A Response to Martin Poenie on Protein Evolution
Douglas Axe recently responded to University of Texas cell biologist Martin Poenie (here and here) regarding Stephen Meyer's new book Darwin's Doubt: The Explosive Origin of Animal Life and the Case for Intelligent Design. Dr. Poenie has a particular interest in the subject of proteins, and thus the discussion has tended to focus on the arguments against their origin by Darwinian means, advanced in Chapters 8-11 of Dr. Meyer's book.
Writing here at Evolution News & Views, Dr. Poenie has now contributed a rebuttal to Axe's remarks. Poenie's critiques are a welcome oasis from the uncivil and ad hominem attacks that have come to characterize the debate surrounding Darwin's Doubt. Unlike many of our critics, Poenie seems to have read the book. We would therefore like to thank and congratulate Dr. Poenie for engaging these issues in a serious way.
Here, I offer my own thoughts on his response.
Critiquing the work of Gauger and Axe (2011) on the evolutionary accessibility of novel enzyme functionality, Poenie writes:
As Axe notes, there are 250 out of 381 amino acids that differ between BioF and Kbl such that the two sequences are 34% identical over 381 aligned positions. From an evolutionary perspective, this is a long distance. It is not surprising that his efforts to swap functionality by swapping a few residues failed. Approaches similar to what Axe have done have failed repeatedly. As noted by Harms and Thornton, "The reason studies of this type fall short is that they ignore history" (Harms and Thornton, 2010, Current opinion in structural biology 20: 360-366).
What Poenie neglects to mention is how similar the two enzymes are. For an illustration of the structural similarity between Kbl2 and BioF2, take a look at the figure below (excerpted from Gauger and Axe, 2011).
The top two quadrants, labeled "A," show the structures of the BioF2 (left) and Kbl2 (right) dimeric enzymes. The bottom left quadrant, labeled "B," shows the aligned backbones of the BioF and Kbl monomers. The bottom right quadrant, labeled "C", depicts the catalytic side-chains present in the active sites of the two enzymes (BioF2 shown in blue, and Kbl2 in green). The external aldimine of Kbl2 is shown in orange, and red for BioF2.
Poenie also cites the criticism of Todd Wood that nobody thinks that BioF2 evolved from Kbl2 or vice versa. In other words, as he quotes, "The reason studies of this type fall short is that they ignore history" (Harms and Thornton, 2010).
This criticism, however, is simply to misconstrue the nature of the work. The purpose of the study was not to recreate history. Indeed, the paper itself makes this point clear. Rather it was to test the feasibility of converting enzyme function from any starting point with a common structure. For the evolutionary paradigm to be true, a conversion of this type -- between two highly structurally similar enzymes -- ought to be feasible, but was not. Ann Gauger explained why her study is intended to be a "disproof of concept," thereby rebutting this common objection, in an ID the Future interview here. Gauger and Axe also respond to this objection here, here and here.
But let's take that statement at face value. Is the only kind of evolutionary change that is possible that which has already occurred? If there are so few paths in sequence space that allow functional conversion, this would be suggestive of design and fine-tuning. The PLP-dependent transferase superfamily, for example, is an extremely diverse group of proteins. How did the first PLP-dependent enzyme evolve into enzymes capable of so many different chemistries, if evolution is so constrained?
In fact, the reason Gauger and Axe chose Kbl2 and BioF2 to study is because this should be among the simplest of conversions to new chemistry in the PLP-dependent family, given their structural similarity, even down to the catalytic residues in the active site. Yet for this case, and in similar studies, it has proven to be prohibitively difficult (e.g. see Gerlt and Babbitt, 2009).
On another point, Poenie argues, "[Gauger and Axe] have no idea what their substitutions actually did to their protein. For example, these mutations could have simply led to an unfolded protein." But Gauger and Axe deliberately kept their residue changes to a minimum, at a level unlikely to result in an unfolded protein. As the paper states:
We think it unlikely, however, that the Kbl variants in this study have been destabilized enough to disrupt folding--even the more highly altered ones like Kblg1,g2,g3 and Kblg1,g2,g3,N155H. There are two reasons for this. First, apart from substitutions that fall into the potentially highly disruptive classes mentioned previously (random replacements of buried side chains or glycines, or introduction of prolines at random locations), it appears that about 10% or more of the residues in natural proteins need to be changed before the cumulative structural disruption can be expected to cause complete loss of function. The twenty substitutions in Kblg1,g2,g3,N155H alter only 5% of the protein, and because the replacements come from corresponding positions in a protein with a very similar overall structure (BioF), they are not apt to be drastically disruptive. For example, the single proline introduced in this twenty position mutant (Y357P; see Table 1 legend) is at a turn with very similar backbone geometry to the turn at proline 350 in BioF. Consequently, we would not expect the Y357P substitution in Kbl to have anything like the disruptive effect that a randomly introduced proline might have. Second, if the conventional role distinction between scaffold and active site has any validity, side chains forming the ligand interface (i.e., the "front line" of the active site) must carry little or no responsibility for stabilizing the scaffold. The thinking, in other words, is that these residues are free to be optimized for substrate binding and catalysis precisely because the scaffold residues have been optimized to stabilize the overall folded structure. If this is true, or even approximately true, then the locations of the changes introduced in Kblg1,g2,g3,N155H imply that they are even less apt to cause structural destabilization.
Poenie contends that "efforts to change enzyme/binding specificity using ancestral reconstruction actually do work." In support of this, he references two papers (Bridgham et al., 2009; Smith et al., 2013). But do these papers support Poenie's contention? The first of these studies (Bridgham et al., 2009) has been addressed previously here at ENV by biochemist Michael Behe and by Ann Gauger. Here are links to these articles:
- Nature Publishes Paper on the Edge of Evolution
- Nature Paper Reaches "Edge of Evolution" and Finds Darwinian Processes Lacking
- Dollo's law, the symmetry of time, and the edge of evolution
- Piddling Pebbles and Empty Promises: Response to Carl Zimmer and Joseph Thornton
- Not So Many Pathways: Response to Carl Zimmer and Joseph Thornton
- Severe Limits to Darwinian Evolution: Response to Carl Zimmer and Joseph Thornton
- Probability and Controversy: Response to Carl Zimmer and Joseph Thornton
- Corticosteroid Receptors in Vertebrates: Luck or Design?
The Bridgham et al. paper actually demonstrates quite well the difficulties in shifting proteins to new functions. Beginning with a modern glucocorticoid hormone receptor, the team demonstrated that "the evolutionary path by which this protein acquired its new function soon became inaccessible to reverse exploration," and that "five subsequent 'restrictive' mutations, which optimized the new specificity of the glucocorticoid receptor, also destabilized elements of the protein structure that were required to support the ancestral conformation" (Bridgham et al., 2009). In other words, owing to epistatic interactions, the modern receptor is unable to regain the ancestral activity, showing exactly why such conversions are hard. In any case, the study only deals with modifications in steroid binding (i.e. modification of the preference for one hormone over another). This isn't a case of novel functionality, but rather only a tweaking of existing function. As Gauger and Axe write in their paper on enzyme conversion,
We focus not on minor functional adjustments, like shifts in substrate profiles, but rather on true innovations -- the jumps to new chemistry that must have happened but which seem to defy gradualistic explanation.
Transitions that require only minor modification (such as modifying a residue in a binding pocket to make it bigger) are a walk in the park compared to recruiting proteins to fundamentally new chemistry.
The other paper cited by Poenie is by Smith et al. (2013). This study examined a gene encoding an anthocyanin pathway enzyme called dihydroflavonol-4-reductase (Dfr) -- a gene found in the plant genus Iochroma that is known to be involved in the transition from blue to red colored flowers. Two functionally distinctive versions of the enzyme are found in Iochroma cyaneum and Iochroma gesnerioides, which possess blue and red flowers respectively. The authors identify three substitutions that facilitated color transition and two subsequent substitutions that "resulted in an additional increase in specificity for red precursors." The study was an exercise in phylogenetic inference. The study asks: What was the sequence of the most recent common ancestor between the two versions of the gene? The researchers then used a site-directed mutagenesis technique to create Dfr alleles that corresponded to the inferred ancestral sequence. Given that the two enzymes differ by only 12 amino acids, however, and three substitutions are necessary to facilitate the transition in flower color, such a conversion is much easier than a Kbl2-to-BioF2 transition. In any case, given that plant breeders regularly encounter variations of plants with different colored flowers, the notion that a different flower color could arise by mutational processes would hardly be surprising to ID theorists.
Poenie also claims that the work of Gauger and Axe ignores or "leaves out a major mechanism for change involving recombination" and that "recombination can do all the things that Axe thinks are impossible." Rather than reiterate what I have already said on this subject, I will refer readers to my previous relevant articles available here and here.
Poenie rounds out his response with a discussion of ORFan genes. He writes:
Finally, in regard to ORFans, in my view, Axe's argument simply backfires. ORFans are found in all genomes; prokaryotes, eukaryotes, bacteriophages and animal viruses. Remarkably, the recently discovered megavirus "Pandoravirus" has 2500 genes, almost all of which are ORFans. The fact that the number of ORFans tends to be constant from one type of organism to the next in prokaryotes and eukaryotes indicates that they not uniquely associated with the Cambrian explosion and that they are likely formed by mundane genetic mechanisms that operate in all organisms. In one particularly informative example, Toll-Riera et al. (2009. Mol. Biol. Evol. 26, 603-612) identified 270 primate-specific ORFans. Of these, 70% contained a transposable element. In other cases, where a particular gene appeared to be ORFan, the same organism had a paralogue that did show homology with other organisms suggesting that the ORFan in question underwent rapid divergence. Contrary to Axe and Meyer, the fact that ORFans could represent new genes generated by genetic mechanisms such as transposition really throws a monkey wrench into their arguments.
Neither Axe nor Meyer (or, for that matter, anyone else I know of) has argued that ORFan genes are "uniquely associated with the Cambrian explosion." Moreover, it's not at all clear why the fact that prokaryotes and eukaryotes have similar proportions of ORFans should do anything to bolster an unguided evolutionary origin for those genes. The point made by Axe in his first article responding to Poenie was that "we can expect to find many dedicated protein folds [with 'unique fold structures'] in each specific kind of animal, right down to the level of species." It has yet to be shown that haphazard mechanisms such as transposition of mobile elements are likely to construct functional proteins. Most of the evidence bearing on the question is sequence-based and indirect. The few studies that are experimental highlight the severe difficulties with such a mechanism (as I discuss here).
To conclude, Dr. Poenie's criticisms of ID research on protein evolution are hardly new, and have been addressed at length before (sometimes even in the paper he's critiquing). If these are the best criticisms that are available, then there is reason to be encouraged.