QTL Mapping for Leaf Rust Resistance in a Wheat Recombinant Inbred Line Population of Liang66-S/Hengmai28 Authors: Cunyao Bo, Chenyang Pang, Xianghai Meng, Ding Li, Minghui Zhao, Chuncai Shen, Qiang Li, Wenchen Qiao Published: 2026 ABSTRACT Leaf rust (LR), caused by Puccinia triticina, is a major constraint to global wheat production. Identifying quantitative trait loci (QTLs), conferring adult-plant resistance (APR), and developing breeder-friendly markers are essential for durable disease control. In this study, a recombinant inbred line (RIL) population derived from Liang66-S × Hengmai28 was evaluated for leaf rust severity across four environments. Maximum disease severity (MDS) showed continuous variation and moderate to high correlations among environments, indicating polygenic inheritance. Five APR QTLs were mapped on chromosomes 2BL, 3BS, 3BL, 7BL and 7DL, explaining 4.5–9.8% of the phenotypic variance (PVE) with LOD scores of 2.9–7.9. Among these, QLr.DFI-2BL, QLr.DFI-3BS, and QLr.DFI-7BL correspond to previously reported APR loci, whereas QLr.DFI-3BL and QLr.DFI-7DL represent potentially novel loci. QLr.DFI-3BL was detected across all four environments, while QLr.DFI-3BS and QLr.DFI-7BL showed the highest R2 (7.7–8.5% and 8.4–9.8%). Multiple defense-related genes within the QTL intervals for LR were identified, including the putative NBS-LRR, the putative BTB/POZ-MATH, F-box, receptor-like kinase, polyphenol oxidase, calcium-dependent protein kinase, peroxidase, and ethylene-responsive transcription factors. Based on the flanking markers, five KASP markers were developed. K-LR-3BS (for QLr.DFI-3BS) and K-LR-7BL (for QLr.DFI-7BL) successfully genotyped a natural population with 120 accessions, with call rates of 100% and 94.2%, and showed significant association with reduced leaf rust severity. However, further validation in diverse genetic backgrounds and populations is required before they can be considered for routine marker-assisted selection. These results provide two KASP markers as promising markers that require further validation in diverse backgrounds and highlight QLr.DFI-3BS and QLr.DFI-7BL as promising targets for fine mapping and eventual cloning of APR genes in wheat. 1. Introduction Leaf rust, caused by the fungal pathogen Puccinia triticina (Pt), is a devastating disease characterized by wide distribution, high frequency, and sudden outbreaks, posing a serious threat to wheat production. Leaf rust occurs in nearly all wheat-growing regions worldwide, with particularly severe impacts in warm and humid environments such as the southeastern United States, the Pampas of South America, the Indo-Gangetic Plains of South Asia, and the North China Plain. The pathogen can cause yield losses exceeding 50% under severe epidemics [1]. More critically, global warming has intensified the epidemic of leaf rust in major wheat-growing regions worldwide, including the southern United States, South America, Canada, Eastern Europe, Egypt, and the entirety of China [2]. Rising temperatures and altered precipitation patterns have extended the seasonal window conducive to pathogen reproduction, enhanced overwintering survival of urediniospores, and accelerated disease cycles, leading to earlier onset and more frequent severe outbreaks in many temperate zones. Severe yield losses due to leaf rust outbreaks occurred globally in 2008, 2009, 2012, 2015, and 2023 [3]. Therefore, effective control of leaf rust is essential for safeguarding global food security. Breeding resistant varieties is the most economical and effective strategy for managing leaf rust [4]. Wheat leaf rust resistance is primarily categorized into all-stage resistance (ASR) and adult-plant resistance (APR). ASR is typically governed by one or a few major genes and confers race-specific resistance based on the gene-for-gene theory, often resulting in high resistance or immunity to specific pathogen races [5]. This type of resistance imposes strong selection pressure on the pathogen and can rapidly lose effectiveness upon the emergence of new virulent races. For instance, the widely deployed genes Lr1, Lr13, and Lr26 have successively become ineffective in many regions [6]. APR is characterized by susceptibility at the seedling stage but resistance at the adult stage and is controlled by multiple minor genes acting in concert [7]. Typically race-nonspecific, APR reduces disease severity through a prolonged latent period, reduced infection frequency, and decreased spore production, thus imposing low selection pressure on the pathogen and conferring durable and stable resistance [8]. Although individual APR genes have limited effects, pyramiding four to five such genes can achieve high to near-immune resistance levels. Since the early 2000s, the International Maize and Wheat Improvement Center (CIMMYT) has successfully developed high-yielding wheat materials with durable, high-level APR through gene pyramiding [9]. Resistance breeding is severely constrained by the narrow genetic basis and the limited availability of effective resistance sources. Since the initiation of genetic research on wheat leaf rust resistance in 1926, a total of 83 leaf rust resistance genes have been officially named internationally, most of which are ASR genes [10]. The well-characterized APR genes include Lr12, Lr13, Lr22a, Lr22b, Lr34, Lr35, Lr37, Lr46, Lr48, Lr49, Lr67, and Lr68 [9,11,12]. Over ten Lr genes have been cloned, with most ASR genes encoding NLR proteins (e.g., Lr1, Lr10, Lr21), while APR genes encode diverse protein types [13,14]. Notably, Lr34 (ABC transporter) and Lr67 (hexose transporter) exhibit pleiotropic effects, conferring broad-spectrum and durable resistance to multiple rust diseases [13,15]. Among the ASR genes, Lr9, Lr19, Lr24, Lr28, Lr29, Lr38, Lr47, Lr51, and Lr53 remain effective in many regions, whereas Lr2c, Lr3, Lr16, Lr17, LrB, Lr3bg, Lr14b, Lr23, and Lr39 have largely lost effectiveness against most current races [6,10]. Due to its durability and broad-spectrum characteristics, APR has become a key focus in modern resistance breeding. However, its breeding application has been constrained by the complex genetic basis and minor individual effects of its component genes. Recent advances in genetic research and molecular marker technologies have enabled the mapping of over 200 QTLs for leaf rust resistance across diverse wheat populations, offering valuable resources for pyramiding breeding [16,17]. Widely grown resistant varieties such as Zhou 8425B and Pvon76 carry multiple APR genes, demonstrating the feasibility and value of this strategy in breeding programs [18]. With leaf rust posing an escalating threat under climate change, and given the concomitant challenges of resistance breakdown and linkage drag in existing genes, the ongoing discovery and deployment of novel resistance sources have become a pressing global imperative [19]. Therefore, identifying and cloning new genes, elucidating their genetic mechanisms, and developing efficient molecular tools for marker-assisted selection (MAS) represent critical tasks in current leaf rust resistance breeding [20]. In this study, we used a population of 253 recombinant inbred lines (RILs) derived from a cross between Liang66-S (moderate to high susceptibility to leaf rust) and Hengmai28 (high leaf rust resistance). The objectives were to: (1) identify novel leaf rust resistance loci and tightly linked SNP markers through genetic analysis; (2) develop user-friendly Kompetitive Allele Specific PCR (KASP) markers to facilitate efficient MAS for wheat leaf rust resistance. 2. Results 2.1. Phenotypic Analysis A RIL population consisting of 253 lines derived from the Liang66-S/Hengmai28 cross was evaluated for maximum disease severity (MDS) across four environments. Descriptive statistics revealed continuous variation with wide ranges (minimum 1.0–5.0, maximum 95.0–100.0) and mean values between 46.4 and 55.2, indicating moderate to high disease pressure (Table 1 and Figure 1). The standard deviations ranged from 20.6 to 22.4, and coefficients of variation (CV) were 40–46% across all environments, suggesting similar relative variability. The presence of transgressive segregation was evident as some RILs exhibited lower MDS than the resistant parent (Hengmai28) and others higher than the susceptible parent (Liang66-S). Frequency distributions were approximately normal, consistent with polygenic inheritance of adult-plant leaf rust resistance. Pairwise Pearson correlation coefficients among the four environments were moderately positive, ranging from 0.60 (Xinxiang2022 vs. Xinxiang2023) to 0.70 (Baoding2023 vs. Xinxiang2022 and Baoding2023 vs. Xinxiang2023) (Table A1). All correlations were highly significant (p < 0.05). These moderate correlation levels indicate that while the same genetic factors largely influenced disease severity across environments, genotype-by-environment (G × E) interactions were also present (Table A2). Seedling inoculation results revealed that Liang66-S and Hengmai28 exhibited moderate to high susceptibility (infection type 3–4) to the leaf rust races THTT, PHTT, and THTS. Adult plant stage evaluations showed that Hengmai28 displayed high resistance (MDS 18–32% across all environments) to the leaf rust races, while Liang66-S exhibited moderate to high susceptibility (MDS = 42–58%). The analysis of variance showed that the genotype effect remained significant when tested against the genotype × environment interaction mean square (p = 0.001), while the genotype × environment interaction was highly significant (p < 0.001). Based on the estimated variance components (σ2g = 21.39, σ2ge = 241.37, σ2e = 65.1), the broad-sense heritability (Hb) on a line-mean basis was 0.25, indicating low-to-moderate genetic control. The genotype × environment interaction variance alone accounted for 73.6% of the total phenotypic variance (σ2g + σ2ge + σ2e), confirming that genotype performance was substantially environment-dependent. 2.2. QTL Mapping for Leaf Rust Resistance A total of five QTLs imparting APR to leaf rust were identified across four environments in the Liang66-S/Hengmai28 RIL population. These QTLs were distributed on chromosomes 2BL, 3BS, 3BL, 7BL, and 7DL (Table 2 and Figure 2). The phenotypic variation explained (R2) by each QTL ranged from 4.5% to 9.8%, and the LOD scores ranged from 2.9 to 7.9. On chromosome 2BL, a QTL designated QLr.DFI-2BL was detected in environments Baoding2022 and Xinxiang2022 with a 3.9-LOD of 4.6 and an R2 of 4.5–5.2% (additive effect: −3.8–−4.2). The resistance allele was contributed by Liang66-S. On chromosome 3B, two QTLs were identified. QLr.DFI-3BS was detected in Baoding2022 and Xinxiang2022 with an LOD of 5.9–7.0 and R2 of 7.7–8.5% (additive effect: 5.9–6.8), and mapping to chromosome 3BS with the resistance allele was contributed by Hengmai28. QLr.DFI-3BL was the most environmentally stable QTL in this study, being detected across all four environments (Baoding2022-Xinxiang2023). It exhibited an LOD of 3.8–4.5 and an R2 of 5.1–6.0% (additive effect: 5.0–5.8), with the resistance allele also derived from Hengmai28. On chromosome 7BL, QLr.DFI-7BL was the most robustly detected QTL in this study, with an LOD of 6.3–7.9 and an R2 of 8.4–9.8% (additive effect: 5.5–6.7). This QTL was identified in Baoding2022 and Xinxiang2023, with the resistance allele from Hengmai28. Finally, a minor QTL QLr.DFI-7DL was identified on chromosome 7D in Baoding2022 and Xinxiang2023, with an LOD of 2.9–3.2 and an R2 of 4.7–5.2% (additive effect: −3.8–−4.8). The resistance allele at this locus was contributed by Liang66-S. 2.3. Candidate Gene Analysis of Leaf Rust Resistance QTLs To dissect the molecular basis of the five stable QTLs for wheat leaf rust resistance, we prioritized candidate genes involved in plant immunity based on four criteria: (i) high-confidence annotation from the IWGSC v1.0; (ii) functional annotation matching known immunity-related families (including NBS-LRR, RLK, CDPK, BTB/POZ, F-box, PPO, peroxidase, and transcription factors); (iii) leaf-preferential or stress-responsive expression patterns from the wheat-expression.com database; and (iv) physical positions within the QTL intervals delimited by flanking markers. Of the 51 genes initially identified (Table A3 and Figure A1), only 18 met the expression-based prioritization criteria (Table 3 and Figure 3). For QLr.DFI-2BL, the physical interval contained several defense-related genes, including TraesCS2B01G492800 (BTB/POZ-MATH), TraesCS2B01G491500 (F-box), TraesCS2B01G493200 and TraesCS2B01G493300 (LRR-like protein), and multiple putative NBS-LRR disease resistance proteins (TraesCS2B01G496700, TraesCS2B01G496800, TraesCS2B01G496900, TraesCS2B01G497000, TraesCS2B01G497200). The potential BTB/POZ proteins have been implicated in immune signaling regulation in plants [21]. For QLr.DFI-3BS, the interval harbored TraesCS3B01G002100 and TraesCS3B01G004400 (F-box), TraesCS3B01G005000 and TraesCS3B01G005100 (RLKs), TraesCS3B01G005600 (F-box), TraesCS3B01G005700 (F-box), TraesCS3B01G007000 and TraesCS3B01G007100 (RLKs), and TraesCS3B01G008600 (RLKs). For QLr.DFI-3BL, the QTL detected across all four environments contained TraesCS3B01G352300 (serine/threonine-protein kinase), TraesCS3B01G354000 (peroxidase), TraesCS3B01G354400 (the putative NBS-LRR), TraesCS3B01G354600 (CDPK), TraesCS3B01G355100 (the putative BTB/POZ-MATH), TraesCS3B01G355500 (F-box), TraesCS3B01G355900 (PPO), TraesCS3B01G357400 (ubiquitin-protein ligase), and TraesCS3B01G357500 (ethylene-responsive factor-like transcription factor). The presence of a CDPK-like protein is noteworthy, as CDPKs are key Ca2+ decoders that activate downstream defense responses [22]. For QLr.DFI-7BL, which showed the strongest phenotypic effect, the interval included TraesCS7B01G438200, TraesCS7B01G438300, and TraesCS7B01G438800 (ubiquitin system component Cue proteins), TraesCS7B01G440000 (protein kinase), TraesCS7B01G440800 (the putative TIR-NBS-LRR), TraesCS7B01G441200 (RLKs), TraesCS7B01G441600 (the putative NBS-LRR), TraesCS7B01G441700 (the putative NBS-LRR), TraesCS7B01G442100 (ubiquitin-protein ligase MARCH6), and TraesCS7B01G443800, TraesCS7B01G443900, TraesCS7B01G444000 (MYB transcription factors). For QLr.DFI-7DL, the minor QTL contained TraesCS7D01G426700 (ethylene-responsive transcription factor), TraesCS7D01G427100, TraesCS7D01G427200, TraesCS7D01G427600, TraesCS7D01G430100, TraesCS7D01G430200, TraesCS7D01G432500, TraesCS7D01G432600 (F-box), TraesCS7D01G429500 and TraesCS7D01G429600 (the putative TIR-NBS-LRR), and TraesCS7D01G430600 (the putative BTB/POZ). Collectively, these candidates span multiple layers of plant immunity, from pathogen recognition to downstream defense execution, suggesting that the resistance conferred by these QTLs is genetically complex. To further prioritize the candidate genes within the five stable leaf rust resistance QTL, we examined their expression profiles using the public wheat expression database (wheat-expression.com) and retained those showing leaf-preferential or stress-responsive patterns. This filtering yielded 18 high-confidence genes distributed across all five QTL intervals. In QLr.DFI-2BL, TraesCS2B01G491400 encoding PPO was notably expressed. The QLr.DFI-3BS interval contained three differentially expressed receptor-like kinases (TraesCS3B01G005000, TraesCS3B01G005100, TraesCS3B01G008600). The QLr.DFI-3BL QTL was consistently detected across all environments and harbored six expressed candidates: TraesCS3B01G352300, TraesCS3B01G354000, TraesCS3B01G354400, TraesCS3B01G355500, TraesCS3B01G355900, and TraesCS3B01G357400. For the major-effect QTL QLr.DFI-7BL, six highly expressed genes were identified: TraesCS7B01G438200 and TraesCS7B01G438800, TraesCS7B01G440000, TraesCS7B01G441600 and TraesCS7B01G441700, and TraesCS7B01G442100. In QLr.DFI-7DL, the expressed candidates included TraesCS7D01G426700 and TraesCS7D01G429600. 2.4. Development and Validation of KASP Markers Five KASP markers were designed from the flanking sequences of the five stable QTLs (Table 4 and Table A4; Figure 4). These markers were genotyped in a natural validation panel comprising 120 diverse wheat varieties from the Huang-Huai region. The association between marker genotypes and leaf rust severity, measured by MDS, was evaluated across multiple environments. For K-LR-3BS, developed from QLr.DFI-3BS, varieties carrying the GG genotype (n = 59) exhibited significantly lower mean MDS of 32.8%, compared with 42.1% in those lines carrying the AA genotype (n = 47) (p < 0.05), confirming that the GG allele represents the resistant haplotype at this locus, consistent with the genetic effect observed in the mapping population (Table A5). Similarly, the marker K-LR_7BL developed from QLr.DFI-7BL on chromosome 7B revealed that varieties with the AA genotype (n = 52) displayed significantly reduced disease severity (33.7%) relative to those with the GG genotype (n = 45; 42.5%; p < 0.05), supporting the resistance-enhancing role of the AA allele, which was derived from the resistant parent Hengmai28. In addition, For K-LR-3BS, among the 120 accessions, 14 were heterozygous (AG), and 0 had missing data (failed calls); for K-LR-7BL, 16 were heterozygous (AG) and 7 had missing data. The call rates were 100% for K-LR-3BS and 94.2% for K-LR-7BL. For the remaining three markers: K-LR_2BL for QLr.DFI-2BL successfully genotyped the natural population but showed no significant association with MDS (p > 0.05); K-LR_3BL for QLr.DFI-3BL successfully genotyped the population but lacked significant association; and K-LR_7DL for QLr.DFI-7DL failed to produce reliable genotyping in the natural population. These results indicate that K-LR-3BS and K-LR-7BL are significantly associated with reduced leaf rust severity in the tested natural panel and represent promising markers for MAS. Given that only five targeted KASP markers were tested, no correction for multiple testing (e.g., Bonferroni or FDR) was applied. It should be noted that these associations were derived from a natural panel genotyped only for these five markers; therefore, population structure and relatedness could not be accounted for, and the results should be considered preliminary until confirmed with genome-wide marker data. 3. Discussion Since the 1960s, CIMMYT has been engaged in breeding for APR and has successfully developed a set of multi-resistant varieties, such as Chapio, Cook, Pavon 76, Sonoita 81, and Tukuru, which have been widely adopted in developing countries [8]. In 2005, CIMMYT pyramided four to five adult plant resistance genes into single wheat varieties, generating APR materials with high levels of durable resistance. In recent years, developing APR varieties has become a major focus of international disease resistance breeding efforts [8]. Recently, rapid advances in genomics and molecular breeding technologies have provided powerful genetic tools for wheat breeders to dissect the genetic architecture of rust resistance [23]. QTL mapping, meta-analysis, and transcriptome profiling are now integrated to identify stable, environmentally robust loci [24]. In this context, mining resistance alleles from elite germplasm, developing functional markers, and applying marker-assisted or genomic selection to engineer novel germplasm are critical for advancing wheat breeding for enhanced leaf rust resistance. In this study, although the phenotypic correlation coefficients among environments were moderately to highly positive (r = 0.60–0.70), the Hb was only 0.25, which is considered low to moderate. ANOVA revealed that the genotype main effect was significant (F = 1.33, p = 0.001) when tested against the G×E interaction, but this significance was substantially lower than when tested against the residual error, highlighting that much of the apparent genotypic differentiation was confounded by genotype-by-environment interactions. The G × E interaction itself was highly significant (F = 12.12, p < 0.001), indicating that genotype performance varied considerably across environments. Consistently, the estimated variance components showed that the G × E interaction (σ2ge = 241.37) accounted for 73.6% of the total phenotypic variance, while genotypic variance (σ2g = 21.39) contributed only about 6.5%. These results suggest that although genuine genetic differentiation exists among the RILs, the resistance in this population is strongly influenced by environmental fluctuations and G × E interactions. It should be noted that our experimental design (ANOVA) (Icimapping V4.1) does not separate additive from non-additive genetic effects; therefore, the reported genotypic variance should be interpreted as total heritable variation. In this study, we identified five APR QTLs for leaf rust resistance using an RIL population derived from Liang66-S × Hengmai28, which were evaluated across four environments. Among them, QLr.DFI-3BS and QLr.DFI-7BL consistently reduced disease seve rity in the RILs and were validated in an independent natural population using newly developed KASP markers K-LR-3BS and K-LR_7BL. These two QTLs, together with the three additional loci, provide valuable genetic resources for understanding the molecular basis of durable resistance. 3.1. Comparison with Previously Reported QTLs QLr.DFI-2BL is located in a genomic region frequently associated with APR to leaf rust. Several independent studies have confirmed the presence of major QTLs for leaf rust APR on 2BL (Table A6). Zhang et al. [25] mapped a major APR QTL, QLr.cau-2BL, on 2BL between IWB3854 (531.14 Mb) and IWB21922 (616.48 Mb), and developed a diagnostic SSR marker (Ta2BL_ssr7) for MAS breeding. Buerstmayr et al. [26] also detected QLr.ifa-2BL (around 690.0 Mb) conferring both leaf rust and stripe rust resistance in the Austrian winter wheat cultivar Capo. Thus, QLr.DFI-2BL is a genuine and validated APR QTL, likely representing the same or a closely linked locus to previously reported 2BL QTLs. QLr.DFI-3BS is among the most consistently reported genomic regions for adult plant leaf rust resistance. AC Taber, a hard red spring wheat cultivar with long-lasting leaf rust resistance, was found to carry a QTL (around 33.1 Mb) on chromosome 3BS that maps to the same region as the adult plant resistance gene Lr74 (tightly linked with cfb5006 and IWA6651, around 10–35 Mb) [27]. Furthermore, co-localized QTLs (5.0–13.5 Mb) for leaf rust and stripe rust on 3BS were also identified [27], indicating either closely linked genes or pleiotropic gene action. QLr.DFI-3BS is therefore judged as a genuine, validated QTL. Notably, its explained PVE is consistent with the moderate-effect QTLs commonly reported for this chromosome arm. QLr.DFI-3BL demonstrated the greatest environmental stability among all identified QTLs, being detected across all four environments. This stability suggests that this locus contributes to constitutive, broad-spectrum resistance. A novel leaf rust resistance QTLs (around 760 Mb) on 3BL in a wheat doubled haploid population, with the 3BL QTL conferring seedling resistance and different from the QLr.DFI-3BL identified in this study [28]. QLr.DFI-7BL is the most robust QTL in this study. The 7BL chromosome is a well-characterized genomic region harboring multiple rust resistance genes. The known slow-rusting gene Lr68 (Xgwm146, csSNP856, Xbarc32, around 640–705 Mb) resides on 7BL and has been widely deployed in wheat breeding programs worldwide [9]. Lr68 were mapped on chromosome 7B and developed tightly linked markers for MAS breeding [29]. Additionally, QLr.usw-7BL (nearly with QLr.csiro-7B at 730.1 Mb) was detected on the 7BL in the durum wheat Arnacoris, accounting for up to 65.9% of the disease severity variance [30]. CI13227, a wheat accession with slow-rusting resistance, was also reported to carry a QTL on 7BL (around 640 Mb) explaining 6.8% to 28.5% of the phenotypic variance [31]. The 7BL QTL has also been shown to confer simultaneous resistance to both leaf rust and stripe rust [26]. Given its strong linkage to established resistance loci and high LOD value, QLr.DFI-7BL is considered a genuine, potentially major APR QTL. Finally, QLr.DFI-7DL resides in a genomic region known to carry important rust resistance genes. The long arm of chromosome 7D harbors the well-characterized leaf rust resistance gene Lr19 (XsdauK3734 and XsdauK2839 around 650–712.0 Mb), which was originally transferred from Thinopyrum ponticum and has been widely used in wheat breeding [32]. Although QLr.DFI-7DL is detected as a minor QTL in this study, its chromosomal location on 7DL supports its authenticity. Further fine-mapping will be necessary to determine whether QLr.DFI-7DL corresponds to Lr19 or represents a different resistance locus on 7DL (Table A3). Among the identified QTLs, QLr.DFI-2BL, QLr.DFI-3BS, and QLr.DFI-7BL are considered previously reported and validated loci as they consistently map to genomic regions frequently associated with APR to leaf rust and have been documented in multiple independent studies. In contrast, QLr.DFI-3BL and QLr.DFI-7DL are regarded as new: QLr.DFI-3BL was reported as a novel QTL [28], while QLr.DFI-7DL, though residing on a chromosome arm known to carry Lr19, requires further fine-mapping to determine its relationship to known resistance genes and is detected as a minor QTL in this study, suggesting it may represent a distinct or uncharacterized locus. In summary, the most robust QTLs are QLr.DFI-3BL and QLr.DFI-7BL, which should be prioritized for further fine-mapping and potential MAS breeding. 3.2. Candidate Gene Functions and Resistance Mechanisms In this study, we identified 18 high-confidence candidate genes within the five QTL intervals, encompassing classical resistance genes, signaling components, and metabolic enzymes involved in diverse immune pathways. Canonical NBS-LRR genes were found in QLr.DFI-3BS, QLr.DFI-7BL, and QLr.DFI-2BL, consistent with the fact that most major R genes encode putative NBS-LRR proteins [7]. Notably, QLr.DFI-7BL contains two adjacent NBS-LRR genes (TraesCS7B01G441600 and TraesCS7B01G441700), reminiscent of paired NLRs recently shown to function cooperatively in wheat stripe rust resistance [33]. Beyond canonical R genes, we identified several regulatory and metabolic candidates. A CDPK gene (TraesCS3B01G354600) within QLr.DFI-3BL encodes a key Ca2+ decoder that activates downstream defense responses, and multiple wheat CDPKs are induced by leaf rust infection [22,24]. BTB/POZ-MATH domain proteins, involved in ubiquitin-proteasome regulation and R-protein stability [34], were detected in three independent QTLs (QLr.DFI-2BL, -3BL, and -7DL), while F-box proteins, also part of the ubiquitin ligase complex, appeared in multiple intervals (QLr.DFI-2BL, -3BS, -3BL, -7DL), further underscoring the importance of ubiquitination in tuning defense signaling [21,35]. Oxidative-burst enzymes—polyphenol oxidases (PPO), peroxidases, and germin-like proteins (GLP), were located in specific QTLs (TraesCS2B01G491400 in QLr.DFI-2BL; TraesCS3B01G354000 peroxidase and TraesCS3B01G355900 PPO of QLr.DFI-3BL; TraesCS7B01G442000 GLP in QLr.DFI-7BL), and their leaf-biased expression supports a role in reactive oxygen species generation and cell wall reinforcement against biotrophic fungi, including Puccinia species [36,37]. In addition, multiple LRR-RLKs (e.g., TraesCS2B01G493200, -493300, TraesCS3B01G007100) and ethylene-responsive factors (ERFs) suggest that both PAMP-triggered immunity (PTI) and hormone signaling are integrated into the resistance network defined by these QTLs [24]. Overall, the 18 prioritized candidates, which show significant expression in leaf tissues or under stress conditions, provide strong functional relevance to the leaf rust resistance conferred by these QTLs. The convergence of expression evidence with known immune functions justifies future functional validation through virus-induced gene silencing (VIGS) or genome editing, with particular emphasis on the paired NBS-LRR genes in QLr.DFI-7BL and the multi-component signaling network in QLr.DFI-3BL, to elucidate their precise roles and facilitate marker-assisted breeding for durable leaf rust resistance in wheat. 3.3. Limitations and Future Perspectives Despite the identification of five QTLs and the validation of two KASP markers, several limitations of this study should be acknowledged, including the moderate population size and marker density that resulted in large confidence intervals, the lack of functional characterization for candidate genes, and the limited environmental and population breadth for marker validation. Critically, because the natural diversity panel was genotyped exclusively for the five target KASP markers derived from previous QTL mapping, we were unable to correct for population stratification or familial relatedness, which may confound the observed marker–trait associations. Therefore, while K-LR-3BS and K-LR-7BL showed statistically significant associations, they should be regarded as preliminarily validated rather than fully diagnostic. The stability of these associations should be further tested in broader germplasm sets using genome-wide markers with proper correction for population structure. To address these genetic and mechanistic gaps, high-resolution fine mapping of QLr.DFI-3BS and QLr.DFI-7BL using large segregating populations is required to narrow candidate intervals to a few hundred kilobases, followed by positional cloning coupled with transcriptomic profiling and gene silencing/overexpression in wheat to identify and validate causal genes. The recently available wheat pan-genome resources and multi-omics approaches will greatly accelerate the cloning of adult plant resistance genes [23,38,39,40]. Ultimately, integrating validated diagnostic markers for these QTLs into genomic selection models, combined with marker-assisted pyramiding of other cloned APR genes (e.g., Lr34, Lr46, Lr68, and Lr74), will enable the development of wheat cultivars with durable, broad-spectrum leaf rust resistance in the Huang-Huai region and beyond. 3.4. Perspectives for Adult-Plant Leaf Rust Resistance Breeding The two validated KASP markers provide immediate, cost-effective tools for integrating leaf rust resistance into winter wheat varieties adapted to the Huang-Huai region. Their direct application in MAS enables breeders to efficiently select progenies carrying favorable alleles without time-consuming field phenotyping. Critically, the resistance alleles at QLr.DFI-3BS and QLr.DFI-7BL are derived from the resistant parents used in our RIL population, which can serve as valuable donor germplasm for crossing programs. By crossing these donors with elite but susceptible cultivars, followed by marker-assisted backcrossing, breeders can rapidly transfer the QTLs into adapted backgrounds while retaining agronomic performance. For durable resistance, the validated markers should be multiplexed with other known resistance genes, particularly the widely deployed adult-plant genes Lr34 and Lr46, as additive or synergistic effects have been reported [8,12]. Moreover, the integration of K-LR-3BS and K-LR_7BL into a low-density, breeder-friendly genotyping panel would allow routine screening of early-generation populations, enabling the selection of resistant individuals before field evaluation and significantly reducing breeding cycle time. Beyond immediate MAS applications, the two QTL regions offer promising targets for positional cloning and candidate gene validation; the validated markers can serve as anchor points for fine-mapping and for developing more tightly linked markers to further enhance selection precision. Collectively, our findings not only enrich the genetic resources for leaf rust resistance but also provide actionable tools to accelerate the development of resistant cultivars, thereby contributing to sustainable wheat production in the face of increasing disease pressure. 4. Materials and Methods 4.1. Plant Materials and Phenotypic Identification The parental line Liang66-S exhibited adult-plant leaf rust responses varying between moderate susceptibility and moderate resistance, whereas Hengmai28 displayed levels ranging from moderate to high resistance. A recombinant inbred line (RIL) population consisting of 253 F2:8 lines, derived from a cross between Liang66-S and Hengmai28, was used for field experiments. This population was grown during the 2021–2022 and 2022–2023 growing seasons at two locations: Xinxiang (35°18′ N, 113°52′ E, Henan Province) and Baoding (38°52′ N, 115°28′ E, Hebei Province). The field trial employed a randomized complete block design with three replications, and each plot row was 3 m long. The susceptible cultivar Zhengzhou 5389 served as both the disease control and the inoculum spreader; one spreader row was planted after every ten rows of test material to guarantee uniform inoculation. To validate the genetic effects of the leaf rust resistance quantitative trait loci (QTLs) identified, a natural panel comprising 120 wheat accessions originating from the Huang-Huai wheat region was employed. This validation panel was planted during the 2020–2021 and 2021–2022 growing seasons at Baoding (38°52′ N, 115°28′ E, Hebei Province) and Zhengzhou (35°44′ N, 113°42′ E, Henan Province), also following a randomized complete block design with three replicates, a row length of 2 m, and a row spacing of 20 cm. Xinxiang and Zhengzhou in Henan, together with Baoding in Hebei, are endemic regions for wheat leaf rust, and the predominant, prevailing physiological races in these areas are THTT, PHTT, and THTS [41]. At the wheat jointing stage, urediniospores of the three races were mixed in equal proportions and suspended in 0.03% Tween-20 solution for spray inoculation. Approximately four weeks after inoculation, when the leaves of the susceptible check Zhengzhou 5389 were nearly fully covered with uredinia, disease severity assessments were conducted on the mapping population. Field evaluations were performed weekly, recording the leaf rust severity of both parents and all progeny lines, using the percentage of leaf area covered by uredinia as the scoring criterion. In each environment, the disease assessment was repeated two to three times, and the MDS was retained as the phenotypic data for subsequent linkage analysis [42]. 4.2. Statistical Analysis and Linkage Mapping Analysis of variance (ANOVA) was performed using IciMapping v4.2 (http://www.isbreeding.net/, accessed on 25 December 2025) [43], and the Pearson correlation coefficient was used to evaluate the correlation of maximum disease severity across different environments. Hb2 was estimated as Hb2 = σ2g/σ2p, where σ2g is the genotypic variance and σ2p is the phenotypic variance (σ2p = σ2g + σ2ge + σ2e), with σ2ge and σ2e representing genotype-by-environment interaction variance and residual (environmental) variance, respectively. Variance components were estimated from the ANOVA mean squares using the expected mean square method. For the RIL population evaluated across four environments (e = 4) with three replicates per environment (r = 3), the components were derived as follows: σ2g = (MSg − MSge)/(e × r) = (1045.9 − 789.2)/12 = 21.39, σ2ge = (MSge − MSe)/r = (789.2 − 65.1)/3 = 241.37, σ2e = MSe = 65.1. Hb on a line-mean basis was calculated as: Hb = σ2g/(σ2g + σ2ge/e + σ2e/(e × r)) = 21.39/(21.39 + 60.34 + 5.43) ≈ 0.25. All calculations are based on the variance components derived from the corrected ANOVA (Table A2). Genotyping of Liang66-S, Hengmai28, and all RIL lines was performed using the wheat 90K SNP array [44]. The genetic map was constructed [45], with 6526 high-quality polymorphic markers. The inclusive composite interval mapping (ICIM) method implemented in IciMapping v4.2 was used to perform QTL mapping for maximum disease severity of the Liang66-S/Hengmai28 RIL population across four environments. This method combines the advantages of interval mapping and multiple regression analysis. QTLs were mapped separately for each environment using the ICIM method, and the LOD and R2 values are presented in Table 2. A 1000-permutation test was performed to determine the significance threshold (p = 0.05), with an LOD threshold of 2.6 and a step size of 1 cM, to generate LOD curves and identify QTL positions. In this study, a QTL detected in two or more environments was defined as a stable QTL. In this study, “DFI” denotes the Dryland Farming Institute, Hebei Academy of Agriculture and Forestry Sciences, and our QTL naming follows the standard wheat convention (QTL + trait abbreviation + “DFI” + chromosome). 4.3. Development of KASP Markers KASP primers were designed using the flanking and internal markers of the identified QTLs for leaf rust resistance. Primer design was carried out via the PolyMarker web interface (http://www.polymarker.info/, accessed on 15 April 2026) [46], and all primer sequences are provided in Table A4. Each KASP primer set comprises two allele-specific competing primers and a single common primer, with only one nucleotide difference at the 3′ end of the two competing primers. After selecting locus-specific or semi-specific primers, the 5′ ends of the two competing primers were separately tagged with FAM and HEX sequences. All primers were synthesized by Sangon Biotech Co., Ltd. (Shanghai, China). The KASP PCR reaction mixture contained: 2.0 µL of 2 × KASP Master Mix, 0.0336 µL of mixed KASP primers, and 2 µL of template DNA (50 ng/µL). Amplification was conducted using a water-bath PCR system from LGC Limited (UK). The thermal cycling conditions were: one cycle of pre-denaturation at 94 °C for 15 min; followed by 10 touchdown cycles consisting of denaturation at 94 °C for 20 s and annealing for 1 min per cycle, with the annealing temperature decreasing from 65 °C to 57 °C; and finally 35 cycles of denaturation at 94 °C for 20 s and annealing at 57 °C for 1 min. Following amplification, fluorescence intensity was measured using a PHEstar plate reader (BMG Labtech GmbH, Germany) with the 384-well KASP scanning program. Genotyping data were analyzed using KLUSTER v1.0 (LGC, London, UK). A two-tailed Welch’s t-test to compare the mean MDS between the two homozygous genotype classes (e.g., AA vs. GG) for each marker. For the natural panel, the MDS used for association is the average MDS across the two evaluation years (2020–2021 and 2021–2022) at both Baoding and Zhengzhou. Heterozygous calls (e.g., AG) were not included in the t-test comparisons between the two homozygous classes because their phenotypic expectation is ambiguous. The number of heterozygous accessions is reported alongside the call rate for complete transparency. Because only the five KASP markers were genotyped in this natural panel and genome-wide SNP profiles were not obtained, formal correction for population structure and relatedness could not be performed. For the five KASP markers tested, no adjustment for multiple comparisons was performed as these markers were chosen as targeted candidates from prior QTL mapping rather than as a genome-wide screen. 4.4. Candidate Gene Prediction Candidate genes were identified based on IWGSC v1.0 [47]. Candidate genes were selected based on the physical interval defined by the flanking markers of each QTL, using the Chinese Spring reference genome (IWGSC v1.0). The criteria for candidate gene selection followed: (i) high-confidence genes located within the mapped QTL interval based on IWGSC v1.0 annotation; (ii) genes with protein domains/functional annotations consistent with known disease resistance genes (e.g., NBS-LRR, RLK, CDPK, BTB/POZ); (iii) genes showing leaf-preferred expression and/or stress-responsive expression patterns based on public database (wheat-expression.com) [48,49], focusing on expression patterns across different developmental stages and under stress conditions relevant to disease response. 5. Conclusions Wheat leaf rust seriously threatens wheat production and causes significant losses to both yield and quality. Breeding resistant cultivars is currently the most economically and environmentally sustainable strategy for leaf rust control. In this study, five APR QTLs for leaf rust were identified in the Liang66-S/Hengmai28 RIL population, and they explained 4.5–9.8% of the PVE. Two developed KASP markers (K-LR-3BS and K-LR_7BL) were validated in a panel of 120 Chinese wheat cultivars and confirmed to be significantly associated with leaf rust resistance, whereas the remaining three markers, K-LR_2BL for QLr.DFI-2BL, K-LR_3BL for QLr.DFI-3BL, and K-LR_7DL for QLr.DFI-7DL, either showed no significant association with MDS (p > 0.05) or failed to produce reliable genotyping in the natural population. Both K-LR-3BS and K-LR-7BL are promising candidates for leaf rust resistance, but their routine application in MAS awaits independent validation in diverse genetic backgrounds and with proper control for population structure. Furthermore, multiple candidate genes involved in various immune pathways were identified based on public gene expression data, including NBS-LRR, CDPK, BTB/POZ, F-box, PPO, peroxidase, and ERF. The functional QTLs, breeding-applicable KASP markers, and candidate gene resources obtained in this study provide important support for molecular marker-assisted selection for wheat leaf rust resistance breeding.