Introduction

Many recognition-dependent plant disease resistance (R) genes encode proteins comprised an amino-terminal domain, a nucleotide binding site (NBS) domain, and leucine-rich repeats (LRRs) (Bent 1996; Hammond-Kosack and Jones 1997; Hulbert et al. 2001). NBS-LRR encoding genes are often highly duplicated and evolutionarily diverse, with hundreds of family members in plant genomes (Meyers et al. 1999, 2003; Dangl and Jones 2001; Goff et al. 2002; Richly et al. 2002; McHale et al. 2006; Jones and Dangl 2006; Ameline-Torregrosa et al. 2007; Kohler et al. 2008). NBS-LRR resistance proteins directly or indirectly recognize pathogen avirulence factors, which trigger signal transduction cascades leading to rapid defense responses and frequently to hypersensitive reactions and programmed cell death (Dangl and Jones 2001; Bent 1996; Hammond-Kosack and Jones 1997; Hulbert et al. 2001). Of the 40 or so R-genes known to confer resistance to bacterial, fungal, viral, and nematode pathogens in plants, 75% encode NBS-LRR proteins. LRRs are determinants of response specificity and are typically under strong diversifying selection, whereas NBS domains are not (Parniske et al. 1997; McDowell et al. 1998; Meyers et al. 1998a; Sun et al. 2001; Kuang et al. 2004).

Conserved amino acid sequence motifs in the NBS domain have been widely used to isolate and classify NBS-LRR encoding genes (Meyers et al. 1999, 2002, 2003; Pan et al. 2000; Bai et al. 2002; He et al. 2004; Cannon et al. 2002). Two superfamilies have been described in dicots, one with an N-terminal structural domain homologous to the intracellular signaling domains of the Drosophila Toll and mammalian interleukin1 receptor (TIR) and the other lacking the TIR structural domain (non-TIR) (Meyers et al. 1999; Pan et al. 2000). Coiled-coil (CC) motifs are commonly found in the N-terminal structural domains of genes encoding non-TIR-NBS-LRR proteins, and members of the two superfamilies can be distinguished using conserved amino acid sequence motifs in the NBS domain (Meyers et al. 1999; Pan et al. 2000). Hundreds of NBS-LRR encoding genes have been identified in sequenced plant genomes and create a complex pathogen defense network—the Arabidopsis genome harbors ca. 150, the Populus trichocarpa and Medicago truncatula genomes harbor ca. 400, and the rice (Orzya sativa L.) genome harbors ca. 600 NBS-LRR encoding genes (Meyers et al. 1999, 2003; Dangl and Jones 2001; Cannon et al. 2002; Goff et al. 2002; Richly et al. 2002; Holt et al. 2003; Jones and Dangl 2006; Ameline-Torregrosa et al. 2007; Kohler et al. 2008). The number and diversity of NBS-LRR encoding genes in the sunflower (Helianthus annuus L.; 2n = 2x = 34) genome are not known.

NBS-LRR encoding genes have been isolated from numerous species by cloning and sequencing genomic DNA fragments amplified by degenerate oligonucleotide primers complementary to conserved amino acid sequence motifs in the NBS domain, typically from the P-loop to the GLPALP motif (Kanazin et al. 1996; Leister et al. 1996; Yu et al. 1996; Shen et al. 1998). The spectrum of resistance gene candidates (RGCs) isolated from sunflower using the degenerate primer strategy has been limited, and only a few of the previously identified RGCs have been genetically mapped, primarily using RFLP markers (Gentzbittel et al. 1998; Gedil et al. 2001). The latter facilitated the discovery and cloning of downy mildew (Plasmopara halstedii (Farl.) Berl. & De Toni 1888) R-genes found in two tightly linked families of NBS-LRR encoding genes (Gentzbittel et al. 1998; Bert et al. 2001; Gedil et al. 2001; Bouzidi et al. 2002; Radwan et al. 2003, 2004; Slabaugh et al. 2003).

EST databases are another resource for identifying NBS-LRR encoding genes (Zhu et al. 2002; Rossi et al. 2003). The sunflower genome is physically large (3,500 Mbp; Baack et al. 2005) and has not been sequenced; however, 326,421 ESTs have been produced from H. annuus, Jerusalem artichoke (Helianthus tuberosus L.; 2n = 6x = 102), silverleaf sunflower (Helianthus argophyllus Torr. and Gray; 2n = 2x = 34), and other wild species (http://cgpdb.ucdavis.edu/cgpdb2/). We identified several hundred RGCs by mining the sunflower EST database and sequencing genomic DNA fragments amplified from H. annuus, H. tuberosus, and other wild sunflower species using degenerate primers complementary to conserved amino acid sequence motifs in the NBS domain. The newly isolated RGCs are described here. Nearly one-fourth of the unique NBS-LRR encoding genes identified in the present study were genetically mapped using insertion–deletion (INDEL) or single-strand conformational polymorphism (SSCP) markers (Orita et al. 1989; Kukita et al. 1997; Larsen et al. 1999, 2007). The analyses described here build a more complete picture of the genetic diversity and genomic distribution of NBS-LRR encoding genes in sunflower and supply a genomic framework for identifying and isolating recognition-dependent R-genes conferring resistance to a wide range of pathogens of sunflower.

Materials and methods

Plant materials

Genomic DNAs were isolated from young leaves harvested from greenhouse grown plants of an elite oilseed inbred line (RHA373) and a wild self-incompatible outbred population (ANN1811; PI 494567) of H. annuus and wild, self-incompatible, outbred populations of H. argophyllus (ARG-1807, PI 494573), Helianthus deserticola (DES-2345, PI 649872), Helianthus paradoxus (PAR-Cibola), and H. tuberosus (TUB-49, PI 650095). Seeds of PAR-Cibola were supplied by Dr. Loren Rieseberg (Indiana University, Bloomington, Indiana). Seeds of the other germplasm accessions were supplied by the United States Department of Agriculture (USDA) Agricultural Research Service (ARS) National Plant Germplasm System (http://www.ars-grin.gov/npgs/) or the USDA-ARS Northern Crop Science Research Laboratory. DNA was isolated from manually ground fresh or frozen leaf tissues using a modified CTAB (cetyltrimethylammonium bromide) method (Murray and Thompson 1980).

NBS fragment isolation by PCR amplification using degenerate primers

Genomic DNA fragments spanning conserved NBS sequences were isolated from common and wild sunflower species using two pairs of degenerate forward and reverse oligonucleotide primers (NBS-F1/NBS-R1 and NBS-F2/NBS-R2) complementary to conserved P-loop and GLPLAL motifs in the NBS domain (Table 1; Leister et al. 1996; He et al. 2004) and were predicted to amplify ca. 500 bp DNA fragments. Standard PCR methods were used, amplicons were separated on 1.5% agarose gels (Sigma, USA), and bands of the predicted length were excised and purified using the QIAquick method (Qiagen USA), cloned into the pCR 4-TOPO vector (Invitrogen, USA), chemically transformed into TOP10 E. coli cells (Invitrogen, USA), and plated on Luria broth agar containing 100 mg of ampicillin/l. We randomly selected 192 clones from the ANN1811, ARG-1807, DES-2345, PAR-Cibola, and TUB-49 libraries and 288 clones from the RHA373 library for DNA sequencing. The selected clones were bidirectionally Sanger sequenced on an ABI3730XL using ABI Big-Dye technology (Applied Biosystems, USA). Sequences of clones harboring putative pseudogenes were substantiated by resequencing.

Table 1 Conserved NBS amino acid sequences targeted by degenerate oligonucleotide primers for isolating genomic DNA fragments from sunflower
Full size table

Mining sunflower and lettuce EST databases for NBS-LRR encoding resistance gene homologs

Sunflower and lettuce transcript assemblies (TAs) in the Compositae Genome Database (CGPdb; http://www.cgpdb.ucdavis.edu/) were screened to identify unigenes homologous to NBS-LRR encoding genes using “NBS”, “LRR”, and “NBS-LRR” keyword searches (TAs were BLAST annotated). Homologs with BLASTX scores ≤e−0.15 were selected, cDNA sequences were translated in six frames using the CGPdb pipeline (http://cgpdb.ucdavis.edu/database/sms/translation.php), and BLASTX analyses were performed against the NCBI Protein Database to substantiate the homology of the selected sunflower and lettuce unigenes to NBS-LRR encoding R-genes identified in other plant species (Altschul et al. 1997; http://www.ncbi.nlm.nih.gov). Sunflower NBS-LRR sequences were used to search for lettuce homologs in the CGPdb EST database.

Sequence and genetic diversity analyses

Nucleotide binding site nucleotide and amino acid sequences were aligned using Clustal_X (Thompson et al. 1997). JALVIEW was used to display and identify redundant sequences (Clamp et al. 2004; http://www.jalview.org/). NBS amino acid sequences were screened for motifs characteristic of proteins encoded by plant disease resistance genes and were classified as non-TIR or TIR using amino acid sequence motifs characteristic of the two superfamilies (Bent 1996; Hammond-Kosack and Jones 1997; Meyers et al. 1999, 2003; Pan et al. 2000). TIR sequences were identified by the presence of the RNBS-A TIR motif (FLENIRExSKKHGLEHLQKKLLSKLL) and aspartic acid (D) as the last amino acid residue in the Kin-2 motif, whereas non-TIR sequences were identified by the presence of the RNBS-A non-TIR motif (FDLxAWVCVSQxF) and tryptophan (W) as the last amino acid residue in the Kin-2 motif (Meyers et al. 1999, 2003; Pan et al. 2000).

Because many of the newly isolated NBS sequences (n = 758) were closely related (>90% amino acid similarity), sequences with <90% amino acid similarity (n = 69) were selected for in-depth analysis. The NBS nucleotide sequence alignment for the newly isolated RGCs (n = 758) and the amino acid sequence alignment for the selected sunflower NBS sequences (n = 69) have been supplied as GeneDoc input files [Appendices (S1) and (S2)]; GeneDoc is freeware and can be downloaded from http://www.nrbsc.org/gfx/genedoc/index.html or http://www.nrbsc.org/downloads/. Neighbor-joining (NJ) trees were produced from analyses of NBS amino acid similarity matrices with bootstrap resampling (k = 1,000 permutations) (Saitou and Nei 1987; Thompson et al. 1997). Two trees were produced, one from an alignment and analysis of the 69 selected sunflower sequences with <90% amino acid similarity and 24 lettuce (Lactuca sativa L.) sequences selected from previously identified families (Meyers et al. 1998b; Kuang et al. 2004) and one from an alignment and analysis of sunflower and lettuce sequences with sequences from diverse NBS-LRR encoding R-genes isolated from other plant species: RPS4 (AJ243468), RPP1 (AF098962), ADR1 (AJ581996), RPS5 (AF074916), RPM1 (X87851), RPP13 (AF209730), RPS2 (ATU14158), RPP5 (ATRPP5LE1), and RPP8 (AF089711) from Arabidopsis; L6 (LUU27081) from flax (Linum usitatissimum L.); N (NGU15605) from tobacco (Nicotiana tabacum L.) and NRG1 (DQ054580) from Nicotiana benthamiana; Mi (AF039681) and Rx (AJ011801) from potato (Solanum tuberosum L.); I2 (AF004878) from tomato (Solanum lycopersicon L.); Xa1 (AB002266), Pi-ta (AF207842), and Pib (AB013449) from rice (Oryza sativa L.); Rp1 (AF107293) from maize (Zea mays L.); and Dm3 (LSRGC2B1) from lettuce (GenBank accession numbers are shown in parentheses).

DNA marker development, genotyping, and genetic mapping

Single-strand conformational polymorphism and INDEL markers were developed for genotyping NBS-LRR loci by designing flanking oligonucleotide primers complementary to cDNA and genomic DNA reference sequences. Whenever possible, nucleotide polymorphisms were used to identify paralog-specific sequences for designing primers for INDEL and SSCP marker genotyping. INDEL markers were manually genotyped on 1.5% NuSieve agarose gels (FMC, USA). SSCP markers were manually genotyped by silver-staining polyacrylamide gels (Bassam et al. 1991; Slabaugh et al. 2003). PCRs were performed using 30 ng of DNA template, 0.65 U Taq polymerase (Qiagen, USA), 1× PCR reaction buffer, 2.5 mM Mgcl2, 0.2 mM dNTPs, 0.16 μM of each primer (one pair/marker), and a ‘touchdown’ PCR protocol (Don et al. 1991; Hecker and Roux 1996) with an initial denaturation temperature of 94 C for 3 min, followed by 1 cycle of 94 C for 30 s, 68 C for 30 s and 72 C for 60 s; annealing temperatures in subsequent cycles were decreased by 1 C until reaching 58 C, products were amplified for 35 cycles at 94 C for 30 s/cycle, 58 C for 30 s/cycle, and 72 C for 60 s/cycle, and products were extended in the final cycle at 72 C for 15 min. Genomic DNA amplicons were separated on agarose to screen for INDEL polymorphisms and, when absent, were separated and genotyped using SSCP analysis. Polymorphic INDEL and SSCP markers were genotyped on 94 RHA280 × RHA801 recombinant inbred lines (RILs) (Tang et al. 2002; Yu et al. 2003) or 94 NMS373 × (NMS373 × ANN1238) BC1 progeny (Gandhi et al. 2005). NBS-LRR INDEL and SSCP marker loci were anchored, grouped, and ordered against a previously mapped genome-wide backbone of SSR and INDEL markers (Tang et al. 2002; Yu et al. 2003; Gandhi et al. 2005). Genetic mapping analyses were done using MAPMAKER (Lander et al. 1987) as previously described (Tang et al. 2002). DNA marker loci were ordered by allowing 0 or 2% genotyping errors (Lincoln and Lander 1992), primarily for comparing likelihoods of locus orders in genomic regions densely populated with NBS-LRR marker loci.

Results

Sunflower RGCs identified by targeting conserved NBS sequence motifs and mining the EST database

Genomic DNA fragments were amplified from H. annuus, H. argophyllus, desert sunflower (H. deserticola Heiser; 2n = 2x = 34), Pecos sunflower (H. paradoxus Heiser; 2n = 2x = 34), and H. tuberosus using two pairs of degenerate oligonucleotide primers complementary to highly conserved NBS sequence motifs (Leister et al. 1996; Meyers et al. 1999; He et al. 2004, Tables 1, 2). Both primer pairs (NBS-F1/R1 and NBS-F2/R2) were predicted to amplify ca. 500 bp fragments from TIR and non-TIR-NBS-LRR encoding genes. Collectively, 1,248 amplicons were cloned and sequenced (196–288/genotype), 1,052 of the 1,248 resequenced amplicons (RSAs) were homologous to NBS sequences in NBS-LRR encoding genes previously isolated from sunflower or other plant species, 784 RSAs had uninterrupted ORFs from the P-loop to the GLPLAL motif, 542 RSAs were unique and lacked stop codons or frame shift mutations, and 242 RSAs harbored stop codons or frame shift mutations and were suspected to be pseudogenes [Table 2; GenBank Acc. No. EF559385-EF560169; Appendix (S3)]. Sequencing errors should have been minimal because each clone was bidirectionally sequenced and putative pseudogenes were resequenced. The 542 unique NBS sequences ranged in length from 164 to 173 amino acids (GenBank Acc. No. ABQ58077-ABQ57529). Three-fourths of the newly isolated NBS sequences lacking stop codons were TIR class (Meyers et al. 1999, 2003; Pan et al. 2000). Nearly half of the NBS sequences isolated from H. annuus were TIR (53.6%) class, whereas 93.9% of the NBS sequences isolated from H. tuberosus were TIR class. Of the NBS sequences harboring stop codons, 95.3% were TIR class [Appendix (S3)].

Table 2 Number of unique NBS sequences isolated from four annual diploid (2n = 2x = 34) species of sunflower (H. annuus, H. argophyllus, H. deserticola, and H. paradoxus) and Jerusalem artichoke (H. tuberosus), a perennial hexaploid (2n = 6x = 102), by sequencing genomic DNA fragments amplified by PCR using degenerate oligonucleotide primers complementary to conserved NBS sequences
Full size table

Eighty-eight unigenes homologous to NBS-LRR encoding R-genes were identified by screening transcript assemblies (TAs) developed from 284,745 Helianthus ESTs in the Compositae Genome Program Database (CGPdb; http://cgpdb.ucdavis.edu/cgpdb2/). The ESTs spanned pre-NBS, NBS, LRR-, or post-LRR domains or combinations thereof [Appendix (S3)]. ESTs spanning NBS domains were aligned with previously and newly isolated NBS domain sequences (Gedil et al. 2001; Radwan et al. 2003; Plocik et al. 2004). Nucleotide identities ranged from 28 to 100%, while amino acid similarities ranged from 16 to 100%.

Diversity of sunflower homologs encoding NBS-LRR resistance proteins

Sunflower NBS sequences (n = 758) spanning the P-loop to the GLPLAL motif were translated and aligned to identify and eliminate redundant and closely related (>90% amino acid similarity) sequences and select a subset of unique sequences (<90% amino acid similarity) for genetic diversity analyses; 69 NBS sequences were selected for cluster analyses [Appendix (S1, S2)]. Of the latter, 15 were previously identified (Gedil et al. 2001; Radwan et al. 2003; Plocik et al. 2004) and belonged to a single TIR family (see group T3 in Fig. 1), eight were ESTs identified in the CGPdb, and 46 were newly isolated genomic DNA sequences [Appendix (S3)]. The 69 sunflower NBS sequences were aligned with 28 lettuce NBS sequences. Of the latter, 21 were previously identified and shared <98% identity (Meyers et al. 1998b; Kuang et al. 2004) and seven were identified by mining the lettuce EST database (http://cgpdb.ucdavis.edu/cgpdb2/). Several other lettuce ESTs homologous to NBS-LRR encoding genes were identified, but spanned non-NBS sequences.

Fig. 1
figure 1

Neighbor-joining tree produced from an analysis of the amino acid similarity matrix for 69 sunflower (Ha prefix) nucleotide binding site (NBS) sequences with <90% amino acid identity and 26 lettuce (Ls prefix) NBS sequences. Seven non-TIR (N1–N7) and eight TIR groups (T1–T8) were identified. Numbers shown on branches are percentages of bootstrap replications supporting nodes

Full size image

The sunflower NBS sequences clustered into 14 groups with strong bootstrap support, eight solely comprised TIR (T1–T8) and six solely comprised of non-TIR (N1–N6) NBS sequences (Fig. 1). Two groups (T7 and N6) were populated with ESTs only. Nine groups (T1–5, T8, and N3–5) lacked ESTs; however, most of the ESTs spanned non-NBS sequences and could not be classified [Appendix (S3)]. One additional non-TIR group was identified (N7) and was solely populated with lettuce sequences (Fig. 1). The latter belong to a large and diverse cluster of NBS-LRR encoding genes harboring downy mildew R-genes in lettuce (Kuang et al. 2004). The other lettuce RGCs clustered with sunflower RGCs in two of the seven non-TIR groups (N1 and N6) and two of the eight TIR (T4 and T6) groups, with 35–98% identity among sequences within a group. Sunflower RGCs homologous to lettuce RGCs in group N7 have not been identified and lettuce RGCs homologous to sunflower RGCs in several groups (N2–5, T1–3, T5, and T7–8) have not been identified.

Sunflower and lettuce NBS sequences selected to sample diversity across the 15 groups (Fig. 1) were clustered with NBS sequences from 20 previously identified monocot and dicot R-genes (Fig. 2). The selected RGCs were highly similar to one or more known TIR and non-TIR-NBS-LRR encoding R-genes. Two non-TIR groups were identified, one harboring ADR1, NRG1, and a single sunflower (Ha-RGC185) and lettuce (Ls-BQ866268) RGC each (non-TIR1), and the other harboring the other non-TIR R-genes and RGCs (non-TIR2). The latter harbored sunflower and lettuce RGCs from each of the previously identified non-TIR groups other than N3 (Figs. 1, 2). Sunflower and lettuce RGCs in the TIR superfamily were most closely related to N, a TIR-NBS-LRR encoding R-gene isolated from tobacco (N. tabacum L.) (GenBank Acc. No. NGU15605; Whitham et al. 1994; Lawrence et al. 1995).

Fig. 2
figure 2

Neighbor-joining tree produced from an analysis of the amino acid similarity matrix for 26 sunflower (Ha prefix) and six lettuce (Ls prefix) NBS sequences and NBS sequences for 20 plant R-genes (see text)

Full size image

Genetic mapping and genomic distribution of sunflower homologs encoding NBS-LRR resistance proteins

Single-strand conformational polymorphism and INDEL markers were developed for 196 RGCs, 88 from ESTs and 108 from RSAs [Appendix (S4)]. The parents of the RHA280 × RHA801 RIL (Tang et al. 2002) and NMS373 × ANN1811 BC1 (Gandhi et al. 2005) mapping populations were screened for presence-absence polymorphisms (dominant INDELs), SSCPs, or both (Fig. 3). Because NBS-LRR encoding genes are commonly duplicated, allelic and non-allelic strands were separated by SSCP analysis to facilitate paralog discovery and mapping. The number of loci amplified by a particular SSCP marker was inferred from the number of strands identified, e.g., two loci were inferred when four strands were amplified and three loci were inferred when six strands were amplified from one or both parents, assuming two complementary strands per allele (Bassam et al. 1991). Of the 196 DNA markers developed and screened in the present study, 43 amplified a single locus, 153 amplified two or more loci, and 150 (76.5%) were polymorphic (10 were polymorphic for two or three loci each).

Fig. 3
figure 3

Genotypes for two NBS-LRR INDEL (RGC124 and 126) and 10 NBS-LRR SSCP (RGC38, 44, 116, 125, 129, 137, 139, 141, 165, and 185) markers among RHA280 × RHA801 recombinant inbred line (RIL) or NMS373 × (NMS373 × ANN1811) BC1 mapping population parents and progeny

Full size image

Collectively, 167 NBS-LRR loci were genetically mapped, nine by genotyping INDEL and 158 by genotyping SSCP markers (Figs. 3, 4). INDEL and SSCP markers for NBS-LRR loci were integrated into a framework of previously mapped SSR and INDEL marker loci distributed throughout the sunflower genome (Tang et al. 2002; Yu et al. 2003; Gandhi et al. 2005). One-third of the mapped NBS-LRR markers (58/167) were developed from ESTs, 66.9% of the SSCP markers developed from ESTs were polymorphic (58/88), and 88.9% of the SSCP markers developed from RSAs were polymorphic (96/108). NBS-LRR loci were found throughout the genome—the upper and lower segments of most linkage groups harbored at least one NBS-LRR locus (Fig. 4). The 17 linkage groups (x = 17) displayed in Fig. 4 were developed by integrating NBS-LRR loci mapped in the NMS373 × ANN1811 BC1 population into the framework of DNA marker loci mapped in the RHA280 × RHA801 RIL population by interpolating genetic distances and binning loci within the shortest interval possible flanked by common DNA marker loci–individual linkage groups for the two populations are displayed in a comparative mapping database (http://www.sunflower.uga.edu/CMAP) developed using CMAP (http://www.gmod.org/wiki/index.php/CMAP; Gonzales et al. 2005).

Fig. 4
figure 4figure 4figure 4

Genomic locations for 167 NBS-LRR loci in sunflower (2n = 2x = 34) mapped using INDEL or SSCP markers. Species sources for DNA markers developed from ESTs are shown as prefixes in locus names (ANN, H. annuus; ARG, H. argophyllus; PET H. petiolaris; TUB, H. tuberosus; LSE, L. serriola). Groups (clusters), when known, are identified by alphanumeric suffixes (T1–T8 for TIR and N1–N7 for non-TIR-NBS-LRR groups) in locus names (groups for SSCP or INDEL markers developed from non-NBS sequences were not known). Putative paralogous loci found on different linkage groups are identified by numerical suffixes, e.g., the RGC131 (a member of group T1) SSCP marker identified paralogous loci on linkage groups 9 and 10 (RGC131-9-T1 and RGC131-10-T1). Putative paralogous loci found on a single linkage group are identified by alphanumeric suffixes, e.g., the RGC145 (a member of group T3) SSCP marker identified paralogous loci (A and B) on linkage group 9 (RGC145-9A-N3 and RGC145-9B-N3)

Full size image

Wild species ESTs were a particularly rich source of novel RGCs, with 31 originating from H. tuberosus, 20 originating from domesticated and wild H. annuus, three originating from prairie sunflower (H. petiolaris Nutt.; 2n = 2x = 34), one originating from H. argophyllus, and one originating from prickly lettuce (L. serriola L.). ESTs facilitated the discovery of 14 new NBS-LRR clusters or singletons, which were found on the upper and lower segments of linkage group (LG) 1, upper segments of linkage groups 2, 3, 4, 5, 9, 10, and 14, and lower segments of linkage groups 6, 8, 12, 15, 16, and 17 (Fig. 4). Several NBS-LRR homologs identified from ESTs mapped to the two largest and most complex and well known clusters in sunflower, a TIR cluster (T3) on the upper end of LG 8 and a non-TIR cluster (N1) on the lower end of LG 13 (Gentzbittel et al. 1998; Bert et al. 2001; Gedil et al. 2001; Bouzidi et al. 2002; Radwan et al. 2003; Slabaugh et al. 2003). NBS-LRR encoding genes from each of the 14 groups identified from analyses of the NBS amino acid sequence alignment [Fig. 1; Appendix (S2)], other than N7, were genetically mapped, and members of several groups were found on two or more linkage groups. Collectively, 44 NBS-LRR clusters or singletons were identified by genetic mapping (Fig. 4).

Discussion

The NBS-LRR markers mapped in the present study supply landmarks for identifying and isolating R-genes conferring resistance to diverse pathogens in sunflower [Fig. 4; Appendix (S3, S4)]. Of the 14 NBS-LRR groups identified by cluster analysis (Figs. 1, 2), only four had previously been identified (Gedil et al. 2001; Radwan et al. 2003; Plocik et al. 2004). Moreover, of the 44 NBS-LRR clusters and singletons mapped in the present study (Fig. 4), only four had previously been mapped and only two were linked to previously identified R-genes, the T3 cluster on LG 8 and the N1 cluster on LG 13 harboring downy mildew R-genes (Gentzbittel et al. 1998; Bert et al. 2001; Gedil et al. 2001; Bouzidi et al. 2002; Radwan et al. 2003; Slabaugh et al. 2003). Previously, RGCs had not been identified for downy mildew R-genes on LG 1 (Pl ARG), black rust (Puccinia helianthi Schw.) R-genes on LGs 8 (R 1) and 13 (R ADV), a broomrape [Orobanche cernua var. cumana (Wallr.) G. Beck] R-gene on LG 3 (Or 5), and a chlorotic mottle virus R-gene on LG 14 (Rcmo-1) (Gentzbittel et al. 1998; Bert et al. 2001; Gedil et al. 2001; Bouzidi et al. 2002; Radwan et al. 2003, 2004; Slabaugh et al. 2003; Tang et al. 2003; Yu et al. 2003; Dußle et al. 2004; Pérez-Vich et al. 2004; Lenardon et al. 2005); the locations of R 1 and R ADV were inferred from linkages of both loci to SCAR markers (Lawson et al. 1998) mapped by Yu et al. (2003) and cross-referenced in the present study. Other than Rcmo-1 (Lenardon et al. 2005), newly identified and mapped NBS-LRR loci were linked to each of the previously mapped recognition-dependent R-genes in sunflower (Pl ARG, Pl 8, R 1, R ADV, and Or 5). RGCs linked to Pl 8, a downy mildew resistance gene on LG 13, were previously identified by mapping RFLP markers (Radwan et al. 2003, 2004). The DNA sequences and markers described here [Appendix (S1–S4)] should greatly facilitate the isolation of Pl 8 and other R-genes in sunflower.

While 167 NBS-LRR loci were mapped in the present study, including 100% of the polymorphic INDEL or SSCP markers developed from ESTs and members of each of the 14 sunflower NBS-LRR groups (Fig. 1), the genomic locations of several hundred of the newly identified RGCs are not known, although most of the unmapped RGCs were highly similar (>91% amino acid similarity) to one or more of the mapped RGCs [Appendix (S3); Fig. 4]. The development and genetic mapping of additional NBS-LRR markers could be accelerated using multiplex capillary SSCP genotyping (Larsen et al. 1999, 2007) or other high-throughput DNA fragment analysis methods (Slabaugh et al. 2003) and might identify unique loci or clusters of loci, although one or more members of nearly every group have been mapped. Until the sunflower genome has been sequenced, the discovery of RGCs will be limited to the sorts of forward genetic and comparative genomic approaches described here, as has been done in many other plant species (Kanazin et al. 1996; Leister et al. 1996; Yu et al. 1996; Collins et al. 1998; Shen et al. 1998; Pan et al. 2000; Wang et al. 2001; Wei et al. 2002; Zhu et al. 2002; Lee et al. 2003; Madsen et al. 2003; Meyers et al. 2003; He et al. 2004; Rossi et al. 2003). On the basis of analyses of sequenced plant genomes (Meyers et al. 2003; Ameline-Torregrosa et al. 2007; Kohler et al. 2008), the full complement of NBS-LRR encoding genes in the sunflower genome is undoubtedly larger than the complement isolated so far [Appendix (S1–S3)].

Wild species are rich sources of genetic diversity for disease resistance and other traits in domesticated plants (Tanksley and McCouch 1997; Michelmore 2003) and have been an important source of R-genes in sunflower (Pustovoit and Kroknin 1978; Škoric 1985; Miller and Gulya 1988, 1991; Seiler 1991a, b; Tan et al. 1992; Quresh et al. 1993; Degener et al. 1999; Viguié et al. 2000; Langar et al. 2002; Miller et al. 2002; Vear et al. 2003; Slabaugh et al. 2003; Dußle et al. 2004). Several novel sunflower NBS-LRR loci were identified in the present study by mining wild species ESTs and isolating NBS sequences from wild species. We sequenced more deeply than had been done in earlier studies and targeted wild species to increase the depth and breath of sequences discovered by the degenerate primer approach, particularly since wild species have played an important role in disease resistance breeding in sunflower, e.g., the downy mildew R-genes found on three chromosomes (Pl ARG on LG 1, Pl 7 on LG 8, and Pl 8 on LG 13) were introgressed from wild species (Miller et al. 2002; Slabaugh et al. 2003; Dußle et al. 2004).

Between ESTs and RSAs, at least 918 unique NBS-LRR sequences have been identified in sunflower so far [Gentzbittel et al. 1998; Bert et al. 2001; Gedil et al. 2001; Bouzidi et al. 2002; Radwan et al. 2003, 2004; Slabaugh et al. 2003; Plocik et al. 2004; Appendix (S3)]. Wild species ESTs were a particularly rich source of novel RGCs, were the sole source of NBS-LRR encoding genes in four of 10 newly identified groups (T2, T4, T8, and N5), and were the sole source of RGCs linked to previously mapped R-genes for which RGCs have not been identified (Figs. 1, 4). H. tuberosus ESTs supplied the largest number of novel RGCs, several of which were linked to R-genes previously identified from resistance phenotypes. H. tuberosus has been a historically important source of R-genes transferred across the diploid × hexaploid bridge, e.g., the first dominant race-specific broomrape R-gene (Or 1) was discovered in H. tuberosus, perhaps as early as 1916, and was introgressed into H. annuus through backcross breeding (Vranceanu et al.1980; Parker and Riches 1993). Several clusters of NBS-LRR loci identified from ESTs and RSAs are not linked to R-genes identified from resistance phenotypes, but could be the source of as yet undiscovered R-genes.

The NBS-LRR copy numbers often vary widely within species, and NBS-LRR loci frequently rearrange and evolve through recombination, unequal crossing-over, gene conversion, insertions–deletions, and point mutations (Leister et al. 1998; Michelmore and Meyers 1998; Meyers et al. 1998a; Noel et al. 1999; Chin et al. 2001; Noir et al. 2001; Zhu et al. 2002; Ashfield et al. 2004; Kuang et al. 2004). The copy number in a large cluster of NBS-LRR loci harboring Pl 1 and other downy mildew R-genes on linkage group 8 is highly variable in sunflower, e.g., 2 to 24 copies/individual have been identified among elite and exotic individuals using INDEL fingerprinting (Slabaugh et al. 2003). The LG 8 cluster is the largest and most complex discovered so far in sunflower, with 54 NBS-LRR loci spanning a 36 cM segment in three subclusters (Fig. 4). One subcluster spans 12 cM, is primarily populated with T3 group members, has two T4 group members in the middle of the cluster (RGC138 and RGC139), and is demarcated on the upper end by two apparently telomeric or near-telomeric loci (RGC230 and RGC237) and lower end by additional T3 group NBS-LRR loci (e.g., RGC244). The second subcluster is slightly downstream, spans 7 cM, and harbors four T3 group members. Finally, additional unclassified NBS-LRR loci were found on the lower end of LG 8, two of which were identified from ESTs and could not be classified (ANN-RGC198 and PET-RGC221).

The second largest NBS-LRR cluster identified so far in sunflower is found on the lower end of LG 13, with 27 loci in two subclusters and a few singletons spanning a 31 cM segment (Fig. 4). This cluster harbors TIR- and non-TIR-NBS-LRR encoding genes. Other than three TIR-NBS-LRR loci, two from the T3 group at the upper end of the segment and one from the T1 group, loci in this cluster belong to a single non-TIR group (N1). This segment harbors downy mildew (Pl 5 and Pl 8) and black rust (R ADV) R-genes (Lawson et al. 1998; Radwan et al. 2003; Slabaugh et al. 2003; Yu et al. 2003), and is now much more densely populated with RGCs, which should facilitate the identification and cloning of R-genes for Pl 8 and R ADV (Fig. 4). The presence of complex mixed clusters of NBS-LRR loci, such as the TIR and non-TIR cluster found on LG 13, is common in the genomes of higher plants (Meyers et al. 1999; Zhu et al. 2002).

While an unknown number of genes encoding NBS-LRR proteins have presumably not yet been discovered in sunflower, a significant number have and supply resources for cataloging and cross-referencing phenotypic and quantitative trait loci conferring resistance to diverse pathogens in sunflower and, when coupled with the steadily growing collection of NBS-LRR sequences and mapped NBS-LRR loci, should greatly facilitate the discovery of additional R-genes and a greater understanding of R-gene diversity in domesticated and wild sunflower species (Michelmore and Meyers 1998; Noel et al. 1999; Ellis et al. 2000; Noir et al. 2001; Cannon et al. 2002; Michelmore 2003). Wild species have been an important source of R-genes in sunflower and, not coincidentally, RGCs identified from wild species ESTs were linked to several important R-genes in sunflower (Pl ARG, Or 5, Pl 8, R 1, and R ADV). Genetic analyses of RGCs for many of these R-genes are underway and should lead to the discovery of functionally important determinants of resistance and the development of additional high-throughput DNA markers for accelerating disease resistance breeding in sunflower through marker-assisted selection.