Genome Wide Association Mapping for Arabinoxylan Content in a Collection of Tetraploid Wheats

Abstract

Background: Arabinoxylans (AXs) are major components of plant cell walls in bread wheat and are important in bread-making and starch extraction. Furthermore, arabinoxylans are components of soluble dietary fibre that has potential health-promoting effects in human nutrition. Despite their high value for human health, few studies have been carried out on the genetics of AX content in durum wheat.

Results: The genetic variability of AX content was investigated in a set of 104 tetraploid wheat genotypes and regions attributable to AX content were identified through a genome wide association study (GWAS). The amount of arabinoxylan, expressed as percentage (w/w) of the dry weight of the kernel, ranged from 1.8% to 5.5% with a mean value of 4.0%. The GWAS revealed a total of 37 significant marker-trait associations (MTA), identifying 19 quantitative trait loci (QTL) associated with AX content. The highest number of MTAs was identified on chromosome 5A (seven), where three QTL regions were associated with AX content, while the lowest number of MTAs was detected on chromosomes 2B and 4B, where only one MTA identified a single locus. Conservation of synteny between SNP marker sequences and the annotated genes and proteins in Brachypodium distachyon, Oryza sativa and Sorghum bicolor allowed the identification of nine QTL coincident with candidate genes. These included a glycosyl hydrolase GH35, which encodes Gal7 and a glucosyltransferase GT31 on chromosome 1A; a cluster of GT1 genes on chromosome 2B that includes TaUGT1 and cisZog1; a glycosyl hydrolase that encodes a CelC gene on chromosome 3A; Ugt12887 and TaUGT1genes on chromosome 5A; a (1,3)-β-D-glucan synthase (Gsl12 gene) and a glucosyl hydrolase (Cel8 gene) on chromosome 7A.

Conclusions: This study identifies significant MTAs for the AX content in the grain of tetraploid wheat genotypes. We propose that these may be used for molecular breeding of durum wheat varieties with higher soluble fibre content.

Fig 1. Arabinoxylan content evaluated in a collection of 104 Triticum turgidum genotypes.

The reported values are the mean of three biological replicates grown at Valenzano 2012.

Fig 2. Unrooted Bayesian tree based on 22,106 SNPs spread across the wheat A and B genomes.

The tree was constructed from Maximum Likelihood and Bayesian Inference algorithms. Bootstrap values are a percentage of 1000 replicates.

Fig 3. Manhattan plot of grain AX content from GWAS using the Kinship relationship model.

The-log10 (p-values) from the GWAS are plotted according to genetic position on each of the 7 wheat chromosome pairs. Positions of candidate genes for AX content are indicated by black downward arrows.

Fig 4. Schematic representation of the seven tetraploid wheat homoeologous chromosome groups analyzed for genome wide association scan.

The seven tetraploid wheat homoeologous chromosome groups are characterized by the 22,106 markers used for the genome wide association scan. Chromosome scales are based on the marker position reported by Wang et al. (2014) and each figure was obtained using GenStat 15^th edition. On the left side of each chromosome, the putative QTL (indicated by the blue arrows) are reported. The first and the last marker of the QTL region (in black), SNPs with the LOD score ≥3 (in green), and SNPs located in gene sequences encoding AX biosynthesis enzymes (in pink) are indicated on the right hand side.

Fig 5. Expression analysis from PLEXdb database of candidate genes for AX biosynthesis.

Gene expression measurements are reported on three developmental tissues (caryopsis, embryo and endosperm) for Chinese Spring using the Affymetrix Wheat GeneChip. The Affymetrix code TaAffx.17044.1.S1_at corresponded to the gene cisZog2B, Ta.23909.1.S1_at and Ta.23909.1.S1_a_at for TaUGT1 gene, Ta.25520.1.S1_at and Ta.21296.1.S1_at for Gsl12, Ta.4878.1.S1_at and TaAffx.6809.1.S1_at for Utg12887.