Boross Gábor, Papp Balázs
Synthetic and Systems Biology Unit, Institute of Biochemistry, Biological Research Centre, Szeged, Hungary.
Mol Biol Evol. 2017 Feb 1;34(2):380-390. doi: 10.1093/molbev/msw236.
Changes in gene expression can affect phenotypes and therefore both its level and stochastic variability are frequently under selection. It has recently been proposed that epistatic interactions influence gene expression evolution: gene pairs where simultaneous knockout is more deleterious than expected should evolve reduced expression noise to avoid concurrent low expression of both proteins. In apparent support, yeast genes with many epistatic partners have low expression variation both among isogenic individuals and between species. However, the specific predictions and basic assumptions of this verbal model remain untested. Using bioinformatics analysis, we first demonstrate that the model's predictions are unsupported by available large-scale data. Based on quantitative biochemical modeling, we then show that epistasis between expression reductions (epigenetic epistasis) is not expected to aggravate the fitness cost of stochastic expression, which is in sharp contrast to the verbal argument. This nonintuitive result can be readily explained by the typical diminishing return of fitness on gene activity and by the fact that expression noise not only decreases but also increases the abundance of proteins. Overall, we conclude that stochastic variation in epistatic partners is unlikely to drive noise minimization or constrain gene expression divergence on a genomic scale.
基因表达的变化会影响表型，因此其水平和随机变异性常常处于选择之下。最近有人提出，上位性相互作用会影响基因表达进化：同时敲除比预期更具有害性的基因对，应该进化出降低的表达噪声，以避免两种蛋白质同时出现低表达。表面上的证据是，具有许多上位性伙伴的酵母基因在同基因个体之间以及物种之间都有较低的表达变异。然而，这个文字模型的具体预测和基本假设仍未得到验证。通过生物信息学分析，我们首先证明该模型的预测没有得到现有大规模数据的支持。基于定量生化建模，我们接着表明，表达降低之间的上位性（表观遗传上位性）预计不会加剧随机表达的适应性成本，这与文字观点形成鲜明对比。这个非直观的结果可以很容易地通过适应性对基因活性的典型递减回报以及表达噪声不仅会降低而且会增加蛋白质丰度这一事实来解释。总体而言，我们得出结论，上位性伙伴中的随机变异不太可能在基因组尺度上推动噪声最小化或限制基因表达差异。
