Bacterial Motility - flagella and nanotechnology
The diagram above shows a model of a close-up view of the basal components of a bacterial flagella.
These are the wheels at the roots of the flagella as mentioned in the introduction to bacteria, if you have
not read this introduction then
click here to learn the basics of bacterial structure. The diagrams below add
some labels to this structure. Note that only a tiny portion of the filament is shown here, since this is the
long helix-like propeller that we saw previously. Remember the flagellum rotates like a propeller to propel
the bacterial cell along, as shown by the arrows in the diagram below.
Above: the flagellum emerges from the bacterial surface or wall (or cell envelope)` which consists of
three main layers: the
outer membrane (OM), the periplasm (P) which contains a mesh of very strong
fibrous material called
peptidoglycan (or murein) and an inner cytoplasmic membrane (CM). (Click
here to learn about cell membranes). Note how different this arrangement is from an animal cell which
has a single cell membrane rather than this double membrane structure. Beneath these layers is the cell
interior (the intracellular compartment) or protoplast (consisting of cytoplasm and nucleoid) and external
to these layers is the extracellular (external) environment, such as the water the bacterium is swimming
in. (Additional layers may exist outside the OM, including a slime capsule, but we shall look at these
possibilities later). This type of wall structure is particular to some types of bacteria called
negative bacteria
, but other wall structures occur, as we shall see later. The root of the flagellum
consists of a series of rings (made of proteins) that anchor the structure in the cell envelope.
Above: the flagellar basal complex with some outer structures (Mot proteins) removed to show the
internal structure of the motor. The labels are shown in the figure below:
Above: the flagellar motor rotates, causing the flagellum to rotate (and the cell to rotate in the opposite
sense). Notice the scale bar: the line illustrates a real-life length of 20 nanometres (20 millionths of a
millimetre!) so we are dealing with a minute machine - a nanomachine! These minute electric motors
evolved by natural means, on Earth, long before human beings existed.
The basal structure consists of a series of rings connected by a rod, the rings are the L ring (embedded in
the lipid bilayer of the outer membrane), the P ring (embedded in the periplasm), the S and the M rings
(the rotating motor or rotor) and the C ring.
The L and P rings act as a bushing (a bushing is a ring-like
structure that constrains moving mechanical parts, in this case the rotating rod, and may also be lubricated
to reduce friction). The
motor proteins (Mot) conduct electric current carried by positively charged
protons ('positive electricity' as opposed to negative electricity in which the current is carried by negatively
charged electrons as in a metal wire) from the periplasm into the cell cytoplasm. The electric charge is
thought to flow into the M (motor) ring where it is converted into rotary mechanical motion, causing the
M-ring to rotate. The M ring is attached to the rod, causing the rod to rotate. The M ring acts against the
fixed Mot (
stator) ring, which rotates slowly in the opposite direction to the M ring - slowly because it is
fixed to the bacterial cell wall and so causes the whole bacterial cell to rotate in a direction opposite to the
M ring and rod. The S ring is now known to be part of the rotor, along with the M ring, and forms a socket
for the rod, but is not part of the stator. Confusion may arise when the stator ring is referred to as the

The rod is attached to the filament via a
flexible hook (which acts as a universal joint, transferring rotary
motion to the filament via the
hook associated proteins (HAPs)). The filament is actually much too long
to show more than a tiny segment of it in these diagrams, it is about 20 nanometres in diameter, but 10 -
15 micrometres (10 - 15 thousand nanometres) long, which is longer than a typical bacterial cell which is
about 2 micrometres long. The filament is made up of about 30 000 subunits of a protein called
and is a corkscrew or helix shape. (The flagellin is arranged into typically 11 strands that are twined
together). This shape is important, mutants with straight filaments are immotile - the filament is the
propeller driven by this remarkable microscopic electric motor! (The cell body may contribute to thrust in
some forms in which the body is also helical). The helical filament is hollow, and flagellin is transported
from inside the cell, through the C ring (which has a hole in its centre) and along the filament, in its hollow
core, to its tip to which they are added - the filament constantly grows, as it must do to compensate for
breakages. A cap protein forms the tip and stabilises the filament.
Above a diagram of the bacterial flagellum showing the detailed structure of the flagellum basal complex
(in a Gram negative bacterium). The units of protein FliG (about 25-45 units) form a ring extension to
the M-ring (the M ring is shown in section here)
or an extension on top of the C-ring according to the
. Units of the proteins FliM (about 35 units) and FliN (about 110 units) form the 45 nm diameter C
ring, which together with the M and S rings forms the rotor. This rotor drives the rod, which is a
rotor-shaft connected through the centre of the L and P rings to the hook.

The proteins MotA and MotB form a ring of 10 studs embedded in the cytoplasmic membrane,
forming the stator, which is tethered by connections to the rigid Peptidoglycan layer (PG). The L
and P rings act as bushing.The C ring is made up of the proteins FliM and FliN
(and perhaps also FliG)
and the Mot proteins consist of two subunits: Mot A and MotB. Studies suggest that 4 MotA subunits
complex with 2 MotB subunits and that mostly MotA, but also part of MotB contributes to the formation of
a proton channel.

The hook associated proteins HAP3 and HAP1 are also known as FlgL and FlgK respectively. The rod is
made up of a variety of proteins and the protein FliF spans the region between the M and S rings. It is
thought that as protons (H+ or hydrogen nuclei) move through the Mot ring complex, MotA undergoes a
shape change, exerting a torque (rotary force) on FliG which connects to the M ring.

The basal complex anchors the flagellum into the surface of the cell. There are two structural variations
according to the type of bacterial wall it is anchored in. Gram positive bacteria (stain purple with Gram's
stain) like
Bacillus possess a thick peptidoglycan wall (about 80 nm) overlying the bilipid cytoplasmic
membrane (CM). In this case the basal body has three rings, the 26 nm diameter M (membrane) ring
embedded in the CM, the S (supramembrane) ring or socket attached to the inner surface of the
peptidoglycan wall by techoic acids and the C (cytoplasmic) ring. The S ring is an extension of the
M-ring, to which it is attached, and both are composed of the same ring of protein FliF, thus the M and S
rings are sometimes considered to be a single double-ring, the MS ring. A rod (7nm diameter) passes
into the S ring socket and its other (distal) end attaches to the hook. More details of these structures are
illustrated below:
Above: illustrating the flow of protons (carrying positive electric charge) across the Mot ring from the
periplasm into the cell, powering the motor.
Comment on this article!
bacterial flagellum schematic
More articles on bacterial motility:

Models of flagella rotation - how does the rotor work? Click here to find out!

Learn about motility in the strange spirochaetes which drill through materials (including people!).

Learn about gliding and other alternative mechanisms of locomotion in bacteria.

Download an illustrated essay on bacterial motility and navigation in pdf format: Prokaryotes_motility.

How can something as complex as a bacterial flagellum evolve?
Inside the 'circle' of flagella, driven by their CCW rotation, a CW vortex would be established. At this low
Re we would expect no turbulence, so these are orderly vortices and the central CW vortex will spiral
outward toward us (the central dot indicates the head of its velocity vector coming out of the page
toward us). This core of fluid will be displaced, essentially drilled out of the viscous medium and the 6
flagella will close together to fill the void it leaves behind it (fluid flow from outside the flagella, passing
in-between them to fill this void would not readily occur, since the flows of neighbouring vortices tend to
cancel midway between them) and the flagella will bundle together. Fluid will continue to spiral CW
around the bundle as it rotates CCW, propelling the cell.

We have to ask the question: do the flagella or the cell body generate most of the propulsive thrust? For
a spherical or rod-shaped bacteria powered by a single CCW rotating flagellum the cell body will rotate
CW. This rotation of the cell body will not, by itself, generate thrust since, as already discussed, in order
to generate thrust in a high viscous fluid it is necessary to have time-non-reversible flow, which requires
a helix of definite handedness. Perhaps the spiral S-layer of protein that encases many bacteria
(forming an outer layer around the cell envelope) helps break the symmetry and generate thrust. In
coma (vibrioid) and spiral forms, the cell body will indeed displace fluid in a non-reversible manner and
so contribute to the thrust, effectively drilling its way through the medium.

So, what happens when the flagella rotate CW?

If the bundle remained intact then it would begin to pull the cell backwards by displacing a helix of fluid
toward the cell body. However, the external fluid will flow towards the end of the flagellum bundle and
perhaps this drives the bundle apart by forcing its way between the flagella. If this was so, might the cell
be observed to reverse momentarily before tumbling? This raises the question as to how many flagella
must rotate CW for the bundle to separate. If one flagellum only reverses direction, then its flow would
couple with the neighbouring flagellar vortices.

Consider two neighbouring vortices, one rotating CW and one rotating CCW, as shown below:
bacterial flagellar vortices
Vortices generated by the 6 rotating flagella of Escherichis coli
as viewed from behind the cell looking forward along its
direction of locomotion.
This vortex pair will drive fluid in one direction between them and around them – the two vortices are
coupled together. Viscosity will cause the vortex pair to move in the direction of this flow. Some of the
pairs will try to move toward the centre (which they can not do since other flagella in the bundle will block
their path) and other pairs will tend to move away from the bundle – the bundle will disperse.

In addition to changing rotation sense, the flagella also change pitch and this may also contribute to the
decoupling of the flagella and dispersal of the bundle.

Swarmer Cells

Most bacteria alternate between a single-celled stage and some form of multicellular stage during their
life-cycle. Although bacteria rarely form true multicellular organisms, in which cells are in intimate contact
and communicating directly with their neighbours through cell-to-cell junctions, most do form multicellular
aggregates in which cells are separately embedded in a common slime layer, which is attached to a solid
surface and forming a
biofilm. Cells in the biofilm are often immotile and certainly lose their ability to swim
as they shed their flagella. Swarmer cells, however, are a dispersive stage and these may float, swim,
crawl or glide away from the biofilm before eventually settling down, adhering to a surface and
establishing a new
microcolony which may eventually develop into a biofilm. Sometimes the flagella
assist in adhesion by sticking to a suitable surface they come into contact with. However, soon after
adhering the swarmer cells lose their flagella and establish a microcolony. Bacteria in both the swarmer
cell stage and the biofilm stage maintain a degree of individuality and both are capable of dividing and

Biofilms also disperse by shedding slimy
streamers, containing cells, downstream. Generally, dispersal
is more efficient if the biofilm sits upstream of fluid flow. How do the swarmer cells, which are often
released into the stream some 400 or so micrometres above the surface swim in these currents?
Research has shown that Escherichia coli, at least, has a neat trick. In those strains under study it was
found that swimming Escherichia coli (swarmer cells) preferentially circle to the left until they find a
sheltered crevice in a surface, where flow is less, and they swim upstream to establish a new microcolony
and biofilm!

Swarming and crawling

The normal swimming flagellated cells that we have considered so far are described as swarmer cells, or
planktonic cells, distinguishing them from the non-flagellated cells that are integral residents in biofilms.
However, ‘swarming’ is usually used for yet another, though closely related, phenomenon – the mass
migration of colonies of bacteria across a solid surface. This swarming behaviour is regulated by quorum
sensing and in
Escherichia coli and Salmonella typhimurium involves a switch to a multinucleoid elongate
hyperflagellated phenotype. These multinucleoid cells, called filaments, are up to 50
micrometres long and have several nucleoids, dispersed at intervals throughout the filament (nucleoid
division and cell wall growth have continued in the absence of cell division). They move as a colony with
the outer layers spiraling outward with the evacuated space inside the colony becoming filled with new
cells. This can give rise to fast colony expansion at rates up to ~3 mm/s (1 cm/h). Spiral and 2D-
branching patterns of colonial growth result and probably acts to optimize nutrient uptake on a solid
substrate in which diffusion is limited (in a solid sheet that completely covers the surface, many cells may
become starved of nutrients due to competition with their neighbours – similar considerations have shown
to predict the branching growth of sponges in 3D in computer models utilizing the diffusion equation).

The hyperflagellated filaments use their many flagella to crawl over solid surfaces, such as agar in a petri
dish. The flagella still bundle, but the cells do not tumble, instead when the flagella disperse the filament
simply stops and the flagella may reform a bundle of the opposite sense resulting in a reversal of
direction of movement. Synthesis of so many flagella appears to be triggered when the bacteria sense
the presence of a highly viscous medium (such as the surface of an agar plate).

Update on Swarming

Research shows that bacteria only swarm when
moving in groups and evidence suggests that they
interlock their flagella to move as rafts. Cells may detach and reattach to the raft, but detached cells do
not swarm but remain immotile instead. They can also secrete surfactant to reduce surface tension and
facilitate gliding of the rafts. Research also suggests that both the flagella and pili can function as
mechanoreceptors, sensing contact with a surface to trigger microcolony formation.
A pair of oppositely rotating vortices couple
together and move downwards in this case.
Detailed Mechanism of Torque Generation - A Model

Quite a number of ingenious models have been proposed over the years, attempting to explain how the
bacterial flagellum rotor generates torque. The most recent model can be summarised thus:

  1. Protons flow (down their electrochemical gradients) through the MotA/B proton channels.
  2. This induces a conformational change in MotA which extends charged amino acid residues towards
    charged residues on FliG (presumably without the protons the charged residues on MotA are
    internalised and screened).
  3. The electrostatic force (Coulomb force) of repulsion pushes the Ms and c rings to which the FliG is

Certainly, electrostatic repulsion can easily generate sufficient force, as elementary calculations show.

Under typical conditions the filament (at least in Escherichia coli) is a left-handed super-helix (the
flagellin protein subunits that make up the filament are arranged in a helix which is itself coiled into a larger
helix). Counterclockwise rotation of this helix exerts force on the cell body (due to fluid viscosity resisting
the moving filament) causing it to rotate as it is pushed along. In peritrichously flagellated bacteria,
hydrodynamic forces draw the flagella into a bundle when they rotate counterclockwise. Clockwise rotation
of the filaments causes them to fly apart in the bundles of peritrichous enteric bacteria or pulls the cell in
reverse in polarly flagellated bacteria.

Performance of the bacterial flagella motor

Not all bacteria are motile and of those that are more than half use one or more helical flagella, like the
one we have described above. Compared to other bacterial propulsion systems (which we shall look at
later) the flagella propellers have the advantage of speed. For example, a cell of the bacterium
is about 2 micrometres (2 thousandths of a millimetre) long and has six flagella that originate from
various points on the cell surface but the filaments come together to form a bundle which propels the cell
at about 20 micrometres per second, or ten body lengths per second! This is a clear advantage that more
than pays back the high cost of flagella - each flagellum contains about 1% of the bacterium's total protein
and requires some 50 genes for its production (about 2 % of the genome of about 2 500 genes). The
flagella enable swarmer cells to disperse from the biofilm (colony) and locate new sites for colonisation.
They enable the bacteria to swim to a source of nutrients and to avoid harmful irritants. The drawback of
flagella is that they require a fluid medium in which to work most effectively, although bacteria can use
them to move across moist surfaces. Flagella also fail to work well in highly viscous media, such as the
muddy ooze at the bottom of ponds, and bacteria may employ different propulsion systems in these

The helical filament rotates in a rigid manner as determined by attaching latex beads along the filaments of
mutants in which the filaments are straight. This ruled out the possibility that they undulate like whips (as
do the flagella of many non-bacterial cells) or that they move by winding and unwinding. More recent
experiments using laser dark-field microscopy have enabled individual flagella to be directly observed
rotating. They can flex, however, by virtue of the flexible hook. When the six flagella of
Escherichia coli
rotate counterclockwise (CCW) forces exerted on them by the surrounding water (hydrodynamic forces)
force them to come together into a single bundle, trailing behind the tail end of the cell.

The flagella of
Vibrio alginolyticus can rotate at about 1000 revolutions per second (rps) and propel the
cells at up to 116 micrometres per second (compare to the flagellum of
Escherichia coli with 270 rps and a
top speed of 36 micrometres per second). However, removal of the filament (which loads the motor) may
increase the engine rotation rate to 200 000 rps!

Flagella Bundles

Although many bacteria have a single falgellum at one end or side of the cell, or perhaps one at each end,
polar flagellation,  many also have multiple flagella that entwine together to form a rotating
flagellum bundle. This is the type seen in
Escherichia coli and Salmonella typhimurium. These bacteria are
peritrichously falgellated - they have several or many flagella dispersed over the cell's surface. Obviously if
flagellum rotated independently then the cell would go nowhere fast! Instead the many flagella bundle
together at the rear of the cell. In order to turn around the flagella change the sense of their rotation,
rotating in the opposite direction, and this mysteriously causes the flagella bundle to fly apart and the cell
tumbles, randomly changing direction, before the flagella switch rotation direction again and come
together as a bundle and the cell moves off in a new direction. A flagella bundle creates the opportunity for
more locomotive force, since more motors are working together to propel the cell and are perhaps also
useful for moving through more viscous fluids.

What causes flagella to bundle or to fly apart?

It is said that ‘hydrodynamic forces’ cause the flagella to bundle, for example during CCW (counter-
clockwise) rotation in the left-handed flagella of
Escherichia coli the flagella bundle together. Recall that a
helix or screw can be either left-handed or right-handed. If you look down the axis of a left-handed helix
from behind and then rotate it CCW then it will move forwards as it rotates. A right-handed helix, on the
other hand, will move forwards when rotating CW (clockwise) when viewed in the same manner.

We propose a model using elemntary fluid mechanics. Perhaps we can picture what could be happening
with a simple diagram. The diagram below is the view we would have looking from behind a bacterium with
6 flagella, like
Escherichia coli, as it swims forward straight away from us (into the page). The six left-
handed (LH) flagella rotate CCW. Since these are LH helices, the flagella will drive forward into the page (X
marks the tail-end of their velocity vectors). Remember that we have a very low
Reynolds number (Re) for
such a small organism and so the water medium is behaving as a very viscous (thick and sticky) fluid.
flagella structure in detail
Note that it is difficult to map proteins from their predicted or measured shapes to electron density
maps, such as those generated by electron cryomicroscopy. Some newer models have FliG as the
top of the C-ring, the U-shaped structure at top in yellow in the fig. above, rather than as part of the
M ring, whilst other models have it spanning the C and M rings. Whatever its exact position its
interactions with Mot are thought to drive the rotor, including both the C and MS rings, which must,
therefore, be connected in some way.

There are some 25 or 26 FliG molecules in the flagellum base and about 11 Mot complexes, each
containing 4 MotA and 2 MotB. It should also be realised that different bacterial species have
different flagellum base architectures. An updated and more detailed model is shown below:
Page updated: 1/9/2013. Details about flagellar bundles, swarming, swarmer cells
and protein structure added. Errata concerning the nature of the stator and
end-cap proteins applied.

Page updated: 24/8/2018. Recent discoveries on the mechanism of flagella
rotation and the mechanism of swarming added.
hyperflagellated filament
Above: a hyperflagfellated bacterial filament (this one is 4 times the normal length at about 8 micrometres
and has 20 flagella). When fully developed this cell may have 100 flagella and measure about 50
micrometres in length!
The table below summarises the various substructures shown in the figures above and their
encoding genes.
Bacterial flagellum model 2