Introduction

The Leguminosae family constitutes about 650 genera and 18,000 species and ranks third among families of flowering plants. Legumes are unique because of their capability of fixing atmospheric nitrogen in soil through symbiosis. Among legumes there is extensive literature on genetic studies for several species, but the existence of a wide range of genomic tools and genome information is limited only to Medicago truncatula (Mt), Lotus japonicus (Lj) and Glycine max.

Chickpea (Cicer arietinum L.) is an economically important crop with a life cycle of 3–4 months, and a genome size of 740 Mbp that is only 1.5 times larger than M. truncatula. Although genomic research in chickpea is feasible, it was not until the beginning of twenty-first century that this aspect of research began in this cool season legume crop. Since then, several genomic tools that include bacterial artificial chromosome (BAC) libraries (Rajesh et al. 2004; Lichtenzveig et al. 2005), expressed sequences tags (EST) (Buhariwalla et al. 2005; Corum and Pang 2005), Targeted Induced Local Lesions IN the Genome (TILLING) mutants (Rajesh P. N. et al. unpublished data) and gene technology (Sarmah et al. 2004; Sanyal et al. 2005) have been developed in chickpea, and made available to the research community. Utilization of these tools for crop improvement should be the next focus as it will provide precedence for effective application of genomics for crop improvement of crop legumes. With the availability of suitable genomic tools, genetic materials and also the shorter life cycle with relatively smaller genome than other crop legumes, chickpea can be considered as a model crop legume.

Current knowledge of the chickpea genome is based mostly on seven linkage maps that are comprised the less informative RAPD, ISSR, AFLP and SSR markers. No gene specific markers have been mapped to the genome and thus these maps have limited use in comparative genomic studies.

Genome organization and composition analysis between other model legumes, such as Mt and Lj, has revealed several interesting features of their genomes. For example, 18,844 and 20,800 genes (excluding transposons) were predicted to be present in Mt and Lj, respectively. Also, the average gene density was estimated to be 12.6 genes per 100 kb in Mt and 17.4 genes per 100 kb for Lj. In addition, 38 and 19% of the Mt and Lj genomes, respectively, were predicted to contain repetitive sequences (Cannon et al. 2006). Availability of genome sequences in these legumes (149 and 121 Mbp in Mt and Lj, respectively) facilitated the analysis of synteny and colinearity among them, which will have long-term application in transferring genetic information between different species. Phylogenetically, Mt and Lj belong to galegoid phylum and were separated by 37–38 million years ago (MYA) (Choi et al. 2004). Microsynteny analysis between Mt and Lj genome sequences discovered colinearity at different synteny blocks (Cannon et al. 2006).

Since chickpea is closer phylogenetically to Mt than Lj is to Mt, and both belong to the galegoid phylum, microsynteny analysis between these two legumes will be useful for transferability of genetic information between them. The lack of genomic sequences has limited the information on genome organization and microsynteny in chickpea. Also, the absence of common markers, or cross-species specific markers, in genetic linkage maps has restricted the macrosynteny establishment between chickpea and model legumes.

Ascochyta blight [caused by Ascochyta rabiei (Pass). Lab.] is a devastating disease worldwide and has the potential to cause up to 100% yield loss to chickpea. Earlier genetic studies identified two QTL (ABR-QTL1 and ABR-QTL2) that confer resistance to this disease (Santra et al. 2000). To dissect ascochyta blight resistance QTL and narrow the genetic distance between the flanking markers, it is necessary to increase the marker density at this genomic region.

Therefore, in this study, we sequenced 471,638 bp from 11 BAC clones that include three clones representing ABR-QTL1 and analyzed genome organization and composition in this economically important crop legume. Microsynteny analysis between chickpea with M. truncatula was also performed. In addition to genome organization and microsynteny analysis, we characterized ABR-QTL1 genomic region by increasing marker density developed from the ends of BAC clones. Our results provide new insight into the chickpea genome, and the sequences representing different genomic regions will be a potential resource for the development of additional markers to increase marker density in important regions of the chickpea genetic map.

Materials and methods

BAC selection and sequencing

The BAC clones used in this study (Table 1) can be classified into the following categories: (i) targeted selection and (ii) random selection. BAC clones in the first category represent QTL1 accounting for ascochyta blight resistance. BAC clones representing the latter category were selected randomly.

Table 1 Predicted gene content in chickpea BAC clones. Values in the parentheses are without retroelements
Full size table

The detailed procedures for cloned, large insert genomic DNA isolation, random shot-gun cloning, fluorescent-based DNA sequencing and subsequent analysis were used as described earlier (Bodenteich et al. 1993; Pan et al. 1994; Chissoe et al. 1995; Roe et al. 1996; Roe 2004). Fifty microgram portions of purified BAC DNA was randomly sheared and made blunt-ended, and after kinase treatment and gel purification, fragments in the 2–4 kb range were ligated into SmaI-cut, bacterial alkaline phosphatase (BAP)-treated pUC18 (Pharmacia) and transformed by electroporation into Escherichia coli, strain XL1BlueMRF’ (Stratagene) competent cells. A random library of approximately 1,200 colonies was picked from each transformation, grown in terrific broth (TB) medium (Roe 2004) supplemented with 100 μg of ampicillin for 14 h at 37°C with shaking at 250 rpm in a HiGro incubator (Genomic Solutions), and the sequencing templates were isolated by a semi-automated cleared lysate-based procedure on a Capilar Life Sciences Sciclone ALH 3000 workstation equipped with a Twister II robotic arm and four microtiter plate shakers.

Sequencing reactions were performed as previously described (Chissoe et al. 1995; Roe et al. 1996; Roe 2004) using the Amersham ET Terminator kit (US-81070) or Applied Biosystem BigDyes version 3.1 (4336921) sequencing reaction mixes diluted 1:32 with TM buffer (20 ml Tris-HCl pH9.0, 0.5 ml 1 M MgCl2 and 29.5 ml ddH2O). The reactions were incubated for 60 cycles in a Perkin–Elmer Cetus DNA Thermocycler 9600 under the cycle conditions recommended by the manufacturer. Any unincorporated dye terminators were removed by ethanol precipitation at room temperature, and after dissolving the fluorescent-labeled nested fragment sets in 0.1 mM EDTA pH 7.4, the nested fragment sets were resolved by electrophoresis on an ABI 3730 Capillary DNA Sequencer. After base calling with the ABI Analysis software, the analyzed data were transferred to a Sun Workstation Cluster, and assembled using Phred and Phrap (Ewing and Green 1998; Ewing et al. 1998). Overlapping sequences and contigs were analyzed using Consed (Gordon et al. 1998). Gap closure and proofreading were performed either using custom primer walking or using PCR amplification of the region corresponding to the gap in the sequence followed by sequencing directly using the amplification or nested primers, or by sub-cloning into pUC18 and cycle sequencing with the universal pUC-primers (Roe 2004). In some instances, additional synthetic custom primers and PCRs with 7deaza-dGTP replacing dGTP (Roe 2004) or by rolling circle amplification (Detter et al. 2002) were necessary to obtain at least threefold coverage for each base. The sequenced BACs along with their GenBank accession numbers are listed in Table 1 and the genome browser-based annotation (Gbrowse) is available at http://genome.ou.edu/plants_totals.html.

Sequence annotation

The chickpea BACs sequenced were annotated using Repeatmasker (Jurka 2000), Genscan (Burge and Karlin 1997) and FgenesH (Salamov and Solovyev 2000), and gene predictions were compared to the Arabidopsis genome and the Plant unigene database (Wheeler et al. 2005). Results from this are displayed for each BAC on the Advanced Center for Genome Research’s web site (http://www.genome.ou.edu) using Gbrowse (Stein et al. 2002).

Further analysis was performed to determine genomic organization of the BACs sequenced. Genic information was determined using the FgenesH dicot model, as FgenesH has been shown to be the most accurate gene modeling program for plants (Yu et al. 2002). Identification of these predicted genes were performed via a basic local alignment search tool (BLAST) (Zhang and Madden 1997) against the NCBI databases. Genes that have no significant homology (an e value of less than e−10) were designated as “unknown,” if they have EST support, “hypothetical” if the gene is conserved in multiple organisms and “putative” if there is no support other than prediction by a gene model. Other genomic organization information, such as GC content and repetitive sequences present (simple repeats, retroviral insertions, and transposons) were determined using Repeatmasker (Jurka 2000), against the Arabidopsis matrix. The values of calculated parameters are in Table 1 and the predicted genes in Table 2.

Table 2 Estimated values for various genomic parameters derived from BAC sequences
Full size table

Plant materials

Chickpea accession FLIP84-92C and wild relative PI599072 that are resistant and susceptible to Ascochyta rabiei (Pass). Lab. respectively were used in this study. The F2 population from this cross was advanced by single seed descent to the F8 to produce F7-derived RILs (Santra et al. 2000; Tekeoglu et al. 2002). A population of 142 F7:8 derived recombinant inbred lines (RILs) from the FLIP 84-92C (resistant) x ×C. reticulatum PI 599072 (susceptible) cross was used for genetic mapping. DNA was extracted from leaf tissue of each RIL and the parental lines according to Doyle and Doyle (1987).

Linkage analysis

Segregation of marker loci was tested for goodness of fit to the expected Mendelian ratio of 1:1 using Chi-square analysis (P < 0.05). Markers with distorted distribution were also used for linkage analysis. Linkage analysis was performed using Mapmaker/Exp 3.0 (Lander et al. 1987). Linkage groups were established at a constant LOD score of 4.0 and a recombination value of 0.25 by two point analyses using the ‘group’ command. The most possible order of loci within a group was determined using multipoint ‘compare’ command, and these orders were verified using the ‘ripple’ command. The Kosambi mapping function was used to determine cM distances between markers (Kosambi 1944). Double crossovers were checked by ‘double crossover’ command in Map Manager QTb (version 2.8) (Manly 1998). QTL analysis for blight resistance was carried out with the simple interval mapping function using Qgene (Nelson 1997) at an LOD score of 3.0. Single-point regression analysis was used to identify markers significantly associated with blight resistance.

Results

BAC selection and genetic analysis

The BAC clones AC161102, AC145766 and AC145456 were selected by targeted screening of the BAC library. The BAC clone AC161102 was identified by screening our BAC library with a single locus SCAR733b marker derived from a RAPD marker UBC733b. AC145766 and AC145456 clones were identified with a single locus marker OPS06-01 (Rakshit et al. 2003). UBC733b and OPS06-01 were genetically proved to be flanking ABR-QTL1 in earlier studies (Santra et al. 2000; Tekeoglu et al. 2002; Rakshit et al. 2003). In this study, we generated six markers from the ends of identified BAC clones and a gene (Flavonoid 3-0-galactosyl transferase) present in AC145456. Since these PCR-based markers are monomorphic in amplicon length between the parental lines (FLIP84-92C and PI599072), we developed cleaved amplified polymorphic site (CAPS) and derived CAPS (dCAPS) markers by utilizing the SNPs to increase marker density at QTL1 (Rajesh et al. 2005).

Genome organization

FgenesH program predicted 51 genes in 11 BAC clones representing 471,638 bp of chickpea genome (Table 1). Of these 51 predicted genes, 20 were with known function, 18 were putative genes and the remaining 13 were retro elements. Of 51 predicted genes, 12 genes identified ESTs with other plants at http://www.ncbi.nlm.nih.gov database. The average gene length of 2.5 kb, protein length of 452aa, 8.68% of repetitive elements and the one long terminal repeat (LTR) for every 13.5 kb were estimated (Table 2). The gene density in chickpea is estimated to be ranging from 1 gene for 5.3 kb to 1 gene per 63.3 kb and averaged 1 gene for every 9 kb (11 genes per 100 kb) which is similar to Mt in which 12.6 genes per 100 kb were found and relatively higher than Lj in which 17.4 genes per 100 kb were found (Cannon et al. 2006). No gene duplication was detected among these chickpea BAC clones. Interestingly, the overall percentage of repetitive elements such as simple and low complexity repeats is small (8.68%) in chickpea genome ranging from 2.12 to 20.31%. In our study, we calculated that 25% of the genes were retro elements in chickpea which are commonly considered to be abundant in the genome and also considered to play a role in genome expansion (Neumann et al. 2006). Among the 11 BAC clones, unequal distribution of gene density was detected from a relatively high density of 1 gene per 5.3 kb (AC145766) to a relatively low density of 1 gene in 63.5 kb (AC145459).

We characterize these BAC clones into two main categories based on selection.

Targeted selection: BAC clones representing ABR-QTL1 region

Three BAC clones (AC161102, AC145766 and AC145456) represent a genomic region where ABR-QTL1 was mapped previously (Tekeoglu et al. 2002). Interestingly, these BAC clones of 80 kb consist of only 6.55% of repetitive sequences, a gene density of 1 gene per 5.7 kb and 14 genes. Of all the genes present in this agronomically important genomic region, only one retrotransposon was detected.

Random selection

The remaining randomly selected BAC clones were further sub-divided into three types based on the presence of retroelements.

BAC clones containing high retrotransposons

In AC161103 (74.4 kb) and AC161104 (56.6 kb), gene islands where genes surrounded by retro elements were observed. This 130,965 bp revealed gene density averaging 1 gene per 6.8 kb which is slightly lower than ABR-QTL1 genomic region and higher percentage level of short tandem repetitive elements (16%).

BAC clones containing low retrotransposons

Three BAC clones (AC145458, AC161105 and AC145459) consist of relatively low amount of retrotransposons with four retroelements of seven genes predicted. The average gene density of one gene for every 11.1 kb which is lower than previous two categories was observed in this 123,087 bp. Also, the moderate percentage level of short tandem repetitive elements (11.2%) was detected.

BAC clones containing no retrotransposons

AC161101 (68.3 kb), AC145457 (23.3 kb), and AC145781 (45.8 kb) are the BAC clones that contained no retrotransposons and had a very low level of short tandem repetitive elements as well (5.4%). These four BAC clones contained an average low gene density of 1 gene per 31.4 kb ranging from 1 gene per 5.8 kb to 1 gene per 65.8 kb.

Microsynteny

Blast analyses of gene sequences from each chickpea BAC clone with Mt genome sequences and expressed sequence tags (ESTs) were performed for microsynteny studies at http://www.medicago.org and http://www.tigr.org (currently http://www.jcvi.org), respectively. Significant similarities of the genes, representing ABR-QTL1 genomic region, such as protein phosphatase, 40S ribosomal protein, flavanoid galactosyl transferase and eukaryotic transcription factor with e-values ranging from 2e-28 to 1e-134, were observed (Fig. 1a). The physically linked genes in AC145456 and AC145766 represent one chromosome in chickpea while the genes in these BAC clones represented different chromosomes in Mt on blastn analysis of genes from ABR-QTL1 (Fig. 1a). Nevertheless, three of four genes from AC161102 displayed a high degree of microsynteny between chickpea and M. truncatula. Although the genes structures were different, more than 85% homology was detected among these three genes (Fig. 1b).

Fig. 1
figure 1

a Blastn analysis of the genes present in BAC contigs (AC161102, AC145456 and AC145766) representing ABR-QTL1 region with Mt genome sequences at http://www.medicago.org/genome. b Microsynteny analysis of chickpea BAC clone AC161102 and Mt BAC clone AC146861

Full size image

No significant similarities were detected for genes from the remaining eight BAC clones (AC161103, AC161104, AC145457, AC161101, AC145781, AC145458, AC161105 and AC145459). Twenty-one genes out of 51 predicted genes in chickpea did not identify homologs in genome sequences and ESTs of Mt.

Discussion

Genome organization and composition

Sequencing of 471,638 bp from 11 BAC clones provides significant new insights into chickpea genomic organization. Considering an average gene density of one gene per 9.2 kb and the genome size of 740 Mb, if the genes were evenly distributed throughout the genome, 80,000 genes could be present in chickpea genome. However, since the size of the heterochromatic region in the chickpea genome is unknown, it is clear that the actual number of genes will be much less than this number. Considering the average length of a gene and estimated number of genes (80,000) if they are evenly distributed, gene-rich region is predicted to be occupying 27% of chickpea genome. On the basis of these BAC sequences, the average exon per gene in chickpea is 4.7, which is close to Lj (4.8) and Arabidopsis (5.2) but higher than Mt (3.1). The average predicted exon size (401 bp) in these chickpea BACS is significantly larger than Mt (318 bp), Lj (296 bp) and Arabidopsis (268 bp) and the average intron size of 316 bp, is close to Mt (364 bp), Lj (378 bp) and significantly larger than Arabidopsis (165 bp). Taken together with other parameters such as gene length, gene density, GC content and protein length, chickpea, Mt, Lj and Arabidopsis, all closely resemble one another with the exceptions of gene density and intron length values of Arabidopsis (Table 3). The smaller intron length in Arabidopsis is due to its small genome size, as species with smaller genomes tend to have smaller introns (Dubcovsky et al. 2001). Although the amount of sequence analyzed to predict the genome organization of chickpea (~500 kb) is less than Mt (149 Mbp) and Lj (121 Mbp), the estimated values of the three different, legume genomes appear not to be significantly different from one another (Table 3). This would indicate that the extrapolation of the results obtained from the representative 500 kb sequences, especially the estimates of gene-rich region to the whole chickpea genome, is probably correct. As a first step, gene-rich regions need to be identified in chickpea genome by genetically mapping expressed gene tags (ESTs) and/or by integrating genetic and physical maps.

Table 3 Comparison of chickpea genomic parameters with Medicago truncatula, Lotus japonicus and Arabidopsis thaliana
Full size table

Genetic analysis, genome organization and composition of ABR-QTL1 region

Our efforts to clone ABR-QTL1 using map-based cloning approach have unraveled the architecture of this agronomically important genomic region. Our genetic analysis using six markers derived from AC161102, AC145766 and AC145456 increased the marker density at QTL1. Marker-trait association using QGene software program discovered 20(T)l12-Right, a CAPS marker derived from the end of AC161102 that accounted 56% of the variation in ascochyta blight resistance at LOD value 19.98, an improvement of the previously reported (35% and LOD 13.40) (Rajesh and Muehlbauer 2008). Low repetitive sequences, high gene density and 14 candidate genes are the characteristics of ABR- QTL1.

Another interesting observation was that none of the genes at this resistant locus resemble NBS-LRR type genes which may hypothesize that the ascochyta blight resistance in chickpea involves a different mechanism other than classical R genes where NBS-LRR motifs are present. This could be due to the fact that Ascochyta rabiei is a necrotrophic fungus and resistance to necrotrophic fungus is poorly understood unlike resistance to biotrophic fungi. However, several candidate genes were predicted at ABR-QTL1 such as Flavonoid 3-0-galactosyl transferase, protein phosphatase and putative eukaryotic transcription initiation factor (TIF6) and are interesting candidate genes for ascochyta blight resistance. The former two genes (Flavonoid 3-0-galactosyl transferase and protein phosphatase) identified ESTs from the database and there were no corresponding ESTs for putative eukaryotic transcription initiation factor (TIF6). There are reports available on possible involvement of flavonoids in disease resistance in other plant systems (Jeandet et al. 2002; Yu et al. 2003); however, further investigation is required to correlate the function of these genes with blight resistance using suitable reverse genetic approaches or genetic transformation.

Microsynteny

The genes in Mt were annotated by Eugene gene prediction program which uses a combined ab initio and homology-based approach while FgenesH program was used in chickpea. Comparison of the Gbrowse displays of Mt at http://www.genome.ou.edu and http://www.tigr.org (currently http://www.jcvi.org) that used Eugene and the final annotation to the raw FgenesH predictions have found that the differences were at either end in the 5′ and/or 3′ non-translated regions that did not effect the gene call.

Choi et al. (2004) estimated the macrosynteny among legumes by genetically mapping orthologous genes and discovered broad genome conservation among them. However, genes that are genetically close to one another might physically be far away from one another (Goff et al. 2002). Although macrosynteny identifies the orthologous genomic location for the genes and determines the gene order along the chromosomes, microsynteny based on genomic sequences unravels the actual evolutionary relationship among the plant species.

A high level of microsynteny was reported among Mt and Lj by comparing the genome sequences of 149 and 121 Mbp, respectively (Cannon et al. 2006). In their study, they estimated a minimum of ten large-scale synteny blocks with substantial colinearity between these two genomes. Phylogenetically, the tribe Cicereae, where chickpea belongs, falls within Trifolieae, which includes both Medicago and Loteae. Hence, we expected microsynteny between chickpea and Mt as found between Mt and Lj.

Surprisingly, our comparison of 11 chickpea BAC clones reveals that Mt and chickpea display limited synteny in this 471,638 kb of genomic sequences. This lack of synteny might be due to one or more of the following reasons: (1) Sampling bias—the representative BAC clones selected are not orthologous between chickpea and M. truncatula. (2) Chickpea BAC selection—8 of 11 chickpea BAC clones are short genomic sequences representing different genomic regions and are not actually contiguous. (3) Incompleteness of the Mt genome—the gene-rich regions of the Mt genome are not sequenced completely and therefore extensive microsynteny was not found between chickpea and M. truncatula. (4) Genomic differences—although, time wise, Lj diverged earlier than chickpea, it is possible that chickpea has undergone evolutionary processes that have caused divergence since its split with Mt, which has resulted in less microsynteny, than Mt has with Lj. Perhaps, by specifically choosing orthologous regions, as well as longer contigs, microsynteny between chickpea and M. truncatula can be investigated in more detail.

In contrast to clear absence of microsynteny with most of the BAC sequences between chickpea and Mt, we observed one syntenous block between chickpea BAC clone AC161102 and Mt BAC clone AC146861. If the orientation of the sequences AC161102 of chickpea is reversed, the gene direction and order will be identical in both chickpea and Mt (Fig. 1b). All four predicted genes in AC161102 of chickpea identified ESTs of Mt from the database at http://www.ncbi.nlm.nih.gov which indicates that these genes are present in the genome. We identified homologs to three of four genes (AC161102.2.fg.1, AC161102.2.fg.2 and AC161102.2.fg.4) from the 23 kb BAC clone, AC161102, in the 125.3 kb Mt BAC clone, AC146861 (Fig. 2). Although these three chickpea genes showed between 88 and 92% homology with their Mt counterparts, they differed in structure with respect to the number of exons in the protein phosphatase gene and the position of exons and introns in the unknown and hypothetical protein genes (Fig. 2). However, it should be noted that these discrepancies could be due to the gene prediction software, and not due to actual differences in intron/exon borders. As the matrix was not specifically designed for chickpea, it could be splitting the genes incorrectly.

Fig. 2
figure 2

Comparison of gene structure between chickpea BAC clone AC161102 and Mt BAC clone AC146861 that exhibited microsynteny

Full size image

One of the four predicted chickpea genes (AC161102.2.fg.3) failed to identify its homolog in Mt and similarly the predicted gene AC146861.24.fg.8 in Mt did not have its homolog in chickpea. Since the sequences that are flanking AC161102 are not available in chickpea, and only “one gene loss and one gene gain” was observed between chickpea and Mt, it is impossible to infer if they are the result of genomic rearrangements. However, by genetically mapping the orthologous genes that are predicted in the flanking Mt BAC sequences (AC146861) in chickpea, microsynteny can be analyzed in depth between chickpea and Mt. Although sequence analysis using 10 of 11 BAC clones did not exhibit microsynteny between chickpea and Mt, the chickpea BAC clone AC161102 showed that comparative genomics between chickpea and Mt should not be completely ruled out.

In conclusion, our analysis of approximately 500 kb of genomic sequence from 11 BAC clones revealed several interesting features of the chickpea genome architecture. Sequencing of BAC clones representing ABR-QTL1 identified candidate genes for future experiments to elucidate the genes that control this agronomically important trait. Also, microsynteny was established between chickpea and Mt at ABR-QTL1 genomic region, suggesting that the genomic data from Mt may be a useful source for identification of additional candidate genes in chickpea. The new markers flanking ABR-QTL1 with increased contribution to resistance can be exploited in marker-assisted selection. Having discovered information on genome composition and organization, our representative work will have significant impact on chickpea genome sequencing efforts in the future.