Figure 1.
Residue-level structure-evolution relationships.
(A) A cartoon diagram of a protein shown in cross section, highlighting three residues in different relative solvent accessibility (RSA) microenvironments. (B) Evolutionary rate (as measured by dN/dS) scales in a strong, positive, linear manner with RSA, as previously demonstrated [7].
Figure 2.
Protein core size, expression level, and evolutionary rate.
(A) An illustration of protein core size. A large-core protein (blue) contains a greater proportion of buried residues than a small-core protein (red). (B) An illustration of protein expression level. (C) We divided the 795 proteins in our study into small-core and large-core groups corresponding to the top and bottom third of proteins ranked by average relative solvent accessibility (RSA). Average RSA (core size) is a relatively weak predictor of whole protein evolutionary rate. (D) We divided proteins into high-expression and low-expression groups corresponding to the top and bottom third of proteins ranked by codon adaptation index (CAI). CAI is a strong predictor of whole protein evolutionary rate. Rankings of whole-protein evolutionary rate are from [13]; a larger rank implies a faster evolutionary rate (relaxed selective constraint).
Figure 3.
Qualitatively different effects of protein core size and expression level on residue evolution.
(A) Exposed residues from large-core proteins evolve faster than exposed residues from small-core proteins, while buried residues are similarly constrained. (B) Residues from high-expression proteins always evolve more slowly than similarly exposed residues from low-expression proteins. (C) Fold change in evolutionary rate varies significantly as a function of solvent accessibility for large-core versus small-core proteins. (D) Fold change in evolutionary rate does not vary significantly as a function of solvent accessibility for low-expression versus high-expression proteins. Error measures for slope and intercept reflect the standard error.
Figure 4.
Fold change in selection at the amino acid sequence level as a function of solvent accessibility across protein core size and expression contexts.
(A) Fold change in the rate of amino acid sequence evolution (dN) varies significantly as a function of solvent accessibility for large-core versus small-core proteins. (B) Fold change in dN does not vary significantly as a function of solvent accessibility for low-expression versus high-expression proteins.
Figure 5.
Fold change in selection at the synonymous codon level as a function of solvent accessibility across protein core size and expression contexts.
(A) Large-core versus small-core proteins experience minimal difference in the degree of synonymous codon selection that they experience. (B) Highly expressed proteins experience a uniform increase in selection at the level of synonymous codons throughout their coding sequences. (C) Large-core versus small-core proteins experience minimal difference in their use of preferred codons (as measured by codon adaptation index, CAI). (D) Highly expressed proteins experience a relatively uniform increase in their use of preferred codons.
Figure 6.
Independent effects of protein core size and expression on residue evolution.
The dN/dS versus solvent accessibility relationship for four different protein groups: (A) small-core and high-expression, (B) large-core and high-expression, (C) small-core and low-expression, and (D) large-core and low-expression. Increasing core size and decreasing expression level simultaneously in (D) results in significant increases to the slope of the trend relative to either separate change. The best-fit lines for all four groups of proteins are replicated in each panel for comparison.