When no oxygen is bound to hemoglobin, the iron atom of each subunit is a little too fat to fit completely comfortably into the hole in the middle of the heme where it resides. However, when an oxygen molecule comes along and binds to it, for chemical reasons the iron shrinks slightly. The modest slimming allows the iron to sink perfectly into the middle of the heme. Remember that “H” that was attached to the iron in myoglobin? (I knew you would!) Well, there also is an “H” attached in hemoglobin. As the iron sinks, it physically pulls along the attached “H.” The “H” itself is part of one of the helical segments of the subunit, so when the “H” moves, it pulls the whole helix along with it. Now, at the interface of the subunits of hemoglobin, where alpha and beta chains contact each other, there are several positively charged amino acids across from negatively charged ones; of course they attract each other. But when the helix is pulled away by the “H” that’s attached to the sinking iron, the oppositely charged groups are pulled away from each other (Figure A.2). What’s more, the shape of the subunits is such that when one moves, they all have to move together. So hemoglobin changes shape into a somewhat distorted pyramid when oxygen binds, and electrostatic interactions between all of the subunits of hemoglobin are broken.
That takes energy. The energy to break all those electrical attractions comes from the avid binding of the oxygen to the iron. But here’s the catch. Just as only one quarter dropped into the slot of a soda machine can’t release the can, the binding of just one oxygen doesn’t provide enough energy to break all those interactions. Instead, several subunits must each bind oxygen almost simultaneously to provide enough power. That only happens efficiently in a high-oxygen environment like the lungs. Conversely, when a hemoglobin that has four oxygen molecules attached to it is transported by the circulating blood from your lungs to the low-oxygen environment of, say, your big toe, when one of the oxygens falls off, the others aren’t strong enough to keep the hemoglobin from snapping back. The electrostatic attractions between subunits reform, which yanks back the helix, which tugs up the “H,” which pushes off the oxygens. As a result, the remaining several oxygens are unceremoniously dumped off, exactly where they are needed.
FIGURE A.2
Schematic drawing of how the binding of a small molecule can cause a protein to switch shape. (Top) Structural elements of a protein, represented by the cylinders and stippled surface, are apart. (Bottom) A small molecule, represented by the hexagon, binds to both surfaces, tilting the cylinders and bringing them closer to the stippled surface. The ability to switch its shape allows hemoglobin to deliver oxygen efficiently.(Drawing by Celeste Behe.)
My point in discussing the intricacies of the relatively simple molecular machine that is hemoglobin is not to tax the reader with details. Rather, the point is to drive home the fact that the machinery of the nanobot works by intricate physical mechanisms. Robots in our everyday, large-scale world (such as, say, robots in automobile factories that help assemble cars) function only if very many exactly shaped and precisely positioned parts—nuts, bolts, levers, wires, screws—are all in place and working. If they are ever built, artificial nanobots will also have to work by excruciatingly detailed physical mechanisms. Biological nanobots must do the same. There is no respite from mechanical complexity except in idle dreams or Just-So stories.
STICKING TOGETHER
Many molecular machines in the cell are much more complex than hemoglobin, but all work in the same mechanistic way. There are proteins that act as automatic gatekeepers, regulating the flow of small molecules or ions into and out of the cell. There are proteins that act as timing devices; others that are molecular trucks to ferry supplies to different parts of the cell; still others that act as cables and winches, pulling on cellular parts that need to be together. One of my favorites is a protein called gyrase, which can literally tie DNA into knots. In terms of our big, everyday world, gyrase is somewhat like a machine that could tie shoelaces. In developing an intuition for how such molecular machines act, a good start is to ask yourself how a shoelace-tying machine might work in our big world, or how a clock might work, or a delivery system, or a regulated gate. As you might suspect, they all would work by mechanical principles, and none of them would be simple.
Yet intuition can be insufficient. There is also a subtle but critical difference between molecular machines and everyday machines that needs to be kept in mind, a dissimilarity that underscores the much greater difficulty of making a molecular machine. One crucial way in which machinery in the nanobot differs from machinery in our everyday experience is that cellular machines have to assemble themselves. There is no conscious agent walking around in the cell putting pieces of machinery together, as there might be in a factory making, say, flashlights. Needless to say, the requirement for self-assembly enormously complicates the task of building a functional nanomachine.
How do cellular nanomachines build themselves? Here’s a very simplified description. A protein binds to its correct partner(s) in the cell by having an area(s) on its surface that is closely complementary in shape and chemical properties with the other member(s) of the team. Let’s think how that might work. Consider a protein with a positively charged amino acid on its surface. Of course the positive charge might attract a negative charge on the surface of another protein. However, there are thousands of different kinds of proteins in the cell and almost all have many negative charges. The interaction of just one positive and negative charge isn’t enough to allow a protein to distinguish its partner from the many other proteins in the cell. So suppose that, next to the positive charge, the protein had an oily amino acid. Then it could match other proteins that had an oily patch next to a negative charge. Yet there will still be a lot of proteins in the cell with those two simple features, so even more specificity is needed. Further suppose next to the positive charge and oily patch there was a large amino acid sticking out from the surface. Then it could match a protein that had a negative charge, oily patch, and indentation in its surface. That combination of features decreases the number of potential partner proteins that it would match even further (see Figure 7.1).
In the last paragraph we worried about getting enough distinguishing features on the surface of a protein to allow it to discriminate between its correct binding partner and the thousands of other proteins in the cell that it should not stick to. But we also have to worry about the strength of the attraction. The reason is that a protein has to search blindly through the cell for its partner. It does so by randomly bumping into many surfaces, like pieces of flotsam and jetsam colliding with each other in a flowing stream, until the protein accidentally hits the complementary surface of its partner and sticks. However, suppose that the attraction between one positive and one negative charge were so overwhelmingly strong that whenever two opposite charges were close to each other they’d glom together, never to separate. If that were the case, the contents of the cell would congeal in an instant, killing it. The lesson is: Individual interactions can’t be too strong. On the other hand, the total interaction strength of two proteins can’t be too weak either, or the protein pieces might not form a stable entity, and might fall apart after a short time. The solution is to have a number of weak interactions between two proteins. Like Velcro fasteners, each individual interaction is rather delicate but the sum is strong. In the cell, multiple weak interactions make for strong binding. In general, the more interactions there are, the more specificity and strength there is to the binding between two proteins.
MAGNETS IN A SWIMMING POOL
As an illustration, imagine that a flashlight had to automatically self-assemble. To make the example closer to what happens in the cell, let’s further imagine that the parts of the flashlight were floating in a big, well-stirred swimming pool, so that they could randomly bump against each other. Also imagine that thousands of other parts were floating in the pool, parts that belonged to other kinds of machinery. On the surface of all of the parts were tiny, rather weak, bar magnets, s
ome with their north pole facing outward and south pole inward (buried inside the part, where it couldn’t interact with other magnets), others with the south pole out and north in. As the water is stirred parts bump against other parts, some stick briefly when one or two magnets are in the right place to touch, but quickly break apart. When two pieces that are part of the future flashlight happen to collide in the correct orientation, they stick. The reason they stick, of course, is that there are multiple magnets (say, five to ten) on their surfaces in just the right positions, with just the right pattern, with exposed north poles arranged to be opposite exposed south poles. As with Velcro, the multiple weak interactions add up to a stable, strong binding. Then a third piece of the flashlight can stick to the growing conglomerate, and a fourth, until the flashlight is assembled. (Notice that the third and fourth pieces can’t have the same pattern of magnets as the first and second pieces, or you wouldn’t get the correct parts in the right order—for example, the battery might be stuck to the outside of the case!)
Let me make a few simple, interrelated points from this illustration. The first point is that of course parts of the flashlight all have to have patterns of magnets that match their binding partners. Put another way, even if all the correct pieces of the flashlight were floating in the pool, if none had magnet patterns to match each other, no flashlight would be made—the parts would occasionally bump, but wouldn’t stick and thus wouldn’t self-assemble into a flashlight. A further point is that the magnet features needed to form a binding pattern for a molecular machine to self-assemble are beyond the requirements for the function of the machine itself. In other words, the pattern of magnets that helps assemble the flashlight doesn’t at all address the other aspects of the parts that allow them to act as a flashlight when assembled. Another point is that the binding patterns on a piece can’t match incorrect parts. If the magnets on a piece of flashlight matched those on a piece of toaster, a mishmash would likely result, and would interfere with the construction of both flashlight and toaster.
A final, more subtle point is especially important for evaluating what Darwinian evolution can and can’t do. Suppose we had a piece of one type of machine that we would also like to use in a different machine. Maybe we had a general-purpose part like a nut or bolt or gear. In our everyday world, of course, we could happily use the same type of nut or bolt in a thousand different machines. For example, a child’s Lego building set can be used to make many different constructs. But when we’re talking about self-assembling machinery, there’s a major-league hitch. If a part has to attach to a partner different from its usual one, then the self-assembly instructions have to change. That is, the pattern of magnets on the surface of the part would have to be changed to match the new target. That might require multiple coherent changes before the part could assemble with the new target. What’s more, if the assembly instructions changed, the part would lose its ability to assemble into the old system. To keep its old role while also gaining a new one, a near-duplicate of the old part would have to be made that had luckily acquired altered assembly instructions. For a process supposedly driven by random mutation, that would be a very tall evolutionary order.
Appendix B
Malaria Drug Resistance
In order to understand malaria’s strengths, let’s briefly look at a few examples of medicines that worked in ways different from chloroquine and are now being brushed aside by new mutations.
One set of treatments that was developed to take the place of chloroquine is abbreviated S/P, which stands for two different drugs, sulfadoxine with pyrimethamine. Both of these drugs target a vital metabolic pathway in malaria that builds components of DNA. It turns out that the four kinds of building blocks in DNA are of two types, called purines and pyrimidines. The parasite can obtain one type, purines, from the host it’s invading, but has to make its own pyrimidines. So if its ability to make pyrimidines can be undercut, the bug is stymied. In order to make pyrimidines, the parasite, like other organisms, needs first to make several forms of a vitamin called folic acid. The two drugs in S/P, which both resemble natural chemicals in the metabolic pathway, block separate steps in the multistep pathway that makes pyrimidines. They do so by binding to the enzymes that normally catalyze the chemical conversions. However, mutations in the enzymes can make the drugs ineffective, probably by stopping them from binding. In the case of pyrimethamine (the “P” in the “S/P”), the drug interferes with an enzyme abbreviated DHFR. However, when a mutation appears in the enzyme and changes the amino acid at position number 108 from serine to asparagine, the drug loses its effectiveness. Similarly, when a mutation in an enzyme abbreviated DHPS changes the alanine normally found at position number 437 to a glycine, sulfadoxine (the “S” in the “S/P”) fails.1
A hopeful note amid the gloom is that, about five years after the use of chloroquine was discontinued and S/P substituted in Malawi in 1993, the malaria there became susceptible to chloroquine once again. Some scientists have speculated that, if we’re lucky, maybe drugs can be rotated; ineffective ones can be shelved for a while in the hope that they’ll regain effectiveness sometime down the road.2
A relatively new drug, atovaquone, which interfered with a different step in P. falciparum metabolism, can be countered by a single amino acid mutation in a protein called cytochrome b.3 The very latest drug, artemisinin, is derived from the Chinese sweet wormwood plant. Resistance to artemisinin has not yet been seen in clinics, but has been reported in laboratory investigations, and will almost certainly develop in the field eventually.4 Nicholas White of Mahidol University in Thailand worries, “If we lose artemisinins to resistance, we may be faced with untreatable malaria.”5 Quinine, the natural drug that first turned the tide of battle toward humanity’s side, is still pretty effective against P. falciparum. But the bug is slowly gaining ground, apparently by many little changes in a number of separate genes (like sickle hemoglobin and HPFH on the human side) rather than in one gene, as for chloroquine resistance.6
Appendix C
Assembling the Bacterial Flagellum
THE OUTBOARD MOTOR
The cilium is an elegant molecular machine that powers the swimming of cells as diverse as sperm and pond algae. As we’ve seen, not only is the cilium itself enormously complex, but IFT—the system that builds the cilium—is also highly sophisticated, intricate, and dynamic. Without the assembly system, no working machinery gets built. The need to spontaneously assemble intricate machinery enormously complicates any putative Darwinian explanation for the foundation of life, which has to select from tiny, random steps the size of the sickle cell mutation. Yet IFT is not some fantastic aberration. In a cellular nanobot, where machines run the show without the help of conscious agents, everything has to be assembled automatically. To drive home the complexity of self-assembly, let’s look at just one more example—the bacterial flagellum.
The flagellum is a cellular propulsion system that is completely different from the cilium. Rather than acting as an oar that goes back and forth like the cilium, the flagellum is a rotary motor—literally an outboard motor that bacteria use to swim. And just like the familiar outboard motor that powers a boat on a lake, the flagellum needs many different parts to work. Although it consists of dozens of different protein parts, when I wrote about the flagellum in Darwin’s Black Box I focused on just the several mechanical parts—propeller, motor, and stator—that all rotary motors need to work, to show the system was irreducibly complex. As one biochemistry textbook put it, the bacterial rotary motor “must have the same mechanical elements as other rotary devices: a rotor (the rotating element) and a stator (the stationary element.)”1
However, not all rotary devices are equal. For example, the rotary device that spins the wheels on my son’s toy car is a far cry from the kind that operates a real motorboat. In turn, the motorboat’s engine is quite different in many details from one that powers an ocean liner. These different rotary systems all have a large number of parts—not just two
or three—all of which are necessarily precision-machined to the right shapes for the job. If one were to try to realistically sketch out the kind of automated assembly machinery that would put together any one of these, it would be quite different from the assembly machinery for any other of them. The assembly machinery would have to be different because the details of the assembly itself are different—the distance that one newly made part is from another in the staging area, which nut goes onto which bolt, what size clamp is needed to grasp a part, and so on. So when we are thinking about the assembly of the flagellum, we have to think about all the specific details of the particular machinery we’re making. Let’s briefly consider the structure of the bacterial flagellum.
Figure C.1 shows a sketch of a flagellum taken from a recent article in a science journal describing how a flagellum is built.2 A flagellum contains several dozen different kinds of protein parts, many of which are labeled with their scientific names in Figure C.1. The labels give a taste of the complexity of the parts, but in the following description I won’t use those labels—I’ll use more reader-friendly terms. Again, don’t think you have to memorize the details—just taste the complexity.
The Edge of Evolution Page 28