THE NEGLECTED FLAGELLUM
By Mike Gene (
The first example of an IC system that Michael Behe provided in his book, Darwin’s Black Box, was the flagellum. While the bacterial flagellum has indeed become a focal point in the debates concerning design, Behe focused most of his attention on the eukaryotic flagellum. How is it that the bacterial flagellum quickly overshadowed the eukaryotic flagellum in the debates about irreducible complexity and design?
A charitable interpretation would note that the bacterial flagellum has received more attention because it looks so much like a product of design. Functioning as a nanoscale version of an outboard rotary motor, even flagellar expert David DeRosier acknowledges “the flagellum resembles a machine designed by humans.” (Cell 93: 17-20). Given the bacterial flagellum looks like something a human might design, more so than many other molecular machines, it is understandable why this particular machine would attract more interest.
A cynical interpretation would note that the irreducible complexity of the eukaryotic flagellum does not really pose an insurmountable obstacle to Darwinian evolution. Two powerful counterarguments have been developed and many in the ID community have thus quietly shied away from this example. Let’s consider the two arguments.
Different Patterns
When Cavalier-Smith reviewed Behe’s book in 1997, he responded to Behe’s claims about the flagellum by essentially setting up a straw man. Cavalier-Smith was under the impression that Behe was arguing that the 9+2 pattern of microtubules represented the IC essence of the flagellum. Cavalier-Smith then easily knocked this down by pointing out that there are functioning flagella that don’t show this pattern. Ken Miller picked it up from here. In his book, Finding Darwin’s God, Miller finds himself “amused” at this argument, adding, “A phone call to any biologist who had ever actually studied cilia and flagella would have told Behe that he’s wrong in his contention that the 9+2 structure is the only way to make a working cilium or flagellum.” (p.141). Miller then cites several examples of flagella that don’t show this 9+2 pattern, including the 9+0 arrangement from eel sperm and the much reduced 3+0 arrangement from the protozoan Diplauxis.
Unfortunately, Behe’s argument never rested on the 9+2 flagellar arrangement. Contrary to Miller’s assertion, not once did Behe contend that the 9+2 structure is the only way to make a functioning flagellum. Behe does include an illustration of the 9+2 arrangement, but this is a standard illustration borrowed from a standard biochemistry textbook. That’s hardly justification for portraying Behe’s argument as being dependent on the 9+2 arrangement as the focus of his discussion of irreducible complexity. This is especially clear in light of the fact that Behe spells out the “IC recipe” of the eukaryotic flagellum on pp. 64-65:
What components are needed for a
cilium to work? Ciliary motion certainly
requires microtubules; otherwise, there would be no strands to slide. Additionally, it requires a motor, or else
the microtubules of the flagellum would lie stiff and motionless. Furthermore,
it requires linkers to tug on neighboring strands, converting the sliding
motion into bending motion, and preventing the structure from falling
apart. All of these are required to
perform one function: ciliary motion.
Behe thus reduces the cilium to a three-component IC system. And given that the examples cited by Miller all include these three parts, they hardly amount to an effective response to this argument. The Cavalier-Smith/Miller response seems to be a classic example of raising a straw man for the purpose of knocking it down. However, there is something here that I will explore below.
The Cooptable Toolkit
If we take Behe’s list of IC components, they really don’t seem to pose much of an obstacle for Darwinian evolution. This is because two of the three components are also known to function apart from the flagellum. Microtubules (composed of the protein tubulin) are needed for ciliary function, as Behe notes, but they also play many essential roles in the cytoplasm of all eukaryotes. For example, microtubules are intimately involved in the organization of the cytoplasm and the separation of chromosomes during mitosis. This means that this IC component could have theoretically existed long before flagella existed. The same holds true for the motor, dynein. Cytoplasmic versions have been identified and they function to transport cargo inside the cell. This basic “house-keeping” function could very well have also existed before the existence of flagella. All that is needed are linker proteins, and there are scores of proteins that fulfill this function inside the cytoplasm. Thus, it would not be hard to imagine these three components, or duplicated (and tweaked) versions of these components coming together through cooption. The standard story here is that some form of protrusion on the cell surface could first be formed, creating some type of selectable function (adherence, increasing cell surface area, enhancing the ability of cells to remain suspended, etc.) We could point to the axopodia [1], of certain protozoa as examples. This protrusion would then associate with the motor proteins to begin movement. Natural selection would then fine tune the structure over billions of years.
While the story appears quite plausible, it is also worth mentioning that Behe anticipated this explanation in his book and addressed it accordingly (pp. 66-67). He points out that the story is much too vague and it is easy to envision multiple problems were we to get more specific. In another words, if the story breaks down upon attempting to flesh it out, the story ceases to function as a counter-argument against IC. Thus, from Behe’s perspective, until someone fleshes out the story with details, there is no reason to think it maps to reality and the IC obstacle remains.
One can be both critical and sympathetic to this type of response to the cooption story. From the critical perspective, the demand for such details seems excessive, as much of this detail will have been lost in history. Furthermore, the shift to details detracts from the simplicity of the IC-to-ID inference, as we’ve moved from simply scoring IC components to accounting for their history. From the sympathetic perspective, certainly the vague account as listed above cannot be considered a robust explanation. As it stands, it really is a “just-so” story that may only reflect the human brain’s tremendous ability to imagine accounts. After all, the causal explanation for the origin of flagella must be situated in history. Depending on one’s expectations concerning evolutionary explanations, the challenge posed by the IC flagellum will be assessed differently.
But before moving on, I would like to address an element of the story not discussed by Behe. The cooption story conveys the impression that evolving cilia/flagella would be rather easy. All the parts already exist in the cytoplasm and the selective benefit from a protrusion would be widely enjoyed by all sorts of protozoa, who all could benefit from improved adherence, increasing cell surface area, and enhancing the ability of cells to remain suspended. Thus, the story would seem to predict that such proto-flagellar protrusions would be fairly ubiquitous among the various major groups of protists, even independently evolving on many different occasions across deep time. Yet as far as I can tell, such structures are not common at all. That people must point to axopodia actually undercuts the argument.
There are three things worth mentioning about axopodia. First, they appear to be largely restricted
to two of the eighteen protozoan phyla, Heliozoa and
Radiozoa [2]. Secondly, axopodia are not
primitive protrusions in any sense. They
are very effective organelles used to capture prey, acting much like molecular
tentacles. For example, if an protozoan
comes into contact with an axopod, the axopod quickly retracts, reeling in the
protist for delivery to food vacuoles.
In other words, axopodia are rather specialized features[/URL]. The limited distribution and specialization of
such structures argue that they are derived, not primitive. What’s more, the contractile activity of
axopodia is not mediated by dynein nor is it ATP-dependent, but instead it
appears to be dependent on calcium [3], giving us no reason to connect axopodia
to flagella.
The third point related to
axopodia is to wonder why they have not evolved into flagella, given they have
all the ingredients of the cooption story in play. Long cell extensions built around
microtubules. Bending has been observed during contraction.[4] And as one page
notes:
Although this microbe lacks locomotory structures like cilia and flagella, it can move slowly by motion of stiff axopodia or by floating along in the water current. [5]
The axopodia are also moonlighting as motility organelles. There would thus appear to be a smooth slope up the fitness landscape to some form of flagellum. Simply follow the cooption story – duplicate a motor protein and a linker, toss them into axopod, it will start wiggling, and the rest is smooth selective sailing. Yet Actinophryidae have no flagella. Neither do Radiolaria (except when specialized reproductive cells are formed). But there are Helizoans that have both flagella and axopods, indicating that the existence of axopods does not preclude the existence of flagella.[6] The fact that many protozoa with axopodia have not evolved flagella, when the stage has been set to do so for a billion years or so, clearly indicates something is missing from the cooption story. Behe was right in asking for more details.
Return to IC
Behe scores only three part for the flagellum: microtubule, motor, and linker. Yet this list would seem to seriously underestimate the complexity of the flagellum. Behe later recognizes this:
Above I noted that the cilium contains tubulin, dynein, nexin, and several other connector proteins. If you take these and inject them into a cell that lacks a cilium, however, they do not assemble to give a functioning cilium. Much more is required to obtain a cilium in a cell. A thorough biochemical analysis shows that a cilium contains over two hundred different kinds of proteins; the actual complexity of the cilium is enormously greater than what we have considered. (p. 72)
This count actually comes from the study of the green algae, Chlamydomonas reinhardtii, where approximately 250 proteins were detected in the flagellum.[7] However, about half of these have not been detected through genetic screens using nonmotile mutant strains, suggesting that approximately 125 are both essential and specific to the flagellum. If only 10% of these components form an IC core that resists explanations thought cooption and/or duplication, it would seem IC does indeed pose a serious problem for a Darwinian explanation.
Let’s focus briefly on dynein. Dynein itself is actually a complex motor, typically consisting of two heavy chains, 2-4 intermediate chains, and several light chains. The heavy chains constitute the motor and function in an ATP-dependent fashion, while the other chains function as adaptors for binding/transporting cargo. The modular construct makes sense from a design perspective and would also seem to allow dynein to be a good target for gene duplication/divergence, extending its reach with evolution. Thus, it is not surprising that there are many different variants of dynein playing roles in different processes. Yet all the variants fall into two basic categories – axonemal (flagellar) and cytoplasmic. As noted above, this makes it easy to envision a cytoplasmic variant being duplicated and tweaked to become a axonemal motor. But perhaps things are a little more complicated than this.
If we consider only the dynein heavy chains, Chlamydomonas possesses 16 different versions. Two are cytoplasmic and the other fourteen are associated with the flagellum. [8] The dyneins are situated in specific regions of the flagellum and play different functional roles. This raises the question of whether or not a truly “wiggling” flagellum could be formed from using only one type of dynein (the original duplicated version of the ancestral cytosplasmic form in the cooption story). For example, mutations in two of the axonemal heavy chains resulted in short, paralyzed flagella that would cause the cells to sink in liquid medium. [8] And thus far, comparative data indicate multiple axonemal dynein forms coincided with the origin of flagella:
The size of the Dhc gene family in Chlamydomonas is comparable with that found in other species such as sea urchin (14), Paramecium (12), Drosophila (>7), rat (13–15), mouse (11), and humans (>8) (Asai et al., 1994 ; Gibbons et al., 1994 ; Rasmusson et al., 1994 ; Tanaka et al., 1995 ; Andrews et al., 1996 ; Vaisberg et al., 1996 ; Vaughan et al., 1996 ; Neesen et al., 1997 ). The remarkable conservation of the Dhc gene family between such diverse organisms is consistent with the proposal that the Dhc gene family diverged into a small number of groups relatively early in the evolution of eucaryotes, but after these groups were established, they remained largely unchanged (Gibbons, 1995 ). [9]
Furthermore, it is also becoming clear that some of the dyneins are not directly involved in motility, but rather in the assembly of the flagellum. Consider the light chain LC8 of outer arm axonemal dynein in Chlamydomonas. It’s a small, highly conserved protein as evidenced by the 90% sequence identity seen in humans, Drosophila melanogaster, Caenorhabditis elegans, and Chlamydomonas. [10]. Genetic deletion of this gene product in Chlamydomonas results in the formation of stumpy, non-functional flagella. Yet the morphology, cell division, and growth of these cells is not altered. A similar phenotype was seen when one of the heavy chain genes, cDhc1b, was deleted. [9]
While gene duplication followed by divergence remains a very plausible explanation for the origin of flagellar dyneins, one has to wonder if a more detailed analysis would detract from the plausibility of this explanation. For example, if a flagellum requires two distinct dyneins, one for flagellar function and one for flagellar assembly, might we have to invoke to separate, concurrent gene duplication events to provide this machinery?
IC Assembly
As we begin to better understand flagellar formation, a distinct IC theme is appearing. Eukaryotic flagella form at the tip, posing a logistical assembly problem. Basically, how do you get about 250 different proteins into the flagellum such that they assemble in a very specific fashion? Many of the pieces are partly assembled at the base of the flagellum and then transported into the flagellum by a specialized pathway known as Intraflagellar Transport (IFT). The basic ingredients are a kinesin-II motor that shuttles precursors into the flagellum, a dynein motor (cDhc1b), and an IFT “raft” that functions much like the trailer on a truck, carrying the material in/out of the flagellum [11]. Click here for a schematic from Cole.
The kinesin and dynein motors could be explained by gene duplication. Yet here we would seem to require concurrent duplications, as both motors are essential for flagellar assembly. If you knock out the kenesin-II (also known as fla10), no flagella assemble (the phenotype of called “bald”), as the machine for moving the material into the flagellum is missing. If you knock out the transport dynein, stumpy, non-functional flagella form, as material (including the kinesins) is not trucked out of the flagella and thus accumulate into a disordered tangle.
Even more intriguing are the IFT rafts. These particles are composed of 16 different proteins, some large and some small. They can be isolated as two distinct complexes: A and B. Complex A contains IFT 144, IFT140, IFT139, IFT122A, IFT122B, and IFT43. Complex B contains IFT172, IFT88, IFT81, IFT80, IFT74/72, IFT57/55, IFT52, IFT46, IFT27, and IFT20. The sequences of many of these proteins are conserved among the protozoa, nematodes, and vertebrates [12]. Genetic manipulations in green algae, whereby the genes are essentially knocked out, demonstrate the important of these proteins. For example, removal of IFT88, IFT172, IFT140, and IFT52, result in the same bald phenotype as removal of the kinesin transporter [11]. Removal of the IFT proteins have no other effect on the cell, indicating a flagellum-specific function. (I’ll discuss the IFT proteins in more detail at a later time).
If the assembly of flagella require multiple, independent parts shuttling material in and out of the flagellum, then the serious IC challenge posed by the flagellum may not so much reside in the structure described by Behe, but in the manner this structure is assembled.
More later…..
Cites:
