Variation in AX Structure
AX comprises a backbone of β-D-xylopyranosyl residues linked through (1,4) glycosidic linkages, with some residues being substituted with α-L-arabinofuranosyl residues at either position 3 or positions 2 and 3. Another key attribute of AX struc-
ture is the degree of feruloylation of some of the three-linked arabinose. The solubility of AX is dictated by degree of arabinose substitution (increasing solubility), cross-linking of AX chains by intermolecular dimerization of ferulate (decreasing solubility) and the chain length of AX backbone (longer chains are less soluble) (Saulnier et al. 2007). Any of these attributes may influence the benefits that AX confers in the diet, either directly or by altering the solubility of AX. There is genetic variation in the proportions of monosubstituted and disubstituted xylose residues (Saulnier et al. 2007) but less is known about genetic variation of the other AX structural attributes. AX feruloylation, measured as total bound ester-linked ferulic acid, is highly variable with low heritability, but the extent of genetic variation within white flour (where levels are much lower) is unknown. More research is needed both of the structural attributes of AX which confer greatest benefit in human diet and on the wheat genes responsible for them to determine the importance and feasibility of genetic improvement of AX structure.
Genes Responsible for AX Synthesis in Wheat Endosperm
Identification of Candidate Genes
Xylan is more abundant in primary cell walls of grasses than those of dicots and arabinose substitution is far more common (the main substitution in dicot xylan being glucuronic acid). Feruloylation of AX occurs exclusively in grasses and other commelinid monocots. Since xylan is a highly abundant polymer within plant biomass, we expect transcripts encoding the synthetic enzymes to be highly abundant; exploiting this fact and the differences between grasses and dicots led us to identify candidate genes for the synthesis and feruloylation of AX (Mitchell et al. 2007). The results are summarized in Fig. 46.1; we identified clades within the glycosyl transferase (GT) gene families 43, 47 and 61 as likely involved in the synthesis of AX, and a clade from the BAHD acyl coA transferase superfamily as likely responsible for AX feruloylation.
Using RNA-Seq transcriptome analyses of developing wheat starchy endosperm, we found transcripts within all of these clades during the period of greatest AX synthesis (Pellny et al. 2012). In order to demonstrate their function, we transformed wheat to suppress their expression using an endosperm-specific promoter driving RNAi constructs.
Role of GT61 Genes in AX Synthesis
We compared five homozygous transgenic lines suppressing the most highly expressed GT61 gene in wheat endosperm with null segregant controls. Digestion of AX with xylanase and HPAEC quantification of resultant oligosaccharides showed that those oligosaccharides with single three-linked arabinose substitution were substantially decreased and 1H-NMR showed almost complete abolition of this linkage in undigested soluble AX (Anders et al. 2012; Fig. 46.2). Furthermore our collaborators (Dupree group, University of Cambridge, UK) showed that expression of similar wheat and rice genes in Arabidopsis introduced arabinose linkages onto Arabidopsis xylan (Anders et al. 2012). Therefore these GT61 genes encode xylan arabinosyl transferases, consistent with their far greater expression in grasses than dicots (Fig. 46.1).
Fig. 46.1 Histogram of results of analyses in Mitchell et al. (2007). The number of orthologous groups are shown for each bin of log ratio of expression in cereals to that in dicots. Clades from families with greater expression in cereals and correct characteristics for AX synthesis (inverting GT families) and feruloylation (acyl transferases) are indicated
Fig. 46.2 Analysis of xylan structure in endosperm samples from homozygous TaGT61_1 RNAi transgenic wheat. (a) Oligosaccharide abundance from transgenic samples relative to corresponding azygous controls after xylanase digest; mean of five independent lines ±95 % confidence intervals. Columns are colored according to oligosaccharide substitution: unsubstituted (green), mono-substituted only (red), di-substituted only (blue), and monoand di-substituted (purple). (b) 1H-NMR spectra for transgenic (red) and azygous control (blue) samples showing H1 signals for arabinose in AX: A3-Xmono denotes α-(1,3)–linked to mono-substituted xylose (Redrawn from Anders et al. 2012)
Role of GT43 and GT47 Genes in AX Synthesis
We compared homozygous transgenic lines suppressing either the most highly expressed GT43 (TaGT43_2) or GT47 (TaGT47_2) genes in wheat endosperm with their null segregant controls. Suppression of either gene had similar effects with AX content decreased to about 50 % of controls as determined by monosaccharide composition of non-starch polysaccharide (Fig. 46.3a; Lovegrove et al. 2013).
Fig. 46.3 Summary of results in Lovegrove et al. (2013) from RNAi suppression of TaGT43_2 and TaGT47_2 genes in wheat endosperm. (a, b) Total AX abundance (a) and arabinose to xylose ratio (b) from transgenic samples and null-segregant controls determined by monosaccharide analyses of non-starch polysaccharide; mean of three independent lines ±SE. (c) Log intrinsic viscosity of soluble AX fractions (which is proportional to log AX chain length; Dervilly-Pinel et al. 2001) separated by SE-HPLC versus concentration for samples from TaGT43_2 and a TaGT47_2 RNAi lines and corresponding controls
In contrast to suppression of TaGT61_1, the amount of arabinose substitution was increased (Fig. 46.3b). The TaGT43_2 and TaGT47_2 genes are orthologous to IRX9 and IRX10 in Arabidopsis, respectively. Mutations in IRX9 and IRX10 lead to decreased xylan and xylan chain length in Arabidopsis (Pena et al. 2007; Brown et al. 2009) and we obtained similar results by suppressing their orthologues in wheat endosperm. However, size-exclusion profiles of AX showed that suppressing TaGT43_2 had a greater effect on the longest chains of AX than suppressing TaGT47_2 (Fig. 46.3c).