Let us now turn our attention to the evolution of motility. The EFM hypothesis essentially proposes:

At this point, the EFM hypothesis becomes so vague and so speculative that it borders on the vacuous. Such assertions are without any independent support and have all the appearance of an ad hoc explanation invoked to rationalize a previously held belief in the cooptional origin of the bacterial flagellum. Nevertheless,let us consider current thinking about the flagellar motor and then come back to consider such claims in that light.

As of today, we have a fairly good handle on the components of the bacterial motor: Mot A, Mot B, FliG, FliM, and FliN. The motor itself is broken down into two basic components - a stator and a rotor. The rotor is the component that turns and the stator is an adjacent component against which the rotor turns.

The rotor is composed of FliM, FliN, and FliG. The proteins form a C-ring structure at the base of the flagellum just underneath the cytoplasmic membrane. (Figure 1).

This C-ring is composed of about 25-45 copies of FliG, about 35 copies of FliM, and around 110 copies of FliN. [2] It actually has three functions. [3] First, it plays an essential role in flagellar assembly. Secondly, it is part of the motor. And third, it is also part of the switch complex that mediates clockwise and counterclockwise spinning. A recent study from the March 23,2001 issue of Science helped to clarify the assembly-essential role of the C-ring. Apparently, it functions as a "quantized measuring cup" for the hook monomers. The current model is that the interior chamber of the C-ring is loaded with about 120 hook monomers (where FliM, FliN, and FliG each have four binding sites for the monomer) and these are then secreted en bloc to form a hook of a distinct length (helping to explain how the flagellum controls the assembly of its hook, as the heterogeneous natural of the hook/filament is often overlooked by some). After the hook monomers vacate the C-ring, another protein enters the chamber and converts it from a hook-secretor to a flagellin-secretor.

When it comes to motility, it appears that FliG plays the primary role, as FliM and FliN are more involved in switching. FliG is protein that is approximately 331 amino acids long. It is thought be directly involved in the generation of torque as a consequence of its specific interactions with the stator.

The stator is composed of motA and motB. These are both membrane proteins, where removal of either one abolishes motility. MotA has four membrane-spanning regions and most of its bulk is found on the cytoplasmic side of the membrane. MotB has only one membrane spanning domain and most of its bulk is in the periplasm, where it is anchored to the underside of the bacterial cell wall. Together, they form the torque generating units, as the not only form a stationary structure against which the rotor can move, but conduct sodium ions or protons (depending on the flagellum) from the periplasm to the cytoplasm and this ion/proton flow generates the spinning of the C-ring. There are at least 8 copies of the motA/motB torque generators that surround the C-ring (keep in mind that Figure 1 is a cross section through the flagellum).

The actual mechanism is torque generation is poorly understood. Let me briefly describe three models.

The first model, the "proton turbine model," proposes electrostatic interactions between motA/B and FliG, as can be shown by the right illustration in Figure 2 (keeping in mind only one torque generator is shown).

This model proposes that the flow of protons or sodium ions interact with carefully positioned charged residues on the FliG component of the C-ring, creating a dynamic electrostatic field that moves the rotor. The second model is sometimes called the "turnstile" (shown in the illustration on the right). Protons or ions enter the motA/B complex and are passed on to specific components of the rotor. Yet the rotor must spin to again pass on the protons/ions for entry into the cytoplasm. The third model (not shown) is called the "water turbine model." [5] In this model, protons or sodium ions bind to residues on motA/B. Normally, the protons/ions are complexed by a surrounding sphere of water molecules. Their binding to amino acid residues causes the water molecules to be vectorially ejected: "the binding of protons (or Na+ ions) to specific groups on MotA leads to the vectorial ejection of water molecules tangentially to the C-ring, thus causing its rotation."[5] In this model, the water molecules become an active participant in rotary motion. As of today, the first model, built around electrostatic interactions, appears most popular.

What all models share in common is the theme of specificity, whereby specific interactions between the stator and rotor are required to elicit rotation. For example,

Electrostatic interactions are weaker in water than in a less polar milieu. They could still exert significant forces, however, particularly if the interacting groups are positioned in the motor so as to ensure that they will approach each other closely at some step(s) in motor rotation. Also, water might be partially excluded from the rotor-stator interface, which would make the interactions stronger. (emphasis added) [6]

Or consider FliG. The torque generation function associated with this protein is restricted its C-terminal domain. Random mutations were introduced into the gene coding for this protein and yielded a set of mutants with flagella that did not rotate. In fact, even if the mutant protein was overexpressed, motility was not restored. All but one of these mutants involved the loss of a hydrophobic amino acid and were found to make the proteins subject to degradation. [7] This likely means that these mutations altered the conformation of FliG such that its charged residues directly involved in torque generation were no longer properly positioned. The mutations involved residues at positions 234, 237, 249, 252, 257, and 306. Remember that a mutation in any one of these sites resulted in complete loss of motility. In addition, small deletions also had the same effect: deletions of residues 280-285 and 292-295. With this in mind, let us return to the EFM hypothesis.

The EFM hypothesis envisions some ill-defined fortuitous interaction between some unknown ion channel and the basal body of the non-motile filament. Then somehow, motility spontaneously emerges and selection takes over from here. Once again, pure chance must bring about the new function. But how likely is this?

First, I'm having a hard time envisioning this. The rotor must have access to the proton/ion through the ion channel. What type of fortuitous change is going to pry open this ion channel and then glom it onto the proto-rotor? It would seem that of all the ways to mutate an ion channel, such a change would represent only a very small minority of all possible changes.

In fact, this very same theme repeats itself with all necessary parts of this fortuitous interaction. Of all the ways to mutate an ion channel, the number of ways that would result in its interacting with the base of some filament is surely in the distinct minority. And of all the ways to mutate an ion channel that gloms onto a filament, the number of ways to mutate it such that rotation does not occur is probably much higher than the number of ways to elicit some rotation.

Thus, as with the first cooption event, we need another special mutation. This one allows some ion channel to glom onto the base of a filament and open its channel and expose the ion flow to the proto-rotor in such a way that a set of electrostatic interactions just happen to form and elicit significant rotation. Suffice it to say that such an improbable mutation has never been observed in nature or the lab.

Of course, just because such a mutation may be highly improbable does not mean it did not occur. But if your thesis is built around the existence of improbable events, you need some type of independent evidence to support such claims. Especially as the whole problem of IC again raises its head.

The appeal to cooption to explain the origin of rotation still fails to escape the grasp of IC. This is because the motor depends on three proteins: fliG, motA, and motB. Remove any one of these three proteins and you do not have a motor that is 2/3 as good as the whole. Remove any one of these three and you have a non-functioning motor.

Consider fliG. The set of hydrophobic mutations above abolished only motility and not flagellation. Keep in mind that this set of mutations does not represent an exhaustive screen of all the residues crucial to fliG's torque generation (we'd have to add in other residues important for structure and those that play a role in torque generation). Yet they (or some similar class) are essential for motility. Yet without the motA/motB interactions, they have no known function. MotA/MotB, on the other hand, could plausibly exist as some ion channel prior to the existence of the flagella, but there is no evidence of this. And there is no reason to think that the residues crucial for motility (important for positioning and direct interactions) would be under any selective constraint prior to the supposed fortuitous interaction. Thus, as far as we can tell, all the information necessary for motor function provided no selective benefit until motility appeared. And we're left with the intuitive awareness that of all the various sequences that can be shared by any randomly chosen three proteins, the number of ways to hook up into some motor structure are far less likely than the ways that do not elicit a motor function.

Of course, just because an interaction may be highly improbable does not mean it did not occur. But if your thesis is built around the existence of improbable events, you need some type of independent evidence to support such claims.

Another aspect of this motility component of the EFM hypothesis worthy of a critical look is the assumption that some kind of primitive, proto-motility function would be selectively advantageous. While a crude Darwinian "common sense" would seem to indicate this, I am not so sure. To appreciate why, we need to ask why it is that modern day bacteria move in a series of straight runs and tumbles. Why don't they simply swim straight for a food source instead of taking a convoluted path involving short bursts of straight runs interspersed with tumbles that randomly reorient them? In fact, bacteria will only be propelled by their flagella spinning about 100-300 times/sec for about 3-4 seconds. Why?

We sometimes forget that the small-scale world of bacteria is much different from our macro-world. Bacteria are constantly being buffeted by water molecules and thus live in a "Brownian storm." The simple fact is that because bacteria are so small, they swim through a Brownian storm. Brownian motion will knock bacteria off course after 3-4 seconds. [4] And this highlights a serious problem with the EFM hypothesis. The flagellum is a highly sophisticated machine. Even if one believes it evolved, what we study today is the product of billions of years of evolutionary modification. Yet even this high sophisticated/highly evolved system barely overcomes the Brownian storm. Thus, just how advantageous would some proto-wiggle really be? Imagine a boat in the ocean during a tropical storm. Would a propeller that spun once every second really be any better than no propeller? In other words, it is possible that biologically significant motility on these scales depends on a minimal amount of system complexity and output that is out of reach in a Darwinian search beginning with simple states. To assure myself this was not the case, I did a PubMed search with the following search words: " partial motility flagella selective advantage" and it returned 0 hits. I obtained one hit with the search words partial motility selective advantage" and this was not a relevant study. Thus, this essential feature of the EFM hypothesis is without any evidential support.

To sum this section up, let's consider more problems inherent in the EFM hypothesis

• The fortuitous interaction component of the story is incredibly vague and without any independent support and appears to be an ad hoc component invented merely to prop up an a priori belief that cooption was behind the origin of the flagellum.
• Flagellar motor function appears to require a high degree of specific interaction among the parts. The information content involved is likely to high.
• The fortuitous interactions invoked to elicit rotation appear to be highly improbable, in light of the likelihood that such functional interactions would represent a very small fraction of all possible types of interactions. Also, we're faced with positing special mutations again.
• The fortuitous interaction does not eliminate the problem of IC even at this micro-level.
• Invoking a selective benefit for some ill-defined "partial motility" is completely without evidential support. It would seem that some form of minimal motility function is required to cross the threshold of selective benefit.

NEXT: In the next essay, I will consider the generic appeal to cooption in this example to determine if teleologists ought to abandon their thesis and adopt it instead.

1. The bacterial flagellar motor.
2. DeRosier, DJ. 1998. The Turn of the Screw: The Bacterial Flagellar Motor. Cell 93, 17-20.
3. Yamaguchi S, Fujita H, Ishihara A, Aizawa S, Macnab RM. 1986. Subdivision of flagellar genes of Salmonella typhimurium into regions responsible for assembly, rotation, and switching. J Bacteriol 166(1):187-93.
4. Bacterial Flagella: Flagellar Motor.
5. Oplatka A. 1998. Do the bacterial flagellar motor and ATP synthase operate as water turbines? Biochem Biophys Res Commun. 249(3):573-8.
6. Jiadong Zhou, Scott A. Lloyd, and David F. Blair. 1998. Electrostatic interactions between rotor and stator in the bacterial flagellar motor. PNAS, 95, 6436-6441
7. Lloyd, SA and Blair DF. 1997. Charged residues of the rotor protein FliG essential for torque generation in the flagellar motor of E. coli. JMB. 266, 733-744.