This is the second of five posts in which I reply to Dr. Ian Musgrave’s “Open Letter to Dr. Michael Behe” on the Panda’s Thumb blog.
Musgrave:
But by far the worst, you ignored her core argument. That in the space of a decade HIV-1 Vpu developed a series of binding sites that made it a viroporin, a multisubunit structure with a function previously absent from HIV-1.
Behe:
It is not clear to me why you call that Smith’s “core argument.” In her post, her writing meanders quite a bit; it’s hard for me to glean what she thinks is most important. After sneering a bit at me, Smith began her post by asserting that vpu is a “new” gene (even though it is found in SIVcpz across several primate families, as her own citations show.) (1) She then spent several paragraphs making the point that Vpu’s from chimps and humans share only 37% sequence identity. That was followed by the declaration, “Turns out a LOT of evolution has been going on in HIV-1 since it was transferred to humans 50-60 years ago.” So, rather naturally I think, I concluded that she thought the simple fact of 37% sequence identity was compelling evidence for the power of Darwinian evolution. In reply, I pointed out that proteins with similar structures and functions have such sequence identity (trypsin and chymotrypsin are ~40% identical), so mere degree of sequence identity means little.
After that point she went on to speak of the function of Vpu to assist in degrading CD4 (again, not the “core argument” you assert). She claimed that this involved the evolution of two protein-protein binding sites. In my reply to her, I pointed out that in my book I placed viral protein – cellular proteins binding in a different category (more about that later), so I didn’t reply further to that. However, I now think that reply wasn’t the best one. Rather, I should have simply pointed out that Smith’s references themselves show that “Vpu-mediated CD4 down-regulation and degradation is conserved among highly divergent SIVcpz strains.” (1) Now, why are we even talking about the evolution of a process which has been conserved from chimps and other primates to humans? Where was it shown that the Vpu-CD4 degradation process evolved two protein binding sites? We have primate viral ancestors that have that function and human viruses that have that function. What was the point here again?
She then has two paragraphs on what you call the “core” argument. I’ll get back to those, because right after those paragraphs she reverts to discussing CD4 degradation. She writes:
HIV-1 Vpu requires two casein kinase II sites. … Yet some SIVcpz Vpus have only one CKII site, and instead utilize a simple string of negatively charged amino acids in place of the second site. Different ways of performing similar tricks with totally different amino acids. I think that’s biochemically significant as well.
I disagree with her assessment; I think this is a trivial biochemical change given HIV’s mutation rate. Casein Kinase II sites have the consensus sequence (S/T-X-X-D/E). (2) (Notice one of the consensus amino acids is already acidic.) Now, since the mutation rate of HIV is about 10-4, and since there are 109-1010 viruses in an infected individual, that means any particular double mutation would happen in every individual with AIDS every day. Thus in every hundred persons with HIV, every day there would occur at least one virus with that consensus sequence mutated to, say D-D-X-E. (That would allow for three nucleotide changes, since two are needed to go from S/T to D/E.) If all that was necessary at the site was a blob of negative charge to replace the phosphate negative charge, you’d get that every day in a group of a hundred infected individuals. Smith thinks “that’s biochemically significant.” I think she hasn’t done her math. I view that as sequence drift — the feature stays the same (negatively charged residues), but the sequence drifts within limits.
The same goes for the handful of other changes that impress her.
For instance, Subtype C Vpus are characteristically longer than the others, have key phosphorylation sites shifted, have an extra CKII site, and its tertiary structure is totally different (Subtype B Vpus have an Mr of 43,000 in an SDS-PAGE gel, while Subtype C is 34,000).
Since casein kinase II sites consist of just two specific amino acid residues (separated by any two residues), phosphorylation sites can be shifted, created, or destroyed daily as easily as the calculations from the last paragraph show. By themselves they are no more significant than the simple drift in the sequence of proteins. The fact that the Vpu-fusion proteins molecular weight shifts from 43,000 to 34,000 in SDS-PAGE is difficult to interpret. It may or may not mean anything significant about the protein’s native structure.