The reason they fit so closely, of course, is that the process that made one part depended on the other part. The shape of the ice simply reflected the shape of the object that marked the boundary of the water; the water froze around the rock. As a vase breaks, the two sides of a crack are necessarily reflections of each other. A zig for one side is automatically a zag for the other. So in order to conclude that two closely matched parts were purposely intended to fit each other, not only do they have to be complex, but the process that made one has to have been independent of the process that made the other. One reason scientists initially hypothesized that antibodies “molded” themselves to the molecules they bound was that it seemed the easiest way to explain the match in shape—the shape of the antibody would be determined by the shape of what it bound. But when that simple explanation didn’t pan out, further research revealed the elegant immune system, which independently and efficiently covers all of shape space.
With a couple of interesting exceptions,6 protein-protein binding isn’t the result of processes analogous to breaking a vase or water freezing around a complex shape. It arises either from searching a huge shape-space library, as the immune system does, or by some nonrandom mechanism.
GIVE ME JUST A LITTLE MORE TIME
Time has always figured prominently in Darwinian explanations. Although few changes can be noticed in our own age, Darwinists say, over vast stretches of geological time imperceptible modifications of life can add up to profound ones. It’s no wonder that we don’t see much coherent variation going on in the biology of our everyday world—evolutionary processes are so slow that a human lifetime is like a moment. The work on malaria and HIV upon which I base much of the argument for the edge of evolution has mostly been done in just the past fifty years. So how can it tell us anything reliable about what could happen over millions or even billions of years?
Time is actually not the chief factor in evolution—population numbers are. In calculating how quickly a beneficial mutation might appear, evolutionary biologists multiply the mutation rate by the population size. Since for many kinds of organisms the mutation rate is pretty similar, the waiting time for the appearance of helpful mutations depends mostly on numbers of organisms: The bigger the population or the faster the reproduction cycle, the more quickly a particular mutation will show up. The numbers of malaria cells and HIV in just the past fifty years have probably greatly surpassed the number of mammals that have lived on the earth in the past several hundred million years. So the evolutionary behavior of the pathogens in even such a short time as a half century gives us a clear indication of what can happen with larger organisms over enormous time spans. The fact that no new cellular protein-protein interactions were fashioned, that mutations were incoherent, that changes in only a few genes were able to help, and that those changes were only relatively (not absolutely) beneficial—all that gives us strong reason to expect the same for larger organisms over longer times.
Still, are the numbers we’ve examined enough? A hundred billion billion (1020) malarial cells and HIV viruses is certainly a lot, but it’s minuscule compared to the number of microorganisms that have lived on the earth since it first formed. Workers at the University of Georgia estimate that 1030 single-celled organisms are produced every year; over the billion-year-plus history of the earth, the total number of cells that have existed may be close to 1040. Looked at another way, for each malarial cell in the past fifty years there have been about 1020 other microorganisms throughout history. Can we extrapolate from malaria and HIV to all of bacteria? To all of life?
Sure. We do of course have to be cautious and keep in mind that we are indeed extrapolating, but science routinely extrapolates from what we see happening now to what happened in the past. The same laws of physics that work here and now are used to estimate broadly how the universe developed over billions of years. So we can also use current biology to infer generally what happened over the course of life on earth. Since we see no new protein-protein interactions developing in 1020 cells, we can be reasonably confident that, at the least, no new cellular systems needing two new protein-protein interactions would develop in 1040cells—in the entire history of life, as illustrated in Figure 7.4. The principle we use to make the extrapolation—that the odds against two independent events is the multiple of the odds against each event—is very well tested.
We can be even more confident of extrapolating over all of life, because in some ways HIV itself has mutated as much as all the cells that have ever existed on earth. The mutation rate of HIV (and other retroviruses) is at least ten thousand times greater than the mutation rate of cells. The much higher mutation rate of HIV gives it an evolutionary advantage over cells that increases dramatically if multiple changes are needed. For cells of higher organisms, each nucleotide of DNA has at most a one in a hundred million (108) chance of mutating.7 The odds of getting any two particular nucleotides to change in a cell in the same generation is that number squared, or one in 1016. Any good bookie could do the math to see that it would take about 1040cells to generate all possible six-nucleotide mutations.8 On the other hand, when HIV replicates, each of its nucleotides has a one in ten thousand (104) chance of mutating. Two particular nucleotides changing at the same time in the virus would have odds of that number squared, one in 108, and so on. So to generate all possible six-nucleotide mutations in HIV would require only 1020 viruses, which have in fact appeared on earth in recent decades. In other words, while we have studied it, HIV has run the gamut of all the possible substitution mutations, a gamut that would require billions of years for cells to experience. Yet all those mutations have changed the virus very little. Our experience with HIV gives good reason to think that Darwinism doesn’t do much—even with billions of years and all the cells in the world at its disposal.
Incidentally, the results with HIV also shed light on the topic of the origin of life on earth. It has been speculated that life started out modestly, as viral-like strings of RNA, and then increased in complexity to yield cells. The extremely modest changes in HIV throw cold water on that idea. In 1020 copies, HIV developed nothing significantly new or complex. Extrapolating from what we know, such ambitious Darwinian early-earth scenarios appear to be ruled out.
E PLURIBUS UNUM
In trying to determine where lies the edge of evolution, I’ve relied heavily on one organism, the malarial parasite, with support from two other microbes, HIV and to a much lesser extent E. coli. Yet, even though malaria does attain enormous population sizes, still it’s only one kind of organism. There are millions of species of animals, and many more species of plants and microbes. Is it possible that some other organism might have a greater evolutionary potential than malaria or HIV or E. coli? Could it be that, unluckily, the best-studied examples just happened to be evolutionary laggards? That some bacterium or plant hidden away in an unexplored forest or ocean could run Darwinian rings around the million-murdering death?
Yes, in a logical sense it is possible. One can never completely rule out the unknown. Bare possibility, however, is a poor basis for forming a judgment about nature. A rational person doesn’t give credence to a claim based on bare possibility—a rational person demands positive reasons to believe something. Until an organism is found that is demonstrated to be much more adept than the malarial parasite at building coherent molecular machinery by random mutation and natural selection, there is no positive reason to believe it can be done. And the best evidence we have from malaria and HIV argues it is biologically unreasonable to think so.
What’s more, there are compelling reasons to suppose that the results we have in hand for malaria and HIV are broadly representative of what is possible for all organisms. In the past fifty years biology has unexpectedly shown that to a remarkable degree all of life uses very similar cellular machinery: With a few minor exceptions the genetic code is the same for all the millions of species on earth; proteins are made of the same kinds of amino acids; nucleic acids are made of the sa
me kind of nucleotides; and many, many other basic similarities. A biochemistry textbook typically observes, “Although living organisms…are enormously diverse in their macroscopic properties, there is a remarkable similarity in their biochemistry that provides a unifying theme with which to study them.”9
The physical forces between proteins do not vary from organism to organism, nor does protein shape space depend on species. Since the criterion we are using to determine the edge of evolution is the development of specific protein-protein interactions, which is one of the most fundamental features of life, in that regard malaria is no different from any other organism.
Another possible objection is that malaria and HIV were just trying to get rid of poisons—to counter antibiotics—any way they could. Since the problem they were trying to solve is so narrow, it’s not surprising (one might say) that changes were concentrated in a few proteins, and that nothing at all complex was produced. Yet that objection would run up against a contradiction. It is widely thought that when it first appeared, atmospheric oxygen itself was poisonous to cells. But it is also widely thought in Darwinian circles that random mutation and natural selection allowed cells not only to tolerate the poison, but to construct enormously complex cellular mechanisms to take advantage of oxygen. Richard Dawkins opined that arms races build complex coherent machinery—where is the complex new machinery to deal with chloroquine? If Darwinism could spin gold out of once-deadly oxygen, why can natural selection do nothing with modern antibiotics? The obvious answer is that the premise is wrong: Random mutation did not build either the complex cellular machinery of respiration or any other. Left to its own devices, mutation and selection produce the disjointed, limited responses we see for the case of modern antibiotics.
ONE AT A TIME
The conclusion from Chapter 7—that the development of two new intracellular protein-protein binding sites at the same time is beyond Darwinian reach—leaves open, at least as a formal possibility, that some multiprotein structures (at least ones that aren’t irreducibly complex, in the sense defined in Darwin’s Black Box) might be built by adding one protein at a time, each of which is an improvement. But there are strong grounds to consider even that biologically unreasonable. First, the formation of even one helpful intracellular protein-protein binding site may be unattainable by random mutation. The work with malaria and HIV, which showed the development of no such features, puts a floor under the difficulty of the problem, but doesn’t set a ceiling. Maybe my conservative estimate of the problem of getting even a single useful binding site is much too low. What we know from the best evolutionary data available is compatible with not even a single kind of specific, beneficial, cellular protein-protein interaction evolving in a Darwinian fashion in the history of life.
A second reason to doubt a one-protein-at-a-time scenario is the demanding criterion of coherence. The longer an evolutionary pathway, the much more likely that incoherent, momentarily-helpful-but-dead-end mutations will sidetrack things. The pathway to just one binding site is long, so the pathway to a second one is even longer. That means many more opportunities to take a wrong turn and get stuck on some tiny hill in a rugged landscape. As noted earlier, Allen Orr showed that on average just one or two steps would land an organism at a local evolutionary optimum, unable to progress further. Although Orr was discussing nucleotide sequences, it is reasonable to think the same consideration operates at other biological levels, so that at best one or two new protein-binding sites would present a local evolutionary peak, resistant to further change.
A third reason for doubt is the overlooked problem of restricted choice. That is, not only do new protein interactions have to develop, there has to be some protein available that would actually do some good. Malaria makes about fifty-three hundred kinds of proteins. Of those only a very few help in its fight against antibiotics, and just two are effective against chloroquine. If those two proteins weren’t available or weren’t helpful, then, much to the joy of humanity, the malarial parasite might have no effective evolutionary response to chloroquine. Similarly, in its frantic mutating, HIV has almost certainly altered its proteins at one point or another in the past few decades enough to cover all of shape space. So new surfaces on HIV proteins would have been made that could bind to any other viral protein in every orientation. Yet of all the many molecules its mutated proteins must have bound, none seem to have helped it; no new protein-protein interactions have been reported. Apparently the choice of proteins for binding is restricted only to unhelpful ones.
Restricted choice is a problem not only in fighting antibiotics, but also in fighting the environment and other organisms. Although malaria has only had a few decades to deal with manmade antibiotics, it has pretty much trashed them all, because only one or a few point mutations were needed. Yet it’s had about ten thousand years to deal with the sickle hemoglobin mutation and has been unable to get around it. The same with other human genetic responses to malaria—thalassemia, hemoglobin E, and so on. It may be that there simply is no effective mutational response that is available to malaria. The same with its vulnerability to chilly temperatures. Even though Antarctic fish cobbled together an antifreeze system by random mutation to survive in icy waters, with many more chances P. falciparum hasn’t learned to even knit itself a figurative sweater. As with sickle hemoglobin, it seems likely that there simply is no available evolutionary response. Nothing helps.
When you are building a fine-tuned, multicomponent cellular structure, the problem gets exponentially more severe at each step, as many specialized components are required. The bottom line is, it’s reasonable to think that building multiprotein complexes one protein at a time is also well beyond the edge of evolution.
WHAT LIES BEYOND THE EDGE?
Although Darwin’s is one theory of how unintelligent forces may mimic intent, it isn’t the only one. So if random mutation and natural selection can’t do the trick, maybe some other unintelligent process can. Although Darwin’s theory is far and away most biologists’ favored account for the appearance of design in life, a minority of biologists think it’s woefully inadequate and prefer other unintelligent explanations.
One of the more popular minority views, called “complexity theory” or “self-organization,” has been championed for decades by Stuart Kauffman, currently of the University of Calgary. The use of the term “self-organizing” can be a bit confusing because all of biology is profoundly self-organizing, as we saw with the example of IFT. But that’s not what’s meant here. Self-organization theorists use the term in a more general way. For example, one nasty example of self-organization from our everyday world is a hurricane—when conditions are right, the ocean, atmosphere, and heat combine to forge a highly organized storm that can persist for weeks. But most of the physical details of the system aren’t critically precise. It’s also completely unclear how the concept would apply to evolution. While it’s certainly plausible that in some instances biological systems can self-organize in Kauffman’s sense, there’s no reason to think that self-organization explains how complex genetic systems arose.10 Here’s an illustration from everyday life. Some very simple rush hour traffic patterns are self-organizing, but self-organization does not explain where very complex carburetors, steering wheels, and all the other physical parts of a car came from, let alone how “cars could be manufactured by merely tumbling their parts onto the factory floor.” In the same way intraflagellar transport might be self-organizing in the sense that it self-assembles, but self-organization doesn’t explain how the structures that IFT depends on arose.
A second rival to Darwin has been dubbed “natural genetic engineering” by its most prominent proponent, University of Chicago biologist James Shapiro. The gist of the idea is that cells contain the same tools that human genetic engineers use to manipulate genes, to clone, and generally to tinker with life. In fact, in most cases that’s where the human engineers got the tools—from cells. Cells have proteins that can cut pieces out of DNA, mov
e them to different places in the cell, repeatedly duplicate genes, and so on. What’s more, the cell itself “knows” where critical regions in the DNA are: where genes start and stop, which regions are inactive and which are active, and so on. The cell “knows” this because it contains proteins that sense all those features. Since cells contain sophisticated tools, the argument suggests, evolution doesn’t have to proceed in a Darwinian manner by tiny random changes. It can progress in big steps, just as human genetic engineers take big steps when manipulating cells.
In many ways Shapiro has a higher, more respectful view of the genome than do Darwinists. Over the years, some Darwinists have derided portions of DNA where sequences are repeated many times as “junk.” Shapiro disagrees:
Despite its abundance, the repetitive component of the genome is often called “junk,” “selfish,” or “parasitic” DNA…. We feel it is timely to present an alternative “functionalist” point of view. The discovery of repetitive DNA presents a conceptual problem for traditional gene-based notions of hereditary information…. Weargue here that a more fruitful interpretation of sequence data may result from thinking about genomes as information storage systems with parallels to electronic information storage systems. From this informatics perspective, repetitive DNA is an essential component of genomes; it is required for formatting coding information so that it can be accurately expressed and for formatting DNA molecules for transmission to new generations of cells.11
Shapiro thinks the genome is much more sophisticated than we had supposed; it’s like a computer that contains not only specific programs, but an entire operating system. Shapiro’s thinking makes random (although not “Darwinian”) evolution more plausible, because the randomness includes steps that are more likely to be helpful.12
Unfortunately, in my view, natural genetic engineering proponents mistake cause for effect. Although big changes in repetitive DNA sequences certainly may affect gene expression13 and animal shapes14 (just as point mutations in proteins and more “traditional” Darwinian processes may do),15 natural genetic engineering does not explain where the engineering tools came from, or how they can be employed coherently, or how formatting came about, or how it might change coherently, or a host of other pressing questions.
The Edge of Evolution Page 17