BIOC3390 Tutorials: 16 & 18 November 1999

Flagella in E. coli

Gram negative bacteria (such as E. coli and S. typhimurium) have 3 distinct surface layers: inner and outer phospholipid membranes, between which lies the rigid peptidoglycan cell wall. The space between the inner and the outer membranes is called the periplasmic space. The image on the left was constructed from superimposed EM pictures, and the contours show image density measurements.

The torque generating assemblies consist of the S, M and C rings. These protrude from the inner membrane into the cytoplasm, and are anchored by ten studs to the peptidoglycan layer. Imediately outside the cell is a flexible protein "hook" which functions as a universal joint. This allows the flagellum to adopt any orientation with respect to the cell wall. Between the hook and the motor there is a rigid drive shaft which rotates within molecular bearings located in the bacterial membranes: the L ring in the outer membrane and the P ring in the peptidoglycan layer. The walls of gram positive bacteria lack both outer layers, and in these species the L and P rings are also absent.

In addition to its impressive rotation speed, each flagellar assembly is hollow. It is synthesised from the inside outwards, and most of the new protein subunits are transported down the hollow core and then fitted into place at the growing tip. The protein component supply changes during flagellar synthesis as various parts of the structure are completed. It is a bit like building a tall chimney by hauling pre-fabricated components up the flue.

A typical E. coli or S. typhimurium cell has about 6 flagellae, which are about 10nm long. In these species the maximum rotation speed is about 18,000 rpm. Most of the time they all rotate in a counter-clockwise direction. This allows the individual rotating corkscrews to mesh comfortably together, forming a tight bundle which is an effective propeller. About once per second one or more of them changes direction and rotates briefly in a clockwise sense. This does not mesh with its neighbours and causes the bundle to fly apart, so that the bacterium tumbles randomly, and adopts a new direction.

This behaviour forms part of a strategy for finding food and avoiding noxious compounds. Bacteria are too small to detect any concentration gradients from one end of the cell to the other. Their only method for deciding which way to swim, is to move a short distance at random, and then check to see whether things have improved or not. If the situation is getting better then they keep going a little longer in the current direction, but if things are worse then they tumble and try a new tack.

(Eukaryotic cells are much larger and are able to detect tiny concentration gradients from one side of the cell to another. [See Parent, CA & Devreotes, PN (1999) A cell's sense of direction. Science 284(5415), 765-770.] Eukaryotes consequently know which way to swim, but they use larger dynein-based cilia and flagella and a different internal signalling system.)

Eukaryotes can measure spatial concentration gradients, but bacteria measure concentration versus time, and then equate time with distance moved. They can remember the previous situation for about 3 seconds. We know this from rapid mixing experiments, where added attractants briefly suppress tumbling behaviour, even though there is no gradient to follow. There is no point in bacteria trying to swim in a "straight" line for more than a few seconds, because random Brownian motion will eventually knock them off course, whether they tumble or not.

Read through the following section which lists the main components and describes the signalling system. This has been reasonably well understood for the last ten years. Your task, however, is to do more work on the motor, where many questions remain unanswered.

Bacteria can follow more than one gradient at the same time, so they must have more than one type of detetector. They can also respond to concentrations of attractants spanning five orders of magnitude. This requires some method to adjust the sensitivity of the detector systems so that they do not become swamped at very high ligand concentrations.

The central components in the bacterial chemotaxis system are four varieties of methyl-accepting chemotaxis proteins or MCPs. All four proteins span the bacterial inner membrane. Their conformation changes from a clockwise signalling [CWS] to a counter-clockwise signalling [CCWS] state after attractant binding, either directly, or via one of the periplasmic receptor proteins. It should be noted that the effector ligand does not enter the bacterial cell cytoplasm in order to exert its effects. The ligand binding affinities of the MCPs can be adjusted by multiple methylation of glutamate residues within a cytoplasmic domain of these proteins.

Many components of the bacterial chemotaxis system were identified from mutants lacking various features of the wild type system. About 50 genes are involved, some of which are listed in the following table:




effect of gene deletion

MCP I (tsr)

inner membrane

binds serine (attractant), or leucine (repellant) directly

no response to either serine or leucine, remainder OK

MCP II (tar)

inner membrane

binds aspartate, the maltose receptor protein (attractants) and Ni, Co (repellants)

no response to aspartate, maltose, Ni or Co, but all other systems operate OK

MCP III (trg)

inner membrane

binds the galactose and ribose receptor proteins

no response to galactose or ribose, all other systems OK

MCP IV (tap)

inner membrane

binds the dipeptide receptor protein

no response to dipeptides, but all other systems are OK

maltose receptor


binds maltose to MCP II

no response to maltose, but all other systems are OK

galactose receptor


binds galactose to MCP III

no response to galactose, but all other systems are OK

ribose receptor


binds ribose to MCP III

no response to ribose, but all other systems are OK

dipeptide receptor


binds dipeptides to MCP IV

no response to dipeptides, but all other systems are OK

Che A


phosphorylates itself if MCP empty, then Che Y & Che B

counterclockwise bias: swim smoothly, cannot tumble

Che B


MCP methyl esterase when phosphorylated

insensitive, no adaptation, no memory

Che R


MCP methylator (uses SAM)

hypersensitive, no adaptation

Che W


connects MCPs to Che A

counterclockwise bias

Che Y


motor turns clockwise if this protein is phosphorylated

counterclockwise bias

Che Z


Che Y phosphatase

incessant tumbling

Mot A

inner membrane

proton channel (4 crossings) part of stator creates torque

flagellae assembled but do not turn

Mot B

inner membrane

proton channel (1 crossing) part of stator creates torque

flagellae assembled but do not turn


M & S rings

major rotor component

incomplete flagellae


"M" ring

part of rotor: creates torque? causes direction switching?

incomplete flagellae


"C" ring

part of rotor?

incomplete flagellae


"C" ring

part of rotor?

incomplete flagellae

FlgB, FlgC and FlgG

minor rod components

parts of the hollow drive shaft (6 copies each per rod)

incomplete flagellae


major rod component

part of the hollow drive shaft (about 26 copies per rod)

incomplete flagellae


P ring

peptidoglycan bearing

incomplete flagellae


L ring

outer membrane bearing

incomplete flagellae



universal joint

incomplete flagellae


hook tip

scaffolding: controls hook assembly, then discarded

incomplete flagellae

FlgK [HAP1]

hook tip

caps hook, joins to flagellin

incomplete flagellae

FlgL [HAP3]

hook tip

caps hook, joins to flagellin

incomplete flagellae


flagella spiral

major structural component

incomplete flagellae


caps flagella

prevents loss of subunits

new subunits drop off

Conversion of an MCP into the CWS state results in the activation of the Che A protein [via the Che W adaptor protein]. Che A is an autokinase which initially phosphorylates itself using ATP, but then proceeds to phosphorylate two other regulator proteins: Che Y and Che B. Remember that both of these proteins are phosphorylated when the attractant concentration FALLS.

Phosphorylated Che Y interacts with the motor control system at the base of the flagellum, and increases the probability of clockwise rotation and tumbling.

Che Z, promotes the de-phosphorylation of Che Y, thereby terminating this part of the response. If the above mechanism were the whole story it would merely inhibit tumbling at high attractant concentrations. It has no memory and would not be sufficient for the bacteria to follow a gradient. The memory is supplied by the second regulator, Che B.

Che R continuously methylates the MCPs, using S-adenosyl methionine as the methyl donor. It methylates multiple glutamate residues in both the CWS and CCWS states and progressively reduces the attractant binding affinity of the MCPs.

Phosphorylated Che B is a methyl esterase that removes these methyl groups from the MCPs. Removal of the methyl groups from the MCPs restores their high binding affinities for their attractants.

There are two effects of this arrangement.

The first effect is that MCP methylation continuously tracks the effector concentration. At low attractant concentrations, or when a repellant is bound, most of the MCPs are in the CWS state. Under these conditions Che A is activated, and Che B removes most of the methyl groups. In this state the MCPs have high affinities for the attractants and are able to detect very low concentrations of these effectors. If the attractant concentration increases, the MCPs tend to adopt the CCWS conformation, Che A and Che B thereby lose their activities, and methyl group removal becomes progressively less efficient.

With increasing attractant concentrations the MCPs are progressively converted into the fully methylated state with a low affinity for the attractants. At a constant intermediate ligand concentration the system reaches a steady state, where there is just sufficient residual Che B activity to prevent full methylation and keep the MCPs partially methylated and partially loaded with attractant. Although the percentage of the MCPs in the CCWS state increases in proportion to the average ligand concentration, the negative feedback tracking system opposes this effect and spreads these changes over a much greater concentration range.

The second effect is that the system has memory. It takes the tracking system a few seconds to catch up with any alteration in the ligand concentrations, and during that period the MCPs generate very large output signals via Che A. Instead of signalling absolute concentrations of the attractants, the detector systems have first derivative outputs, and signal changes in the attractant concentrations. This vital information enables bacteria to follow the concentration gradients, and is important for cell growth and survival.

In addition, the tracking system greatly extends the dynamic range of the MCPs as ligand detectors. The changing receptor affinities allow these proteins to respond to small fractional changes in attractant concentration, over a very wide range of absolute ligand concentrations.


Your presentation should concentrate on the motor, rather than the signalling system, but the reversal mechanism would be relevant. We want to hear about torque, thrust, work output, step size, proton or sodium stochiometry and efficiency. What are the unsolved problems, and how might the motor work? Draw some diagrams. Make sure that you read the two reviews on bacterial chemotaxis mentioned in the general introduction. In addition, the following papers may be helpful:

Aizawa, SI (1996) Flagellar assembly in Salmonella typhimurium Molecular Microbiology 19(1), 1-5. [This journal is available electronically. You must quote your normal BIDS username and password. Write down the volume and page numbers, which you need later, before clicking HERE to view the electronic copy.]

Berg, HC & Turner, L (1993) Torque generated by the flagellar motor of Escherichia coli. Biophys. J. 65, 2201-2216. [Click HERE for the abstract.]

Berry, RM & Berg, HC (1997) Absence of a barrier to backwards rotation of the bacterial flagellar motor demonstrated with optical tweezers. Proc. Natl. Acad. Sci. USA 94(26), 14433-14437. [Click HERE for the HTML version, or HERE for PDF format.]

Berry, RM & Berg, HC (1999) Torque generated by the flagellar motor of Escherichia coli while driven backwards. Biophys. J. 76, 580-587. [Click HERE for the HTML version, or HERE for PDF format.]

Bourret et al (1991) Signal transduction in prokaryotes Annual Rev. Biochem. 60, 401-441.

Hughes, MP & Morgan, H (1999) Measurement of Bacterial Flagellar Thrust by Negative Dielectrophoresis Biotechnol. Prog., 15(2), 245 -249. [Click HERE for the HTML version, or HERE for PDF format.]

Lloyd et al (1999) Structure of the C-terminal domain of FliG, a component of the rotor in the bacterial flagellar motor Nature 400, 472 - 475. [Write down the volume and page numbers for future reference. Click HERE for access instructions. Click HERE to gain access via the Leeds University network.]

Thomas et al (1999) Rotational symmetry of the C ring and a mechanism for the flagellar rotary motor Proc. Natl. Acad. Sci. USA 96(18), 10134-10139. [Click HERE for the HTML version, or HERE for PDF format.]

Most electronic journals check your browser address and require an institutional subscription to the printed version of the journal. Consequently, many of the hypertext links in this reference list will only work from a university campus computer.

Site Navigation:
or use the buttons below.


If you have comments, queries or suggestions, email me at