A flagellum can be conceptually broken down into three subsystems: the base (which contains the motor), the “hook” (which acts as a universal joint), and the filament (which is the propeller). Within each subsystem, however, are multiple precision-made parts. The base contains the motor that drives the rotation of the flagellum. It also contains protein parts that act as the stator (to clamp the structure firmly in place), as well as bushings and a protein pump that, as we’ll see below, is critical to the assembly of the flagellum. The structure of the base is made of several rings, one of which (the MS ring) is in the cell membrane, the next of which (the P ring) is in the cell wall, and the next of which (the L ring) is in the outer membrane. Each of the three rings is made up of about twenty-six copies of its particular protein component.
FIGURE C.1
The bacterial flagellum. Protein components of the system are labeled in detail. (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.)
As shown in Figure C.1, through the rings is placed a rod, which acts as the drive shaft for the flagellum, transmitting the rotation of the motor to the filament-propeller. The rod contains several different kinds of proteins. The three proteins that compose the part of the rod closest to the cell are present in six copies each, and the protein that makes up the farther part of the rod is present in about twenty-six copies. The proteins of the interior ring have cylindrical symmetry, like balls arranged on a hula hoop, while the rod has helical symmetry, like the thread of a wood screw. Since the two symmetries are mismatched, there is another protein part—present in nine copies—that seems to act as an adaptor between them, reconciling the discordant symmetries. Also in the base is the protein that acts as the motor, as well as three kinds of proteins that act as molecular switches, which allow the motor to change from spinning in a clockwise rotation to spinning in a counterclockwise one.
The hook is the region that connects the base to the propeller. It consists of 120 copies of another type of protein. When it is being assembled, the length of the hook has to be tightly controlled so it isn’t too short or too long. The measurement of the hook length seems to be the job of another protein part. How it measures is not yet clear. After the hook comes the propeller. But it turns out that the mechanical properties needed by something that acts as a universal joint (like the hook does) are not the same as the mechanical properties needed for a propeller. So between the hook and the propeller in the flagellum is a very small but critical region called the junction zone, where several other protein parts (present in copies of a baker’s dozen apiece) act as adaptors to fit the two disparate pieces together. In other words, “It seems very likely that the junction zone acts as a buffering structure connecting two filamentous structures with distinct mechanical characteristics.”3
The propeller itself is made of tens of thousands of copies of flagellin, a sophisticated protein that can switch between several different shapes. The different shapes then give the elongated propeller a different curl, with varying swimming properties. Although the word “flagellum” comes from the Latin for “whip,” the propeller turns out not to be a solid structure like a bullwhip. Instead, it’s hollow like a drinking straw. This feature is critical for the assembly of the flagellum, as we’ll now see.
BUILDING THE OUTBOARD MOTOR
In just the past ten years or so, through the hard work of many scientists in many labs in many countries, details of how a flagellum is built in a bacterial cell have been pieced together. Although many aspects remain hazy, enough is now clear to give a fascinating overview of the elegance and complexity of the assembly process. An animation of the construction of a flagellum has been produced by the “Protonic NanoMachine Project” of the Japan Science and Technology Corporation in a remarkable video, which can be viewed on the Web.4
Like the cilium and the tower at Iacocca Hall, the flagellum is built from the bottom up. The first component to be laid down is the basement—the protein ring in the inner cell membrane (the MS ring). Then, using that structure as a foundation, a sort of housing unit is built on the inside of the cell (called the C ring). Inside the housing is then assembled a machine, called a Type III export apparatus. The export machinery is like a gun that grabs the correct proteins (which are suitably labeled so the automated machinery can distinguish them from proteins that are not part of the flagellum) and pushes them out to the end of the growing structure. The first proteins to be pushed through are those that make up the rod, along with a special protein that can chew through the cell wall. That is needed so the flagellum can grow beyond the stiff boundary of the cell.
The next stage is the assembly of the other rings, L and P. The proteins that make up these structures don’t come through the regular way, however; they are pushed out of the cell by a different set of machinery that is used for the secretion of a variety of other proteins. The protein that makes up the P ring can’t get to the incipient flagellum by itself—it needs another protein called a chaperone to shepherd it over to the construction site; otherwise, the protein loses its way and never arrives. After escorting the P ring protein to its proper destination, the chaperone floats away.
Once the rod is finished, another protein is pushed through the middle of the growing structure to start the hook. The protein isn’t one that will be part of the final structure, however. Rather, it’s called the “hook cap” protein; it helps keep the actual building components in place as the flagellum grows. After the hook is assembled, the hook cap falls off and floats away. The proteins that make up the junction zone are then grabbed by the export machinery and sent through the export channel to the end of the nascent flagellum.
Finally, we’re just about ready to start the business end of the flagellum, the propeller that actually pushes the bacterium forward. But before we do, there’s another critical step. Just as the construction of the hook region needed a “cap” at the end, so does the propeller. But it’s not the hook cap; it’s a different cap. So before the protein pieces that make up the propeller are sent through the export machinery, a “filament cap” precedes them. The cap fits on the end of the hollow flagellum, and as each of the tens of thousands of copies of the propeller protein are pushed down the center to the end, the cap prevents them from spewing out into the surrounding liquid and being lost. In order to traverse the rather thin, hollow central channel of the flagellum, the flagellar proteins have to be kept in an extended shape. When they arrive at the far end, the cap also helps all the copies of the propeller protein to fold into the correct, compact shape—the shape needed to form the propeller.
THE BALLERINA
While describing the structure of the flagellum in this section I’ve written rather blandly of “protein parts” for this and that, as if the individual proteins were like so many simple nuts and bolts. That is not the case at all. Like hemoglobin, all of the dozens of proteins involved in building the flagellum are themselves quite intricate and wonderfully suited to their jobs. To drive home the point, for illustration let’s look at just one example—the filament cap.
The filament cap is made up of five copies of a single protein whose official name is “FliD” but I’ll call it “Twinkletoes.” For comparison, remember that hemoglobin has four parts—two alpha and two beta chains. When stuck together, the five protein parts give Twinkletoes a shape that might best be described as a starfish on stilts. The leglike stilts point vertically down from the horizontal pentagonal starfish. Now, the hollow filament of the flagellum is made of multiple copies of flagellin protein arranged in eleven strands, so the fivefold symmetrical cap is slightly mismatched to the ends of the filament. One leg of the cap can fit in a crease between every other strand, but two times five is ten, not eleven, so one crease does not have a cap leg stuck in it.
B
ut the mismatch is not some mistake; it’s part of the elegant design of the assembly system. As a copy of flagellin protein is pushed down the hollow tube to be added to the growing end of the filament, it is prevented from floating out into space by the filament cap. The cap allows the flagellin time to fold to its functional shape, and then directs it to fill the empty space on the growing filament. So the “mismatch” actually directs the protein to the correct, available position. As the flagellin fills the proper vacant position, the pentagonal cap rotates, so that the next available slot is now in position to be filled. To do this, Twinkletoes lifts one of its legs and moves it over a notch. The next copy of flagellin then comes down the follow tube of the filament and is directed to the right spot, Twinkletoes rotates again to the next space, and the next leg swings over. Tens of thousands of times the dancing machinery5 automatically directs the right building blocks to the right positions, lifts its supple legs, and spins to the next position.
ANOTHER MATTER
How do Darwinists explain the flagellum? In the same way as they explain the cilium—usually by a tactful silence, occasionally by Just-So stories. There is currently a lively discussion going on in the professional science literature about the flagellum and another structure called a “Type III secretory system” (TTSS), which contains a number of protein parts that resemble those of the flagellum. The TTSS is used by bacteria as a protein pump; since parts of the flagellum also act as a pump in order to build the flagellum, some workers reasonably think that the two are related by common descent. Whether the TTSS or the flagellum came first is the point of controversy.6 But none of the papers seriously addresses how either structure could be assembled by random mutation and natural selection, or even how one structure could be derived from the other by Darwinian processes.7 Consider a review of flagellar assembly written by the eminent Yale biologist Robert Macnab shortly before his premature death in 2003. The article of course shows great erudition, and it nicely summarizes the startlingly complex pathway of flagellum assembly.
How did such a pathway evolve by random mutation? In the approximately seven-thousand-word review, the phrase “natural selection” does not appear. The word “evolution” or any of its derivatives occurs just once, in the very last sentence of the article. Speaking of the flagellum and the TTSS, Macnab writes: “Clearly, nature has found two good uses for this sophisticated type of apparatus. How [the TTSS and the flagellum] evolved is another matter, although it has been proposed that the flagellum is the more ancient device, since it exists in bacterial genera that diverged long before eukaryotic hosts existed as virulence targets.”
Darwinism has little more of substance to say.
Appendix D
The Cardsharp
STACKING THE DECK
One intriguing possibility for Darwinian construction of cellular machines that has been much discussed in the scientific literature recently is the shuffling around of binding sites, to bring different proteins close to one another.1 To illustrate, suppose there were two large pegboards on the wall of a carpentry shop, with chalk outlines drawn of which tools were supposed to be hung on which pegs, with a different set of tools on each of the two different pegboards. If we cut the two pegboards down the middle and switched two halves of the two boards, we’d have different tools next to each other than we had before, without having to draw a new outline of a tool in a new position.
Something like that is thought to explain some features of cells of higher organisms (eukaryotic cells). Some proteins resemble several proteins that have been stitched together. Such proteins have discrete regions called “domains”2 that can each fold up into compact shapes, the way myoglobin does. The domains are often connected by short, thin lengths of the amino acid sequence of these multidomain proteins. The thin lengths look like they do little more than just tie the domains together. In some proteins, several or all of the domains have binding sites for other proteins, with a different kind of protein binding to each domain. The apparent purpose of these particular multidomain proteins is just to bring the other proteins close together (Figure D.1).
FIGURE D.1
Cartoon illustrating “domain swapping” of proteins. A) Two proteins, each consisting of two linked domains. Each domain has a binding site for a separate, different protein, indicated by the differently shaped depressions. B) Mutational processes rearrange the genes for the proteins, generating novel combinations of binding sites.
What do such so-called “scaffold” proteins do? Fascinatingly, many seem to act as little computer circuits, signaling a cell to make appropriate “decisions” in a changing world. The cell is an extremely complex system that has to respond suitably to a variety of circumstances. It has to grow at the right time, defend itself when necessary, search for food, even self-destruct sometimes for the greater good. To be able to do all of this, the cellular nanobot has to collect information about the environment, weigh it, and then use the information to take effective action. As one group of scientists notes, “Cells require a remarkable array of sophisticated signal processing behaviors that rivals or surpasses that of modern computers.”3 So one scaffold protein might have binding sites for proteins that indicate some condition (like, say, it’s time to grow now) as well as binding sites for proteins that will then take the appropriate action (like, say, sending a definitive signal to the nucleus to start replicating). Another scaffold protein might have binding sites for proteins that tell a cell to kill itself (perhaps sent by immune cells that “perceive” the doomed cell has been invaded by a virus) as well as proteins that begin the autodestruct sequence.
Scaffold proteins have been likened to parts of computer programs4 called “AND” gates or “OR” gates.5 It’s common for a human programmer to write some computer code that in English says “IF (one condition is true) AND (another condition is true) THEN (execute action number one).” A scaffold protein that conveys a certain signal only IF one certain protein AND another particular protein are bound to it is acting like that computer statement. A programmer might also write “IF (one condition is true) OR (another condition is true) THEN (execute action number two).” A scaffold protein that conveys a signal IF either one certain protein OR another particular protein is bound is acting like that computer statement. As one might imagine,6 more complex computer or protein circuits could easily be generated.
Suppose, however, that the two computer statements got mixed up. Suppose that through some glitch we got the statement “IF (one condition is true) AND (another condition is true) THEN (execute action number two).” If the statements somehow got mixed up, the input conditions for the first AND statement would be linked to the output condition for the second OR statement. If something like that happened for two scaffold proteins, a new biological circuit might be made without having to produce any new protein-binding sites.
Exactly that scenario has been modeled by the group of Wendell Lim, a biologist at the University of California at San Francisco.7 Using clever laboratory techniques, in one experiment Lim and coworkers spliced a yeast scaffold protein that normally binds a protein that allows the yeast to mate with a second protein that receives a signal that tells the yeast to brace itself against extra-salty water. As hoped, the result was a new signaling circuit—yeast that had the hybrid protein could only survive in concentrated salt solutions in the laboratory if they were exposed to the mating signal protein. In another set of experiments8 Lim’s group constructed artificial proteins using multiple different domains. One of the domains regulated the formation of actin fibers; the other domains bound various other proteins. Under some conditions in the test tube the artificial scaffold proteins either didn’t work at all or were turned on all the time. But in other conditions some proteins could act as either an AND circuit or an OR circuit, just as the scientists planned.
Lim thinks such results will help us both to engineer cells and to understand evolution:
These findings demonstrate that scaffolds are highly flexible
organizing factors that can facilitate pathway evolution and engineering…. [P]rimitive tethering scaffolds generated by recombination or fusion events could in principle [emphasis added] be sufficient to generate new pathways…. [T]hese organizing structures thus appear to be optimized for evolvability, a property that may provide increased fitness in the face of constantly changing environmental challenges and signaling needs…. [S]caffold engineering may allow for systematic manipulation of cytoplasmic signaling pathways.
Although the results do show great promise for the productive engineering of cells by intelligent agents, I do not believe they indicate that an incoherent process could build new, complex, helpful genetic circuits by randomly rearranging old parts. The simple point that even superb scientists like Lim—who assume a Darwinian framework—do not seem to grasp is that the purposeful arrangement of parts (including by scientists in laboratories) is the hallmark of intelligence. It does not mimic random mutation. It is the exact opposite of random mutation.
Lim of course doesn’t claim his work is an actual example of evolution in action, but he does view it as a sort of proof of principle that such a phenomenon is theoretically possible. So it’s worth recalling the key insight of evolutionary biologists Jerry Coyne and Allen Orr that “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.” What do the lab results tell us about whether random-yet-productive shuffling of domains “occurs with significant frequency under conditions that are likely to occur in nature”? About whether that is biologically reasonable? Nothing at all. When a scientist intentionally arranges fragments of genes in order to maximize the chances of their interacting productively, he has left Darwin far, far behind. You don’t learn much about the fair odds of winning at poker by watching a cardsharp deal himself a royal flush, and you don’t learn much about random mutation by arranging genes in the lab on purpose.
The Edge of Evolution Page 29