One Basic Blueprint, Many Different Motors

The cytoplasmic C ring of the bacterial flagellum is known as the switch complex. It binds the response regulator phospho-CheY to control the direction of flagellar rotation.

ABSTRACT

The cytoplasmic C ring of the bacterial flagellum is known as the switch complex. It binds the response regulator phospho-CheY to control the direction of flagellar rotation. The C ring of enteric bacteria is well characterized. However, no Gram-positive switch complex had been modeled. Ward et al. (E. Ward, E. A. Kim, J. Panushka, T. Botelho, et al., J Bacteriol 201:e00626-18, 2019, https://doi.org/10.1128/JB.00626-18) propose a structure for the Bacillus subtilis switch complex based on extensive biochemical studies. The work demonstrates that a similar architecture can accommodate different proteins and a reversed signaling logic.

COMMENTARY

The study by Ward et al. (1) published in this issue used two-hybrid protein interaction (bacterial adenylate cyclase two-hybrid [BACTH]) experiments, cysteine-mediated protein cross-linking, and analysis of mutant phenotypes to infer the structure of the cytoplasmic C ring of the flagellar switch complex of Bacillus subtilis. It can accurately be described as heroic. It is also a tribute to its senior—in the true sense of the word—author, George Ordal, who bravely persevered with the study of chemotaxis and motility in B. subtilis when everyone else focused on Escherichia coli and Salmonella enterica.

The indirect inductive approach was necessitated by the lack of a protocol for isolating intact flagellar basal bodies from B. subtilis and the difficulty in carrying out electron cryotomography on such large cells. The well-characterized architecture of the C ring in the enteric bacteria, based on direct electron microscopic observation of both isolated and in situ flagellar basal bodies, provided a blueprint for reconstructing the B. subtilis switch complex. The structure that emerges is familiar, with the main framework composed of many subunits of the three-domain FliG and FliM proteins organized in a cylindrical configuration, as in the enteric motor. However, the replacement of the small FliN protein by the much larger FliY protein introduces profound changes. The C ring must have a larger diameter in B. subtilis, the FliY and FliM proteins are intimately intertwined at their C termini, and the FliY protein provides a built-in, motor-localized phospho-CheY (CheY-P) phosphatase.

The proposed model will serve at least two important roles. First, it will inform follow-up studies of the detailed architecture of Gram-positive flagellar motors, which have hitherto been poorly visualized. Second, it will provide insight into how the basic logic of chemotaxis signaling can be modified during evolution. The signaling pathways in Gram-negative enteric bacteria and Gram-positive B. subtilis involve many of the same players. However, the molecular mechanisms that enable chemotaxis differ in fundamental ways that may be shaped by different environmental pressures and that certainly reflect different evolutionary histories.

The communication between the sensory and motility systems is the generation of CheY-P by the CheA kinase at the receptor array and the binding of CheY-P to the switch complex of the flagellar rotary motor (2). Binding of CheY-P changes flagellar rotation from its default direction—which, in different species, can be either clockwise (CW) or counterclockwise (CCW)—to the opposite direction. In peritrichously flagellated bacteria, one sense of rotation allows the helical flagellar filaments to form a bundle that propels the cells in a relatively straight path called a run. Reversal of the direction of rotation by one or more flagella disrupts the bundle and causes an abrupt change in heading known as a tumble. In both the enteric bacteria and B. subtilis, the filaments are left-handed helices, so CCW rotation leads to running and CW rotation leads to tumbling. An alternation of runs and tumbles in a swimming cell caused by small fluctuations in intracellular levels of CheY-P (3) produces a three-dimensional random walk. There are several dozen CheY-P binding sites per switch complex. Although the binding is not cooperative (4), there is a very steep dependence of the rotational state on the occupancy by CheY-P (5), suggesting an allosteric coupling among CheY-P binding sites around the C ring (6). Thus, the flagellum changes its direction of rotation almost instantaneously and reaches full speed, which is the same either CW or CCW, extremely rapidly.

Cells carry out chemotaxis by biasing the random walk to extend runs in the favorable direction by suppressing tumbles. E. coli and B. subtilis both do this, but in diametrically different ways. In E. coli, the default direction of rotation of the motor in the absence of CheY-P is CCW, and the default behavior is running. Binding of CheY-P to the motor increases the probability of switching to CW rotation so that the cells tumble. Attractants inhibit CheA activity, resulting in a suppression of tumbles. Repellents increase CheA activity, resulting in tumbling, but removal of repellents inhibits CheA activity to suppress tumbling. In B. subtilis, the default direction of rotation of the flagellar motor in the absence of CheY-P is CW, and the default behavior is tumbling. Binding of CheY-P to the motor promotes CCW rotation, leading to an increase in the length of runs.

CheA and CheY are shared by all chemotactic bacteria. The composition of the switch complexes, however, shows considerable variation among highly divergent bacterial species (7), as shown in Fig. 1. The C rings of E. coli and other enteric bacteria contain three proteins: FliG, FliM, and FliN. FliG and FliM are found in the switch complexes of all the bacteria whose flagellar motors have thus far been characterized. They both contain three structural domains (N, M, and C, standing for N terminal, middle, and C terminal, respectively), but they are otherwise dissimilar. FliG_N attaches the C ring to the flagellar basal body by binding to FliF, which oligomerizes to form the MS ring (membrane and supramembraneous ring) of the flagellar basal body (see Fig. 1 in the article by Ward et al. [1]). CheY-P binds to the short FliM_N domain and then interacts with FliN, and perhaps FliM_M, to transmit conformational changes to FliG_M, which is in intimate contact with FliM_M (8). FliG_C, which interacts directly with the ion-conducting MotA₄MotB₂ stator complexes (9), rests upon FliG_M (10). Changes in the conformation of FliG_M are transmitted to FliG_C to alter the way it interacts with the cytoplasmic loop of MotA to change the direction of flagellar rotation (11).

FIG 1.

Schematic representation of the switch complex and the soluble phosphatases in four distantly related bacterial species. All four contain similar FliG and FliM proteins that make up the core of the C ring. E. coli also contains the small FliN protein in its switch complex and has the soluble CheY-P phosphatase CheZ. H. pylori has the same arrangement but in addition includes the FliY protein, which, however, lacks the FliY_N domain and phosphatase activity. B. subtilis possesses neither FliN nor CheZ and replaces them with an unrelated soluble phosphatase, CheC, and a FliY protein that possesses two active phosphatase sites in its middle domain that are related to those of CheC. T. maritima has both CheC and a phosphatase-competent FliY, in addition to a second soluble phosphatase, CheX, which is also related to CheC. Domains that share evolutionary ancestry are shown in the same color: gray for CheZ, yellow for the CXY family of domains, blue for FliN-related domains, navy for CheY-P binding N-terminal domains, pink for the FliG_M domain, and purple for the FliG_C domain. The FliG_N domain, which interacts with the MS ring, is not shown. The location of each GXXN active site for phosphatase activity is indicated with an asterisk. (Adapted from reference 7 with permission of the publisher.)

FliN has a fold similar to that of the FliM_C domain and has presumably evolved from a common ancestor (12). FliN interacts with FliM_C in a stoichiometry of about 4 to 1 to form the membrane-distal lower edge of the C ring. FliN plays a role in the switch function of FliM. It has the additional role of interacting with the FliH protein of the protein export apparatus to position and stabilize the FliI ATPase within the C ring.

The B. subtilis switch complex does not contain a homolog of FliN. Instead, it contains the much larger FliY protein, which has a three-domain structure much like that of FliM, with N, M, and C regions that have folds similar to those of their counterparts in FliM. The single FliY_N helix is similar to FliM_N, which binds CheY-P, although it is uncertain whether FliY_N binds CheY-P. However, the functions of the FliM_M and FliY_M domains are very different. Rather than interacting with FliG_M, like FliM_M does, FliY_M contains two active sites for phosphatase activity against CheY-P. The more C-terminal of the two sites has a higher activity. FliY_N seems to be largely dispensable for the phosphatase activity of FliY_M and is in fact missing from the FliY homolog of Helicobacter pylori (Fig. 1). However, the domain corresponding to FliY_M in H. pylori also lacks phosphatase activity. FliY_C has a fold similar to that of FliM_C and FliN, and it shares with FliN the ability to interact with FliH. It also can form heterodimers with FliM_C. Notably, expression of B. subtilis FliY_C in a nonmotile fliN knockout mutant of S. enterica can partially complement the defect, probably because FliY_C can serve as a FliN mimic (13).

The presence of an additional large protein, FliY, requires that the C ring of B. subtilis have a significantly greater diameter than the C ring of enteric bacteria. This larger diameter is visible in the electron cryomicrographs of the FliY-containing C ring of the spirochete Leptospira interrogans (14). In the model by Ward et al. of the B. subtilis switch complex (see Fig. 1 of reference 1), the bulky FliY_M domains are arranged peripherally around a ring of FliM_M domains, with intertwined FliM_C/FliY_C heterodimers forming the bottom edge of the ring. As cross-linking studies also indicate that there are FliY_C homodimers, it is possible that some FliY_M domains are also incorporated into the FliM_M ring, although seemingly without interacting with FliG_M. The overall picture that emerges from the study is that the fundamental organization of the switch complex is similar to that in the enteric bacteria but with considerably greater elaboration to accommodate the FliY_M domains that provide the phosphatase activity at the B. subtilis motor.

The structure and location of the phosphatases are among the most variable aspects of chemotaxis signal transduction systems (Fig. 1). In E. coli, the CheY-P phosphatase CheZ is a soluble protein that can also associate with the chemoreceptor patch through its interaction with CheA_short, an N-terminally truncated variant of CheA that is translated from an internal start site. H. pylori contains FliY in its switch complex, but H. pylori possesses a soluble CheZ phosphatase (19), and its FliY does not possess phosphatase activity. B. subtilis has two phosphatases: CheC, which is associated with the receptor patch and activated as a phosphatase through its interaction with the receptor deamidase CheD, and FliY_M within the switch complex. CheC and FliY_M both have two phosphatase active sites, and their overall protein folds are similar, suggesting a common evolutionary origin. Thermotoga maritima, a very deep-branching member of the Gram-negative bacteria, has an even more complicated arrangement, with two soluble phosphatases, CheC and CheX, and a FliY protein with a domain structure like that of B. subtilis. CheC, CheX, and FliY_M are similar enough in their folds and disposition of their active sites that they can be placed in one group, the CXY protein family. CheZ has a completely different fold, although the pair of Glu and Asn residues at the active site is the same as in the CXY phosphatases. It is noteworthy that E. coli and H. pylori have one soluble phosphatase active site, that B. subtilis has two soluble phosphatase active sites and two in its switch complex, and that T. maritima has three soluble phosphatase active sites and two in its switch complex. Perhaps regulation of phosphatase activity plays a bigger role in chemotaxis by the latter two species than in that by the former two.

Four questions of biological significance arise from this diversity of structures. The first is why some bacteria have FliY instead of FliN or, in the case of H. pylori, have both FliY and FliN. FliY is probably the ancestral form, and FliN likely arose through the segregation of the DNA region encoding FliY_C as a separate gene. It will be interesting to determine whether motors containing FliY can generate higher torque, perhaps as an adaption to swimming in highly viscous environments, gliding on surfaces, or having to turn within the confines of the periplasmic space. (Many spirochete flagellar motors contain FliY.)

The second question is whether the reversed logic of B. subtilis chemotaxis compared to that of the enteric bacteria has a selective advantage. In B. subtilis, the default motility pattern in the absence of CheY-P is tumbling, CheY-P levels increase in response to attractants, and higher levels of CheY-P cause smooth swimming. In enteric bacteria, the default motility pattern in the absence of CheY-P is smooth swimming, CheY-P levels decrease in response to attractants, and lower levels of CheY-P cause smooth swimming. A starving E. coli cell with low levels of ATP should have decreased levels of CheY-P and wander aimlessly, perhaps randomly finding its way to greener pastures. Would a starving B. subtilis cell with low levels of ATP, and thus diminished CheY-P levels, just tumble in one place, perhaps tearing itself loose from a nutrient-poor biofilm or substrate? Or would it simply disengage its motors, as B. subtilis does in a biofilm (15), in order to save energy?

The third question concerns the intracellular location of the chemotaxis phosphatases. In E. coli, CheZ either is located within the receptor patch or is free in the cytoplasm. Much of the CheY-P produced in the patch is dephosphorylated before it can diffuse away and interact with a flagellar switch. A consequence of the localization of CheZ at the receptor patch is that CheY-P concentrations are more evenly distributed within the cell. This situation may serve to synchronize the behavior of flagella that are different distances from the pole so that pole-proximal flagella are not always spinning CW and pole-distal flagella are not always spinning CCW. Because FliY is part of the switch complex in B. subtilis, CheY-P binding to FliM and binding to FliY are in direct and local competition. This arrangement might serve to dampen the response to changes in CheY-P, as higher concentrations of CheY-P would both increase binding to FliM to promote smooth swimming and increase the rate of dephosphorylation of CheY-P to achieve a sort of adaptation at the motor.

Finally, it has been reported that CheY-P is produced in a pulsatile fashion in E. coli (16) and that these fluctuations may help to generate correlated switching of multiple flagellar motors that may be separated by some distance along the cell wall (17). It will be interesting to see whether a similar situation obtains in B. subtilis.

It will be fascinating to model these two different variants of chemotaxis signaling to see how they affect steady-state motile behavior and responses to attractant and repellent gradients. One thing to keep in mind is that B. subtilis has five times more (∼25 versus ∼5) peritrichous flagella per cell than E. coli. Also, B. subtilis flagella are organized in a grid-like pattern over the lateral surface of the cell rather than being randomly placed (18). Do these differences in anatomy favor one or the other type of signal-response coupling? It will then be an even greater challenge to discover how any differences in responsiveness that may exist provide advantages in the lifestyles of the respective organisms. Ultimately, the goal will be to understand what ecological niches these two bacteria inhabit and what selective constraints exist in those various environments. They have had untold millions of years to optimize their chemotaxis behavior. A whole new field of bacterial ethology suggests itself. Who will do for E. coli and B. subtilis what Jane Goodall did for chimpanzees and George Schaller did for lions?

The views expressed in this article do not necessarily reflect the views of the journal or of ASM.