Gene Expression Dominance in Allopolyploids: Hypotheses and Models

Abstract

The classical example of nonadditive contributions of the two parents to allopolyploids is nucleolar dominance, which entails silencing of one parental set of ribosomal RNA genes. This has been observed for many other loci. The prevailing explanation for this genome-wide expression disparity is that the two merged genomes differ in their transposable element (TE) complement and in their level of TE-mediated repression of gene expression. Alternatively, and not exclusively, gene expression dominance may arise from mismatches between trans effectors and their targets. Here, we explore quantitative models of regulatory mismatches leading to gene expression dominance. We also suggest that, when pairs of merged genomes are similar from one allopolyploidization event to another, gene-level and genome dominance patterns should also be similar.

The causes and consequences of subgenome dominance in hybrids and recent polyploids

“…Furthermore, gene expression dominance increased over successive generations with the greatest bias between subgenomes observed in the naturally established 140‐yr‐old allopolyploid. This finding is in line with newly developed mathematical models of progressive transcription factor binding and regulatory mismatch leading to expression dominance over successive generations (Bottani et al ., ). In all cases, the TE‐dense subgenome was less expressed than the TE‐sparse subgenome.…”

Cis–trans controls and regulatory novelty accompanying allopolyploidization

“…One possible insight is offered by Bottani et al . (), who demonstrated how regulatory variation in TF binding and chromatin state can propagate to the level of differential expression between homoeologs. At the single‐gene level, when homoeologs are regulated by a common set of TFs, parental differences in binding site affinity to TFs (modeled by dissociation constant K of the Hill equation), rather than in TF expression levels, were shown to be a key driver of differential transcriptional response.…”

“…Bottani et al . () presented a two‐step model to interrogate the causal mechanisms of expression bias (Fig. d).…”

“…Thus, following genomic merger and doubling, genes that harbor binding sites with high affinity become preferentially expressed, hence becoming the ‘dominant’ subgenome in allopolyploids. Despite the support offered by this mathematical model and simulation (Bottani et al ., ), the underlying biology remains largely unexplored. Do homoeologous chromatin states mainly reflect the cis divergence of parental euchromatic contents, or are trans effects on chromatin important?…”

“…If the parental conditions in TE adjacency and epigenetic accessibility () are predominantly inherited to explain the extent and direction of homoeolog expression bias, the TE model (Steige & Slotte, ; Bird et al ., ; Cheng et al ., ; Wendel et al ., ) is supported. Not exclusively, if the parental divergence in TF affinity is inherited () and correlates with homoeolog expression levels, the euchromatin/TF model (Bottani et al ., ) is supported. (e) Aggregated expression patterns are categorized by the contrasting total homoeolog expression to parental and mid‐parental expression levels, which can be interpreted as cis – trans regulatory interactions.…”