Fine Mapping of the Wheat Leaf Rust Resistance Gene LrLC10 (Lr13) and Validation of Its Co-segregation Markers

Wheat leaf rust, caused by the fungus Puccinia triticina Eriks. (Pt), is a destructive disease found throughout common wheat production areas worldwide. At its adult stage, wheat cultivar Liaochun10 is resistant to leaf rust and the gene for that resistance has been mapped on chromosome 2BS. It was designated LrLC10 and is the same gene as cataloged gene Lr13 by pedigree analysis and allelism test. We fine-mapped it using recessive class analysis (RCA) of the homozygous susceptible F2 plants derived from crosses using Liaochun10 as the resistant, male parent. Taking advantage of the re-sequencing data of Liaochun10 and its counterpart susceptible parent, we converted nucleotide polymorphisms in the LrLC10 interval between the resistant and susceptible parents into molecular markers to saturate the LrLC10 genetic linkage map. Four indel markers were added in the 1.65 cM map of LrLC10 flanked by markers CAUT163 and Lseq22. Thirty-two recombinants were identified by those two markers from the 984 F2 homozygous susceptible plants and were further genotyped with additional ten markers. LrLC10 was finally placed in a 314.3 kb region on the Chinese Spring reference sequence (RefSeq v1.0) that contains three high confidence genes: TraesCS2B01G182800, TraesCS2B01G182900, and TraesCS2B01G183000. Sequence analysis showed several variations in TraesCS2B01G182800 and TraesCS2B01G183000 between resistant and susceptible parents. One KASP marker and an indel marker were designed based on the differences in those two genes, respectively, and were validated to be diagnostic co-segregating markers for LrLC10. Our results both improve marker-assisted selection and help with the map-based cloning of LrLC10.

Wheat leaf rust, caused by the fungus Puccinia triticina Eriks. (Pt), is a destructive disease found throughout common wheat production areas worldwide. At its adult stage, wheat cultivar Liaochun10 is resistant to leaf rust and the gene for that resistance has been mapped on chromosome 2BS. It was designated LrLC10 and is the same gene as cataloged gene Lr13 by pedigree analysis and allelism test. We fine-mapped it using recessive class analysis (RCA) of the homozygous susceptible F 2 plants derived from crosses using Liaochun10 as the resistant, male parent. Taking advantage of the re-sequencing data of Liaochun10 and its counterpart susceptible parent, we converted nucleotide polymorphisms in the LrLC10 interval between the resistant and susceptible parents into molecular markers to saturate the LrLC10 genetic linkage map. Four indel markers were added in the 1.65 cM map of LrLC10 flanked by markers CAUT163 and Lseq22. Thirty-two recombinants were identified by those two markers from the 984 F 2 homozygous susceptible plants and were further genotyped with additional ten markers. LrLC10 was finally placed in a 314.3 kb region on the Chinese Spring reference sequence (RefSeq v1.0) that contains three high confidence genes: TraesCS2B01G182800, TraesCS2B01G182900, and TraesCS2B01G183000. Sequence analysis showed several variations in TraesCS2B01G182800 and TraesCS2B01G183000 between resistant and susceptible parents. One KASP marker and an indel marker were designed based on the differences in those two genes, respectively, and were validated to be diagnostic co-segregating markers for LrLC10. Our results both improve markerassisted selection and help with the map-based cloning of LrLC10.

INTRODUCTION
Globally, common wheat (Triticum aestivum) is one of the most commonly cultivated crops, comprising 20% of human caloric intake and 15% of cultivated area in the world (FAOSTAT, 2015;WAP, 2017). Wheat leaf rust, caused by Puccinia triticina Eriks. (Pt) is one of the most damaging diseases of wheat, especially in coastal regions or areas with high temperatures and humidity during the wheat maturing seasons (Kolmer et al., 2018). In China, past widespread wheat leaf rust epidemics have caused severe yield losses (Dong, 2001;Zhou et al., 2013). In the future, due to impending climate change, leaf rust is expected to damage wheat production even more (Jiang et al., 2018). Utilization of wheat resistant cultivars considered a most effective, economical and environmentallyfriendly strategy for controlling this disease (Bariana et al., 2007;Singh et al., 2013).
Currently, about 80 leaf rust resistant genes have been reported and formally named in common wheat or its relatives (McIntosh et al., 2017;Qureshi et al., 2018), and by using different types of molecular markers, most of these genes have been mapped on the wheat chromosomes 1 . Development of robust molecular markers linked to resistance genes is essential in wheat disease resistance breeding, especially for resistance gene pyramiding. Nevertheless, among these designated leaf rust resistance genes, only a few have tightly linked molecular markers for marker-assisted selection 2 .
Because of limited wheat genomic sequence data, developing molecular markers for wheat genes has been difficult. But now, by combining the T. aestivum 'Chinese Spring' (CS) IWGSC RefSeq v1.0 genome 3 with annotations of high-quality gene models, these difficulties have been reduced, especially for gene location and markers development (Clavijo et al., 2017). Discovery of the highly abundant, locus-specific wheat nucleotide variations that can be used to identify the relevant genes are now within grasp because of affordable next-generation sequencing (Varshney et al., 2014;Xu et al., 2017). Compared with other kinds of markers, kompetitive allele-specific PCR (KASP) assays accelerate the conversion of DNA variations into available gene-linked markers. Taking advantage of such whole genomic sequences, Wu et al. (2018b) mapped wheat yellow rust resistance gene Yr26 on a 0.003-cM interval on chromosome 1B near the centromere, Narang et al. (2019) defined wheat leaf rust resistance gene LrP and yellow rust resistance gene YrP on a 15.71 Mb region on 5DS in the CS RefSeq v1.0 genome assembly, and Wu et al. (2019) localized the Pm52 locus within a 5.6 Mb interval on the long arm of chromosome 2B (2BL).
Bulked segregant analysis (BSA) can rapidly identify markers linked to target genes (Michelmore et al., 1991), and it has been improved by bulking homozygous recessive plants and using recessive class analysis (RCA) to map specific genes (Zhang et al., 1994). RCA is highly efficient, with a lower probability of misclassification and more reliability than using a random F 2 population. Furthermore, this approach avoids creating F 2:3 families and screening the entire F 2 population, thus saving time in fine mapping and map-based cloning. RCA has been proven to map genes efficiently and reliably (Zhang et al., 1994;Yao et al., 1997;Mei et al., 1999;Chen et al., 2006;Kiswara et al., 2014). In wheat, rapid gene mapping using RCA has been used to map a sterile female gene (Dou et al., 2009) and a stripe rust resistance gene, YrLM168a (Feng et al., 2015), and to successfully map and clone the powdery mildew resistance gene Pm60 (Zou et al., 2018).
In wheat, Lr13, first identified in the Canadian cultivar "Manitou" in 1966, is an important adult-plant leaf rust resistance gene (McIntosh et al., 1995) that is widely found in wheat cultivars (e.g., 'Frontana, ' 'Frondoso, ' and 'Fronteria') and used in many breeding programs throughout the world (Roelfs, 1988;Pathan and Park, 2006). In China, Lr13 is one of the main resistance genes and confers effective resistance to leaf rust (Singh et al., 1999;Yuan et al., 2007;Yuan and Chen, 2011;Ren et al., 2015;Zhang et al., 2019). Previous studies indicated that Lr13 was located on chromosome 2BS (McIntosh et al., 1995), and Bansal et al. (2008) reported Lr13 was delimited to a 13.8 cM interval flanked by markers Xksm58 and Xstm773b. Recently, using a segregating population of Lr13 near-isogenic lines with simple sequence repeat and KASP markers, Zhang et al. (2016) mapped Lr13 to a small interval of 10.7 cM, and the closest marker was kwh37 (4.9 cM). A morphological marker, hybrid necrosis gene Ne2m, was found linked to Lr13 by Singh and Gupta (1991), but it cannot be used to accurately detect Lr13 (Anand et al., 1991). Therefore, a co-segregated and diagnostic marker for Lr13 in molecular breeding is yet unavailable.
In this study, we confirmed that the leaf rust resistance gene LrLC10 in Liaochun10 is Lr13 by pedigree analysis, and finely mapped it to a close interval with recessive class analysis (RCA) through the markers developed according to the resequencing data of the parental lines. We also developed molecular markers that were closely linked to LrLC10 and that can be used to facilitate marker-assisted selection of LrLC10 in wheat resistance breeding.

Plant Materials
The spring wheat cultivar Liaochun10, is highly resistant to leaf rust, was crossed with two susceptible wheat lines Han 87-1 (87-1) and 7D49 (a wild emmer wheat introgression line created by the crossing IW123/Zheng98//87-1 * 2), to construct two F 2 segregating populations of 3,908 plants. Wild

Field Evaluation of Leaf Rust Symptoms at the Adult Stage
Wheat leaf rust isolate PHT (provided by the Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing, Y, indicates progressive necrosis; N, indicates the cross was not tested; A, indicates the band is identical with resistant parent; B, indicates the band is identical with susceptible parent. China) was used as the inoculum. PHT isolate was avirulent on Liaochun10 and RL4031. The populations were sown at the experiment farm of China Agriculture University, Beijing, China. At the late tillering stage (Feekes stage 5) at least one tiller of each plant was inoculated by injecting urediniospore suspended in 0.1% Tween 20 into the leaf bundle with a 10 mL syringe. The urediniospore was propagated in the greenhouse on the susceptible control, Xuezao.
The infection type of the flag leaf and the top second leaf of each individual was evaluated about 1-2 month postinoculation when the susceptible control was fully infected, based on an infection type scale of 0-4, where 0 indicated no visible symptoms, 0; indicated hypersensitive flecks, and 1-4, indicated small uredinia with necrosis, small-to medium-sized uredinia with green islands and surrounded by necrosis or chlorosis, medium-to large-sized uredinia with chlorosis, and large uredinia without chlorosis, respectively. Values 0-2 were categorized as resistant and 3-4 were classified as susceptible (Roelfs et al., 1992). A second assessment was conducted for each plant 4 days after the first examination.

Allelism Tests
We used an F 2 population derived from Liaochun10 × RL4031 to determine the allelic relationships between genes Lr13 and LrLC10. The responses of each F 2 plant to Pt race PHT was determined by the rust response method described above.

Development of Molecular Markers
The sequences of all the markers anchored in the LrLC10 (Lr13) genetic linkage map were used as queries to search against the Chinese Spring reference genome sequence (RefSeq v1.0) to define the genome interval of the resistance gene on chromosome arm 2BS. Near the LrLC10 locus, singlenucleotide polymorphisms (SNPs) or insertion/deletion (indel) polymorphisms were found based on re-sequencing result of the two parents (the concrete method refer to Chai et al., 2018) and the 300 bp flanking sequence of those indel sites which were ≥5 bp that were obtained from the Chinese Spring reference genome sequence 4 , Primer3 (v.0.4.0) 5 was used to design the indel markers. The SNPs or indels (<5 bp) were converted into kompetitive allele-specific PCR (KASP) markers, which were designed using PolyMarker 6 . The markers used in this study were listed in Table 2.

Marker Genotyping Assays
PCR amplification was conducted in a 10 µL reaction volume consisting of 5 µL 2 × Tag PCR StarMix with loading dye, 50-100 ng/µL DNA 1.5 µL, 1.5 µL primer (mixture of forward and reverse primer, 2 µM), and 2 µL H 2 O. PCR proceeded with initial denaturation at 94 • C for 5 min, then 35 cycles at 94 • C for 30 s, 30 s at 50-60 • C for primer annealing (depending on the specific primers), 72 • C for 30 s of extension; and the final extension at 72 • C for 5 min. The PCR product was separated in either 8% or 10% non-denaturing polyacrylamide gels (acrylamide:bisacrylamide = 39:1) that were silver stained and photographed. KASP assays were performed following the protocol described in Wu et al. (2018a).

Construction of the Genetic Linkage Map
We performed chi-square analysis of the leaf rust test data from the segregating F 2 populations to confirm the goodnessof-fit of the observed ratios to theoretical expectations. The recombination frequencies of the resistance gene and the markers were calculated according to Chen et al. (2006). Using the Kosambi mapping function, we converted the recombination frequencies to centimorgans (Kosambi, 1943) and drew the genetic map using Mapdraw v2.1 (Liu and Meng, 2003).

Genetic Analysis of Wheat Leaf Rust Resistance Gene LrLC10 in Two Segregating Populations
The parental lines 87-1 and 7D49 were highly susceptible to Pt race PHT, [infection type (IT) = 3], whereas Liaochun10 was highly resistant (IT = 0, Figure 1). We examined the two segregating F 2 populations that grew from crossing the susceptible lines with Liaochun10. The 87-1 crossed with Liaochun10 produced the F 2 population including 3,057 plants, of which 2,300 were resistant and 757 were susceptible to Pt isolate PHT (χ 2 3:1 = 0.092, P > 0.05). In the F 2 population derived from 7D49 crossed with Liaochun10, 624 plants were resistant and 227 were susceptible to Pt race PHT (χ 2 3:1 = 1.268, P > 0.05). These results indicated that leaf rust resistance in Liaochun10 is controlled by a single dominant gene.

Allelism Test of LrLC10 and Lr13
Liaochun10, RL4031, and the 1,395 F 2 plants from the cross of Liaochun10 × RL4031 were evaluated against Pt race PHT. We found no susceptible plants, thus confirming that LrLC10 in Liaochun10 was on the same locus as Lr13. Since there was Lr13 donor parents Frontanan and UP301 in the Liaochun10 pedigree (Singh and Gupta, 1991;He et al., 2001;Pathan and Park, 2006), we concluded that leaf rust resistance gene LrLC10 in Liaochun10 is Lr13.

Molecular Mapping of Leaf Rust Resistance Gene LrLC10 (Lr13)
We chose 92 extremely susceptible individuals from the 7D49 × Liaochun10 F 2 population to be re-genotyped using markers linked to LrLC10 that were established by Lv et al. (2017) (Figures 2A,B). To define the LrLC10 physical interval, we searched the sequences of all markers anchored in the genetic map against the Chinese Spring reference genomic sequence (RefSeq v1.0) and found that the relative physical positions of those markers were generally consistent with the genetic map (Figures 2B,C). Two flanking markers, CAUT163 and Xbarc18, spanned an approximately 100 Mb region (153,676,348,323) in the reference genome, and here we detected numerous sequence variations between the parents when we analyzed the re-sequencing data. Twenty indel primer pairs were designed based on those insertion/deletion polymorphisms we found between parental lines in the 11 Mb (159,000,000-170,000,000) section that was 6 Mb from marker CAUT163 going toward LrLC10. Among these, 4 markers (Lseq22, Lseq29, Lseq31, and Lseq35) were successfully added to the genetic map and the LrLC10 gene was delimited within a 1.65 cM area between markers CAUT163 and Lseq22, an interval corresponding to a 5.7 Mb (153,676,602-159,302,377) region in the CS reference genome (Figure 2C).

Development of Tightly Linked Markers to LrLC10 (Lr13)
We used those co-dominant flanking markers, CAUT163 and Lseq22, to identify recombinants in the 984 homozygous, susceptible F 2 plants. Thirty-two recombinant plants were identified and then used for fine mapping of LrLC10.
Based on the parents' re-sequencing data that corresponded to the 5.7 Mb interval of the CS RefSeq v1.0, we designed 80 indel primers and a KASP marker. They were tested on 3 parental lines and 10 markers were polymorphic between the parents and used to finely map LrLC10 (Figure 3).
Among the 32 recombinant plants, 28 showed recombination between marker CAUT163 and LrLC10, while 4 recombination events were detected between marker Lseq22 and LrLC10. We used the 10 polymorphic markers located between the flanking markers to examine these recombinants and found that the closest flanking markers were Lseq301 (with 1 recombination event) and Lseq85 (with 2 recombination events) and markers Lseq302 and Lseq102 co-segregated with LrLC10 (Figure 3). These results suggest that LrLC10 locus is located in a 314.3kb region between markers Lseq85 and Lseq301 (157,688,415-158,002,717) in the CS RefSeq v1.0 (Figure 3). Markers Lseq85 and Lseq301 were designed based on the 5 bp and 6 bp deletions, respectively, in Liaochun10 as compared to 7D49. The KASP marker Lseq302 was based on the SNP (A/T) detected in exon 2 of TraesCS2B01G182800 between Liaochun10 and 7D49 ( Figure 4A and Table 3), while Lseq102 was developed based on a 9-bp deletion in the TraesCS2B01G183000 coding region between Liaochun10 and 7D49 ( Figure 4B and Table 3).

Validation of LrLC10-Co-segregating Markers for Marker-Assisted Selection
We wanted to test if these co-segregating markers (Lseq302 and Lseq102) could be used for marker-assisted selection of LrLC10 in different backgrounds. We tested 25 wheat leaf rust resistant accessions and 10 susceptible cultivars with those 2 markers to evaluate their utility. Twenty-three of the resistant accessions had the same marker genotypes as Liaochun10 and all the susceptible cultivars' genotypes were identical to 87-1 and 7D49 (Figures 5A,B and Table 1). Moreover, we found that the F 1 plants of the crosses of those 14 of the 23 resistant accessions with Xuezao, which was proved to carry the hybrid necrosis gene Ne1 (unpublished results), showed progressive necrosis ( Table 1). According to Zhang et al. (2016), Lr13 and Ne2m are the same gene; so by inference, the F 1 plants' phenotypes indicate that those 14 cultivars have leaf rust resistance gene Lr13. These results suggest that markers Lseq302 and Lseq102 can be used to identify Lr13.
To evaluate the distribution of Lseq302-L and Lseq102-L (the marker alleles of Lseq302/Lseq102 in Liaochun10) in China, a panel of 524 common wheat accessions/landraces from China was tested with these markers. Lseq302-L and Lseq102-L always co-existed in all the cultivars, forming a specific haplotype block. The haplotype of Lseq302-L/Lseq102-L was present in various frequencies in 4 of 10 agro-ecological production zones: I, North China winter wheat region (30.76%); II, Yellow and Huai River valleys winter wheat region (24.56%); III, middle and lower Yangtze River valley winter wheat region (14.81%); and V, South China winter wheat region (33.33%) (Figure 5C and Table 4).
Frontiers in Plant Science | www.frontiersin.org
In the process of fine mapping LrLC10, enough polymorphic markers were found to narrow down the genetic interval covering the targeted gene. We accomplished this by re-sequencing the parental lines and developing indel and SNP markers based on sequence information in the targeted region. Four indel markers revealed polymorphisms and localized LrLC10 gene in a 1.65 cM genetic interval, which corresponded to a 5.7 Mb interval on the Chinese Spring reference genomic sequence (Figures 2B,C). Based on sequence diversities between the parental lines on the candidate interval, we designed 9 indel markers and a KASP marker and used these to test the 32 recombinants that we identified using LrLC10-flanking markers CAUT163 and Lseq22 derived from 984 homozygous, susceptible F 2 individuals. LrLC10 was finally delimited into a 314.3-kb genomic interval on the Chinese Spring reference sequence v1.0 by markers Lseq301 and Lseq85 (Figure 3). This mapping of LrLC10 demonstrated our methods efficiently developed molecular markers based from the re-sequencing data of the parents.
Lr13 is one of the most widely distributed leaf rust resistance genes in wheat (McIntosh et al., 1995), but it has become ineffective in some regions, such as Mexico and South America (Singh and Rajaram, 1992). However, it is effective in combination with other resistance genes, such as Lr3ka, Lr34, and Lr16 (Kolmer, 1992). Therefore, we need diagnostic molecular markers for Lr13 to facilitate selection or stacking it with other Lr genes. In our study, KASP marker Lseq302 and indel marker Lseq102 co-segregated with Lr13 and were effective in diverse wheat backgrounds (Figures 5A,B, Table 1, and Supplementary Table S1). Therefore, these diagnostic markers may be used for efficient marker-assisted selection of Lr13, thus enabling researchers to either pyramid it with other adult plant resistant genes to achieve durable leaf rust resistance or stack it with stripe rust, stem rust, and powdery mildew resistance genes (e.g., Yr27, Sr40, and pm42) on chromosome 2BS, to create multiresistance accessions (McDonald et al., 2004;Hua et al., 2009;Wu et al., 2009).
In this study, we delimited LrLC10 to a 314.3 kb region on short arm of chromosome 2B in Chinese Spring reference genome sequence (RefSeq v1.0). Three high confidence genes (TraesCS2B01G182800, TraesCS2B01G182900, and TraesCS2B01G183000) have different functions based on the annotation of Chinese Spring reference genome 7 , which encode a typical NBS-LRR protein, Ribonuclease, and an F-box domain containing Leucine-rich repeats protein were located in this region. DNA sequence comparison showed that the parents did not differ in TraesCS2B01G182900. Compared to 7D49, Liaochun10 had a 9 bp deletion in its TraesCS2B01G183000 coding region, resulting in a deletion of three amino acids, and another 1-bp deletion that led to a translational frame shift ( Figure 4B and Table 3). In TraesCS2B01G182800, we found many sequence variations, including 56 SNPs and 4 indels, between Liaochun10 and its rust-susceptible parent. Among them, we found 18 SNPs and one indel in the intron region, and 38 SNPs and three indels in exon 2. Among those 38 SNPs, 26 caused amino acid substitutions, and two of the indels led to a translational frame shift. A 3-bp insertion in 7D49 resulted in an amino acid insertion that was not found in Liaochun10 ( Figure 4A and Table 3). Based on the TraesCS2B01G182800 and TraesCS2B01G183000 sequence polymorphisms between the parental lines, we developed the markers Lseq102 and Lseq302 and found that they co-segregated with LrLC10 (Figure 3). In total, our results suggest that LrLC10 (Lr13) might be one of those two annotated genes. However, there is still a chance that the sequence corresponding to LrLC10 is absent in the CS genomic sequence. Therefore, our analysis of re-sequencing data based on the wheat reference genome sequence is not enough to be absolutely certain of its identity. Because of this, a library of the resistant parent must be constructed so that a physical map would enable the cloning of LrLC10. Recently, some alternative methods (e.g., MutRenSeq, TACCA, and MutChromSeq) have been used to clone wheat disease resistance genes