The origin of new proteins is a major focus of BioLogos author and biologist Dennis Venema’s arguments in Adam and the Genome, which is the subject of this review series. He devotes nearly an entire chapter to arguing that there are biochemical “stepping stones” showing that “forming new protein binding sites, structures, and functions includes processes readily accessible to evolution — even if numerous mutations are necessary — without invoking miraculous intervention.” (p. 70) None of these discussions are relevant to whether Adam and Eve existed, but they are relevant to intelligent design. Consider Venema’s main examples of gene evolution: the p24-2 gene, nylonase, and the lambda phage.
Venema starts by discussing the origin of the p24-2 gene found only in one species of fruit fly, Drosophila melanogaster. He believes that it is derived from another fruit fly gene, Éclair. He argues that they arose when a primordial copy of Éclair, found in many fruit fly species, duplicated in Drosophila melanogaster. It then evolved into p24-2. There are only five amino acids that are different between Éclair and p24-2.
Keep in mind the raw data here:
- A gene exists in Drosophila melanogaster that is similar to another fruit fly gene, except for five amino acid differences.
- We don’t know what this gene does but we know it is essential.
Based upon this evidence, Venema assumes that this gene uses a novel binding site, and that the binding site evolved via Darwinian evolution. How does that follow from the raw data? It doesn’t.
Indeed, Michael Behe commented on this study when it first came out, making these same points. To his credit, Venema quotes Behe pointing out that we really don’t know how the five mutation differences arose. For all we know, as Behe stated, “those five point mutations may have required guidance or design in their appearance.” That’s a very reasonable point. Here’s what Behe said in his response to this alleged example of gene evolution (“At BioLogos, Confusion over the Meaning of ‘Irreducibly Complex’”):
Professor Venema discusses several proteins from the fruit fly called “p24-2” and “Éclair,” which are very similar to each other and very likely arose through gene duplication. Venema claims the proteins are part of an irreducibly complex system, and that the new, duplicated protein is a new part of that system. Some pro-ID commenters at the site rightly take Venema to task for not spelling out what the ID system is, but that point should be pounded home much more. The putative new binding site of p24-2 that Venema emphasizes is likely not a necessary part of the overall IC system.
Let me explain. Suppose as Professor Venema does, that Eclair and p24-2 are parts of systems that transport specific other proteins that they bind in the cell. Imagine an analogous mechanical system that has a rotating metal wheel at the top of a pole. To two places on the wheel are attached wires that at the bottom hold a claw that has an oval hole in it. When the wheel rotates, every half turn it encounters a metal bump which pushes up the wire and causes the claw to open. Once it passes the brief bump, the claw closes. During that time the claw can pick up an object that has a part of it shaped like the oval hole in the claw. As the wheel makes another half turn, it encounters a small depression, which causes the wire to dip down, opening the claw, and releasing the object. At the other end of the oval-shaped object is a little array of magnets, that match a second object. That second object gets transported with the claw, attached to the magnetic end of the oval shaped object.
Now what we have here is an already-intact transport system. At best, the amino acid changes in p24-2 would be analogous to re-arranging the magnets at the end of the oval-shaped object, so that it could transport a different object than before. But that is just taking advantage of an already-existing IC system; it is not a new one. As the old adage goes, Darwinian evolution is a tinkerer. It can fiddle a bit with pre-existing IC systems, but is incompetent to build new ones.
A final point is that there are five specific amino acid differences between Eclair and p24-2. Professor Venema emphasizes that they apparently comprise a new binding site. However, he seems not to notice that they may not have arisen by random mutation. Once one gets beyond one or two random mutations, one can’t assume that multiple further mutations arose by chance. In other words, as far as anyone knows, those five point mutations (or a subset of them) may have required guidance or design in their appearance.
Another possibility is that the amino acid changes that produced p24-2 were selectively beneficial. One or two changes may have changed p24-2’s affinity for a new product, and then each sequential change after that improved things more. But the fact that p24-2 is essential in its function argues more for design than serendipity. To evolve a new essential function is a mysterious process. Presumably the organism could do without it at first, but then began to find it to be essential somehow. This could be done by design all at once; there is no clear evolutionary model.
Cataloguing the differences between two genes is not an argument that one gene could or did evolve from the other by Darwinian evolution. To establish that an evolutionary pathway is possible, one must evaluate the fitness advantages associated with each change and determine if there are sufficient trials (in terms of population size, mutation rate, and generation time) to produce the changes. If there are, then fine, Darwinian evolution can do the job. If there are not, then Darwinian evolution is not the best explanation.
Venema doesn’t do any of this necessary scientific analysis, and instead descends into more God wouldn’t have done it that way arguments. As he writes in Adam and the Genome: “If a new function was needed, why design the required gene to appear as a slightly modified version of a gene right next to it in the genome? Wouldn’t that potentially fool researchers into thinking they were observing the results of natural processes?” (p. 73) The obvious answer is that the designer isn’t trying to fool anyone, and perhaps there are functional reasons why the two genes need to be near one another.
The truth is that nobody knows exactly how p24-2 arose. Venema would have done better to adopt the cautious tone found in the paper he cites regarding p24-2. That paper states:
It is unclear how essential genes arise and how new genes accumulate essential functions.
(Chen et al. 2010, “New Genes in Drosophila Quickly Become Essential,” Science 330: 1682-1685 (Dec. 17, 2010).)
That could hardly be improved upon.
In Adam and the Genome, as well as in other writings in which he critiques intelligent design, Venema relies heavily on the example of nylonase which he claims shows that new functional proteins, and even new protein folds, can be created by random mutations. Venema thinks that we can see this because nylonase underwent a “frameshift” mutation, which is a kind of mutation that inserts or removes a base pair from a gene. This shifts the triplet reading frame by one, which completely changes the sequence of the resulting protein. Essentially, it garbles the gene. If something as randomizing as a frameshift mutation can produce a useful protein, Venema argues, then new functional proteins and in particular new protein folds are not all that difficult to evolve. This, he argues, refutes a core ID argument.
To summarize Venema’s argument against ID, it goes like this:
- Venema writes that a “major argument proffered by the ID movement is that new biological information cannot be produced by evolutionary mechanisms.” (p. 80)
- He then describes Stephen Meyer’s argument as saying “evolution cannot account for ‘specified, complex information’ that we observe in living systems, and that ID is the only known cause for information.” (p. 81) Although Venema fails to precisely define what he means by “evolution,” this is basically a correct description of Meyer’s viewpoint.
- Venema of course disagrees with this argument, and writes that there’s evidence of blind evolution producing new information through whole genome duplications (WGD).
- Venema then writes, “Even more dramatic are cases when a sequence that was not a gene at all becomes one, something called ‘de novo’ gene origination.” (p. 81)
- He proceeds to given an example of what he believes is “de novo” gene origination — nylonase. This is supposed to be his knock-down drag-out argument to show that evolution can produce new information.
The problem with Venema’s argument is that nylonase doesn’t seem to have arisen de novo from a frameshift mutation. Biologist Ann Gauger has shown that Venema’s arguments about nylonase are false, and that nylonase probably arose from a very similar pre-existing enzyme that could break down a similar molecule.
To see (or listen to) Gauger’s arguments, please see:
- “The Nylonase Story: When Imagination and Facts Collide”
- “The Nylonase Story: How Unusual Is That?”
- “The Nylonase Story: The Information Enigma”
- “Gauger: Is It Easy to Get A New Protein?”
To review, Gauger went back to the original literature and showed that nylonase did NOT arise by a frameshift mutation. In “The Nylonase Story: When Imagination and Facts Collide,” she writes:
Unfortunately, Venema doesn’t have the story straight. Nylonase has a particular fold, a particular three-dimensional, stable shape. Most proteins have a distinct fold — there are several thousand kinds of folds known so far, each with a distinct topology and structure. Folds are typically made up of small secondary structures called alpha helices and beta strands, which help to assemble the tertiary structure — the fold as a whole. Venema seems unclear about what a protein fold is, and the distinction between secondary and tertiary structures. Nylonase is not “chock full of folds.” No structural biologist would describe nylonase as “chock full of protein folds.” Indeed, no protein is “chock full of folds.” Perhaps Venema was referring to the smaller units of secondary structure I mentioned above, the alpha helices or beta strands. But it would appear he doesn’t know what a protein fold is.
Maybe that explains why Venema missed the essential point of the paper describing nylonase’s structure. The crystal structure of EII-EII’ (a nylonase hybrid necessary to be able to crystalize the protein) revealed that it is not a new kind of fold, but a member of the beta-lactamase fold family. More specifically, it resembles carboxylesterases, a subgrouping of that family. In addition, when the scientists checked EII′ and EII, they found that both enzymes had previously undetected carboxylesterase activity. In other words, the EII’ and EII enzymes were carboxylesterases. If it looks like a duck and quacks like a duck, it is a duck.
Thus, EII′ and EII did not have frameshifted new folds. They had pre-existing folds with activity characteristic of their fold type. There was no brand-new protein. No novel protein fold had emerged. And no frameshift mutation was required to produce nylonase.
Where did the nylon-eating ability come from? Carboxylesterases are enzymes with broad substrate specificities; they can carry out a variety of reactions. Their binding pocket is large and can accommodate a lot of different substrates. They are “promiscuous” enzymes, in other words. Furthermore, the carboxylesterase reaction hydrolyzes a chemical bond similar to the one hydrolyzed by nylonase. Tests revealed that both the EII and EII′ enzymes have carboxylesterase and nylonase activity. They can hydrolyze both substrates. In fact it is possible both had carboxylesterase activity and a low level of nylonase activity from the beginning, even before the appearance of nylon.
nylB′ may be the original gene from which nylB came. Apparently there was a gene duplication at some point in the past. The two genes appear to have acquired mutations since then — they differ by 47 amino acids out of 392. The time of that duplication is unknown, but not recent, because it takes time to accumulate that many mutations. However, at least some of those mutations must confer a high level of nylonase activity on EII, the enzyme made by nylB. The enzyme EII’ made by nylB’ has only a low ability to degrade nylon, while EII degrades nylon 1000 fold better. So one or more of those 47 amino acid differences must be the cause of the high level of nylonase activity in EII. Through careful work, the Japanese workers Kato et al. identified which amino acid changes were responsible for the increased nylonase activity. Just two step-wise mutations present in EII, when introduced into EII’, could convert the weak enzyme EII’ to full nylonase activity.
From Kato et al. (1991): “Our studies demonstrated that among the 47 amino acids altered between the EII and EII’ proteins, a single amino acid substitution at position 181 was essential for the activity of 6-aminohexanoate-dimer hydrolase [nylonase] and substitution at position 266 enhanced the effect.”
So. This is not the story of a highly improbable frame-shift producing a new functional enzyme. This is the story of a pre-existing enzyme with a low level of promiscuous nylonase activity, which improved its activity toward nylon by first one, then another selectable mutation. In other words this is a completely plausible case of gene duplication, mutation, and selection operating on a pre-existing enzyme to improve a pre-existing low-level activity, exactly the kind of event that Meyer and Axe specifically acknowledge as a possibility, given the time and probabilistic resources available. Indeed, the origin of nylonase actually provides a nice example of the optimization of a pre-existing fold’s function, not the innovation or creation of a novel fold.
As the scientists who carried out the structural determination for nylonase themselves note: “Here, we propose that amino acid replacements in the catalytic cleft of a preexisting esterase with the beta-lactamase fold resulted in the evolution of the nylon oligomer hydrolase.” (Emphasis added.)
Let’s put to bed the fable that the nylon oligomer hydrolase EII, colloquially known as nylonase, arose by a frame-shift mutation, leading to the creation of a new functional protein fold. There is absolutely no need to postulate such a highly improbable event, and no justification for making this extravagant claim. Instead, there is a much more parsimonious explanation — that nylonase arose by a gene duplication event some time in the past, followed by a series of two mutations occurring after the introduction of nylon into the environment, which increased the nylon oligomer hydrolase activity of the nylB gene product to current levels. Could this series of events happen in forty years? Most certainly. Probably in much less time. In fact, it has been reported to happen in the lab under the right selective conditions. And most definitely, the evolution of nylonase does not call for the creation of a novel protein fold, nor did one arise. EII’s fold is part of the carboxylesterase fold family. Carboxylesterases serve many functions and have been around much longer than forty years.
Thus, nylonase seems to have arisen from a pre-existing highly similar protein that already had an ability to degrade materials similar to nylon. Nylonase did not arise from a radical garbling of protein in a frameshift mutation. Instead, it is highly similar to a pre-existing protein. We already know that proteins can slightly change targets like this and so there is literally “nothing new” here, on multiple levels.
Keep in mind that in both Adam and the Genome and in some of his online article series on ID, nylonase is Venema’s central argument against intelligent design. He’s made a big deal about it. So you would think that if his case has been refuted, then Venema would have something to say on the topic. And how did Venema reply to Ann Gauger? See this post in which Gauger reviews Venema’s response:
Gauger explains that Venema did NOT in any way contest Gauger’s main points. Nor did he admit that he was wrong. Instead he tried to change the subject by citing a new paper — effectively he gave up on the nylonase argument and tried to use a new one. Gauger writes:
He responded by saying, essentially, “Move on, nothing to see here.” A combination of selective retelling, sleight of hand, redirection, and downgrading the importance of the nylonase story were his main techniques. … He begins with the canonical [nylonase] story, simply restated. No mention of the contrary facts. … But Venema did notice I was writing about the subject. … But at this point, oddly enough, rather than describe and refute my arguments, he changes the subject. He spends the rest of the post discussing a new paper that deals with other data regarding the ease with which random sequences benefit bacteria. He only mentions nylonase to dismiss it, in the closing paragraphs. … This is a classic tactic. When shown to be wrong about nylonase, Venema changes the subject. Now nylonase is something that happened in the past, he says, implying we can’t reliably make inferences about it. He even says that nylonase is a “distraction” — despite his own repeated use of the nylonase story in trying to argue against Stephen Meyer and Doug Axe.
Here, Gauger is exactly right. Rather than admit an error, Venema tried to deflect attention from what happened.
As a final example of gene evolution, Venema cites a study in which he believes a protein evolved a new binding site. He argues that this refutes Michael Behe’s thesis:
[T]hese mutations did not happen simultaneously, but rather sequentially. As it turns out, these single mutations allowed the protein J to bind more tightly to LamB, which was a significant advantage since hosts with LamB were so scarce in the experiment. … [T]he transition from one system to another required no less than four mutation events. Behe bases his entire argument on the assumption that these mutations must occur simultaneously. The probability of these four changes happening at once in one virus is about one in a thousand trillion trillion-far, far beyond anything that is remotely possible, and far beyond what Behe sets as his “limit” for what evolution can accomplish. The reason that the virus could repeatedly skip past Behe’s so-called limit was that the mutations happened sequentially, not simultaneously. Thus Behe is now faced with a concrete example of a new protein-binding site arising through multiple mutations, with that new binding event replacing a previously essential part of a complex system-and all documented to a level of detail that cannot be disputed.
(Adam and the Genome, pp. 79-80)
But Venema has misunderstood Behe.
Behe never denies that when each mutation provides an advantage, that an evolutionary pathway is possible. In fact, Behe has written that “if only one mutation is needed to confer some ability then Darwinian evolution has little problem finding it.” If this pathway is as Venema describes it, where “these mutations did not happen simultaneously, but rather sequentially” as “single mutations allowed the protein J to bind more tightly to LamB,” then it’s exactly the kind of system that Behe claims is evolvable.
However, when multiple mutations are required before any functional advantage is gained, Darwinian evolution tends to get stuck. As Behe explains, “if more than one [mutation] is needed, the probability of getting all the right ones grows exponentially worse.” Behe argues that if a system requires two or more mutations before providing any advantage, then it would be difficult to produce by Darwinian evolution unless the available probabilistic resources are very high. But this lamba-phage example apparently does not involve the origin of that kind of system because, according to Venema, each mutation provided a stepwise advantage.
In The Edge of Evolution, Behe explains that these types of systems are possible to evolve, provided that there is a smooth, stepwise pathway for evolution to traverse:
In order to forge the many complex structures of life, a Darwinian process would have to take numerous coherent steps, a series of beneficial mutations that successively build on each other, leading to a complex outcome. In order to do so in the real world, rather than just in our imaginations, there must be a biological route to the structure that stands a reasonable chance of success in nature. In other words, variation, selection, and inheritance will only work if there is also a smooth evolutionary pathway leading from biological point A to biological point B.
(The Edge of Evolution, p. 5)
Behe acknowledges that such systems can exist, as he further writes:
[W]e shouldn’t be at all surprised to see resistance of mosquitoes to the new insecticides arise and spread by Darwinian processes. The necessary preconditions are all there: tiny, incremental steps-amino acid by amino acid-leading from one biological level to another.
(The Edge of Evolution, p. 76)
Thus Venema has not raised a challenge to a system that Behe ever claimed could not evolve, nor did it arise in a manner that Behe says is impossible.
Venema might reply that because this example involves the origin of a new binding site, it produced a class of functionality that Behe had previously argued was difficult to evolve. But there’s more to the story.
In fact, Behe reviewed this very study soon after it was published in 2012. See “More from Lenski’s Lab, Still Spinning Furiously” in which he explains that it seems likely that no new binding site arose, and that this is modification of function rather than gain of a new function:
As the authors state, however, the mutated viral J protein can still bind to the original protein, LamB, which strongly suggests the same binding site (that is, the same location on the J protein) is being used. It turns out that both LamB and OmpF have similar three-dimensional structures, so that strengthening the binding to one fortuitously led to binding to the other.
In my review (Behe 2010) I discussed why this should be considered a “modification of function” event rather than a gain-of-function one. The bottom line is that the results are interesting and well done, but not particularly novel, nor particularly significant.
To me, the much more significant results of the new paper, although briefly mentioned, were not stressed as they deserved to be. The virus was not the only microbe evolving in the lab. The E. coli also underwent several mutations. Unlike for lambda, these were not modification-of-function mutations — they were complete loss-of-function mutations.
The mechanism the bacterium used to turn off LamB in 99 percent of cells to resist initial lambda infection was to mutate to destroy its own gene locus called malT, which is normally useful to the cell. After acquiring the fourth mutation the virus could potentially invade and kill all cells. However, E. coli itself then mutated to prevent this, too. It mutated by destroying some genes involved in importing the sugar mannose into the bacterium. It turns out that this “mannose permease” is used by the virus to enter the interior of the cell. In its absence, infection cannot proceed.
So at the end of the day there was left the mutated bacteriophage lambda, still incompetent to invade most E. coli cells, plus mutated E. coli, now with broken genes which remove its ability to metabolize maltose and mannose. It seems Darwinian evolution took a little step sideways and two big steps backwards.
In discussing the topic in Adam and the Genome, Venema doesn’t mention any of Behe’s commentary on this paper.
To repeat, Behe makes a strong case that no new binding site arose and that this is a “modification of function” rather than a gain of a new one. At the very least, it doesn’t show the evolution of a multimutation feature which requires multiple mutations before giving an advantage. It shows the evolution of a stepwise feature which, according to Behe, is not a challenge to Darwinism.
Photo: Drosophila melanogaster, by Botaurus (Own work) [Public domain], via Wikimedia Commons.