Genetic dissection and identification of stripe rust resistance genes in the wheat cultivar Lanhangxuan 121, a cultivar selected from a space mutation population

Abstract

Stripe rust is a devastating disease of wheat worldwide. Chinese wheat cultivar Lanhangxuan 121 (LHX121), selected from an advanced line L92-47 population that had been subjected to space mutation breeding displayed a consistently higher level of resistance to stipe rust than its parent in multiple field environments. The aim of this research was to establish the number and types of resistance genes in parental lines L92-47 and LHX121 using separate segregating populations. The first population developed from a cross between LHX121 and susceptible cultivar Xinong 822 comprised 278 F_2:3 lines. The second validation population comprised 301 F_2:3 lines from a cross between L92-47 and susceptible cultivar Xinong 979. Lines of two population were evaluated for stripe rust response at three sites during the 2018–2020 cropping season. Affymetrix 660 K SNP arrays were used to genotype the lines and parents. Inclusive composite interval mapping detected QTL QYrLHX.nwafu-2BS, QYrLHX.nwafu-3BS, and QYrLHX.nwafu-5BS for resistance in all three environments. Based on previous studies and pedigree information, QYrLHX.nwafu-2BS and QYrLHX.nwafu-3BS were likely to be Yr27 and Yr30 that are present in the L92-47 parent. QYrLHX.nwafu-5BS (YrL121) detected only in LHX121 was mapped to a 7.60 cM interval and explained 10.67–22.57% of the phenotypic variation. Compared to stripe rust resistance genes previously mapped to chromosome 5B, YrL121 might be a new adult plant resistance QTL. Furthermore, there were a number of variations signals using 35 K SNP array and differentially expressed genes using RNA-seq between L92-47 and LHX121 in the YrL121 region, indicating that they probably impair the presence and/or function of YrL121.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11032-024-01461-0.

Introduction

Stripe rust or yellow rust (YR), caused by Puccinia striiformis f. sp. tritici (Pst), is one of the most devastating wheat diseases. Yield losses caused by stripe rust in individual crops can vary between 5 and 50%, and in some cases, even 100% in highly susceptible cultivars (Chen 2005). Although traditionally occurring on wheat in cooler and wetter regions, including Asia and Europe (Brown and Hovmøller 2002) new aggressive races apparently adapted to warmer climates have become established after year 2000 in regions where stripe rust had never been present (Wellings 2011; Chen et al. 2014). A consequence of these variation means that stripe rust is now a worldwide problem. Strategies for reducing the effects of wheat rusts are predominantly host resistance and chemical fungicides (Chen 2014). Fungicides increase production costs, pose human and environmental risk, and like single resistance genes, may lose effectiveness by evolutionary changes in pathogen populations. Generally, resistance type is categorized as seedling or all-stage resistance (ASR) and adult-plant resistance (APR) or high-temperature adult-plant (HTAP) resistance (Chen 2013). ASR is broadly applied in wheat breeding, which is race-specific and easily loss. The breakdown of major resistance genes shortly after their use in wheat production has resulted in a decline in resistance gene numbers and a rethinking of strategies to prolong the duration of effectiveness of individual resistance genes (McIntosh et al. 2018).

Rust pathogens have a higher relative fitness due to their rich genetic diversities. This feature of stripe rust is highly associated with its rapid variations, including known mutation and somatic exchange patterns, as the newly found sexual recombination (Chen and Kang 2017). More recently, Berberis and Mahonia identified as alternate hosts of Pst to complete its life cycle (Jin et al. 2010; Wang and Chen 2013; Zhao et al. 2013, 2016). These findings emphasized the need for more breeding efforts of resistant varieties to increase resistance gene diversity and reinforce integrated management practices to reduce the evolvability of pathogens and overcome the stripe rust epidemic worldwide.

The International Maize and Wheat Improvement Center (CIMMYT) is an advanced institution in the world for publicly-funded research on maize and wheat, as well as the associated agricultural systems. CIMMYT-derived germplasms are widely used worldwide because of their high yield potential, broad adaptation, biotic and abiotic stress resistance (Lantican et al. 2016). The collaboration between China and CIMMYT in the field of wheat improvement can be dated back to the 1970s. This involved the introduction and use of germplasm, the commencement of shuttle breeding, and the creation of germplasms that exhibit APR (He et al. 2018). Thousands of CIMMYT advanced lines are imported to China from international nurseries and breeder selection each year. After testing the adaptation of these lines in key sites in China, some are directly used in production, and others served as parents in breeding programs (He and Xia 2016). To date, over 330 CIMMYT-derived cultivars have been released in China, including many famous cultivars [e.g., Chuanmai (42, 82, 86), Emai 18, Han 6172, Jinan 17, Jinmai 19, Kefeng 3, Ningchun 4 and some prefixed as Xinchun (2, 6)] and are widely used in various regions (He et al. 2019). CIMMYT wheat has been essential in improving stripe rust resistance in the history of Chinese wheat breeding. Additionally, its resistance gene pool includes several well-known Yr genes such as Yr17 + Lr37 + Sr38, Yr18, Yr29, Yr30, Yr46, Yr54, and Yr78 (Rosewarne et al. 2013; Basnet et al. 2014; Huerta-Espino et al. 2019) and some stable quantitative trait loci (QTL) on chromosomes 2BS (QYr.nwafu-2BS, Yr27 = YrKK) (Li et al. 2013; Wu et al. 2017b), 4BL (QYr.nwafu-4BL, YrRC) (Wu et al. 2018), and 7BL (QYr.nwafu-7BL = QYratt.csiro-7BL/QYrchi.cim-7BL/QYrpas.cim-7BL, YrKB) (Rosewarne et al. 2008; Ponce-Molina et al. 2018; Huang et al. 2019). Some of these genes have been used in Chinese wheat resistance breeding programs (He et al. 2011; Han and Kang 2018). In recent decades ‘space-flight mutation breeding’ has become a new line of enquiry and many new cultivars/strains of plants (including 100 wheat cultivars) (Lv et al. 2016; Zhang et al. 2019) and microorganisms were developed from populations that had been subjected to space flights. Additionally, this technique has high mutagenesis frequency, vast variation extent, short breeding period, and more elite mutants than traditional breeding. It also accelerated the development of Chinese wheat breeding with outstanding achievements. Most of new mutation cultivars were directly selected from space mutagenesis progenies or indirectly from the offspring of cross-breeding by space-induced. For example, Prof. Jiuyuan Du and his colleagues have engaged in wheat space mutagenesis for many years and brought several common wheat cultivars into space mutation projects (Du et al. 2012). After the return flight, they performed a systematic field evaluation of mutant progenies of wheat for stripe rust resistance, including stripe rust resistance variations and tracking their resistance stability by pedigree selection (Du et al. 2012). Subsequently, various space mutant-derived lines with different resistance levels were screened, further proving that they conferred stable APR when tested with different Pst races (Lv et al. 2016).

Few studies have attempted to determine the nature of advantageous genetic changes in cultivars produced by space breeding (or indeed by more conventional mutation breeding, compared to the pre-space-exposed genotypes). Lanhangxuan 121 (LHX121) is a commercial cultivar developed from space-mutagenized line L92-47 (pedigree: (CIMMYT-derived line ZHF1/Shaannong 7859) F₄ Seln. // Lantian 3). Although L92-47 has considerable “slow rusting” resistance, LHX121 has showed a consistently higher level (Lv et al. 2016). Now the question was whether this additional resistance was novel and caused by the exposure to radiation in space. A comprehensive dissection of resistance in both lines was undertaken to address these questions.

Materials and methods

Plant materials

Stripe rust resistant parents employed in this study were Lanhangxuan 121 (LHX121) and Line 92–47 (L92-47), whereas susceptible lines Xinong 822 (XN822) and Xinong 979 (XN979), served as the other parents. MingXian169 (MX169), a Chinese winter wheat landrace, Avocet S (AvS), and Xiaoyan 22 (XY22) served as controls. Two segregating populations were used to map and compare genes or QTL associated with stripe rust resistance. The first population involved 278 F_2:3 lines from the cross XN822 × LHX121 and the second cross comprised 301 F_2:3 lines from cross XN979 × L92-47. 55 wheat cultivar and nine recombinant inbred lines (RILs) from cross MX169 x Centrum were used for phylogenetic analysis by 35 K single nucleotide polymorphism (SNP) array.

Pathogen materials

Preliminary tests on seedlings showed that both L92-47 and LHX121 were susceptible to currently prevalent Pst races in China, including isolates of CYR29, CYR31, CYR32, CYR33, CYR34, V26/SX, V26/SC, and V26/GS (Fig. 1b, d). More information of these races was listed in Wu et al. (2020). For field trials aimed at addressing adult plant response plots at Yangling were inoculated at flag leaf emergence with a urediniospore:talc (1:10) mixture of races CYR32, CYR33 and CYR34. Trials grown at Tianshui in Gansu province were inoculated under natural infection conditions, as Tianshui is in Pst over-summering region.

Fig. 1.

a, c Adult plant stripe rust reactions of Lanhangxuan121 (LHX121), Xinong822 (XN822), L92-47, and Xinong979 (XN979). b, d Seedling reactions of LHX121 (b) and XN822 (d) to seven Pst races

Field trial

Both segregating populations and parents were grown at Yangling in Shaanxi province and Tianshui in Gansu during cropping seasons 2018–2019 and 2019–2020. The trials were arranged in randomized complete blocks with two replications. Each plot was a single 1 m sown with ~ 25 seeds and 30 cm between rows. The parents and a susceptible control XY22 were sown after every 20 rows. Susceptible control MX169, was planted in blocks of 5 rows every 50 rows and surrounding the field site. Stripe rust responses were visually assessed for infection types (IT) using a 0 (resistant) to 9 (susceptible) scale (Line and Qayoum 1992) and disease severity (DS, percentage leaf area affected) based on the modified Cobb scale with two observations made for each plot.IT and DS assessed when disease levels on controls MX169 and AvS reached the highest levels; 13–21 May at Yangling and 10–16 June at Tianshui. Within plots an average of two or more observations was used for analysis, and a single value was used to record the responses of non-segregating lines.

Statistical analysis

The values of IT and DS from each environment and the mean values across replicates each line were used in analysis of variance (ANOVA) and QTL mapping. The effects of genotypes, environments, and genotype-environment interaction were investigated through the implementation of ANOVA utilizing both IT and DS. The "AOV" function in software QTL IciMapping 4.1 with default parameters (Meng et al. 2015) was used to calculate Pearson's correlation coefficients (r) and best linear unbiased evaluations (BLUEs). BLUEs were estimated using genotype and environment data for each line as fixed effects and were used to evaluate genetic effects and to find likely QTL positions.

Broad-sense heritability (h2 b) of resistance to stripe rust was estimated with the following equation: h2 b = σ2 g/(σ2 g + σ2 ge/e + σ2 e/re); σ2 g, σ2 ge, and σ2 e were determined as follows: σ2 g = (MS_f – MS_fe)/re, σ2 ge = (MS_fe – MS_e)/r, and σ2 e = MS_e. In the equations, σ2 g represents genetic variance, σ2 ge represents genotype-environment interaction variance, and σ2 erepresents error variance. The values of r and e represent the numbers of replications and environments, respectively.

Bulk segregant and 660 K SNP array analyses

Genomic deoxyribonucleic acid (DNA) was isolated from leaves of each F₂ plant and the parents (Liu et al. 2021). DNA quality and quantity was verified by 1% agarose gel electrophoresis and a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA). Each line of two segregating populations and respective parents were genotyped by AQP™ (Allele-Specific Quantitative PCR based genotyping assay) markers. To identify the target QTL regions, we used Bulk Segregant Analysis (BSA) combined with 660 K SNPs arrays. Equal amount of DNA from 10 extremely resistant F₂ plants (IT 1–2, DS ≤ 10) and 10 extremely susceptible F₂ plants (IT 8–9, DS ≥ 90) along with DNA of the respective parents were used to generate resistant and susceptible pools for BSA (Michelmore et al. 1991). Genotypes of the F₂ pools and parents were identified by 660 K SNP array (CapitalBio Corporation, Beijing; http://www.capitalbio.com). Using software GenomeStudio to call and cluster the SNP genotypes by the polyploid version of Illumina (http://www.illumina.com). To refine the SNP dataset, we used several filtering criteria; specifically, SNPs that were monomorphic and had call rates below 85%, with ambiguous calls, and minor allele frequencies below 5%, were removed. Polymorphic SNPs that linked to resistance in BSA contained homozygous and heterozygous genotypes. However, only homozygous genotypic differences were mapped to chromosomes using the high-density 660 K map, and the number of SNPs per centimorgan (cM) was counted.

AQP marker assays and genotyping

Once polymorphic SNPs were localized on chromosomes, the SNPs within each probable target region were transformed into AQP™ (Allele-Specific Quantitative PCR based genotyping assay) (Liu et al. 2022) markers through the PolyMarker website (Ramirez-Gonzalez et al. 2015). Prior to genotyping the population, parents and bulks were screened using specific-chromosome AQP markers to verify polymorphism. The LGC Genomics protocol was adhered to while conducting AQP assays in 384-well plates. Reaction mixtures were 5 µL, comprising 2.5 µL of genomic DNA (50–100 ng), 2.5 µL of 2 × AQP master mix (V4.0, LGC Genomics), and 0.014 µL of primer mix—a combination of 12 µM of each allele-specific primer and 30 µM of common primer. The conditions of the Veriti 384 well Thermal Cycler (Applied Biosystems) in our experiment was as follows: cycling protocol an initial denaturation stage at 95 °C for 15 min, followed by nine cycles of denaturation at 95 °C for 20 s. A touchdown protocol was subsequently initiated at 65 °C for 60 s, with temperature decreasing by 0.8 °C per cycle. This was followed by 30–40 cycles of amplification at 95 °C for 20 s and 57 °C for 60 s. A microplate reader (FLUOstar Omega, BMG LABTECH, Germany) was used to visualize the end-point fluorescence data, which were subsequently analyzed using Kluster Caller software (LGC, Middlesex, UK).

Data analysis and genetic linkage map

Χ² tests were used to confirm that 1:1 segregation of individual SNP markers (p > 0.05); markers with deviating segregation ratios were eliminated from the analysis. Construction of linkage groups was carried out using QTL IciMapping V4.0 software, utilizing the "MAP" function to create the map based on filtered markers (Van Ooijen 2006). The mean DS for F_2:3 in each environment were used in identifying QTL. QTL detection was facilitated by the inclusive composite interval mapping (ICIM) additive tool within IciMapping V4.0. A QTL was considered significant if the logarithm of odds (LOD) score surpassed the predetermined threshold value of LOD = 2.5 (Sun et al. 2013). Linkage maps were subsequently illustrated using Mapchart V2.3 (Voorrips 2002). Due to minor discrepancies observed in the LOD contours of a single QTL peak across environment, QTL sharing flanking markers or overlapping confidence intervals, as identified by different programs, were considered equivalent. The impact of an individual QTL was gauged using the proportion of phenotypic variance explained (PVE).

Genotyping by the 35 K SNP array and comparative genomics

LHX121, L92-47 and a series of lines derived from cross MX169 × Centurm along with other wheat accessions were genotyped by the 35 K SNP array to perform phylogenetic analysis following methods previously described (Liu et al. 2021). For evaluation of the collinearity relationship among the target QTL region, we performed comparative genomics using published common wheat genomes at the website http://wheat.cau.edu.cn/TGT/ (Chen et al. 2020).

RNA-seq

The ribonucleic acid -sequencing (RNA-seq) experiment included mock-inoculated and Pst-inoculated flag leaves of LHX121 and L92-47. Mock (fluorinated solution applied) and inoculation with mixed Pst races CYR32, CYR33 and CYR34 were performed at the heading stage in a greenhouse. Tissues were collected from LHX121 and L92-47 at 48 hpi using three independent biological replicates.

Leaf tissues from three biological replicates were combined in equal amounts and send to BGI Genomics, BGI-SHENZHEN (https://www.genomics.cn/) for RNA isolation, RNA library construction and RNA-seq. Quality control, alignment, counting fragments, TPM (transcripts per million Kb calculation and differential expression of genes (DEGs) analysis for RNA-seq data as referenced in Hiebert et al. (2020) and Yi et al. (2023).

Results

Phenotypic evaluation

LHX121 and L92-47 demonstrated susceptibility at seedling stage but showed resistance in all field experiments, with IT below 3, and mean DS below 10% at Yangling and Tianshui during 2018–2019. While XN822 and XN979 were susceptible with IT of 8 to 9 and mean DS exceeding 90% (Fig. 1a, c). The segregating populations confirmed high levels of APR against stripe rust, with IT varying from 0 to 9 and DS varying from 0 to 100% at both sites during 2018–2019 and 2019–2020 (Table S1, S2). The frequency distribution of the mean IT and DS classes among the F₂ and F_2:3 lines continuously confirmed that the resistance genes were quantitative in nature and were probably governed by multiple genes/QTL (Figs. 2a, b, c, d and S1a, b, c, d). ANOVA of the F_2:3 lines data confirmed significant phenotypic variation in IT and DS among lines, environments, and line-environment interactions. No significant variation was detected among replicates (Table 1). The h2 b of stripe rust disease severity ranged from 0.67 to 0.77 in different environments, and the APR expression was uniform across environments. Pearson’s correlation coefficients for mean IT and DS in the F_2:3 lines among the three or two field environments ranged from 0.68 to 0.97 and were all significant (p < 0.001, Figs. 2e, f and S1e, f).

Fig. 2.

a, b Frequency distributions of mean infection types (IT), disease severities (DS) and BLUEs for cross XN822 × LHX121 evaluated at Yangling and Tianshui. c, d Violin plots of mean infection types (IT) and disease severities (DS) for cross XN822 × LHX121 evaluated at Yangling and Tianshui. e, f Correlation coefficients (r) for mean IT and DS and BLUEs of the cross XN822 × LHX121 across environments. All r values are significant at p < 0.001

Table 1.

Variance components for infection type (IT) and disease severity (DS) scores in the XN822/LHX121 population grown in three environments

Source of variation	IT				DS
	Df	MS	F-value	Pr > F	Df	MS	F-value	Pr > F
Blocks	3	7.31	9.26	0.00E + 00	3	5023.47	29.55	2.04E-05
Genotypes	277	19.96	25.30	0.00E + 00	277	2954.18	17.38	0.00E + 00
Environments	2	33.47	42.42	0.00E + 00	2	8262.36	48.61	1.29E-05
GE_interaction	554	0.97	1.23	0.0037	554	126.80	0.75	0.99
Error	831	0.79			831	169.99
h2 b	0.77				0.67

Mapping QTL for stripe rust resistance and comparisons with the other genes/QTL

We detected more SNPs on chromosomes 2B, 3B, and 5B than other chromosomes in both populations (Fig. 3). There were 6,160 polymorphic markers in the L92-47 population, including 805, 963, and 684 SNPs grouped on chromosomes 2B, 3B, and 5B, respectively. The other SNPs were distributed across the remaining chromosomes (Fig. 3a). In resistant and susceptible pools from cross LHX121, 7,680 polymorphic SNPs were identified, with 838, 1000, and 1,026 SNPs on chromosomes 2B, 3B, and 5B, respectively. The remaining SNPs were scattered across each chromosome (Fig. 3b). Enrichment indicated that most SNPs were in the same intervals of chromosomes 2B (Fig. 3c, f, i) and 3B (Fig. 3d, g, j). However, the concentration of SNPs located in the 0–50 Mb interval of chromosome 5B occurred only in the XN822/LHX121 population (Fig. 3e, h, k). The polymorphic percentage of the SNP among the 21 chromosomes showed us more reliable results of BSA (Fig. 3b), except chromosomes 2B, 3B and 5B, other chromosomes with higher percentage were detected by selected markers and unlocked with phenotype. Following development of specialized primers were based on the BSA analysis using the PolyMarker website, 21 marker pairs were chosen from an original set of 37 pairs. The sequences of the polymorphic AQPmarkers are listed in Table S4.

Fig. 3.

Bulked segregant analysis (BSA) of XN979 × L92-47 (a) and XN822 × LHX121 (b) and integrated single nucleotide polymorphisms (SNPs) (f, g, h, l, j, k). Line graph (b) showing the polymorphic percentage of the SNP among the 21 chromosomes in cross XN822 × LHX121. Venn diagrams showing counts of common and distinctive SNPs between the two populations (c, d, e)

The genetic maps of APR loci on chromosomes 2BS, 3BS, and 5BS were drawn by Joinmap4.0. The length of the linkage maps was 22.5 cM for chromosome 2B, 54.8 cM for chromosome 3B, and 40.0 cM for chromosome 5B. The identified linkage group was used for QTL analysis using mean DS values obtained from the field experiments. QTL QYrLHX.nwafu-2BS and QYrLHX.nwafu-3BS were identified in both populations and covered all environments. The QTL on chromosome 2BS was located in a 6.2 cM interval spanned by AQP markers AX-86174319 and AX-111140388 which covered the region of cloned gene Yr27 (Fig. 4a). QYrLHX.nwafu-3BS was mapped to a 7.5 cM interval between AQP markers AX-109990245 and AQP-TM, and overlapped the region of Yr30 (Fig. 4b). L92-47 is a close relative of CYMMIT derived line ZHF1, as probable a sister line of P10057 (Fig. 5a). APR in P10057 is controlled by QYrLHX.nwafu-2BS (Yr27) and QYrLHX.nwafu-3BS (Yr30) (Wu et al. 2017a). Hence, it was inferred that the APR in L92-47 and LHX121 was also at least in part controlled by Yr27 and Yr30. QYrLHX.nwafu-5BS (hereafter referred to as YrL121) was identified only in the XN822/LHX121 population. The major QTL on chromosome 5BS contributed to a PVE of 10.67–22.57% in infection type and 12.30–13.84% in mean disease severity. This QTL was located on a 7.6 cM interval spanned by AQP markers AX-89459127 and AX-110578940 corresponding to physical interval of ~ 21.18–30.61 Mb (Table 2; Fig. 4c).

Fig. 4.

Genetic linkage map of chromosomes 2B (a), 3B (b), and 5B (c) in LHX121 based on data from F_2:3 lines and QTL (red bar) identified in this study and previously mapped Pst resistance genes and QTL (blue bars) were positioned based on integrated genetic maps provided by Prof. Cui Fa. Reference was based on Chen and Kang 2017

Fig. 5.

Phylogenetic analysis (a), genome-wide of difference comparisons between wild type L92-47 and LHX121 (b) and comparative genomics (c). In the phylogenetic tree, the branch in blue presents the same cultivar but different plants; the branch in orange presents the progenies derived from the cross of Mingxian 169 and Centurm or the founder parent Zhou 8425B; the branch in red presents the mutants and their wild types

Table 2.

Quantitative trait loci (QTL) on chromosome 5B for stripe rust resistance detected in the XN822 / LHX121 F_2:3 population using infection type (IT) and mean disease severity (DS) data

Environmenta	Marker interval	LOD	Addb	PVE%
5B-IT
IT-19YL	AX-89459127—AX-110578940	7.07	−0.97	10.67
IT-20YL	AX-86178260—AX-110049418	11.65	−1.14	17.74
IT-20TS	AX-89459127—AX-110578940	14.05	−1.18	21.47
IT-BLUE	AX-89459127—AX-110578940	14.61	−1.2	22.57
5B-DS
DS-19YL	AX-89459127—AX-110578940	9.95	−12.76	13.84
DS-20YL	AX-86178260—AX-110049418	8.51	−12.03	13.28
DS-20TS	AX-89459127—AX-110578940	7.15	−11.216	12.62
DS-BLUE	AX-110578940—AX-86178260	7.96	−11.03	12.30
2B-IT
IT-19YL	AX-111140388—AX-111023848	2.80	−0.61	4.30
IT-20YL	AX-111622432—AX-110909599	2.35	−0.48	3.15
IT-20TS	AX-111622432—AX-110909599	3.03	−0.5	3.86
IT-BLUE	AX-111622432—AX-110909599	2.94	−0.49	3.60
2B-DS
DS-19YL	AX-86174319—AX-111622432	6.74	−7.85	9.48
DS-20YL	AX-111622432—AX-110909599	2.50	−5.77	3.69
DS-20TS	AX-86174319—AX-111622432	1.55	−4.69	2.27
DS-BLUE	AX-111622432—AX-110909599	2.16	−5.26	3.07
3B-IT
IT-19YL	AX-109990245—AQP-TM	3.89	−0.79	5.31
IT-20YL	AX-109990245—AQP-TM	1.65	−0.6	3.06
IT-20TS	AX-109990245—AQP-TM	1.42	−0.41	1.78
IT-BLUE	AX-109990245—AQP-TM	1.52	−0.418	1.8874
3B-DS
DS-19YL	AX-109990245—AQP-TM	3.87	−7.82	5.32
DS-20YL	AX-109990245—AQP-TM	2.13	−6.84	3.00
DS-20TS	AX-109990245—AQP-TM	1.91	−6.56	2.79
DS-BLUE	AX-109990245—AQP-TM	2.11	−6.65	3.00

a YL and TS, Yangling and Tianshui, respectively. 19 and 20, 2018–2019 and 2019–2020 cropping seasons, respectively

b All the resistance alleles of QTL are from the parent LHX121

Based on the integrated genetic map (Fig. 4d), chromosome 5B contains three officially named genes (Yr3a, Yr3b, Yr3c, Yr47, and Yr74) and 23 QTL (Q5B.1-Q5B.22) conferring stripe rust resistance in previous studies (Chen and Kang 2017). However, these named genes are related to seedling resistance. Therefore, YrL121 is different from them. Only six QTL with significant effects were identified, specifically QyrPI192252-5BS (Q5B.3), Qyr.tem-5B.1 (Q5B.6), Qyr.tem-5B.2 (Q5B.18), Qyr.inra-5BL.1 (Q5B.7), Qyr.inra-5BL.2 (Q5B.20), and QYrdr.wgp-5BL (Q5B.8), which accounted for over accurately 20% of the phenotypic arm variation on 5B. QYrPI192252-5BS, which was precisely mapped to the 5BS arm, was the only QTL that was linked to HTAP resistance at a high level. Only QYrPI192252-5BS was accurately mapped to the 5BS arm and was related to high-level HTAP resistance. The other QTL were positioned on 5BL (Lu et al. 2014). The pedigree analysis showed that QYrPI192252-5BS was from the Portuguese spring wheat landrace PI 192252. Additionally, the relative distances of all loci did not overlap the YrL121 position on the integrated genetic map (Fig. 4d). These results indicated that YrL121 might be different from the other loci. Nevertheless, whether it is a new resistance locus needs further verification and comparison.

Phylogenetic analysis and comparative genomics

After filtering the 35 K SNP array data, there were 22,796 SNPs with physical location information. Using these SNP data we analyzed the phylogenetic population structure. As shown in Fig. 5a, LHX121, L92-47 and line P10057 were in a unique branch. To ensure the reliability of the results, we also analyzed a series of RILs from the cross MX169 × Centurm. Plants of Centurm and MX169 were clustered into the same group, respectively. And their derived lines were also in the same branch (Fig. 5a). In addition, cultivars prefixed by “Zhoumai” derived from parent Zhou 8425B were also in a common branch. These results indicated that LHX121 was derived from L92-47 and that L92-47 had the closest relationship with line P10057. Interestingly, another wheat cultivar, Lanhangxuan 122, was also proved to be from Lantian 10 by space mutagenesis (Fig. 5a) that was consistent with the result in the previous study (Lv et al. 2016). In addition, we determined genotype differences between LHX121 and L92-47 across the whole genome and there were 2,189 SNPs accounting for 9.6% of 22,796 SNPs (Table S5). The SNPs were clustered on each chromosome (Fig. 5b). In general, the terminal regions of chromosome tend to harbor more SNP differences that is probably related to the activities of chromosome exchange (Cannan and Pederson 2016).

We made a comparative genomics analysis of the YrL121 region which was also the enriched region of differential SNPs (Fig. 5b). First, we extracted all the high confidence genes in the target region within the physical interval (21–27 Mb) in the International Wheat Genome Sequencing Consortium (IWGSC) Chinese spring (CS) reference genome v1.1 and compared it with the other published hexaploid wheat genomes such as 10 + Genome. The target region spanned 6 Mb in the reference genome of CS v1.1, and the corresponding intervals ranged from 5.27–13.57 Mb in other hexaploid genomes (Fig. 5c). The analysis indicated small fragment insertions and deletions in the target region during the evolution and artificial selection. Moreover, there were also some inverted segments in certain genomes. All the evidence suggests that this chromosome segment is more active for crossover and recombination and is likely to be more susceptible to space mutagenesis (Lian et al. 2022; Monroe et al. 2022).

Expression changes in genes in the wild-type and mutant line in responsive to Pst

To pinpoint putative differentially expressed genes we evaluated and compared RNA-seq profiling in uninoculated and Pst-inoculated leaves of L92-47 and LHX121. Only the RNA-seq reads uniquely mapped to the reference genome of Chinese Spring IWGSC v1.1 were used to calculate the relative abundance of transcripts from each gene. A total of 6,472 DEGs associated with 2,486 downregulated genes and 3,986 upregulated genes, respectively, were identified in LHX121 and not L92-47 using a false discovery rate of 5% (Fig. 6a and Table S6). The physical distribution of differentially expressed genes when overlaid on the wheat genome, overlapped most of the regions (Table S6). Gene ontology (GO) analysis showed that the 6,472 DEGs are linked to biological pathways involved in plant biotic stress response (Fig. 6b), including several genes related to defense response signaling pathway, response to salicylic acid, and innate immune response (Table S7). There were five specific genes with expression changes in the genetically mapped region in chromosome arm 5BS (Fig. 6c; Table S8); TraesCS5B02G022800, TraesCS5B02G024700, TraesCS5B02G025300 and TraesCS5B02G026000 were up-regulated in LHX121 compared to L92-47. TraesCS5B02G024700 and TraesCS5B02G025300 encode a CC-NBS-LRR family disease resistance protein and a receptor-like kinase, respectively, both of are related to disease response. TraesCS5B02G023300 encodes a cationic amino acid transporter involved in nutrient substance transport and metabolism which is also associated with adult plant response (Kourelis and van der Hoorn 2018).

Fig. 6.

RNA-seq analysis between mutant LHX121 and wild-type L92-47. a Venn diagrams showing RNA-seq-based differentially expressed genes (DEGs) between LHX121 and L92-47 using a false discovery rate (FDR) of 5%. b Gene ontology (GO) analysis of DEGs only in LHX121. c Heatmap of gene expressions in the target region of YrL121 in L92-47 and LHX121. The gene number in red presents upregulated and the gene in blue presents downregulated comparing LHX121 with L92-47. CK: control check; IN: inoculation

Interaction between the QTL

To explore interaction among Yr27, Yr30 and YrL121 in both populations, we calculated the mean disease severities of lines with different QTL combinations in each population. The mean disease severity for genotypes with resistance alleles from all three genes/QTL in cross XN822/LHX121 was 15.25–18.5% and 15.0–22.7% in cross XN979/L92-47 whereas the double combination and single resistance allele genotypes had higher values and those with no resistance allele were the most susceptible, but clearly lower than the susceptible parents indicating the presence of further resistance genes/QTL (Fig. 7a, b, c, d) again indicating other minor undetected resistance genes. LHX121 had lower disease severity than the mean value for lines with two QTL (Figs. 1a and 7c). The combination of Yr30 and YrL121 conferred the highest resistance among two-QTL combinations. Among the three genes individually, YrL121 was the most effective. QYrLHX.nwafu-3BS explained more disease severity in cross XN979/L92-47, and Yr27 explained more in cross XN822/LHX121.

Fig. 7.

Effects of quantitative trait loci (QTL) combinations on phenotype and final disease severity scores in the XN822/LHX121 (a, c) and XN979/L92-47 (b, d) populations

Discussion

QTL mapping represents a widely-used method for dissecting quantitative genetic variation and serves as a foundation for map-based cloning of associated genes and for marker-assisted selection (MAS). The procedure of QTL mapping involves genotyping a large number of individuals from a biparental cross, which can be a laborious and expensive. The strategy of BSA, as proposed by Michelmore et al. (1991) is a pragmatic method to search for markers that are linked to particular genes or QTL associated with a trait of interest. Since 2000, there has have been a rapid advance in high-throughput genotyping technologies based on microarrays and next-generation sequencing (NGS). By utilizing these technologies, BSA has the capability to detect markers associated with the target genes or QTL (Wu et al. 2018). These markers can then be utilized to directly map target genes or QTL to reference genome sequences. The combination of high-throughput genotyping technologies and reference genome sequences is making BSA an increasingly useful approach for mapping genes or QTL. Through the utilization of an integrated BSA approach in tandem with a 660 K SNP array, we were able to rapidly identify three and two QTL/genes conferring stripe rust resistance in LHX121 and L92-47, respectively.

The resistance gene-rich chromosome 2BS region possesses many genes or QTL for stripe rust resistance (Chen and Kang 2017). Recently, the Lr13 and Yr27 resistance alleles were shown to be different sequences located at the same locus using long-read genome sequencing with optical mapping (Athiyannan et al. 2022). A recent study in our laboratory identified YrNP63/Yr27 in CIMMYT wheat cultivar Napo 63 by map-based cloning (Unpublished). YrNP63/Yr27 conferred a high-level adult plant resistance despite not being detected in seedling tests with Chinese Pst races (Wu et al. 2017b). The presence in LHX121 and L92-47 of the same AQP marker alleles linked with Yr27 in Napo 63 indicated that QYrLHX.nwafu-2BS could be Yr27. Chromosome arm 3BS is recognized as a gene-rich region for resistance. This region contains several stripe rust resistance genes, including the pleiotropic/closely linked Sr2/Lr27/Yr30, here referred to as Yr30. This adult plant resistance gene was presumably introgressed from emmer wheat (Triticum dicoccum) cv. Yaroslav in the breeding of bread wheat cultivars Hope and H-44 (McFadden 1930), but Yr30 was discovered relatively recently (Kota et al. 2006). According to Mago et al. (2011), Yr30 is recessive and confers limited resistance. The presence of Sr2 (and potentially Yr30) is often confirmed by pseudo- black chaff, a blackening of glumes and parts of the upper stem (Kota et al. 2006). Gene Sb3 for resistance spot blotch caused by B. sorokiniana (Sacc.) Shoem. syn. Drechslera sorokiniana (Sacc.) Subrm and Jain (syn. Helminthosporium sativum, teleomorph Cochliobolus sativus) was fine mapped to the same chromosome region (Lu et al. 2016). Tests of LHX121 and L92-47 with the Sr2 marker csSr2 (http://maswheat.ucdavis.edu/protocols/sr2/) and AQP™ marker for Sb3 (DNA sequence provided by Prof Liu Zhiyong, Chinese Academy of Science, Beijing) each gave positive results. The AQP™ marker was also genotyped in the F_2:3 lines of XN822/LHX121 (Fig. 4). These results strongly indicated that QYrLHX.nwafu-3BS is Yr30. QYrLHX.nwafu-5BS (YrL121) requires further validation. It appears to be novel as we have no evidence of a gene previously located gene in the same position in chromosome arm 5BS. We harbor the idea that the gene might have been present at low frequency in L92-47 and selected in the new cultivar. Here, we first suggested that LHX121 was actually selected from the wild type L92-47 using phylogenetic analysis. However, YrL121 was located in 0-50 Mb of chromosome 5BS where the concentration of SNPs occurred only in XN822/LHX121 population. Then we found that the target chromosome segment was more active for crossover and recombination. Finally, there were a large number of differentially expressed genes identified between LHX121 and L92-47 using RNA-seq analysis. Based on the previous studies, we infer that there may be suppressors in L92-47 that inhibited the expression of disease resistance genes. There are some suppressors examples such as Sr44, Pm3 or Pm8, Lr23, Yr18 in wheat breeding history (Nelson et al. 1997; McIntosh et al. 2011; Liu et al. 2013; Hurni et al. 2014; Huang et al. 2019). The dominant suppressor gene SuSr-D1, suppressor of stem rust resistance 1, D-genome, was first cloned on chromosome arm 7DL (Hiebert et al. 2020). The suppression mechanism of SuSr-D1 involves the interaction of RNA polymerase II, general and specific transcription factors through a subunit of the Mediator complex, which together coordinate transcription in eukaryotes. The regulatory processes of coregulation are frequently achieved through modification of chromatin architecture (Lian et al. 2022; Monroe et al. 2022). In the present study, there were several genes in the target region up/down regulated involved in plant disease resistance indicating that space mutations may lead to loss of function for certain suppressors, thus activating the expression of genes that confer disease resistance.

In the present study, we identified genes/QTL and the potentially new YrL121 locus underlying wheat stripe rust resistance using BSA with 660 K assays. This methodology can offer cost-effective and rapid flexibility to generate high-density genetic maps in F₂ and F_2:3 populations and be potentially useful for pyramiding stripe rust APR genes in wheat breeding. Phylogenetic analysis, comparative genomics and RNA-seq analysis reveal that space mutagenesis possibly caused new variation associated with stripe rust response but further research is required to elucidate the underlying cause. In wheat breeding endeavors, it requires genes/QTL that can be easily detected and possess robust and durable disease resistance. The QTL detected in this study all could be efficiently and precisely identified by linked AQP markers for MAS. The combination of these three QTL showed a high-level adult resistance to stripe rust, especially YrL121 which is more inclined to be a new locus, that holds substantial potential for integration into wheat resistance breeding projects.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors’ contribution

QM Wu, JH Wu and CL Li designed and conducted the experiments, analyzed the data, and wrote the manuscript. JY Du, WT Zhang and QL Wang participated in creation of the genetic populations and assisted in analysis of the SNP array data. L Liu, DD Zhang, CC Li, RQ Nie, JL Duan, JF Wan, JW Zhao, JH Cao, D Liu, and SJ Liu participated in greenhouse and field experiments and contributed to genotyping and data analysis. WJ Zheng, Q Yao, ZS Kang, and DJ Han participated in revision of the manuscript. CL Li, CF Wang and JH Wu conceived and directed the project and revised the manuscript.

Data availability

All data, models, or codes generated or used during the study are available by request from the corresponding authors.

Declarations

Conflict of interest

The authors declare no conflicts of interest and all experiments comply with the current laws of China.