Molecular mapping of the Aegilops speltoides-derived leaf rust resistance gene Lr36 in common wheat (Triticum aestivum)

Leaf rust caused by Puccinia triticina Eriks. is the most prevalent wheat rust worldwide and occurs annually wherever wheat is grown. The most economical and environmentally friendly method to control this fungal disease is genetic resistance, which is achieved through deploying effective resistance genes. Tightly-linked molecular markers facilitate gene tagging and their deployment with other resistance genes, which in turn contribute to durable leaf rust resistance. The leaf rust resistance gene Lr36 derived from Aegilops speltoides Taush and introgressed into hexaploid wheat by backcrossing, is located on chromosome 6BS. Despite detection of low frequencies of virulence for this gene, no tightly-linked marker is available in disease resistance breeding. Therefore, this research aimed at analysis of simple sequence repeats (SSR) markers linked with Lr36 in 171 individuals of an F2 population from a cross between the Lr36-carrying line (ER84018) and the susceptible cultivar; Boolani. Of 36 primer pairs on chromosome 6BS tested for polymorphism in parents and resistant and susceptible bulks, eight were polymorphic of which the markers Xgwm88 and Xcfd13 flanked Lr36 by 3.8 and 5.2 cM, respectively. The identified markers were validated on 35 genotypes with different genetic backgrounds including few Australian wheat leaf rust differential sets, Iranian landraces and cultivars, and several cultivars and/or breeding lines from Switzerland and Afghanistan, which confirmed that they can be used in selection for Lr36 in wheat breeding programs.


Introduction
Wheat (Triticum aestivum) is a strategic crop all over the world. However, its production is often threatened by biotic and abiotic stresses (Arzani and Ashraf 2017). Among biotic stresses, fungal diseases including the three rusts; leaf, stripe, and stem rusts, result in significant annual yield losses and have detrimental effects on seed quality (McIntosh 1998). Leaf rust caused by Puccinia triticina Eriks. is the most prevalent wheat rust worldwide and occurs annually wherever wheat is grown (Kolmer 2013). The most economical and environmentally friendly method to control this disease is genetic resistance. However, rust pathotypes can mutate to virulence on wheat Abstract Leaf rust caused by Puccinia triticina Eriks. is the most prevalent wheat rust worldwide and occurs annually wherever wheat is grown. The most economical and environmentally friendly method to control this fungal disease is genetic resistance, which is achieved through deploying effective resistance genes. Tightly-linked molecular markers facilitate gene tagging and their deployment with other resistance genes, which in turn contribute to durable leaf rust resistance. The leaf rust resistance gene Lr36 derived from Aegilops speltoides Taush and introgressed into hexaploid wheat by backcrossing, is located on chromosome 6BS. Despite detection of low frequencies of virulence for this gene, no tightlylinked marker is available in disease resistance breeding. Therefore, this research aimed at analysis of simple sequence repeats (SSR) markers linked with Lr36 in 171 individuals of an F 2 population from a cross between the Lr36-carrying line (ER84018) and the susceptible cultivar; Boolani. Of 36 primer pairs on chromosome 6BS tested for polymorphism in parents genotypes with previously effective resistance genes. Therefore, breeding for rust resistance can be a neverending task that has to be supported by ongoing pathogenicity surveys to confirm continued resistance gene effectiveness.
Stacking multiple resistance genes is often difficult due to the unavailability of appropriate pathotypes and/or problems concerning their presence in combination with other resistance genes conferring lower infection types. Recent advances in molecular biology have facilitated the indirect selection of individual genes based on closely linked markers (Gupta et al. 1999), a procedure known as marker-assisted selection (MAS; Qureshi et al. 2018). Combinations of effective resistance genes not only increase the durability but, in many cases, also increase the degree of resistance. Simple sequence repeats (SSRs) or microsatellites are polymorphic tandem repeats of up to six base pairs (Tautz and Renz 1984) that can be identified and traced in genetic analysis or in selection. Though more comprehensive marker systems have been developed, SSR markers still provide valuable information for MAS in rust resistance programs because they are highly polymorphic, low cost to phenotype, and their analysis can be automated (Hayden et al. 2006). In addition, microsatellite consensus maps enable high-density maps of the wheat genome. In this context, more than 80% of primer sets are genome-specific and detect only a single locus in one of the three sub-genomes of bread wheat (A, B, or D). Moreover, publicly available databases provide opportunities to predict allele sizes in breeding populations and develop molecular and genomics strategies in gene mapping (Röder et al. 1998;Somers et al. 2004).
The seedling stage leaf rust resistance gene, Lr36, which was transferred from Aegilops speltoides Taush (2n = 14) into hexaploid wheat by backcrossing, is located on the short arm of chromosome 6B (Dvořák and Knott 1990). This gene has been rarely deployed in wheat cultivars despite having no linkage to undesirable quality or agronomic characters (Dvorak and Knott 1990) and being effective in the USA, China, Poland, Kazakhstan, Pakistan, Egypt, and Iran (Aktar-Uz-Zaman et al. 2017;Ali et al. 2018;Czajowski and Czembor 2016;Koyshybaev 2019;Li et al. 2016;Safavi and Afshari 2013;Zhang et al. 2019). Hence, it could be pyramided and deployed with other rust resistance genes. However, linked molecular markers that benefit its deployment in wheat breeding programs have not been reported (McIntosh et al. 2017). Therefore, this study describes mapping the gene Lr36 using SSR markers in a bi-parental population and validates the efficiency of closely-linked markers in different genetic backgrounds.

Plant materials and leaf rust inoculation
The wheat line ER84018, carrying the leaf rust resistance gene Lr36 (Dvorak and Knott 1990), was crossed with 'Boolani', the Iranian susceptible cultivar, to produce F 1 seeds. Subsequently, the resulting F 2 population with 171 individuals and their F 3 progenies were used for phenotypic and genetic analysis. Parental genotypes and all F 2 plants were grown in 10 cm diameter pots in a temperature-controlled greenhouse (18 ± 2 °C).
The P. triticina pathotype FHTQQ (isolate no. 92-23; virulent for 2c, 3a, 3bg, 3ka, 10, 11, 14b, 16, 17, 26, 30, B), which produced infection types "0;1 + N" and "33 + " on ER84018 and Boolani (Table 1), respectively, was multiplied on a susceptible line. Fresh urediniospores were then mixed with talcum powder in a ratio of 1:4 and used to inoculate the test populations at the two-leaf stage. The inoculated plants were kept in a plastic-covered container at 100% humidity, 18 °C, and dark condition for 24 h prior to moving them to microclimate rooms at 18-24 °C. Infection types (ITs) were recorded according to the 1-4 scale described by McIntosh et al. (1995) approximately 10-12 days post inoculation (dpi) when the susceptible cultivar Boolani showed an IT of "33 + ". Plants with ITs less than "2" were considered resistant and those with IT "3" and above were classified as susceptible.
After rust scoring, all 171 F 2 plants were transplanted to 20 cm diameter pots and individually harvested. The segregation ratio from F 2 population was confirmed by testing F 3 families with the same pathotype as described before. Ten to 12 seedlings from each F 3 family and both parents were inoculated as described previously, and phenotypes were recorded as homozygous resistant, segregating heterozygous, and homozygous susceptible.

Marker analysis
Rust-free leaf tissue from each F 2 plant and both parents was used to isolate DNA using the CTAB method (Yu et al. 2017). A Nanodrop ND-1000 (Wilmington, USA) and 1% (w/v) agarose gel were used to evaluate the quality and quantity of DNA. Working solutions for both genomic DNA and primers were prepared in 200 µl volumes at a concentration of 50 ng µl −1 . As described by Michelmore et al. (1991), resistant and susceptible DNA bulks were made from equal amounts of DNA from 20 most resistant and 20 most susceptible F 2 individuals.

Statistical and linkage analyses
Alleles for the SSR markers were scored A, B, and H for the resistant, susceptible and heterozygous, respectively, and '-' was used for missing values (PCR failure). The phenotypic data of F 3 families were used to establish the genotypes of each F 2 individual. The χ 2 test was used to evaluate the goodness of fit of observed and predicted ratios. Linkage analysis was conducted using JoinMap v0.5 (Van Ooijen 2019) with a threshold LOD ≥ 3.0 and the Kosambi mapping function (1943) based on the recombination frequencies.

Marker validation
The efficiency of the linked markers to select Lr36carrying genotypes was evaluated in 35 wheat genotypes with different genetic backgrounds. The validating genotypes included two groups; the first group comprised of Australian cultivars with known resistance genes used as differential testers in leaf rust studies. The second group included Iranian cultivars and landraces, Swiss cultivars and breeding lines, and Afghan genotypes. The genotypes in this group were classified as carrying/lacking Lr36 based on their pedigrees or rust response in the greenhouse. The DNA extraction protocol and PCR conditions were the same as mentioned earlier.

Genotyping and linkage map
Of the 36 pairs of SSR primer sets located on chromosome 6B, eight showed polymorphisms between the parents and bulks and hence, were used to genotype the whole F 2 population (Fig. 2). Based on PCR amplifications, the primer gwm88 produced a band of 160 bp in ER84018 while it amplified a 140 bp band in Boolani. The corresponding bands in the resistant and susceptible parents for cfd13 were 480 bp and 220 bp, respectively. The primer pair wmc486 amplified a fragment of 200 bp in the susceptible cultivar; Boolani, the susceptible progenies and the heterozygotes while no band was amplified in the resistant parent and the homozygous resistant offsprings. The remaining markers produced amplicons as presented in Table 4. The χ 2 tests for segregation of these loci, indicated that all primers except wmc486 comply with the expected Mendelian ratio of 1:2:1 for co-dominant inheritance, whereas the wmc486 primer fitted a 3:1 ratio for dominance (Table 4) and at LOD ≥ 3, showed a significant distortion, and was consequently categorized in a separate linkage group. Overall, seven primer sets were mapped to one linkage group (with an LOD of 3 and greater) covering an interval of approximately 37.7 cM on chromosome 6BS. The closest markers were Xgwm88 and Xcfd13 flanking Lr36 at a distance of 3.8 and 5.2 cM, respectively. The markers Xgwm518, Xgwm193, and Xwmc105 were located 9.1, 10, and 15.2 cM from this gene, respectively. The marker Xbarc101 was mapped 17.5 cM proximally to Lr36 while the marker Xbarc198 was placed 20.2 cM distal to it (Fig. 3).

Marker validation
To validate their usefulness, the closely linked markers; Xgwm88 and Xcfd13 were used to evaluate 35 wheat genotypes from Australia, Iran, Switzerland Table 3 Frequencies of different phenotypes in F 2 and F 3 populations from the cross between the leaf rust resistant line carrying Lr36; 'ER84018' and the susceptible cultivar 'Boolani' when infected with Puccinia triticina pathotype FHTQQ (isolate no. 92-23) a The total number of F 3 differs from that of F 2 because some of them did not set seed b R and S represent resistant and susceptible phenotypes, respectively in F 2 . R, Seg and S indicate resistant (Lr36Lr36), segregant (Lr36lr36) and susceptible (lr36lr36) in F 3 , respectively. The resistant individuals showed resistance ITs "0;1" and "1 + 2 N", the susceptible showed the ITs "33 + " and the segregating families represented a combination of the above-mentioned ITs to the Puccinia triticina pathotype FHTQQ c Non-significant at p = 0.01 (df = 1) d Non-significant at p = 0.01 (df = 2) and Afghanistan. The markers Xgwm88 and Xcfd13 amplified 140 and 220 bp amplicons, respectively, in all the genotypes that lacked Lr36 (Supplementary  Table 1).

Discussion
Wheat relatives are valuable resources for resistance breeding to biotic stresses, especially rust diseases. Though the Aegilops speltoides-derived leaf rust resistance gene, Lr36, has been mapped on the short arm of 6B (Dvorak and Knott 1990), it has not been deployed in wheat cultivars likely due to the unavailability of a diagnostic marker for its selection. Therefore, phenotyping and genotyping were performed on an F 2 population and its F 3 progenies to map it using SSR markers.
In the current study, P. triticina pathotype FHTQQ (isolate no. 92-23) showing a very low IT on Lr36-carrying genotype and a high IT on Boolani, was used to phenotype the F 2 and F 3 populations derived from the cross between these two genotypes. Most F 2 individuals were resistant and showed a segregation ratio of 3:1 indicating that the resistance in ER84018 was controlled by a single dominant gene which was confirmed by a segregation ratio of 1:2:1 in the F 3 generation.
To map Lr36, SSR primers specific to chromosome 6B, were evaluated for polymorphism between the two bulks of resistant and susceptible, of which eight revealed clear polymorphisms. As generally expected with SSR markers, all tested markers displayed co-dominant inheritance except Xwmc486 that fitted to a 3:1 genotypic ratio. This marker showed a null allele in the resistant plants (amplified only in Boolani, the susceptible lines and the heterozygotes) and therefore, could not differentiate heterozygotes from the susceptible homozygotes. Altering of the annealing site has likely led to the loss of the amplicon, resulting in null alleles (Naik Vinod et al. 2015).
The most closely-linked markers to Lr36 were the flanking markers; Xgwm88 and Xcfd13, at 3.8 and 5.2 cM followed by Xgwm518 and Xwmc105 at 9.1 and 10 cM, respectively. The order of markers was slightly different from that of the SSR consensus map developed by Somers et al. (2004). This is likely due to the translocation of Lr36 from Aegilops speltoides to wheat, which has led to different recombination frequencies depending on the distance from the breakage points. Additionally, the type and sample Table 4 Segregation of SSR primers in F 2 plants from the cross between the leaf rust resistant line carrying Lr36; 'ER84018' and the susceptible cultivar 'Boolani' on wheat 6BS chromosome a The resistant individuals displayed a range of ITs i.e. "0;1" and "1 + 2 N" to Puccinia triticina pathotype FHTQQ b The susceptible individuals showed high ITs of "33 + " to the same pathotype c The ratios represent the number of individuals in which bands corresponding to homozygous resistant, heterozygous and homozygous susceptible were amplified, respectively d Differences in total number of observed genotypes are due to the non-amplification of the corresponding band (s)  size of the tested population might have contributed to the altered genetic distances (Liu et al. 2013). The gene Lr36 was conclusively mapped on the short arm of 6B similar to the two tightly linked leaf and yellow rust resistance genes Lr53 and Yr35 as reported by Dadkhodaie et al. (2011). According to their findings, these genes segregate independently from the gene Lr36, and the marker Xbarc198 was distal to Lr53 at an approximate distance of 28 cM while our results located this marker at a distance of 20.2 cM from Lr36. Therefore, it could be inferred that the gene Lr36 is located distal to Lr53.
Both markers were tested on a panel of 35 wheat genotypes from different backgrounds. The Australian Fig. 3 Genetic mapping of the leaf rust resistance gene Lr36 on chromosome 6BS based on analysis of an F 2 population derived from a cross between the resistant line 'ER84018', and the susceptible parent 'Boolani', and seven linked SSR mark-ers. The 6B consensus map described by Somers et al. (2004) is also shown. Marker loci common to both maps are connected by solid lines. The genetic distances are indicated in cM on the left side of each map and the markers on the right side 26 Page 8 of 9 Vol:. (1234567890) cultivars are known to carry known resistance genes. The Iranian cultivars are mostly derived from CIM-MYT germplasms, which lack Lr36 in their pedigrees. Previous phenotypic evaluation of these genotypes along with landraces had shown they do not carry this gene (A. Dadkhodaie, unpublished data). Similarly, phenotyping and pedigree information evidently showed the absence of this gene in Swiss and Afghan genotypes. Therefore, the results of genotyping with these markers were in agreement with those of previous studies confirming their reliability in screening for Lr36.
Overall, though sequence-based genotyping and single nucleotide polymorphism (SNP) markers have dominated in genomic selection, SSR markers could facilitate finding neighboring SNPs in gene regions and fine mapping. In the present study, we developed PCR-based molecular markers (SSR) for the leaf rust resistance gene Lr36 for the first time. Both linked markers; Xgwm88 and Xcfd13 gave clearly scorable bands and categorized resistant, susceptible, and heterozygotes. Furthermore, their application in tracking Lr36 in diverse wheat lines and genotypes produced unambiguous and precise outcomes and confirmed the marker-trait association. Therefore, both markers could be utilized for marker-assisted selection in breeding programs. Since this gene is still effective in many countries across the world, these markers not only enable researchers to combine it with other seedling and adult plant rust resistance genes but also lay the foundations for its map-based cloning and fine-mapping.