Genome-wide association analysis of type II resistance to Fusarium head blight in common wheat

Plant materials

The association mapping panel of 205 wheat accessions for GWAS comprised 77 released cultivars, 55 founder parents including two lines from Mexico and France, and 73 breeding lines from 10 provinces that represent the major winter wheat production regions in China (Chen et al., 2016).

Sumai 3 was selected as an FHB highly resistant check, Yangmai158 as a moderately resistant check, and Ningmaizi 22 as a moderately susceptible check.

Growth conditions

The materials were grown in the field and greenhouse of Shandong Agricultural University (117°16′E, 36°17′N) during the 2015–2016 and 2016–2017 cropping seasons, hereafter referred to as 2016 and 2017, respectively. The terms E1, E2, and E3 represent the experimental field of Shandong Agricultural University in 2017, the greenhouse of Shandong Agricultural University in 2017, and the experimental field of Shandong Agricultural University in 2016, respectively. In the field, a randomized block design was used, with two replications. All lines were planted in 2 m plots with three rows uniformly spaced at 25 cm intervals. Each row contained 70 seeds evenly distributed seeds. The local recommended field crop management practices were followed and no pests or diseases were found in the field. However in the greenhouse, the seeds of materials were germinated and vernalized for an additional 4 weeks (4 °C, 12 h light/dark regime) before being transferred to the greenhouse. Plants were potted in a mixture of compost, sand chalk and common soil. Each plant was planted in an individual pot (the diameter was 30 cm, and the depth was 25 cm, with four seedlings) and in three replications (pots). The temperatures of the greenhouse were gradually increased from 15 °C/13 °C during day/night to 20 °C/18 °C, with a 16 h/day photoperiod at the time of anthesis. The growing conditions in the greenhouse have been described in our previous article (Zhao et al., 2023).

Inoculum preparation and inoculation

In this study, the conidiospore suspension of 7136, F301, F609, and F15 virulent strains of F. graminearum, was obtained with the courtesy of Nanjing Agricultural University. The pathogen was propagated in mung bean medium (Buerstmayr et al., 2000) and incubated on a shaker at 150 rpm under 25 °C for 4–5 days. After culturing and filtering, the mass of conidia was examined under a microscope, and then, the four pathogen strains were mixed equally and stored at 4 °C for later use. The preparation of F. graminearum was also the same as in our previous study (Zhao et al., 2023).

Wheat was inoculated with 10 µl of the F. graminearum conidia suspension (50,000 spores/mL) applied to a pair of florets in the middle of the spikeduring flowering (Guo et al., 2015). Ten spikes were inoculated per line from each replicate. The whole wheat spike was then covered with a self-sealing bag to retain moisture, and had the self-sealing bag was removed after 3 days. The disease symptoms were investigated on Day 21 after inoculation, and the diseased spikelet rate (DSR), diseased spike rachis rate, and disease index (DI) were calculated. Both the field nursery and greenhouse experiments followed this method. All spikes were classified into five classes of disease severity according to the diseased spikelet rate (DSR): 0% (class 0), 1–25% (class 1), 26–50% (class 2), 51–75% (class 3), and 76–100% (class 4) (Lu, Cheng & Wang, 2001). The disease index (DI) was calculated based on the rules for monitoring and forecasting wheat head blight (Chinese Standard: GB/T 15796-2011). ${\displaystyle \text{Diseased spikelets rate}=\frac{\text{Number}\mathrm{of}\text{infected spikelets}}{\text{Total spikelets}\mathrm{per spike}}\times 100\text{\%}}$ ${\displaystyle \text{Diseased spike rachis rate}=\frac{\text{Number}\mathrm{of}\text{infected rachis}}{\text{Total rachis}\mathrm{per spike}}\times 100\text{\%}}$ ${\displaystyle \text{DI}\left(\text{\%}\right)=\frac{\sum h_i\times i}{H\times 4}\times 100\text{\%}}$ where i is the class of disease severity, hi is the number of wheat spikes in each class, and H is the number of all investigated wheat spikes.

The standard of wheat FHB resistance was as follows: Immune (DI = 0), High resistant (DI<Sumai 3), Moderately resistant (Sumai 3<DI<Yangmai 158), Moderately susceptible (Yangmai 158<DI<Ningmaizi 22), High susceptible (DI>Ningmaizi 22).

Genome-wide association analysis

The genome-wide positions of SNPs were used in this study based on a consensus genetic map of wheat (Wang et al., 2014). SNP markers, genotyping, and the population structure of the samples have been previously reported (Chen, Chen & Tian, 2015; Chen et al., 2016; Chen et al., 2017). A total of 24,355 mapped SNPs were used for MTA analysis. According to Chen et al. (2017), the association population was divided into four categories using STRUCTURE’s maximum membership probability(comprised of 43, 32, 105 and 25 varieties). Chen et al. (2016) also reported LD values for different chromosomes.

Based on this information, significant marker −trait associations (MTAs) were identified using a mixed linear model (MLM) in TASSEL3.0. The P value was used to determine whether a QTL was associated with a marker, while the phenotypic variation explained(PVE) was used to evaluate the magnitude of the MTA effects. SNPs with P ≤10⁻³were considered to be significantly associated with phenotypic traits, and SNPs with P ≤10⁻⁴ were considered to be extremely significantly associated with phenotypic traits. Furthermore, when the marker was detected in two or more environments at the same time, it was considered a stable MTA.

Statistical analysis

Analysis of variance (ANOVA) and correlations among phenotypic traits were carried out using the statistical software SPSS version 17.0 (SPSS Inc., Chicago, IL, USA).

Forecasting candidate genes for FHB resistance

A BLAST (Basic Local Alignment Search Tool) search was performed on the International Wheat Genome Sequencing Consortium database (RefSeq v1.0; https://urgi.versailles.inra.fr/blast/) using the sequence of the significant SNP markers identified by GWAS. When an SNP marker sequence from the IWGSC was 100% identical to any wheat contig, the sequence was extended by 2 Mb for each marker using the IWGSC BLAST results. Afterward, the extended sequence was used to run a BLAST search on the National Center for Biotechnology Information (NCBI) database (https://www.ncbi.nlm.nih.gov/) and on Ensembl Plants (http://plants.ensembl.org/Triticum_aestivum/Tools/Blast) to confirm possible candidate genes and functions.

Analysis of marker haplotype

The effect of resistance genes was calculated by using the average diseased spikelet rate of various gene combinations. Effect of resistance genes = (Average diseased spikelet rate of materials carrying resistance gene − Average diseased spikelet rate of materials without resistance gene)/Average diseased spikelet rate of materials without resistance gene.

Phenotypic variation analysis of wheat FHB resistance

The variation coefficient of the diseased wheat spikelet rate (DSR) was the highest in E2 (52.96%), followed by that in E3 (44.30%), and E1(36.55%) (Table 1), indicating that DSR genetic variation was abundant. The variance analysis of FHB resistance of the spikelet and spike rachis indicated that significant differences were present between cultivars and environments, and their interactions (Table 2). This illustrated that FHB resistance was a quantitative trait affected not only by genotype but also by the environment. Furthermore, there were significant positive correlation coefficients between the spikelet and spike rachis in the three environments, indicating that the resistance trend of FHB was consistent between the spikelet and spike rachis (Table 3).

Marker–trait associations (MTAs) of FHB resistance

Sixty-six MTAs associated with FHB resistance were distributed on chromosomes 1A, 1B, 2A, 2B, 2D, 3B, 3D, 4A, 5A, 5B, 5D, 6A, 6B, 7A, and 7B (Tables S1 and S2; Fig. 1). The phenotypic variation explained(PVE) of MTA loci to FHB resistance ranged from 5.45% to 11.20%, of which, 11 MTA loci were detected in both the spikelet and spike rachis. On chromosome 7B, a novel genomic region from genetic position 92 to 103, significantly associated with FHB resistance, was detected in all three environments. Moreover, there was one major locus at genetic position 92 of chromosome 7B accounting for 11.20% of the phenotypic variation in the spikelets, namely locus BS00025286_51, which was also detected for the spike rachis, explaining 7.07% of its phenotypic variation. In E3, four loci on chromosome 7B were found to be associated with both diseased spikelet rate and diseased spike rachis rate, but these MTAs are located in the same region and may represent one QTL (Table 4; Table S2). In addition, there were some genomic regions associated with FHB resistance on chromosomes 5B, 6A, 2A, and 3B, but they were found only in a single environment. The other six loci, including D_contig74317_533 on chromosome 5D, Kukri_c14239_1995 on chromosome 6A, Kukri_c7087_896 on chromosome 3B, RAC875_c35801_905 on chromosome 3D, BS00099729_51 on chromosome 5B, and RAC875_c68525_284 on chromosome 6B, were also determined to be associated with both the diseased spikelet rate and the diseased spike rachis rate. The remaining MTA loci were detected only for a single trait in a single environment.

Allelic variation analysis of MTA loci

According to the PVE of MTA loci (Table S2), we selected 10 of markers and analyzed their allelic variation (Table 4). Alleles T and C of the marker, Kukri_c14239_1995 on chromosome 6A were associated with the largest phenotypic difference (0.2297). Specifically, the phenotypic value of the diseased spikelet rate associated with Kukri_c14239_1995-T was significantly higher than that associated with Kukri_c14239_1995-C, indicating that Kukri_c14239_1995-C was better than Kukri_c14239_1995-T for FHB resistance (Table 4). Furthermore, because allele C of D_contig74317_533 showed a significantly higher diseased spikelet rate than D_contig74317_533-T, allele T was deemed to be better for improving FHB resistance. On chromosome 7B, allele C of BS00025286_51 had a higher diseased spikelet rate than allele T; thus, allele T for this locus was favorable for FHB resistance. Nevertheless, for the other four loci on this chromosome, significant differences between the two alleles for the diseased spikelet rate seemed to be at 5%. The smallest difference in diseased spikelet resistance was observed between Kukri_c7087_896-G and Kukri_c7087_896-A, which indicated that this locus affected FHB resistance to a smaller degree. Moreover, on chromosome 3D, RAC875_c35801_905-G yielded better results than RAC875_c35801_905-A for FHB resistance.

Prediction of candidate genes for some important loci

Eight important candidate genes were screened for important loci significantly associated with the diseased spikelet rate and diseased spike rachis rate in wheat (Table S3). One candidate gene, TraesCS3D02G326700 located on chromosome 3D is associated with actin-binding in wheat. The candidate gene, TraesCS5D02G006700 of the marker, D_contig74317_533 on 5DS was predicted in wheat, whose function was intramolecular transferase activity. Two candidate genes TraesCS6A02G013700 and TraesCS6A02G013800, predicted by IAAV9150, participate in the ubiquitin-dependent ER-associated degradation (ERAD) pathway in wheat. The candidate gene, TraesCS6A02G013600 of Excalibur_c20597_509 functions in GTP binding in wheat. On chromosome 7BL, TraesCS7B02G370700 of BS00025286_51 is involved in the biological process of defense response to fungi. There is one candidate gene, TraesCS7B02G340200, for the three loci, RAC875_c18043_369, RAC875_c18043_411, and Kukri_c4143_1055, on chromosome 7BL that was identified because of being in the same physical location. The candidate gene, TraesCS7B02G340100 of RAC875_c5646_774 is associated with the carbohydrate metabolic process in Triticum aestivum. By analyzing the homologous genes of these candidate genes, we found that the functions or biological processes of most homologous genes in other crops, such as Japonica rice, Hordeum vulgare, Oryza sativa Indica, Oryza sativa Japonica, and Arabidopsis thalian, are involved in the defense response of fungi, which indicates that these candidate genes may be related to FHB resistance, but this needs further verification in wheat.

Analysis of marker haplotype and resistance

Among the alleles at the marker −trait associations, the alleles with decreasing diseased spikelet rate were assumed to be the resistance alleles at this site. In this study, five important SNP loci were selected to evaluate the effect of aggregation of their favorable alleles on decreasing the diseased spikelet rate (Table 5). In general, with the increase in the number of favorable alleles, the effect of reducing the rate of diseased spikelets was more obvious, which could improve FHB resistance. By gene combination analysis (Table 6), two samples, B202 and B34, had four favorable alleles with TCTAC and TTTGC haplotypes, respectively. The haplotype TCTAC showed high resistance to FHB (Table 6). Six haplotypes with three favorable alleles were found. The materials with the haplotypes TCCGA and TTTAC showed high resistance to FHB, including B70, B72 and B179. The residue haplotypes showed moderate resistance or moderate susceptibility to FHB. This result indicated that multiple haplotypes of the materials played an important role in the screening of anti-scab materials. In addition, by comparing the reported GWAS results for plant height using the same panel materials (Chen, Chen & Tian, 2015), these five loci had no significant effect on plant height (Tables S4 and S5).