Then they went fishing. They used various molecules—either a different protein or other chemical—as bait and tried to pull out from the mixture of antibodies ones that would bind to the bait. They saw that, like the pieces in the swimming pool that would stick to each other about half of the time, on average the antibodies they isolated could bind to the bait with just moderate strength.8 Over the years they and other laboratories juiced up the size of the antibody library, from a hundred million to a hundred billion and more. They found that the strength of binding improves with the size of the library, so that the best binders from the big library stick very tightly to the bait,9 like the pieces in the swimming pool that spent 99 percent of their time together. This result—the bigger the library the better the binding—is pretty much what you would intuitively expect. After all, if by chance you find a shape in a smaller library that binds something reasonably well, then you have a chance with a larger library to find a shape that fits even better.
The general results from Winter’s lab have been consistently confirmed: In order to get a particular protein to bind to any other one with modest strength, on average you have to wade through about ten to a hundred million binding sites.10 Actually, these and other experimental results are strongly skewed in a way that underestimates library sizes that would be needed if mutation were truly random. In all of these experiments, mutations were deliberately confined to a coherent patch of amino acids that were close to each other on the surface of the protein, to make as many novel, binding-site-sized regions as possible. If the workers had not deliberately directed the changes to a coherent patch on the protein’s surface, most changes would be scattered, unable to effectively interact. In that case a very much larger number of mutations on average would have to be sifted to find one that stuck specifically to a target protein. So we can take the results of these experiments as a very optimistic estimate for the difficulty of searching shape space.
LOST IN SHAPE SPACE
The elegant immune system is designed to saturate shape space. But the situation is entirely different inside the cell. For cellular proteins there is no built-in mechanism to deliberately make new binding sites. Cellular proteins almost always are made with just one sequence, not billions of different sequences like antibodies. In general the only way to get a new sequence for a cellular protein is over many generations by random mutation. Searching through shape space with cellular proteins is glacially slow and abysmally inefficient.
How much of a hurdle is it for Darwinian evolution? Consider a hypothetical case where it would give an organism some advantage if a particular two of its proteins, which had been working separately, bound specifically to each other. Perhaps the two-protein complex would be able to perform some new task, or do an old task much better. The lesson from shape space is that, in order for the one to bind the other, we should expect to have to search through tens of millions of different mutant sequences before luckily happening upon one that would specifically stick with even modest strength, which would allow the two to spend even half of their time together. (This is likely the minimum necessary strength, enough to have a noticeable biological effect.)11 Since the mutation rate is so low—about one mutation at a particular site in a hundred million births—we would expect to have to slog through an enormous number of organisms before striking on that lucky one.
Let’s make a rough calculation for the average number of organisms we would have to slog through to find a new protein-protein binding site. As I said, shape space tells us that about one in ten to a hundred million coherent protein-binding sites must be sifted before finding one that binds specifically and firmly to a given target. The simplest way to alter a protein is by point mutation, where one amino acid is substituted for another at a position in a protein. There are twenty different kinds of amino acids found in proteins. That means that if just five or six positions changed to the right residues—the ones that would allow the two proteins to bind—that would be an event of approximately the right frequency, since twenty multiplied by itself five or six times (205 or 206) is about three million or sixty million, respectively—relatively close to the ten to a hundred million different sites we need.
So one way to get a new binding site would be to change just five or six amino acids in a coherent patch in the right way.12 This very rough estimation fits nicely with studies that have been done on protein structure.13 Five or six amino acids may not sound like very much at first, since proteins are often made of hundreds of amino acids. But five or six amino acid substitutions means that reaching the goal requires five or six coherent mutational steps—just to get two proteins to bind to each other. As we saw in the last chapter, even one missing step makes the job much much tougher for Darwin than when steps are continuous. If multiple steps are missing, the job becomes exponentially more difficult.
Let’s consider one further wrinkle. Most amino acid changes in proteins diminish a protein’s function. But about one-third of possible amino acid changes are like switching a k for a c in “cat” or “candy”; they can be accommodated without too much trouble.14 Such “neutral” changes can occur during evolution and spread around a population by chance. So let’s suppose that of the five or six changes that have to happen to a protein to make a new binding site, a third of them are neutral. They could occur before the other key mutations, as a separate step, without harm. Although finding the right neutral changes would itself be an improbable step, we’ll again err on the conservative side and discount the average number of neutral mutations from the average number of total necessary changes. That leaves three or four amino acid changes that might cause trouble if they occur singly. For the Darwinian step in question, they must occur together. Three or four simultaneous amino acid mutations is like skipping two or three steps on an evolutionary staircase.
Although two or three missing steps doesn’t sound like much, that’s one or two more Darwinian jumps than were required to get chloroquine resistance in malaria. In Chapter 3 I dubbed that level a “CCC,” a “chloroquine-complexity cluster,” and showed that its odds were 1 in 1020 births. In other words (keeping in mind the roughness of the calculation):
Generating a single new cellular protein-protein binding site is of the same order of difficulty or worse than the development of chloroquine resistance in the malarial parasite.
Now suppose that, in order to acquire some new, useful property, not just one but two new protein-binding sites had to develop. A CCC requires, on average, 1020, a hundred billlion billion, organisms—more than the number of mammals that has ever existed on earth. So if other things were equal, the likelihood of getting two new binding sites would be what we called in Chapter 3 a “double CCC”—the square of a CCC, or one in ten to the fortieth power. Since that’s more cells than likely have ever existed on earth, such an event would not be expected to have happened by Darwinian processes in the history of the world. Admittedly, statistics are all about averages, so some freak event like this might happen—it’s not ruled out by force of logic. But it is not biologically reasonable to expect it, or less likely events that occurred in the common descent of life on earth. In short, complexes of just three or more different proteins are beyond the edge of evolution. They are lost in shape space.
And the great majority of proteins in the cell work in complexes of six or more. Far beyond that edge.
TOUCHING BASE
In science as in other areas of life, it’s easy to fool yourself if you aren’t careful. A lot of ideas seem plausible at first blush, but when you check them against the facts they don’t work out. Reasoning in the abstract about shape space and what that implies for Darwinian evolution is a good first step, but how does it square with the data? In the next few sections we’ll survey what we know from the best sources of evolutionary data available.
In its battle with malaria the human genome has been terribly scarred. In the past ten thousand years a number of genes have been broken or their efficiency reduced in order to fend off malari
a (as discussed in Chapter 2). Other than sickle hemoglobin (an exception we’ll discuss in the next chapter), has the war with malaria caused humanity to evolve any new cellular protein-protein interactions? No. A survey of all known human evolutionary responses to the parasite includes no novel protein interactions. Although it can’t be ruled out that some such thing has developed but escaped detection, we can be certain that its effects are (or were) weaker than those of the sickle mutation, thalassemia, and the other simple fractured genes, because they did not prevail over time.
Since malaria first appeared in its most virulent form about a hundred centuries ago, more than a billion humans have been born in infested regions. So, although it’s risky to draw too firm a conclusion from just one example, it appears that the likelihood of the development of a new, useful, specific protein-protein interaction is less than one in a billion organisms.
Conversely, in its battle with poison-wielding humans, the malaria genome has also been terribly scarred. In the past half century a number of genes have been broken or altered to fend off drugs such as chloroquine. As discussed in Chapter 3, none of the changes seem to be improvements in an absolute sense. They disappear once drug therapy is discontinued. Has the war with humanity caused malaria to evolve any new cellular protein-protein interactions? No. A survey of all known malarial evolutionary responses to human drugs includes no novel protein-protein interactions. Although, as above, it can’t be ruled out that some such thing developed in the past, no such change persisted, so none could have been as effective as the damaging changes discussed earlier.
Since widespread drug treatments first appeared about fifty years ago, more than 1020, a hundred billion billion, malarial cells have been born in infested regions. It thus appears that the likelihood of the development of a new, useful, specific protein-protein interaction is less than one in 1020. Since sickle hemoglobin, thalassemia, and other human genetic responses have appeared, probably another thousandfold P. falciparum, 1023, have infected humans, with no known protein-protein interactions, or any other effective response, having developed. So it seems that the odds of the development of a new, useful, specific, protein-protein interaction are less than one in 1023—worse than a CCC.
AIDS AND EVOLUTION
Studies of malaria provide our best data about what Darwinian evolution can do, but there are other studies of interest. One excellent source of information comes from the study of the human immunodeficiency virus HIV, the virus that causes AIDS. Like malaria, HIV is a well-studied scourge and killer that first appeared in Africa. Unlike malaria, HIV is spread by person-to-person contact, so it can survive in mild and even cold climates. Also unlike P. falciparum—which is a eukaryote, the most complex type of cell—HIV is a virus, one of the simplest forms of life. The amount of genetic information in the AIDS virus is less than a thousandth the amount of DNA in the malarial parasite. What’s more, viruses such as HIV mutate much more readily than cells do—about ten thousand times faster. The HIV virus is so small, and the mutation rate is so great, that on average each new copy of the virus contains one change, one mutation, from its parent. HIV mutates at the evolutionary speed limit—Darwinian evolution just can’t go any faster.
FIGURE 7.3
Schematic diagram of the genome of HIV. The black bar represents the intact genome. The gray bars show the approximate location of the nine viral genes in the genome. (Gray bars connected by a dashed line represent genes that are pieced together.) The virus is about one-millionth the size of the human genome. Its basic genetics have changed very little in the past decades, despite an enormous mutation rate and the production of a hundred billion billion copies.
About a hundred billion billion, 1020, malarial cells are born each year. The best current estimate is that a person infected with HIV is burdened with a total of one to ten billion (109 to 1010) virus particles.15 The generation time for virus replication is about a day or two16, so over the course of ten years a single person will produce more than a thousand generations of HIV, or up to 1013 viruses. Since there are approximately fifty million people worldwide infected with the virus, the math points to a total of about 1020 copies of the virus having been produced in the past several decades, when HIV became widespread in human populations—roughly the same as the number of malarial cells produced each year.
But the total number of copies of the virus is only part of the story. The other important factor is the speeded-up evolution of HIV due to its much greater mutation rate. Because of the difference in mutation rates HIV has actually experienced about ten thousand times as many mutations as would a comparable number of malarial cells. The very many copies of HIV in the world would be expected to contain almost every imaginable kind of mutation. As one study put it, “Each and every possible single-point mutation occurs between 104 and 105 times per day in an HIV-infected individual.”17
Every double point mutation, where two amino acids are changed simultaneously, would occur in each person once each day. (This means a chloroquine-type resistance mutation—where two particular amino acids had to appear before there was a net beneficial effect—would occur in each AIDS patient every day. Now that’s mutational firepower!) In fact, just about every possible combination of up to six point mutations would be expected to have occurred in an HIV particle somewhere in the world in the past several decades—double the number that could occur in the slower-mutating P. falciparum. In addition to all those point mutations, enormous numbers of insertions, deletions, duplications, and other sorts of mutations would occur as well.
And exactly what has all that evolution of HIV wrought? Very little. Although news stories rightly emphasize the ability of HIV to quickly develop drug resistance, and although massive publicity makes HIV seem to the public to be an evolutionary powerhouse, on a functional biochemical level the virus has been a complete stick-in-the-mud. Over the years its DNA sequence has certainly changed. HIV has killed millions of people, fended off the human immune system, and become resistant to whatever drug humanity could throw at it. Yet through all that, there have been no significant basic biochemical changes in the virus at all.
With a few apparent exceptions,18 HIV enters its target cells of the immune system by first binding tightly and specifically to one of the many kinds of proteins on their surface, and then reaching over to bind another protein called a coreceptor. (Some humans are resistant to HIV because they burn the bridge that the virus uses to invade the cell: They have a broken copy of the gene for a coreceptor.) A hundred billion billion mutant viruses later, HIV continues to do exactly the same thing, to bind the same way. If a mutant virus developed the ability to enter other kinds of cells by binding to other kinds of proteins, it might replicate more effectively and thus outcompete its siblings. That hasn’t happened.19 Neither has much else happened at a molecular level.20 No new gizmos or basic machinery. There have been no reports of new viral protein-protein interactions developing in an infected cell due to mutations in HIV proteins.21 No gene duplication has occurred leading to a new function. None of the fancy tricks that routinely figure in Darwinian speculations has apparently been of much use to HIV.
But what about its ability to quickly evolve drug resistance and evade the immune system? Doesn’t that show that Darwinian evolution is very powerful? Isn’t that a sophisticated maneuver? No. It turns out that HIV employs the same modest tricks that malaria uses to evade drugs—mostly simple point mutations to decrease the binding of the poison to its pathogen target. For example, a change of just one amino acid at position 184 of one particular HIV enzyme causes a little bump that interferes with one drug.22 Another major drug target is a protein called HIV protease, which is a kind of special scissors needed to cut out some other viral proteins from their immature form. Typically, drugs are made that can stick to the protease and gum it up. And just as typically, point mutations appear that alter the protein shape a bit, so the poison doesn’t stick so well.23 Like the development of resistance to rat poison by ra
ts, resistance of HIV to drugs is a very simple biochemical affair.24
Let’s compare the results of HIV evolution to malaria evolution, and consider the changes both have wrought in humans. The number of copies of both in the last fifty years are roughly comparable—roughly, in the sense that malaria outnumbers HIV by a factor of only ten or a hundred. The number of genes in malaria is in the thousands; HIV has just nine. The mutation rate of HIV is greater than that of malaria by a factor of ten thousand. So the small HIV genome has been riddled by changes to a limited number of genes; malaria has endured a roughly comparable number of mutations, but spread out over a much larger genome. Nonetheless, despite the many differences between them, the evolutionary changes in both in the past fifty years are comparable and—despite their severe consequences for public health—biochemically trivial. A few point mutations, the occasional gene duplication in malaria; but no new, useful protein-protein interactions, no new molecular machines. The biochemical changes they have triggered in humans are comparable as well: a long list of broken genes for the ancient P. falciparum; a shorter list for the recent HIV.
The bottom line: Despite huge population numbers and intense selective pressure, microbes as disparate as malaria and HIV yield similar, minor, evolutionary responses. Darwinists have loudly celebrated studies of finch beaks, showing modest changes in the shapes and sizes of beaks over time, as the finches’ food supplies changed. But here we have genetic studies over thousands upon thousands of generations, of trillions upon trillions of organisms, and little of biochemical significance to show for it.
LAB STAR
The studies of malaria and HIV provide by far the best direct evidence of what evolution can do. The reason is simple: numbers. The greater the number of organisms, the greater the chance that a lucky mutation will come along, to be grabbed by natural selection. But other results with other organisms can help us find the edge of evolution, especially laboratory results where evolutionary changes can be followed closely. The largest, most ambitious, controlled laboratory evolutionary study was begun more than a decade ago in the laboratory of Professor Richard Lenski at Michigan State University. Lenski wanted to follow evolution in real time. He started a project to watch the unfolding of cultures of the common gut bacterium Escherichia coli. E. coli is a favorite laboratory organism that has been studied by many scientists for more than a century. The bug is easy to grow and has a very short generation span of as little as twenty minutes under favorable conditions. Like those of P. falciparum, H. sapiens, and HIV, the entire genome of E. coli has been sequenced.
The Edge of Evolution Page 15