PDF《A biophysical protein folding》

A biophysical protein folding model accounts for most mutational fitness effects in viruses C.

Scott Wylie and Eugene I. Shakhnovich Department of Chemistry and Chemical Biology, Harvard University.

12 Oxford Street, Cambridge, MA 02138, USA classification: Physical sciences: physics, Biological sciences: evolution corresponding author: [email protected] Abstract Fitness effects of mutations fall on a continuum ranging from lethal to deleterious to beneficial. The distribution of fitness effects (DFE) among random mutations is an essential component of every evolutionary model and a mathematical portrait of robustness. Recent experiments on five viral species all revealed a characteristic bimodal shaped DFE, featuring peaks at neutrality and lethality. However, the phenotypic causes underlying observed fitness effects are still unknown, and presumably thought to vary unpredictably from one mutation to another. By combining population genetics simulations with a simple biophysical protein folding model, we show that protein thermodynamic stability accounts for a large fraction of observed mutational effects. We assume that moderately destabilizing mutations inflict a fitness penalty proportional to the reduction in folded protein, which depends continuously on folding free energy (?G). Most mutations in our model affect fitness by altering ?G, while, based on simple estimates, ?10% abolish activity and are unconditionally lethal. Mutations pushing ?G>

0 are also considered lethal. Contrary to neutral network theory, we find that, in mutation/selection/drift steady-state, high mutation rates (m) lead to less stable proteins and a more dispersed DFE, i.e. less mutational robustness. Small population size (N) also decreases stability and robustness. In our model, a continuum of non-lethal mutations reduces fitness by ?2% on average, while ?10-35% of mutations are lethal, depending on N and m. Compensatory mutations are common in small populations with high mutation rates. More broadly, we conclude that interplay between biophysical and population genetic forces shapes the DFE. Introduction What fraction of new mutations is deleterious to organismal fitness? Are most deleterious mutations mild or are they nearly lethal? The answers to these fundamental questions are provided by the distribution of fitness effects (DFE). The DFE quantifies robustness of genomes to random mutations: deleterious mutations have small effects in robust genomes while having large or lethal effects in brittle genomes. The DFE also shapes the pattern and extent of genetic diversity segregating within populations. This diversity, in turn, is crucial to interpreting molecular polymorphism data (1), the evolutionary function of sex/recombination (2), and genomic decay due to DMuller'

s ratchet‖ (3). Finally, the DFE also constrains patterns of nucleotide substitutions between species, e.g. the Dmolecular clock‖ (4). Properties of the DFE have long been estimated by two indirect methods. First, mutation accumulation experiments pass populations through deep bottlenecks, which relaxes selection and causes (mostly) deleterious mutations to accumulate, depressing the population'

s mean fitness (4). The rate and strength of typical mutations can be estimated from the tempo and variability of fitness decline. A second method compares the rate of nucleotide substitutions across species at sites of interest to that of putatively neutral sites (e.g. ref. (5)). Importantly, neither of these methods can detect lethal mutations because they are instantly purged from populations. For a review of these methods and the DFE generally, see ref. (6). Recently, a more direct method utilizing site-directed mutagenesis was applied to viruses (7-11). These studies measured mutant fitness paired with the exact underlying genomic change among an unbiased set of single nucleotide substitutions, finding similarly shaped DFE across five viral species. Most missense mutations probably impact organismal fitness by altering protein activity and/or stability. Predicting which rare mutations dramatically improve protein activity or create new functions remains a formidable challenge that we do not address here. However, estimating the distribution of mutational effects that merely perturb evolved, more-or-less optimized proteins is a more tenable goal. The role of most residues is to maintain a protein'