Reducible complexity

The main difference between creationism/intelligent design and real science is that the proponents of the former are fond to proclaim the impossibility of something (and at the same time believing in myths without a shred of evidence), while scientists are hard at work trying to figure out what is actually true by using empirical means.

A perfect example of this difference is a paper by Bridgham et al. published in the 7 April 2006 issue of Science. They studied a biological system that fits ID supporters’ definition of “irreducible complexity,” and which – accordingly – couldn’t possibly have evolved by natural means. The system in question is a classic example of “lock-and-key” at the biochemical level: an hormone receptor that binds specifically to the hormone cortisol. The question is: how could such a specific system evolve, considering that one had to have both parts (the receptor and the hormone) in place at the same time?

The elegant answer – which I predict right now will not convince any ID supporter, because their position is ideological and has nothing to do with evidence – came out of careful studies involving the reconstruction of the ancestral receptor molecule and the analysis of the kind of mutations that can change its shape.

As Darwin first pointed out, if one wishes to understand the evolution of any biological structure, one has to look as much as possible to its history, not just the current form. The latter may look impossible to explain simply because we have limited access to the intermediate forms that preceded it. So, Bridgham and colleagues used phylogenetic information on the evolution of the relevant genes to reconstruct the potential ancestor of the current receptor, and then study its biochemical characteristics in the laboratory (this is not circular reasoning: one begins by assuming evolution and making predictions: if the predictions are independently verified, this is a confirmation of the evolutionary hypothesis – standard scientific practice, no flim-flamming).

The first thing they found out about the ancestral receptor was rather surprising: not only it bound cortisol, albeit less efficiently than the current version of the molecule, but it also bound aldosterone (another hormone) and DOC (11-deoxycorticosterone). In other words, the ancestral receptor molecule had a generic biochemical ability, which is a key prediction of evolutionary theory (biological functions start out generic and become more complex and specific). How did the specific affinity for cortisol evolve?

It turns out that one needs two mutations, called L111Q and S106P, to modify the original generic receptor to a specific one. But wait! Isn’t that precisely what’s difficult to explain from a Darwinian perspective? How could two advantageous mutations happen simultaneously? And if one happened after the other, how come that the intermediate molecule persisted while waiting for the second mutation, since presumably a half-step is deleterious? (This is the molecular version of the old creationist question: what good is half an eye? The answer, by the way, is: “about half as good as a full eye”)

Bridgham et al. were able to show that in fact it makes a huge difference which of the two mutations happened first: if the L111Q appears first, then the receptor binds to nothing at all, and it becomes molecular junk, an evolutionary dead end. But if the S106P mutation happens first to the original molecule, the result is a receptor that binds only DOC and not the other two hormones. When the second mutation comes, the receptor now binds both DOC and cortisol, but not aldosterone. So one has a complete evolutionary sequence of molecular evolution, in which all the intermediates are functional, and more importantly that shows in action a basic principle of evolution: natural selection builds on previously existing materials, and in the process sometimes changes the function of those materials. No irreducible complexity needed, thank you very much.