The relationship between misfolding avoidance hypothesis and protein evolutionary rates in the light of empirical evidence

For more than a decade the misfolding avoidance hypothesis (MAH) and related theories have dominated evolutionary discussions aimed at explaining the variance of molecular clock across cellular proteins. In this study we use various experimental data to further investigate the consistency of the MAH predictions with empirical evidence. We also critically discuss experimental results that motivated the MAH development and that are often viewed as evidence of its major contribution to constraining protein evolution. We demonstrate, in Escherichia coli and Homo sapiens , the lack of a substantial negative correlation between protein evolutionary rates and Gibbs free energies of unfolding, a direct measure of protein stability. We then analyze multiple new genome-scale datasets describing protein aggregation and interaction propensities, which are likely optimized in evolution to alleviate deleterious effects associated with toxic protein misfolding and misinteractions. Our results demonstrate that the propensity of proteins to aggregate, the fraction of charged amino acids, and protein stickiness do correlate with protein abundances. Nevertheless, across multiple organisms and datasets we do not observe substantial correlations between proteins aggregation- and stability-related properties and evolutionary rates. Therefore, diverse empirical data support the conclusion that the MAH and similar hypotheses are unlikely to play a major role in mediating a strong negative correlation between protein expression and molecular clock, and thus in explaining the variability of evolutionary rates across cellular proteins. Significance statement Evolutionary rates vary substantially across cellular proteins. Understanding the nature of molecular clock and its variability across proteins is a foundational question in molecular evolution. The popular and currently dominant theory to explain the molecular clock variability is the misfolding avoidance hypothesis (MAH). The role of the MAH is currently under active debate. In the manuscript we discuss how to appropriately test the MAH based on available empirical data, and then rigorously test the hypothesis using more than a dozen of new genome-wide datasets of protein stability and aggregation propensities. Our results suggest that the MAH is unlikely to play a major role in explaining the variability of molecular clock across proteins.