Origin, inheritance, and gene regulatory consequences of genome dominance in polyploids

Significance Ancient plant polyploids contain dissimilar subgenomes. Subgenomes that have lost fewer genes are subgenomes that tend to express their genes to higher mRNA levels: “genome dominance.” Genome dominance is heritable through multiple rounds of polyploidy and over tens of millions of years. Twenty-four–nucleotide RNA coverage of noncoding, transposon DNA upstream of genes marks the recessive subgenomes of Brassica rapa and Arabidopsis with only a small effect of gene expression. We hypothesize that the “diploid” parent of a tetraploidy with the lowest transposon load was to become the dominant subgenome. The balance of transposon position effects on genes is important for the regulation of quantity (and perhaps to solve the heterosis and C-value paradoxes). Whole-genome duplications happen repeatedly in a typical flowering plant lineage. Following most ancient tetraploidies, the two subgenomes are distinguishable because one subgenome, the dominant subgenome, tends to have more genes than the other subgenome. Additionally, among retained pairs, the gene on the dominant subgenome tends to be expressed more than its recessive homeolog. Using comparative genomics, we show that genome dominance is heritable. The dominant subgenome of one postpolyploidy event remains dominant through a subsequent polyploidy event. We show that transposon-derived 24-nt RNAs target and cover the upstream region of retained genes preferentially when located on the recessive subgenome, and with little regard for a gene’s level of expression. We hypothesize that small RNA (smRNA)-mediated silencing of transposons near genes causes position-effect down-regulation. Unlike 24-nt smRNA coverage, transposon coverage tracks gene expression, so not all transposons behave identically. We propose that successful ancient tetraploids begin as wide crosses between two lines, each evolved for different tradeoffs between transposon silencing and negative position effects on gene expression. We hypothesize that following a chaotic wide-cross/new tetraploid period, genes acquire their new expression balances based on differences in transposon coverage in the parents. We envision patches of silenceable transposon as quantitative cis-regulators of baseline transcription rate. Attractive solutions to heterosis and the C-value paradox are mentioned.