At his blog University of Chicago professor of evolutionary biology Jerry Coyne has commented on my reply to his analysis of my new review in the Quarterly Review of Biology. This time he has involved two other prominent scientists in the conversation. I’ll discuss the comments of one of them in this post and the other in a second post. The first one is University of Texas professor of molecular biology James J. Bull, who works on the laboratory evolution of bacterial viruses (phages). I reviewed a number of Bull’s fascinating papers in the recent QRB publication. Coyne solicited Prof. Bull’s comments and put them up on his blog. Bull says several nice things about my review, but agrees with Prof. Coyne that he wouldn’t expect “novelty” in the lab evolutionary experiments he and others conducted, and he thinks they are not a good model of how evolution occurs in nature. (I wonder if he mentions this in his grant proposals….)
Prof. Bull states that bacteriophage T7 (which he used in his studies) avoids taking up DNA from its host, E. coli, so it really isn’t an example of a system where novel DNA was available to the phage, despite his initial hopes that it would be. (In the paper describing the work he and his co-authors wrote, “At the outset, our expectation from work in other viral systems was that the loss of ligase activity would … require the [T7] genome to acquire new sequences through recombination or gene duplication.”) But, he writes in his new post, “what we failed to point out in our paper, and is fatal to MB’s criticism, is the fact that T7 degrades E. coli DNA, so even if the phage did incorporate an E. coli gene, it might well destroy itself in the next infection.” This reasoning strikes me as overlooking an obvious problem, and overlooking an obvious solution to the problem.

First, the problem. If T7, and presumably other bacteriophages, find it advantageous to have a mechanism that excludes host DNA from being incorporated into the phage genome, doesn’t this drastically cut down the opportunity for the very mixing of cross-species DNA that Coyne and Bull tout as the Darwinian solution to the problem of developing complex new functions? I suppose they could respond that, well, maybe the phages can’t exclude other, non-host DNA, so that’s where novel DNA would come from. But it seems host DNA would be by far the DNA the phages contacted the most. But if that is essentially excluded as a source, then the sorts of compensatory mutations that Professor Bull observed in his experiments are still by far the most likely ones to occur in nature. (And the grant application is saved!) It’s a matter of rate. The adaptive mutations that come along first will be selected first, and clearly point mutations and deletions come along very rapidly in phage populations.
Next, the solution. If a phage has a mechanism that is preventing it from taking up DNA that could be advantageous to it (such as the gene for a DNA ligase in the case of the experiment of Rokyta et al 2002 then all it has to do is break that mechanism and the opportunity for acquiring DNA is now opened to it. After all, breaking things is what random mutation does best, and, as I reviewed, many of the reported adaptive mutations in lab evolution experiments resulted from broken genes. Broken genes can also be neutral mutations. In the majority of the cultures of E. coli that Richard Lenski has grown for 50,000 generations, “mutator” strains took over. A mutator strain is one which has lost at least part of its ability to repair its DNA. If E. coli can toss out part of its repair ability with impunity, why couldn’t T7 lose its inability to take up some helpful host DNA?
Professor Bull suggests that lab evolution experiments which use whole cells and viruses aren’t needed to show the power of Darwinian processes because that is apparent in experiments using “directed” evolution. I strongly disagree with his assessment. In directed evolution, workers use an experimental set-up so that a single, particular gene or protein must mutate to be adaptive. “Directed” evolution is a much, much more artificial system than ones that use whole cells and/or viruses, as he did. In response to some selective pressure, a cell has potentially very many more ways to adapt to deal with it than does a single protein — a cell has thousands of genes and thousands of regulatory elements that can potentially help the cell adapt by gain- or loss-of-function, or tweaking of pre-existing function. On the other hand, directed evolution artificially constrains the system to mutate the component that the experimenter chooses. It seems a bit inconsistent to me for someone to claim that single species of cells (and/or viruses) are insufficiently complex to produce gain-of-function mutations by Darwinian processes, but that artificially constraining mutation/selection to single genes or proteins shows it clearly. Seems to me this is exactly backwards.
In his post Professor Bull describes an experiment he did with coworkers which, they hoped, would mimic the process of gene duplication and divergence. They placed two copies of the same gene, each on its own kind of plasmid, into the same cell. The gene produced a protein that could disable one kind of antibiotic very well and disable a second kind of antibiotic rather poorly. In the presence of both antibiotics, they expected one of the copies of the gene to stay about the same, degrading the first antibiotic. They expected the second copy of the gene to accumulate point mutations which would help it become more efficient at degrading the second kind of antibiotic (from other publications such mutants were already known to exist.) The system, however, had its own ideas. Bull says that contrary to expectations, one of the genes was deleted and the other gene accumulated point mutations so that it did a decent job degrading both antibiotics.
Professor Bull writes, “This study merely illustrates that the conditions favoring the maintenance of two copies undergoing evolutionary divergence are delicate.” Skeptic that I am, instead of “delicate,” I would say it illustrates that the conditions are “rare.” That is, it demonstrates very nicely that having two copies of a gene under what seem to be ideal conditions for adaptive divergence is not enough. (A similar result using a different system was recently obtained by Gauger et al 2010. Other factors enter into the result as well. Since we don’t know exactly what those other factors are, or how rare they make successful duplication/divergence events, we should not automatically assume that the occurrence of duplicated and diverged genes in nature happened by unguided, Darwinian processes.