Abstract
The avoidance of cytotoxic effects associated with protein misfolding has been proposed as a dominant constraint for the evolution of highly expressed proteins. Recently, Leuenberger et al. developed an elegant experimental approach to measure protein thermal stability at the proteome scale. The collected data allow to rigorously test key predictions of the misfolding avoidance hypothesis. Specifically, that highly expressed proteins are designed to be more stable, and that thermodynamic stability significantly constrains their evolution. Careful re-analyses of the Leuenberger et al. data across four different organisms shows no substantial correlation between protein stability and protein abundance. We also find that protein stability does not substantially contribute to sequence constraints of highly abundant proteins. Therefore, the key prediction of the misfolding toxicity avoidance hypothesis is not supported by the empirical data.