Genomic regions associated with spot blotch resistance in elite barley breeding populations

doi:10.21203/rs.3.rs-3982288/v1

Download PDF

Research Article

Genomic regions associated with spot blotch resistance in elite barley breeding populations

https://doi.org/10.21203/rs.3.rs-3982288/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Spot blotch (SB) is a common foliar disease of barley, caused by the hemibiotrophic fungal pathogen Bipolaris sorokiniana. Predominately found in humid growing regions worldwide, the disease can cause yield losses of up to 30%. Genetic control through deployment of durable resistance in barley cultivars is the preferred, yet the majority of Australian cultivars are susceptible, despite the identification of major resistance loci in bi-parental studies. In this study, we explore the genetic architecture of spot blotch resistance in an Australian barley breeding program. Seedling (glasshouse) and adult (field) experiments using a single conidial isolate (SB61) were conducted across two consecutive years for 337 lines genotyped with 16,824 polymorphic DArT-seq™ markers. Two mapping approaches were investigated, a single-marker genome-wide association study (GWAS) and a haplotype-based local genomic estimated breeding values (GEBV) approach. Both methods detected two major regions on chromosomes 3H and 7H that contribute resistance across growth stages. Additional chromosomal regions on 1H, 3H and 6H were detected only in the local GEBV approach. Haplotype stacking revealed the importance of the 7H region for adult-plant resistance in combination with other resistance haplotypes, suggesting a significant gene by gene interaction and demonstrating the highly quantitative nature of the genetic control. Our study validates the presence of major spot blotch resistance loci in Australian breeding populations, while generating novel insight into the genetic architecture, and highlights the challenges and opportunities to improve resistance levels via haplotype stacking or whole-genome approaches.

Spot blotch

resistance

genome-wide association study

haplotype

barley

Bipolaris sorokiniana (teleomorph: Cochliobolus sativus) is a hemibiotrophic fungal pathogen that causes spot blotch foliar disease and associated yield losses in barley (Hordeum vulgare L.; Knight et al. 2010, Wang 2015, Gupta et al. 2018, Wang et al. 2019, Kaur et al. 2022). The disease is most prevalent in warm and humid growing locations worldwide (Steffenson et al. 1996, Meldrum et al. 2004, Bovill et al. 2010, Grewal et al. 2012, Gupta et al. 2018), and the prevalence and geographical areas impacted by spot blotch is anticipated to expand with the global increase in temperature and rainfall variability (Bastas 2022, Martínez et al. 2022). Increased disease incidence is of concern to the global barley industry, where yield losses of up to 30% are associated with spot blotch infection (Kaur et al. 2022). Among the disease management options, deployment of cultivars with robust genetic resistance is the preferred method of control, both environmentally and economically (Leng et al. 2020). Pathotypes of Cochliobolus sativus are diverse, continuously evolving and display divergent virulence (Meldrum et al. 2004). Previous exploration of this pathotypes virulence demonstrated that only four out of fifteen Australian barley genotypes exhibited resistance, with the majority of the commercial barley cultivars exhibiting susceptibility (Meldrum et al. 2004). To date, genetic characterisation of resistance has focused on quantitative trait loci (QTL) mapping approaches applied to research populations to locate associated loci, with minimal exploration of resistance in breeding populations (Lander and Botstein 1986, Bilgic et al. 2005, Roy et al. 2010).

Using this approach, three resistance genes have been fine mapped thus far. Rcs5 - associated with all-stage resistance - was mapped to chromosome 7H using the Steptoe × Morex double haploid population (Steffenson et al. 1996, Bilgic et al. 2006). Further, the Rcs6 resistance gene was identified on chromosome 1H, using the Calicuchima × Bowman doubled haploid population, conferring both all stage and adult plant resistance (Bilgic et al. 2006). The corresponding susceptibility allele Scs6, located at the same locus, was later identified by (Leng et al. 2020) and deemed to be dominate and contributed by the cultivar Bowman. An interval on the short-arm of chromosome 6H, designated Rbs7, was also consistently detected for spot blotch resistance, containing a small number of both high- and low-confidence genes (Wang et al. 2017, Wang et al. 2019). In recent years, several genome wide association studies (GWAS) have been conducted. In the United States Department of Agriculture (USDA) barley core collection, 1480 accessions were assessed, with six novel QTLs conferring resistance being identified on chromosomes 1H, 2H, 3H, 6H and 7H (Wang et al. 2017). The QTL identified on chromosome 3H was also identified in the International Center for Agricultural Research in the Dry Areas (ICARDA) spring barley collection consisting of 336 lines (Gyawali et al. 2018), suggesting globally importance. Further, four novel all-stage resistance QTL were identified in the ICARDA collection tested in Morocco, while nine novel adult plant stage resistance QTL were detected when assessed in India. The Rcs5 region was consistently detected within the ICARDA collection across growth stages and locations (Gyawali et al. 2018). A similar study exploring resistance in a collection of high input barley lines found a similar result, identifying some novel and some previously reported resistance QTL (Visioni et al. 2020). This research demonstrates that the genetic architecture of spot blotch is quantitative in nature, and thus combining several small and large effect QTLs is likely required to confer robust resistance across growth stages.

The genetic architecture of spot blotch resistance complicates the breeding task to develop resistant varieties. The Rcs6/Scs6 locus described above has been found to be tightly linked, yet not allelic, to Mla8, a major gene conferring resistance to powdery mildew (Leng et al. 2020). Further, the Rcs5 gene is closely linked and in repulsion with the adult plant resistance gene for leaf rust, Rph23, reported by Ziems et al. (2023). For optimum breeding outcomes, careful consideration needs to be taken prior to strongly selecting for or against these major resistance and susceptibility alleles. This further highlights the need for continual detection of resistance loci and preferably with minimal polytropic effects. To date, resistance allele identification has been approached using single marker QTL mapping methodologies that generally ignore linkage or the practical application of detected QTL in breeding programs.

Here, we explore two Australian elite breeding populations for response to spot blotch across growth stages and use advanced haplotype mapping approaches to identify chromosomal regions influencing susceptibility within the breeding program. Insights generated highlight the opportunity to breed for improved genetic resistance and reveal strategies for practical application in a breeding program.

Barley breeding populations

A panel of 337 elite two-row spring barley lines from the Australian Northern Region Barley breeding program (NRBBP) along with a small sub-set of Australian commercial cultivars (Supp. Table 1) were investigated in this study. Two distinct groups of materials were designated as breeding population one (BP1), which consisted of 141 lines sampled from Stage 2 of the NRBBP in 2012, and breeding population two (BP2), which consisted of 232 lines sampled from Stage 2 of the NRBBP in 2013 (Table 1). A total of 33 lines were in common across BP1 and BP2.

Table 1

Summary of the barley breeding populations assessed for spot blotch disease resistance.
Breeding population	Year	Environment	Growth stage	Genotypes
1	2012	Glasshouse, HRF^a, Warwick	Seedling	141
1	2012	Field, RRF^b	Adult	132
2	2013	Glasshouse, HRF^a, Warwick	Seedling	231
2	2013	Field, RRF^b	Adult	232

^α All seedling assessments were performed at Hermitage Research Facility, Warwick (HRF)

^b All field experiments were conducted at Redlands Research Facility (RRF)

Pathogen materials

The breeding populations were assessed for resistance to Bipolaris sorokiniana at both seedling (all stage) and adult plant stages, using pathotype SB61 at the Queensland Department of Agriculture and Fisheries (QDAF) in the Hermitage Research Facility (HRF), Warwick and Redlands Research Facility (RRF), Cleveland, Queensland, Australia respectively. SB61 is well described and prevalent throughout the growing region (Meldrum et al. 2004) and was originally collected on 2/10/1998 from Monto in Queensland.

Fungal inoculum preparation

To prepare conidia for inoculation, barley leaves infected with SB61 were sourced from storage in an ultra-low deep freezer set to -80°C. The leaves were in 2 cm lengths stored in water-tight tubes. To induce heat-shock, tubes were immediately immersed in a 40°C warm water bath for 3 minutes. Three leaf segments were placed in petri dishes containing one filter paper overlying a make-up removal pad that was wetted to runoff with 18 Ω nanopore filtered water. Petri dishes were placed on a temperature-controlled plate set to 19°C with 12 hours diurnal light that used blacklight blue and white fluorescent tubes to induce sporulation. Single conidia were transferred from infected leaves to Starch Nitrate agar (SNA) media (17.5 gm agar, 3.5 gm NaNO3, 1.165 gm K2HPO4, 0.585 gm MgSO4·7H2O, 0.585 gm KCl, 0.128 gm/L FeSO4·7H2O + EDTA, 11.665 gm starch (soluble), 1050 ml H2O (ionised)). All steps were performed in a laminar air flow cabinet. A suturing needle was sterilised with 100% ethanol, flamed, and then used to transfer six single conidia from the leaves to six positions within each plate of SNA media. The plates were placed in dark conditions at 25°C (± 1°C) for 7–9 days to allow for adequate growth of mycelia. V8 agar media was prepared for subculturing the fugus from the SNA culture plates, and was comprised of V8 juice (400 ml), CaCO3 (4 gm), distilled water (1600 ml) and agar (30 gm) to a final volume of 2L. Bipolaris sorokiniana mycelia were transferred from the SNA plates to V8 agar plates by using a sterilised 4 mm hole punch. After sub culturing, the plates were kept in the same dark conditions at 25°C (± 1°C) for 9–12 days for fungal growth for use in the experimental trials. To dislodge conidia into suspension, 5–10 mL of distilled water was applied to each plate and the surface was wiped with an artist’s camel-hair brush. The resulting spore suspensions were then filtered through a fine tea strainer (0.025–0.05 mesh) and diluted with purified Tween®- water (three drops of Tween®20 per 100 mL of purified water) to make up a standardised desired inoculum concentration of 10,000 conidia/ml, which was confirmed using a haemocytometer.

Screening for seedling-plant resistance in the glasshouse

BP1 and BP2 were evaluated for spot blotch resistance at the seedling stage in a temperature-controlled glasshouse facility at the Queensland Department of Agriculture and Fisheries (QDAF) in the Hermitage Research Facility, Warwick, Queensland, in 2012 and 2013. Barley seeds were sown into 10 cm diameter pots, filled with Searles® premium potting mix. The seedlings were inoculated 12 days after sowing, at approximately the Zadok scale growth stage 13 (Z13) (Zadoks et al. 1974), as described by Meldrum et al. (2004). The conidial suspensions was applied at a rate of 3 mL/pot using a spray paint gun, spraying with even passes from all sides. Plants were then transferred into an incubation chamber maintained at 99% humidity using an ultrasonic humidifier. Seedlings were incubated for 24 hours at 19ºC (± 4ºC) in the dark, then transferred back to the glasshouse for disease development. Phenotypic infection responses were recorded 12 days later (Z15 growth stage)

Meldrum et al. 2004) post inoculation using the 1–9 scale, where 1 = resistant and 9 = susceptible (Fig. 1).

Figure 1A. The 1–9 scale for spot blotch disease reactions in seedlings, where 1 = resistant, and 9 = susceptible. B. An elite barley line showing susceptibility to spot blotch in the field.

Field screening for adult-plant resistance

BP1 and BP2 were evaluated for spot blotch resistance under field conditions at the Queensland Department of Agriculture and Fisheries (QDAF) in the Redlands Research Facility, Queensland in 2012 and 2013. The mycelial broth was prepared by blending the V8 agar mycelium until a smooth consistency was reached and then the broth was strained through a fine mesh. Uptima Tween® 20 detergent was also added to aid adhesion to the leaves. The disease spreader blocks were pre-mist irrigated and the mycelial broth applied using a battery powered backpack. Plants were immediately covered with a tarpaulin for approximately 15 hours with regular short interval misting irrigation was conducted. These spreader rows became heavily infected after two months and were then used to inoculate the designated spreader rows within the hill plot screening nursery. Infected straw was frequently and evenly spread to inoculate the spreader rows in conjunction with sprinkler irrigation. Genotype responses to the spot blotch infection were recorded once sufficient differentiation was obtained among a known set of reference varieties.

Genotyping and marker curation

From the BP1 and BP2 populations, 337 unique barley lines were genotyped with DArT-seq markers using the Barley GBS 1.0 platform (http://www.diversityarrays.com), returning a total of 21,364 single nucleotide polymorphic (SNP) markers. SNP marker data was curated by excluding markers with more than 30% missing SNP calls and a minor allele frequency (MAF) of less than or equal to 5% in R (R Development Core Team 2022). Following curation, a total of 4,767 high-quality polymorphic SNP markers were retained and used for all downstream genomic analyses.

Population structure

Genetic distances were calculated between individuals using the modified Roger’s distances (Wright 1984). The resultant genetic distance matrix was then used in a principal component analysis (PCA) to visualise the population structure. Genotypes were assigned to groups for further exploration of the population structure by applying the k-means clustering approach to the first two principal components in R (Hartigan and Wong 1979, Voss-Fels et al. 2015).

Phenotypic data analysis

A linear mixed model was fit to the disease response data to account for spatial variation within the screening nurseries using ASReml-R (Butler et al. 2009). To obtain best linear unbiased estimates (BLUEs), genotype was fitted as a fixed effect and replicate as a random, while a spatial correlation variance structure was applied to the residual variance, using the auto-regressive one variance model (AR1), where genotypes were indexed by their column and row position. Predictions of genotype performance for disease response were estimated by generating BLUEs. Phenotypic correlations between disease response BLUEs obtained across four environments were calculated using the Pearson algorithm in the “PerformanceAnalytics” (v.1.4.3541) package (Peterson et al. 2018) in R.

QTL discovery using a single-marker GWAS approach

GWAS was performed in R using the GenABEL (v1.8-0) package (Aulchenko et al. 2007) for seedling and adult plant BLUEs across the experiments. The mixed linear model was adjusted for population stratification by including a kinship matrix to account for any variation created by the relatedness between genotype pairs and the first two principal components to account for population structure. For each of the four datasets, a Manhattan plot was created to display the association between the peak markers and spot blotch disease response across the seven chromosomes. The Bonferroni method (Sidak 1967) (p < 0.05) was used to calculate the threshold value -log₁₀(p) ≥ 4.78 to identify the significant marker-trait associations and the most strongly associated marker was referred to as the QTL “peak”.

Haplotype-based mapping using the local GEBV method

The local GEBV approach was performed using the SelectionTools 19.3 package in R (Voss-Fels et al. 2019). SNP markers (4,767) used in the single-marker GWAS approach were further filtered, removing those that could not be mapped to a chromosome and/or positioned within the chromosome, resulting in 3,010 markers. SNP markers were assigned to LD blocks by grouping markers based on pairwise r2 with a minimum LD threshold of r2 = 0.5. Prior to assigning LD blocks, the pairwise r2 values between SNP markers were first calculated across each chromosome, allowing selection of the highest LD among adjacent pairs of markers within each chromosome. For each data set, genome-wide single-marker effects were first calculated simultaneously using ridge-regression best linear unbiased predictions (RR-BLUP; Meuwissen et al. 2001). The initial RR-BLUP model ignores SNP effects of neighbouring markers in LD with the marker being estimated, to ensure that the SNP effects are not overestimated. This model attribute is important in the LD block effect calculation, ensuring single-marker effects are only captured once. Individual SNP effects within each LD block was then summed to calculate the individual haplotype effects within each LD block across the genome. The variance between haplotype effects within each LD block was then estimated and plotted as a Manhattan plot, where blocks with high variance were considered most associated with the trait (Voss-Fels et al. 2019).

The top 20 LD blocks for each environment were selected based on the highest variance. The allele contributing resistance or susceptibility at each marker within these LD blocks was determined along with the overall block effects, thus allowing blocks to be assigned to groups as either resistance or susceptibility blocks. Due to having a high number of lines within each nursery, the 2013 datasets were selected for downstream haplotype analysis focusing on the LD blocks with the highest variance for haplotype effects identified on chromosome 3H and 7H.

Haplotype network analysis

Haplotype network analysis was performed for the peak block identified on chromosome 3H. TCS genealogies between haplotype variants (Clement et al. 2000), were calculated using PopART (http://popart.otago.ac.nz.; Leigh and Bryant 2015). The network nodes were coloured according to the average disease rating in the respective haplotype groups.

Spot blotch disease varies across nurseries

A high degree of variation in disease responses was observed for BP1 and BP2 at all growth stages (Fig. 2A). Field nurseries assessing adult-plant stage resistance (SBa12 and SBa13) had a higher frequency of susceptible disease response scores than the seedling glasshouse nurseries (SBs12 and SBs13; Fig. 2B). This trend was consistently observed in the control lines, with the exception of the resistance reference controls (Supp. Table 1.)

All nurseries appear related and trend in the same direction for principal component (PC) 1, which accounts for 87.3% of the variation (Fig. 3A). PC2, accounting for almost the remainder of the variation, demonstrates a clear separation between the adult and seedling nurseries. A statistically significant (p < 0.05) and strong positive correlation was observed between both seedling nurseries (SBs12 and SBs13; r = 0.89; Fig. 3B). Similarly, a high correlation was observed between the disease responses from the adult plant nurseries (SBa12 and SBa13; r = 0.89). Interestingly, the correlation between seedling and adult phenotypes was lower, both in 2012 (SBs12 and SBa12; r = 0.67) and in 2013 (SBs13 and SBa13; r = 0.69), but still statistically significant.

Population structure and genetic relatedness

The PCA of the genetic distance between individuals in the breeding populations revealed three main clusters (Fig. 4). PC1 accounts for 12.9% of the variation and PC2 accounting for 8.9%, suggesting minimal population structure within the breeding population. Genotypes within the three clusters were investigated, where cluster 1 contained 106 lines (BP1 = 32, BP2 = 74; 8 common), cluster 2 contained 114 lines (BP1 = 57, BP2 = 57; 6 common), and cluster 3 contained 90 lines (BP1 = 45, BP2 = 45; 13 common), yet no definitive patterns could be identified in the absence of pedigree information (Fig. 4).

Single-marker genome-wide association mapping

A total of 24 markers were significantly associated with SB disease susceptibility across the four nurseries and detected on chromosomes 1H, 3H, 5H and 7H (Fig. 5; Supplementary Table 2), half of which were detected at the seedling stage and the remaining half at the adult plant stage. Of the QTLs identified, some were environment and growth stage specific, for example two QTLs on 1H were identified in 2012 for adult plant stage resistance. In contrast, the three QTLs on 7H at the peak markers 3398217, 3256980 and 3259148 were detected across all four environments and growth stages. Similarly, the QTL detected on 3H at the peak marker 3665142 was strongly associated to adult plant stage resistance across both years (Fig. 5).

LD blocks associated with spot blotch response

Characterisation of the LD structure across the chromosomes resulted in the classification of 1,610 LD blocks. The top 20 LD blocks explaining the highest proportion of variance for haplotype effect were selected for each seedling and adult plant nursery dataset (Fig. 6), resulting in 80 blocks. Of these, a total of 55 blocks were common across environments. Most interestingly, block b001471 located on 7H was detected across all environments and growth stages (Supplementary Table 3) with consistently high variance for haplotype effect. This block also contained the marker peak 3256980 that was previously detected across all environments in the single-marker analysis. Similarly, block b000531 was detected across both methodologies and across nurseries, yet also only contained a single marker. The block b000552, located on 3H, consistently explained a high proportion of variance across three environments (SBa12, SBs13 and SBa13; Supplementary Table 3) but was not detected in the single-marker analysis. Additional QTL were identified in the local GEBV method but not in the single-marker approach, for example the blocks on 1H in SBs12 and SBa13, and the blocks on the short and long arms of 6H. Higher variance for haplotype effects was observed in the seedling nurseries conducted under controlled conditions compared to the adult stage environments grown in the field.

Haplotype analyses and allele stacking

Of the 80 LD blocks identified, three were selected for further haplotype analysis due to their high variance for haplotype effects and their consistent detection across multiple environments. This included b000531 (3H-1) and b000552 (3H-2) both located on chromosome 3H, as well as b001471 on 7H. Both b000531 and b001471 contained only 1 marker and were also identified in the single-marker analysis (Supplementary Table 2). With all three blocks having an effect in the adult plant nurseries, the absence of susceptibility haplotypes at each block and effect on disease score was explored across both nurseries (Fig. 7). Consistent variance for disease scores was observed across SBa12 and SBa13, where large variance in disease score was observed when susceptibility alleles were absent in only one of the three blocks or in both 3H blocks (Fig. 7A and Fig. 7B). Interestingly, the largest reduction in disease score and variance was achieved when the 7H susceptibility haplotype was absent in combination with either or both 3H block’s susceptibility haplotypes. This suggests that minimising susceptibility haplotypes at the 7H block in combination with reducing other major genetic contributors to susceptibility is highly beneficial. In the breeding panel, 135 of the 337 genotypes carried the 7H susceptibility allele, demonstrating the population is enriched with the resistance allele.

The compounding effect of susceptibility haploblocks on disease score was explored in the SBa12 and SBa13 adult plant nurseries, and consistent results were observed across nurseries (Fig. 7C, D), where disease scores increased in a linear trend with the addition of susceptibility blocks. Very high levels of disease susceptibility were reached and plateaued once 13 and 14 susceptibility blocks were present within an individual in SBa12 and SBa13, respectively. Interestingly, there were a reasonable number of individuals within each nursery with greater than 13 susceptibility haploblocks, with x in SBa12 and x in SBa13. Yet, this is consistent with the general level of spot blotch susceptibility within Australian barley cultivars and is further supported by the limited number of individuals with four or less of the major susceptibility haplotypes across both nurseries. Individuals with four or less susceptibility haplotypes displayed moderately resistant to resistant disease scores across both adult plant nurseries (Fig. 7C, D).

Of the three major blocks selected for haplotype analysis, only 3H-2 (b000552), contained more than one marker and was therefore selected to further explore haplotype variation through network analysis. The high proportion of variance explained by the block along with its consistent effects across years at the adult growth stage also makes it an ideal candidate. Three haplotype variants for the block were evident (Fig. 8), where haplotype 1 (Hap 1) contained 61 genotypes, consisting of a mixture of commercial varieties and NRBBP lines, with a mean disease response of 7.41. In contrast, haplotype 2 (Hap 2) contained 26 lines, predominately originating from the NRBBP and North Dakota and had a mean of 6.52. The smallest haplotype group (Hap 3) consisted of only four lines that all originated from the NRBBP and had a mean disease response of 5.38, the lowest of all haplotypes (Fig. 8; Supplementary Table 4).

Useful sources of spot blotch resistance

Cultivars with durable genetic resistance to spot blotch are the preferred and most effective form of disease control (Gyawali 2010, Wang et al. 2017). Australian cultivars are well described as being susceptible to spot blotch and our understanding of whether Australian germplasm contains the previously described genetic sources of resistance is limited. This study is the first to explore the genetic architecture of spot botch susceptibility in the Australian breeding germplasm using both a single-marker and haplotype-based mapping approach. Two breeding populations from the NRBBP were evaluated with the same pathotype (SB61) in four environments, two of which explored the disease response in seedling glasshouse assays, and the other two in adult plants grown under field conditions. Strong phenotypic correlations were identified between environments, with the strongest found between those assessing similar growth stages and grown under similar conditions. In this study, we demonstrate that the majority of the germplasm evaluated was moderately to highly susceptible to spot blotch, and particularly when assessed as adult plants grown under field conditions. Importantly, variation for disease response was observed in both the seedling glasshouse and adult field-based assays. This suggests that genetic sources of resistance are present in Australian germplasm and below we discuss their nature and how they can be harnessed by applying new technology to enhance selection in breeding.

Genomic regions associated with spot blotch response

Two breeding populations from the NRBBP were evaluated as part of this study, yet the variation driving the three clusters in the PCA (Fig. 5) could not be explained by the two populations. This is likely a result of the two populations having shared ancestry and thus moderate relatedness due to being part of the same breeding program. However, limited pedigree information is available to validate the assumption of relatedness between the populations. Across both mapping approaches, we identified 29 chromosomal regions associated with spot blotch disease susceptibility, with reasonable consistency of results across methodologies. The highly significant QTL identified on chromosome 7H (b001471) using the single-marker approach was also detected using the haplotype-based approach. Furthermore, the 3H QTL identified in three out of the four environments was also identified across methods (b000531, 3H-1). Notably, the local GEBV approach identified additional haplo-blocks on 1H, 3H (3H-2) and 6H, which were not detected using the single-marker method, but were detected across environments. Our results demonstrate that the local GEBV approach in this situation is a superior method, as it detects the same significant QTLs in addition to novel LD blocks missed by the single-marker approach. The moderate level of LD within the population, likely a result of inbreeding and selection, plausibly provides greater power to detect significant marker trait associations across the haplotypes (Nielsen et al. 2004). Further, tightly linked loci generally undergo less recombination events, therefore this haplotype-based approach can be considered a more realistic representation of the effects when using significant marker-trait associations in breeding selection and progeny development.

The haplo-blocks detected on chromosomes 3H and 7H were in close proximity with QTL reported in the literature and catalogued resistance gene Rcs5, respectively (Steffenson et al. 1996, Bilgic et al. 2005, Bovill et al. 2010, Roy et al. 2010, Zhou and Steffenson 2013). More specifically, these two blocks appear to co-locate with two seedling resistance QTLs (SRT-SB54-3 and SRT-ICSB3-10) reported by Visioni et al. (2020) for two differing pathotypes. Interestingly, our study observed resistance at both loci for seedling and adult plant resistance, yet this could be driven by the differing pathotype. Rcs5 is a major gene located on chromosome 7H and it confers seedling resistance to spot blotch with strong origins in barley germplasm bred by the United States North Dakota barley breeding program (Steffenson et al. 1996). It is well-known the NRBBP utilised germplasm from the North Dakota breeding program as foundational parents, so it is not unexpected that Rcs5 is present within this population, but this is the first study to provide confirmation. The additional block on chromosome 1H identified using the LGEBV method also coincide with Rcs6 and several other reported QTLs in the region (Bilgic et al. 2006, Grewal et al. 2012, Zhou and Steffenson 2013, Haas et al. 2016). Similarly, the block identified on the short arm of chromosome 6H may also align with Rbs7, however the block on the long arm of 6H is likely novel. The differing molecular marker genotyping platforms used in the literature made it nearly impossible to accurately align the QTLs detected in this study, however, based on their approximate position on the chromosome the alignments seem plausible, yet this requires validation.

Allele stacking to achieve high levels of resistance

The results of this study along with the earlier literature clearly demonstrate that the genetic architecture of spot blotch is highly complex and quantitative in nature. Although some previous success has been achieved using marker-assisted selection (MAS; Kottapalli et al. 2010, Castro et al. 2012), to achieve a durable level of resistance across growth stages, a multi-allele stacking or whole genome selection approach could be the most effective breeding strategy. Outcomes of this study support a quantitative approach, where high disease scores and variance were observed when each of the three major susceptibility blocks were absent on an individual basis, or when both the 3H susceptibility blocks were absent. Interestingly, when the 7H block was combined with either or both the 3H blocks, a substantial reduction in disease score and variance were observed (Fig. 7A, B). Based on chromosomal location, we speculate that our 7H block contains the major all-stage and adult plant resistance gene, Rcs5, that has been widely identified in a range of global populations and is of high frequency in the germplasm originating from North Dakota. The Rcs5 region, is well described as the principal contributor to all-stage, seedling resistance, however the regions’ ability to confer adult-plant resistance is variable, and previous research suggests a strong gene x environment and/or gene x gene interaction could influence resistance levels (Bovill et al. 2010). With the similarity in results from our two adult plant nurseries, no gene x environmental effects were observed, however a clear gene x gene interaction was observed when combining our 7H block with other major sources of adult-plant resistance. A combinatorial approach is likely the most effective method to harness adult-plant resistance from the Rsc5 region, and our global stacking of resistance haploblocks (Fig. 7C, 7D) suggests a quantitative approach to minimising susceptibility haplotypes is likely the most effective approach to rapidly achieve spot blotch resistance in Australian barley.

A significant proportion of the breeding population in the NRBBP contains susceptibility haplotypes at both the 3H and 7H blocks, despite active selection for spot blotch disease resistance within the breeding program. Since breeding programs routinely select for numerous traits such as yield, multiple disease resistance and end-use grain quality traits, and either explicitly or innately apply a selection weighting, it is possible that the major haplo-blocks could have pleiotropic effects influencing multiple traits. Previous research supports the possible negative pleiotropic nature of key spot blotch susceptibility genes, where Scs6 and Rcs5 are reported to be negatively linked to powdery mildew and leaf rust resistance, respectively, and both of which genes have likely been identified in the NRBBP. As a result, careful consideration needs to be given before strong selection against the major susceptibility haploblocks identified in this study. However, our results demonstrate that genetic resistance is highly quantitative, and as such a reasonable level of disease resistance could still be achieved in the presence of the major susceptibility blocks through the minimisation of other susceptibility alleles (Fig. 7C, D). Further research simulating the presence of the major susceptibility blocks identified on chromosome 1H, 3H, 6H and 7H, and exploring the gradual minimisation of alternative susceptibility haplotypes to determine the number required to meet adequate levels of resistance would be informative for resistance breeding. This potential negative linkage highlights the importance of continually exploring diverse germplasm for novel sources of resistance to spot blotch and using new breeding technologies for rapid introgression. Our study is a clear example of the power of the local GEBV approach to accurately depict the genetic architecture of quantitative traits and it also demonstrates the utility for breeders to select for or against multiple haplotypes.

This is one of the first studies to explore resistance to spot blotch in an Australian barely breeding program, and we demonstrate there is variation for resistance within Australian germplasm that can be harnessed by breeders for cultivar improvement. This is a significant outcome for the industry, where the alternative approach of introgressing diverse resistant germplasm is time consuming and can take numerous years for the germplasm to assimilate within the breeding program. Our study further reinforces the quantitative genetic nature of spot blotch resistance and for the first time explores susceptibility haplotypes within LD blocks, as opposed to the traditional single-marker mapping approach. We demonstrate the power of this haplotype-based approach through the detection of consistent marker-trait associations between methodologies as well as the detection of novel LD blocks not detected in the single-marker approach or across the literature. Finally, we show the utility of a haplotype stacking in selecting against major susceptibility haplotypes, whereby resistance can be achieved in the presence or absence of the major haplo-blocks if the overall number of susceptibility blocks are kept to a minimum. However, considering the negative trait linkage previously reported for spot blotch resistance, selection against major susceptibility haplotypes requires careful consideration and a holistic trait selection approach is likely preferable. In a time when developing spot blotch resistant cultivars is more imperative than ever, our study reveals novel insight for Australian barley breeding and applies an innovative yet practical approach that is highly translatable to the industry.

Acknowledgments

This research was supported by the Grains Research and Development Corporation of Australia (DAQ00190). We thank the Department of Agriculture and Fisheries for technical assistance in the field.

Funding

This research was supported by the Grains Research and Development Corporation of Australia (DAQ00190).

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.

Author Contributions

All authors contributed to the study conception and design. Material preparation and data collection were performed by Ryan Fowler, Greg Platz and Jerry Franckowiak. Analyses were performed by Dipika Roy, Eric Dinglasan, Kai Voss-Fels and Hannah Robinson. The first draft of the manuscript was written by Dipika Roy and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Data Availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Aulchenko YS, Ripke S, Isaacs A, Van Duijn CMJB (2007) GenABEL: an R library for genome-wide association analysis. Bioinformatics 10:1294-1296. https://doi.org/10.1093/bioinformatics/btm108
Bastas KK (2022) Impact of climate change on food security and plant disease. In: Microbial biocontrol: food security and post-harvest management, 2^nd edn. Springer, Berlin, pp 1-22
Bilgic H, Steffenson BJ, Hayes PM (2005) Comprehensive genetic analyses reveal differential expression of spot blotch resistance in four populations of barley. Theor Appl Genet 7:1238-1250. https://doi.org/10.1007/s00122-005-0023-2
Bilgic H, Steffenson BJ, Hayes PM (2006) Molecular mapping of loci conferring resistance to different pathotypes of the spot blotch pathogen in barley." Phytopath 7:699-708. https://doi.org/10.1094/PHYTO-96-0699
Bovill J, Lehmensiek A, Sutherland MW, Platz GJ, Usher T, Franckowiak J, Mace E (2010) Mapping spot blotch resistance genes in four barley populations. Mol Breeding 4:653-666. https://doi.org/10.1007/s11032-010-9401-9
Butler D, Cullis BR, Gilmour A, Gogel BJTSQ (2009) ASReml-R reference manual. Department of Primary Industries and B. Fisheries. https://asreml.kb.vsni.co.uk/wp-content/uploads/sites/3/ASReml-R-Reference-Manual-4.2.pdf. Accessed 20 February 2023
Castro AJ, Gamba F, German S, Gonzalez S, Hayes PM, Pereyra S, Perez C (2012) Quantitative trait locus analysis of spot blotch and leaf rust resistance in the BCD47 × Baronesse barley mapping population. Plant Breeding 2:258-266. https://doi.org/10.1111/j.1439-0523.2011.01930
Grewal TS, Rossnagel BG, Scoles GJ (2012) Mapping quantitative trait loci associated with spot blotch and net blotch resistance in a doubled-haploid barley population. Mol Breeding 1:267-279. https://doi.org/10.1007/s11032-011-9616-4
Gupta PK, Vasistha NK, Aggarwal R, Joshi AK (2018) Biology of B. sorokiniana (syn. Cochliobolus sativus) in genomics era. Plant Biochem Biotechnol. 2:123-138. https://doi.org/10.1007/s13562-017-0426-6
Gyawali S (2010) Association mapping of resistance to common root rot and spot blotch in barley, and population genetics of Cochliobolus sativus. Dissertation, North Dakota State University.
Gyawali S, Chao S, Vaish SS, Singh SP, Rehman S, Vishwakarma SR, Verma RPS (2018) Genome wide association studies (GWAS) of spot blotch resistance at the seedling and the adult plant stages in a collection of spring barley. Mol Breeding 5: 62. https://doi.org/10.1007/s11032-018-0815-0
Haas M, Menke J, Chao S, Steffenson BJ (2016) Mapping quantitative trait loci conferring resistance to a widely virulent isolate of Cochliobolus sativus in wild barley accession PI 466423. Theor Appl Genet 10:1831-1842. https://doi.org/10.1007/s00122-016-2742-y
Hartigan JA, Wong MA (1979) Algorithm AS 136: A k-means clustering algorithm. J R Stat 1:100-108.
Kaur S, Bhardwaj RD, Kaur J, Kaur S (2022) Induction of defense-related enzymes and pathogenesis-related proteins imparts resistance to barley genotypes against spot blotch disease. J Plant Growth Regul 2:682-696. https://doi.org/10.1007/s00344-021-10333-2
Knight N, Platz G, Lehmensiek A, Sutherland MJ (2010) An investigation of genetic variation among Australian isolates of Bipolaris sorokiniana from different cereal tissues and comparison of their abilities to cause spot blotch on barley. Australasian Plant Pathol. 3:207-216. https://doi.org/10.1071/AP09082
Kottapalli KR, Narasu ML, Jena KK (2010) Effective strategy for pyramiding three bacterial blight resistance genes into fine grain rice cultivar, Samba Mahsuri, using sequence tagged site markers. Biotechnol Lett 7:989-996. https://doi.org/10.1007/s10529-010-0249-1
Lander ES, Botstein JP (1986) Strategies for studying heterogeneous genetic traits in humans by using a linkage map of restriction fragment length polymorphisms. Proc Natl Acad Sci USA 19: 7353-7357. https://doi.org/10.1073/pnas.83.19.7353
Leng Y, Zhao M, Fiedler J, Dreiseitl A, Chao S, Li X, Zhong SJP (2020) Molecular mapping of loci conferring susceptibility to spot blotch and resistance to powdery mildew in barley using the sequencing-based genotyping approach. Phytopath 2:440-446. https://doi.org/10.1094/PHYTO-08-19-0292-R
Martínez M, Biganzoli F, Arata A, Dinolfo MI, Rojas D, Cristos D, Stenglein SJA (2022) Warm nights increase Fusarium Head Blight negative impact on barley and wheat grains. Agric Meteorol 318:108909. https://doi.org/10.1016/j.agrformet.2022.108909
Meldrum S, Platz DG, Ogle H (2004) Pathotypes of Cochliobolus sativus on barley in Australia. Australasian Plant Pathol. 1:109-114. https://doi.org/10.1071/AP03088
Meuwissen TH, Hayes BJ, Goddard ME (2001) Prediction of total genetic value using genome-wide dense marker maps. Genetics 4: 1819-29. https://doi.org/1819-1829. 10.1093/genetics/157.4.1819
Mir RR, Zaman-Allah M, Sreenivasulu N, Trethowan R, Varshney RK (2012) Integrated genomics, physiology and breeding approaches for improving drought tolerance in crops. Theor Appl Genet 4:625-645. https://doi.org/10.1007/s00122-012-1904-9
Nielsen DM, Ehm MG, Zaykin DV, Weir BS (2004) Effect of two-and three-locus linkage disequilibrium on the power to detect marker/phenotype associations. Genetics 2: 1029-40. https://doi.org/1029-1040. 10.1534/genetics.103.022335
Peterson BG, Carl P, Boudt K, Bennett R, Ulrich J, Zivot E, Cornilly D, Hung E, Lestel M, Balkissoon KJ (2018) PerformanceAnalytics:Econometric tools for performance and risk analysis. Rpackages. https://cran.rproject.org/web/packages/PerformanceAnalytics/PerformanceAnalytics.pdf
R Development Core Team (2022) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. https://www.R-project.org/.
Roy JK, Smith KP, Muehlbauer GJ, Chao S, Close TJ, Steffenson BJ (2010) Association mapping of spot blotch resistance in wild barley. Mol Breeding 2:243-256. https://doi.org/10.1007/s11032-010-9402-8
Sidak Z (1967) Rectangular confidence regions for the means of multivariate normal distributions. J Am Stat Assoc 318:626-633. https://doi.org/10.2307/2283989
Steffenson BJ, Hayes PM, Kleinhofs A (1996) Genetics of seedling and adult plant resistance to net blotch (Pyrenophora teres f. teres) and spot blotch (Cochliobolus sativus) in barley. Theor Appl Genet 5:552-558. https://doi.org/10.1007/BF00224557
Visioni A, Rehman S, Viash SS, Singh SP, Vishwakarma R, Gyawali S, Al-Abdallat AM, Verma RPS (2020) Genome wide association mapping of spot blotch resistance at seedling and adult plant stages in barley. Front Plant Sci 11: 642. https://doi.org/10.3389/fpls.2020.00642
Voss‐Fels K, Frisch M, Qian L, Kontowski S, Friedt W, Gottwald S, Snowdon RJ (2015) Subgenomic diversity patterns caused by directional selection in bread wheat gene pools. Plant Genome 2. https://doi.org/10.3835/plantgenome2015.03.0013
Voss-Fels KP, Stahl A, Wittkop B, Lichthardt C, Nagler S, Rose T, Chen TW, Zetzsche H, Seddig S, Baig MM (2019) Breeding improves wheat productivity under contrasting agrochemical input levels. Nat Plants 7:706-714. https://doi.org/706-714. 10.1038/s41477-019-0445-5
Wang R (2015) Genetic and molecular characterization of genes involved in barley-Cochliobolus sativus interaction. Dissertation, North Dakota State University.
Wang R, Leng Y, Ali S, Wang M, Zhong S (2017) Genome-wide association mapping of spot blotch resistance to three different pathotypes of Cochliobolus sativus in the USDA barley core collection. Mol Breeding 4:44. https://doi.org/10.1007/s11032-017-0626-8
Wang R, Leng Y, Zhao M, Zhong SJT (2019) Fine mapping of a dominant gene conferring resistance to spot blotch caused by a new pathotype of Bipolaris sorokiniana in barley. Theor Appl Genet 1:41-51. https://doi.org/10.1007/s00122-018-3192-5
Wright S (1984) Evolution and the genetics of populations: variability within and among natural populations, 4^th edn. University of Chicago Press, Chicago.
Zadoks JC, Chang TT, Konzak CF (1974) A decimal code for the growth stages of cereals. Weed Res 6:415-421. https://doi.org/10.1111/j.1365-3180.1974.tb01084.x
Zhou H, Steffenson B (2013) Genome-wide association mapping reveals genetic architecture of durable spot blotch resistance in US barley breeding germplasm. Mol Breeding 1:139-154. https://doi.org/10.1007/s11032-013-9858-4
Ziems LA, Lovepreet S, Dracatos PM, Dieters MJ, Sanchez-Garcia M, Amri A, Verma RPS, Park RF, Singh D (2023) Characterization of leaf rust resistance in international barley germplasm using genome-wide association studies. Plants 4:862. https://doi.org/10.3390/plants12040862

RoyetalSupplementaryTables.xlsx

Download PDF

Reviewers agreed at journal
02 Aug, 2024
Reviewers invited by journal
02 Aug, 2024
Editor assigned by journal
22 Feb, 2024
First submitted to journal
22 Feb, 2024

You are reading this latest preprint version

Genomic regions associated with spot blotch resistance in elite barley breeding populations

Status:

Version 1

Abstract

Figures

Introduction

Materials and Methods

Barley breeding populations

Pathogen materials

Fungal inoculum preparation

Screening for seedling-plant resistance in the glasshouse

Field screening for adult-plant resistance

Genotyping and marker curation

Population structure

Phenotypic data analysis

QTL discovery using a single-marker GWAS approach

Haplotype-based mapping using the local GEBV method

Haplotype network analysis

Results

Spot blotch disease varies across nurseries

Population structure and genetic relatedness

Single-marker genome-wide association mapping

LD blocks associated with spot blotch response

Haplotype analyses and allele stacking

Discussion

Useful sources of spot blotch resistance

Genomic regions associated with spot blotch response

Allele stacking to achieve high levels of resistance

Declarations

References

Supplementary Files

Status:

Version 1