Insight into candidate Fusarium head blight resistance genes associated with a quantitative trait locus on chromosome 1A of Triticum turgidum ssp. Carthlicum

Triticum turgidum ssp. carthlicum line Blackbird is a tetraploid species relative of durum wheat that offers partial Fusarium head blight (FHB) resistance. Through QTL mapping of a population from a cross between durum wheat cv. Strongfield and Blackbird, we previously reported a moderate effect FHB resistance QTL on the telomeric region of chromosome 1AS with consistent expression over multiple environments. The objective of this study was to identify within this QTL candidate FHB resistance genes, and single nucleotide polymorphisms (SNPs) for high resolution mapping. Infection site specific mRNA and exome capture sequencing of Blackbird, Strongfield and a resistant and a susceptible transgressive phenotype progeny of the Strongfield/Blackbird population was performed. Co-expression network analysis identified five gene networks significantly associated with the resistance to Type II FHB resistance. The five candidate hub genes within the QTL interval belonged to three separate gene networks and were orthologues of peroxisomal membrane protein (PEX14), nucleotide binding-leucine-rich repeat (two genes), Bowman-Birk type trypsin inhibitor and proline-rich receptorlike protein kinase (PERK9) genes. PEX14 carried two predicted splice receptor SNPs between cv. Strongfield and Blackbird, which along with its role in plant defense make it a candidate FHB resistance gene. The differential expression analysis allowed identification of 29 additional candidate genes within the interval with an orthologue of powdery mildew 3 (Pm3)-like disease resistance protein carrying a predicted high-impact splice donor variant. Most of these candidate genes were previously detected in the five FHB-wheat interaction studies available in the Wheat Expression Browser (http://www.wheat-expression.com/), e.g. the gene encoding a Bowman-Birk serine proteinase inhibitor was detected in four out of the six studies, while 12 candidate genes were unique to our study. Phenotyping tilling lines with high-impact mutations in the candidate genes along with a set of Near Isogenic lines is expected to contribute to the high-resolution mapping of this QTL.