Protein Evolution |
||
|
The reasons why rates of evolution vary across sites in proteins is not well understood. Assuming each site evolves at a rate that is independent of all other sites, the estimated proportion of the observed site-to-site variability that needs to be explained, p2, is |
|
| . | ||
 |
||
| . | ||
where  is the mean
number of replacements per site and s2 is the variance. The remainder,  /s2, estimates the proportion of
variation attributable to Poisson error. |
||
| . | ||
| We investigated patterns of amino acid replacement with respect to protein structure. Rates of evolution are correlated with (i) distance from the active site (substrate), (ii) solvent accessibility, and (iii) treating glycines in unusual conformations as a separate class. This model explains half the causal variation, p2, with only 3 df. Secondary structure exerts little influence on the pattern and distribution of replacements. | ||
| . | ||
| Dean, Neuhauser, Grenier & Golding. 2002. MBE 19:1846-1864. | ||
|
There is no a priori reason why each site in a protein should evolve independently of others. The concomitantly variable codon (covarion) hypothesis proposes that only a few sites in a protein are free to vary at any one time, and that as changes accumulate so the set of sites changes. To test the covarion hypothesis we constructed chimeras of triosephosphate isomerase from E. coli and P. aeruginosa. Although their peptides differ by 50%, covarions of lethal effect are extremely rare. Only a handfull of lethal concomitantly varying sites were uncovered (colored residues) with most at the homodimer interface, far from the active site. By contrast, sensitive kinetic assays revealed many chimeras with impaired performance, suggesting covarions of small effect are relatively common. |
|
|
||
 |
||
| . | ||
| Merlo & Dean, coming soon. | ||
