Saturday, October 01, 2005

Pax visual   posted by Razib @ 10/01/2005 12:36:00 AM

I have mentioned the Pax6 gene before in relation to myopia, but recently I stumbled onto this review paper/model proposal titled Pax genes in eye development and evolution (you can find the full PDF in the forum files as "kozmikpax"). Many of you are probably aware of the shocking level of conservation of hox genes across animal taxa, but this sort of evolutionary developmental similarity on the genetic level is probably rather common (to the surprise of biologists when this was first discovered a few years back). Though creationists like to point to the eye as an example of an organ which bears all the hallmarks of design, it is an interesting observation that the structure of this particular sensory apparatus varies a great deal from species to species (compound insect eyes vs. the vertebrate eye vs. the cephalopod eye). Some groups of animals, like cave fish, also tend to lose their visual capacity when selection for the character is relaxed.

The Pax family of genes, of which Pax6 is a prominent member, produce transcriptional factors1 which are implicated in a host of regulatory functions as relates to the development of eyes across many animal taxa. The author points out that:
  • The Pax family is found in animals which have no visual apparatus or neural network.
  • The most recent common ancestor of all the Pax genes seems to predate the geological age when vision became a ubiquitous feature of animals.
  • In mammals particular Pax genes that are involved in eye development in invertebrates are to be found in the glial tissue. This is, different family members have different roles across taxa.
  • Though Pax genes seem a necessary condition for normal eye development, they are not a sufficient condition (there are other genes which are involved in eye development).

The phylogenetic and comparative points of evidence suggests to the author that the Pax family originated prior to the evolution of vision, and that it was coopted multiple times over the course of natural history. The implication here is that convergent evolution not only produced "eyes" in a variety of taxa, but the same gene family was recruited for this developmental pathway each time! It is important to note though that the review points out that in vertebrates more functional diversity is squeezed out of these genes via alternative splicing while in invertebrates duplication is the norm.2

Ultimately the author suggests that the reason that Pax has been recruited is rather prosaic. He offers that the minimal characteristics of a visual sensory apparatus include a photoreceptor an dark pigment, no matter what variation is the norm on other structural features. The Pax family tends to have two binding regions3 which can be regulated in a coupled fashion, so the gene can map onto the appropriate functional pathways. The reason that Pax, and not some other gene with two binding regions, is preferentially recruited for its role in the development of the eye is unclear, though one can spin stories of pre-adaptations unconstrained by pleiotropic effects with negative fitness consequence pretty easily.

A more parsimonious explanation is that the Pax family has always played a role in eye development, and in cases where it isn't implicated in eye development it is a loss of function. Though the points that argue against this are highlighted in the paper, I can't quite give up on this avenue because of its simplicity.

There are many philosophical issues wrapped up in the resolution of questions rooted in the "problems" that evolutionarily developmentally salient loci like the Pax family pose. Interesting times....

Addendum: If eyes were each special creations, god seems an old dog that can't learn new tricks.

1 - Transcription is the mapping of DNA to RNA, which is eventually translated into proteins. Transcriptional factors turn up or down the cranks on genes which are normally spinning off RNA into the aether.

2 - Gene duplication is pretty straightforward (relatively), but splicing is weirder.

3 - Binding regions are parts of the DNA helix where transcriptional factors lodge themselves and either turn up or turn down transcription. Turning off is pretty common sense, since a foreign object just shows up on DNA's door. But sometimes the factor has all sorts of weird extensions which might aid in the initiation of transcription.