Introduction

Thus far, more than 50 resistance (R) genes have been cloned from a variety of plant species. These R-genes confer resistance against various pathogens, including bacteria, fungi, viruses, pests, and nematodes (Whitham et al. 1994; Buschges et al. 1997; Collins et al. 1999; Wang et al. 1999; Hulbert et al. 2001). Most isolated R-genes have conformed to the typical ‘gene-for-gene hypothesis’, except for the dominant Hm1 gene in maize as well as several recessive R-genes, such as the mlo in barley and the edr1 in Arabidopsis (Johal and Briggs 1992; Buschges et al. 1997; Frye and Innes 1998; Kim et al. 2002). Despite their broad resistance, most R-genes share limited conserved domains or motifs, such as nucleotide binding site (NBS), leucine-rich repeat (LRR), protein kinase (PK), transmembrane (TM), leucine zipper (LZ), and Toll-Interleukin-1 (TIR) (Traut 1994; Cooley et al. 2000; Liu and Ekramoddoullah 2004). Of 50 R-genes cloned so far, at least 37 share the NBS-LRR domain in their protein sequences, and forming the largest R-gene class (Bent et al. 1994; Lawrence et al. 1995; Kaloshian et al. 1998; Yoshimura et al. 1998). This R-gene class can be further divided into two sub-classes, TIR-NBS-LRR and CC-NBS-LRR, based on their N-terminal structure (Meyers et al. 1999). The second R-gene class equipped with the LRR-TM-PK domains integrates functions of both signal perception and transduction, for example the Xa21 gene in rice (Song et al. 1995; Liu et al. 2002). The third R-gene class, Cf-9 and Cf-4 among others, is characterized with a large and symmetrical extra-cellular LRR domain that can combine with avr-polypeptide or other proteins for signal reception (Jones et al. 1994; Thomas et al. 1997). The fourth R-gene class contains only the PK domain, such as Pto and Rpg1 (Martin et al. 1993; Brueggeman et al. 2002).

Most R-genes have been cloned via two approaches, transposon tagging and map-based cloning (or positional cloning). The first method depends on inactivation of the target gene caused by transposon insertion. Currently, only a few transposons are suitable for tagging, and the likelihood that a transposon is inserted within the target R gene is very low (Brutnell 2002). The second approach requires saturation of the target R gene with molecular markers. In model plants with available genome sequences, such as Arabidopsis and rice, high-density markers can be readily developed (The Arabidopsis Genome Initiative 2000; Yu et al. 2002). In maize, however, development of high-density markers in the target R region is rather difficult, if not impossible, due to lack of availability of the complete genome sequence. Fortunately, most R-genes share limited conserved domains, and these domains can be exploited for designing degenerate primers to amplify other unknown R-gene analogs (RGAs) and mine EST databases in order to recover R-gene-like sequences (Kanazin et al. 1996; Graham et al. 2000; Dilbirligi and Gill 2003; Hunger et al. 2003; Madsen et al. 2003; Rossi et al. 2003). If sufficient RGAs are available to cover genome-wide R-genes for a particular plant species, this reservoir of RGAs can serve as a candidate R-gene resource for cloning of unknown R-genes. This candidate gene approach is particularly productive for cloning R-genes in plants having either large genomes or those with no available genome sequences such as wheat and maize.

So far, PCR amplification using degenerate primers designed based on conserved motifs is a common approach for identifying RGAs in various plant species. A total of 11 classes of non-cross-hybridizing sequences, having high-level amino acid identities with those of NBS-LRR R-genes, have been isolated and mapped in the maize genome (Collins et al. 1998). Based on both sequence comparisons and hybridization patterns, RGAs isolated from maize, rice, and wheat have been assigned into 13 classes, and 12 were mapped onto 17 loci on the barley genome using a doubled-haploid mapping population (Collins et al. 2001). In Arabidopsis, 42 non-redundant R-gene-like ESTs and three Pto-like sequences amplified with degenerate primers have been mapped onto 47 loci (Botella et al. 1997). In rice, 109 R-gene/defense-gene like ESTs were mapped to regions harboring major R-genes or QTLs using a doubled-haploid mapping population (Wang et al. 2001). To date, a sizable collection of ∼800 RGAs have been obtained via PCR amplification from nearly 20 plant species, and one of these RGAs has already been confirmed to correspond to a functional R-gene (DM3) (Shen et al. 2002).

With advances in genomics programs, EST data have been steadily accumulating in many plant species, thus allowing for identification of RGAs from EST databases via the data-mining approach. Moreover, RGAs screened from EST databases are derived from expressed R-genes, and therefore are more desirable for cloning R-genes via the candidate gene approach. In wheat, 220 expressed R-gene candidates have been identified using the data-mining and modified RNA fingerprinting methods (Dilbirligi et al. 2003). In sugarcane, 88 R-gene-like ESTs have been identified via the data-mining approach (Rossi et al. 2003). Ultimately, genome-wide analysis of R-gene homologues has been conducted in Arabidopsis and rice whereby whole genome sequences are now available. Consequently, nearly 200 coding sequences, similar to NBS-LRR R-genes, have been identified in the Arabidopsis genome (Meyers et al. 2003). Approximately 500 NBS-LRR R-genes belonging to the non-TIR sub-class are present in the rice genome (Monosi et al. 2004).

The overall goal of this research is to isolate as many RGAs as possible from the maize genome using three different approaches, including a modified AFLP method to amplify RGAs from maize genome, a modified RACE method to isolate expressed RGAs from maize cDNA population, and data-mining of maize EST databases, available in MaizeGDB and NCBI, to obtain R-gene-like ESTs.

Materials and methods

Plant material

The parental lines ‘1145’ and ‘Y331’ along with their BC1 population (using ‘Y331’ as recurrent parent) were kindly provided by Prof. Shaojiang Chen (National Maize Improvement Center of China) and used in the modified AFLP method. The parent ‘1145’ is a ‘universal’ resistant inbred line conferring resistance to various diseases, including rot stalk, head smut, Northern and Southern blights, and virus diseases, among others. Conversely, the parent ‘Y331’ is a rather susceptible inbred line with no reported resistance to any of the major diseases mentioned above. For the modified RACE method, the European inbred line FAP1360, conferring complete resistance against sugarcane mosaic virus (SCMV), was used. For the data-mining method, R-gene-like ESTs were separately cloned from two parental lines, ‘87-1’ and ‘Zong 3’, from which a recombinant inbred line (RIL) mapping population has been developed.

The modified AFLP method

Extraction of genomic DNA was performed according to the CTAB procedure (Chen and Ling 1996). About 100 ng of genomic DNA was double-digested to completion with restriction enzymes MseI/PstI. Digested DNA fragments were ligated to MseI- and PstI-adapters. Ligation products were then diluted with TE0.1 at a ratio of 1:10 for pre-amplification. PCR products were diluted with TE0.1 at a ratio of 1:50 for selective amplification using the degenerate primer (designed on P-loop motif GVGKTT in the NBS domain) combined with each of the 64 P-xxx selective primers (x represents one of the four nucleotides).

The selective amplification was performed in 10 μl containing 2 μM of each of four dNTPs, 2 μM Mg2+, 1 μl 10×PCR reaction buffer, 25 μM of each of the degenerate primer and of the selective primer, 0.5 U Taq DNA polymerase, and 2.5 μl diluted pre-amplification product. The amplification conditions were denatured at 94°C for 2 min, then cycling for 10 times using a touch-down strategy (an initial cycle of 94°C for 30 s, 65°C for 45 s, 72°C for 1 min, then lowering the annealing temperature for each cycle by 0.7°C during the following 9 cycles), followed by 30 cycles at 94°C for 30 s, 56°C for 45 s, and 72°C for 1 min.

The amplification products were size-separated on 6% denaturing polyacrylamide gels. AFLP bands were sliced from the gel and boiled in 20 μl of TE for 10 min. After centrifugation at 13,000× g for 10 min, the supernatant was used for re-amplification of the isolated DNA band. The PCR mixture was prepared in 50 μl volume containing 25 μM corresponding primers, 5 μl 10×PCR buffer, 2 U Taq DNA polymerase, 2 μM Mg2+, 2 μM each of four dNTP, and 0.5 μl supernatant. The PCR reaction was run at first step of denature at 94°C for 2 min, followed by 36 cycles of 94°C for 30 s, 56°C for 45 s, and 72°C for 1 min.

The modified RACE method

Leaf tissues of 2–3-week-old seedlings of inbred line FAP1360 were collected. Total RNA was isolated and precipitated using the LiCl method, and mRNA isolation was performed using PolyATract System III kit (Promega, Madison, WI, USA). The first-strand cDNA synthesis was catalyzed by PowerScript reverse transcriptase (BD SMARTTM RACE cDNA Amplification kit, BD Biosciences, Palo Alto, CA, USA) according to the manual. The modified RACE reaction was performed using both the degenerate primer (designed based on the P-loop motif) and the 3′-UPM primer (provided in the SMART RACE kit). The PCR mixture, 50 μl, contains the following reagents: 5 μl 10×UPM primer, 2 μl P-loop degenerate primer (10 μM), 1 μl 50×BD Advantage 2 polymerase mix, 5 μl 10×BD Advantage 2 PCR buffer, 2 μl dNTP mixture (10 mM each), and 35 μl PCR-grade water. The following touch-down PCR program was used for the RACE reaction: the first 5 cycles at 94°C for 30 s, 68°C for 30 s, and 72°C for 3 min, the next 5 cycles at 94°C for 30 s, 65°C for 30 s, and 72°C for 3 min, followed by 25 cycles at 94°C for 30 s, 60°C for 30 s, and 72°C for 3 min.

The PCR products from the modified RACE reaction were subjected to electrophoresis on 1% agarose gel. The main bands, ranging in size from 1.5 to 2.0 kb, were excised from the agarose gel and used for identifying RGAs.

Data mining of R-gene-like ESTs

The amino-acid sequences of 48 known R-genes, covering the four major R-gene classes, were used to search for (tBLASTn) homologous ESTs in online available maize EST databases, including MaizeGDB (http://www.maizegdb.org) and NCBI (http://www.ncbi.nlm.nih.gov). Those ESTs with scores ≥ 100 and E-values ≤ 10−10 were deemed tentative R-gene-like ESTs, and all hypothetical R-gene-like ESTs were used to find their corresponding unigenes in MaizeGDB. The resulting unigenes and singletons were in turn used to search the GenBank database by tBLASTx to confirm their putative R-gene-like functions. In addition, all R-gene-like ESTs were used to identify possible genomic sequences in the MAGI database in the Schnable laboratory (http://schnablelab.plantgenomics.iastate.edu/). For each R-gene-like unigene or EST, the sequence was used to design primers for amplification of corresponding sequences from both parental lines ‘87-1’ and ‘Zong 3’.

Cloning, sequencing, and analysis of PCR products

Polymerase chain reaction products from the above three methods were purified using Nucleo Trap Gel Extraction Kit (BD Biosciences), and these were ligated into the pGEM-T vector (Promega) at 16°C for overnight. The ligation mixture, 10 μl, consisted of the following reagents: 5 μl 2×rapid ligation buffer, 1 μl pGEM-T vector, 1 μl T4 DNA ligase, 2 μl purified PCR products, and 1 μl ddH2O. The ligation product was transformed into Escherichia coli strain DH10B by electroporation, and incubated overnight at 37°C in a solidified LB medium containing 50 μg/ml ampicillin, 20 μg/ml X-gal, and 20 μg/ml IPTG.

The positive clones were identified by PCR amplification using the flanking T7 and SP6 primers, and sequenced at the HuaDa Sequencing Center (Beijing, China). The software, ‘sequence assembly’, and ‘multiple sequence alignment’ in DNAMAN Version 4.0, were used to confirm either identities or similarities for all sequences obtained from positive clones. Functional confirmation of obtained sequences were performed using tBLASTx (http://www.ncbi.nlm.nih.gov).

Development of RGA-tagged molecular markers

In maize, 3′-UTRs were reported to show more sequence diversity than coding regions (Rafalski 2002). Therefore, RGAs amplified from the two parental lines ‘87-1’ and ‘Zong 3’ were carefully investigated for sequence divergence (InDels and SNPs) along their 3′-ends. The flanking regions of the InDels and SNPs were used to design primers to re-amplify both ‘87-1’ and ‘Zong 3’. For convenient separation of PCR products on 6% polyacrylamide gels, the distance between the forward and reverse primers was set at ∼500 bp. If distinct and polymorphic PCR bands were observed between ‘87-1’ and ‘Zong 3’, then an STS marker was deemed successfully developed for an InDel. Likewise, if polymorphic PCR bands were observed between ‘87-1’ and ‘Zong 3’ following digestion with a particular restriction enzyme, then a CAPS marker was deemed successful for a SNP. These developed RGA-tagged markers are indispensable in anchoring corresponding RGAs along the maize genome.

Results

The RGAs generated from the modified AFLP method

For each primer pair, 40–50 AFLP bands ranging from 200 to 800 bp in sizes were visualized on a 6% polyacrylamide gel. Most bands appeared in both parental inbred lines, and only a few bands appeared in the ‘universal’ resistant parent ‘1145’ and in resistant progeny, but not in the susceptible parent ‘Y331’. All resistance-related bands and some of common bands were isolated. A total of 672 positive clones were sequenced.

Assembly of sequenced AFLP bands resulted in 284 non-redundant sequences, and these sequences were used to conduct a GenBank database search using tBLASTx to assign their putative protein functions. Of the 284 sequences, 23 were highly homologous to the maize rust R-gene rp3-1 (E ≤ 10−13 to E ≤ 10−140) (Table 1), and 95 sequences contained putative domains similar to those in cloned R-genes, including 46 sequences containing NBS domains, 22 sequences harboring TM domains, and 27 sequences carrying PK domains. The remaining 166 sequences did not have any known domains whereby 132 sequences were highly homologous to maize genomic DNA and 34 sequences were highly homologous to genomic DNA of other plant species.

Table 1 Resistance gene analogs derived from both the modified AFLP and RACE methods
Full size table

The RGAs from the modified RACE method

The PCR products from modified RACE amplifications were subjected to electrophoresis on a 1% agarose gel. As a result, PCR products appeared as smears, and ranging in size from 0.5 to 2.0 kb (Fig. 1). Considering the distance between the P-loop motif (where the degenerate primer was designed on) and poly-A tails of R-gene transcripts, the 1.5–2.0 kb bands most likely contained PCR products amplified from R-genes (Fig. 1). Therefore, bands of 1.5–2.0 kb in size were excised from 1% agarose gel, and cloned into the pGEM-T vector. A total of 672 positive clones were subjected to sequencing.

Fig. 1
figure 1

The 3′-RACE products were subjected to electrophoresis on 1% agarose gel. M marker, DL2000; RP RACE products. The gel slice containing RACE products of 1.5–2.0 kb in sizes (as indicated in the bracket) was excised from the gel and cloned into pGEM-T vector

Full size image

Following assembly of these 672 positive clones, 246 non-redundant sequences were identified. All the unique sequences were used to search for their homologous sequences in GenBank using tBLASTx to identify their hypothetical functions. Among these 246 sequences, 12 were highly homologous (E ≤ 10−12 to E ≤ 10−240) to the maize rust R-gene rp3-1 (Table 1), 48 sequences contained putative NBS domains, 73 sequences contained TM domains, 25 sequences contained PK domains, and 88 sequences were maize mRNAs with unknown functions. Meanwhile, all 246 sequences were compared with the maize EST databases in both MaizeGDB and GenBank, and for every sequence, it’s corresponding high-homologous EST was found. These results indicated that available maize EST databases have good coverage of genome-wide expressed sequences.

The RGAs from the data mining method

Full-length sequences of the 48 known R-genes were used to perform tBLASTn searches against the maize EST databases in MaizeGDB and GenBank. As shown, the same EST could be hit repeatedly by using different R-genes (E ≤ 10−10). Accordingly, all redundant ESTs and their corresponding R-genes were removed, thus resulting in over 600 R-gene-like ESTs. The R-gene-like ESTs identified were further used to identify their corresponding unigenes in the MaizeGDB database. The resultant unigenes together with singletons were in turn compared with the GenBank database to confirm their putative R-gene-like functions. A total of 109 unigenes and 77 singletons were identified as having amino-acid sequences similar to those of R-genes. These 186 non-redundant unigenes/ESTs were cloned and sequenced from both parental lines ‘87-1’ and ‘Zong 3’. Among these 186 unigenes/ESTs, 57 contained putative NBS-LRR domains, 53 contained putative LRR-TM domains, 66 contained putative PK, PK/PK, and PK-LRR domains, and ten contained putative TM domains (Table 2).

Table 2 Resistance-gene-like unigenes/ESTs identified from the data-mining method
Full size table

Integration of RGAs derived from the three methods

All 23 AFLP-derived RGAs and 12 RACE-derived RGAs were used to perform BLASTn search to identify their corresponding homologous ESTs in the maize EST database in MaizeGDB. With an E value setting of E≤ 10−50, ten AFLP-derived RGAs and all 12 RACE-derived RGAs appeared to have corresponding R-gene-like ESTs previously recovered in the data-mining method. However, for the remaining 13 AFLP-derived RGAs, no corresponding ESTs were identified. Since the maize EST databases available at present have a good coverage of maize expressed genes, the R-gene-like ESTs identified from the data-mining method may cover the majority of R-genes present in the maize genome. The integration of RGAs recovered from the above three methods resulted in a total of 199 non-redundant RGAs available in the maize genome, of which at least 186 RGAs were derived from putative expressed R-genes.

Resistance gene analogs-tagged molecular markers

Sequence alignments between two parental lines ‘87-1’ and ‘Zong 3’ revealed frequent occurrences of small InDels and SNPs. Based on small InDels, 16 RGA-tagged STS markers have been successfully developed, and each STS marker gave rise to distinct and polymorphic PCR bands between two parental lines (Table 3). Likewise, 39 RGA-tagged CAPS markers have been obtained from those SNPs related to restriction sites, and each CAPS marker showed unambiguous polymorphic bands following digestion with their corresponding restriction enzyme (Table 4).

Table 3 Primer sequences and sizes of polymorphic bands for the RGA-tagged STS markers
Full size table
Table 4 Primer sequences and restriction enzymes for the RGA-tagged CAPS markers
Full size table

Discussion

In an attempt to efficiently isolate RGAs, a modified AFLP method was developed in the present study whereby the PstI restriction enzyme was used for digesting maize genomic DNA instead of EcoRI, and a degenerate primer based on the conservative P-loop was used for selective amplification. Of 284 unique sequences, only 23 were highly homologous to rp3-1. Although 95 other sequences contained R-gene-like domains, these were not confirmed to be derived from R-genes. In comparison with the commonly used PCR-based RGA isolation, this method was more flexible for isolating RGAs as only a single degenerate primer was needed. Nevertheless, there were some inherent drawbacks for this method. These included the following: (1) only a fraction of PCR products were R-related sequences, thus limiting applicability of this method for large-scale identification of RGAs, and (2) many R-genes might be missed if no PstI/MseI fragments could be generated from their NBS domains used for designing the degenerate primer.

Some inherent limitations are present in isolating RGAs from genomic DNA using degenerate primers. For plant species having small genome sizes such as Arabidopsis, the presence of RGAs lacking intron have been frequently observed (Pan et al. 2000), however, this is not often the case in most plant species of large genome sizes, such as barley (Leister et al. 1999). In addition, all family members in a resistance locus, regardless of the presence of either active genes or pseudogenes, are capable of generating RGAs. Conceivably, RGAs derived from non-expressed members will severely interfere with identification of active R-genes. A viable solution to these problems in RGA isolation can be achieved by replacing genomic DNA with cDNA. This switch of template DNA has been recently reported in wheat (Dilbirligi et al. 2004). In the present study, a modified RACE method using cDNA as template DNA has been set up in an attempt to uncover as many RGAs as possible from a single PCR reaction. Considering the distance between the C terminal poly-A and the P-loop motif in the NBS domain is about 2 kb in length for most R-genes (Mindrinos et al. 1994), RACE products of 1.5–2.0 kb in sizes are sliced from the agarose gel, and cloned into the pGEM-T vector. Sequence analysis has shown only 5% of cloned RACE products are highly homologous to the maize R gene rp3-1, and another 62% of RACE products contain domains similar to those in R-genes, although these could not be confirmed as R-genes derivatives.

To our surprise, both AFLP- and RACE-derived RGAs were highly homologous to a single rp3-1 gene. The rp3-1 locus was very complex and contained at least nine members (Webb et al. 2002). Sequence analysis indicated that this locus was a typical NBS-LRR R-gene. The complete sequence was deposited in MaizeGDB. As known, a majority of R-genes belong to the NBS-LRR class. In the present study, the degenerate primer was designed based on the P-loop in the NBS domain, and generated RGAs corresponding to the NBS-LRR class. This may explain the rapid identification of AFLP or RACE sequences homologous to rp3-1. There must be other RGAs corresponding to either other NBS-LRR R-genes or other resistance classes, however, their functions could not be confirmed by tBLASTx search due to lack of corresponding maize R-genes present in Genbank.

With a steady accumulation of ESTs, an efficient approach to mine expressed RGAs from EST databases has been reported in several plant species (Dilbirligi and Gill 2003; Rossi et al. 2003). In wheat, four data-mining strategies, including individual full-length search, domain search, consensus-sequence search, and multiple-motif search, were compared with each other in mining of RGAs from the wheat EST database, and the first strategy was proven to be the most successful (Dilbirligi and Gill 2003). Accordingly, in the present study, the individual full-length search using each of the 48 known R-genes was adopted to mine RGAs from the maize EST databases accessible in maizeGDB and NCBI websites. In addition, applying different cut-off values in BLAST searches would result in the recovery of different numbers of R-gene-like sequences. In sugarcane, adoption of a very stringent threshold value (E ≤ 10−50) resulted in identifying only 88 RGAs (Rossi et al. 2003). In contrast, large numbers of RGAs, including 243 NBS-LRR-type and 101 other type RGAs, were identified in wheat when applying a BLAST threshold value of E ≤ 10−1 (Dilbirligi and Gill 2003). In the present study, a moderate threshold value of E≤ 10−10 was adopted to perform data mining, and the putative RGAs were in turn compared with the database in GenBank to confirm their identities, leading to the identification of 186 R-gene-like unigenes/ESTs. Since most known R-genes were used in the data-mining method, RGAs belonging to different R-gene classes could be recovered from the EST databases. As reported previously, the 88 RGAs screened from the sugarcane EST database represented three major classes of R-genes carrying respective NBS-LRR, LRR-TM, and Serine/Threonine kinase domains (Rossi et al. 2003). The 184 putative expressed R-genes from wheat consisted of 87 NBS-LRR genes, 16 receptor-like kinases, 13 Pto-like kinases, and 68 genes from other types (Dilbirligi et al. 2004). The number of R-gene-like ESTs identified in the present study was very close compared to those found in wheat. Of 186 ESTs/unigenes, 57 contained putative NBS-LRR domains, 53 contained putative LRR-TM domains, 66 contained putative PK, PK/PK, and PK-LRR domains, and ten contained putative TM domains. In contrast, RGAs amplified from genomic DNA or cDNA using degenerate primers could only recover those R-genes on which the degenerate primer was designed. Generally, degenerate primers were designed based on the conservative NBS domain, and these primers could only generate RGAs corresponding to the NBS domain.

Among the three different ways adopted in this study, the data-mining method is found to be the most efficient to isolate RGAs. The success of this data-mining method depends largely on the coverage of the EST database, but only if the EST database covers all active genes in the genome, will the majority of R-gene-like ESTs be obtained from a particular plant species. In maize, more than 550,000 EST entries (up to August 26, 2005) have been deposited in GenBank and MaizeGDB. All 246 unique sequences from the modified RACE method have corresponding ESTs in the above two EST databases, indicating that all 550,000 ESTs covered almost all expressed genes in the maize genome. While comparing RGAs obtained from the three different methods, all RACE-derived and nearly half AFLP-derived RGAs overlap with R-gene-like ESTs obtained from the data-mining method. For those13 AFLP-derived RGAs without corresponding ESTs, it is likely that they are derived from inactive R-gene sequences.

At this time, only two R-genes have been cloned from maize via the transposon tagging method, Hm1 encoding a toxin reductase that detoxifies the fungal toxin of Cochliobolus carbonum race 1 (Johal and Briggs 1992) and Rp1-D confering resistance to rust in maize (Collins et al. 1999). Because of the presence of highly-repetitive sequences and lack of whole genome sequences for maize, it is very difficult to clone R-genes via a map-based cloning strategy. Considering the distinct R-gene structure and advances made in the maize genome, the candidate R gene approach via RGA is a very promising strategy for cloning R-genes in maize. This approach has been useful in cloning of the R-gene Lr10 in wheat (Feuillet et al. 1997). We are attempting to use this candidate gene approach in cloning of the Scmv1 gene conferring resistance to the SCMV in maize. Among 20 RGAs discovered by Collins et al. (1998), one RGA, pic19, has been mapped to the region whereby one of two SCMV R-genes, Scmv1, resides (Xu et al. 1999). Three pic19 homologous sequences have been identified while screening of the B73 BAC library using pic19 (Quint et al. 2003). Based on conserved regions among pic19 and its three homologous sequences, full-size cDNAs have been obtained using RACE, and these cDNAs have been cloned downstream of the 35S promoter for complementary tests (unpublished results).

To date, only a small fraction of maize RGAs have been cloned and mapped onto the maize genome. However, it is impossible right now to clone other maize genes via the candidate gene approach. This is why we have pursued efforts to isolate genome-wide RGAs and developed RGA-tagged markers. In the next step, an RGA linkage map will be constructed based on RGA-tagged markers and a RIL mapping population. This RGA linkage map will be very useful in identifying a candidate R-gene(s) for a given disease if a certain RGA is co-located with a major resistance locus or QTL.