I have previously discussed the various evolutionary mechanisms proposed to overcome the problem posed by Irreducible Complexity (IC). [1] Evolution through the coincidental cooption of an alternative function (CCAF) remains the best non-teleological explanation for the origin of IC. Thus, it is not surprising that the EFM hypothesis builds around this type of thinking. But there are two ways to envision CCAF: simultaneous cooption or gradual cooption. For the sake of simplicity, let us pretend the flagellum is a six-part IC system, where A,B,C,D,E, and G are essential proteins and F* is rotary motility. A simplistic model of simultaneous cooption is shown in figure 1.

Figure 1. Simultaneous CCAF

Here A, B, C, D, E, and G all pre-existed the flagellum and had alternative functions in the cell. Then, through some type of fortuitous event, they all came together and spawned flagellar function. Generating a novel function through the chance conglomeration of six independent proteins seems highly unlikely for two basic reasons:

1. Cooption really needs the help of gene duplication in order to donate the components to the new system. Presumably A-E and G exist because they provide some other selective benefit to the organism. Thus, titrating off these components, and their functions, onto another system is likely to bring about deleterious effects in six other systems. Thus, to make this explanation more plausible, we'd have to invoke nearly simultaneous gene duplications in six independent spots in the genome and ensure they were expressed at the same time.
2. The flagellum is a machine, thus its components must physically interact in a stable fashion to carry out the series of coordinated movements that reverberate as a function of the initial energy input. A nice example of this can be seen with some FliN mutants that cause a complete loss in motility in bacteria that still had flagella. Such mutants led to the original interpretation that FliN played a direct role in the torque generation of the motor. But later work better explained these mutants as having reduced binding to their sites in the flagellum. [2] Thus, merely unstabilizing the interaction between FliN and its partners resulted in a complete loss of motility, highlighting the importance of these "well-matched" physical interactions. Since components A-E and G were all previously shaped by selection for various non-flagellar functions, it seems highly unlikely that they would just happen to have the "right" match of conformations to sufficiently interact and generate some proto-flagella.

Thus, it is not surprising that evolutionary biologist, H. Allen Orr, dismisses this type of solution to the problem of IC:

First it will do no good to suggest that all the required parts of some biochemical pathway popped up simultaneously by mutation. Although this "solution" yields a functioning system in one fell swoop, it's so hopelessly unlikely that no Darwinian takes it seriously. As Behe rightly says, we gain nothing by replacing a problem with a miracle. [3]

Perhaps we should then turn to the more "Darwinian" version of CCAF, one that merely envisions gradual addition to the IC system while sustaining the components by giving them each an alternative function along the way. The evolutionary origin of the same six-part system might come together gradually as in figure 2.

Here, we invoke 10 alternative non-flagellar functions to sustain individual parts and partial-IC systems. The advantage to this explanation is that it gets us as far away from simultaneous CCAF as possible. It captures the gradualism at the heart of Darwinism and helps to push each addition to the system closer to observable mutation effects, such as those seen in the generation of antibiotic resistance. That is, mutations in one gene at a time can add the component to the system. What's more, gene duplication can now more plausibly contribute to the account, as a duplication could have occurred prior to each cooption event.

Yet Orr is still skeptical of this type of account:

Second, we might think that some of the parts of an irreducibly complex system evolved step by step for some other purpose and were then recruited wholesale to a new function. But this is also unlikely. You may as well hope that half your car's transmission will suddenly help out in the airbag department. Such things might happen very, very rarely, but they surely do not offer a general solution to irreducible complexity. [3]

His solution is to invoke Original Helping Activities [1, 3], which would make the scenario more complex as many cooptional events would simply aid and not yet be part of a truly IC structure (as defined by Behe). Mutations would later alter them and confer an essential status upon them.

When viewed like this, IC ceases to pose any problem. Since IC is a function-dependent concept, and functions can disappear and emerge, IC states simply evolve. In fact, this viewpoint is so seductive that it has led many to believe a replay of life's tape (borrowing from Gould's metaphor) would simply generate something comparable to the bacterial flagellum. That is, while the actual flagellum we see may not exist in some replay, some comparable, complex motility structure would likely exist playing its role in filling the niche afforded by motility.

Yet despite its seductive appeal, we must ask a simple question : does this explanation really account for the origin of the bacterial flagellum? We are not merely engaged in speculative philosophy here. We are talking about something that actually exists and has an actual history. Possible worlds are fun to think about, but historical claims must have more than an intuitive appeal to the possible to support them. History is not about what could have possibly happened; it is about what did happen. Thus, does this gradual CCAF explanation really explain the origin of something that exists, namely, the bacterial flagellum? That we can vaguely envision it happening means only that we should seriously consider it as an explanation. It does not mean we should adopt it as the explanation. It the previous three essays, I highlighted many of the problems with the EFM hypothesis. Let us then step back from this example to see if further problems exist.

Julie Thomas originally surveyed the bacterial flagellum from the perspectives of Ur-IC and thematic IC. [4] Ur-IC is a postulated IC state that existed in the last common ancestral flagellum. Thematic IC focuses on the various functional roles entailed by the components of a machine to determine if the roles themselves exist in an IC relationship. Thematic IC might also be a helpful concept as we try to begin reverse engineering the flagellum.

If we look to the standard E. coli flagellum, it is composed of eight subsystems with distinct functions:

1. The Base: Composed of FliF (the MS ring) and FliG(N-term), FliN, and FliM (the C-ring). These protein rings in the inner membrane are the first structures built. Removal of any of these genes results in the inability to further construct the flagellum, indicating they serve as the foundation of the flagellum. Also, as mentioned previously, the C-ring plays additional function roles. Only the N-terminal 200 amino acids of FliG seem important in these regards, as a mutant FliG with approximately 100 amino acids truncated off its C-terminus still forms flagella, but motility is lost. [2]
2. The Motor: Composed of FliG (C-terminal), MotA, and MotB. MotA/B play two roles as part of the motor: a) They serve as the stator against which the rotor moves and; b) They conduct protons (or sodium ions) that serve as the energy source for motility. The C-terminal domain of FliG directly interacts with the stator and plays an essential role in torque generation.
3. The Switch: Composed of FliN and FliM, along with other proteins that are part of the chemotaxis system. The switch allows the motor to change from a clockwise rotation to a counter-clockwise rotation (and visa versa).
4. The Export Machinery: Composed of flhB, fliQ, fliR, fliP, FliI, and flhA. These proteins form a core of the type III machinery discussed in Part I of this series. Also included, though much less conserved, are FliF, FliG, and FliN. This machinery exports proteins that will form the more distal components outside the inner membrane.
5. The Drive Shaft: Composed of flgB, flgC, fliE, and flgG. These proteins form a tube that crosses the periplasm and transmits the torque generated by the motor to the filament.
6. The Bushings: Composed of two rings, the L ring formed by flgH and the P ring formed by flgI. FlgG is the most distal driveshaft protein that probably interacts with these rings. These rings facilitate the driveshaft's penetration of the outer membrane.
7. The Hook: Composed of flgE. It transmits torque from the drive shaft to the filament. It is connected to the filament through hook-filament junction proteins, flgL and flgK.
8. The Filament: It acts as the "propeller." It is composed of flagellin (fliC) and the cap (fliD) discussed previously in Part II.

We could theoretically eliminate many of the genes listed above if we reverse engineer our way to a simpler flagellar prototype. The base itself could possibly be composed of one protein. The motor could be composed of one membrane protein and use the base as the rotor. We can eliminate the switch, as this is more of luxury item. For example, when the motor is switched such than bacteria tumble, this merely amplifies the effects of Brownian motion to facilitate the reorientation of the cell. Simply turning off the motor could accomplish much of the same effect. In fact, this is how the flagella in Rhodobacter sphaeroides work. [5] The export machinery is hard to reverse engineer since it is still a black box. Nevertheless, let's be radical and assume it could function much the same with only three proteins. The drive-shaft could theoretically be composed of a polymer composed of only one protein. The L and P rings are not needed in gram-positive bacteria, since such bacteria lack outer cell membranes, thus we can eliminate this from the prototype (although it raises an interesting question about the origin of gram-positives and gram-negatives to be explored at a later date). Finally, it would seem you could eliminate the hook and filament proteins and simply have the drive shift polymer extend well past the outer membrane. This would leave us with the Base, the Motor, the Export Machinery, and a Filament, which is awfully similar to the assumption about original flagella inherent in the EFM hypothesis (where the filament and motor are added to the export machinery through CCAF). And this would leave us with a total of only six proteins. Of course, a six-protein IC system still poses an IC problem, but in light of the EFM hypothesis, it does not seem insurmountable.

Yet the problem with the above approach is two-fold. Our ability to reverse engineer, in a realistic manner, is greatly hindered by our lack of understanding about the mechanistic interactions between the various components of the flagellum. Furthermore, our experience with designing nanotechnology is practically non-existent. Both factors suggest this six-part system may very well be over-simplistic and non-workable. Recall, for example, that the flagellum must be assembled and the act of assembly may place constraints on the minimal complexity of a rotary nano-propeller. For example, we have already seen in part II that filament assembly appears to be IC, requiring both flagellin and the cap. That is, the cap is not an added luxury item that merely seals off the filament (as was once thought). It is an essential assembly factor, playing the role of a processive chaperone. Thus, we can predict that further understanding of the flagellar components is likely to expand the list of essential players in the minimal flagellar prototype.

Because of these limitations, it would be helpful to consider the flagellar components last shared by an ancestral flagellum. Thomas [4] originally did this by comparing the flagellar gene content of various distantly related bacteria. Allow me to do likewise and provide a list of flagellar genes found in Aquifex aeolicus, Bacillus subtilis, Escherichia coli, and Treponema pallidum. (Table I)

Table I. Genes Likely Part of the Ur-IC Flagellum [6]

Thus, 21 genes are shared by these four distantly related bacteria. Aquifex aeolicus are the most thermophilic bacteria, growing just below the boiling point of water. They are also thought to represent the earliest lineages to branch off the eubacterial tree. Bacillus subtilis is a gram-positive soil bacterium that can use a wide variety of carbon sources. Very similar bacteria (Clostridium) used to be thought most primitive. Escherichia coli represent the gram-negative proteobacteria and live in the digestive tracts of many organisms. Their flagella are among the most studied. Treponema pallidum is a spirochete whose flagellum is part of a rather specialized motility organelle known as the axial filament. As I mentioned, these four species are very distantly related, as seen by the phylogenetic tree constructed from 16s rRNA sequence (Fig 3). Furthermore, all four bacteria have experienced very different environmental pressures over the last several billions years. This strongly implies that these 21 genes were present in the last common ancestor of all eubacteria, thus comprising the Ur-IC flagellum. To further test this notion, I surveyed the flagellar genes of Thermotoga maritima since it is also a very deeply branching bacterium. According to the TIGR list of flagellar genes, everything in the Ur-IC list is represented, thus confirming what IC would predict.[7] Furthermore, this Ur-IC state has persisted for billions of years since it appeared. That billions of years of microbial evolution, in each lineage, have not imposed significant permutations on this IC core speaks to its true IC state.

Since the last detectable flagellum most likely contained these 21 genes (22, if we split FliG; 23 if we separate FliM and FliN), we can finally turn to the hypothesis of gradual CCAF to understand why it is so unconvincing.

1. TeleoLogic 7
2. Lloyd SA, Tang H, Wang X, Billings S, Blair DF. 1996. Torque generation in the flagellar motor of Escherichia coli: evidence of a direct role for FliG but not for FliM or FliN. J Bacteriol 178(1):223-31
3. H. Allen Orr
4. Julie Thomas
5. Shah DS, and Sockett RE. 1995. Analysis of the motA flagellar motor gene from Rhodobacter sphaeroides, a bacterium with a unidirectional, stop-start flagellum. Mol Microbiol, 17, 961-9.
6. I've split the FliG into its two separate functional domains. I've collapsed FliM/FliN into one species for three reasons reasons: a) Various studies have shown FliN and FliM to be closely related in a functional sense; b) In Bacillus subtilis, FliN is missing, but another gene, FliY, exists that appears to be a fusion of FliM and FliN (Mol Microbiol 1992 Sep;6(18):2715-23; and c) among all the groups compared, bacteria either had FliN, FliM, or both. For example, species lacking FliN had FliM. Species lacking FliM had FliN.
7. TIGR
8. Perry, JJ and Staley, JT. 1997. Microbiology: Dynamics&Diversity. Saunders College Publishing; Fort Worth. p. 405.