Transcription factor retention through multiple polyploidisation steps in wheat.

Whole genome duplication is widespread in plant evolutionary history and is followed by non-random gene loss to return to a diploid state. Across multiple angiosperm species, the retained genes tend to be dosage-sensitive regulatory genes such as transcription factors, yet data for younger polyploid species is sparse. Here we analysed the retention, expression and genetic variation in transcription factors in the recent allohexaploid bread wheat (Triticum aestivum L.). By comparing diploid, tetraploid and hexaploid wheats we found that, following each of two hybridisation and whole genome duplication events, the proportion of transcription factors in the genome increased. Transcription factors were preferentially retained over other genes as homoeologous groups in tetraploid and hexaploid wheat. Across cultivars, transcription factor homoeologs contained fewer deleterious missense mutations than non-transcription factors, suggesting that transcription factors are maintained as three functional homoeologs in hexaploid wheat populations. Transcription factor homoeologs were more strongly co-expressed than non-transcription factors, indicating conservation of function between homoeologs. We found that the B3, MADS-M-type and NAC transcription factor families were less likely to have three homoeologs present than other families, which was associated with low expression levels and high levels of tandem duplication. Together, our results show that transcription factors are preferentially retained in polyploid wheat genomes although there is variation between families. Knocking out one transcription factor homoeolog to alter gene dosage, using TILLING or CRISPR, could generate new phenotypes for wheat breeding.