Protein misinteraction avoidance causes highly expressed proteins to evolve slowly

Abstract

The tempo and mode of protein evolution have been central questions in biology. Genomic data have shown a strong influence of the expression level of a protein on its rate of sequence evolution (E-R anticorrelation), which is currently explained by the protein misfolding avoidance hypothesis. Here, we show that this hypothesis does not fully explain the E-R anticorrelation, especially for protein surface residues. We propose that natural selection against protein–protein misinteraction, which wastes functional molecules and is potentially toxic, constrains the evolution of surface residues. Because highly expressed proteins are under stronger pressures to avoid misinteraction, surface residues are expected to show an E-R anticorrelation. Our molecular-level evolutionary simulation and yeast genomic analysis confirm multiple predictions of the hypothesis. These findings show a pluralistic origin of the E-R anticorrelation and reveal the role of protein misinteraction, an inherent property of complex cellular systems, in constraining protein evolution.

Biophysics of protein evolution and evolutionary protein biophysics

“…Introduction of various protein -protein interactions to this modelling set-up by Shakhnovich and co-workers [345] has provided further rationalization for the emergence of species-like collections of model cells with very similar sequence make-ups, an increased rate of mutation in stress response [346], as well as a trade-off between strengthening functional interactions and avoidance of misinteractions as observed in experimental proteomic data [347]. More recently, Zhang and co-workers [99] developed a related lattice approach to model evolution of protein -protein interactions that offers an explanation of slow evolution of highly expressed proteins in terms of stronger constraints on these proteins to avoid misinteractions.…”

“…Another probable biophysical constraint behind the anticorrelation between protein expression level and evolutionary rate is the need for a protein to avoid misinteraction with other proteins [99]. This selection pressure affects primarily surface residues that can potentially participate in interactions between the protein and other molecules.…”

“…Mutations affect these interactions and can lead to misinteractions [99]. A classic example is the glutamic acid to valine mutation in haemoglobin [100] that causes aggregation of haemoglobin and consequently sickle-cell anaemia [101].…”

“…The expanded concept of molecular phenotype discussed in §2.9 is an attempt towards a better account of this biophysical reality. In this regard, highly simplified yet promising explicit-chain biophysical models have recently been developed to study protein evolution in the context of protein-protein interactions [99,347] ( §3.2.3) We have also summarized explicit-chain biophysical models that take into consideration the ensemble nature of a protein's conformational phenotype, and how these models can provide insights into selection of promiscuous function, bi-stability and structural divergence ( §3.2).…”

“…It would be daunting to account for these interactions in all their complexity at the molecular level. Using simple explicit-chain protein models, conceptual advances were made in elucidating how the biophysical constraint of misinteraction avoidance might impact protein evolution [99,347] ( §3.10). However, investigators have to rely upon abstract descriptions of protein interactions, using model parameters extracted from experimental data on binding and on the effects of enzymatic activities on biochemical reaction rates.…”