Sean Carroll (SC) recently gave a talk at my institution (Penn State University) about the role of cis regulatory regions in the evolution of body form.
SC made a bold conjecture, viz. that all or most evolution of animal form, as distinct from physiology, is achieved through changes in cis regulatory elements (CREs). This is because, in any given evolutionary trajectory, changes in protein coding sequences will likely have pleiotropic consequences and gene duplication would result in dosage problems, while adjustments to modular CREs (which are specific to different tissue types) would bypass pleiotropy and therefore be likely to have higher relative fitnesses than these other changes.
He illustrated this with examples from fruitfly genetics from his lab. He described an island Drosophila species that had lost posterior segment pigmentation due to mutation of a CRE that, in D. melanogaster, responds to the abdominal-B (abd-B) transcription factor (TF) in a pigment gene called yellow. A reporter construct was used to confirm the CRE in melanogaster and its absence in the island species and to confirm the minimal mutation. The yellow coding sequence (CDS) experienced no change at all between pigmented and non-pigmented sister species on the island. See this for more detail on this example.
In a gain-of-function example, he showed how yellow was specifically activated in a wing spot in a relative of melanogaster. He showed a video illustrating how this spot was used by males in courtship. yellow activation in a specific region was acquired in a two-step process in which a CRE of yellow evolved binding sites for an activator: a TF already present in the distal wing tip, and also for a repressor, present in the wing’s dorsal half: engrailed. The joint action of these limited the pigmentation to the distal anterior wing portion. This was determined by deleting portions of the CRE singly and in combination in vivo. See this for more.
SC said that only CRE mutations were found to be significant in all cases of evolution of animal form so far investigated and that such important regulators of form as Hox genes (note: abd-B is a Hox gene) have actually only lost members during hundreds of millions of years of morphological evolution (his periods for consideration were the entire evolution of the arthropods and the emergence of tetrapods from arthropods) – so gene duplication is unlikely to be of any significance – it’s the downstream targets that matter when you are trying to make big changes in bauplan (I am wondering about pleiotropy in the Cambrian now). He emphasised that his conclusions were specific to animals and that he was not talking about physiology.
Questions were revealing of the following points:
1. yellow has functions in the brain and it is therefore difficult to draw a clear boundary between what is related to form and what to physiology.
2. yellow acts in a pathway and the other elements in this pathway need to be present for pigment formation therefore there are more steps than stated above in the evolution of the wing spots. This seems to undermine the point about evolution a bit because it ramps up the improbability a bit – though it arguably does so equally for the alternatives too (such as CDS evolution).
3. yellow is involved in male sexual determination in the brain and therefore there is missing information about the whole adaptive complex in a species that uses the wing spots as a lure. We need a joined-up explanation.
4. is modularity of CREs a cause or consequence of this sort of evolution? Binding sites are easy to add and take away.
5. plants are more tolerant of dose effects because of the frequency of polyploidy (or is this also chicken and egg?) – so gene duplication seems more likely here
I would like to add that I am a bit confused about modularity as, after the sort of evolution he describes, this would be predicated upon the tissue-specificity of TFs – something that we are not explaining. The eminence grise in terms of general explanatory efforts is constraint it would seem.