Evolution Icon Evolution
Intelligent Design Icon Intelligent Design

Harvard Molecular Geneticist Vindicates Michael Behe’s Main Argument in Darwin Devolves

When Michael Behe’s book Darwin Devolves came out last year, critics were quick to pounce. Skeptic Magazine wrote that “In Darwin Devolves, Michael Behe continues to dig himself further into the hole he opened 20 years ago with Darwins Black Box.” Three Quarks Daily stated that Behe’s central thesis in the book, “is clickbait, the book title misleading, and the argument long since rebutted.” 

That thesis is what Behe calls the “first rule of adaptive evolution,” namely that Darwinian processes tend to “Break or blunt any functional gene whose loss would increase the number of a species’s offspring.” A review in the journal Science called Behe’s arguments “quixotic” and charged that “[t]here are indeed many examples of loss-of-function mutations that are advantageous, but Behe is selective in his examples” — so much so that Behe “misrepresents theory and avoids evidence that challenges him.” 

A response in the journal Evolution had this to say:

Concise, catchy and matter-of-fact, Behe’s First Rule makes for a quality sound bite, but it is overly simplistic and untruthful to the data. Darwin Devolves overemphasizes loss-of-function mutations, and brushes off countervailing examples as nothing more than a “sideshow.”

You can find many responses to the critics here, but it’s still worth revisiting some relevant questions: Did Behe mislead readers in emphasizing the importance of loss-of-function mutations? Did Behe misrepresent evolutionary theory and ignore evidence by claiming that adaptations tend to involve degradative mutations? Did Behe dig himself into a hole by claiming that constructive mutations are less common than those that break or diminish functions? If a new article Current Biology means anything, the critics are not just rude but plain wrong:

In laboratory-based experimental evolution of novel phenotypes and the human domestication of crops, the majority of the mutations that lead to adaptation are loss-of-function mutations that impair or eliminate the function of genes rather than gain-of-function mutations that increase or qualitatively alter the function of proteins. Here, I speculate that easier access to loss-of-function mutations has led them to play a major role in the adaptive radiations that occur when populations have access to many unoccupied ecological niches.

Andrew W. Murray, “Can gene-inactivating mutations lead to evolutionary novelty?Current Biology, 30(10); R465-R471; emphasis added.

The Importance of Loss-of-Function

The author, Andrew Murray, is a professor of molecular genetics at Harvard. He gives every indication that he is a typical evolutionary biologist who believes that unguided material mechanisms can produce all the diversity of life. But like Behe, he recognizes that loss-of-function is very important to the evolutionary process:

My guiding hypothesis is that a significant fraction of evolutionary novelty has been produced by mutations that reduce or eliminate gene function.

Also much like Behe, Murray knows what the data say. He reviews multiple examples where Darwinian evolution proceeded via mutations that caused loss or degradation, rather than gain or improvement, of function. After predicting that when “there are the same number of genes where loss-of-function mutations improve fitness (i.e. are deleterious for the protein product of the gene but beneficial for the organism) as the number of genes that are targets for gain-of-function mutations and both classes of mutations confer similar fitness increases, the loss-of-function mutations will dominate the initial response to selection,” he observes that this hypothesis is supported by empirical data:

Two forms of evolutionary novelty support the inferences from these calculations: laboratory evolution and agricultural domestication. A variety of laboratory experiments have selected for novelty. One of the earliest such experiments exposed a bacterial species, Pseudomonas fluorescens, to unstirred liquid medium, instead of the traditional shaken flask. In response, two new phenotypes appeared, clones that formed mats on the surface of the medium and those that formed aggregates at the bottom of the flask, with both phenotypes caused by loss-of-function mutations. Further experiments led to even more complex phenotypes and these too were dominated by loss-of-function mutations. In budding yeast, two novel phenotypes have been experimentally evolved: multicellularity and a circadian oscillator. In both cases, most of the mutations that combined to produce the novel phenotypes were loss-of-function mutations. For the populations that evolved multicellularity to proliferate successfully on low sucrose concentrations, four different biological functions were altered by loss-of-function mutations: the transcriptional network inducing cleavage of the shared cell wall that holds mothers and daughters together after cytokinesis, catabolite repression — the functional module that represses metabolism of other carbon sources when cells are grown on glucose — the transcriptional mediator complex and a module that represses cell growth and division in response to carbon starvation.

He further observes that genetic analysis of domesticated corn crops often differ greatly from their wild ancestors, but these differences are largely the result of loss-of-function mutations:

Genetic analysis of crosses between agricultural and ancestral isolates have identified many of the mutations that gave rise to today’s [maize] crops. Of those that have been characterized in detail, more than half are loss-of-function mutations; in a list of 60 candidate genes involved in the domestication of 14 different crops, 38 of the genes have alleles that are either annotated as being loss-of-function mutations or have a high-probability of damaging gene function (frameshifts and premature stop codons).

“Drivers” of Human Evolution

But it’s not just artificially selected organisms like corn that have experienced degenerative mutations. Murray finds that loss-of-function mutations have been “drivers” of evolution in humans: 

The evolution of humans also offers support for loss-of-function mutations as drivers of evolution. In 1999, a pioneering review argued that in many organisms, loss-of-function mutations were likely to have been drivers of evolution as populations shifted from one environment to another. For example, a human loss-of-function mutation in the promoter of a red blood cell chemoreceptor, DARC, protects against malaria caused by Plasmodium vivax. A second example, also putatively driven by selection imposed by malaria, is the loss of function in CMAH, a gene whose product produces a sialic acid derivative. A comprehensive analysis identified 80 human genes that have inactivating mutations in humans but are intact in chimpanzees and proposed that at least one of these mutations, in the gene CASPASE12, was selected for because it conferred survival after sepsis.

New Niche for a Hopeful Monster

Lest you think that Murray believes that loss-of-function mutations occur but are mere unimportant curiosities in evolutionary history, he goes on to propose that these could be the very mechanisms behind adaptive radiations where new types of organisms evolve rapidly to fill empty niches. In his model, loss-of-function mutations occur more readily than gain-of-function mutations. Thus they may allow “hopeful monsters” to arise, potentially allowing those organisms to occupy a new niche. Normally these mutants get outcompeted by other organisms that already fill the niche. However, if the niche is empty then the organism might persist and undergo further mutations to specialize within the niche. 

Murray admits that it’s not yet known whether the evidence backs up this hypothesis. But he observes that loss-of-function mutations drove diversification in some African cichlid populations:

Another well-studied radiation has produced about 1200 species of cichlids in the rift lakes of East Africa, with an extraordinary variety of morphologies, behaviors, and colorations and in this radiation, inter-species crosses are possible. In three lakes, Victoria, Malawi, and Tanganyika, dark horizontal stripes have arisen, and in all three cases this is due to reduced expression of agouti-related peptide 2 (agrp2), which encodes a peptide that controls pigment distribution. Although the mutations responsible for reduced expression differ between the different lakes, in the youngest lake, Victoria, the causative mutation is a loss-of-function mutation in an agrp2 enhancer.

Even Michael Behe would agree that such lower taxa can diversify via degradative mutations. As he writes in Darwin Devolves,“degradative mutations can help organisms to adapt and in the process can sometimes shift them into new minor categories of genus and species.” (p. 160)

Objections to the Model

Murray is aware of various objections to his model. For example, he notes that some fear “Killing genes might be beneficial in the short term and even visible in natural populations, but should we worry that if evolution keeps doing it there will be no genes left?” The objection is reasonable, and Murray’s response isn’t entirely convincing. He argues that degradative mutations may eventually be reversed, but this would violate Dollo’s Law, a very old postulate in evolutionary biology which holds that evolution is not reversible. In Darwin Devolves, Michael Behe explains that the evidence suggests Dollo’s Law is applicable even at the molecular level: 

In 2009 Joseph Thornton’s group set out to explore whether something like Dollo’s Law applied to steroid receptors. Having shown in 2006 that the reconstructed ancestral steroid receptor could be changed by a few mutations into a weakened one they considered more similar to the modern GR receptor (as discussed above), Thornton’s lab decided to investigate the reverse problem whether, starting from the modern GR receptor (which binds just one kind of steroid), a pathway conducive to Darwinian evolution could be found back to the ancestral one (which binds several kinds including the one bound by modern GR). As an analogy, if we show that a metal rod can be made into a hammer by a series of beneficial steps, can a hammer be turned back into a plain rod in the same manner?

After much impressive technically difficult work, their answer was no. The modern GR receptor is stuck where it is. It can’t go home again, at least not with any reasonable probability by a Darwinian process. The reason is that the modern receptor has accumulated a number of other changes from the ancient one, some positively selected to help its function, others seemingly neutral. 6 Reversing them would be necessary to get the old function back, but changing them individually, one at a time, as Darwinism requires, either doesn’t help or actively hurts, so natural selection would not be expected to favor them. The authors conclude that “the probability of all [necessary mutations occurring] in combination would be virtually zero.”

Although theirs is the first study with the necessary depth to address the question of the reversibility of protein molecular evolution, they are confident that the results will be quite general-that is, most proteins will be stuck in their present roles. In fact, they predict that further work “will support a molecular version of Dollo’s Law.” That is, “as evolution proceeds, shifts in protein structure-function relations become increasingly difficult to reverse.”

Darwin Devolves, pp. 208-209

If this argument is correct, then it’s unlikely that inactivated or degraded genes can be restored even if an organism luckily finds an empty niche to thrive in. 

A Second Workaround

Murray then proposes a second workaround. He cites the standard evolutionary idea that “genes can be produced as well as destroyed, both by duplication and divergence and by the emergence of coding sequences from non-coding DNA.” Even if this mechanism is viable for generating new genes, it must be guaranteed that newly functional genes can be created at a faster rate than they are destroyed. But Behe gives us good reasons to understand why gene duplication is not an efficient mechanism for generating new genes:

Up until now the lighthearted assumption was that the acquisition of significant new abilities by divergence of gene duplicates through random mutation and natural selection was unproblematic. Yet Thornton’s work demonstrating the severe difficulties with even comparatively simple transformations-at least by a Darwinian mechanism-calls into grave question whether the others, even the most familiar, could have developed that way either.

As just one example, the alpha and beta chains of hemoglobin are universally thought to have arisen in the distant past from myoglobin-like precursor by gene duplication and divergence. In fact, more than twenty years ago in Darwin’s Black Box I myself pointed to them as examples of what Darwinian evolution could likely do. They certainly derive from a common gene, but whether that could have happened by a Darwinian process—whether a comparatively simple oxygen-binding protein could without direction yield the sophisticated oxygen-delivery system that is hemoglobin — is now very much an open question….

However, whenever multiple amino-acid substitutions or other mutations were needed to confer a substantially different activity on a duplicated protein, it can no longer be blithely assumed that the transition was navigated by Darwinian evolutionary processes. 

Darwin Devolves, pp. 213-215

Despite the fact that Behe discussed and directly critiqued gene duplication in his book, in a review of Darwin Devolves in the Washington Post, Jerry Coyne falsely claimed that Behe “ignores” mechanisms like gene duplication:

Behe selectively gives a handful of examples in which mutations have produced broken genes that are nevertheless useful, but he simply ignores the large number of adaptive mutations that do not inactivate genes. These include duplications, in which a gene is accidentally copied twice, with the copies diverging in useful ways (this is how primates acquired our three-color vision, as well as different forms of hemoglobin).

The accusation that Behe ignored counterarguments is patently false, thought it was a common theme among critics of his book. In this case Behe specifically mentions the origin of different hemoglobin genes by gene duplication — the very counterexample that Coyne claims Behe “ignores”!

ID Proponents Wouldn’t Disagree 

Perhaps unintentionally, Murray’s article implies a counterargument. It explains that mere gene duplication by itself does not represent novelty:

In molecular evolution, the appearance of a rotary motor, whether as an ATP synthase or a flagellar motor, was clearly novel, whereas most would agree the gene duplication and divergence that produced multiple, paralogous mitotic cyclins in fungi and animals was not.

That is a most curious sentence — curious not because it’s largely correct but curious because it is so correct. There is no reason to believe that Murray supports intelligent design. In fact, from the rest of the article there’s no doubt that he believes blind evolutionary mechanisms alone can do the job. But Murray’s claim here is exactly what ID proponents would say: evolutionary novelty requires explaining how entire systems come into existence, and citing gene duplication alone is not tantamount to explaining evolutionary novelty. As Evolution News has argued in the past, gene duplication does not give you anything new; it only gives you more of something you already had. 

Murray offers this conclusion: “I have argued that loss-of-function mutations may have played a major role in the evolution of novelty.” He doesn’t cite Michael Behe, and there’s no evidence that he got any ideas from reading Darwin Devolves. He would likely disagree with Behe about the efficacy of mechanisms such as gene duplication for generating new genes, as well as with Behe’s support for intelligent design. What’s most interesting here are the points of agreement — namely the hypothesis that degradative mutations are very common in the evolutionary process and may even predominate among adaptive features.

So this is good news: Mainstream evolutionary biologists are independently arriving at very similar conclusions to Behe’s central thesis in Darwin Devolves, the one that drew the most wrath from critics. In light of Murray’s article in Current Biology, Mike Behe appears to be vindicated.