The Edge of Evolution

by Michael J Behe

  IFT exponentially increases the difficulty of explaining the irreducibly complex cilium. It is clear from careful experimental work with all ciliated cells that have been examined, from alga to mice, that a functioning cilium requires a working IFT.12 The problem of the origin of the cilium is now intimately connected to the problem of the origin of IFT. Before its discovery we could be forgiven for overlooking the problem of how a cilium was built. Biologists could vaguely wave off the problem, knowing that some proteins fold by themselves and associate in the cell without help. Just as a century ago Haeckel thought it would be easy for life to originate, a few decades ago one could have been excused for thinking it was probably easy to put a cilium together; the pieces could probably just glom together on their own. But now that the elegant complexity of IFT has been uncovered, we can ignore the question no longer.

  How do Darwinists explain the cilium/IFT? In 1996 in Darwin’s Black Box I surveyed the scientific journals and showed that very few attempts had been made to explain how a cilium might have evolved in a Darwinian fashion—there were only a few attempts. Although Brown University biologist Kenneth Miller argued in response that the two-hundred component cilium is not really irreducibly complex, he offered no Darwinian explanation for the step-by-step origin of the cilium. Miller’s professional field, however, is the study of the structure and function of biological membranes, and his rejoinder appeared in a trade book, not in the scientific literature. An updated search of the science journals, where experts in the field publish their work, again shows no serious progress on a Darwinian explanation for the ultracomplex cilium.13 Despite the amazing advance of molecular biology as a whole, despite the sequencing of hundreds of entire genomes and other leaps in knowledge, despite the provocation of Darwin’s Black Box itself, in the more than ten years since I pointed it out the situation concerning missing Darwinian explanations for the evolution of the cilium is utterly unchanged.14

  On the origin of the cilium/IFT by random mutation, Darwinian theory has little that is serious to say. It is reasonable to conclude, then, that Darwinian theory is a poor framework for understanding the origin of the cilium.

  The cilium is no fluke. The cell is full of structures whose complexity is substantially greater than we knew just ten years ago. (In Appendix C, I discuss intricacies of the bacterial flagellum and its construction, for readers who enjoy plenty of details.) The critical question is, of course, Can mutation of DNA explain this? Or rather, can random mutation explain it? Life descended from a common ancestor, so DNA did mutate—change from species to species. But what drove the crucial changes?

  Repeating Darwin’s own mistakes, modern Darwinists point to evidence of common descent and erroneously assume it to be evidence of the power of random mutation.15 Yet if modern malaria can’t deal with the single amino acid change of sickle hemoglobin, why should we think that the IFT system would be supplied by random mutation in some ancient cell? If the human genome is substantially harmed by its trench warfare with P. falciparum, why do we think competition would build an elegant molecular outboard motor? To ask such questions is to answer them. There is no evidence that Darwinian processes can make anything of the elegance and complexity of cilia.

  TIMING IS EVERYTHING

  If the cilium is likened to the tower of Iacocca Hall, then IFT can be compared to the bulldozers, cranes, and other machinery needed to construct it. But that’s not all that’s needed for bottom up–top down construction. To appreciate the massive challenge that cellular systems present to random mutation, we have to consider more than just physical features, more than the final structures themselves and the construction machinery needed to build them. We also have to consider the molecular planning that goes into the project. Genetic control of planning is in some ways the most difficult aspect of a molecular construction process for scientists to investigate, but is no less critical than the physical parts that make up the final structure.

  A large construction project has to be conducted in an orderly manner. Orderly construction isn’t needed because of some aesthetic obsession with neatness; it’s needed because if there are too many machines and other items on the construction site they can interfere with each other. If all the items needed for a finished office building were present on site from the start, they would get in each other’s way; some would be damaged, machinery might be clogged. If office furniture were scattered over the construction site at the start, at the same time when steam shovels first arrived to dig the foundation, the furniture would likely get scooped up in a shovel or crushed under a tractor tread. The end result would be a mess.

  Physical construction in the cell is almost exclusively the job of proteins. Proteins constitute the molecular bulldozers, steam shovels, train engines, train cars, railroad tracks, and all the other tools, both large and small, needed for construction projects. Of course, the genes that code for the proteins are composed of DNA, so ultimately all the information needed to make all the material required for construction—both the construction machinery and the materials that make up the office tower itself—resides in DNA. In addition to those genes, however, the DNA of a cell also has regions that act as control signals. The control signals of DNA, in conjunction with control proteins, orchestrate the project, to make sure that the proper machinery is made at the proper time in the proper amounts.

  Elucidating how the cell functions is very difficult work, and much remains unknown. Although aspects of IFT have been unveiled in the past decade, the control program for making a cilium is still largely a mystery. However, in that same time remarkable progress has been made in outlining the control program for another large structure, the bacterial flagellum (see Appendix C). Briefly, the bacterial flagellum is an outboard motor that bacteria use to swim. In order to illustrate the planning that molecular construction must involve, over the next few paragraphs I’ll describe what has recently been learned about the control of flagellum construction. (Some readers may wish to skip to the next section.)

  Just as the outboard motor of a motorboat in our everyday world consists of a large number of parts (propeller, spark plugs, and so on), so does the molecular outboard motor. The flagellum has dozens of protein parts that do the particular jobs necessary for the complex system to work. Those dozens of proteins are coded by dozens of genes in a bacterial cell. The genes are grouped into fourteen bunches called “operons.” Next to each operon in the DNA are control signals. The control signals themselves fall into three categories we’ll call class 1, class 2, and class 3. The genes for proteins that have to be made first in the construction process have class 1 control signals, those genes that go second have class 2 signals, and so on.

  Most of the time, a bacterial cell isn’t building a flagellum, because it already has one. However, after cell division a new cell has to start the construction program. To begin, the DNA control regions for class 1 genes mechanically “sense” that the time has come and switch on class 1 genes. There is just one operon in class 1, which contains just two genes. The genes code for two protein chains, which, like the alpha and beta chains of hemoglobin, stick to each other to make a single functioning protein complex. That protein is neither a part of the flagellum nor a part of the construction machinery. Rather, it’s akin to the foreman of a project, who has to tell the other workers what to do. Let’s call it the “boss” protein.

  The boss protein binds specifically to the DNA control regions of the seven class 2 operons, mechanically turning them on. Class 2 genes code for the proteins that make up the foundation of the flagellum (plus some helper proteins), just as you’d expect in bottom-up construction. One class 2 gene, however, isn’t part of the foundation. It’s another control protein. Let’s call it the “subboss” protein. The subboss protein binds to the DNA control region of class 3 genes, which comprise proteins that make the outer parts of the flagellum. So each class of genes contains the gene for a protein that will turn on the next class.

  But that’s not
all. Clever as that part is, the control system is much more finely tuned than just the cascading control proteins. For years researchers knew that if the genes for any of a score of protein parts in class 2—the ones that made up the foundation of the flagellum—were experimentally broken in the lab, the genes for the outer parts of the flagellum would remain switched off. But how could so many genes all control later construction?

  Class 3 contains a gene for a protein that binds tightly to the subboss protein, inactivating it. Let’s call that the “checkpoint” protein. Why turn on the subboss only to immediately inactivate it with the checkpoint protein? Later in the construction project, a clever maneuver gets rid of the checkpoint protein. The flagellum not only is an elegant outboard motor, but also contains a complex pump in its foundation, which actively extrudes class 3 protein parts to form the outer portion of the structure.

  Here’s the elegant trick. When the pump in the foundation of the flagellum is completed and running, one of the first proteins to be extruded is the checkpoint protein. Getting rid of the checkpoint protein releases the subboss protein to bind to the control regions of class 3 operons, switching on the genes for the outer portion of the flagellum. So the completion of the first part of the flagellum is directly linked to the switching on of the genes to make the final parts of the flagellum.

  “MIND-BOGGLING COMPLEXITY”

  In just the past few years a group of Israeli scientists has developed clever new laboratory techniques to analyze in even finer detail the control exerted by DNA control elements on the construction of the flagellum. By successively joining the control elements to the gene for a protein that can be detected by its fluorescence, the scientists showed that, even within classes 2 and 3, the control elements switch the genes on in the order that they are needed for construction. Within class 2, the genes needed for the bottom of the foundation are switched on before the genes for the top of the foundation, and within class 3, genes for the bottom of the top are activated before genes for the top of the top.16

  FIGURE 5.3

  Genes for the construction of the bacterial flagellum are activated in a precisely timed fashion. Those needed for construction of the bottom of the molecular machine are switched on first, followed in order by those needed for more distant parts. (Illustration of the flagellum reprinted courtesy of the Kyoto Encyclopedia of Genes and Genomes, Kanehisa, M., Goto, S., Hattori, M., Aoki-Kinoshita, K. F., Itoh, M., Kawashima, S., Katayama, T., Araki, M., and Hirakawa, M. 2006. From genomics to chemical genomics: new developments in KEGG. Nucleic Acids Res. 34:D354–57.)

  The same group of scientists has examined DNA control elements for other cellular systems and discovered similar elegance there. When they studied cellular biochemical pathways for making amino acids, they discovered what is called “just-in-time” organization, where a protein is made as close to the time it’s needed as possible:

  Mathematical analysis suggests that this “just-in-time” transcription program is optimal under constraints of rapidly reaching a production goal with minimal total enzyme production. Our findings suggest that metabolic regulation networks are designed to generate precision promoter timing and activity programs that can be understood using the engineering principles of production pipelines.17

  What does all this jargon mean? Simply put, the more closely we examine the cell, the more elegant and sophisticated we discover it to be. Complex, functional structures such as the cilium and flagellum are just the beginning. They demand intricate construction machinery and control programs to build them. Without those support systems, the final structures wouldn’t be possible. The bacterial flagellum contains several dozen protein parts. The cilium, which so far has resisted investigation of its DNA control program, has several hundred. There is every reason to think that the control of its construction will have to be much more intricate than that of the flagellum.

  Control of construction projects and other activities in the cell is difficult for scientists to investigate, because “control” is not a physical object like a particular molecule that can be isolated in a test tube. It’s a matter of timing and arrangement. The upshot is that even now in the twenty-first century—more than fifty years after the double helical shape of DNA was discovered by Watson and Crick, and decades after the first X-ray crystal structures of proteins were elucidated—science is still discovering fundamental new mechanisms by which the operation of the cell is controlled.

  Recently—some sixty-five years after George Beadle and Edward Tatum proposed the classic definition of a gene as a region of DNA that codes for an enzyme—an issue of the journal Nature ran a feature with the remarkable title “What Is a Gene?” The gist of the article was that the control systems that affect when, where, and how much of a particular protein is made are becoming so complex, and their distribution in the DNA so widespread, that the very concept of a “gene” as a discrete region of DNA is no longer adequate. Marvels the writer, “The picture these studies paint is one of mind-boggling complexity.”18

  DELIMITING THE EDGE

  Where is it reasonable to draw the edge of evolution? In this chapter and the preceding one I intended to circumscribe that question—show examples of what I think clearly can and what clearly cannot be explained by random mutation and natural selection. Somewhere between those extremes, then, lies the edge.

  On the one side are our very best examples—from humanity’s trench war with parasites—of what random mutation and natural selection are known to do. We know that single changes to single genes can sometimes elicit a significant beneficial effect. The classic example, taught in virtually all biology textbooks, is that of sickle cell hemoglobin, where a change of one amino acid confers resistance to malaria, saving many children from premature deaths. Other examples fit the single-change profile, such as HbC and HbE, warfarin and DDT resistance, and so on. Random mutation also produced a long list of broken genes that can be beneficial in dire circumstances: thalassemia, G6PD deficiency, CCR5 deletion, and so on.

  More rarely, several mutations can sequentially add to each other to improve an organism’s chances of survival. An example is the breaking of the regulatory controls of fetal hemoglobin to help alleviate sickle cell disease. Very, very rarely, several amino acid mutations appear simultaneously to confer a beneficial effect, such as in chloroquine resistance with mutant PfCRT. Changing multiple amino acids of a protein at the same time requires a population size of an enormous number of organisms. In the case of the malarial parasite, those numbers are available. In the case of larger creatures, they aren’t.

  On the other side are the examples of what random mutation and natural selection clearly cannot do. In this chapter I discussed several illustrations—IFT and the control of bacterial flagellum construction—of the kind of astonishingly complex, coherent systems that fill the cell. Those systems aren’t built from just one or two amino acid changes to random proteins of systems doing other jobs—they consist of dozens of different proteins dedicated to their tasks. They didn’t arise by breaking genes; they required the coordinated construction of many new genes. Cilia and flagella are not only stupendously complex systems in their own right, but they have complicated systems dedicated to their construction, and genetic control systems coordinating that construction, whose intricacy science is only now beginning to appreciate.

  The structural elegance of systems such as the cilium, the functional sophistication of the pathways that construct them, and the total lack of serious Darwinian explanations all point insistently to the same conclusion: They are far past the edge of evolution. Such coherent, complex, cellular systems did not arise by random mutation and natural selection, any more than the Hoover Dam was built by the random accumulation of twigs, leaves, and mud.

  6

  BENCHMARKS

  It’s time to consider some general principles. How do we decide if some biological feature is unlikely to have been produced by random mutation and natural selection? Writing of other matter
s in their book Speciation, evolutionary biologists Jerry Coyne and Allen Orr pinpoint the key principle:

  The goal of theory, however, is to determine not just whether a phenomenon is theoretically possible, but whether it is biologically reasonable—that is, whether it occurs with significant frequency under conditions that are likely to occur in nature.1

  In this book we’ll apply the paramount Coyne-Orr principle to Darwinian evolution as a whole (which they do not).2 In light of the recent tremendous progress of science, can we determine not what is merely theoretically possible for Darwinian evolution, not what may happen only in some fanciful Just-So story, but rather what is biologically reasonable to expect of random mutation and natural selection at the molecular level? If we can decide what is biologically reasonable to expect of unguided evolution, then we can also determine what is unreasonable to expect of it.

  Since we’ll be looking at borderline, marginal cases, determining the ragged edge of evolution will necessarily be more tentative than finding clear-cut examples of what certainly can and cannot be done by Darwinian processes. Sickle hemoglobin can inarguably be explained by mutation and selection, the bacterial flagellum cannot. Is the edge of Darwinian evolution closer to sickle hemoglobin, or closer to the flagellum?

  In this chapter I develop two criteria by which to judge whether random mutation hitched to natural selection is a biologically reasonable explanation for any given molecular phenomenon. The criteria, spelled out in more detail over the rest of the chapter, are the following.

  First, steps. The more intermediate evolutionary steps that must be climbed to achieve some biological goal without reaping a net benefit, the more unlikely a Darwinian explanation.