Genome-Wide Linkage Mapping for Preharvest Sprouting Resistance in Wheat Using 15K Single-Nucleotide Polymorphism Arrays

Preharvest sprouting (PHS) significantly reduces grain yield and quality. Identification of genetic loci for PHS resistance will facilitate breeding sprouting-resistant wheat cultivars. In this study, we constructed a genetic map comprising 1,702 non-redundant markers in a recombinant inbred line (RIL) population derived from cross Yangxiaomai/Zhongyou9507 using the wheat 15K single-nucleotide polymorphism (SNP) assay. Four quantitative trait loci (QTL) for germination index (GI), a major indicator of PHS, were identified, explaining 4.6–18.5% of the phenotypic variances. Resistance alleles of Qphs.caas-3AL, Qphs.caas-3DL, and Qphs.caas-7BL were from Yangxiaomai, and Zhongyou9507 contributed a resistance allele in Qphs.caas-4AL. No epistatic effects were detected among the QTL, and combined resistance alleles significantly increased PHS resistance. Sequencing and linkage mapping showed that Qphs.caas-3AL and Qphs.caas-3DL corresponded to grain color genes Tamyb10-A and Tamyb10-D, respectively, whereas Qphs.caas-4AL and Qphs.caas-7BL were probably new QTL for PHS. We further developed cost-effective, high-throughput kompetitive allele-specific PCR (KASP) markers tightly linked to Qphs.caas-4AL and Qphs.caas-7BL and validated their association with GI in a test panel of cultivars. The resistance alleles at the Qphs.caas-4AL and Qphs.caas-7BL loci were present in 72.2 and 16.5% cultivars, respectively, suggesting that the former might be subjected to positive selection in wheat breeding. The findings provide not only genetic resources for PHS resistance but also breeding tools for marker-assisted selection.


INTRODUCTION
Preharvest sprouting (PHS) refers to the germination of physiologically mature grains in spikes before harvest under rainy weather or humid environment (Groos et al., 2002). PHS is a major problem in cereal production and causes losses in seed vitality, yield, and quality (Xu et al., 2019). Wheat (Triticum aestivum L.) is one of the most important staple crops. The average annual loss of wheat caused by PHS exceeds $1 billion worldwide (Shao et al., 2018). Identification of genetic loci for PHS should be helpful for breeding resistant wheat cultivars.
Preharvest sprouting is a complex trait influenced by genetic and environmental factors (Barrero et al., 2015;Wang et al., 2019). Seed dormancy, an adaptive trait that prevents seeds from germinating, even under favorable conditions, is a major genetic factor for PHS (Née et al., 2017). Germination index (GI) is a common parameter to quantify genetic mechanisms underlying seed dormancy and PHS (Barrero et al., 2015). Some non-dormancy factors, such as spike erectness, spike and awn structure, and openness of florets, are also associated with PHS (Zhu et al., 2019).
Yangxiaomai, a red-seeded Chinese landrace, has a high level of PHS resistance, whereas white-seeded Zhongyou9507 with good processing quality is susceptible to PHS. The objectives of this study are to mine QTL for PHS resistance in a recombinant inbred line (RIL) population derived from a Yangxiaomai/Zhongyou9507 cross and to develop breedingfriendly markers for selection of PHS-resistant varieties.

Plant Materials and Field Trials
The parents Yangxiaomai and Zhongyou9507 and 194 F 6 RILs were planted at Beijing and Shijiazhuang (Hebei Province) in the 2011-2012 cropping season and at Gaoyi (Hebei Province) and Xinxiang (Henan Province) in the 2019-2020 cropping season. Field experiments were arranged in randomized complete blocks with three replications. Each plot was 1 m single row in which 30 seeds were sown. A panel of 101 wheat cultivars (Zhang et al., 2017) was used to determine the genetic effects of the QTL of interest.

Evaluation of PHS Resistance
The GI was used as an indicator of PHS. Five spikes were harvested from each plot at physiological maturity characterized by loss of green color from the spike . The harvested spikes were air-dried for 2 days at room temperature, hand-threshed to avoid damage to embryos, and then stored in a refrigerator at −20 • C to maintain dormancy until phenotyping (Zhang et al., 2017). Seeds were sterilized with 1% (V/V) of NaClO for 20 min, followed by three rinses with sterile water. Notably, 100 healthy seeds of each line were incubated in a 90 mm Petri dish containing a filter paper and 8 ml of distilled water at 20 • C for 7 days. Germinated seeds were counted every day and removed. GI was calculated according to the following formula (Walker-Simmons, 1988 7×total grains ×100, where n 1 , n 2 , . . . , n 7 are the number of seeds germinated on the first, second, and subsequent days until the seventh day.

Statistical Analyses
Phenotypic correlation coefficients among environments, the best linear unbiased prediction (BLUP) values, ANOVA, and t-tests were carried out using SAS 9.4 software (SAS Institute Inc., Cary, NC, USA). Broad-sense heritability (H 2 ) for PHS was calculated using the following formula: H 2 = σ 2 g /(σ 2 g + σ 2 ge /e+ σ 2 ε /re), where σ 2 g , σ 2 ge , and σ 2 ε are the variances of genotype, genotype-environment interaction, and residual error, respectively, r is the number of replicates, and e is the number of environments (Nyquist and Baker, 1991).

Genotyping and Linkage Map Construction
The 194 RILs and parents were genotyped with the wheat 15K single-nucleotide polymorphism (SNP) chips containing 13,947 SNP markers at China Golden Marker (Beijing) Biotech Co., Ltd. (http://www.cgmb.com.cn/). To reduce the impact of low-quality SNPs on mapping results, SNP data were processed as follows: (1) Heterozygous loci were treated as missing data, and (2) SNPs with low minor allele frequencies (<0.3) and missing values (>0.2) were excluded using Tassel version 5.0 (Bradbury et al., 2007). Redundant markers were eliminated by the BIN function in QTL IciMapping version 4.2 (Meng et al., 2015). Joinmap version 4.0 was used for linkage map construction (Stam, 1993), and genetic distances between markers were calculated according to the Kosambi mapping function (Kosambi, 1943).

QTL Analysis
Composite interval mapping (CIM) was used to search QTL of phenotypic traits from each environment and BLUP value by Windows QTL Cartographer version 2.5 (Zeng, 1994;Wang et al., 2012). Significant QTLs were identified if the logarithm of odds (LOD) values were more than the threshold of 2.5 (Yan et al., 2006). According to International Wheat Genome Sequencing Consortium (IWGSC) RefSeq 1.0 [(International Wheat Genome Sequencing Consortium (IWGSC), 2018) http://plants.ensembl. org/index.html], the physical positions of QTL were figured out by the closely linked flanking markers. The genetic maps of QTL were drawn using MapChart version 2.3 (Voorrips, 2002). The analysis of epistatic effects among the QTL was performed using IciMapping version 4.2.

Phenotypic Evaluation
The parents Yangxiaomai and Zhongyou9507 and RILs were evaluated for PHS resistance in four environments. The phenotypes of seed germination in parents Yangxiaomai and Zhongyou9507 were depicted in Supplementary Figure 1. Yangxiaomai had a significantly lower GI (4.3%) than Zhongyou9507 (72.3%) across environments (Supplementary Figure 2). GI for the RIL population showed continuous variation, indicating polygenic inheritance (Supplementary Figure 2). The GI frequencies were skewed toward resistance, suggesting the presence of major genetic loci. GI was significantly correlated among environments with correlation coefficients of 0.53-0.73 (Supplementary Table 1). ANOVA indicated that genotypes and environments, as well as their interactions, had significant effects on GI (Supplementary Table 2). The broad-sense heritability of GI was high (0.88) across environments, denoting that GI variation was mainly determined by genotypes.

Linkage Map Construction and QTL Analysis
The RIL population was genotyped by 15K SNP chips, and 4,515 polymorphic markers were used to construct a genetic map with 1,702 bin markers, spanning 2,630.9 cM on 21 wheat chromosomes (Supplementary Table 3 and Supplementary Figure 3). The average linkage group was 125.3 cM with an average marker interval of 1.6 cM. Overall, 1,743 (38.6%), 1,750 (38.8%), and 1,022 (22.6%) markers were mapped to the A, B, and D sub-genomes with average marker densities of 1.5, 1.2, and 2.2 cM, respectively (Supplementary Table 3 and Supplementary Figure 3).

Combinational Effects of the Stable QTL for PHS Resistance
Quantitative trait loci for a given trait detected in more than one-half of tested environments can be considered stable genetic loci (Cao et al., 2020). The QTL Qphs.caas-3AL, Qphs.caas-3DL, and Qphs.caas-4AL fulfilled that criterion. To confirm their genetic effects on PHS, the population was classified into eight groups based on the closest flanking SNPs for each QTL (Supplementary Table 4). Qphs.caas-3AL, Qphs.caas-3DL, and Qphs.caas-4AL were temporarily designated as the loci 1, 2, and 3, respectively, and R and S represented resistance and susceptible alleles, respectively. The GI values of eight groups (i.e., 1 R 2 R 3 R , 1 R 2 R 3 S , 1 R 2 S 3 R , 1 S 2 R 3 R , 1 R 2 S 3 S , 1 S 2 R 3 S , 1 S 2 S 3 R , and 1 S 2 S 3 S ) were compared across four environments (Figure 2 and Supplementary Table 4).
The eight groups were ranked according to the GI in four environments and BLUP values. In general, more resistance alleles conferred lower GI, demonstrating cumulative effects of resistance alleles at the three loci (Figure 2). RILs with genotype 1 R 2 R 3 R had the lowest GI, and those with 1 S 2 S 3 S exhibited the highest GI across all environments. However, the GI of RILs with 1 R 2 R 3 S were higher than those with 1 S 2 R 3 S in Beijing 2012, suggesting that the genetic effect of locus 1, Qphs.caas-3AL, was significantly affected by the environment in some cases. No epistatic effects were detected among the QTL. Regression analysis also showed that the lines carrying more resistance alleles had higher PHS resistance in individual environments and the BLUP data (Supplementary Figure 4). Thus, the pyramiding of resistance alleles was effective in improving PHS resistance.

Relationship of Qphs.caas-3AL and Qphs.caas-3DL With Tamyb10
According to IWGSC RefSeq 1.0, Qphs.caas-3AL and Qphs.caas-3DL were delimited in the intervals of 700.4-709.2 Mb and  . The x-axis shows the genotypic groups, and the y-axis indicates GI (%). The numbers 1, 2, and 3 represent Qphs.caas-3AL, Qphs.caas-3DL, and Qphs.caas-4AL, respectively; the superscript letters R and S represent resistance and susceptible alleles, respectively. Genotypes with different letters indicate significant differences (P < 0.05) in GI, and those with the same letters show no significant differences (P > 0.05). Error bar, standard error of each group mean.

Relationship Between Qphs.caas-4AL and Reported PHS Resistance Genes on Chromosome 4AL
TaMKK3-A was reported as a major gene controlling seed dormancy on chromosome 4AL (Torada et al., 2016). Based on the IWGSC RefSeq 1.0, TaMKK3-A is located at the site of ∼609.1 Mb (GenBank accession number: LC091368.1) (Liton et al., 2021  two parents in all exons, but three SNPs were detected in the introns (Supplementary Figure 5). A KASP marker KASP-6464 was developed for TaMKK3-A. Linkage mapping showed that KASP-6464 was out of the target region of Qphs.caas-4AL (Figure 1). Association analysis also indicated that TaMKK3-A had no significant effect on GI in three environments (Supplementary Table 6). Therefore, TaMKK3-A was not a candidate gene in Qphs.caas-4AL. PM19-A1 is a second PHS-related gene in chromosome 4AL. However, it is positioned at ∼604.1 Mb, which is far from Qphs.caas-4AL (489.1-532.2 Mb) according to the IWGSC RefSeq 1.0. We also compared the sequences of PM19-A1 between Yangxiaomai and Zhongyou9507 and found no polymorphic sites in the open reading frame and the previously reported 18 bp indel in the promoter (Barrero et al., 2015;Shorinola et al., 2016Shorinola et al., , 2017. Therefore, PM19-A1 was not the causal gene in Qphs.caas-4AL.
No genes related to PHS were isolated on 7B so far, so no candidate genes could be used to perform sequencing and genetic mapping analyses for Qphs.caas-7BL. Qphs.caas-7BL is defined in the interval of 522.6-529.7 Mb based on its flanking markers according to IWGSC RefSeq 1.0. QTL related to PHS, which have been reported on 7B, were summarized in Supplementary Table 7. The location of Qphs.caas-7BL is different from those of the previously reported QTL based on their flanking markers.

Genetic Effects of Qphs.caas-4AL and Qphs.caas-7BL on GI in a Panel of Cultivars
The causal genes of Qphs.caas-4AL and Qphs.caas-7BL remained unknown although they could explain 4.6-10.6% of the phenotypic variances. To further decipher the importance of Qphs.caas-4AL and Qphs.caas-7BL, we attempted to identify their genetic effects in a panel of cultivars. An SNP AX-89597750 closely linked with Qphs.caas-4AL was converted to a KASP marker. The KASP marker was mapped at the position of AX-89597750, indicating that it could act as a closely linked marker of Qphs.caas-4AL. We detected the genetic effect of Qphs.caas-4AL on GI using the KASP marker in the cultivar panel (Supplementary Table 8). Genotypes with the resistance allele had significantly lower GI than those with the susceptible allele ( Table 2). Moreover, a majority of genotypes (72.2%) possessed the resistance allele of Qphs.caas-4AL, suggesting that it had been subjected to selection in wheat breeding ( Table 2). Another KASP marker was developed based on SNP AX-9496498 closely linked to Qphs.caas-7BL. The KASP marker was mapped to the target region of Qphs.caas-7BL in the mapping population. Like Qphs.caas-4AL, Qphs.caas-7BL was also significantly associated with PHS resistance ( Table 2), but only 16.5% of cultivars carried the resistance allele.

Grain Color Is a Major Factor Modulating PHS
The red-grain Yangxiaomai and white-grain Zhongyou9507 have a relatively low and higher GI, respectively. In this study, we confirmed that Qphs.caas-3AL and Qphs.caas-3DL co-localized with Tamyb10-A1 and Tamyb10-D1, respectively, at the R loci for grain color (Himi et al., 2011;Lang et al., 2021;Mares and Himi, 2021). Wang et al. (2016) also observed that Tamyb10-D1 was significantly (P < 0.001) associated with GI in a natural population. Thus, grain color is probably regulated by Tamyb10 alleles in Qphs.caas-3AL and Qphs.caas-3DL, which also cause significant differences in GI between Yangxiaomai and Zhongyou9507. Pleiotropic QTL for grain color and PHS resistance on chromosomes 3AL and 3DL were identified in a genome-wide association study conducted by Lin et al. (2016). These findings also confirm that grain color has a great effect on PHS resistance in wheat breeding.

Qphs.caas-4AL Has Potential Value for Wheat Breeding
Qphs.caas-4AL, as a stable QTL, accounted for 4.6-10.6% of the phenotypic variances. Although quite a few QTL for wheat PHS resistance on chromosome 4A have also been reported (Kato et al., 2001;Mares et al., 2005;Torada et al., 2005;Imtiaz et al., 2008;Rasul et al., 2009;Singh et al., 2010;Kulwal et al., 2012;Cabral et al., 2014;Jiang et al., 2015;Cao et al., 2016;Zhou et al., 2017;Martinez et al., 2018;Zuo et al., 2019;Liton et al., 2021), Qphs.caas-4AL appears to be unique based on genetic mapping and physical locations of the flanking SNPs according to IWGSC RefSeq 1.0 (Supplementary Table 9). QTL pyramiding plays an important role in breeding, and resistance allele combinations of QTL for PHS have been reported previously Shao et al., 2018;Liton et al., 2021). In this study, we identified that the RILs combining resistance alleles in Qphs.caas-3AL, Qphs.caas-3DL, and Qphs.caas-4AL had the lowest GI (Figure 2). Qphs.caas-4AL also improved resistance to PHS in the absence of the alleles for red-grain color (Figure 2 and Supplementary Table 4). This indicated that Qphs.caas-4AL could function independently of grain color. We converted an SNP tightly linked to Qphs.caas-4AL into a KASP marker. Genotyping by the KASP marker showed that most of the cultivars (72.2%) carried the resistance allele in Qphs.caas-4AL in the test panel ( Table 2), indicating that the resistance allele of Qphs.caas-4AL might undergo positive selection in breeding programs. Qphs.caas-4AL is also significantly associated with GI (Table 2). Thus, the KASP marker will be useful for MAS to improve PHS tolerance in wheat. Grain color is an important parameter for wheat appearance quality. Red-grain wheat usually has high resistance to PHS but is adverse to make Chinese traditional food, such as steamed bread and noodles (Fakthongphan et al., 2016;Shao et al., 2018). Thus, Qphs.caas-4AL is a better choice for improvement of PHS than Qphs.caas-3AL and Qphs.caas-3DL at least in Chinese wheat breeding. Overall, these findings indicate that Qphs.caas-4AL is a valuable genetic locus for PHS in wheat breeding.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Materials, further inquiries can be directed to the corresponding authors.

AUTHOR CONTRIBUTIONS
LL and SC wrote the manuscript. LL, YiZ, ML, DX, XT, JS, and XL performed the experiments. SC and LL analyzed the data. YoZ, LX, and DW participated in the field trials. SC and YaZ designed the experiments. XX and ZH assisted in writing the manuscript. All authors read and approved the final manuscript.