Dawkins programmed the computer first to generate many separate strings (sequences) of English letters. He then programmed it to compare each string to the Shakespearean target phrase and select only the string that most closely resembled that target.2 The program then generated variant versions of that newly selected string and compared those sequences to the target, selecting, again, only the one that most closely resembled the desired target. This eventually generated—after many iterations—a string that matched the target perfectly.
Axe recognized immediately the role that Dawkins’s own intelligence had played. Not only did Dawkins provide the program with the information that he wanted it to generate (“Me thinks it is like a weasel”), he imbued the computer with a kind of foresight by directing it to compare the variant sequences of letters with the desired target. Axe realized that Dawkins’s program did not simulate natural selection, which by definition is neither guided toward nor given information about a desired outcome generations in the future.
Axe began to wonder if there was some other way to assess the creative power of the mutation and natural selection mechanism, not with clever analogies or simplistic computer simulations, but with experimental and mathematical rigor.
Axe realized that Dawkins was right about one thing: the importance of genetic information. Like his fellow engineer Murray Eden, Axe’s tendency to view biology as an engineer led him to ask whether selection and mutation could actually build new organisms. Axe’s own research explored the connection between process control (a field of study in engineering) and genetic regulation, a sophisticated version of automated process control at work on a molecular scale inside living cells. Since cells use proteins to perform various feats of regulation, Axe was acutely aware that building new organisms necessarily involved building new proteins, which would in turn require new genetic information.
FIGURE 10.1
Douglas Axe. Courtesy Brittnay Landoe.
But could mutation and selection generate the precise arrays of nucleotide bases necessary to build fundamentally new protein structures? Axe’s interest in this question eventually led him to the scientific papers of Robert Sauer and the proceedings of a seemingly obscure 1966 Philadelphia conference called “Mathematical Challenges to the Neo-Darwinian Interpretation of Evolution.”
Unresolved Issues
As Axe read the papers that Sauer’s research group had produced, he realized their importance as a first step toward answering the questions Murray Eden had raised at Wistar. If Sauer’s quantitative measures of rarity held up, then Axe thought it obvious that mutation and selection could not adequately search a space that large. If, on the other hand, subsequent mutagenesis experiments overturned Sauer’s work and showed that protein function was largely indifferent to changes in amino-acid sequence, then the number of functional sequences might be large enough that mutation and selection would have a good chance of finding new functional genes and proteins in a reasonable amount of time.
After completing his Ph.D., Axe made inquiries about doing postdoctoral research in a top research lab where he could address these unanswered questions. He was soon invited by Alan Fersht, a professor at the University of Cambridge and director of the Centre for Protein Engineering, part of the world-famous Medical Research Council (MRC) Centre at Cambridge, to join his research group.
The decision to accept Fersht’s offer was an easy one. The star-studded history of the adjoining Laboratory for Molecular Biology (LMB), arguably the birthplace of molecular biology, included such luminaries as James Watson, Francis Crick, Max Perutz, John Kendrew, Sidney Brenner, and Fred Sanger. Starting in the chemistry department and then moving to the MRC Centre, Axe hoped to apply his research training to resolve the uncertainty that surrounded the interpretation of Sauer’s results. Specifically, he wanted to eliminate what he saw as two sources of error in Sauer’s method in order to get a more definitive estimation of the frequency of functional sequences in sequence space.
Axe thought, first, that Sauer’s team might have underestimated the rarity of functional proteins. In their experiments, Sauer’s group tested the tolerance of proteins to amino-acid substitution by changing amino acids at one or a few consecutive sites without making any other changes to other sites at the same time, much like a typist introducing an isolated typographic error in an otherwise accurately transcribed text. Not surprisingly, Sauer and his colleagues found that many sites along a protein chain could tolerate these isolated amino-acid substitutions, just as the reader of a text with only a few typos can often make out its meaning. Sauer’s team seemed to be assuming that a similar tolerance would have emerged if they had changed many sites simultaneously.
Axe thought that this assumption ignored the importance of the larger context provided by the mostly unaltered protein. A single typographic error typically will not totally destroy the meaning of a section of English text, because of the surrounding context provided by the other words, as well as the correct letters in the altered word. That does not mean, however, that specificity of sequence doesn’t matter. Instead, the meaning of a sentence with a typographic error can be discerned only because the rest of the letters are specifically arranged into meaningful words and phrases that provide a context for determining the meaning of the incorrectly spelled word. For just this reason, however, the meaning of a sentence is rapidly degraded if errors are allowed to accumulate at multiple sites.
Axe wondered if much the same could be true of genes and proteins. He wondered whether multiple, as opposed to single, position changes would quickly degrade function and whether a tolerance for substitutions at individual sites was itself context dependent—whether the tolerance for substitution at one site might depend upon having highly specific sequences at other sites. Thus, without questioning Sauer’s experimental findings, Axe thought Sauer’s result lent itself to misinterpretation. To many molecular biologists, it suggested that proteins could readily accommodate many simultaneous changes to their amino-acid sequences at many positions and remain functional.
As it turns out, Sauer recognized the potential for misinterpretation of his results. As he explained in the very paper in which he developed his quantitative estimate of rarity, “this calculation overestimates the number of functional sequences, since changes at individual positions are less likely to be independent of one another as more positions are allowed to vary.”3
Another assumption in Sauer’s approach had potentially the opposite effect—exaggerating the rarity of functional proteins. Axe thought the test that Sauer and his colleagues used to decide whether their mutant proteins were functional required a higher level of function than natural selection might require. Sauer and his team judged proteins with less than about 5–10 percent of the function seen in the natural protein to be nonfunctional. Yet Axe knew that even damaged enzymes with less than 5 percent of normal activity could add significantly more benefit than no enzymatic activity at all. Thus, from a neo-Darwinian point of view, the emergence of even such handicapped proteins might confer a selectable advantage on an organism. Axe thought that by rejecting as nonfunctional such mutated sequences, Sauer’s team probably had introduced another estimation error. These competing errors made it hard to know if the estimate made by Sauer’s team was too high or too low or whether perhaps they might neatly cancel each other out. To eliminate both sources of possible error, Axe carefully designed a new series of experiments.
The Importance of Folds
Axe had a key insight that animated the development of his experimental program. He wanted to focus on the problem of the origin of new protein folds and the genetic information necessary to produce them as a critical test of the neo-Darwinian mechanism. Proteins comprise at least three distinct levels of structure:4 primary, secondary, and tertiary, the latter corresponding to a protein fold. The specific sequence of amino acids in a protein or polypeptide chain make up its primary structure. The recurring structural motifs such as alpha helices and beta strands
that arise from specific sequences of amino acids constitute its secondary structure. The larger folds or “domains” that form from these secondary structures are called tertiary structures (see Fig. 10.2).
Axe knew that as new life-forms arose during the history of life—in events such as the Cambrian explosion—many new proteins must also have arisen. New animals typically have new organs and cell types, and new cell types often call for new proteins to service them. In some cases new proteins, while functionally new, would perform their different functions with essentially the same fold or tertiary structure as earlier proteins. But more often, proteins capable of performing new functions require new folds to perform these functions. That means that explosions of new life-forms must have involved bursts of new protein folds as well.
FIGURE 10.2
Different levels of protein structure. The first panel at the top shows the primary structure of a protein: a sequence of amino acids forming a polypeptide chain. The second panel depicts, in two different ways, two secondary structures: an alpha helix (left), and beta strands forming a beta sheet (right). The third panel at the bottom shows, in two different ways, a tertiary structure—that is, a protein fold.
The late geneticist and evolutionary biologist Susumu Ohno noted that Cambrian animals required complex new proteins such as, for example, lysyl oxidase in order to support their stout body structures. When these molecules originated in Cambrian animals, they also likely represented a completely novel folded structure unlike anything present in Precambrian forms of life such as sponges or one-celled organisms. Thus, Axe was convinced that explaining the kind of innovation that occurred during the Cambrian explosion and many other events in the history of life required a mechanism that could produce, at least, distinctly new protein folds.
He had another reason for thinking that the ability to produce novel protein folds provided a critical test for the creative power of the mutation and selection mechanism. As an engineer, Axe understood that building a new animal required innovation in form and structure. As a protein scientist, he understood that new protein folds could be viewed as the smallest unit of structural innovation in the history of life.
It follows that new protein folds represent the smallest unit of structural innovation that natural selection can select. Of course, natural selection can operate on smaller units of change—individual amino-acid changes that result in slight functional advantages or fitness gains, but not new folds, for example. But what if the functional or fitness gains that natural selection preserves and passes on never generate structural innovations? What if, instead, it only preserves slight differences in the sequence or function of proteins that confer an advantage without altering structure? Then, clearly, fundamental changes in the form of an organism will not occur. Building fundamentally new forms of life requires structural innovation. And new protein folds represent the smallest selectable unit of such innovation. Therefore, mutations must generate new protein folds for natural selection to have an opportunity to preserve and accumulate structural innovations. Thus, Axe realized that the ability to produce new protein folds represents a sine qua non of macroevolutionary innovation.
Could random mutations generate such novel protein folds? Axe realized that answering this question depended upon measuring the rarity of functional genes and proteins in sequence space and determining whether random genetic mutations would have enough opportunities to search the relevant sequence spaces within evolutionary time.
Axe’s Initial Results
Axe read the paper by Sauer and his colleague John Reidhaar-Olson that estimated the proportion of functional protein sequences to be extremely low (1 in 1063). He noticed that the authors chose not to emphasize this measure of rarity, however, but instead the variety of amino-acid substitutions that the protein under study could tolerate.
In their paper, Reidhaar-Olson and Sauer also repeated a then popular idea that the amino acids buried in the interior of a folded protein (forming what is known as the hydrophobic core) are most important for specifying the structure, while the arrangement of the exterior amino acids did not matter nearly as much.5 They thought that the amino acids buried within folded proteins typically need only to be hydrophobic (water repelling), whereas the amino acids on the exterior, for the most part, need to be hydrophylic (water attracting). In fact, some protein scientists thought that these simple restrictions might be the whole story—that a functional protein fold might require nothing more than an appropriate arrangement of hydrophobic and hydrophylic amino acids in a given sequence.
At Fersht’s lab in Cambridge, Axe conducted an experimental test of this idea, and surprised himself with the first result. In a paper he coauthored in the Proceedings of the National Academy of Sciences in 1996, he reported his findings. When he replaced the entire thirteen-residue hydrophobic core of a small enzyme with random combinations of other hydrophobic amino acids, a high fraction of the randomized proteins (about one-fifth) still performed their original function. This suggested that proteins were, perhaps, less susceptible to functional loss as a result of sequence changes than Axe had thought.
Next he focused on the exterior of proteins, randomizing portions of two different proteins’ exteriors in much the same way that he had randomly changed the interior of one of them. This time his approach failed to produce any functional variants at all. Realizing that this seemed to contradict what Sauer and others had supposed, Axe decided to make only much more restrictive changes in the next trial. He replaced each exterior amino-acid residue only with its most similar amino-acid alternative. Nevertheless, both of the proteins that he studied still lost all function by the time he had replaced one-fifth of their exterior residues. Thus he concluded that the exterior parts of the proteins were much more susceptible to functional loss as a result of amino-acid changes than had been widely assumed.
In all this work, Axe designed his experiments to remedy the two sources of estimation error inherent in Sauer’s method. First, by studying amino-acid changes in combination rather than in isolation, he determined that the surrounding context typically did influence whether a particular amino-acid change at a particular site caused functional loss. In other words, he discovered that the neglect of the influence of the surrounding context had the effect of exaggerating the tolerance to amino-acid changes at particular sites, as he (and Sauer) had suspected.
Second, the proteins that Axe chose for his study made it possible for function to be detected at much lower levels than was possible in Sauer’s studies. For one protein Axe studied, the more sensitive test of function did indeed allow a greater proportion of single mutants to retain some function, with about 95 percent of the mutant proteins achieving the designation “active.” This suggested that Sauer’s less sensitive screen did contribute to another source of estimation error, this time in the opposite direction. Nevertheless, Axe’s more sensitive screen also enabled him to establish that even though single mutations allow many proteins to retain some function, they still diminish or damage the function of the protein—often enough to ensure that they will be eliminated by the purifying effect of natural selection. Further, because of the extreme sensitivity of his test for function, Axe learned that any single mutation that failed his test was single-handedly destroying function. He determined that fully 5 percent of such changes did destroy protein function.
Overall, therefore, he showed that despite some allowable variability, proteins (and the genes that produce them) are indeed highly specified relative to their biological functions, especially in their crucial exterior portions. Axe showed that whereas proteins will admit some variation at most sites if the rest of the protein is left unchanged, multiple as opposed to single amino-acid substitutions consistently result in rapid loss of protein function. This was the case even when these changes occur at sites that allow variation when altered in isolation.6 His new experiment also roughly confirmed Sauer’s earlier quantitative assessment of the rarity of functional proteins, desp
ite the estimation errors inherent in Sauer’s method. Why? Because it appeared that Sauer’s two estimation errors—ignoring context and using an insufficiently sensitive screen for function—did, in fact, roughly cancel each other out.
Despite these advances in understanding, Axe had not yet determined whether the greatly restricted picture of tolerance that his work had exposed would cause problems for the evolution of new protein folds. In order to answer that question he would need to obtain a more precise quantitative estimate of the rarity of proteins in sequence space.
Having developed a method that eliminated the main sources of estimation error in earlier mutagenesis experiments, Axe was now in a position to answer that question with unprecedented rigor. Once he did, he could determine whether random genetic changes would have enough opportunities—even on the scale of evolutionary time—to search the relevant sequence spaces for functional genes and proteins.
Don’t Leave the Fold
Of course, Axe understood that neo-Darwinists do not envision a completely random journey through nucleotide or amino-acid sequence space. They see natural selection acting to preserve useful mutational variations and to eliminate deleterious ones. Richard Dawkins, for example, likens an organism to a high mountain peak.7 He compares climbing the sheer precipice up the front side of the mountain to building a new organism purely by chance—random mutations alone. He acknowledges that this approach up “Mount Improbable” will not succeed. Nevertheless, he asserts that there is a gradual slope up the back side of the mountain that can be climbed in small incremental steps. In his analogy, the back side up “Mount Improbable” corresponds to the process of natural selection acting on many small random changes in the genetic text. What chance alone cannot do, natural selection acting on random mutations can accomplish through the cumulative effect of many slight successive steps.
Darwin's Doubt Page 21