Adaptive Traits to Improve Durum Wheat Yield in Drought and Crown Rot Environments

Abstract

Durum wheat (Triticum turgidum L. ssp. durum) production can experience significant yield losses due to crown rot (CR) disease. Losses are usually exacerbated when disease infection coincides with terminal drought. Durum wheat is very susceptible to CR, and resistant germplasm is not currently available in elite breeding pools. We hypothesize that deploying physiological traits for drought adaptation, such as optimal root system architecture to reduce water stress, might minimize losses due to CR infection. This study evaluated a subset of lines from a nested association mapping population for stay-green traits, CR incidence and yield in field experiments as well as root traits under controlled conditions. Weekly measurements of normalized difference vegetative index (NDVI) in the field were used to model canopy senescence and to determine stay-green traits for each genotype. Genome-wide association studies using DArTseq molecular markers identified quantitative trait loci (QTLs) on chromosome 6B (qCR-6B) associated with CR tolerance and stay-green. We explored the value of qCR-6B and a major QTL for root angle QTL qSRA-6A using yield datasets from six rainfed environments, including two environments with high CR disease pressure. In the absence of CR, the favorable allele for qSRA-6A provided an average yield advantage of 0.57 t·ha^-1, whereas in the presence of CR, the combination of favorable alleles for both qSRA-6A and qCR-6B resulted in a yield advantage of 0.90 t·ha^-1. Results of this study highlight the value of combining above- and belowground physiological traits to enhance yield potential. We anticipate that these insights will assist breeders to design improved durum varieties that mitigate production losses due to water deficit and CR.

Keywords: association mapping; drought adaptation; fusarium; root architecture; stay-green; water use.

Figure 1

Illustration displaying the effect of crown rot infection on yield under different scenarios: (A) Minimum yield losses when water is available through the growing season, i.e., intermittent rainfall events. (B) Maximum yield losses when water is limited, and the root system architecture is not designed to reach moisture at depth. (C) Less severe yield losses achieved under water-limited conditions when there are optimised below and aboveground trait combinations (i.e., root system architecture and stay-green).

Figure 2

The Pearson’s correlation matrix displays (A) stay-green trait correlations with yield in the presence of and (B) absence of crown rot in 2017 Warwick, Queensland. The colour gradient of ellipses indicates positive correlations (green) and negative correlations (brown) while the absence of an ellipse indicates that the correlation was not significant (p ≥ 0.05). Traits include yield (GY) in tonnes per hectare; PH, plant height; DTF, number of days to flowering; TKW, thousand kernel weight; OnS, number of days from flowering until 10% senescence; MidS, number of days from flowering time until 50% senescence; EndS, number of days from flowering time until 90% senescence; and SGint, area under the curve modelled senescence curve from flowering time to complete senescence.

Figure 3

Haplotype effects of seminal root angle quantitative trait loci (QTL) qSRA-6A on (A) stay-green (SGint) and (B) yield (GY) in the absence of crown rot (CR): Haplotype effects of qSRA-6A on (C) SGint and (D) GY in the presence of CR. Significance levels for comparisons of two major haplotypes are indicated at the levels p ≤ 0.001 (***) and p ≤ 0.01 (**). No significant difference between qSRA-6A-hap1 and qSRA-6A-hap2 was noted for SGint as well as GY in the presence of CR. In the boxplots, the line is the median, the box are the bounds for the lower and upper quartile values Q1 = 25% and Q3 = 75 respectively, while the lines below and above indicate the extreme values; values outside the lines are outliers.

Figure 4

The panel illustrates (A) crown rot (CR) severity during the grain filling stage with 5% (left) and 95% whiteheads (right). (B) The Manhattan plot shows significant marker-trait association (green) at an arbitrary threshold: −log10(P) ≥ 3 (blue horizontal line). The x-axis displays the DArTseq markers on 14 chromosomes; the y-axis is −log10(P). The local linkage disequilibrium (LD) block for 5 significant markers representing the CR QTL “qCR-6AB” was used for constructing the haplotype network. A total of 26 haplotype variants of the qCR-6B for 168 genotypes was observed, and the two major haplotype groups were used for investigating root biomass, stay-green and yield performance under different environments. (C) The significant differences between the hap1 variant (favorable allele in green) and hap2 variant (unfavorable allele in yellow) of qCR-6B is presented in boxplots. For haplotype trait comparisons, the significance level is indicated as *** (p < 0.001), ** (p < 0.01) and * (p < 0.05). (D) A section of chromosome 6B (60–125 cm) displaying the location of QTL was identified in this study along with a previously reported QTL positioned on the Svevo durum physical map. Genomic regions controlling CR severity and symptoms (whiteheads), stay-green and yield from this study (green) were aligned with previously reported QTL associated with traits such as root growth angle, Fusarium head blight (FHB), grain quality and yield (blue).

Figure 5

Comparison of genotypes that are segregating for root angle QTL (qSRA-6A) and CR QTL (qCR-6B): The genotypes were evaluated in the field across six environments (four in the absence of CR and two in the presence of CR). Green indicates the performance of genotypes that carry the resistant allele CRhap1 and wide root angle RAhap2. Orange indicates the performance of genotypes that carry the susceptible allele CRhap2 and narrow root angle allele RAhap1. Statistical tests were performed for haplotype groups within each environment, where the significance level is indicated as ** (p < 0.01) and * (p < 0.05). The mean yield benefit in the presence of CR was 1.1 t·h⁻¹, whereas the mean yield benefit in the absence of CR across the four environments was 0.57 t·h⁻¹.