Cell Reports W o

requirement for highly expressed proteins to evolve slowly. Lastly, we predict from multiscale evolutionary simulations thatproteins. The primary hypothesis is that misfolded proteins are detrimental to cellular fitness (Drummond et al., 2005). Conse-quently, more abundant proteins, being prone to produce more constructed a monoclonal population of 104 model cells (Fig-ure 1). Each cell consisted of 103 genes that were characterized by their folding stability (DG =Gfolded –Gunfolded) and abundance.toxic molecules (Bucciantini et al., 2002), experience stronger selection pressure, hence will evolve slowly (Drummond et al., The abundances of the 103 genes were randomly drawn from the distribution of measured protein abundances in yeastUnderstanding the variation of protein structures and sequences in nature is a fundamental problem in biology. Crucial to address-ing this question is knowing what determines the rate of protein evolution. A major advance came from the observation that highly expressed proteins consistently evolve slowly in bacteria, yeast, worm, fly, mouse, and humans (Drummond et al., 2005; Drummond andWilke, 2008; Pál et al., 2001). This anticorrelation between a protein’s cellular abundance and its evolutionary rate (denoted herein as ER) found a unified possible explanation in the global selection against the cytotoxic effect due to misfolded the strength of the abundance-ER correlation will increase with divergence time. Systematic analysis of species in a class of bacteria confirms this prediction