Mendelization and Molecular Mapping of a Large-effect QTL Conferring Durable Adult-plant Resistance to Stripe Rust in a Chinese Wheat Landrace ‘Gaoxianguangtoumai’

The Chinese wheat landrace ‘Gaoxianguangtoumai’ (GX) has exhibited a high degree of adult-plant resistance (APR) to stripe rust in eld environments for more than a decade. To reveal the genetic basis for APR to stripe rust in GX, a set of 249 F 6:8 recombinant inbred lines (RILs) was developed from a cross between GX and the susceptible cultivar ‘Taichung 29’. The parents and RILs were evaluated for disease severity at the adult-plant stage in eld environments by articial inoculation with the currently predominant Chinese Puccinia striiformis f. sp. tritici races during three cropping seasons, and genotyped using the Wheat 55K single-nucleotide polymorphism (SNP) array to construct a genetic map with 1,871 SNP markers. Two stable APR quantitative trait loci (QTL), QYr.GX-2AS and QYr.GX-7DS from GX, were detected on chromosomes 2AS and 7DS, which explained 15.5–27.0% and 9.6–15.6% of the total phenotypic variation, respectively. Compared with published genes and QTL, QYr.GX-7DS is likely Yr18, whereas QYr.GX-2AS is probably novel. Haplotype analysis revealed that QYr.GX-2AS is likely to be rare which present in 5.3% of the 325 surveyed Chinese wheat landraces. By analyzing a near-isogenic line population, QYr.GX-2AS was further mapped to an interval with a physical distance of about 1.37 Mb and co-segregated with a Kompetitive allele-specic PCR (KASP) marker. Furthermore, three tightly linked KASP markers were highly polymorphic among 109 Chinese wheat cultivars. The short physical interval and tightly linked KASP markers developed in this study will facilitate marker-assisted selection and map-based cloning of QYr.GX-2AS. effective resistance than Yr18. The objectives of the present study were to (1) identify the QTL conferring APR to stripe rust in a recombinant inbred line (RIL) population developed from the cross between GX and a susceptible cultivar, ‘Taichung 29’ (TC 29), (2) validate and mendelize the novel QTL in a near-isogenic line (NIL) derived population, and (3) develop tightly linked Kompetitive allele-specic PCR (KASP) markers for use in marker-assisted selection in breeding programs. Assessments of adult-plant stripe rust responses The F 6:8 RIL population and the parental lines were evaluated for APR to stripe rust during the 2017–2018, 2018–2019, and 2019–2020 growing seasons (referred to as CZ2018, CZ2019, and CZ2020, respectively). The NIL-derived population (F 2 ) was evaluated for APR to stripe rust during the 2020–2021. In all tests, 20 seeds of each line were planted in rows 2 m in length and spaced 30 cm apart, with individual plants spaced 10 cm apart. The susceptible cultivar TC 29 was planted in every 20th row as a susceptible control. To provide inoculum for infection, the susceptible cultivars SY95-71 and AvS were planted around the perimeter of the experimental area as spreaders. Articial inoculation was conducted using a mixture of currently predominant Pst races in China (comprising CYR32, CYR33, CYR34, G22-14, Su11-4, Su11-5, and Su11-7). Stripe rust response was rst recorded by scoring the IT and disease severity (DS) when the susceptible checks SY95-71 and AvS showed more than 80% DS, and was followed by two additional evaluations at 7-day intervals (i.e., three evaluations in total) for three randomly selected individual plants. The IT was recorded based on the 0–9 scale of Line and Qayoum (1992). The DS was scored as the percentage infected leaf area (0, 5%, 10%, 20%, 40%, 60%, 80%, or 100%) in accordance with the Chinese National Standard, GB/T 15797-2011. The nal DS (FDS) was used for phenotypic analysis. FDS of accessions carrying the different haplotypes. Haplotype data were combined with provenance information to examine the geographic distribution of the superior haplotypes in the 10 major agro-ecological production zones of Chinese wheat landraces.


Introduction
Stripe rust (also known as yellow rust), caused by Puccinia striiformis f. sp. tritici (Pst), is among the most harmful and widespread obligate pathogens of common wheat (Triticum aestivum L.) worldwide (Knott 1989;Wellings 2011). China has the largest stripe rust epidemic area in the world (Zeng and Luo 2006;, and frequent epidemics have been reported . Since the 1950s, four severe epidemics of wheat stripe rust have occurred in China in 1950China in , 1964China in , 1990China in , and 2002, resulting in yield losses of 6.0, 3.2, 1.8, and 1.4 million tonnes, respectively (Li and Zeng 2000; Wan et al. 2004). The main cause of the outbreaks is the emergence of new virulent races that overcome the widely deployed resistance genes (Chen and Kang 2017). At present, CYR32 and CYR34 are the most virulent and predominant Pst races in China ; Wang et al. 2018). Continuous improvement in the resistance of wheat cultivars to cope with evolving races of Pst is a high priority to control stripe rust (Manickavelu et al. 2016).
To date, more than 300 genes or quantitative trait loci (QTL) for stripe rust resistance on the 21 wheat chromosomes have been reported (Rosewarne et al. 2013;McIntosh et al. 2018). In general, these resistance genes and QTL can be classi ed into two major classes: all-stage resistance (ASR) and adult-plant resistance (APR). ASR usually confers complete resistance during all growth stages and is simple to select during breeding. However, most ASR genes are race speci c and encode nucleotide-binding and leucine-rich repeat (NLR) proteins, and therefore are effective against only a subset of Pst races. With regard to the dynamic rust pathogen populations of the virulent races, only a small number of the characterized ASR genes, such as Yr5 (Marchal et Wu et al. 2018). In contrast, APR is effective starting at adult-plant growth stages and typically provides a degree of partial resistance, but it is usually non-race-speci c and provides durable resistance to Pst. Of the three APR genes cloned to date, Yr18 encodes a putative ATP-binding cassette transporter (Krattinger et al. 2009), Yr36 encodes a kinase domain and a lipid-binding domain (Fu et al. 2009), and Yr46 encodes a predicted hexose transporter (Moore et al. 2015). These genes represent different protein families compared with classical ASR genes (the NLR family) and provide unique mechanisms effective against a broader range of pathogens. As an example, Yr18 has been globally used as a component of durable rust resistance in breeding programs and no evolution of increased virulence has been observed for almost 100 years (Krattinger et al. 2009). To achieve a high degree of durable resistance, combining multiple APR genes into the same background has been considered as an important strategy for improvement of stripe rust resistance in wheat breeding.
Chinese wheat landraces are farmer-developed and maintained as traditional cultivars in China. These landraces harbor rich genetic diversity for stripe rust resistance. Numerous stripe rust genes or QTL have been identi ed, such as Yr1 (Bansal et al Many resistant accessions of Chinese wheat landraces continually display APR to stripe rust in eld environments, providing a novel resistance resource for the breeding of wheat cultivars with durable resistance to stripe rust. The novel APR genes require further research for identi cation, validation, and mendelization to facilitate their use in wheat breeding. Gaoxianguangtoumai (GX) is a spring wheat landrace from Sichuan Province in southwest China, which is a regional center for oversummering and overwintering of the stripe rust pathogen. This landrace has exhibited a high degree of APR to stripe rust in eld environments for more than a decade, but little information is available on the genetic basis of resistance in this landrace. The objectives of the present study were to (1) identify the QTL conferring APR to stripe rust in a recombinant inbred line (RIL) population developed from the cross between GX and a susceptible cultivar, 'Taichung 29' (TC 29), (2) validate and mendelize the novel QTL in a near-isogenic line (NIL) derived population, and (3) develop tightly linked Kompetitive allele-speci c PCR (KASP) markers for use in marker-assisted selection in breeding programs.

Plant materials and races
The Chinese wheat landrace GX (accession number ZM7854 in National Germplasm Bank, China (NGBC) and AS1579 in Triticeae Research Institute, Sichuan Agricultural University) originating from Sichuan Province, was crossed (as the female parent) with the highly stripe rust susceptible wheat cultivar TC 29. In total, 249 F 6:8 RILs derived from a single F 1 seed were developed by single-seed descent. A NIL-derived population of 130 individuals (F 2 ), derived from a residual heterozygous line Evaluation of resistance to stripe rust Seedling tests to evaluate the stripe rust resistance of GX and TC 29 were conducted in a greenhouse using two prevalent Chinese Pst races (CYR32 and CYR34). Five plants of each line were sown in a plastic pot lled with nutrient soil and grown in a controlled environment in the greenhouse. Seedlings were inoculated at the two-leaf stage with each Pst race in accordance with the protocol of Hickey et al. (2012). Inoculated plants were placed in a dew chamber at 10 °C and 100% relative humidity for 24 h in the dark, and then moved to separate growth chambers at 15-16 °C with 12-14 h of light daily.
When the susceptible check 'Mingxian 169' showed full sporulation, the infection type (IT) on the second leaf (approximately 15-18 days after inoculation) was scored using a 0-9 scale (Line and Qayoum 1992). Plants with IT scores of 0 to 6 were considered resistant, whereas plants with IT scores of 7 to 9 were considered susceptible.
Assessments of adult-plant stripe rust responses were conducted at the Chongzhou Experimental Station (30°33′N, 103°39′E), Sichuan Agricultural University, Chengdu, China. The F 6:8 RIL population and the parental lines were evaluated for APR to stripe rust during the 2017-2018, 2018-2019, and 2019-2020 growing seasons (referred to as CZ2018, CZ2019, and CZ2020, respectively). The NIL-derived population (F 2 ) was evaluated for APR to stripe rust during the 2020-2021. In all tests, 20 seeds of each line were planted in rows 2 m in length and spaced 30 cm apart, with individual plants spaced 10 cm apart. The susceptible cultivar TC 29 was planted in every 20th row as a susceptible control. To provide inoculum for infection, the susceptible cultivars SY95-71 and AvS were planted around the perimeter of the experimental area as spreaders. Arti cial inoculation was conducted using a mixture of currently predominant Pst races in China (comprising CYR32, CYR33, CYR34, G22-14, Su11-4, Su11-5, and Su11-7). Stripe rust response was rst recorded by scoring the IT and disease severity (DS) when the susceptible checks SY95-71 and AvS showed more than 80% DS, and was followed by two additional evaluations at 7-day intervals (i.e., three evaluations in total) for three randomly selected individual plants. The IT was recorded based on the 0-9 scale of Line and Qayoum (1992). The DS was scored as the percentage infected leaf area Haplotype variants were detected using Haploview v4.2 (http://www.broad.mit.edu/mpg/haploview/). The haplotypes detected in at least 10 accessions were considered to be major haplotypes. Boxplots were generated to display the average FDS of accessions carrying the different haplotypes. Haplotype data were combined with provenance information to examine the geographic distribution of the superior haplotypes in the 10 major agro-ecological production zones of Chinese wheat landraces.
Exome capture sequencing, development of KASP markers, and genetic mapping Genomic DNA of the parents, GX and TC 29, was sequenced using the wheat exome capture sequencing protocol described by Dong et al. (2020). The sequence variants were identi ed using the variant calling pipeline GATK4 (Heldenbrand et al. 2019). The SNPs in the target region for QYr.GX-2AS detected by exome capture sequencing and the Wheat 55K array were converted to KASP markers using the PolyMarker online tool (Ramirez-Gonzalez et al. 2015). The KASP markers were used to screen the parents and NILs to con rm polymorphism before genotyping the NIL-derived population. The KASP assays were performed in 96-well format as 10 μL reactions containing 2 μL of 50-100 ng genomic DNA, 5 μL of HiGeno 2× Probe Mix B, 0.24 μM of each forward primer, 0.6 μM of the common reverse primer, and double distilled water to make up the volume to 10 μL. Each PCR was conducted using the BIO-RAD CFX96 qPCR system. Thermocycling was performed with a touchdown

Stripe rust response of the parents and RILs
Plants of GX were susceptible (IT = 8) to CYR32 and CYR34 at the seedling stage (Fig. 1a), but exhibited strong resistance (IT = 3, FDS < 10%) to mixed Pst races at the adult-plant stage in three crop seasons from 2018 to 2020 (Fig. 1b, Fig. 2, Table  S1). These results indicated that GX showed non-race-speci c APR to stripe rust. The frequency distributions of RILs for FDS were continuous with a pronounced skewness towards resistance, and the average FDS of RILs for GX × TC 29 was 12.5%-15.7% in the eld tests, suggesting the presence of a large-effect QTL in the RIL population (Fig. 1c, Fig. 2, Table S1). Broadsense heritability (H 2 ) was 96.7% for FDS in all tests (Table 1). Correlation coe cients (R 2 ) for FDS of the RILs among the different environments were signi cant (P < 0.01) and ranged from 0.82 to 0.95 (Table S2). Table 1 The summary of nal disease severity (FDS) data for the RILs population from the Gaoxianguangtoumai (GX) × Taichung 29 (TC 29) recorded in elds at Chongzhou in 2018-2020  (Table   S3). The map consisted of 21 linkage groups de ned with representatives from each of the 21 chromosomes.
Two major QTL conferring APR to stripe rust were detected from the resistant parent GX in each of the three eld tests and BLUP data ( Table 2, Fig. 3a, b). The most highly signi cant QTL, designated QYr.GX-2AS, was mapped to the short arm of chromosome 2AS and explained up to 27.0% of the phenotypic variation with a LOD score of 8.1 (Table 2, Fig. 3a). A second QTL, QYr.GX-7DS, was anked by the SNP markers AX-109379249 and AX-110431109 on chromosome 7DS and overlapped with Yr18, and explained 9.6%-15.6% of the phenotypic variation in all trials and BLUP data ( Table 2, Fig. 3b). We concluded that it was highly likely that QYr.GX-7DS corresponded to Yr18.
To determine the effects of the QTL, the RILs were divided into four groups based on the presence/absence of the most closely linked anking markers of QYr.GX-2AS and QYr.GX-7DS. Clearly, the RILs that carried one of the QTL showed a lower FDS than those without any QTL (average FDS = 63.4%). In particular, the RILs carrying only QYr.GX-2AS showed only 9.3% of the average FDS, which was similar to the effect of both QTL in combination (average FDS = 7.1%) (Fig. 3c). This result indicated that QYr.GX-2AS had a large effect on stripe rust resistance and provided relatively stronger resistance than QYr.GX-7DS. Haplotype analysis of QYr.GX-2AS To assess the distribution of QYr.GX-2AS among 325 Chinese wheat landraces, the favorable haplotype was identi ed by haplotype analysis and seven SNP markers tightly linked to QYr.GX-2AS were screened from the Wheat 55K or 660K SNP arrays (Fig. 4a, b, c). Eight major haplotypes (n > 10) were detected in the panel (Fig. 4a, b). GX and 15 other accessions clustered with Hap1 (Table S4), which showed a frequency of about 5.3% in the total population (Fig. 4a) 4c). These results revealed that the favorable haplotype of QYr.GX-2AS was Hap1, which was relatively rare among the Chinese wheat landraces.
Validation and mapping of QYr.GX-2AS To further validate and map the location of QYr.GX-2AS, the SNPs in the target region for QYr.GX-2AS identi ed by exome capture sequencing and the Wheat 55K array were selected for conversion to KASP markers. Eleven markers were con rmed to be polymorphic between GX and TC 29 (Table S5). Combined with the KASP markers for QYr.GX-2AS, a NIL-derived population of 130 individuals (F 2 ) for QYr.GX-2AS was developed from a residual heterozygous plant in the F 8 generation of RILs. No signi cant phenotypic differences were observed in the NIL-derived population, except for APR to stripe rust (Fig.   5a). With regard to stripe rust response in the eld test, the F 2 plants of the NIL-derived population were clearly classi able into 97 resistant (IT ≤ 6) and 33 susceptible (IT ≥ 7) individuals, which ts the expected ratio (3:1) for a single Mendelian factor (chi-square goodness-of-t test, χ 2 = 0.01, P = 0.92) (Table S6). Using the newly developed eleven KASP markers (Table  S6) to construct the genetic map, QYr.GX-2AS was localized to a 1.37 Mb interval between KP2A_36.85 and KP2A_38.22, and co-segregated with the KASP marker KP2A_37.09 (Fig. 5b).

Validation of KASP markers for marker-assisted selection
To check the speci city and polymorphism of the linked marker of QYr.GX-2AS for marker-assisted selection, a set of 109 Chinese wheat cultivars was tested with the markers KP2A_36.85, KP2A_37.09, and KP2A_38.22 (Fig. S1, Table S7). Most of the lines ampli ed the susceptible-speci c alleles in the three markers, which showed 85.3%, 99.1%, and 95.4% polymorphism, respectively, in the cultivars (Table S7). Thus, these KASP markers can be used as the speci city markers for marker-assisted selection of QYr.GX-2AS in the vast majority of Chinese wheat cultivars.

Discussion
Breeding for durable resistance to stripe rust has been among the highest priorities for wheat breeding in the last decade (Chen 2013). A large number of genes or QTL that confer various degrees of APR to stripe rust have been identi ed, but most only have minor effects on stripe rust response and are therefore di cult to use in breeding. Thus, identi cation of genes or QTL with a high degree of APR that are useful in breeding programs is required. The Chinese wheat landrace GX has displayed a high degree of APR to stripe rust in eld environments for more than a decade in southwest China. The strong APR to stripe rust of GX is conferred by two QTL identi ed on chromosomes 2AS and 7DS, respectively. The QTL on 2AS (QYr.GX-2AS), being more effective than the QTL on 7DS (Yr18), had a large effect in the reduction of stripe rust severity at adult-plant stages, and thus shows great potential for use in breeding durable resistance to stripe rust in wheat.
QTL analysis is a useful procedure to reveal possible multiple loci when analyzing complex genetic traits, such as APR to stripe rust, in resistant germplasm. However, this procedure only allows approximate mapping of the QTL (Tanksley and Hewitt 1988) owing to the heterogeneity in genetic backgrounds. The con dence interval of many QTL spans a considerable genetic distance and, as a result, molecular markers for these QTL may not be reliably used in marker-assisted selection. As a strategy for accurate mapping of QTL in genetic analysis, NIL-derived populations that allow the conversion of a quantitative trait into a Mendelian factor have been widely used for ne mapping and cloning of many important QTL in wheat, such as Yr18 (Krattinger et al. 2009