Resistance genes mediate dominant resistance to pathogens possessing corresponding avirulence (Avr) genes (Collier and Moffett, 2009; Jones and Dangl, 2006). Single dominant R genes typically respond differentially to pathogen strains or races. The typical dominant resistance response is associated with several defense-related events, including rapid induction of reactive oxygen species, phytoalexin accumulation, and activation of salicylic acid biosynthesis and pathogenesis-related genes, which often results in localized necrotic response (Hammond-Kosack and Jones, 1996).
Rsv3, one of three known genes that confer resistance to the Soybean mosaic virus (SMV) in soybean [Glycine max (L.) Merr.], is unlike the well-characterized Rsv1 alleles in terms of the patterns of resistance to seven SMV strain groups (G1–G7 classified on the basis of their virulence; Cho and Goodman, 1979). Various alleles of the Rsv1 locus generally confer extreme resistance to the lower numbered (G1 through G4) strain groups and condition necrotic or mosaic reactions to higher numbered (G5 through G7) groups (Chen et al., 1991). Rsv3 alleles from diverse soybean cultivars including ‘Columbia’, ‘Hardee’, ‘Tousan 140’, and ‘Harosoy’ confer extreme resistance to the higher numbered strain groups (G5 through G7) and condition stem-tip necrosis and/or mosaic symptoms to the lower numbered groups (Tu and Buzzell, 1987; Buzzell and Tu, 1989; Bowers et al., 1992; Gunduz et al., 2002; Ma et al., 2002). Although stem-tip necrosis was proposed as a representative symptom conditioned by Rsv3 in a line derived from Columbia × Harosoy at the time of the first description of this gene (Tu and Buzzell, 1987; Buzzell and Tu, 1989), this symptom has not been observed in ‘L29’, a ‘Williams’ isoline derived from Hardee (Bernard et al., 1991; Gunduz et al., 2000). Pyramiding the Rsv3 gene from L29 with Rsv1 conferred resistance to all strains of SMV, demonstrating the value of this gene for developing durable SMV resistant soybean lines (Saghai Maroof et al., 2008). Rsv1 is associated with a NB-LRR gene cluster on the soybean molecular linkage group (MLG) F (chromosome 13) (Hayes et al., 2004) where multiple disease resistance genes have been identified and there are multiple NB-LRR clusters (Innes et al., 2008; Wawrzynski et al., 2008). Jeong et al. (2002) mapped the Rsv3 locus between markers A519F/R and M3Satt on MLG B2 (chromosome 14). Although the disease responses of the Rsv3 gene against SMV, including extreme resistance and stem-tip necrosis, were typical of those conditioned by NB-LRR genes, the molecular nature of the Rsv3 gene is largely unknown. Interestingly, one end sequence of the restriction fragment length polymorphism marker M1a, which is closely linked to the Rsv3 gene, was reported (Jeong et al., 2002) to contain an LRR consensus sequence highly similar to that of the extracellular LRR domain of resistance genes, Cf-9 and Xa21 (Jones et al., 1994; Song et al., 1995).
In 2008, preliminary soybean whole genome shotgun sequence assembly was released (version “Glyma0”) by the USDOE-Joint Genome Institute Community Sequencing Program (www.phytozome.net/soybean.php [verified 7 Jan. 2011]) and then an improved version Glyma1.0 was released and reported in 2010 (Schmutz et al., 2010). Integration of the soybean sequence and physical maps with the dense genetic marker map would allow the association of mapped phenotypic effectors with the causal DNA sequence (Jackson et al., 2006). In this study, using a sequence-based marker development strategy in three populations, we determined that the Rsv3 gene cosegregates with a cluster of the coiled-coil nucleotide-binding leucine-rich repeat (CC-NB-LRR) resistance genes, which is located in the middle of a heterogeneous cluster containing multiple CC-NB-LRR and leucine-rich repeat receptor-like kinase (LRR-RLK) genes.
Materials and Methods
Plant Genetic Materials and Disease Reactions
A BC3F2 population of 188 individuals from a cross between L29 (Rsv3) and ‘Sowon’ (rsv3) (hereafter referred to as the LS population) was used to investigate the genetic linkage relationship between Rsv3 and molecular markers. This LS population was previously used to develop an Rsv3-linked sequence-based marker, and disease reactions of its F2 individuals were determined by inoculation with the SMV strain G6 (Yu et al., 2005).
To substantiate the genetic relationship between Rsv3 and the microsatellite markers, two additional populations were used: an F2 population of 183 individuals from a cross between L29 (Rsv3) and ‘Lee68’ (rsv3) (hereafter referred to as the LL population) and an F2 population of 61 individuals from a cross between Tousan 140 (Rsv3) and Lee68 (rsv3) (hereafter referred to as the TL population). These LL and TL populations were previously used to locate the Rsv3 gene in the context of soybean molecular linkage groups (Jeong et al., 2002). The LL and TL populations were previously screened for resistance to SMV strain G7, as described by Jeong et al. (2002) and Gunduz et al. (2002). Briefly, SMV G7 cultures were maintained on the soybean cultivar York, which is resistant to SMV G1 but susceptible to SMV G7, to ensure a uniform source of inoculum. Twenty plants for each F2:3 line were inoculated at the unifoliate stage using a carborundum rub method and scored at 14 and 28 d post inoculation. Plants were designated resistant if no symptoms were present and susceptible if mosaic symptoms appeared.
Alignment of Sequences of Rsv3-Linked Markers against the Soybean Whole Genome Sequence
To locate the Rsv3 region in the soybean whole genome sequence, BLASTN searches of sequences of molecular markers that had been mapped near the Rsv3 locus were performed initially against the whole genome shotgun sequence release Glyma0 (http://www.phytozome.net). Sequence analysis was subsequently repeated against Glyma1.0. The markers used for BLASTN searches included A519F/R, M3Satt, M1a, Satt063, Satt560, and Gm-r-Z20a (Jeong et al., 2002; Yu et al., 2005). The predicted gene models from the region delimited by these markers were retrieved from the soybean gene annotation database (accessible at Phytozome v5.0, http://www.phytozome.net, accessed April 2010) for further analysis. Open-reading frame and conserved protein domains were obtained from the Glyma1.0 annotations with the help of the “GBrowse” function of Phytozome (Stein et al., 2002).
To confirm the genetic and physical concordance of the region of the soybean genome sequence corresponding to the predicted Rsv3-residing region, several novel markers were generated from a single NB-LRR gene (Glyma14 g38500.1) and microsatellite repeat sites in the sequence region delimited by A519F/R and M3Satt and from a single LRR-RLK gene (Glyma14 g38650.1) and microsatellite repeat sites flanking the region. For the NB-LRR gene and LRR-RLK gene sequences, primer sets were designed to amplify genomic DNA of the parental soybean lines Sowon and L29 to detect sequence polymorphisms between them (Supplemental Table S1). Polymerase chain reaction (PCR) products were prepared for sequencing by excising a band of expected size from an agarose gel followed by purification by an Accuprep Gel Purification Kit (Bioneer, Daejeon, Korea). When necessary, a given PCR product was subcloned into a plasmid and multiple clones were sequenced. Primers were designed using the Primer3 (http://frodo.wi.mit.edu/primer3/input.htm [verified 7 Jan. 2011]) program. Amplification by means of microsatellite primer sets (Supplemental Table S2) was performed as described by Jeong and Saghai Maroof (2004). Polymerase chain reaction products were resolved by using 3% agarose or 6.5% polyacrylamide gel electrophoresis. Sequence analysis was performed using the BioEdit program (Hall, 1999). Linkage analysis of the markers was performed using MapMaker 3.0b (Lander et al., 1987).
The 1600 NB sequences used in phylogenetic analyses to subdivide NB-LRR proteins from the diverse plant texa into the functionally distinct TIR-domain-containing and CC-domain-containing subfamilies (McHale et al., 2006) were downloaded to serve as a local protein database. BLASTP searches of NB-LRR proteins located at the Rsv3-residing chromosomal region were performed against the local protein database using the “Local Blast” option implemented in BioEdit (Hall, 1999). The NB sequence of NB-LRR proteins near or cosegregating with the Rsv3 locus, the hit sequences, and their close relatives were used to construct a gene tree.
Full-length amino acid sequences of 194 LRR-RLKs and their phylogeny, representing most of the LRR-RLK genes in the Arabidopsis thaliana (L.) Heynh. genome (Gou et al., 2010), were downloaded to serve as a local protein database. TBLASTN searches of LRR-RLK proteins located near the Rsv3 locus were performed as above against this database (Hall, 1999).
For multiple sequence alignment and phylogenetic analysis, protein sequences were analyzed by using ClustalW and the Neighbor-joining and bootstrap methods implemented in MEGA 4 (Kumar et al., 2008). The weighing matrix used for ClustalW alignment was BLOSUM with the penalty of gap opening 10 and gap extension 0.2. The bootstrap consensus trees were inferred from 1000 replicates.
Phytozome Annotation Map
Sequences of molecular markers that have been mapped in three Rsv3-segregating populations (Jeong et al., 2002; Yu et al., 2005) were positioned on soybean chromosome 14 sequence (pseudomolecule) using BLASTN searches against the soybean genome sequence database. The sequential order of the markers determined by genetic maps was concordant to the physical positions of the markers on the soybean chromosome 14 sequence (Fig. 1). Inspection of the soybean gene annotation database revealed that the Rsv3 chromosomal region contains multiple members of two gene families: NB-LRR and LRR-RLK (Fig. 2; see Table 1 for the gene and marker annotation). Therefore, the sequence region (154 kbp) between A519F/R and M3Satt that brackets the Rsv3 locus on the soybean MLG B2 (chromosome 14) (Jeong et al., 2002; Yu et al., 2005) and the surrounding regions were further analyzed by sequence comparison and sequence-based marker development.
|Gene name||Marker name||Position||Gene annotation|
|Glyma14g36630.1||45988517..45993144||Leucine-rich repeat receptor-like kinase|
|Glyma14g36660.1||46001094..46004267||Leucine-rich repeat receptor-like kinase|
|Glyma14g36810.1||46091207..46093272||Leucine-rich repeat receptor-like kinase|
|Glyma14g37590.1||46863817..46868269||Leucine-rich repeat receptor-like kinase|
|Glyma14g37630.1||46908176..46914671||Leucine-rich repeat receptor-like kinase|
|Glyma14g37860.1||47141015..47143938||Coiled-coil nucleotide-binding leucine-rich repeat protein|
|Glyma14g38390.1||47508501..47512747||Leucine-rich repeat receptor-like kinase|
|Glyma14g38430.1||47535945..47538954||Expansin 45, endoglucanase-like|
|Glyma14g38460.1||47566439..47569354||Transcription factor RF2b, putative|
|Glyma14g38490.1||47615588..47619943||Transcriptional factor B3, putative|
|Glyma14g38500.1||47630519..47633501||Coiled-coil nucleotide-binding leucine-rich repeat protein|
|Glyma14g38510.1||47648153..47651727||Coiled-coil nucleotide-binding leucine-rich repeat protein|
|Glyma14g38540.1||47670152..47672833||Coiled-coil nucleotide-binding leucine-rich repeat protein|
|Glyma14g38560.1||47691826..47695095||Coiled-coil nucleotide-binding leucine-rich repeat protein|
|Glyma14g38570.1||47701161..47706197||DNA double-strand break repair RAD50 ATPase|
|Glyma14g38580.1||47721318..47725350||Cinnamate 4-hydroxylase, putative|
|Glyma14g38590.1||47728270..47730998||Coiled-coil nucleotide-binding leucine-rich repeat protein (partial pseudogene)|
|Glyma14g38600.1||47741179..47745790||Translation initiation factor IF5|
|Glyma14g38610.1||47758657..47760317||AP2 domain transcription factor|
|Glyma14g38620.1||47764961..47770195||Ubiquitin-conjugating enzyme E2, putative|
|Glyma14g38630.1||47775271..47778947||Leucine-rich repeat receptor-like kinase|
|Glyma14g38640.1||47800987..47804143||Root phototropism protein, putative|
|Glyma14g38650.1||47816591..47827883||Leucine-rich repeat receptor-like kinase|
|Glyma14g38670.1||47836300..47846561||Leucine-rich repeat receptor-like kinase|
|Glyma14g38700.1||47872961..47876633||Coiled-coil nucleotide-binding leucine-rich repeat protein|
|Glyma14g38740.1||47897628..47899939||Coiled-coil nucleotide-binding leucine-rich repeat protein|
|Glyma14g39180.1||48293479..48297011||Leucine-rich repeat receptor-like kinase|
|Glyma14g39290.1||48411264..48415224||Leucine-rich repeat receptor-like kinase|
|Glyma14g39550.1||48641632..48644930||Leucine-rich repeat receptor-like kinase|
|Glyma14g39690.1||48721778..48724358||Leucine-rich repeat receptor-like kinase|
Sequence Evaluation of the Chromosomal Region between A519F/R and M3Satt through New Marker Development
To further substantiate that the soybean genome sequence region delimited by A519F/R and M3Satt is correctly assembled and corresponds to the predicted Rsv3-residing region, six new markers were developed based on the regional sequence information. Four markers, S156a, S156b, S156c, and S156e, were microsatellite based and two (NB500pro1 and NB500pro2) were designed from the gene model Glyma14 g38500.1 (see below). The locations of the markers were confirmed by mapping in three populations: LS, LL, and TL. The microsatellite markers S156a and S156b cosegregated with Rsv3 in the LS population (Fig. 1). The markers S156a, S156b, and S156c mapped 0.3 cM away from Rsv3 and S156e cosegregated with Rsv3 in the LL population (Fig. 1). In the TL population, markers S156b and S156c cosegregated with Rsv3 (Fig. 1). Marker BARCSOYSSR_14_1417 (hereafter referred to as BS1417), which was in silico identified from the soybean genome sequence (Song et al., 2010), was mapped 0.3 cM away from Rsv3 in the LS population and cosegregated with Rsv3 in the LL population (Fig. 1).
Of the four NB-LRR genes located between A519F/R and M3Satt, Glyma14 g38500.1 was genetically mapped by using two markers generated from its promoter region. A set of primers was designed to PCR amplify the promoter and 5′-end-coding region (Supplemental Table S1). The PCR products amplified from the soybean parental lines using this primer set gave an expected size of 2.3 kbp and their end sequences were aligned, as expected, to the Glyma14 g38500.1 sequence region with greater than 99% similarity. Then, sequences of promoter parts of the PCR products from the parental lines L29 and Sowon were determined and then aligned. The comparison showed several single nucleotide polymorphic sites. Two of the polymorphic sites were used to generate the sequence-based markers NB500pro1 and NB500pro2 (Supplemental Table S2). The markers cosegregated with Rsv3 in the LS population (Fig. 1).
Sequence Evaluation of the Chromosomal Regions Flanking the A519F/R to M3Satt Region through New Marker Development
Alignment of the soybean genome sequence north of the Rsv3-containing chromosomal region delimited by A519F/R and M3Satt was examined through mapping microsatellite markers generated from its corresponding sequence region. Three microsatellite markers, S156g, S156i, and S156l, were generated from the A519F/R-flanking sequence within 100 kbp (Supplemental Table S2). S156g and S156i cosegregated in the LS population with A519F/R, and S156l mapped 0.3 cM away from A519F/R (Fig. 1). S156g was 0.2 cM away from A519F/R in the LL population. Markers S156g and S156i cosegregated with A519 in the TL population. Additional microsatellite markers located over 100 kbp away from A519F/R were developed: GMC, S156t, GME, S156v, and S156w (Supplemental Table S2). All of the additional markers mapped in the three populations at the genetic locations predicted by the soybean genome sequence.
Alignment of the soybean genome sequence south of the Rsv3-containing chromosomal region delimited by A519F/R and M3Satt was examined by mapping microsatellite markers generated from its corresponding sequence region. One microsatellite marker, S156h, generated from the M3Satt-flanking sequence was within 100 kbp. Marker S156h mapped 0.5 cM south of M3Satt in the LS population, cosegregated with Rsv3 in the LL population, and cosegregated with M3Satt in the TL population (Fig. 1). One of the LRR-RLK genes, Glyma14 g38650.1, which is located near the M1a-hit Glyma14 g38630.1 gene, was genetically mapped using a marker generated from one of its intron–exon junction regions. A set of primers was designed to PCR amplify the intron–exon region (Supplemental Table S1). Sequences of the PCR products from L29 and Sowon were aligned, as expected, to the Glyma14 g38650.1 sequence region with greater than 99% similarity. One sequence-based marker RLK650.p1, which was generated from a T/C single nucleotide polymorphism site between L29 and Sowon, cosegregated with S156h and Satt560 in the LS population (Fig. 1).
Construction of a Phylogenetic Tree of NB-LRR Genes
Examination of the list of the gene models for the sequence region between A519F/R and M3Satt on Gm14 indicated that the region contains four full-length NB-LRR genes and one NB-LRR pseudogene, which are members of the disease resistance gene superfamily (Table 1). An additional three NB-LRR genes were observed outside of the A519F/R to M3Satt sequence region (Fig. 2). The NB sequences of the NB-LRR proteins near or cosegregating with the Rsv3 locus, the sequences BLAST-hit against the NB sequence database created by McHale et al. (2006), and their close relatives were used to construct a gene tree. The previously described full-length soybean CC-NB-LRR proteins were also included to determine their relationships with the Rsv3-associated CC-NB-LRR proteins: 3gG2, 5gG3, and 6gG9 associated with the Rsv1 locus on chromosome 13 (Hayes et al., 2004), Rpg1-b encoded by Rpg1-b on chromosome 13 (Ashfield et al., 2004), and Rps1-k-1 and Rps1-k-2 associated with the Rps1-k locus on chromosome 3 (Gao and Bhattacharyya, 2008). Overall amino-acid-sequence identity between the Rsv3-associated NB-LRR proteins and the previously described soybean CC-NB-LRR proteins was less than 25% and amino-acid-sequence identity between their N-terminal domains ranged from 20 to 50%. The NB sequences of the five NB-LRR genes (including the one pseudogene) found in the sequence between A519F/R and M3Satt are highly similar to each other and formed a subclade in the Neighbor-joining tree (Fig. 3). The subclade was designated as the Rsv3-associated NB in Fig. 3. The NB sequences of two (Glyma14 g38700.1 and Glyma14 g38740.1) of the three NB-LRR genes outside of the A519F/R to M3Satt sequence region are sisters to the Rsv3-associated NB subclade. Collectively, the seven NB sequences inside and outside of the A519F/R to M3Satt sequence region formed a well-supported monophyletic group with a bootstrap support of 98%. Branch lengths indicated that the Rsv3-associated NBs are probably the consequence of recent duplications. We defined the clade as the GmCC-NB I (the red box in Fig. 3). The gene model Glyma14 g37860.1 is an outlier out of the three NB-LRR genes outside of the A519F/R to M3Satt sequence region (below the red box in Fig. 3). In BLAST searches, N-terminal and C-terminal parts of the NB domain in Glyma14 g37860.1 best hit, respectively, different groups of NB sequences that belong to two distantly related nonlegume clades, in the phylogenetic tree of McHale et al. (2006). The best-hit sequences (GenBank gene identification [GI] number 16933577 and GenBank GI number 53680944) for each of the two groups were included in the present phylogenetic tree. The six previously described full-length-cloned soybean CC-NB-LRR proteins formed a monophyletic clade, which was defined as the GmCC-NB II (the blue box in Fig. 3). Thus, the results from phylogenetic analysis of NB-LRR genes indicate that the Rsv3-associated CC-NB-LRR genes appear to be members of a novel CC-NB-LRR class that has not been functionally characterized in soybean.
The Cluster of NB-LRR Genes Cosegregating with Rsv3 is Located in the Middle of an LRR-RLK Gene Cluster
The restriction fragment length polymorphism marker M1a is tightly linked to Rsv3 (TL in Fig. 1), and the end sequence of M1a contains the extracellular LRR domain (Jeong et al., 2002). Because the consensus sequence of the M1a LRR is identical to that of the extracellular LRR resistance genes Cf-9 and Xa21 (Jones et al., 1994; Song et al., 1995) and because the clustering of disease resistance genes has been reported in many plants (Michelmore and Meyers, 1998), it was hypothesized that the Rsv3 region might contain a member of the extracellular LRR class of disease resistance genes (Jeong et al., 2002). The BLASTN search of the M1a end sequence (GenBank accession no. AF348333) against the soybean genome sequence identified significantly several chromosomal regions, including Gm11 (32.6 Mbp position), Gm18 (4.4 Mbp), Gm02 (45.6 Mbp), Gm14 (47.8 Mbp), and Gm20 (23.1 Mbp), with an E value smaller than 1.3e−113. The M1a sequence showed the highest similarity (99.9% identity) to a part of Gm11 and 80.6% similarity to a part of Gm14. The M1a-hit Gm14 sequence was located near the M3Satt locus and outside of the sequence region delimited by M3Satt and A519F/R (Fig. 2A and 2B). The M1a-containing full-length gene (Glyma14 g38630.1; Glyma1.0 release at http://www.phytozome.net [verified 13 Jan. 2011]) is a member of the LRR-RLK gene family. Interestingly, members of this gene family are repeated 13 times in the vicinity of the Rsv3 locus (Fig. 2; Table 1), thereby supporting the previous thought that extracellular LRR domain-containing sequences might be clustered in this region of the chromosome (Jeong et al., 2002). However, none of these genes appears to be located between M3Satt and A519F/R. The predicted full-length amino acid sequences of the 13 LRR-RLK genes were compared using TBLASTN searches against a local protein database, which contains 194 A. thaliana LRR-RLK sequences reported by Gou et al. (2010). All the sequences hit the A. thaliana LRR-RLK sequences with an E value smaller than 3e–20. However, the G. max sequences dispersed into eight subfamilies of the phylogenetic tree constructed using the 194 A. thaliana LRR-RLK sequences (Table 2).
|Query sequence||Subject sequence||E value||Subfamily|
More than 40 R genes have been functionally characterized over the past two decades, the majority of which belong to the NB-LRR family (Lukasik and Takken, 2009). The NB-LRR genes tend to cluster in many plant genomes (Michelmore and Meyers, 1998). In this study, we showed that a cluster of the four NB-LRR genes is cosegregating with the Rsv3 locus in our three mapping populations segregating for Rsv3. Despite the lack of physical mapping, parallel alignment between the genetic maps (constructed using public and novel markers) and the genome sequence map (constructed by placing the marker sequences on the soybean genome sequence) are strong evidence that the NB-LRR genes or their variants (as Williams 82 is SMV susceptible) are candidate(s) for Rsv3. Furthermore, none of the other types of genes in the sequence region delimited by A519F/R and M3Satt have been reported to be involved in classical disease resistance response mechanisms (Table 1). Although some members of the LRR-RLK gene family have been reported to be disease resistance genes (Parniske and Jones, 1999; Song et al., 1997), our genetic mapping results indicated that these sequences are outside of the sequence region delimited by A519F/R and M3Satt and are unlikely candidate genes for Rsv3.
The Rsv3-residing chromosomal region is of great interest with respect to the evolution of multigene clusters because members of the CC-NB-LRR and LRR-RLK multigene families constitute a heterogeneous cluster. Nucleotide-binding leucine-rich repeat (NB-LRR) or LRR-RLK genes often occur in clusters that consist of several copies of homologous gene sequences arising from a single gene subfamily (simple clusters) or colocalized gene sequences derived from two or more unrelated subfamilies (complex clusters) and may also contain unrelated single genes interspersed between the homologs (Shiu and Bleecker, 2001; Friedman and Baker, 2007). It has been suggested that intergenic unequal crossover and intragenic mispairing contribute to altered gene copy number within the cluster (e.g., Parniske and Jones, 1999; Kuang et al., 2004, 2005; for a review, see Friedman and Baker, 2007). Although it has been reported that a single Prf gene, a member of the NB-LRR superfamily, is embedded within a cluster of five Pto kinase homologs (Salmeron et al., 1996), colocalization of multiple members of both of the two disease resistance gene superfamilies has not been reported to the best of our knowledge. Surprisingly, the members of the two gene superfamilies appear to be interspersed with each other in the Rsv3-residing chromosomal region on Gm14. Our results suggest that the NB-LRR genes in the Rsv3-containing chromosomal region likely arose from a single gene subfamily (Fig. 3) and that the LRR-RLK genes likely arose from eight unrelated subfamilies reported by Gou et al. (2010) (Table 2). Feature(s) of this chromosomal region that resulted in the coevolution of the two gene clusters at the same chromosomal region or interactions between the two gene clusters during the evolution remain unclear. Interestingly, our examination of the duplication blocks at the SoyBase Browser (http://soybase.org [verified 13 Jan. 2011]) indicated that only LRR-RLK genes cluster on the region on chromosome 2 that is homeologous to the Rsv3 region (data not shown). Although our genetic analysis clearly excluded an Rsv3 candidacy of LRR-RLK genes, the possibility of interaction between NB-LRR and LRR-RLK proteins in conferring resistance to SMV at the Rsv3 locus cannot be dismissed in light of the case of the Prf and Pto interaction (Salmeron et al., 1996). Further analyses including crosses and sequence comparison between soybean cultivars and heterologous expression will help resolve these issues.
The LRR domain structure in the NB-LRR genes would lead to the expectation that, in determining recognition specificity, either Avr proteins or recognition cofactors bind to this domain. At the same time, an accumulating body of evidence suggests that the N-terminal domains of NB-LRR proteins also play a role in Avr recognition (reviewed by Collier and Moffett, 2009). Rsv1, one of the three known SMV resistance genes, is associated with the CC-NB-LRR gene cluster on the soybean chromosome 13 (MLG F) (Jeong et al., 2001; Hayes et al., 2004), the NB domains of which belong to the CC-NB-LRR superfamily (Ashfield et al., 2004). Because the findings of this study suggest that Rsv3 may also encode a member of the CC-NB-LRR gene family, which is distantly related to the Rsv1 locus-associated CC-NB-LRR genes, it is hypothesized that the different disease responses of Rsv1 and Rsv3 against a spectrum of SMV strains may be due to the different structures of the CC-NB-LRR genes. It will be interesting to further elucidate what structural difference between the Rsv1-associated and Rsv3-associated CC-NB-LRR genes make the Rsv1 and Rsv3 genes confer different disease responses to the same SMV strain.
Recently the elicitors or pathogenic determinants governing the disease reactions of Rsv1- or Rsv3-genotype soybeans to different SMV strains have been identified. Chimeric clones, constructed by exchanging genomic sequences from virulent and avirulent SMV strains, were used to map the region within the SMV genome that induces a defense response in Rsv1- and Rsv3-containing cultivars. It was shown that the helper component-protease (HC-Pro) and P3 proteins are independently recognized by Rsv1-genotype soybean and elicit the extreme resistance phenotype (Hajimorad et al., 2006; Eggenberger et al., 2008). Similarly, the cytoplasmic inclusion (CI) protein was shown to be the elicitor or pathogenic determinant recognized by Rsv3-genotype soybean (Seo et al., 2009; Zhang et al., 2009) and that a single amino acid substitution was responsible for strains that can avoid Rsv3 recognition to become avirulent (Seo et al., 2009). Thus, two lines of evidence, phylogenetic difference between Rsv1 and Rsv3 shown in this work and distinct viral genomic regions as elicitors and as determinants of pathogenicity, demonstrate the complexity of the interactions in the soybean–SMV pathosystem that are yet to be elucidated.
Alleles of R genes confer different resistance reactions to pathogens (e.g., Jones and Dangl, 2006). For example, when inoculated with SMV strain G1, Rsv1-n, an allele of Rsv1, confers a severe or lethal necrotic reaction that is a typical resistance reaction but is not desirable in a commercial cultivar (Tucker et al., 2009). To remove or to replace these genes in a soybean cultivar, markers located in the middle of the resistance gene clusters, which span over several centiMorgans in many cases, would be essential for marker-assisted selection of desirable line(s) from a breeding population. Among the three known R genes conferring resistance to SMV, several studies have developed molecular markers tightly linked to Rsv1 (e.g., Gore et al., 2002; Shi et al., 2010) and Rsv4 (Hwang et al., 2006; Saghai Maroof et al., 2010). In particular, Saghai Maroof et al. (2010) used the soybean genome sequence to develop numerous Rsv4-linked molecular markers as well as to show that the Rsv4 gene likely belongs to a new class of resistance genes. The molecular markers developed in this study, and prediction of the molecular nature of the Rsv3 gene, should provide additional tools for pyramiding SMV-resistance genes to obtain durable SMV-resistant soybean cultivars and for elucidating the structure and function of the Rsv3 gene.