Leaf rust of wheat (Triticum aestivum L.) caused by Puccinia triticina Eriks., is found worldwide in all wheat growing areas, often resulting in the yield losses up to 50%. In India too, it causes substantial economic losses during epidemic years (Nayar et al. 2002). Although more than 60 leaf rust resistance genes have been identified at the seedling and adult plant stages (McIntosh et al. 2008), these resistances are very often overcome by the pathogen; therefore, incorporation of resistant genes is a continuous process to counter evolving pathotypes. Combining several genes, into one cultivar, (‘pyramiding’) is an effective strategy to enhance the durability of the resistance against rusts. Identification of molecular markers for the resistance genes and their use in breeding can facilitate gene pyramiding into valuable backgrounds in short time, making it more cost effective.

Leaf rust resistance (Lr) genes such as Lr9, Lr19, Lr24, Lr28 and Lr32 provide complete protection against leaf rust pathotypes in wheat-growing regions of Asia (Tomar and Menon 2001). The resistance gene Lr28 located on chromosome 4AL was transferred into wheat from Aegilops speltoides (McIntosh et al. 2008). It has been transferred into different genetic backgrounds of Indian collection (Tomar and Menon 1998), and no undesirable linkage drag or negative effect on yield has been observed (Kumar and Raghavaiah 2004). Therefore Lr28, in combination with other resistance genes, has a potential to reduce yield losses caused by P. triticina. Three markers tagging Lr28 have been hitherto reported by Naik et al. (1998), Vikal et al. (2004) and Cherukuri et al. (2005). For widespread utilization of the molecular markers in marker assisted breeding, validation in multiple backgrounds is essential. The STS marker derived by Naik et al. (1998) was found to be loosely linked to Lr28 in previous studies (Sharp et al. 2001; Chełkowski et al. 2003) and hence not considered in our study. The objectives of this study were to (1) test available microsatellite markers for Lr28, (2) identify markers closely linked to Lr28 and (3) validate their potential in marker-assisted selection.

In the study, two F2 derived F3 populations were used. The first population consisted of 110 F2:3 families from a cross between a near- isogenic line (NIL) HW2037 and its recurrent parent HD2329. Secondly, a F1 harvest of 145 seeds of a cross between cultivar WL711 and CS + Lr28: 2D/2M 3/8 (acc. 2956) (kindly provided by Prof. R. G. Saini, Punjab Agricultural University, Ludhiana, India) was advanced to generate a F2:3 population. Six NILs for Lr28 developed at the Regional Research Station, IARI, Wellington, India, namely HW2031 (C306 + Lr28), HW2032 (LOK1 + Lr28), HW2033 (WH147 + Lr28), HW2035 (NI5439 + Lr28), HW2037 (HD2329 + Lr28) and HW2061 (Sonalika + Lr28), along with their recurrent parents were used in validation studies. Additionally a set of 48 wheat cultivars (postulated to carry leaf rust resistance genes) from advanced varietal trials under the All India Co-ordinated Wheat and Barley Program were used in the final marker validation studies.

The leaf rust pathotype 77-5 (121R63-1), avirulent on Lr9, Lr18, Lr19, Lr24, Lr25, Lr28, Lr29, Lr32, Lr39, Lr41, Lr42, Lr45 genes was used for phenotyping. Seedlings were grown at 18 ± 2°C and >85% RH in the controlled condition facility. Seven-day-old seedlings were inoculated using the standard procedure as described by Browder (1971)). Reaction to the disease, i.e. infection type (IT) was noted 14 days after inoculation, using the 0 - 4 Stakman’s scale (Stakman et al. 1962). For testing F2:3 families, about 16 - 30 plants of each family were inoculated. Based on IT of F2:3 families, the F2 genotypes were classified as homozygous resistant (HR), segregating (Seg) and homozygous susceptible (HS).

Total DNA was extracted from healthy leaves, using the modified CTAB method (Rogers and Bendich 1985). Based on available wheat genetic maps of chromosome 4AL (Somers et al. 2004; Song et al. 2005), a total of 26 microsatellite markers viz. barc70, barc78, cfa2256, cfd2, cfd30, cfd88, cfd257, gwm162, gwm397, gwm494, gwm610, gwm637, psp3030, wmc96, wmc161, wmc219, wmc232, wmc262, wmc313, wmc497, wmc500, wmc513, wmc617, wmc722, wmc776 along with the reported linked marker gwm160 (Vikal et al. 2004) were screened. Amplifications were performed in 10 μl reaction volume, using protocol reported in GrainGenes database (http://wheat.pw.usda.gov). The amplification products were electrophoresed on 5% native polyacrylamide gels (PAGE) stained in ethidium bromide and documented using digital gel documentation system. For amplification of marker TPSCAR SCS421570, the reaction conditions reported by Cherukuri et al. (2005) were used. Only HR and HS plants were considered for linkage analysis. Chi squared distribution analyses were carried out to test if the observed segregation ratios for leaf rust resistance and marker loci fit in the expected ratio. Genetic linkage analysis was performed using Mapmaker program version 3.0 (Lander et al. 1987).

The parental lines HW2037 and CS + Lr28: 2D/2M 3/8 (acc. 2956) were highly resistant to race 77-5 (IT fleck ;), while HD2329 and WL711 were susceptible (IT 3+ and 4). Both the populations segregated in a 3:1 ratio in the F2 generation indicating a monogenic inheritance of the resistance (Table 1). In the marker analysis, seven out of the 26 tested microsatellite markers, showed polymorphism between HD2329 and HW2037. In selective genotyping using homozygous plants, six promising markers were identified and further tested in the population.

Table 1 Seedling reaction in F2 and F3 generations of HD2329 × HW2037 and WL711 × acc. 2956 to the pathotype 77-5 under controlled conditions
Full size table

With marker wmc313, the resistant parent and resistant F2 plants showed two bands (320 bp and 309 bp), while the susceptible plants amplified only a single band (309 bp). The polymorphic fragment of 320 bp was amplified only in the resistant F2-segregants and co-segregated with the resistance. Based on the segregation data (Table 2), wmc313 was linked at a distance of 5.0 cM in the population HD2329 ×HW2037 (Fig. 1). The results were further validated in the population WL711 × CS + Lr28: 2D/2M 3/8 (acc. 2956) where a linkage distance of 4.5 cM was observed (Table 2). Marker gwm160 (Vikal et al. 2004) was loosely linked at a distance of 16.8 cM in the HD2329 × HW2037 population and remained unlinked in the second population (data not shown). Marker SCS421570 showed a polymorphic band of 570 bp in the resistant lines and mapped at 15.3 cM and 6.7 cM in the populations HD2329 × HW 2037 and WL711 × CS + Lr28: 2D/2M 3/8 (acc. 2956), respectively (Table 2). The distance of about 11 cM has also been reported in the previous study using HW4002 × Agra local population (Cherukuri et al. 2005).

Table 2 Test of linkage between Lr28, wmc313 and SCS421570 marker in the F2 generation of crosses HD2329 × HW2037 and WL711 × acc. 2956
Full size table
Fig. 1
figure 1

Linkage map of the microsatellite and SCAR markers linked to the Lr28 locus in the population HD2329 × HW2037 of wheat. The locus names are indicated on right and Kosambi map distance (cM) on left side of map

Full size image

The markers wmc313 and SCS421570 were further validated in a few NILs carrying the gene Lr28 in different genetic backgrounds (Figs. 2a, b). Both the markers amplified the expected polymorphic bands in the NILs except HW2033 and HW2035. Additionally, a set of 48 wheat cultivars, postulated to carry different Lr genes, was genotyped with the markers wmc313 and SCS421570. With wmc313, expected double band pattern was observed only in the cultivar MACS6145 carrying the gene Lr28, whereas other cultivars not carrying the gene showed a single band (Fig. 2c). No cross amplification with other resistance genes was observed. With marker SCS421570, the expected band at 570 bp was also observed only in MACS6145.

Fig. 2
figure 2

Validation of markers linked to gene Lr28 on a set of NILs carrying Lr28 (a, b) and wheat cultivars (c). Marker wmc313 is resolved on 5% native PAGE and TPSCAR SCS421570 on 1.5% agarose gel. The polymorphic band is indicated by an arrow. The postulated genes are given in the parentheses. The sample order for Figs. 2a, b : M. 100 bp ladder / Low range DNA ruler plus (Bangalore Genei); (1) HW2031; (2) C306; (3) HW2032; (4) LOK1; (5) HW2037; (6) HD2329; (7) HW2061; (8) Sonalika; (9) HW2033; (10) WH147; (11) HW2035; (12) Lr28 source (CS2D/2M 3/8). The order for Fig. 2c : M. 100 bp ladder; (1) Lr28 source (CS2D/2M 3/8); (2) HD2189 (Lr13 + Lr34); (3) HI977 (Lr23 + Lr24); (4) NI5439 (Lr34); (5) MACS2496 (Lr23 + Lr26); (6) WH147 (Lr13); (7) HW2044 (Lr24); (8) UAS305; (9) Kalyansona (Lr34); (10) HD4672 (Lr23); (11) MACS6145 (Lr28)

Full size image

The markers SCS421570 and wmc313 (identified in this study) were able to identify the gene Lr28 in diverse genetic backgrounds. Based on the linkage distances of 4.5 cM and 5.0 cM in both populations, wmc313 is the most closely linked marker for Lr28 reported so far. The accuracy of wmc313 was established on a set of 48 cultivars as well as NILs for Lr28. Among the NILs, HW2033 and HW2035 failed to amplify the expected amplicon with either marker. The failure of amplification in HW2035 has also been reported by Prabhu et al. (2003). However, the absence of the polymorphic amplicon in HW2033 was unexpected. The results for HW2033 were consistent for two different seed stocks in our collection, indicating the need to verify every seed stock. Such mis-identity of NILs has been reported earlier by Prabhu et al. (2003). Although the marker SCS421570 gave accurate amplification in the NILs and the elite germplasm, when tested in the two populations it showed more distance from the gene as compared to wmc313. Since SCS421570 acts as a dominant marker, chances of misclassification are high. In contrast, although wmc313 also acts as a dominant marker, it still gives a common amplicon (309 bp) in all genotypes and thus has an advantage of reducing false negatives. Interestingly, in both the populations SCS421570 and wmc313 were found to flank the Lr28 gene. Hence, the use of such flanking markers makes the selection for the target gene in a breeding programme more accurate in comparison to a single marker.

The Lr28 gene provides a high level of resistance against 77-5, the most predominant pathotype in India. Under a heavy leaf rust infection, it contributes to a higher 1000-grain weight and number of effective tillers per plant with no negative effects on yield and bread-making quality (Kumar and Raghavaiah 2004). However, this gene has not been widely used in breeding programmes in India. So far only one cultivar, MACS6145, carrying Lr28 has been released in 2005 by the Agharkar Research Institute, India and Indian Agricultural Research Institute, Regional Station, Wellington, India. This could be due to the fact that conventional pathological testing is a tedious and time-consuming task. Results of the present study suggest that this can be effectively overcome by the use of markers wmc313 and SCS421570 in screening breeding materials for Lr28. Since tight linkage of a marker with the desired gene is important for success of MAS in breeding, the marker wmc313 should be more effective in MAS for Lr28 and facilitate pyramiding to confer a durable resistance to leaf rust.