BOTTOM UP, TOP DOWN
I was in grammar school when the observation tower of Iacocca Hall at Lehigh University was being constructed. Busy studying the three Rs in Harrisburg about a hundred miles away, I never got the chance to see the cranes, cement mixers, dump trucks, and steel I-beams, to see all the machinery and supplies being carried around to the right places, to be joined in the right way with the right complementary pieces, to make the building where I now work. But like most of us, I’ve seen other buildings being constructed, so I can infer how the Iacocca Hall tower was put together. Like all such buildings, it was built in what could be called a “bottom up–top down” fashion. By bottom up I mean that of course the foundation of the building had to be poured first, the ground floor next, and so on, all the way to the zenith at the sixth floor. Successive floors have to be built on preceding ones.
By top down I mean that the building was planned. Blueprints were followed, supplies ordered, ground purchased, equipment moved in, and so on—all with the final structure of the observation tower in mind. Of course, minor features of the building might not be explicitly intended. For example, the exact color of the concrete might not matter, as long as it was an invigorating shade of gray. Or the exact placement of the handrails leading up the interior steps might not be important, as long as they were within a certain distance from the floor. Nonetheless, major structural aspects of the building were conceptualized in advance of the start of construction, and then preparations were taken to carry out the project. The need for bottom up–top down construction extends far beyond the buildings of Bethlehem, Pennsylvania. All major construction projects are conducted that way. So whenever we see a well-framed structure we may be sure it was planned.
In just the past decade or so science has unexpectedly discovered bottom up–top down construction in a location that wasn’t visible just a few years earlier. It wasn’t visible because the optical equipment needed to see it wasn’t available. It was spotted by a powerful new microscope scrutinizing the green alga Chlamydomonas, a favorite laboratory organism affectionately known as Chlammy. Since then the same type of construction has been spied in a very wide variety of cells.
In Darwin’s Black Box, I discussed large cellular structures called the cilium and the flagellum, both of which help cells move around in liquid, acting like propellers. I had no idea how complex they really were. Both the cilium and the flagellum are big pieces of cellular machinery—big, that is, compared to the cell itself. Although they are both quite thin, their lengths can be many times longer than that of the cells to which they are attached. It turns out that the construction of big structures in the cell requires the same degree of planning—the same foresight, the same laying in of supplies, the same sophisticated tools—as did the building of the observation tower at Iacocca Hall. Actually, it requires much more sophistication, because the whole process is carried out by unseeing molecular robots rather than the conscious construction workers who assemble buildings in our everyday world.
FIGURE 5.1
Computer-generated image of a section of a cilium, cut away to reveal component parts. Each small sphere is a protein of roughly the complexity of hemoglobin. The cilium is comprised of about two hundred different kinds of proteins. (Reprinted from Taylor, H. C., and Holwill, M. E. J. 1999. Axonemal dynein—a natural molecular motor. Nanotechnology10:237–43. Courtesy of IOP Publishing.)
GOOSEBUMPS
In 1993 Keith Kozminski, then a graduate student at Yale, was trying out a flashy new microscope.1 The scope had all sorts of bells and whistles, including the ability to videotape cells in real time. Kozminski focused the scope on a cilium of the single-celled alga Chlamydomonas and filmed what no one in the history of the world had ever seen before. Moving up one side of the cilium and down the other were a series of bumps—traveling goosebumps! A videotape of such “intraflagellar transport” (abbreviated IFT; confusingly, cilia are also sometimes called “flagella,” hence “intraflagellar”) can be seen on the web.2
Kozminski and his coworkers knew right away that there must be a lot of complex machinery behind the simple-looking, moving bumps. They hypothesized that the bumps were actually akin to traveling train cars, moving freight up the length of the cilium, and powered by various kinds of motor proteins. The bumps moved at different speeds; they went twice as fast coming back as they did going out from the cell to the tip of the cilium. So the investigators deduced that there were two separate mechanisms responsible for the outward trip and the return. Switching from videotapes to still pictures taken by higher-resolution microscopes, the workers were able to make out some details of the bumps. They saw groups—later called “rafts”—of up to forty lollipop-shaped particles situated between the outer circumference of the protein part of the cilium and the membrane that encloses it. Unlike some other types of transport machinery in the cell, the lollipops were not “vesicles.” That is, they were not enclosed spaces wrapped by a protein or membranous coat.
With a combination of serendipity and skill, a window was opened onto elegant and unsuspected cellular machinery. In the decade since Keith Kozminski first glimpsed IFT, tremendous progress has been made in detailing the many protein players in the mechanism, as well as the often dire consequences when mutations disrupt its work. Over the next several pages we’ll look at some details of IFT.
BUILDING A TOWER
IFT is the machinery that builds and maintains the cilium. If a cilium is cut off a Chlammy cell, another one will be generated over the course of an hour or so. During that hour little IFT rafts can be spotted busily flowing up one side of the growing structure and down the other. If, however, by clever laboratory manipulations, one or more of the protein components of IFT are deliberately broken, an amputated cilium will no longer be rebuilt.
In bottom up–top down construction, for convenience materials are often gathered in advance and brought to the building site. That was surely the case for Iacocca Hall’s tower, and it’s also the case for the cilium. Before starting to build a new cilium, cellular materials are brought to a staging site near the bottom of what will be the new structure. Of course, in human construction projects the conscious workers know which materials they need, recognize them, and bring only needed materials into the building site. In the cell, however, that all has to be done by highly sophisticated, automated mechanisms. It had been hypothesized that, in an area near the base of the new cilium, things called “transition fibers” act as filters to keep out unwanted, potentially disruptive materials. Douglas Cole of the University of Idaho reasoned that if that were indeed the case—if construction materials needed an admission ticket to get into the cilium—then no new materials would be allowed past the transition zone into the cilium if IFT were experimentally halted. That is precisely what was seen in several Chlammy mutants.3 The exact details of the filtering mechanism aren’t yet known, but you can be sure they won’t be simple.
Like all analogies, the comparison of the building of a cilium to a human construction project fails in a number of respects, all of which emphasize the much greater sophistication of cellular construction. Here I’ll mention just one aspect. Although a human construction crew leaves a building project once it’s completed, that’s not the case with the cell. If IFT is experimentally interrupted in a cell that already has a full, finished cilium, the cilium immediately starts to shorten until it disappears. IFT continues throughout the lifetime of the cilium, not only constantly bringing in new copies of ciliary components, but also removing old material. Experiments have shown that in apparently stable cilia whose length remains constant, in a period of several hours over eighty different kinds of proteins amounting to 20 percent of the mass of the cilium are exchanged.4
The current model for IFT pictures the freight cars at the beginning of construction to be mostly full. (Figure 5.2) After construction is completed the trains keep coming at about the same rate, but now some of the cars are empty. Apparently some as-yet-unknown switch
ing mechanism senses how much material the cilium needs at any particular moment and changes the proportion of freight cars between “cargo-capable” and “cargo-incapable” as the need arises. Unlike the tower of Iacocca Hall, the cilium is a dynamic structure, in which many of its protein parts are actively altered in response to changing internal and external conditions.
FIGURE 5.2
Intraflagellar transport (IFT). Molecular containers carry protein cargo from the cell to the tip of the flagellum. Containers return empty. To maintain the correct length of the cilium after it is built, a greater percentage of containers are believed to switch from a “cargo-capable” to a “cargo-incapable” form. (Reprinted from Snell, W J., Pan, J., and Wang, Q. 2004. Cilia and flagella revealed: from flagellar assembly in Chlamydomonas to human obesity disorders. Cell117:693–97. Courtesy of Elsevier Publishing.)
THE FULL MONTY
Writing of IFT as using little “train cars” shaped like “lollipops” that run along molecular “railroad tracks” is of course baby talk. The baby talk has a serious purpose—to abstract some important, overarching points without getting bogged down for the moment in too many details. But the moment comes when details have to be fully faced. A real train, say a steam locomotive, contains very many parts that all have to be working in order for the train to operate. An engineer who blithely ignored the details of those parts would soon find himself in charge of an immobile, hundred-ton paperweight. In the same way, the IFT apparatus contains many protein parts. It directly contains at least sixteen kinds of proteins, each of which is itself roughly the complexity of hemoglobin. And just as a mutation in one of the hundred-plus amino acids of either the alpha or beta chains of hemoglobin can cause it to malfunction, the same is true of the many protein parts of IFT. In the next few paragraphs we’ll stare directly into the maw of the biochemical complexity of IFT, and then come back up for air. Don’t worry about remembering the names of components or other details. The point is to see how elegant and interdependent the coherent system is—to see how different it is from the broken genes and desperate measures that random mutation routinely involves. Readers who don’t feel the need for this level of detail may wish to skip to the next section.
Biochemical studies show that IFT can be conceptually broken down into several parts. The first part consists of the motor proteins that carry the IFT particles along the interior of the cilium. The motor protein that carries the particle toward the tip of the cilium is different from the one that carries it back. The trip out is powered by kinesin-II, one member of a family of kinesin motor proteins that perform a variety of jobs in the cell. Kinesins come in a range of structural variants. Kinesin-II is found only in cells that have IFT, but not in cells such as those of yeasts and higher plants that don’t. (Yeasts and higher plants don’t have cilia.) One study showed that cells that contain a mutant, fragile kinesin-II can form cilia at lower temperature (about 68°F) where the mutant protein works. But at higher temperature (90°F) where the protein is unstable, IFT stopped and cilia began to be resorbed. The trip back is powered by a dynein motor protein. When a mutant, disabled dynein was placed in Chlammy cells that didn’t have cilia, new cilia that were formed were very short and bulging with IFT particles that contained kinesin-II. Apparently, the machinery for getting particles in was working fine, but the machinery for getting particles back out was broken, so the incipient cilium became overstuffed. Exactly what causes IFT to shift from kinesin-powered transport to dynein transport at the tip of the cilium remains unknown.
The second conceptual part of IFT is called the IFT particle. It’s the container that grabs hold of the correct proteins to be carried in or out and releases them at the proper point. The IFT particle consists of sixteen separate proteins that bind together in one aggregate. Under some experimental conditions the sixteen-protein complex can be separated into two complexes—called A and B—that contain six and ten proteins respectively. It’s not certain, but it seems that complexes A and B may play distinct roles in the cell.5 The proteins of complexes A and B contain substructures that are known to be particularly good at binding diverse proteins—exactly what you need to transport the many kinds of protein cargo that travel by IFT along the cilium.
TRAIN WRECK
When parts of a railroad transportation system are missing or broken—when a railroad tie is misaligned, a rivet or two missing, a bolt holding a wheel on the engine broken—disaster may not be far behind. So, too, with IFT. Cilia aren’t just oars flapping in the water—they participate in a wide range of critical biological functions. If they aren’t well maintained, a lot of things can go wrong. In the past decade defects in IFT have been shown to affect a number of important processes.
The earliest hint that cilia have a number of hidden but vital tasks came in the mid-1970s when Swedish scientist Björn Afzelius reported the cases of four men who suffered from infertility and chronic sinusitis. Since the tail of a human sperm is a modified cilium that powers its swimming, and since ciliated cells line the sinus cavities, Afzelius examined respiratory tissue from the patients and checked their cilia. Although cilia were there, they lacked the dynein that’s present in normal cilia, and thus were unable to move. Afzelius also noted something odd about his patients—several of them had situs inversus, that is, their hearts were on the right sides of their bodies and their livers on the left, the opposite of normal. Afzelius’s observation suggested that anything that broke a cilium might cause the left-right mixup. In 1999 some Japanese workers genetically manipulated mice to be missing one of the proteins that forms the kinesin motor of IFT. The mice died before birth. Examination of the embryos showed many to have situs inversus. So one conclusion is that a properly working IFT is necessary for correct embryonic development.6
Another area affected by IFT is vision. In the photoreceptor cells of the retina of vertebrates, a large inner segment (IS) is connected by a thin neck to a large outer segment (OS). The IS harbors the guts of the cell—nucleus, ribosomes, and so on—while the OS has the specialized machinery dedicated to vision. Since all the supplies needed for the OS are first constructed in the IS, they have to be shuttled from one compartment to the next. The connecting neck is actually a modified, nonmotile cilium, so it is suspected that supplies reach the OS by IFT. That hypothesis has been strengthened by recent work showing that a mutation in just one of the sixteen IFT proteins in mice causes the rodent retinas to degenerate.7 In another study the kinesin IFT motor was intentionally broken in lab mice; proteins that normally are shipped out to the OS became stuck in the IS. Eventually, as many improperly functioning cells do, the defective photoreceptor cells activated their self-destruct program and committed suicide.8
People who suffer from polycystic kidney disease develop large cysts on their kidneys (and other organs, too) and gradually lose kidney tissue, leading to kidney failure. Since kidneys are necessary to filter blood, the consequences can be deadly. Studies on humans showed that mutations in the genes for either of two proteins, called polycystin-1 and polycystin-2, were associated with the disease. Polycystin-2 is found in certain cilia of kidneys. In experiments with mice, deliberate breaking of one of the proteins of IFT eliminated the construction of those cilia and led to polycystic kidney disease in the animal.9 The conclusion is that IFT is needed for proper kidney function, too.
Besides its role in embryo development and eye and kidney function, IFT likely plays a number of other roles in the cell. Experiments point to functions in sensing the concentration of chemicals in liquid (osmotic sensing), receiving chemical signals, mating behavior in worms, and more.
IRREDUCIBLE COMPLEXITY SQUARED
When we’re children, life seems simple. We don’t know how the world really works, and don’t even know enough to ask questions about it. Our parents meet all our needs; our country can do no wrong; our school is the best. But while growing up, most of us discover that things aren’t so straightforward as they first appeared. A school bully punches us in
the nose; we hear some of our country’s actions denounced by people whose opinions we respect; Dad tells us we have to earn the money ourselves—he won’t just give us a car. Life gets more and more complicated. So, too, with biochemistry. In Darwin’s era in the nineteenth century the cell seemed boringly simple. The eminent embryologist Ernst Haeckel called it a “simple little lump of albuminous combination of carbon”10—in other words, just some gray goo. As it grew up over the years science has learned that the cell is tremendously more complex than Haeckel thought.
Now we realize that the cilium, too, is tremendously complex. Now we know that a cilium is more than just a flapping oar, useful for swimming or keeping liquid moving through a tissue. It’s also a sophisticated chemical sensor involved in a wide array of biological processes. It is dynamic in multiple, independent ways—not just mechanically dynamic, but also functionally dynamic, continuously being rebuilt to better reflect and respond to its environment.
And now the problem of its irreducible complexity has been enormously compounded. Let’s reconsider the mousetrap—the paradigm of irreducible complexity I discussed in Darwin’s Black Box. A standard mechanical mousetrap needs multiple parts to work. If the spring is removed or a metal bar broken, the trap won’t catch any mice. Despite the imaginative but dubious efforts of Darwin fans over the past decade,11 it’s extremely difficult to see how something like a mousetrap could actually evolve by something akin to a blind Darwinian search process. But now let’s move beyond the structure of just the mousetrap itself. Imagine an automated mousetrap factory that assembled the parts of the trap, set it, and reset the trap each time it went off. Clearly, the complexity of such a system is much greater than the complexity of the mousetrap alone. And just as the odds against winning a Powerball lottery skyrocket the more numbers you have to match, the difficulty of explaining how a mousetrap-making system could arise by “numerous, successive, slight modifications” (as Darwin required of his theory) rises exponentially the more separate kinds of parts the system contains.
The Edge of Evolution Page 10