Synthetic hexaploid wheat: to use or not to use?

Two polyploidization events were involved in the evolution of current day bread wheat (Triticum aestivum;
AABBDD). The chance hybridization events involving limited Aegilops tauschii individuals resulted in
narrow genetic diversity of the ‘D’ subgenome relative to the other subgenomes, and both domestication
and breeding processes further depleted this already limited diversity. Synthetic hexaploid wheat (SHW)
lines are generated to overcome this genetic bottleneck and make use of the untapped genetic diversity
of the Ae. tauschii genepool. However, the phenotypes observed in the parents are not always retrieved
in the SHW backgrounds, possibly due to inter-subgenome interactions. We generated RNA-Seq data
from four SHW lines and their corresponding tetraploid and diploid parents to investigate this postpolyploidization genome reprogramming scenario. This transcriptomics study involved eight genotypes,
ten tissues and three biological replicates for a total of 240 RNA samples. The expression bias among
homoeologues in triad state (1:1:1) was compared between the SHW lines and their parents, which would
be the in-silico SHW-like scenario expected if there were no inter-subgenome interactions. The results
suggest a large-scale suppression of homoeoalleles of the ‘D’ subgenome upon polyploidization, i.e., in
SHWs. The largest proportion of the >18,000 triads analysed did not display any tissue-specific
homoeologues, while the next largest fractions were triads where all homoeologues showed same tissue
specificity and triads where only one homoeologue showed tissue specificity. In all four SHW lines, the
largest subsets of triads where all homoeologues showed tissue-specificity were in the root and mature
anther tissues. A moderate relationship was also observed between the tissue-specificity of the triads and
the expression bias in multiple SHW-tissue contexts. A major fraction of triads also maintained similarly
biased patterns in the domesticated landrace Chinese Spring and the cultivar Azhurnaya. Comparative
transcriptome analysis of the complete gene set further validated the large-scale suppression of ‘D’
homoeoalleles upon polyploidization. Further, differential splicing analysis of the high-confidence genes
identified all five splice events, suggesting that not only gene suppression but also qualitative differences
in transcripts occur upon polyploidization. Both large-scale suppression and splice site variation have
implication in SHW-derived breeding because they can impede trait expression in bread wheat.