A non-scientist colleague asked me to write a brief, non-technical summary of my last post. Here’s my best shot.
Over time, natural selection will tend to shape a protein as closely as possible to the job it performs. As an analogy, think of a screwdriver. Suppose we started out with a shapeless blob of metal that, by chance, had a little edge sticking out of it. Theoretically we might be able to use that as an inefficient screwdriver of sorts. As time went by and improvements were made (somehow), the inefficient tool could be shaped better and better to fit its job. The edge could be made straighter with a longer and longer handle, until it was the shape of an ordinary, store-bought screwdriver. But suppose we then fancied to use that perfectly good screwdriver (or an identical copy) as a wrench. The fact that selection had honed the shape of the tool to do one particular task (driving screws) makes it difficult to use it for a different task that requires a differently shaped tool. The same reasoning applies in the reverse direction. A well-shaped, store-bought wrench would be hard to use as a screwdriver.
That’s what the work of Thornton’s group shows for proteins. Turning a protein shaped to do one particular job into a protein that does just a slightly different job (which most biologists, including myself, had thought would be as easy as pie) turned out to be much more difficult than expected, even though the two proteins are thought to be related as ancestor and descendant. Turning the modern protein into one that worked like the ancient one encountered unexpected problems, and vice versa — turning the ancient one into a protein that worked like the modern one also ran into some considerable bumps in the road. Since all proteins — both ancient and modern — were/are under the whip of natural selection, the upshot is that teaching old proteins to do new tricks is harder than for an old dog. And the plausible power of the Darwinian process rapidly dwindles each time it is realistically tested.
Erratum: In my last post I dubbed the problem of a protein changing jobs the “Time-Asymmetric Dollo’s Law.” It should be “Time-Symmetric.” That is, the problems for Darwinism are the same whether one is trying to convert a modern protein to an ancient one or vice versa.
(1.) Carroll, S. M., E. A. Ortlund, and J. W. Thornton, 2011 “Mechanisms for the evolution of a derived function in the ancestral glucocorticoid receptor.” PloS. Genet. 7: e1002117.
(2.) Bridgham, J. T., E. A. Ortlund, and J. W. Thornton, 2009 “An epistatic ratchet constrains the direction of glucocorticoid receptor evolution.” Nature 461: 515-519.