The Bacterial Flagellar Motor Continues to Amaze

Main Text

The flagellar motor from Escherichia coli has a number of remarkable features. It can rotate at ∼300 Hz and is driven by proton motive force. With constant proton motive force, the motor can switch its sense of rotation from counterclockwise (CCW) to clockwise (CW) in response to interaction with the phosphorylated form of the response regulator CheY (CheY∼P). The transition from CCW to CW rotation is a highly cooperative process with a Hill coefficient of ∼20, with the midpoint of the transition (K_1/2) occurring near 3 mM CheY∼P (1). CheY∼P interacts with the motor by binding to FliM and FliN, two components of the C-ring of the motor (2). Recent work has shown that the motor can adapt by changing the number of FliM and FliN protein subunits in the motor, with resulting changes in the value of K_1/2 for CheY∼P, while maintaining the steeply cooperative nature of the transition from CCW to CW (3). In this issue of the Biophysical Journal, Hosu and Berg use homo-Förster resonance energy transfer (homoFRET) to examine the structural consequences of changes in the number of FliN and FliM subunits that participate in the assembled motor.

Torque is exerted on the rotor component of the motor by the action of the stator proteins MotA and MotB acting on the FliG protein of the rotor. FliG is associated with two other rotor components, FliM and FliN. Pioneering cryotomography from the DeRosier laboratory showed that FliG, FliM, and FliN are arranged in a ring, with rotational symmetry suggesting ∼26 copies of FliG and ∼34 copies each of the FliN tetramer and the FliM monomer (4). This symmetry mismatch has been a longstanding conundrum in the field. That conundrum has only deepened with the observation that the number of FliM and FliN tetramers is not fixed but can increase to 45 in fully CW motors, whereas the number of FliG subunits appears to be constant (5, 6). A recent study suggests a plausible model to explain the variable number of FliM/FliN subunits while FliG is held constant (7).

A typical E. coli cell has six to eight motors. When all are operating in the CCW mode, the cell is propelled in a relatively straight “run.” When one (or more) motor(s) reverses its sense of rotation to CW, the cell undergoes a turning maneuver or “tumble.” In a constant environment, the typical adapted cell is seen to move with a mixture of runs and tumbles, implying that the adapted level of CheY∼P is maintained very close to the transition midpoint concentration of CheY∼P. This allows the cells to effectively diffuse with runs to explore new territory while being poised to undergo more tumbles to test new directions if the CheY∼P level increases as a result of moving into a less favorable environment or to increase runs if the CheY∼P drops by moving into a more favorable environment.

The level of CheY∼P is established by the combined activities of the chemotaxis histidine autokinase, CheA, and the CheY∼P phosphatase, CheZ. As a result of receptor-mediated adaptation, CheA kinase activity is regulated by chemoreceptors that form large membrane-associated arrays with CheA and a coupling protein, CheW. The chemoreceptors control CheA activity as a result of the opposing effects of chemoattractant ligand binding and methylation of certain glutamate residues of the chemoreceptors. Methylation by the methyltransferase CheR stabilizes the active form (ON) of the kinase, whereas binding of chemoattractant ligands stabilizes the inactive form (OFF) of the kinase. Removal of methyl groups by the methylesterase CheB destabilizes the kinase ON state (8, 9). Feedback control of the kinase activity is provided, since CheB activity is regulated by CheA phosphorylation of CheB. This system provides transient changes in the CheY∼P level in response to transient changes in chemoattractant concentration and then returns CheY∼P to its prestimulus level as the cell adapts to a constant but different chemoattractant concentration (10). It is easy to imagine that small changes in the levels of CheA, CheZ, and the other components could change the concentration of CheY∼P generated by adapted cells. Given the steep nature of the CCW/CW transition, even small changes in the adapted level of CheY∼P concentration could result in cells that are fully tumbly or fully smooth rather than the desired mixture of smooth and tumbly. Either situation would hamper the cell’s ability to effectively respond to temporal stimuli.

The adaptation of the motor is a reflection of the fraction of time the motors spend in the CCW or CW state rather than the concentration of CheY∼P per se. Predominantly CCW motors add FliM and FliN subunits, whereas predominantly CW motors remove them. Addition of FliM and FliN moves the K_1/2 for CheY∼P to lower concentrations, making CW rotation more likely and driving the system toward a suitable smooth/tumbly ratio (3). Thus, motor adaptation makes great biological sense by maintaining a favorable smooth/tumbly ratio and potentially tuning all the motors in a given cell to that cell’s current adapted level of CheY∼P.

The notion that continuously variable numbers of subunits in the flagellar motor can be accommodated while maintaining the cooperativity and functionality of the system is counterintuitive to those of us who have marveled at the exquisite precision of the structural changes that give rise in cooperative interactions in well-studied systems with fewer protein subunits. How can this occur?

Hosu and Berg provide the first glimpses of the potential structural changes associated with the addition of FliM and FliN to the motor during motor adaptation. They used yellow fluorescent protein (YFP) fused to either FliN or FliM and then measured the homoFRET between the YFP moieties in the assembled motor by monitoring the YFP fluorescence anisotropy under predominantly CCW or CW conditions. HomoFRET can occur in a structure where multiple copies of a chromophore such as YFP are within the Förster distance of each other, so long as they have different static orientations relative to the laboratory frame. In such a case, polarized light selectively excites certain transition dipoles, energy transfer occurs between the chromophores, and fluorescence is emitted from the different orientations of emission dipoles, resulting in decreased fluorescence anisotropy even if the chromophores are completely static. To distinguish homoFRET from other sources of fluorescence anisotropy, Hosu and Berg used photobleaching to reduce the number of active chromophores in a given motor. They show that fluorescence anisotropy increases with photobleaching for both FliM-YFP and FliN-YFP in motors, as would be expected if the average distance between chromophores increased with photobleaching with concomitant reduction in homoFRET between the chromophores.

The observed fluorescence anisotropy of FliM-YFP does not appear to change in CW or CCW motors. This suggests that the FliM units in the CCW motor are organized with the same intersubunit distance as those in the smaller CW motor. This could be accomplished by increasing the number of FliM subunits with a corresponding increase in the symmetry of the FliM in the C-ring.

The observed fluorescence anisotropy of FliN-YFP does show changes between CW and CCW motors, suggesting increases in the effective spacing or relative orientations between the FliN monomers. The FliN tetramer is organized as a dimer of dimers. In principle, homoFRET changes could reflect reorganizations with a given dimer, within the tetramer, or between tetramers, and the authors discuss the various alternative explanations.

The new work by Hosu and Berg raises fascinating new questions, and we look forward to the next chapters revealing the secrets of the bacterial flagellar motor.

Editor: Dennis Bray.