Three filament systems, one moving cell
The cytoskeleton is built from three filament systems — actin microfilaments, microtubules, and intermediate filaments — but a crawling cell is powered almost entirely by the first. At the leading edge, a dense network of actin filaments polymerises and pushes the membrane forward as a broad lamellipodium or as thin, finger-like filopodia; at the trailing edge, myosin motors contract the network and pull the rear along; adhesions to the underlying substrate cycle between gripping and releasing throughout. That same basic engine drives immune cells chasing an infection, skin cells closing a wound, and — less helpfully — cancer cells spreading through the body.
Actin treadmilling: polymerising forward
Actin filaments are polar: the barbed end grows quickly, the pointed end shrinks slowly. ATP-bound actin monomers add preferentially at the barbed end; once incorporated, the bound ATP is gradually hydrolysed to ADP over the following seconds, and ADP-actin subunits dissociate preferentially from the pointed end. Under steady-state conditions this produces treadmilling: the filament's length stays roughly constant even as individual subunits continuously flow through it from the barbed end to the pointed end.
barbed (+) end: fast ATP-actin addition (growth)
↓ ATP hydrolysis inside the filament, over seconds
pointed (-) end: ADP-actin dissociation (net loss)
net effect: roughly constant filament length, but a steady monomer flux from + to -
The Brownian ratchet: how a soft polymer pushes a membrane
A single actin filament is far too thin and flexible to shove a membrane forward the way a rigid rod would. Peskin, Odell and Oster's 1993 Brownian ratchet model explains the protrusion mechanism differently: thermal fluctuations of both the membrane and the filament tip transiently open a gap wide enough — roughly the length of one monomer, about 2.7 nanometres — for a new actin subunit to insert. Once inserted, the filament is a little longer, and the membrane can no longer drift back as far as before. Net protrusion accumulates from an enormous number of these small, stochastic insertion events rather than from any single continuous push.
Branching networks: Arp2/3 and the dendritic array
At the lamellipodium's leading edge, the Arp2/3 complex — activated by WASP and WAVE family proteins downstream of Rac and Cdc42 signalling — nucleates new actin branches off existing filaments at a characteristic angle of about 70 degrees, producing a dense, tree-like dendritic meshwork rather than a bundle of isolated filaments. Spreading the protrusive force across thousands of filament tips at once is what lets the leading edge advance as a broad, coherent sheet rather than a single spike; the thin, unbranched, bundled filaments built by formins instead form filopodia, the finger-like sensory protrusions that probe ahead of the main front.
Contraction and traction: myosin II and focal adhesions
At the rear of the cell and along internal stress fibres, myosin II motor proteins walk along actin filaments using energy from ATP hydrolysis, generating a contractile force that hauls the trailing edge forward and helps disassemble the adhesions holding it back. Meanwhile, at the front, nascent integrin-based focal adhesions transiently grip the substrate — without that traction, the pushing actin network would simply deform in place rather than translocate the whole cell body. The cycle of protrude, adhere, contract, release repeats continuously as the cell crawls forward.
Steering the crawl: gradients and signalling
Directional crawling, or chemotaxis, comes from spatial gradients of a chemoattractant detected by surface receptors: the side of the cell facing the higher concentration activates Rac, Cdc42 and PI3K signalling more strongly, biasing where Arp2/3-driven branching and protrusion happen and effectively steering net polymerisation — and so net cell movement — up the gradient. This link between a chemical signal and the cytoskeleton's mechanics is what lets an immune cell home in on an infection site or a neuronal growth cone find its target.
Frequently asked questions
What does it mean for actin to treadmill?
Monomers add preferentially at the fast-growing barbed end and are lost preferentially from the slower pointed end, so under steady-state conditions the filament's overall length stays roughly constant even though individual actin subunits are constantly flowing through it from one end to the other.
How can a soft actin filament physically push a stiff membrane forward?
It doesn't push like a rigid rod. The Brownian ratchet model shows thermal fluctuations transiently open a gap between the filament tip and the membrane just large enough for a new monomer to insert, after which the longer filament prevents the membrane from returning to its old position — protrusion accumulates from many small stochastic insertion events rather than a continuous shove.
Why does a crawling cell need adhesions as well as actin polymerization?
Actin growing at the leading edge needs something to push against. Without adhesions gripping the substrate near the front, the protruding actin network would simply deform or slip in place rather than translocating the whole cell body forward.
Try it live
Everything above runs in your browser — open Cytoskeleton and Cell Motility and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.
▶ Open Cytoskeleton and Cell Motility simulation