Ann Gauger already provided an excellent series of articles discussing her recent paper co-authored in BIO-Complexity. She explains why it is important for demonstrating intelligent design (see here, here, here, here, and here). However, I wanted to give a slightly different framing of the new data. My purpose here and in a follow-up post will be to explain how this latest ID research (as well as prior work in the field) addresses fundamental questions in the debate over Darwinian evolution and ID -- and comes up with very positive results.
When Michael Behe published Darwin's Black Box in 1996, he outlined the concept of irreducible complexity as a biochemical challenge to Darwinian evolution. According to his argument, some structures require all of their parts (or a certain core minimum number) in order to function. In this game of all or nothing, such "irreducibly complex" structures cannot be built in the step-by-step manner of Darwinian evolution because they provide no advantage, of the kind evolution requires, until all their parts are present.
At the time and since, evolutionists responded by arguing that irreducibly complex features can be built through co-option. Under this view, biological parts might exist elsewhere in an organism where they are used for a different function. Sometimes, when a gene is duplicated, the extra copy could then be borrowed, retooled, and "co-opted" to perform a new function in a new system. Sounds easy, right?
Not so fast. Darwin observed that his theory "breaks down" when large leaps of complexity are needed in order to provide some advantage. What if retooling proteins to perform new functions requires such leaps? What if multiple mutations are needed to convert a protein to perform some new function? Should we follow many evolutionists and simply wave our hands and assume the co-option model of protein evolution is sufficient, or should we behave like scientists and test the viability of that model? Researchers at Biologic Institute have decided to take the latter approach, and to test the co-option model of protein evolution.
In 2011, Ann Gauger and Douglas Axe published a paper in BIO-Complexity, "The Evolutionary Accessibility of New Enzymes Functions: A Case Study from the Biotin Pathway." They reported results of their laboratory experiments trying to convert one enzyme (Kbl2) to perform the function of a very similar enzyme (BioF2), thought to be very closely related to Kbl2. Because these proteins are both members of the GABA-aminotransferase-like (GAT) family, and are believed to be very closely related, this is the sort of evolutionary conversion that evolutionists say ought to be easily accomplished under the standard co-option model. However, after trying multiple combinations of different mutations, they found otherwise:
We infer from the mutants examined that successful functional conversion would in this case require seven or more nucleotide substitutions.
This presents a serious problem for Darwinian evolution since a 2010 paper by Axe found that a feature that would require more than two maladaptive mutations, or more than six neutral mutations, before providing an advantage could not arise in the entire history of the earth. Gauger and Axe (2011) concluded:[E]volutionary innovations requiring that many changes would be extraordinarily rare, becoming probable only on timescales much longer than the age of life on earth. Considering that Kbl2 and BioF2 are judged to be close homologs by the usual similarity measures, this result and others like it challenge the conventional practice of inferring from similarity alone that transitions to new functions occurred by Darwinian evolution.
Their 2011 study thus provided a "disproof of concept" of the co-option model. However, it only looked at two proteins. What if other proteins are more easily convertible? Last December, in a landmark peer-reviewed paper published in BIO-Complexity, Axe, Gauger, and biologist Mariclair Reeves presented new research on additional proteins in the same family. They showed that these proteins too are not amenable to an evolutionary conversion to perform the function of BioF2.
Now in their new study, "Enzyme Families-Shared Evolutionary History or Shared Design? A Study of the GABA-Aminotransferase [GAT] Family," Reeves, Gauger, and Axe examine nine other enzymes from the same GAT family. Once again, the idea was to see if it is possible to convert them to perform the function of BioF2. They tested proteins that are closer to BioF2, or more distant from BioF2, than the enzyme they tested in their prior study (Kbl2). But all of the proteins studied are in the same family, and are thought to be closely related.
First, they sought to determine if the enzymes could be converted to perform the function of BioF2 through a single mutation. They created mutation libraries with every single possible mutation in those nine enzymes. No BioF2 function was ever detected. As they explain:
The present study has added to our previous examination of these problems in several respects. We have shown, based on sequence alignment of α-oxoamine synthases (a subset of the GAT family), that our previous use of rational design did indeed target regions of Kbl2 that are likely to be functionally significant. Furthermore we have now shown that the lack of a simple evolutionary transition to BioF2 function is not at all unique to our initial choice of Kbl2 as the starting point. Single mutations cannot convert any of eight other members of the GAT family to that function, despite the fact that all of these enzymes are regarded as close evolutionary relatives.
Then they tried double-mutation libraries. They were able to try 70 percent of the double-mutations in two enzymes that are thought to be the most likely candidates for conversion to function like BioF2: Kbl2 and another enzyme, BIKB, which is said to have both BioF2 and Kbl2 functions. They write:[W]e have demonstrated that converting either Kbl2 or BIKB to perform the function BioF2 with two DNA base substitutions is at least mildly unlikely, in that neither conversion was found after examining over two thirds of the possibilities. Of course, many possibilities remain unexamined. Although it is certainly possible for a working combination to be among those unchecked possibilities, we think it is more informative at this point to ask the fundamental question of whether the available evidence as a whole really supports the idea that evolutionary recruitment is the cause of functional diversity in enzyme families.
This suggests that at least two mutations would be necessary to convert a protein in this family to perform the function of BioF2. But it's likely that additional mutations would be necessary for these evolutionary conversions. To be specific, the co-option model requires that a gene become duplicated, and then overexpressed before it can evolve some new function. Thus, at least two more mutations are needed -- one to duplicate the gene and another to overexpress it.
However, one of their citations makes a compelling case that the gene duplication step poses a major obstacle to gene recruitment via gene duplication and mutation. They cite a paper from PLOS Genetics, "The Extinction Dynamics of Bacterial Pseudogenes," which notes:
In bacteria, however, pseudogenes are deleted rapidly from genomes, suggesting that their presence is somehow deleterious. The distribution of pseudogenes among sequenced strains of Salmonella indicates that removal of many of these apparently functionless regions is attributable to their deleterious effects in cell fitness, suggesting that a sizeable fraction of pseudogenes are under selection.
It concludes, "Although pseudogenes have long been considered the paradigm of neutral evolution, the distribution of pseudogenes among Salmonella strains indicates that removal of many of these apparently functionless regions is attributable to positive selection."
Don't miss the profound importance of this. What it means is that there is very likely a fitness cost associated with carrying an extra, useless copy of a gene, and therefore it can be advantageous to delete duplicate version. This has implications for the co-option model of protein evolution, because it shows that producing a new protein does not involve "neutral evolution," but rather requires steps that impose a deleterious effect upon the organism.
Similar results were obtained in a 2010 BIO-Complexity paper reporting ID research, "Reductive Evolution Can Prevent Populations from Taking Simple Adaptive Paths to High Fitness," by Gauger and biologist Ralph Seelke. This paper found that when merely two mutations along a stepwise pathway were required to restore function to a bacterial gene, even then the Darwinian mechanism failed. But the reason why Darwinism failed is important.
Although one mutation restored partial function, the function was very slight. In the experiments, the gene was deleted before the second mutation could occur and restore full function. It was more advantageous to delete a weakly functional gene than to continue to express it in the hope that it would "find" the mutations that fixed the gene. This means that carrying a weakly functional gene is disadvantageous, and that it's more advantageous to just get rid of a gene duplicate that isn't contributing very much to the success of the organism.
So a fundamental component of the co-option model -- duplicating a gene -- very likely involves a deleterious step. What does that say about the potential for evolving complex traits? ID research has already addressed that question.
In a 2010 research paper that Douglas Axe published in BIO-Complexity, "The Limits of Complex Adaptation: An Analysis Based on a Simple Model of Structured Bacterial Populations," he determined that when deleterious mutations are involved, a trait that requires more than two disadvantageous mutations could never form over the history of the earth. In other words, if you are trying to evolve a trait that requires more than two deleterious mutations, it's not going to evolve.
Let's return now to the 2014 paper by Reeves, Gauger, and Axe. They experimentally found that at least two mutations would be necessary to convert the most likely co-option candidates in this enzyme family to function like BioF2. But other mutations would be necessary as well -- at least one to duplicate the enzyme and another to overexpress it. This suggests at least four mutations would be required for this conversion. Given that, as we just saw, some of these mutations (such as duplication) would initially impose a disadvantage, the math says it would take some 1015 years for the necessary mutations to arise to co-opt a protein to function like BioF2 -- over 100,000 times longer than the age of the earth! They thus conclude:
Based on these results, we conclude that conversion to BioF2 function would require at least two changes in the starting gene and probably more, since most double mutations do not work for two promising starting genes. The most favorable recruitment scenario would therefore require three genetic changes after the duplication event: two to achieve low-level BioF2 activity and one to boost that activity by overexpression. But even this best case would require about 1015 years in a natural population, making it unrealistic. Considering this along with the whole body of evidence on enzyme conversions, we think structural similarities among enzymes with distinct functions are better interpreted as supporting shared design principles than shared evolutionary histories.
This new paper thus provides a robust disproof of the co-option model, overturning a cornerstone argument that evolutionists have long used when trying to answer ID arguments like irreducible complexity. By testing the co-option model, Biologic Institute is not just asking the right questions and doing innovative research that addresses key issues in the debate over Darwinian evolution and intelligent design. They're also finding data that confirms that ID's earliest arguments were right all along. 
