or 98% to 100% (light blue color) sequence identity. The accession numbers of the sequence fragments of Nipponbare are AC146937 and AC116367, and the accession numbers of sequence fragments of 93-11 are AAAA02033155, AAAA02033250, AAAA02033253, AAAA02032784, and AAAA02033267 (from left to right).

A BAC clone 26N11 from another indica variety Teqing carrying the Xa4 gene, which confers resistance to bacterial blight disease, was identified using the sequence of the Xa26 family of Minghui 63 as a probe. Sequence analysis of approximately 120 kb of 26N11 identified 10 genes in one family that showed sequence homology with the Xa26 family members and were distributed in tandem in a more than 90-kb region. Four members were alleles of MRKa, Xa26, MRKc, and MRKd in Minghui 63, respectively, according to their sequence similarity and corresponding locations in the family, and thus were named TRKa, TRKb, TRKc, and TRKd. The other six members were designated as TRKe, TRKf, TRKg, TRKh, TRKi, and TRKj (Fig. 1). Four of these, TRKa, TRKb, TRKc, and TRKe, were intact genes and had structures similar to those of MRKa, Xa26, and MRKc (Supplemental Fig. 1). The other six were either truncated or pseudogenes. TRKf, TRKh, TRKi, and TRKj were truncated in the kinase domain, LRR domain, or the region between the LRR and kinase domains (using the paralogs TRKa, TRKb, TRKc, and TRKe as the reference; Supplemental Fig. 1). TRKd, TRKf, TRKg, TRKh, and TRKi each carried one or two in-frame stop codons, and TRKi also contained one frame-shift site. In addition, the predicted coding regions of TRKd, TRKf, and TRKh each were interrupted by an insert of 3,954, 580, and 167 bp, respectively. The predicted coding regions of TRKa, TRKb, TRKc, TRKd, TRKe, TRKf, and TRKg shared 63% to 80% nucleotide sequence identity. The coding regions of the three heavily truncated paralogs, TRKh, TRKi, and TRKj, also showed sequence similarity with other paralogs of this family in their corresponding regions.

A BLAST search of the genomic sequences of indica variety 93-11 (http://www.genomics.org.cn), using the sequences of the Xa26 family from Teqing, identified five long homologous sequences: AAAA02033155, AAAA02033250, AAAA02033253, AAAA02032784, and AAAA02033267. Analysis of the homologous sequences of the Xa26 family from 93-11 identified eight putative genes in one family. These eight genes in 93-11 showed high degree of sequence similarity respective to TRKa, TRKb, TRKc, TRKd, TRKe, TRKf, TRKg, and TRKh; they were designated 9RKa, 9RKb, 9RKc, 9RKd, 9RKe, 9RKf, 9RKg, and 9RKh. The eight paralogs of 93-11 were arranged in the same way as their orthologs in Teqing in terms of their linear order of distribution in the family and the transcription orientations, and they covered a region of more than 83 kb (Fig. 1). In addition, most of the intergenic regions of the Xa26 family also shared 99% to 100% sequence identity between 93-11 and Teqing (Fig. 1). The sequences of all the members in 93-11, except 9RKc and 9RKg, were identical to that of their orthologs in Teqing. 9RKg shared 99.7% sequence identity with TRKg. The intactness of 9RKc could not be determined because of the unfinished sequencing of the 93-11 genome in this region; however, 9RKc showed 100% sequence identity to the corresponding region of TRKc in Teqing. 9RKa, 9RKb, and 9RKe were predicted to be intact genes (Supplemental Fig. 1). The other four members, 9RKd, 9RKf, 9RKg, and 9RKh, were pseudogenes, each with an in-frame stop codon, a frame-shift site, and/or an insert or truncation (Supplemental Fig. 1). The predicted coding regions of the untruncated paralogs in 93-11 shared 63% to 80% nucleotide sequence identity among themselves.

Two overlapping sequences, AC146937 and AC116367 harboring the Xa26 family, were identified from the finishing genomic sequences of the japonica variety Nipponbare by BLAST analysis, using the Xa26 family sequences from the other three varieties described above as queries. The Xa26 family in Nipponbare covered a region of approximately 230 kb and consisted of 12 paralogs that were named NRKa1, NRKa2, NRKb1, NRKb2, NRKc1, NRKc2, NRKd1, NRKd2, NRKe, NRKf1, NRKf2, and NRKf3 on the basis of sequence similarity to their orthologs, a, b, c, d, e, and f of the other three varieties (Fig. 1). NRKa2, NRKe, NRKf1, and NRKf3 were predicted to be intact genes, whereas the other eight members—NRKa1, NRKb1, NRKb2, NRKc1, NRKc2, NRKd1, NRKd2, and NRKf2—were pseudogenes with an in-frame stop codon, a frame-shift site, and/or an insert or truncation (Supplemental Fig. 1). The predicted coding regions of the 12 paralogs shared 61% to 98% sequence identity with each other.

Compared with the organization of the Xa26 family in the other three varieties, it was easy to recognize that the 12 paralogs of Nipponbare could be divided into two clusters according to their physical location. NRKa1, NRKb1, NRKd1, NRKe, and NRKf1 were in cluster 1, and NRKa2, NRKb2, NRKc1, NRKc2, NRKd2, NRKf2, and NRKf3 were in cluster 2. Although most of the intergenic regions of the Xa26 family shared very high degree of sequence identity between Teqing and 93-11, the corresponding intergenic regions between Minghui 63 and Nipponbare, as well as between Minghui 63 or Nipponbare and Teqing or 93-11, generally had very low sequence similarity. However, some intergenic regions with more than 95% sequence identity were identified among the Xa26 family of Teqing or 93-11 and cluster 1 of Nipponbare (Fig. 1), which indicates that cluster 1 could be generated by a tandem duplication through an unequal crossover between indica and japonica subspecies in recent cross-breeding.

Comparison of the Sequences of the Xa26 Family Members

Analysis of the sequences showed that all the members of the Xa26 family in the four rice varieties were composed of two exons and one short intron located in the kinase domain (Supplemental Fig. 1). Comparative sequence analyses of the predicted coding regions and deduced proteins of the paralogs in each variety revealed pronounced divergence in rice lines 93-11, Teqing, and Minghui 63 (Fig. 2). Most of the paralogous comparisons in Nipponbare also showed pronounced divergence, but a few showed limited sequence variance. Comparison of the nucleotide-site difference and amino acid-site difference among orthologs showed that the average divergences within each ortholog comparison of a, b, c, and d were much lower than that within each paralog comparison (Fig. 2). In addition, the orthologs within the b and d groups had lower rates of sequence divergence than orthologs within the a and c groups.

Figure 2.

Figure 2.

Open in a new tab

The percentage differences of nucleotide sites in the predicted coding region and amino acid sites in deduced proteins of the Xa26 family in four varieties. The results in each catalog are from pairwise comparisons. The sequences were aligned using the ClustalX program (Thompson et al., 1997), and sequence differences were calculated using the MEGA program (Nei and Kumar, 2000). The first four columns (93, Nip, TQ, and MH) presenting nucleotide-site difference or amino acid-site difference show comparisons of paralogs within each variety, and the subsequent four (a, b, c, and d) are comparison of orthologs across varieties. 93, 93-11; all members except the truncated member (9RKh) were included for paralog comparison. TQ, Teqing; all members except the truncated members (TRKh, TRKi, and TRKj) were included for paralog comparison. Nip, Nipponbare; MH, Minghui 63; all members in Nipponbare and Minghui 63 were included for paralog comparisons. a, b, c, and d, The a, b, c, and d members in the four rice varieties. The two a, b, c, and d members in Nipponbare were considered as orthologs in the ortholog comparison. −, Average percentage difference of each comparison.

Phylogenetic analysis classified the Xa26 family members into two groups by either the coding regions or introns (Fig. 3). Group I consisted of the members of a, b, d, and e, whereas group II was composed of the members of c, f, g, h, and j. TRKi in variety Teqing was excluded from the analysis because none of its domains was intact (Supplemental Fig. 1). It is worth noting that all the members in the same phylogenetic group had adjacent locations and the same transcription orientation in three indica varieties and cluster 1 of the japonica variety Nipponbare (Fig. 1). However, the family members of the same phylogenetic group in cluster 2 of Nipponbare had different transcription orientations and locations. The difference between cluster 2 of Nipponbare and the indica varieties was caused by two changes: (1) the relative positions of NRKb2 and NRKf2, and (2) the transcription orientation of NRKb2, NRKf2, and NRKf3 (Fig. 1). Two 465-bp regions located near NRKb2 of Nipponbare and TRKb of Teqing shared 94% sequence identity with each other, which suggests that an inversion event of a fragment carrying NRKb2 and NRKf2 might have occurred in Nipponbare during its evolution (Fig. 1). There was a 12% nucleotide difference between NRKc1 and NRKc2 and an 18% difference between NRKf2 and NRKf3, which demonstrated low sequence divergences among the pairwise-paralog comparisons of the Xa26 family members in Nipponbare (Fig. 2). Thus, NRKf3 and NRKc2 could be generated through tandem duplication of the fragment carrying NRKf2, NRKb2, and NRKc1, followed by a deletion of NRKb2 in the duplicated fragment and subsequent modifications of the original and duplicated members during the course of evolution (Fig. 1).

Figure 3.

Figure 3.

Open in a new tab

Phylogenetic analysis of the Xa26 family members in four rice varieties. The phylogenetic trees are neighbor-joining trees for nucleotide sequences of different regions. The trees for the kinase domain (A), LRR domain (B), and complete coding region (C) were constructed using Jukes-Cantor distances. The tree for the intron (D) was constructed using p distance. The numbers for interior branches indicate the bootstrap values for 1,000 replications. The scale at the bottom is in units of nucleotide substitutions per site. The rice bacterial blight R gene Xa21 (Song et al., 1995) was used as the outgroup.

The phylogenetic analysis also showed that the orthologs of the Xa26 family frequently formed a monophyletic cluster with high bootstrap values in the phylogenetic trees (Fig. 3). In general, the orthologs of the family were more similar to each other than to the paralogs, not only among the three indica varieties but also between indica and japonica varieties.

Comparative Sequence Analysis Revealed Both Intralocus and Interlocus Recombination during the Evolution of the Xa26 Family

Comparative sequence analysis of the family members in the four rice varieties showed that some family members were generated by intralocus (between orthologs) or interlocus (between paralogs) genetic exchange. Unequal crossovers resulted in duplication or deletion of some members in different rice varieties. The members of the Xa26 family in Nipponbare were divided into two clusters. Only cluster 1 shows high degree of sequence similarity in the intergenic regions flanking the a, b, d, and e orthologs with that in Teqing and 93-11 (Fig. 1), suggesting that the cluster 1 may have been generated by unequal crossover between the progenitors of Teqing or 93-11 and Nipponbare. Cluster 2 of Nipponbare may represent the original japonica Xa26 family because the divergence time between indica and japonica rice should be reflected by these differences between the intergenic regions of the Xa26 family members. Both TRKf and 9RKf had a truncated 3′ exon, which could have resulted from an unequal crossover before the divergence of the two rice varieties. This unequal recombination most likely occurred through an intralocus exchange, given that orthologs of the Xa26 family were more similar to each other than to the paralogs. The large truncation of TRKh, TRKi, and 9RKh may also have been the result of unequal crossovers (Fig. 1; Supplemental Fig. 1).

A 3.6-kb region including part of TRKg and its 3′ flanking region showed 99% sequence identity to TRKj and its 3′ flanking region, which indicates that an interlocus unequal crossover generated TRKj (Fig. 1). Among the orthologs of the Xa26 family in the four rice varieties, the members of the a group showed the greatest sequence divergence (Fig. 2). NRKa2 of Nipponbare had an approximately 14% nucleotide difference from NRKa1, MRKa, TRKa, or 9RKa, whereas nucleotide differences among NRKa1, MRKa, TRKa, and 9RKa were less than 1.6%. The pronounced difference between NRKa2 and the other four a members occurred mainly in the LRR region (Fig. 3). The nucleotide differences in the kinase domain among the five a members were less than 2.3%, but the differences in the LRR domain between NRKa2 and the other four a members—NRKa1, MRKa, TRKa, and 9RKa—ranged from 19.7% to 19.9%. These results imply that the LRR region of NRKa2 might have been formed by a crossover with a paralog or another gene that does not belong to the Xa26 family.

The Geneconv program was used to determine gene conversion tracts among the coding sequences of all genes in the Xa26 gene family. A total of 70 significant gene-conversion events were detected (Supplemental Fig. 2). The gene-conversion fragments ranged from 25 to 1,999 bp, and about one-third of the fragments were less than 100 bp. Most of the gene conversions were found in the kinase domain, and only 12 were detected in the LRR domain, which included three gene-conversion events spanning the LRR domain and the region between the LRR and kinase domains. This result implies that gene conversion may not be the major evolutionary force for the diversification of the LRR domain of the Xa26 family.

The Hypervariable Sites of the LRR Region

The LRR domain of R proteins of plants is suggested to interact directly or indirectly with pathogen elicitors to determine race specificity. All the members of the Xa26 family in the four rice varieties, except those with a truncation in the LRR region, were predicted to encode 26 LRR repeats. The LRR consensus sequence consisted of 24 amino acids with a characteristic structure of LxxLxxLxxLxLxxNxLxGxIPxx (or xxLxLxxNxLxGxIPxxLxxLxxL). The xxLxLxx motifs in the LRRs are predicted to form a solvent-exposed parallel β-sheet. The conserved Leu (L) project into the hydrophobic core, whereas the five interstitial residues (x) form a solvent-exposed surface that is involved in ligand binding (Kobe and Deisenhofer, 1995; Jones and Jones, 1997). It has been reported that the hypervariability of solvent-exposed amino acids correlates with differential pathogen recognition in some R proteins of the LRR-TM and NBS-LRR types (Parniske et al., 1997; Thomas et al., 1997; Botella et al., 1998; Dixon et al., 1998; McDowell et al., 1998; Meyers et al., 1998; Ellis et al., 1999). Different regions of the deduced protein sequences of the Xa26 family in the four rice varieties were aligned and compared by pairwise analysis. The results showed that the xx(L)x(L)xx motifs (excluding the conserved aliphatic residues in parentheses) of LRRs were the most highly divergent region, with 61.0% average difference of amino acid sites. The average differences of amino acid sites in the rest of the region [(N)x(L)x(G)x(IP)xx(L)xx(L)xx (excluding the conserved aliphatic residues in parentheses)] of the LRR domain, the conserved aliphatic residues of LRR domain, the region between the LRR domain and kinase domain, and the kinase domain were 44.4%, 20.5%, 40.2%, and 20.7%, respectively.

The number of alternate amino acid residues at a given site was further examined by alignment of deduced amino acid sequences of the Xa26 family members in the four rice varieties using the ClustalX program. The alignment results showed that all the regions of these LRR receptor-kinase proteins of the family contained variable amino acid sites (Supplemental Table I). However, the xx(L)x(L)xx motif was more variable than other regions. All of the 45 hypervariable sites (Parniske et al., 1997; with seven or more alternate amino acids per site) were located in the LRR domain. In the LRR domain, approximately 78% (35) of the hypervariable sites were solvent-exposed amino acid sites, and another 13% (six) hypervariable sites were distributed at the xIPxx region (Fig. 4). Most of the hypervariable sites (71%) clustered in the region between LRR8 and LRR18. It is notable that two hypervariable sites were found at the conserved aliphatic residues of LRR domain (Fig. 4).

Figure 4.

Figure 4.

Open in a new tab

Hypervariable sites and positively selected sites in the LRR domain of the Xa26 family. Only the sequence of the LRR domain of XA26 (or MRKB) is shown. The LRR alignments of the XA26 are according to the reports of Sun et al. (2004). Hypervariable sites and positively selected sites of the Xa26 family are superimposed on the sequence of XA26. Hypervariable sites (seven or more different amino acids found per site) are underlined. The positively selected sites in which posterior probability is more than 0.9 are in bold, and the sites in which posterior probability is between 0.5 and 0.9 are in lowercase.

Identification of Amino Acid Sites under Positive Selection

The whole coding regions of the Xa26 family members, except those encoding the signal peptide (corresponding to the amino acid residues from N terminal to the Ala at position 33 of XA26) and the COOH end (corresponding to the amino acid residues from Ile at position 1,086 to C terminal of XA26), were tested for positive selection at individual amino acid sites. Two likelihood ratio (LR) tests were applied. The first LR test compared M3, the free-ratio model that assumes independent ω (the ratio of nonsynonymous nucleotide substitution [dN] to synonymous nucleotide substitution [dS], dN:dS) values for every branch of the phylogeny, versus M0, the null one-ratio model that constrains ω to be equal on all phylogenetic branches (Yang, 1997; Nielsen and Yang, 1998). This test was significant at the P < 0.01 level, indicating that selective pressure indeed varies among amino acid sites. The proportion of amino acid sites potentially under positive selection was 7.8%, with ω = 2.08 under M3. Almost all of the positive-selection sites were found in the LRR domain, except six sites in the region between the signal peptide and the LRR domain.

The second LR test, a more specific test that examined variation in ω among sites, compared models M8 with M7 (Yang et al., 2000). M8 fitted the data significantly (at P < 0.01 level) better than M7. Approximately 7.0% of the amino acid sites of the Xa26 family were potentially under positive selection with ω = 2.15 under M8 (Supplemental Table II). All of the positive-selection sites, except four, were located in the LRR domain. The other four potentially positive-selection sites were at amino acid positions 33, 82, 83, and 85 in the region between the signal peptide and the LRR domain. Forty-six of 58 sites that were potentially under positive selection in the LRR domain were solvent-exposed amino acid residues (Fig. 4). Approximately 73% (33) of the hypervariable sites detected in the LRR domain were subject to positive selection.

For comparison, the positive-selection sites of other R-gene families, the rice Xa21 family and the tomato Cf-4/9 family, which all encode extracytoplasmic LRR domains, were also analyzed using the same two LR tests (Supplemental Table II). Eight sites were potentially under positive selection in the LRR domain of the Xa21 family, and six of these eight were solvent-exposed amino acid sites (Supplemental Fig. 3). Fifty-two of 68 sites that were potentially under positive selection were located in the LRR domain of the Cf-4/9 family (Supplemental Fig. 3). Approximately 62% (32) of sites under positive selection in the LRR domain were solvent-exposed amino acid residues. Approximately 24% (16) of the positive-selection sites were discovered outside the LRR region in the Cf-4/9 family. Among the 24% sites, one was in the signal peptide; 10, most of which had posterior probability >0.95, were clustered in the region between the signal peptide and the LRR domain; and five were in the C-terminal region of the proteins (Supplemental Table II).

DISCUSSION

The Xa26 Family Has Extensive Paralog Diversity

Several R-gene families encoding the extracytoplasmic LRR domain show a high degree paralog similarity (Parniske et al., 1997; Song et al., 1997). Although the average sequence similarity between the two classes (Xa21 and A2) of the rice Xa21 family is only 72.4%, the average nucleotide identities of the members within the Xa21 and A2 classes are 98.0% and 95.2%, respectively (Song et al., 1997). The nucleotide sequence identities between paralogs range from 92.1% to 95.5% in the Cf-4-carrying haplotypes and from 92.3% to 95.3% in the Cf-9-carrying haplotypes in the tomato Cf-4/9 family (Parniske et al., 1997). The high degree of sequence similarity between paralogs in the Cf-4/9 family and the classes of the Xa21 family may imply that a series of recent duplication and unequal crossover produced most of the members of these gene families or that a large amount of gene conversion among the paralogs resulted in homogenization.

This study identified a high degree of nucleotide sequence divergence between paralogs in all four rice haplotypes (Fig. 2). Like the Xa21 family, the paralogs of the Xa26 family in each rice variety were also classified into two groups. However, the sequences of paralogs were highly divergent, even in the same group. The average nucleotide sequence identities of paralogs were 77.9%, 80.3%, 79.0%, and 79.6% for phylogenetic group I in Minghui 63, Nipponbare, Teqing, and 93-11, and 73.3%, 67.2%, and 67.8% for phylogenetic group II in Nipponbare, Teqing, and 93-11 (the heavily truncated members TRKh, TRKi, TRKj, and 9RKh were excluded), respectively. The average nucleotide sequence identities between the two phylogenetic groups were even lower—64.5%, 63.2%, 64.0%, and 64.2% in Minghui 63, Nipponbare, Teqing, and 93-11, respectively. Duplication and the subsequent divergence of a progenitor have been suggested to be a way to form a multigene family with clustered homologous members. Thus, the degree of intraspecific sequence variation of a gene family should represent the relative age, the mutation rate, or the level of unequal recombination (or gene conversion) of a family. Since the LRR domain was highly variable compared with the kinase domain and most of the gene-conversion events detected were in the kinase domain of the Xa26 family, the extensive paralog diversity of the family may suggest that it is an evolutionary old family and/or it has been subject to a higher rate of mutation and a lower level of unequal recombination or gene conversion in the LRR domain.

Only a Small Number of Amino Acid Sites of R Proteins Carrying the Extracytoplasmic LRR Domain Are Subject to Positive Selection

Evidence for positive evolution has been identified in the xx(L)x(L)xx motif of the LRR domain in some R-gene families by use of the average ω analysis in which the ratio ω of dN:dS was averaged over all sites (Parniske et al., 1997; Botella et al., 1998; Meyers et al., 1998; Wang et al., 1998; Noël et al., 1999). However, the average ω analysis cannot examine the proportions and exact locations of the sites for positive selection. Although some studies have reported hypervariable sites in the LRR domain, the present results show that hypervariable sites do not completely overlap with the sites identified under positive selection. The codon-based models implemented in the maximum-likelihood framework are powerful methods to test for positive selection at individual amino acid sites (Yang, 1997; Nielsen and Yang, 1998; Yang et al., 2000). These methods have been used to detect the positive selection in NBS-LRR and receptor-like kinase proteins in Arabidopsis and lettuce (Mondragón-Palomino et al., 2002; Kuang et al., 2004; Strain and Muse, 2005). Using these models, we identified that only 7% of amino acid sites of the Xa26 family were potentially under positive selection. Approximately 74% of the positively selected sites were solvent-exposed amino acid residues of the xx(L)x(L)xx motif (Fig. 4). These results suggest that only very small amounts of amino acid sites putatively involved in pathogen recognition specificity are subject to positive selection for diversification in the Xa26 family. The other regions of the Xa26 family, especially the kinase domain, are under strong purifying selection, owing to functional constraint.

Analysis of the rice Xa21 family and tomato Cf-4/9 family also revealed that only a small number of amino acid sites, most of which were solvent-exposed amino acid residues, were under positive selection (Supplemental Table II). However, positively selected sites were also observed in the region between the signal peptide and the LRR domain in both the Xa26 family and the Cf-4/9 family. Domain-swap or recombination studies revealed that pathogen resistance specificity was affected by the Toll/interleukin-1 receptor homology region in flax (Linum usitatissimum) L6 protein and by the determinants residing in the region between the signal peptide and the LRR domain in the tomato Cf-4 protein (Ellis et al., 1999; Luck et al., 2000; Van der Hoorn et al., 2001). The present results suggest that the positively selected sites locating between the signal peptide and LRR domain may be also involved in race specificity in the Xa26 and Cf-4/9 families.

The Xa26 Family Clearly Experienced Evolution by a Birth-and-Death Process

It has been suggested that a multigene family would be subject to evolution by a birth-and-death mechanism in which the genes in a family are formed first by duplication, followed by diversification, deletion, or dysfunction of the original genes (Nei et al., 1997). The feature of the birth-and-death model is that the alleles (or orthologs) of a multigene family from different species or haplotypes are more closely related to the members (or paralogs) of the same family from the same species or haplotype. The vertebrate major histocompatibility complex (MHC) gene family and the immunoglobulin gene family, which encode products that interact with invading parasites, have been subject to the birth-and-death evolution (Nei et al., 1997). Like its functional counterpart in vertebrates, plant complex R-gene families have also been proposed to evolve by a birth-and-death process (Michelmore and Meyers, 1998). However, most of the evidence came from the deduction by analyzing one haplotype; only the RPP5 gene family seems to provide some clues to support the model (Noël et al., 1999). The present results provide more evidence to support the hypothesis of the evolution of R-gene families. Comparison of the sequences of the Xa26 family members clearly showed that orthologs were more similar than paralogs (Fig. 3).

The Majority of the Sequence Diversity of the Xa26 Family Members Was Caused by Accumulation of Mutations and Positive Selection

The specificity of R genes may be generated by several mechanisms, such as unequal crossover, gene conversion, intergenic recombination, and gene mutations, but there is still controversy about the major mechanism. Michelmore and Meyers (1998) suggest that the majority of changes in specificity are caused by interallelic recombination and gene conversion that alter the combinations and/or orientations of the arrays of solvent-exposed residues in the LRR region. This mechanism could make the R genes generating considerable variation and new ligand-binding points rapidly without the need for high rates of nucleotide substitution. There is some evidence showing that the interallelic recombination and gene conversion are more important than point mutation on the generation of polymorphism in mammalian MHC genes (Watkins et al., 1992; Parham and Ohta, 1996), but less evidence supports the point of view on plant R-gene families because of only a few papers concerning the comparisons between different haplotypes (Parniske et al., 1997; Dixon et al., 1998; Noël et al., 1999). Furthermore, it is difficult to discern which is more important between orthologous or paralogous recombination. The analysis of the Xa26 gene family in four haplotypes showed only a few recombination events occurred in the LRR region and a lot of the events in kinase region. Therefore, the gene recombination should not be the major force that causes the diversity of the Xa26 gene family members.

The solvent-exposed residues of the LRR domain are hypervariable in the Xa26 family. Such changes are more likely to have arisen from point mutations than from sequence exchanges. The cultivated rice is a self-pollination plant, and most of the genes in its genome are homozygous. Thus, there are a few chances to alter the gene composition through gene recombination. An inbreeding mating system will tend to favor the accumulation of structural variants and point mutations because they would rapidly become homozygous. This mechanism with positive selection would result in the accumulation of point mutations and the different physical organization of the gene family among rice varieties, as shown in the Xa26 gene family. The fact that most of the hypervariable sites in the LRR region of the Xa26 gene family are overlapping with the positively selected sites also supports this inference.

Stability of the Xa26 Family and Generation of Novel Resistance Specificity in a Complex R Locus

An inbreeding plant would tend to promote genetic instability in crosses with near relatives but repress recombination in more distant crosses. Outbreeding species, such as maize (Zea mays), may be more unstable as duplications would tend to be hemizygous, therefore promoting a variety of pairing possibilities (Michelmore and Meyers, 1998). The instability of R genes has been observed in the Rp1 locus of maize as well as in the Cf-2/5 and Cf-4/9 loci of tomato (Sudupak et al., 1993; Parniske et al., 1997; Dixon et al., 1998). We had also noticed the instability of the Xa26 family. Five of 241 F9 recombinant inbred lines, developed by a single seed descendent from a cross between Zhenshan 97 and Minghui 63, carry a new member but a loss of member MRKa (Yang et al., 2003; Sun et al., 2004). In an F2 population segregating for the Xa4 gene that is likely the member of the Xa26 family, one of 640 susceptible F2 plants lost the a and d members of this family (X. Sun and S. Wang, unpublished data). Rice lines IRBB3 and Zhachanglong were originally known to carry Xa3 and Xa22(t) genes for bacterial blight resistance, respectively (Yoshimura et al., 1995; Lin et al., 1996). Further study has revealed that IRBB3 and Zhachanglong carry Xa26, and that Xa3, Xa22(t), and Xa26 are likely the same gene (Sun et al., 2004). Some susceptible plants were identified in the progenies of the cross between IRBB4 (Xa4 haplotype) and IRBB3 (Xa26 haplotype), as well as of the cross between Zhachanglong (Xa26 haplotype) and IRBB4 (Ogawa et al., 1986; Lin et al., 1996; Sun et al., 2004).

Although recombination easily causes the loss of resistance, this mechanism could be used to generate novel R genes. The clustered R-gene family members provide a continually horizontal sequence reservoir for generating novel recognition specificity. Crossing with different rice varieties and providing more chances for intergenic recombination could promote the generation of new R genes for different pathogens. In addition, because the cultivated rice infrequently cross-pollinates and wild rice (Zizania palustris) occurs more frequently than cultivar, the cross-breeding strategy between cultivated rice and wild rice for rice improvement has been used widely during recent years.

Toward a Model of Xa26 Family Evolution

The sequences from four haplotypes provide clear evidence to propose a model for the evolution of the Xa26 family. Duplication of a progenitor gene through an event, which perhaps involves aberrant behavior of the DNA replication fork, as proposed by Noël et al. (1999), produced two tandem repeated genes in inverse orientation. Divergence of the duplicated genes generated the ancestors of the two phylogenetic groups with opposite transcription orientations. Further tandem duplication of the ancestors through a series of unequal recombination events created the members of the two groups. All four haplotypes carry a, b, c, and d members, and three haplotypes, including both indica and japonica subspecies, carry f members (Fig. 1), which indicates that phylogenetic group I (containing at least a, b, and d members) and group II (containing at least c and f members) formed before the separation of the indica and japonica subspecies. The four haplotypes were generated through further duplication to create e, g, h, i, and j members, deletion, recombination, and point mutation of the Xa26 family members after the formation of indica and japonica subspecies. The major events for the formation of these haplotypes include the following events. The f member was deleted from the ancestor indica rice during the generation of haplotype Minghui 63. A block duplication of the Xa26 family occurred from an unequal crossover between ancestor indica and japonica rice during the generation of haplotype Nipponbare. Inversion of a fragment containing the ancestors of NRKb2 and NRKf2 and subsequent duplication of a fragment containing the ancestors of NRKf2, NRKb2, and NRKc1, followed by deletion of the duplicated ancestor of NRKb2, completed the final major changes in the generation of Nipponbare. The g members were more closely associated with f members, and h members were associated with c members according to phylogenetic analysis. Thus, the g and h members were generated by tandem duplication of f and c members, respectively, through unequal interallelic recombination before the formation of haplotypes Teqing and 93-11. Most of the regions of the Xa26 family, including the coding and intergenic regions, in Teqing and 93-11 shared approximately 99% to 100% sequence identity. Therefore, the formation of Teqing and 93-11 may be due to shared ancestry or recent cross-breeding.

CONCLUSION

Compared with the Xa21 and Cf4/9 gene families, the Xa26 family exhibits much higher divergence among paralogs and different physical structure among haplotypes. Phylogenetic analysis indicates that the Xa26 gene family clearly experienced evolution by a birth-and-death process. Most positively selected sites locate on the solvent-exposed residues in the LRR region and are overlapping with the hypervariable sites. These results suggest that the majority of changes in the diversity of Xa26 family members were caused by accumulation of mutations with positive selection. Recombination between orthologs or paralogs and unequal crossover were observed, but these contribute less to the generation of novel variation in the Xa26 gene family. The Xa26 family would tend to promote instability in crosses with near relatives, which could be used to generate the novel resistance specificity in this complex R locus.

MATERIALS AND METHODS

Materials

A BAC clone, 3H8, from rice variety Minghui 63 (Oryza sativa subsp. indica) carrying the Xa26 gene (Sun et al., 2004) was used for sequencing. A BAC library constructed using rice variety Teqing (O. sativa subsp. indica, kindly provided by Dr. Hongbin Zhang of Texas A&M) was used for screening BAC clones containing homologs of the Xa26 family.

DNA Sequencing

The nucleotide sequences of rice BAC clones were determined using a shotgun approach. A partial digestion of the BAC clones with the restriction enzyme Sau3AI was also used to construct bigger fragment subclones of the BAC clones. The M13 universal forward and reverse primers were used for sequencing. The sequences were assembled using the computer program Sequencher 4.1.2 (Gene Codes).

DNA and Protein Sequence Analysis

The similarity analyses of DNA and protein sequences were performed using BLAST programs, including BLASTN, BLASTX, and BLASTP (Altschul et al., 1997). The gene prediction programs used were GENSCAN (Burge and Karlin, 1997) and FGENESH (http://www.softberry.com). The BLASTX program was frequently used to find possible coding regions of a gene. The results of BLASTX analyses were compared with those of GENSCAN and FGENESH analyses to determine the coding region, intron, insert, in-frame stop codon, and frame-shift site. Multiple-sequence alignment was performed using the ClustalX program (Thompson et al., 1997). The Molecular Evolutionary Genetic Analysis (MEGA) program (Nei and Kumar, 2000) was used to generate phylogenetic trees using Jukes-Cantor distances and p distance, and to display informative polymorphic sites in multiple-sequence alignment. Geneconv program (http://www.math.wustl.edu/∼sawyer/geneconv) was used to identify gene conversion events (Sawyer, 1989), and the default settings with 10,000 permutations were used for the analysis. The statistical significance of gene conversion events was defined as a global permutation P value of ≤0.05.

Analysis of Positive Selection

The deduced amino acid sequences of the Xa26 family members were first aligned using the ClustalX program. The amino acid alignment was then used to guide the alignment of nucleotide sequences of the family using the protal2dna program (http://bioweb.pasteur.fr/seqanal/interfaces/protal2dna.html). The codon-based models implemented in the maximum-likelihood method were applied to infer amino acid sites under positive selection by estimating the ratio ω (dN:dS; Nielsen and Yang, 1998; Yang et al., 2000). The Codeml program in Phylogenetic Analysis by Maximum Likelihood (http://abacus.gene.ucl.ac.uk/software/paml.html) was used to calculate ω. The analysis consisted of two major steps. The first major step used the LR test to test for positive selection by comparing a null model that did not allow for sites with ω > 1 against a model that did. Two LR tests were used. The first compared models M3 with M0, a test for variation of ω among sites, but positive selection was implicated when ω > 1 under M3. The second LR test compared models M8 with M7, a more specific approach for detection of site under positive selection. This test has been applied widely (Yang et al., 2000; Swanson et al., 2001; Mondragón-Palomino et al., 2002; Kuang et al., 2004; Strain and Muse, 2005). The second major analysis step was to examine the probability of the sites with ω > 1. This was achieved using Bayes theorem to calculate the posterior probability. The sites with a high posterior probability were likely to be under positive selection. The statistical description of the above models and notation, as well as probability analysis, refer to the reports by Nielsen and Yang (1998), Yang et al. (2000), and Swanson et al. (2001).

Sequence data from this article can be found in the EMBL/GenBank data libraries under accession numbers DQ355952 to DQ355956.

Supplementary Material

Supplemental Data
plntphys_pp.105.073080_index.html (1.3KB, html)
1

This work was supported by grants from the National Program on the development of Basic Research in China and the National Natural Science Foundation of China.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantphysiol.org) is: Shiping Wang (swang@mail.hzau.edu.cn).

[W]

The online version of this article contains Web-only data.

Article, publication date, and citation information can be found at www.plantphysiol.org/cgi/doi/10.1104/pp.105.073080.

References

  1. Ahn SN, Kim YK, Han SS, Choi HC, Moon HP, McCouch SR (1996) Molecular mapping of a gene for resistance to a Korean isolate of rice blast. Rice Genet Newsl 13: 74–76 [Google Scholar]
  2. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25: 3389–3402 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Botella MA, Parker JE, Frost LN, Bittner-Eddy PD, Beynon JL, Daniels MJ, Holub EB, Jones JDG (1998) Three genes of the Arabidopsis RPP1 complex resistance locus recognize distinct Peronospora parasitica avirulence determinants. Plant Cell 10: 1847–1860 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Burge C, Karlin S (1997) Prediction of complete gene structures in human genomic DNA. J Mol Biol 268: 78–94 [DOI] [PubMed] [Google Scholar]
  5. Caicedo AL, Schaal BA (2004) Heterogeneous evolutionary processes affect R gene diversity in natural populations of Solanum pimpinellifolium. Proc Natl Acad Sci USA 101: 17444–17449 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Century KS, Lagman RA, Adkisson M, Morlan J, Tobias R, Schwartz K, Smith A, Love J, Ronald PC, Whalen MC (1999) Developmental control of Xa21-mediated disease resistance in rice. Plant J 20: 231–236 [DOI] [PubMed] [Google Scholar]
  7. Chen DH, Vina M, Inukai T (1999) Molecular mapping of the blast resistance gene, Pi44(t), in a line derived from a durably resistant rice cultivar. Theor Appl Genet 98: 1046–1053 [Google Scholar]
  8. Dangl JL, Jones JD (2001) Plant pathogens and integrated defence responses to infection. Nature 411: 826–833 [DOI] [PubMed] [Google Scholar]
  9. Dixon MS, Hatzixanthis K, Jones DA, Harrison K, Jones JDG (1998) The tomato Cf-5 disease resistance gene and six homologs show pronounced allelic variation in leucine-rich repeat copy number. Plant Cell 10: 1915–1925 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dodds PN, Lawrence GJ, Ellis JG (2001. a) Six amino acid changes confined to the leucine-rich repeat beta-strand/beta-turn motif determine the difference between the P and P2 rust resistance specificities in flax. Plant Cell 13: 163–178 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dodds PN, Lawrence GJ, Ellis JG (2001. b) Contrasting modes of evolution acting on the complex N locus for rust resistance in flax. Plant J 27: 439–453 [DOI] [PubMed] [Google Scholar]
  12. Ellis JG, Lawrence GJ, Luck JE, Dodds PN (1999) The identification of regions in alleles of the flax rust resistance gene L that determine differences in gene-for-gene specificity. Plant Cell 11: 495–506 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gomez-Gomez L, Boller T (2000) FLS2: an LRR receptor-like kinase involved in the perception of the bacterial elicitor flagellin in Arabidopsis. Mol Cell 5: 1003–1011 [DOI] [PubMed] [Google Scholar]
  14. Hammond-Kosack KE, Parker JE (2003) Deciphering plant-pathogen communication: fresh perspectives for molecular resistance breeding. Curr Opin Biotechnol 14: 177–193 [DOI] [PubMed] [Google Scholar]
  15. Hulbert SH, Webb CA, Smith SM, Sun Q (2001) Resistance gene complexes: evolution and utilization. Annu Rev Phytopathol 39: 285–312 [DOI] [PubMed] [Google Scholar]
  16. Jones DA, Jones JDG (1997) The role of leucine-rich repeat proteins in plant defense. Adv Bot Res 24: 89–167 [Google Scholar]
  17. Kobe B, Deisenhofer J (1995) A structure basis of the interactions between leucine-rich repeats and protein ligands. Nature 374: 183–186 [DOI] [PubMed] [Google Scholar]
  18. Kuang H, Woo S-S, Meyers BC, Nevo E, Michelmore RW (2004) Multiple genetic processes result in heterogeneous rates of evolution within the major cluster disease resistance genes in lettuce. Plant Cell 16: 2870–2894 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lin XH, Zhang DP, Xie YF, Gao HP, Zhang Q (1996) Identifying and mapping a new gene for bacterial blight resistance in rice based on RFLP markers. Phytopathology 86: 1156–1159 [Google Scholar]
  20. Luck JE, Lawrence GJ, Dodds PN, Shepherd KW, Ellis JG (2000) Regions outside of the leucine-rich repeats of flax rust resistance proteins play a role in specificity determination. Plant Cell 12: 1367–1377 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Martin GB, Bogdanove AJ, Sessa G (2003) Understanding the functions of plant disease resistance proteins. Annu Rev Plant Biol 54: 23–61 [DOI] [PubMed] [Google Scholar]
  22. McDowell JM, Dhandaydham M, Long TA, Aarts MGM, Goff S, Holub EB, Dangl JL (1998) Intragenic recombination and diversifying selection contribute to the evolution of downy mildew resistance at the RPP8 locus of Arabidopsis. Plant Cell 10: 1861–1887 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Meyers BC, Shen KA, Rohani P, Gaut BS, Michelmore RW (1998) Receptor-like genes in the major resistance locus of lettuce are subject to divergent selection. Plant Cell 10: 1833–1846 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Michelmore RW, Meyers BC (1998) Clusters of resistance genes in plants evolve by divergent selection and a birth-and-death process. Genome Res 8: 1113–1130 [DOI] [PubMed] [Google Scholar]
  25. Mondragón-Palomino M, Meyers BC, Michelmore RW, Gaut BS (2002) Patterns of positive selection in the complete NBS-LRR gene family of Arabidopsis thaliana. Genome Res 12: 1305–1315 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Nei M, Gu X, Sitnikova T (1997) Evolution by the birth-and-death process in multigene families of the vertebrate immune system. Proc Natl Acad Sci USA 94: 7799–7806 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Nei M, Kumar S (2000) Molecular Evolution and Phylogenetics. Oxford University Press, Oxford
  28. Nielsen R, Yang Z (1998) Likelihood models for detecting positively selected amino acid sites and applications to the HIV-1 envelop gene. Genetics 148: 929–936 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Noël L, Moores TL, van der Biezen EA, Parniske M, Daniels MJ, Parker JE, Jones JDG (1999) Pronounced intraspecific haplotype divergence at the RPP5 complex disease resistance locus of Arabidopsis. Plant Cell 11: 2099–2111 [PMC free article] [PubMed] [Google Scholar]
  30. Ogawa T, Tabien RE, Busto GA, Khush GS, Mew TW (1986) The relationship between Xa-3, Xa-4, and Xa-4b for resistances to rice bacterial blight. Rice Genet Newsl 3: 83–84 [Google Scholar]
  31. Parham P, Ohta T (1996) Population biology of antigen presentation by MHC class I molecules. Science 272: 67–74 [DOI] [PubMed] [Google Scholar]
  32. Parniske M, Hammond-Kosack KE, Goldstein C, Thomas CM, Jones DA, Harrison K, Wulff BBH, Jones JDG (1997) Novel resistance specificities result from sequence exchange between tandemly repeated genes at the Cf4/Cf9 locus of tomato. Cell 91: 821–832 [DOI] [PubMed] [Google Scholar]
  33. Sawyer S (1989) Statistical tests for detecting gene conversion. Mol Biol Evol 6: 526–538 [DOI] [PubMed] [Google Scholar]
  34. Song W-Y, Wang G-L, Chen L-L, Kim H-S, Pi L-Y, Holsten T, Gardner J, Wang B, Zhai W-X, Zhu L-H, et al (1995) A receptor kinase-like protein encoded by the rice disease resistance gene, Xa21. Science 270: 1804–1806 [DOI] [PubMed] [Google Scholar]
  35. Song W-Y, Pi L-Y, Wang G-L, Gardner J, Holsten T, Ronald PC (1997) Evolution of the rice Xa21 disease resistance gene family. Plant Cell 9: 1279–1287 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Strain E, Muse SV (2005) Positively selected sites in the Arabidopsis receptor-like kinase gene family. J Mol Evol 61: 325–332 [DOI] [PubMed] [Google Scholar]
  37. Sudupak MA, Bennetzen JL, Hulbert SH (1993) Unequal exchange and meiotic instability of disease-resistance genes in the Rp1 region of maize. Genetics 133: 119–125 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Sun X, Cao Y, Yang Z, Xu C, Li X, Wang S, Zhang Q (2004) Xa26, a gene conferring resistance to Xanthomonas oryzae pv. oryzae in rice, encodes an LRR receptor kinase-like protein. Plant J 37: 517–527 [DOI] [PubMed] [Google Scholar]
  39. Swanson WJ, Yang Z, Wolfner MF, Aquadro CF (2001) Positive Darwinian selection drives the evolution of several female reproductive proteins in mammals. Proc Natl Acad Sci USA 98: 2509–2514 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Thomas CM, Jones DA, Parniske M, Harrison K, Balint-Kurti PJ, Hatzixanthis K, Jones JDG (1997) Characterization of the tomato Cf-4 gene for resistance to Cladosporium fulvum identifies sequences that determine recognitional specificity in Cf-4 and Cf-9. Plant Cell 9: 2209–2224 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG (1997) The ClustalX windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 24: 4876–4882 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Van der Hoorn RA, Roth R, de Wit PJ (2001) Identification of distinct specificity determinants in resistance protein Cf-4 allows construction of a Cf-9 mutant that confers recognition of avirulence protein Avr4. Plant Cell 13: 273–285 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Wang G-L, Ruan D-L, Song W-Y, Sideris S, Chen L, Pi L-Y, Zhang S, Zhang Z, Fauquet C, Gaut BS, Whalen MC, Ronald PC (1998) Xa21D encodes a receptor-like molecule with a leucine-rich repeat domain that determines race-specific recognition and is subject to adaptive evolution. Plant Cell 10: 765–779 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Wang GL, Song WY, Ruan DL, Sideris S, Ronald PC (1996) The cloned gene, Xa21, confers resistance to multiple Xanthomonas oryzae pv. oryzae isolates in transgenic plants. Mol Plant Microbe Interact 9: 850–855 [DOI] [PubMed] [Google Scholar]
  45. Watkins DI, McAdam SN, Liu X, Strang CR, Milford EL, Levine CG, Garber TL, Dogon AL, Lord CI, Ghim SH, et al (1992) New recombinant HLA-B alleles in a tribe of South American Amerindians indicate rapid evolution of MHC class I loci. Nature 357: 329–333 [DOI] [PubMed] [Google Scholar]
  46. Yang Z (1997) PAML: a program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci 13: 555–556 [DOI] [PubMed] [Google Scholar]
  47. Yang Z, Sun X, Wang S, Zhang Q (2003) Genetic and physical mapping of a new gene for bacterial blight resistance in rice. Theor Appl Genet 106: 1467–1472 [DOI] [PubMed] [Google Scholar]
  48. Yang ZH, Nielsen R, Goldman N, Pedersen AMK (2000) Codon-substitution models for heterogeneous selection pressure at amino acid sites. Genetics 155: 431–449 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Yoshimura S, Yoshimura A, Iwata N, McCouch SR, Abenes ML, Baraoidan MR, Mew TW, Nelson RJ (1995) Tagging and combining bacterial blight resistance genes in rice using RAPD and RFLP markers. Mol Breed 1: 375–387 [Google Scholar]
  50. Yu ZH, Mackill DJ, Bonman JM, McCouch SR, Guiderdoni E, Notteghem JL, Tanksley SD (1996) Molecular mapping of genes for resistance to rice blast (Pyricularia grisea Sacc.). Theor Appl Genet 93: 859–863 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Data
plntphys_pp.105.073080_index.html (1.3KB, html)
plntphys_pp.105.073080_73080Figure_S1.ppt (62KB, ppt)
plntphys_pp.105.073080_73080Figure_S2.ppt (49.5KB, ppt)
plntphys_pp.105.073080_73080Figure_S3.ppt (34KB, ppt)
plntphys_pp.105.073080_73080Table_S1.doc (35KB, doc)
plntphys_pp.105.073080_73080Table_S2.doc (34.5KB, doc)

Articles from Plant Physiology are provided here courtesy of Oxford University Press

RESOURCES