Evolutionary biochemistry: revealing the historical and physical causes of protein properties

Abstract

The repertoire of proteins and nucleic acids in the living world is determined by evolution; their properties are determined by the laws of physics and chemistry. Explanations of these two kinds of causality - the purviews of evolutionary biology and biochemistry, respectively - are typically pursued in isolation, but many fundamental questions fall squarely at the interface of fields. Here we articulate the paradigm of evolutionary biochemistry, which aims to dissect the physical mechanisms and evolutionary processes by which biological molecules diversified and to reveal how their physical architecture facilitates and constrains their evolution. We show how an integration of evolution with biochemistry moves us towards a more complete understanding of why biological molecules have the properties that they do.

Figure 1. Parallel evolution due to biophysical constraints

a | Distribution of mutations observed during the evolution of the visual pigment opsin in fish. Opsin absorbs light at specific wavelengths through its bound retinal (yellow). Mutations can alter its absorption properties by changing the environment of the chromophore. Spheres highlight residues that changed as fish adapted to different light environments. Large-effect mutations acquired in parallel on multiple lineages (green) border the retinal; small-effect, lineage-specific mutations (orange) are more distant. b-d | Mechanism for the parallel evolution of herbicide resistance via the same mutation in 68 different species of weeds. b | Crystal structure of one half of the symmetrical multi-subunit complex of photosystem II (PSII). The site of the Ser264Gly mutation, which confers resistance, is shown in green, and the endogenous cofactor plastoquinone is yellow. c | Cofactor plastoquinone forms hydrogen bonds to Ser264 and to His215. d | The herbicide terbutryn (shown in yellow) directly hydrogen bonds only to Ser264. The Ser264Gly mutation abolishes all hydrogen bonds from the side chain of this residue, radically reducing terbutryn binding while only partially compromising plastoquinone binding. No other mutations are known that affect herbicide resistance without having a concomitantly large effect on plastoquinone binding.

Figure 2. Molecular mechanisms of evolutionary epistasis

A | Epistasis mediated by effects on global stability. Aa | The schematic shows the effects on the evolution of a new function (blue) of two interacting mutations (green and purple) with different effects on stability and function. Proteins with stabilities below a given threshold are unstructured and non-functional. Ab | Epistasis in the evolution of bacterial resistance to the antibiotic ceftazidime is mediated by effects on global stability. Analysis of the major-effect mutations shows that stability modulates resistance. Each platform is an allele; its location along the y axis shows its melting temperature of unfolding (T_m). Bar graphs show the enzymatic activity of each allele (K_cat/K_M) relative to the ancestral protein (yellow bar) and antibiotic resistance (blue bar, measured as inverse halo diameter). The E104K and G238S mutations (purple spheres in the structure) confer high enzymatic activity but low resistance because the protein is unstable. The distant mutation M182T (green sphere) confers high stability by addition of a new hydrogen bond but does not change activity. Their combination yields resistance. The antibiotic is shown as yellow spheres. B | Specific epistasis mediated by a direct interaction. Ba | A schematic showing a direct, physical conformational change (blue). Bb | An example of direct epistasis from engineered Streptococcus spp. protein G domains that differ at two residues but have radically different folds. These residues form a packing interaction only when both are aromatic residues, driving the transition between folds. C | Specific epistasis indirectly mediated by a conformational change. Ca | The schematic shows how two mutations that do not physically interact can genetically interact in the evolution of a new function. One mutation creates the potential for a new interaction (green), which is realised only if the first residue is repositioned by a conformational change triggered by the other mutation (purple). Cb | An example of conformational epistasis from the evolution of ligand sensitivity in the vertebrate glucocorticoid receptor. Crystal structure of the ancestral (orange) and derived (blue) forms of the glucocorticoid receptor. Novel specificity for glucocorticoid ligand (yellow) evolved because of the interaction of historical substitutions L111Q (green), which introduces a hydrogen bond acceptor, and S106P (purple), which repositioned the helix on which the L111Q is located (arrows), allowing L111Q to form a novel hydrogen bond with the ligand.

Figure 3. The position of a protein in its neutral network determines which mutational path it takes to a derived function

Protein sequences (ovals) are connected by point mutations (arrows). The colours represent functions: ancestral (orange) or derived (blue). Transitional colours represent transitional functions. Nearby sequences in the ancestral neutral network follow the same ‘deterministic’ pathways (dark arrows) when selection for the derived function is applied. Some sequences in the neutral network cannot achieve the derived function without first taking a permissive functionally neutral step through the network (ovals with dashed outlines).