Last Author

Genetic mutation is central to evolution, but mutations that improve one aspect of a protein's function can compromise another. Until recently, gene duplication was thought to lead to the development of genes with novel functions, as one copy of the gene would be free to evolve while the other performed its original function. But this, it turns out, rarely happens — more often, an original gene's functions are simply split between the two copies. Sean Carroll, at the University of Wisconsin–Madison, and his student Chris Hittinger devised a series of assays to trace how two genes — GAL1 and GAL3 — in the yeast Saccharomyces cerevisiae evolved from a single gene after duplication in a distant ancestor (see page 677). They suggest that genes that evolve in this way do so to overcome constraints present in the original gene.

Most of your work has been with fruitflies. Why do this work in yeast?

To find out how duplicate genes had changed, we needed to be able to measure very small differences in organismal fitness. We didn't think this would be possible in fruitflies — we needed the greater power of billions of offspring quickly, and yeast provided this.

Was it surprising that most of the two genes' differences were regulatory in nature?

Yes. The protein Gal1 is galactokinase — an enzyme — and Gal3 acts to regulate its transcription. The original protein would have performed both these functions. We thought the divergence of these proteins was the main story, but the data told us otherwise. Gal1 still has a lot of regulatory activity, and Gal3 only recently lost the last of its enzymatic activity.

This duplication is thought to have occurred 100 million years ago — why did it take so long for enzymatic activity to be lost?

I can think of many examples of genes deteriorating relatively quickly. But in this case, Gal3 still has to bind galactose in its regulatory role, so perhaps that constraint limited the loss of enzymatic activity.

Besides genetic duplication and divergence, might there be other evolutionary 'tools'?

There are other mutational pathways, so yes. For instance, Chris showed that one way to make a better regulatory element is to delete bases so that transcription-factor binding sites get closer together.

Why did you write The Making of the Fittest , an evolution book for the general public?

The pace of discovery in evolutionary science has really quickened. The quality of evidence and the clarity of how evolution works are now easier to talk about. And, happily, people want to know about evolution.