Give the lady her due, Olivia Judson can lay down some serious exposition when she's on:
There are a couple of interesting things about this discovery. The first is that the molecular basis of the change from pelvis to no pelvis does not involve a mutation to the protein-coding region of the Pitx1 gene itself. In other words, the protein made from the gene hasn't changed. What has changed is the way the gene is expressed. This is in contrast to the sorts of mutations one often reads about as being involved in evolution, which typically involve changes to the protein itself.
A second interesting feature of the stickleback pelvis is that -- unlike the armor plates -- the loss is probably due to mutations having occurred independently in the different populations. What's more, changes to the use of Pitx1 are also implicated in pelvic loss in nine-spine sticklebacks (Pungitius pungitius) -- yet nine-spine and three-spine sticklebacks have been going their own evolutionary ways for at least 10 million years. Mice that have been genetically engineered to lack Pitx1 have a suite of abnormalities, including crushed faces and abnormal pituitaries, that cause them to die young. Intriguingly, they also have a reduced pelvis and hind limbs, and as with the sticklebacks, the reduction is lopsided and shows a greater loss on the right than on the left.
Recall my post about HERC2, a mutation on an intron on this gene might have resulted in a change in regulation of OCA2, which induces depigmentation. Speaking of which, human skin color seems to have gotten light independently at least three times, in Europe, East Asia and among the Neandertals (that we know of). Remember, loss of function is a lot easier than gain of function. Dark-skinned peoples tend to exhibit a consensus sequence which is relatively similar no matter their phylogenetic relationship (e.g., the residents of the Solomon Islands are closer to East Asians than Africans in terms of ancestry, but on the loci which control skin color they are nearly identical by state with Africans). Why model organisms? Because we have faith that Nature is One.
- Log in to post comments
That is not why we use model organisms and you know it.
On a serious note, do you know if protein-altering mutations are more or less constrained than expression-altering ones?
Caledonian,
My guess would be that expression is less constrained than protein structure, since any given protein is usually involved in all kinds of pathways, while in principle there could be multiple genes governing expression of the same protein for different pathways. It makes a priori biological sense to design organisms this way, but I don't know of any data that would answer this question (yet).