. "Evolution of codon usage bias in Drosophila." (NAS Colloquium) Genetics and the Origin of Species: From Darwin to Molecular Biology 60 Years After Dobzhansky. Washington, DC: The National Academies Press, 1997.
The following HTML text is provided to enhance online
readability. Many aspects of typography translate only awkwardly to HTML.
Please use the page image
as the authoritative form to ensure accuracy.
Proceedings of the National Academy of Sciences of the United States of America
the Ks being on average about an order of magnitude lower than for the other two pairs. Also, it is possible sharing of polymorphic alleles in closely related species may also be obscuring the picture.
Does a similar phenomenon occur for intraspecific polymorphisms—i.e., do more highly biased genes have less synonymous polymorphism within a species? We observed the opposite for 21 genes in D. melanogaster for which data on intraspecific variation were available (29): there is a statistically significant positive correlation between codon usage bias and level of synonymous polymorphism in a gene. How can this be explained? We speculate this may be due to the effect of variation in recombination. Fig. 6 outlines the argument. Genes vary in their levels of recombination dependent upon position in the genome. Increased recombination can have two effects, one of which is to increase codon usage bias (19) and the other is to increase synonymous polymorphisms (30). Both are due to a decrease in hitchhiking effects of linked genes. As mentioned previously, selection at single nucleotide positions is more effective in regions of high recombination, thus allowing for an increase in selection for optimal codons. The effect of recombination on levels of synonymous polymorphism is thought to be due to selective “sweeps” at linked positions; such sweeps take along with them linked sites which then become less variable. Such selection may be positive when a new linked favorable mutation arises and goes to fixation (31), or may be due to negative “background selection” against deleterious mutations (32); it is not clear which of these processes best fits the data, but their effects are similar. From the observation in D. melanogaster that highly biased genes tend to have higher synonymous polymorphism, the arrows at the bottom of Fig. 6 would seem not to have equal strength in their effects. The expected decrease in synonymous polymorphism caused by codon usage bias is not great enough to overcome the expected increase in such polymorphisms due to lessening effects of selective sweeps.
While the information available on codon usage bias of both microorganisms and Drosophila provides good evidence that selection can act on what had been considered prime candidates for neutral mutations, are all synonymous substitutions detected at all times? This is highly unlikely, and we argue elsewhere that there is likely a continuum in Drosophila (and other organisms) with codon usage in highly biased genes being primarily affected by selection, whereas other genes may have codon usage controlled primarily by mutation and drift along the lines of models previously proposed (33, 34). This is in agreement with Akashi’s (35) observation that the selection for optimal codons in D. simulans has been more effective than in D. melanogaster. D. simulans is thought to have an effective population size greater than D. melanogaster, so the selection coefficients on synonymous mutations (at least on some genes) are sufficiently small as to be sensitive to population size differences among species of Drosophila. This implies the selection coefficients on synonymous mutations are on the order of Nes equal to 1 (Ne is the effective population size and s is the selection coefficient), consistent with previous studies (35, 36). Further, we note that when selection is ineffective due to reduced recombination, codon usage may well reflect mutation/drift (Table 2). Nevertheless, at times selection for codon usage must be very effective, as exemplified by the phylogenetic persistence of avoidance of specific codons in specific genes for very long evolutionary periods (Table 1).
We thank the organizers of this colloquium, Francisco J.Ayala and Walter M.Fitch, for the opportunity to commemorate Theodosius Dobzhansky and the publication of arguably the most important book on evolution in the 20th Century. We appreciate the helpful review provided by Charles F.Aquadro. This work was supported by National Science Foundation Grant DEB9318836.