Two genes for resistance to SMV, Rsv1 and Rsv3, are strain-specific and have been associated with a typical hypersensitive response (HR), resulting in local or systemic necrosis to certain strains of the seven U.S. strain groups (Saghai Maroof et al., 2008). Nine Rsv1 resistance alleles have been defined in Ogden, York, Marshall, Kwanggyo, Raiden, Suweon97, PI 486355, PI 507389, and FT-10, respectively (Moon et al., 2009; Tucker et al., 2009). These alleles confer resistance to lower numbered SMV strain groups (SMV G1-G3), but susceptibility is often characterized by a mosaic or necrotic reaction to higher numbered strains (SMV G5-G7) (Chen et al., 1991). In contrast, lines containing Rsv3 are susceptible to lower numbered strain groups but resistant to the higher numbered ones (Gunduz et al., 2002). Rsv1 and Rsv3 have been mapped to molecular linkage groups (MLGs) -F (chromosome 13) and -B2 (chromosome 14), respectively (Gore et al., 2002; Jeong et al., 2002).
The Rsv4 gene is of great interest because it was initially reported to confer resistance to all SMV strains tested in the U.S., namely SMV-G1 through -G7 (Chen et al., 1993; Ma et al., 1995). This gene was first recognized in PI 486355 in association with an Rsv1 allele and was genetically isolated in a line from an F3:4 plant, derived from crosses with ‘Essex’ (rsv) (Ma et al., 1995). This line, initially designated ‘D26’ and then ‘LR2’, was advanced and later released as V94-5152 (Buss et al., 1997). The same resistance to SMV-G1 through -G7 was also observed in ‘Columbia’ (Ma et al., 2002) and PI 88788 (Gunduz et al., 2004). Columbia was found to also carry two genes, Rsv3 and a gene later designated Rsv4. In populations consisting of Columbia (Rsv3Rsv4) by ‘Lee68’ (rsv), Ma et al. (2002) observed a response to SMV-G1 designated as a late susceptible (LS) phenotype, because symptoms were not observed until approximately three weeks after initial inoculation. However, plants remained resistant to SMV-G7. Gunduz et al. (2004) showed that PI 88788 has the same single, dominant gene, in crosses with LR2 and V94-5152. The mechanism of resistance was suggested to be reduced infection and local cell-to-cell movement along with reduced movement into and from the vascular system (Gunduz et al., 2004). Recently, the LS reaction was also documented in pyramided Essex isolines developed through marker-assisted selection (MAS) using flanking molecular markers and disease reaction assays (Saghai Maroof et al., 2008). The two gene pyramid of Rsv3Rsv4 displayed a LS reaction upon inoculation with strains SMV-G1 and -G2. In none of the above studies has necrosis been observed, which is the typical hypersensitive resistance gene response seen with Rsv1 and Rsv3 (Ma et al., 2002; Gunduz et al., 2004; Saghai Maroof et al., 2008), suggesting that Rsv4 may have unique properties compared to other R genes.
Hayes et al. (2000) mapped Rsv4 to MLG-D1b (chromosome 2) between SSR markers Satt558 and Satt542 at distances of 7.8 and 4.7 cM, respectively, in a population of D26 by Lee68. Further fine mapping by Hwang et al. (2006) employed a comparative mapping strategy for targeting Rsv4 using genomic sequences from Lotus japonicus. Through this comparative genomics approach, two soybean expressed sequence tags (ESTs) were identified that flanked Rsv4, each at a distance of approximately 2.5 cM. However, additional fine mapping of Rsv4 followed by identification of candidate genes is necessary to understand the underlying genetic mechanism of this unique form of resistance.
The Williams82 whole-genome shotgun sequence (WGS) (Joint Genome Institute, 2008) is a new resource available for the soybean community for genomics-related studies. Three studies have taken advantage of this resource to design primers for fine mapping and sequencing of putative genes related to seed composition traits of low stachyose and raffinose (Dierking and Bilyeu, 2008; Skoneczka et al., 2009) and low phytate (Saghai Maroof et al., 2009) in soybean. Utilizing the WGS, Saghai Maroof et al. (2009) identified genes encoding multi-drug resistance associated proteins (MRPs) which mapped near previously identified quantitative trait loci controlling seed phytate content in the soybean mutant line, CX1834 (Walker et al., 2006; Gao et al., 2008). Segments of gene sequences from one MRP gene on MLG-N (chromosome 3) were compared between the low-phyate mutant line, CX1834, and lines with normal phytate levels. A single mutation from an A to T was identified in CX1834 resulting in a stop codon. This mutation is thought to contribute to the low-phytate phenotype expressed in this mutant line. To date, these methods have not been used to further fine map or identify candidate genes contributing to SMV resistance in soybean. Therefore, in this study, the WGS was used to conduct fine mapping of Rsv4 in two separate mapping populations. Also, sequence information from parental lines was analyzed for design of additional, closely linked markers for Rsv4 and for identification of possible gene candidates associated with resistance to SMV.
MATERIALS AND METHODS
Two separate populations were used for fine mapping of Rsv4. The first was an F2:3 D26 (Rsv4) by Lee68 (rsv) population consisting of 254 individuals created by Hayes et al. (2000) for the initial mapping of Rsv4 (Table 1). The second population consisted of 212 and 56 F5 and F7 recombinant inbred lines (RILs), respectively, of V94-5152 (Rsv4) by Lee68 (rsv) (Table 1). V94-5152 is a F6:7 selection from a segregating Essex (rsv) by D26 (Rsv4) population (Buss et al., 1997).
|Population||Generation||Gene||No. of individuals or RILs||Disease screening results|
|D26 x Lee68||F2:3||Rsv4||254||56||135||63|
|V94-5152 x Lee68||F5/F7||Rsv4||268||123||29||116|
Virus Cultures and Inoculation
The G1 strain of SMV was used in this study as it was the same strain used for mapping Rsv4 to MLG-D1B by Hayes et al. (2000). In addition, SMV-G1 produces a distinct, easily scored mosaic phenotype on the susceptible Lee68 (rsv) parental line and has been used in previous studies (Saghai Maroof et al., 2008; Tucker et al., 2009). Cultures were maintained in the greenhouse by periodic transfer to Essex (rsv). Inoculations were performed as previously described (Saghai Maroof et al., 2008) by grinding infected trifoliolate leaflets in 0.01 mol L−1 neutral sodium phosphate buffer (1:10; w/v) in a mortar and pestle, and using the pestle to rub unifoliolate leaves (V-1 stage) previously dusted with 600 mesh carborundum (Buehler, Lake Bluff, IL). Plants were rinsed with tap water immediately thereafter. The pathotype of the SMV-G1 strain was confirmed prior to and at the completion of this study by inoculation to a standard set of differential cultivars (‘York’, ‘Marshall’, ‘Ogden’ and ‘Kwanggyo’ and PI96983) (Chen et al., 1991).
Two parental lines (V94-5152 and Lee68), Essex, and each RIL were tested for reaction to SMV-G1. For each RIL, four to five plastic pots (7.6-cm diameter, 7.6-cm depth) each containing five seeds (for a total of 15–20 seeds per RIL) were tested. The RIL were divided into four sets and the experiment was performed over a 3-mo period screening approximately 63 RILs in each set. Pots were filled with MetroMix360 (Scotts-Sierra Horticultural Products Co., Marysville, OH), seeded, and randomly distributed by RIL on greenhouse benches. Plants in each of the four sets were inoculated as described above in one day. Osmocote 18–6–12 NPK (Scotts-Sierra Horticultural Products Co., Marysville, OH) was used as a supplemental fertilizer at the rate of 16 gm/pot every two weeks. A non-inoculated control pot of each RIL and parental lines were included. However, if a specific RIL had poor germination the control pots were inoculated to ensure that at least 15 plants per RIL were inoculated.
After virus inoculation, each set of RILs was observed for 5 wk, with detailed scoring at 7, 10, 14, 21, 28 and 35 d post inoculation (dpi). Digital images of symptoms were taken over time for selected RILs displaying varying reactions upon inoculation. Throughout these observation periods, two methods were used to test for presence or absence of virus in order to confirm disease ratings for each RIL. Whole leaf immunoprints were taken of a subset of RILs according to Gunduz et al. (2004) to examine virus distribution throughout the leaf. Tissue immunoblot assays (TBIA) were also conducted as previously described by Ma et al. (2002) on selected RILs with a minor modification. In order to prevent the destruction of an entire leaf, a small syringe was used to punch a section of leaf tissue approximately 3 mm in diameter onto the surface of the membrane for testing purposes. This allowed the remaining portion of the leaf to be immunoprinted as needed to examine distribution of viral antigen over a larger area.
Soybean leaf disc tissue was bulk sampled from approximately 20 seedlings for each RIL. Leaf discs (20, one from each seedling) approximately 3 mm in diameter were taken one day prior to inoculation using a single, hand-held paper punch. Tissue was sampled at the edges of the first true leaf to prevent major damage prior to inoculation. Leaf discs were collected into 2 ml eppendorf tubes, ground in liquid nitrogen, and extracted using the CTAB technique following methods similar to Edwards et al. (1991). Samples with poor quality or low quantity of DNA were either re-grown from seed in the greenhouse or collected from field plots using a similar bulk method. DNA for mapping of additional markers in the F2:3 D26 by Lee68 population was the same as reported by Hayes et al. (2000).
Molecular Marker Design and Analysis
Molecular markers used in linkage analysis include those previously reported by Cregan et al. (1999), as well as new markers developed in this study based on the microsatellite motifs identified from the WGS (Joint Genome Institute, 2008). At the beginning of this study, the current WGS release was Glyma0. Microsatellite markers Satt634 and Satt542 were previously found to flank the Rsv4 gene (Hayes et al., 2000). Therefore, blast searches using microsatellite markers Satt634 and Satt542 as queries determined the approximate location of Rsv4 within the Glyma0 release (Table 2). Additional gene predictions using various gene prediction programs were used such as Genmark, Genezilla, FgeneSH, and Blastx, apart from mapping EST-derived unigenes into the Satt634 to Satt542 location (Supplemental Table 1). False positives carrying peptide length less than 30 aa and coding potential (Staden and McLachlan, 1982) less than 0 were discarded. Predicted miRNAs in the region using pssRNAMiner were as described by Dai and Zhao (2008). Microsatellite primers (Table 3) were designed from sequences of this region as described by Skoneczka et al. (2009) and Saghai Maroof et al. (2009). SSR marker assay and polyacrylamide gel electrophoresis were performed according to Saghai Maroof et al. (1994). Primers that yielded polymorphic products among the parents were mapped in the original F2:3 population of D26 by Lee68 (Hayes et al., 2000) and the newly developed RILs of V94-5152 by Lee68. JoinMap version 3.0 (Van Ooijen and Voorrips, 2001) was used in linkage group construction at a logarithm of odds ratio (LOD) of 3.0. Recombination values were converted to genetic distances using Kosambi's mapping function in JoinMap 3.0. Currently, in the Glyma1 release (8X 1.01 Pseudomolecule assembly), the chromosome corresponding to MLG-D1B, which contains Rsv4 (Hayes et al., 2000), is designated as chromosome 2 (Joint Genome Institute, 2008).
|Marker ID||Chromosome||F2:3 D26xLee68||Song et al. (2004)||SoyBase (cM)||Chromosome range (bp)|
|Primer name||Marker type||Forward sequence (5′-3′)||Reverse sequence (5′-3′)|
|STF5 (Exon 1)||SNP||TCCTTATCTGTCAATTTCCTTGT||GCGGATAATGTTCCATAGTAAAA|
|STF5 (Exon 2)||SNP||TAGTCTTTGGCTCAGAATACACC||AAAGCTTTCAAGACTTTGGTTAC|
|STF5 (Exon 3)||SNP||TACTAATGGAACTCTTGTTGCAG||CACGCATTCAATATTAGTTACCTC|
|STF4 (Exon 1)||SNP||TACATGTATATGCCTGAATTTGG||AACAAACCCTGATATTTTTAATTG|
|SUNk1 (Exon 1)||SNP||AGCATCAAAGAAGAAGAAGTCAA||AGGATCAGTTGAAATCATGAAAC|
|STF3 (Exon 1)||SNP||GGAAAGAGATATAAAGCAAGCAA||GATGGGGTTTGTGTGTGTATT|
|STF3 (Exon 2)||SNP||TACAGGATCTTGGTTGTTTTACC||CATATCCTACGTTTTCCAGTTTC|
|STF3 (Exon 3)||SNP||CTAACATATGTCACCCCATGC||GATTTGCTTCAGGGAAGAGTC|
|STF3 (Exon 4)||SNP||AGTTGTGTTCTTAAATACTCATTTTT||GGATAAACTAAAAGAGAGGGGATT|
|STF2 (Exon 1)||SNP||AGCTTCAGAGTTTGGAGCAG||AAGTTAAGCATCGATTACTTTGA|
|STF2 (Exon 2)||SNP||CTTAACAACATGTGTGATACGAAA||TCCAACTCAGAAGAAACAACTTT|
|USP1 (Exon 1)||SNP||TCCAACTACATTGCAAGACAAC||GAGCCACATGAAGTATAAAGAGAA|
|USP1 (Exon 2)||SNP||AGGTTGAAGTTAATTGCATGACT||TTCTGAATGCAGTACTCTACCAC|
Gene-specific primers were designed from the Glyma0.1b gene models (Joint Genome Institute, 2008). PCR products were amplified from V94-5152 (Rsv4) and Lee68 (rsv), sequenced, and aligned as described by Skoneczka et al. (2009) with some annealing temperatures modified to limit non-specific amplification as needed.
Responses to Virus Inoculation and Accumulation in Parental Lines and Recombinant Inbred Lines
Within 6 to 7 dpi, Lee68 (rsv) plants inoculated with SMV-G1 displayed chlorotic spots on inoculated leaves and vein clearing on emerging, non-inoculated trifoliolate leaves (T-1). Lee68 (rsv) displayed a severe leaf curl on emerging trifoliolate leaves 10 to 14 dpi (Fig. 1A). V94-5152 (Rsv4) showed no symptoms after virus inoculation during the course of the 35 day experiment and no virus was detected by TBIA or leaf immunoprints.
Symptoms on susceptible RILs (rsv4rsv4) included chlorotic spots on inoculated leaves within 7 to 10 dpi. However, some RILs displayed veinal and inter-veinal necrosis on the inoculated leaves. On emerging trifoliate leaves (T-1), vein clearing developed within 6 to 7 dpi, followed by slight necrosis in the majority of susceptible RILs. Those RILs that were classified as homozygous susceptible (rsv4rsv4) were discarded after the 14 dpi disease rating because of greenhouse space constraints, while RILs that were homozygous resistant or heterozygous were retained for further observation at 35 dpi.
Heterozygous RILs (Rsv4rsv4), containing both resistant and susceptible individuals in the same pots at the 7 and 10 day screenings, were retained to allow the LS phenotype to develop, as observed in previous studies (Ma et al., 2002; Gunduz et al., 2004; Hwang et al., 2006; Saghai Maroof et al., 2008). The LS phenotype began to be observed between 14 to 21 dpi on previously classified resistant individual plants of segregating heterozygous RILs. For these segregating (Rsv4rsv4) RILs, the proportion of initially resistant to susceptible plants per RIL fits a 3 to 1 genetic ratio at 7 to 10 dpi scoring. Later disease scoring of previously characterized resistant RILs, from a total of 29 that were segregating (Table 1), developed the LS phenotype by 35 dpi.
The LS phenotype first appeared as small chlorotic to bronze lesions on one or more leaflets of T-1 at 14 or more dpi. Leaf immunoprints of these LS phenotypes (Fig. 1B-C) confirmed that the virus was present only in symptomatic areas of these T-3 leaves, demonstrating that virus accumulation following long distance movement was restricted to initially invaded cells to chlorotic spots (Fig. 1C). A small number (less than 10) of the LS phenotype plants of the heterozygous RILs exhibited a darkened pigmentation of the petiole and major leaf veins in T-1 leaflets between 14 and 21 dpi, as well as a faint chlorosis along minor veins (Fig. 1D). Leaf immunoprints of these leaves showed that the virus was mainly confined to minor and major leaf vein areas (Fig. 1E). In all LS phenotypes, later emerging trifoliate leaves (21 to 28 dpi; T-2 to T-3) showed severe leaf distortion and a green island (GI) effect, showing dark green tissue surrounded by a lighter green tissue (Fig. 1F). At the last scoring of 35 dpi, LS plants had bronze lesions on lower trifoliolates (T-1 and T-2) that continued to enlarge (Fig. 1G), and all newly emerged trifoliates (T-3, T-4) displayed severe GI effects.
RILs classified as homozygous resistant (Rsv4Rsv4) typically remained symptomless during this study but were sampled periodically by TBIA or leaf immunoprints if virus symptoms were suspected. Approximately half of the resistant RILs showed a faint chlorotic flecking as early as 10 dpi on one or two plants of 15 that were inoculated. However, no virus was detected in tissues from these plants throughout the course of this study by either TBIA or leaf immunoprint assay. In a small number of RILs (Rsv4Rsv4) tested (one plant from a total of four different RILs), the virus displayed some movement from the original inoculated leaves. In these four plants, a small brownish lesion developed on a first trifoliolate leaflet (T-1) (Fig. 1H). The lesions did not expand on the leaves and virus symptoms did not appear on other portions of the plant unlike a typical heterozygous line.
Molecular marker mapping data near the Rsv4 region enabled us to differentiate between the reactions described above. Lines that remained homozygous resistant (Rsv4Rsv4) had V94-5152 alleles, whereas lines that were homozygous susceptible (rsv4rsv4) had Lee68 alleles for markers mapping near Rsv4. RILs that displayed the LS phenotype had both V94-5152 and Lee68 alleles for Rsv4-linked markers. Individual RILs that had unique recombination events between closely linked markers and the Rsv4 locus were rescreened in additional greenhouse assays to confirm their original scoring.
Fine Mapping of Rsv4
Blast analysis of microsatellite markers Satt634 and Satt542 (6 cM interval in V94-5152 by Lee68 population) identified a 1.5 MB physical distance in the Glyma0 release of the WGS between these two markers (Table 2). New microsatellite markers designed from within this region enabled us to shorten both genetic and physical distances between markers flanking Rsv4 (Fig. 2; Table 2). Traditional protocol for microsatellite marker development involves small insert library construction, screening with oligos, sequencing, and primer design. The availability of the WGS resources provides easily accessible microsatellite sites for efficient primer design (e.g., Skoneczka et al., 2009; Saghai Maroof et al., 2009). A total of 20 markers were designed from the region of interest around Rsv4, six (30%) of which were polymorphic both in the D26 by Lee68 and V94-5152 by Lee68 populations. Six new SSR-based markers designed from the WGS were added to the existing Hayes et al. (2000) map and newly created RIL population in a 5.8-cM region containing Rsv4 (Table 3; Fig. 2). These markers maintained expected mapping order based on their physical location (Glyma1 release) in both mapping populations (Table 3). The closest flanking markers (212MAT16 and 212MTATA11) shortened the genetic interval containing Rsv4 to 0.7 cM and 1.3 cM in the V94-5152 by Lee68 and D26 by Lee68 mapping populations, respectively (Fig. 2). The physical distance between these two markers based on the available Williams82 sequence is approximately 100 kb.
Candidate Gene Discovery
The identified 100-kb region containing Rsv4 was examined for predicted gene models utilizing the Glyma1.0 annotations. Three different gene prediction programs and Blastx were used to search for candidate genes in this area. A total of 122 genes including 8 miRNAs, were identified between SATT634 (Gm02: 11344159) and AW471852 (Gm02: 12164901). Out of 114 genes, 63 are supported by EST and 51 are predicted genes (Supplemental Table 1; Supplemental Fig. 1). Nine gene models were predicted in the 100 kb sequence region. Five of these genes have significant homology to transcription factors (TFs) in Arabidopsis and rice (Oryza sativa), while the other four showed homology with a zinc finger protein, nucleic acid binding protein, and two putative genes with unknown protein function (Table 4; Supplemental Table 1).
|Chromosome location||Glyma1.0 model||Arabidopsis homolog||Description||Primer||Polymorphism|
|Gm02:11651991..11661761||Glyma02g13360||AT4G38440||unknown protein (RNA Polymerase II Associated Protein Domains)||STF5||None|
|Gm02:11671023..11677144||Glyma02g13370 (GmaAffx.88326.1.S1_at)||AT1G20980||SPL14 (SQUAMOSA PROMOTER BINDING PROTEIN-LIKE 14); DNA binding / transcription factor||STF4||None|
|Gm02:11692905..11694242||Glyma02g13380 (GmaAffx.33294.1.S1_at)||AT1G69160||Unknown protein (At1g69160/F4N2_9; Putative myb-like protein)||SUNK1||2 SNPs|
|Gm02: 11729891..11730198||Glyma02g13390 (GmaAffx.70788.1.S1_at)||AT1G01530||AGL28 (AGAMOUS-LIKE 28); DNA binding / transcription factor||STF3||None|
|Gm02:11734620..11736851||Glyma02g13400||AT2G24840||SEP1 (SEPALLATA1); DNA binding / transcription factor||-||-|
|Gm02:11746279..11746718||Glyma02g13410||AAM21342||MADS-box protein 2 transcription factor [Vitis vinifera]||-||-|
|Gm02:11749041..11752996||Glyma02g13420||AT1G01530||AGL28 (AGAMOUS-LIKE 28); DNA binding / transcription factor||STF2||2 SNPs|
|Gm02:11757828..11758642||Glyma02g13430||AT5G38895||Zinc finger (C3HC4-type RING finger) family protein||-||-|
|Gm02:11766761..11768963||Glyma02g13440||AT2G01050||Nucleic acid binding / zinc ion binding||-||-|
|Gm02:11770447..11771944||Glyma02g13450||AT1G69080||Universal stress protein (USP) family protein||USP1||None|
Nucleotide binding site-leucine rich repeat (NBS-LRR) type resistance genes have been implicated for resistance at the Rsv1 locus (Hayes et al., 2004). Therefore, the identified 100-kb interval containing Rsv4 was strictly evaluated for this disease resistance motif. While there were no NBS-LRR sequences found between 212MAT16 and 212MTATA11, approximately 4 cM below, between markers AW307114 and Satt542, a NBS-LRR-Toll/interleukin-1 (TIR) sequence with homology to Peronospora parasitica (RPP1) was located.
To further investigate candidate genes near markers 212MAT16 and 212MTATA11, primers were designed to amplify from V94-5152 and Lee68 predicted gene exons based on Glyma0.1b gene models (Table 4). Although a lack of polymorphism between parents in model (predicted) exons would not exclude a gene for consideration as a candidate gene, observed polymorphisms would provide gene-specific markers for Rsv4.
Thirteen SNP primer pairs (Table 3) were used to amplify coding sequences from six of these putative genes based on the Glyma0.1b gene models (Joint Genome Institute, 2008). Sequencing these fragments from predicted coding regions revealed two modeled genes in the Rsv4 mapping interval with SNPs detected between V94-5152 (Rsv4) and Lee68 (rsv) sequences (Table 4). One is a putative TF with homology to AGL28; AGAMOUS-LIKE 28 and the other is a putative gene with unknown function with a slight similarity to a myb-like protein.
As expected, Rsv4 delayed virus replication and movement in plants (Gunduz et al., 2004; Ma et al., 1995), rather than conferring a hypersensitive response (HR) or an extreme resistance (ER) response as seen with Rsv1 or Rsv3. In ER reactions, no symptoms are induced and there is no detectable virus accumulation (Hajimorad and Hill, 2001; Hajimorad et al., 2006; Zhang et al., 2009). Also no predicted NBS-LRR type resistance gene was observed within the 1.3-cM region in the Williams82 sequence. Similar to reports by Gunduz et al (2004) and Hwang et al. (2006), the LS disease phenotype was expressed only in lines that were heterozygous at the Rsv4 locus. As Rsv4 does not confer a HR response or ER as seen with Rsv1, the gene is predicted to place less selection pressure on pathogen populations possibly creating a durable type of resistance which can be utilized by breeders (Saghai Maroof et al., 2008).
Hayes et al. (2004) identified a strong candidate gene, 3gG2, for Rsv1 from PI 96983 on MLG-F (chromosome 13, Joint Genome Institute, 2008). The 3gG2 ORF sequence encodes a 3,390 bp gene with a deduced protein product similar to previously cloned nonTIR-NBS-LRR disease resistance genes. Additionally, various unique resistant and necrotic reactions were observed coincident with the presence or absence of the other or complement genes contained in this resistance gene cluster on MLG-F. Lines lacking complement genes produced unique reactions to tested SMV strains similar to Ogden and Marshall carrying alleles Rsv1-t and Rsv1-m, respectively. Therefore, multiple and different combinations of these NBS-LRR genes at this locus condition resistance to SMV. In the current study, it is unclear whether a single or multiple genes that are closely linked are contributing to resistance at the Rsv4 locus. Additional studies similar to Hayes et al. (2004) utilizing unique recombinants in a large RIL population may be needed to examine contributions of individual and multiple genes on resistance conferred by the Rsv4 locus. Gene dosage and partial dominance may also be involved in the LS response associated with Rsv4, as suggested by Ma et al. (2002). Still other methods to determine the gene(s) governing resistance to SMV in this region may involve gene expression experiments such as microarray analysis with rsv4 and Rsv4 lines inoculated vs non-inoculated. Previously, Saghai Maroof et al. (2008) developed near isogenic lines of Rsv4 in a susceptible Essex (rsv) background, using marker-assisted selection. These available materials represent excellent starting points for gene expression profiling and high-throughput transcriptome sequencing experiments.
Virus resistance genes that function by restricting virus movement and accumulation are usually recessive in nature (Maule et al., 2007). For many potyviruses, this recessive resistance is under the control of eukaryotic translation initiation factor, eIF4e or its paralogue eI(iso) F4e. The sbm1 gene in pea (Pisum sativum) conferring resistance to Pea seedborne mosaic virus (PSbMV) has been characterized as an eIF4E controlling virus translation, replication, and cell-to-cell trafficking (Gao et al., 2004a, 2004b) in a legume system. This particular locus also confers resistance against Bean yellow mosaic virus (BYMV) and can be overcome by mutations in the VPg protein of the virus (Bruun-Rasmussen et al., 2007). However, although Rsv4 acts to restrict viral movement and accumulation, in addition to exhibiting a LS phenotype, its mode of inheritance is dominant (Gunduz et al., 2004). Further evidence that Rsv4 differs from recessive genes restricting virus movement is that no eIF4E candidate genes were found to be present in the 100 kb region which Rsv4 appears to reside.
Analysis of Williams82 gene model sequences in the 100-kb region believed to be the location of Rsv4 predicted the presence of several TFs, tentatively belonging to MADS or MYB families. Although TFs have been shown to be important in disease resistance in plants, they are members of large families associated with stress defense responses, which were not among those identified near Rsv4 (Lin et al., 2007; Century et al., 2008; Mukhtar et al., 2008). Dai et al. (2008) recently demonstrated the importance of the involvement of two rice TFs in rice tungro virus replication and infection. Over-expression of these two TFs (RF2a and RF2b) led to reduced virus accumulation in tissues and overall reduced symptoms in transgenic rice lines. In our study, leaf immunoprints of SMV-inoculated leaves suggest that replication at the cellular level is not impaired, but initial establishment of an infection site and entry into the vascular system are severely restricted (Gunduz et al., 2004). In this paper, immunoprints of LS leaves confirm that cells in areas of leaflets showing chlorotic spots have a high level of SMV, but further spread within these trifoliolate leaflets is limited. Currently, it remains unknown whether any TF in the identified 100 kb region containing Rsv4 are essential for SMV accumulation and cell-to-cell or long-distance movement.
This study demonstrates that the Williams82 WGS sequence is a powerful tool for targeted generation of markers for use in fine mapping. Each of the 20 markers designed in this study amplified a product of their expected length. This high rate of primer viability facilitated the rapid screening of markers for polymorphism between parents and demonstrates the effectiveness of the WGS sequence as a template. Use of the WGS resources enabled mapping Rsv4 in a physical interval of less than 100 kb, but not to determine which gene or genes are governing or regulating resistance. There were no predicted genes encoding NBS-LRR or eukaryotic initiation factor (eIf4E) motifs in this region. Although this observation contributes to the hypothesis that Rsv4 is a unique and previously uncharacterized type or class of resistance gene, it leaves open the possibility that the genetic entity contributing to resistance in this region may not be observable in the Williams82 sequence, which is derived from a susceptible line. A similar observation was reported by Grant et al. (1995) for the RPM1 resistance gene in Arabidopsis. An annotated sequence of the Rsv4 interval from a resistant parent may reveal unique features that could contribute to resistance. Future studies to characterize this gene, focusing on unique genetic regions in the resistant lines described here could elucidate whether Rsv4 represents, in fact, a novel resistance mechanism.