Validation of Molecular Markers Linked to Fusarium Wilt Resistance in Recombinant Inbred Lines of Chickpea (Cicer arietinum L.)

Molecular markers closely linked to Fusarium wilt resistance in chickpea were identied from earlier studies with an objective to validate them. The recombinant inbred line (RIL) mapping population (F 11 ) derived from JG62 (FW susceptible) and WR315 (FW resistant) was used for validation of linked markers. The RILs were phenotyped for FW resistance over two seasons in a wilt sick eld at ICRISAT and in wilt sick pots under the green house condition. A total of 42 markers linked to Fusarium wilt (FW) resistance in chickpea were selected from earlier reports for validation. Among 42 markers selected 23 markers were polymorphic among the parental lines of RIL population and hence these markers were used for genotyping the mapping population. The genotyping data together with phenotyping data over two seasons and wilt sick pots identied 12 markers linked to wilt resistance by single marker analysis. The markers TA96, CS27, TA110 and TA59 were more consistently related to FW resistance in this and previous studies. The composite interval mapping has identied two major loci; one each in LG1 and LG2 anked by markers H4G11 and CaM1402 and H3A12 and CS27, respectively. These two loci are considered as major loci for FW resistance explaining up to 76.66% of phenotypic variation for FW resistance. The results presented here and those published earlier conrm the emerging picture of two hotspots in chickpea genome for resistance to Fusarium wilt. The QTL anking markers, H4G11, CaM1402, CS27, and H3A12 are most promising and reliable for deployment in chickpea breeding.


Introduction
Chickpea (Cicer arietinum L.), is the second most important cool season legume crop of the world, belongs to the family Fabaceae and sub family Faboideae. It is a self-pollinated and diploid (2n = 2x = 16) pulse with a genome size of 740 Mb (Jingade and Ravikumar 2015). Chickpea seeds are important source of protein, vitamins and essential minerals for human diet and the crop enhances soil fertility by xing atmospheric nitrogen through symbiotic association. The cultivated chickpea is often divided into two types: Desi and Kabuli. Desi chickpea (purple ower, small, dark and angular seeds) are mainly grown in the Indian subcontinent, East Africa, Central Asia and to a limited extent in Mediterranean basin whereas; Kabuli chickpea (white ower, large and cream-coloured seeds) have traditionally been grown in the Mediterranean basin and Central Asia (Karami et al. 2015).
One of the major constraint in chickpea production is the Fusarium wilt (FW) caused by the deuteromycetes fungal pathogen Fusarium oxysporum f. sp. ciceri (Foc). The disease is wide spread in chickpea growing areas of the world (Dubey et al. 2007). In India the annual yield loss to wilt disease varies from 10-15% under normal conditions (Jalali and Chand 1992) to nearly 100% when conditions are favourable for pathogen development (Jimenez-Diaz et al. 1993). Eight different physiological races of Foc are characterised based on disease severity symptoms on different chickpea cultivars. Races 0 and 1B/C cause a yellowing symptoms and races 1A, 2, 3, 4, 5 and 6 induce a wilting symptom (Sharma et al. 2005;Jimenez-Fernandez et al. 2013). Four out of these eight reported races have signi cant presence in India race 1 (synonymous 1A, Indian isolate) in Peninsular India, races 2 and 3 in North India, and race 4 in Central India (Singh et al. 2008;Pande et al. 2007). The pathogen is a soil-borne therefore, the agronomic and plant protection measures are not very effective in controlling the disease incidences.
Rather cultivation of varieties possessing resistance to speci c races of the pathogen prevalent in a region is the most economical management strategy for this disease (Arvayo-ortiz et al. 2012). Even though high level of resistance to Foc race 1A is available in the cultivated species, the progress in the introgression of resistance to elite varieties is very slow owing to the di culty in the identi cation and selection of wilt resistant genotypes in the breeding populations. To overcome such problems, molecular breeding strategies have been successfully deployed in several crop species. Therefore, the QTL based marker-assisted selection (MAS) appears one of the e cient strategies to accelerate pyramiding of different resistance genes for FW resistance in chickpea.
Furthermore, the identi ed QTLs were not validated in different seasons and genetic backgrounds potentially limiting their utility as most of the markers are not functional in different genetic backgrounds.
Another factor confounding the FW resistance is temporal variation in response to FW resistance among chickpea genotypes. Some susceptible genotypes exhibit late wilting against different Foc races (Upadhyaya et al. 1983a, b;Singh et al. 1987). For example, based on the genetics of resistance to FW race 1, even though the resistance was governed by 2-3 major genes, any one of these genes alone can cause late wilting (Sharma et al. 2005;Brinda and Ravikumar 2005). A combination of two genes in recessive form confers complete resistance, while one of these genes in recessive form produce late wilting (Castro et al. 2010). Despite the importance of early and late wilting responses among chickpea genotypes, so far only limited attempts have been made to identify QTLs in uencing the two distinct resistance types (Jingade and Ravikumar 2015). Towards this direction, it is important to further con rm the linked markers in different genetic background to prove their e ciency to characterize susceptible and resistance chickpea genotypes. Thus, the present study was undertaken with an objective to validate previous reported QTLs in a set of recombinant inbred lines (RILs) developed from intraspeci c cross between highly susceptible JG62 and resistant WR315, segregating for resistance to Foc race1 genotypes. This information will aid chickpea breeders in selecting markers for use in MAS and gene pyramiding to enhance the wilt resistance.

Plant materials
One hundred and twenty ve F 10 derived F 11 RILs from intraspeci c cross between JG62 and WR315 were used in study. The female parent JG62 is an early wilting genotype highly susceptible to Foc race 1, while WR315 is resistant to race 1A, race 2, race 3, race 4 and race 5 of FW (Mayer et al. 1997;Sharma et al., 2005). The mapping population (RILs) was developed by single seed descent method.

Phenotyping
The 125  Screening RILs for FW resistance in wilt sick pot The wilt sick pots were developed according to Brinda and Ravikumar (2005) using Foc race 1 inoculum obtained from ICRISAT, Hyderabad. The presence of pathogen in the pots was con rmed by growing the susceptible JG62 plants (6 plants per pot) and observing wilting symptoms and death before 30 days.
Only the pots in which all the 6 plants wilted within 30 days were used for evaluation of RILs (F 11 ). Six seeds of each RIL were sown in a single pot along with one seed of JG62 (susceptible check in the centre). The pots were regularly watered and the number of plants showing wilt symptoms followed by death in each pot was recorded on 30 th and 60 th day after sowing (Barman et al. 2014). All the experimental pots the susceptible check JG62 showed wilting within 30 days ( Figure 1). For each RIL, the percent wilt incidence was measured on the basis of number of wilted plants to the total number of plants in each pot as given below-Screening for FW resistance in the eld The F 7 and F 8 RILs were also screened for FW resistance in the wilt sick eld at ICRISAT, Patancheru, India over two years (Season I: 2007-08 and Season II: 2008-09) in two replications using randomised complete block design. Each RIL was sown in single rows of two-meter length with two replica tions during 2007-2008 along with parents were repeated after every twenty rows in each replication.
Observations for disease incidence were recorded on 30 th and 60 th day after sowing (DAS) by observing individual plants for wilting (Haware et al. 1992). The wilt incidence was measured on the basis number of plants wilted and total number of plants in each row as mentioned above.

DNA extraction and PCR ampli cation
The genomic DNA was extracted from vegetative buds and young leaves of RILs and parental lines by CTAB method of DNA extraction with slight modi cations (Doyle and Doyle 1990). The isolated DNA was quanti ed by nanodrop (Eppendorf Bio-spectrometer). Forty two markers linked to FW resistance previously (Mayer et al. 1997;Winter et al. 2000;Sharma et al. 2004;Cobos et al. 2005;Radhika et al, 2007;Gowda et al. 2009;Millan et al. 2010;Soregaon 2011;Sabbavarapu et al. 2013;Barman et al. 2014;Jingade and Ravikumar 2015) were selected for the present study. The parental lines were screened for polymorphism and 23 markers were found to be polymorphic which were used for genotyping of all the 125 RILs. . The PCR ampli cation for SSR primers was performed according to Mayer et al. (1997) with certain modi cations.
PCR products along with 5 μl of loading dye were electrophoresed on either 3.0% agarose or 7.5% denaturing polyacrylamide gel (PAGE) in Tris-borate EDTA (TBE) running buffer. Out of 23 markers, 20 markers were resolved on 3% agarose gel and three markers on 7.5% PAGE. The resolved PCR products were detected by ethidium bromide staining in agarose gels and by silver staining in PAGE. Gel scoring was done by taking in to considerations banding pattern of individual RILs, and comparing with banding pattern of parental genotypes. RILs were scored as "2" for parent JG62, "0" corresponding to parent WR 315 and "-1" for missing band.

Statistical analysis
The 23 polymorphic markers were used for genotyping the RILs and the segregation of individual markers for goodness-of-t to the expected Mendelian 1:1 ratio was tested by Chi-square test.

Linkage map construction
The linkage analysis was performed using ICIM QTL Ici Mapping version 4.00 software (http://www.isbreeding.net). A minimum threshold log likelihood ratio (LOD) scores 3.0 and maximum recombination fraction of 0.3 was set as threshold value for linkage groups (LGs) determination. 'Ripple' command was used after adding each marker locus to con rm marker order. Recombination frequencies were converted into genetic distances in centimorgan (cM) using Kosambi mapping function (Kosambi 1943). Further the marker order and the linkage distances obtained were used for QTL analysis.

QTL analysis
The genotypic data combined with phenotypic data obtained from wilt sick pot and eld experiments on 125 RILs were analysed for identi cation of the QTLs using ICIM. The positions and effects of QTLs for wilt resistance in each of the stage as well as two seasons eld and pot data were determined following single-marker analysis (SMA) and composite interval mapping (CIM) (Jansen and Stam 1994). The proportion of observed phenotypic variation explained (PVE) due to a particular QTL was estimated by the coe cient of determination from the corresponding linear model (single marker) analysis and using maximum likelihood for CIM (Basten et al. 1997). A minimum LOD score of 2.5 for SMA and 3.0 for CIM was used for declaring the presence of a putative QTL.

Wilt reaction for RILs
Signi cant differences among RILs for FW were observed on 30 th and 60 th days in both eld and pot testing. The susceptible JG62 recorded complete wilting in 30 days in both eld and pot screenings.
Likewise, in pots the highly susceptible RILs showed complete wilting symptoms within 30 days, while the resistant RILs did not show wilting even after pod lling stage beyond 30 days (Figure 1). The mean wilting per cent in both eld and pot screenings ranged from 0.00 -100.00% at both times of scoring ( Figure 2). The mean wilt percentage was signi cantly lower on 30 th day compared to 60 th day in both 2007 and 2008 seasons suggesting that the RILs segregate for early and late wilting and resistance phenotypes. Furthermore, plants in the eld exhibited relatively less wilt damage compared to those tested in pots. There was a signi cant correlation in the reaction of RILs to wilt in two seasons under eld conditions (Table 1). The association between eld and pot screening was also positive and signi cant.

Development of linkage maps
Forty two markers linked to wilt resistance towards Foc races in different mapping populations were selected for study. Among them twenty three (54.76%) markers were polymorphic in parents, JG62 and WR315, while the remaining 19 markers were not polymorphic, hence were not used in study. The mapping population consisting of 125 RILs were screened with the 23 polymorphic markers. The segregation of individual markers for the expected monogenic 1:1 ratio in the RILs was tested using χ² test. Twenty markers did not show segregation distortion from the expected ratio. However, 23 markers were used for linkage map construction, of which 20 were mapped in to three LGs spanning a total length of 195.26 cM with an average marker density of 9.53 cM. The length of the LG ranged from smallest 19.87 cM (LG3) to largest 102.17 cM (LG2). The number of markers mapped per LG varied from 2 (LG3) to 11 markers (LG2). The highest marker density was observed in LG 2 with an average marker density of 9.29 cM and the lowest marker density was observed in LG 1 with an average marker density 10.46 cM (Figure 3).

Single Marker Analysis
The SMA that assess the segregation of FW resistance at two stages (30 th and 60 th days) with respect to a marker genotype in RILs indicated 12 markers (SSR14, H3A12, H5A08, H4G11, CS27, CaM1402, TR19, TA59, TA110, TA96, TA200, H4E09) were associated with FW resistance across both the stages, seasons, and eld and pot experiments. In the present investigation marker having the strongest relationship was measured using PVE% value (Table 2), which indicates the overall percentage of variability of Foc resistance at a particular scoring time. In this study the PVE% ranged from 8.89% to 16.67% ( Table   2).The FW resistance on 30 th day in eld was associated with two markers in LG1 and eight markers in LG2 in season I and three markers in LG2 in season II. Two markers H4G11 and CAM1402 showed association in both seasons. Only one marker was associated with FW resistance on 60 th day in season I and two markers in season II. In pot culture experiment three markers were associated with wilt resistance on 30 th day and four on 60 th day. The marker H4G11 consistently showed linkage with wilt resistance in both seasons and the pot experiment. The markers H3A12, H5A08 and CS27 showed stable association in season I and the pot experiment.
Both late wilting QTLs (qW60-07-1-1 and qW60-08-1-1) identi ed from the eld experiment across two consecutive seasons (2007 and 2008 rabi) are located on LG1 and anked by same pair of markers CaM1402 and H4G11; the QTL qW60-07-1-1 had a LOD score of 3.38 while QTL qW60-08-1-1 had a LOD score of 4.37 and explained phenotypic variation of 12.37% and 12.30% respectively. Both QTLs had a positive additive effect with the contribution of resistance alleles from male parent. However, the QTL (qW60-pot-2-1) for late wilting from the pot experiment is located on LG2 and anked by markers CS27 and H3A12. The QTL had a LOD score of 5.17 and explained a phenotypic variance of 26.62% with contribution of resistance alleles from male parent.

Discussion
Marker validation is referred to determining the target phenotype in independent populations and different genetic backgrounds (Cakir et al. 2003). Marker assisted selection for disease resistance in chickpea, like any other crops, is a two-step process. The process starts with the detection of molecular markers linked to FW resistance and markers anking QTLs in uencing resistance. Subsequently those markers ought to be validated in different mapping populations and genotypes, so that effectiveness of the markers in identifying the desired genotypes can be evaluated in advance.
The current study evaluates the previously reported Foc resistance linked markers in a different mapping population for their application. Among eight different races of Fusarium oxysporum f. sp. ciceri reported earlier (Jimenez-Fernandez et al. 2013), races 1-4 are wide spread in India (Pande et al. 2007;Singh et al. 2008). The inheritance of resistance indicated that resistance to race 1, 2 and 4 is controlled by two or three genes (Upadhyaya et al. 1983a, b;Brinda and Ravikumar 2005), while the race 3 is controlled by a single gene. So also the case with race 5 (Sharma et al. 2005). However, it has been observed that the inheritance pattern is not always clear for some races, as for example there are at least three independent loci, H 1 , H 2 and H 3 , with complex inheritance pattern reported for the race 1 (Singh et al. 1987). Besides even involvement of many polygenes have been proposed for the observed variation and timing of wilting (Brinda and Ravikumar 2005). To overcome the problem posed by complex inheritance pattern for the resistance to certain races, it has been suggested that markers that are tightly linked to resistance QTL is an effective solution in breeding programmes via MAS. Accordingly, several efforts were made previously to identify markers that are tightly linked to FW resistance in chickpea. For example Mayer et al. (1997) developed two allele-speci c associated primer (ASAP) pairs, namely UBC170 and CS27, from RAPD bands that were linked at 7% recombination to the locus for resistance to FW race 1. Subsequently, several linkage maps were used to map FW resistance genes in chickpea (Winter et al. 2000;Sharma et al. 2004;Gowda et al. 2009;Millan et al. 2010;Sabbavarapu et al. 2013;Barman et al. 2014;Jingade and Ravikumar 2015). These studies revealed that the majority of the markers used in tagging wilt resistance are co-dominant microsatellite markers and clustered on the LG2. Thus a consensus map of LG2 is turned out to be a hot spot for resistance genes relating to wilt resistance for many races (Benko-Iseppon et al. 2003;CHO et al. 2004). In addition, earlier reports also mapped a substantial number of QTLs associated with resistance.
For the present study, the mapping population (RILs) was developed by hybridization of highly susceptible and resistant parents producing that potentially generated a full range of variability for wilt resistance. Out of 42 markers reported to be linked to wilt resistance, only 23 were found to be polymorphic in the parental lines. Contrary to some of the earlier reports (Gowda et al. 2009;Millan et al. 2010;Sabbavarapu et al. 2013) we did not observe a tight linkage between nineteen markers to disease resistance loci in our mapping population. The predictive power of the 19 markers was limited when tested in a different set of parents/genotypes. This observed discrepancy might be because of the nature of mapping populations used between studies. Most of the markers developed earlier were using a speci c mapping population and not all of the markers were polymorphic in the parental lines used for this study. Nevertheless, these markers could be of some use for certain breeding populations even though they are not diagnostic across genotypes.
The genetic map constructed with 23 markers consisted of 20 markers on three LGs spanning a length of 195.26 cM. Nearly 50% of the markers (11) mapped to LG2. The LGs in the present study were compared to earlier maps with common markers and the groups agree with the earlier maps with little changes in the order. The LG2 of the present study was same as that of LG2 reported in Gowda et al. (2009); reference map of chickpea (Winter et al. 2000) and LG3 of Radhika et al. (2007). Two clusters of Foc resistance to tested race 1 were observed in LG2 in the present study. The previous studies also indicated two clusters of Foc resistance in LG2. The underlying pattern from all these studies (Winter et al. 2000;Sharma et al. 2004;Sharma and Muehlbauer 2007;Gowda et al. 2009) including the present investigation is that LG2 contains some of the major loci for wilt resistance and they appear to be stable in many genotypes and mapping populations. In addition, we also identi ed another major locus of Foc resistance on LG1 in this study, which is independent of the rst locus (Brinda and Ravikumar 2005). The LG1 of the present study was same as that of LG6 of Sabbavarapu et al. (2013); LG2 of Radhika et al. (2007) and LG1 of Jingade and Ravikumar (2015). The linkage map of chickpea is generally conserved demonstrating low level of genetic recombination and rearrangement in intraspeci c crosses (Flandez-Galvez et al. 2003). Present study could con rm and support the earlier reports (Upadhyaya et al. 1983a, b;Singh et al. 1987), that two three major genes confer resistance to FW resistance in chickpea The single marker analysis identi ed 12 markers -SSR14, H3A12, H5A08, H4G11, CS27, CaM1402, TR19, TA59, TA110, TA96, TA200 and H4E09 -linked to wilt resistance. The markers TA96 (Sharma and Muehlbauer 2007;Winter et al. 2000;Sharma et al. 2004;Millan et al. 2006;Millan et al. 2010;Gowda et al. 2009); CS27 (Winter et al. 2000;Sharma et al. 2004;Soregaon and Ravikumar 2010); TA110 (Gowda et al. 2009) and TA59 (Millan et al. 2006) which were more closely related to FW resistance in different independent studies and genotypes showed close linkage in this study also. This consistent observed linkage pattern across investigations is extremely useful for the breeders, who could use such markers in their breeding programs with high degree of certainties when screening germplasm for wilt resistance. Only exception we noted was for the marker TA27 that showed association in many studies (Winter et al. 2000;Sharma et al. 2004;Millan et al. 2006;Millan et al. 2010) did not show linkage in our re-evaluation. Such a contradictory result in associating the marker with resistance to Foc race1 might be due to change in genetic population (Gurjar et al. 2009).
Among the eight QTLs identi ed using CIM in this study, four were in the same locus in LG1 anked by markers H4G11 and CaM1402 and two were in the same locus in LG2 anked by markers H3A12 and CS27. The QTLs identi ed in the present study are stable and reliable as they are based on both eld testing over years and pot screening of RILs for wilt reaction and con rm the loci identi ed earlier. The results presented here and those published earlier contribute to the emerging picture of two hotspots in chickpea genome for resistance to Fusarium wilt resistance. A majority of studies (Winter et al. 2000;Sharma et al. 2004;Millan et al. 2006;Millan et al. 2010), including ours, consider the two loci as major resistance loci with a PVE (%) values ranging from 12.36 to 76.76. Two additional QTLs identi ed in this study were purely based on a pot experiment, it has to be con rmed and validated in the future in replicated eld experiments.
Present study could con rm and validate several earlier ndings about the nature and strength of linkage between genetic markers and wilt resistant loci in chickpea, thereby enhancing the reliability of linked markers for their use in chickpea breeding for wilt resistance. Besides, validity and reliability of QTL marker data come from wilt sick eld over seasons and arti cial wilt sick pot data. The use of con rmed linked markers for the Foc resistance genes, identi ed in this study, facilitate selection of resistance genotypes unambiguously in chickpea breeding programmes to develop resistant genotypes. The QTL anking markers, H4G11, CaM1402, CS27, and H3A12 can be effectively used in the development of wilt resistant germplasm and variety development programs. They represent the genomic region with gene clusters for resistance against many Foc races including Foc race 1 which is more prevalent in India (Sharma and Muehlbauer 2007). Moreover, anchoring genomic areas of interest with reliable SSR markers has been a very pro table strategy allowing saturation of the genomic region surrounding the Foc resistance genes on LG2 and LG1 to target speci c genes involved in resistance. In addition we also con rm the association of other previously reported markers TA96, CS27, TA59, TA110 and H3A12 with FW by single marker analysis. The QTLs and markers identi ed and con rmed in the present study are useful resource for genomics assisted breeding for resistance to Fusarium wilt in chickpea.