On the Decoupling of Evolutionary Changes in mRNA and Protein Levels

Abstract

Variation in gene expression across lineages is thought to explain much of the observed phenotypic variation and adaptation. The protein is closer to the target of natural selection but gene expression is typically measured as the amount of mRNA. The broad assumption that mRNA levels are good proxies for protein levels has been undermined by a number of studies reporting moderate or weak correlations between the two measures across species. One biological explanation for this discrepancy is that there has been compensatory evolution between the mRNA level and regulation of translation. However, we do not understand the evolutionary conditions necessary for this to occur nor the expected strength of the correlation between mRNA and protein levels. Here, we develop a theoretical model for the coevolution of mRNA and protein levels and investigate the dynamics of the model over time. We find that compensatory evolution is widespread when there is stabilizing selection on the protein level; this observation held true across a variety of regulatory pathways. When the protein level is under directional selection, the mRNA level of a gene and the translation rate of the same gene were negatively correlated across lineages but positively correlated across genes. These findings help explain results from comparative studies of gene expression and potentially enable researchers to disentangle biological and statistical hypotheses for the mismatch between transcriptomic and proteomic data.

Keywords: evolutionary theory; gene expression; mRNA–protein correlation; phylogenetic comparative analysis.

Fig. 1.

Coevolution of the mRNA level, the rate of translation, and the protein level when the protein level is under stabilizing selection. (A) End-point transcription–translation correlation across lineages. (B) End-point mRNA–protein correlation. In (A) and (B), each data point represents a replicate lineage (i.e., species), and the position of the point represents the lineage’s phenotype at the end of the simulation. Lines in (A) and (B) are least-squares regression lines. (C) Variances of the mRNA level, the rate of translation, and the protein level across lineages through time. (D,E) End-point transcription–translation correlation (D) and mRNA–protein correlation (E) under different combinations of ${\mathcal{N}}_e$ and shape of the fitness function. y-Axis of (D) and (E) is reciprocal of the fitness function’s standard deviation (SD). A greater reciprocal means a smaller SD, a narrower fitness function, and stronger selection. All phenotypes plotted are in log scale.

Fig. 2.

(A) Phylogenetic tree used for the simulation. The root edge is only shown to indicate the root’s location. (B) Pairwise phenotypic divergence plotted against pairwise divergence time when the protein level is under stabilizing selection. Each data point represents a combination of species pair and trait. The y-axis value of each point is the absolute phenotypic average difference between the two species (i.e., $|\ln R_i|$, $|\ln {\beta }_i|$, and $|\ln P_i|$ for species i and j), averaged across 500 simulations. Each curve is a locally estimated scatterplot smoothing curve for the corresponding trait.

Fig. 3.

(A) Schematic illustration for the model of between-gene interaction considered in this study. (B–G) Transcription–translation correlation (B, D, and F) and mRNA–protein correlation (C, E, and G) of interacting genes. Axes are genes’ regulatory effects on each other. (B,C) A gene directly subject to stabilizing selection (i.e., has an optimal protein level). (D,E) A regulator gene that is not directly subject to selection. Transcription of the target gene in (B) and (C) is regulated by the protein of the regulator in (D) and (E). (F,G) Correlations observed for one gene (gene 1) when both genes are directly subject to stabilizing selection.

Fig. 4.

Coevolution of the mRNA level, the rate of translation, and the protein level when the protein level is under directional selection. (A) End-point transcription–translation correlation. (B) End-point mRNA–protein correlation. (C) A schematic illustration of the difference between across-species and across-gene correlations, using mRNA–protein correlation as an example. Correlation across species is calculated from mRNA and protein levels of the same gene in different species, whereas correlation across genes is calculated from mRNA and protein levels of different genes in the same species. (D) End-point transcription–translation correlation among multiple genes with different optimal protein levels. (E) End-point mRNA–protein correlation among multiple genes with different optimal protein levels. In (D) and (E), each gene is represented by a cloud of points of a distinct color. Solid lines of different colors are least-squares regression lines of different genes, while the dashed lines are least-squares regression lines based on all data points. Correlation coefficient shown in each panel is based on all data points in the panel.