I’ve been reading the recently published book Microbes and Evolution: The World that Darwin Never Saw, which combines my two primary areas of interest: microbiology and evolution. Chapter 38 of the book is written by Kelly Hughes and David Blair of the University of Utah, two of the world’s leading experts on bacterial flagellar assembly. Having followed the work of Kelly Hughes and his colleagues for a few years, I hold their work in the highest regard. I myself have a deep fascination with the subject of bacterial gene expression. I was intrigued, therefore, when I discovered the title of Hughes and Blair’s chapter: “Irreducible Complexity? Not!”
Following a very basic overview of flagellar structure and function (also described in my own detailed review of the subject), Hughes and Blair ask, “Is the flagellum irreducibly complex, or just complex?” They write,
“It is clear that the flagellum is a complex structure and that its assembly and operation depend upon many interdependent components and processes. This complexity has been suggested to pose problems for the theory of evolution; specifically, it has been suggested that the ancestral flagellum could not have provided a significant advantage unless all of the parts were generated simultaneously. Hence, the flagellum has been described as “irreducibly complex,” implying that it is impossible or at least very difficult to envision a much simpler, but still useful, ancestral form that would have been the raw material for evolution.”
Having thus represented the problem, they proceed to show that sub-components within the flagellar structure are homologous to other bacterial organelles. For example, they correctly point out that the stator proteins MotA and MotB are homologs of ExbB and ExbD, which form part of the TonB-dependent active transport system and which serve to energize transport of vitamin B12, and iron-chelating compounds called siderophores, across the outer membrane of Gram-negative bacteria. ExbB/D and MotA/B are also known to be homologous to TolQ/R, which play an important role in the maintenance of outer membrane stability. The energizing of these systems by proton movement across the inner membrane resembles the setup of the flagellum’s stator, which couples proton transport across the inner membrane to the motor’s rotation. Hughes and Blair also note that the rotor protein FliG is homologous to the magnesium transporter MgtE. The sequence identities of FliG and MgtE, however, appear to be rather weak (<20% as reported by PSI-BLAST), although there is clear similarity to the N-terminal domain of MgtE.
But is a demonstration of molecular homology of flagellar components to proteins found in other cellular organelles really an adequate defeater to the argument from irreducible complexity? I’m not convinced. Homology does nothing to demonstrate that the necessary transitions are evolutionarily feasible (Gauger and Axe, 2011), and it has been shown that the process of gene duplication and recruitment, as a source of evolutionary novelty, is extremely limited (Axe, 2010).
The central challenge posed by irreducible complexity is that functional utility is separated by discontinuous leaps in complexity, which cannot be scaled by a blind search. It is the need for multiple, coordinated changes that delivers a substantive challenge to neo-Darwinian evolutionary theory. This challenge stands regardless of whether sub-components within the flagellar apparatus can serve functions in other organelles.
A further point that is worth noting is that certain irreducibly complex subsystems in flagella are also found in other organelles, and perform essentially identical roles. Pointing to such homologies, therefore, only succeeds in pushing the problem back a stage — at some point, these systems require explanation. Take, for example, the proteins FliK and FlhB, which function in hook-length determination. FliK is critical for both the ability to switch and export filament and the hook-length control. fliK mutants result in polyhooks without filaments, while polyhook-filaments result from second-site suppressor mutations in FlhB (Williams et al., 1996). The hook-length determining protein FliK serves as a “secreted molecular ruler” (Erhardt et al., 2011) inasmuch as it “takes measurements of rod-hook length while being intermittently secreted through the assembly process of the HBB [Hook-basal-body] complex and the number of secreted FliK ruler molecules per time it takes to complete the HBB defines the ultimate length of the flagellar hook” (Erhardt et al., 2010). FliK and FlhB are homologous to YscP and YScU respectively, which also regulate substrate specificity and needle-length of the type III secretion system in Yersinia (Wood et al., 2008; Edqvistet al., 2003).
Moreover, there are a number of flagellar components that are presently not known to have homologs in non-flagellar systems. Examples include the rod cap FlgJ, the L and P ring proteins FlgH and FlgI, the MS ring-rod junction protein FliE, the filament capping protein FliD, and the anti-sigma factor FlgM. A number of these components form part of irreducibly complex subsystems of the flagellum. Let’s consider these proteins in turn. Consider, first, the rod cap FlgJ. The C-terminal domain of FlgJ possesses peptidoglycan hydrolysing (muramidase) activity, which is important for rod formation (Nambu et al., 1999). This is because the peptidoglycan layer has to be locally digested to allow for the penetration of the rod. In fact, experiments with mutant FlgJ proteins with a broken C-terminal (muramidase) domain result in a basal body lacking the L ring and hook (Hirano et al., 2001). Penetration of the peptidoglycan layer is a prerequisite for rod formation (Dijkstra and Keck, 1996; Fein, 1979). The C-terminal muramidase domain of FlgJ, however, may be dispensable, since the muramidase domain is absent in homologs of FlgJ in some bacterial phyla, and it appears that the rod structure “occasionally and fortuitously finds a hole in the peptidoglycan layer by chance; when it does so, it proceeds to assemble a flagellum that includes the outer membrane L ring,” (Hirano et al., 2001).
FlgJ also possesses binding affinity for rod proteins (Hirano et al., 2001). It is thought that FlgJ is the first protein to assemble since its N-terminal serves as a capping protein for the rod structure. Besides the genes that encode the rod subunit proteins (FlgB, FlgC, FlgF, and FlgG), more than 10 genes are necessary for rod assembly (Kubori et al., 1992). One study reported that “In the enteric bacteria, flgJ null mutants fail to produce the flagellar rods, hooks, and filaments but still assemble the integral membrane-supramembrane (MS) rings. These mutants are nonmotile” (Zhang et al., 2012).
The FlgH and FlgI proteins, which make up the outer (LP) ring complex are essential in Gram-negative bacteria, although the L and P rings are obviously not present in Gram-positive bacteria (which possess only a single membrane). In the absence of the LP ring in Gram-negative bacteria, the hook’s “proximal part formation occurs, but elongation does not occur” (Kubori et al., 1992).
The filament cap protein FliD is also an indispensable component. FliD migrates outwards as flagellin monomers are progressively added. The first and second hook-filament junction proteins remain in place and serve to connect the hook to the filament. Of the capping proteins involved in construction of the rod, hook and filament (FlgJ, FlgD and FliD respectively), only FliD remains at the tip of the filament in the finished product. One study, using electron microscopy, examined the structure of the cap-filament complex and isolated cap dimer, reporting that “five leg-like anchor domains of the pentameric cap flexibly adjusted their conformations to keep just one flagellin binding site open, indicating a cap rotation mechanism to promote the flagellin self-assembly. This represents one of the most dynamic movements in protein structures” (Yonekura et al., 2000). FliD is critical to filament assembly. Without the presence of FliD, the flagellin monomers are lost (Kim et al., 1999). As one paper explained, “A FliD-deficient mutant becomes non-motile because it lacks flagellar filaments and leaks flagellin monomer out into the medium” (Ikeda et al., 1996).
Finally, the anti-sigma FlgM forms a component of another irreducibly complex system, for it plays a crucial role in the timing of expression of flagellar genes. Obviously, it makes no sense to assemble the flagellin monomers before completion of the hook-basal-body construction. The expression of the hook-basal-body subunits and associated regulators (encoded by 35 genes spread across eight operons) is under the control of the class II promoters. This includes the regulatory genes fliA andflgM. The former codes for a sigma factor, σ28, which is required to activate the class III promoters; the latter codes for the anti-sigma factor FlgM, which prohibits σ28 from interacting with the RNA polymerase holoenzyme complex (Saini et al., 2011; Ding et al., 2009; Kutsukake et al., 1994;Ohnishi et al., 1990).
Upon completion of the hook-basal-body, the anti-sigma factor FlgM is secreted through the flagellar structures that are produced by the expression of the class II hook-basal-body genes. The class III promoters (which are responsible for the expression of flagellin monomers, the chemotaxis system and the motorforce generators) are then activated by σ28 and the flagellum can be completed. There is, of course, variation on this flagellar setup from species to species. Indeed, “these systems differ from each other by the existence of specific sigma factors and transcriptional activators, by motive force and the efficiency of motors” (Soutourina and Bertin, 2003). A component that is absent in one system, however, does not necessarily entail its redundancy in another system. Liu and Ochman (2007) explain that,
“The absence of certain of these genes from a genome is understandable once the characteristics of the particular bacteria are considered. For example, the L and P ring proteins FlgH and FlgI are not necessary in the Firmicutes because these bacteria lack the outer membrane in which these proteins are typically situated in Gram-negative bacteria. FlgH and FlgI are also not necessary in Spirochaetes, which have a periplasmic flagellum located inside of the outer membrane.”
Hughes and Blair also observe the flagellum’s relatedness to the type III secretion injectisome, noting that,
“It is not yet clear whether the flagellum or the injectisome came first or whether both are descended from something else that either has been lost or is present only in niches not yet explored by genome sequencing.”
Although there is room for debate, I am personally persuaded that the type III secretion injectisome evolved from the flagellar subunit exporters, largely based on, among other things, two key observations:
- While type III injectisomes are found only in gram-negative bacteria that possess both an inner and outer membrane, flagella are found also in gram-positive bacteria that possess only a single membrane: Flagella are thus more widely distributed.
- Multi-cellular organisms on which type III injectisomes are used emerged on the scene much later than bacteria, so evolutionary pressure for such an organelle would proceed the evolutionary pressure for motility.
“In any case,” argue Hughes and Blair,
“it is clear that export is the essential function carried out by both systems; before the flagellar rotor could have served as a useful organelle for motility, it would have provided an exterior appendage that conferred some useful function, such as (for example) adhesion.”
Hughes and Blair argue that “such an ancestor need not have been as complex as the present-day flagellar rotor or the injectisome,” since “it would not have required homologs of the three proteins FliH, FliI, and FliJ, which are dispensable for flagellar export,” and which were “presumably borrowed from the ATP synthase” as a “later refinement.”
I would concede that the chaperone protein FliJ, the ATPase FliI, and its regulator FliH are dispensable for flagellar export. Indeed, Minamino et al. (2000) report that “spontaneous mutants with defects in the 147-amino-acid Salmonella protein FliJ…have the ability to form swarms on motility agar plates after prolonged incubation at 30°C; i.e., they display a leaky motile phenotype.” They also observe that “At 42°C, however, export was inhibited, indicating that the mutant FliJ protein was temperature sensitive.” FliI is an ATPase, which converts the energy produced by hydrolyzing ATP into the energy required for protein export. fliI mutants, however, can still retain functional flagella (Paul et al., 2008; Minamino and Namba, 2008). The same is true of its regulator, FliH (Minamino et al., 2003).
Hughes and Blair further argue that flagellar protein export can be accomplished by only a subset of the ten proteins once thought to be indispensable (FliF, FliG, FliM, FliN, FliO, FliP, FliQ, FliR, FlhA, and FlhB). According to Hughes and Blair, the export apparatus’s “functional core” is comprised of just FliP, FliQ, FliR and FlhA. I would be interested in seeing the literature that Hughes and Blair are drawing on at this point (no citation is offered). My own reading would certainly indicate that all of the above proteins are indispensable for flagellar protein export, with the exception of FliO (which is missing in some systems — see Barker et al., 2010). In the case of FliF, flagellar assembly requires “a short stretch of amino acids at the immediate C terminus” (Grünenfelder et al., 2003). Mutant studies on FliG, FliM and FliN (which are present in 26, 34 and over 100 copies respectively in E. coliand Salmonella) demonstrate that they are also essential for flagellar assembly (Brown et al., 2007,Paul et al., 2006, Tang and Blair, 1995).
To conclude, the claim of Hughes and Blair to have refuted Behe on the bacterial flagellum is unfounded. Although there are sub-components of the flagellum that are indeed dispensable for assembly and motility, there are numerous subsystems within the flagellum that require multiple coordinated mutations. The flagellar motor is not the kind of structure that one can at all readily envision being produced in Darwinian step-wise fashion.