Cell Signaling · Unit 5: Signals that build and defend an animal

Integrins, Rho GTPases and Cell Movement

How does a cell know it is holding on to something, and how does it use that grip to move? This lesson follows the integrin, a receptor with no enzyme that signals in both directions, and then the three Rho-family switches that shape actin. Afterwards you can explain how a cell crawls, and why a missing integrin or a blocked one changes what blood cells can do.

Watch first A Grip Switched On From Inside · 2:36

A white blood cell carried along in your blood can stop on the wall of a vessel in about a second. The receptor that stops it has no enzyme activity of its own.

Most receptors in this course read a signal that drifts in from somewhere else. This one reads something solid: the material the cell is standing on. So it has to answer two questions that a hormone receptor never faces. Am I holding on to anything? How hard is it pulling back? The answers help decide whether a cell stays where it is or crawls away.

The idea: a receptor for something solid

Most cells in your body sit on or in the extracellular matrix, a mesh of protein fibres and sugar chains that cells secrete around themselves. Collagen is its rope. Fibronectin is one of its glues, a long protein that sticks to collagen with one part and to cells with another.

Which part do cells hold? In 1984 Michael Pierschbacher and Erkki Ruoslahti tested short synthetic pieces of fibronectin’s cell-binding region and found that a peptide of four amino acids, Arg-Gly-Asp-Ser, could stand in for it. The first three, RGD in one-letter code, are the handle. Add free RGD peptide to a dish and cells let go of fibronectin, because the peptide fills the site that the matrix would have filled.

The receptor that holds the handle is an integrin. The name was proposed in 1986 by the laboratories of Richard Hynes, Clayton Buck and Alan Horwitz, for an integral membrane protein that ties the matrix outside to the cytoskeleton inside. An integrin is two different chains, α and β, and each crosses the membrane once. Their large outer parts together form a head that grips the matrix. Inside the cell each chain has only a short tail, a few dozen amino acids in nearly every case, and linker proteins join the tails to actin filaments (one integrin, α6β4, joins keratin filaments instead). There is no kinase domain and no channel.

Humans have 18 α chains and 8 β chains, which pair to make 24 known integrins. About a third of them recognise RGD. Others grip collagen or laminin at different sites, and the integrins of white blood cells grip proteins on the surface of other cells.

So an integrin is a mechanical link first. Its signalling comes from the proteins that gather on its tails, and it runs in both directions.

A grip that is switched on from inside

A hormone receptor waits for its ligand. An integrin is often surrounded by ligand, so the cell controls the receptor instead.

In the integrins studied most closely, those of platelets and white blood cells, the outer part can fold at a knee so that the head tips back toward the membrane. Bent, with the head closed, it holds its ligand weakly. Extended, with the head open, it holds tightly.

The switch is thrown from the cytoplasm. A protein called talin binds the tail of the β chain. That separates the two tails, the change passes through the membrane, and the outer part straightens. A second tail-binding protein, kindlin, is needed as well. The message starts inside the cell and changes the grip outside, so this is called inside-out signalling.

A platelet shows why it matters. It carries its integrin, αIIbβ3, on its surface all the time, switched off, while blood full of the integrin’s ligand, fibrinogen, flows past. When a vessel is cut, signals from the wound switch the integrin on within seconds. Fibrinogen has two ends, so it bridges one platelet to the next, and a plug forms. A white blood cell faces a harder version of the same problem.

Design problem

A white blood cell is carried along in flowing blood. It must stop and crawl out of the vessel only where there is an infection. So its integrins must be off while it flows and on within about a second at the right place. Design the control, using a signal on the vessel wall and parts from earlier lessons. Then predict what happens to a child whose white cells cannot make this integrin.

One way to do it

A second is too short to make a new protein, so the integrin must already be on the surface, held in its bent form. The “on” signal must sit on the vessel wall, and only at the infected place.

The address. Cells lining a vessel next to an infection put two things on their surface. Selectins are sticky proteins whose short-lived bonds make a passing white cell slow down and roll. Chemokines are small signal proteins that stay stuck to the wall.

The receptor. The white cell carries a G-protein-coupled receptor for the chemokine, a fast architecture whose parts are already at the membrane.

The relay. The G protein turns on phospholipase C, so calcium and diacylglycerol rise. Together they turn on a GEF for Rap1, a GTPase of the Ras family (CalDAG-GEFI in platelets and mouse neutrophils; other blood cells use other GEFs). Rap1 with GTP bound, “turn on”, helps bring talin to the integrin tail. The details differ between cell types.

The grip. Talin binds the β tail, the integrin extends, and it grabs its partner on the vessel wall, a protein called ICAM-1. The cell stops. Then it crawls out between the cells of the wall.

Nothing has to be made, so the relay can run in about a second.

The child. White cells without this integrin roll but cannot stop. They stay in the blood, so the blood count of white cells is very high while infected tissue has almost none. You would predict repeated bacterial infections that form little pus, since pus is mostly dead white cells. That is what is seen.

The focal adhesion: a signalling machine that can feel a pull

Once integrins hold the matrix they cluster, and their tails collect a crowd of proteins. Talin joins the β tail to actin. Vinculin binds both talin and actin and reinforces the join. Paxillin is a scaffold, with many docking sites and no enzyme activity. Focal adhesion kinase (FAK) is a tyrosine kinase that floats in the cytoplasm until an adhesion forms. The whole assembly is a focal adhesion.

This is how a receptor with no enzyme signals inward: it borrows one, as the cytokine receptors in the JAK and STAT lesson do. Resting FAK is folded shut, with one of its own domains lying against the kinase. At a cluster of integrins it opens and phosphorylates itself on one tyrosine, Tyr397. How clustering opens it is not fully worked out. That phosphotyrosine is a docking site for the SH2 domain of Src. Docked Src phosphorylates FAK further, and the pair then phosphorylates its neighbours, paxillin among them, which makes more docking sites.

One site on FAK binds Grb2, the adaptor that leads to Ras. The next step follows a rule you met with Ste5 and KSR: “Scaffolds can localize ras-mapk.” In cells responding to hepatocyte growth factor through its receptor c-Met, paxillin holds Raf, MEK and ERK next to FAK at the focal adhesion. The cascade is the one you know, running at one address.

The adhesion also measures force. A cell pulls on its adhesions with myosin, and a stiff surface resists more than a soft one. Michael Sheetz’s laboratory spent years asking where that pull becomes chemistry. In 1997 Daniel Choquet, Daniel Felsenfeld and Sheetz held fibronectin-coated beads on fibroblasts with a laser trap. The harder the trap resisted, the more firmly the cell tied that patch of integrins to its actin.

In 2009 Armando del Rio and colleagues went down to one molecule. The long rod of talin is a string of helix bundles, and helices that vinculin binds are folded away inside the bundles. They fixed a piece of the rod to glass at one end and to a magnetic bead at the other, and pulled. Unpulled, a rod bound at most one vinculin. Pulled with about 12 piconewtons, it bound up to three.

No phosphate was added and no GTP was exchanged. Force alone uncovered binding sites. Each new vinculin is another link to actin, so an adhesion that is pulled gets stronger. The experiment used purified protein, so its place in a living cell is an inference, and whether stretch uncovers sites in many other proteins is still open.

Three switches for the actin skeleton

Watch Three Switches, Three Shapes · 1:44

A grip is no use to a cell that cannot change shape. Shape is the work of actin filaments, and three small GTPases of the Rho family decide what the filaments build. They work as Ras does: GTP-bound, “turn on”; GDP-bound, “turn off”; with GEFs and GAPs setting the balance.

In 1992 Anne Ridley and Alan Hall injected purified proteins into fibroblasts that had been starved of serum, which leaves them with little organised actin. Active Rho filled the cells with thick actin cables. Active Rac made their edges ruffle. In 1995 Catherine Nobes and Hall, and separately the laboratory of Louis Lim, showed that Cdc42 produces thin spikes.

  • Rho builds stress fibres: cables of actin and myosin II that end in focal adhesions and contract. One effector is Rho kinase (ROCK), which raises the activity of myosin II. Rho is also the switch that G12/13-coupled receptors turn on.
  • Rac builds lamellipodia: broad, thin sheets pushed forward by a branching mesh of filaments. Rac-GTP turns on the WAVE complex, and WAVE turns on Arp2/3, which starts each new filament as a branch on the side of an old one.
  • Cdc42 builds filopodia: thin fingers with a core of parallel filaments. One effector is WASP, which is folded shut until Cdc42-GTP binds it and frees the part that turns on Arp2/3.

The adhesion and the switches are wired together: proteins phosphorylated by FAK and Src recruit GEFs for Rac, so a new grip at the front calls for more protrusion there.

How a cell crawls, and what happens when it cannot

Put the parts in order and the cell moves.

  1. Push. At the front, Rac drives a lamellipodium forward.
  2. Grip. Under the new sheet, integrins catch the matrix and small adhesions form.
  3. Pull. Rho, through ROCK and myosin II, contracts the cell body against those adhesions and hauls it forward.
  4. Let go. At the back, old adhesions come apart and the tail is drawn in.

Rac in front and Rho behind is the short version. It is a simplification: sensors that report Rho activity in living cells find some at the front edge too.

The child in the design problem is real. In leukocyte adhesion deficiency type I, the gene for β2, the chain shared by the white-cell integrins, is mutated. White cells cannot stop on the vessel wall, and affected children have severe bacterial infections starting in infancy. A bone marrow transplant can cure it.

Platelets have a matching disease. People without working αIIbβ3 bleed easily, a condition called Glanzmann thrombasthenia. A drug can copy that state on purpose. Abciximab, a fragment of an antibody that bound αIIbβ3 and blocked it, is no longer supplied. Eptifibatide and tirofiban are small molecules that sit in its ligand pocket. They are given by vein, in hospital, to some patients with a heart attack or a threatened one, most often around the catheter procedure that reopens the blocked artery, so that platelets cannot bridge. The main risk is the one the genetics predicts: bleeding.

How we measure it

Microinjection with actin staining

A purified protein is injected into one cultured cell through a fine glass needle. The cell is later fixed and stained with fluorescent phalloidin, a mushroom toxin that binds actin filaments, so the microscope shows what the injected protein made the actin do. Ridley and Hall tested Rho and Rac this way, one protein at a time.

Peptide competition

Short synthetic peptides that copy part of a matrix protein are added to cells in solution. If a peptide stops cells from attaching to the whole protein, and a control peptide with one residue changed does not, then its sequence is what the receptor recognises. This is how the RGD sequence was pinned down.

Optical trap (laser tweezers)

A focused laser beam holds a small bead and pulls it back toward the focus with a force the experimenter sets, in the piconewton range. With a matrix-coated bead resting on a cell, the trap reports how the cell responds when the thing it is holding resists.

Magnetic tweezers with single-molecule counting

One protein molecule is tethered between a glass surface and a magnetic bead, and a magnet pulls on the bead with a chosen force. Fluorescent partner proteins that bind the stretched molecule are counted one at a time, because each dye goes dark in a single step.

Finish line

Test yourself

8 questions

Practice and MCAT-style, with explained answers.

Next · Lesson 28

Innate Immunity: Toll-like Receptors and NF-κB

How does a cell recognize an enemy it has never met, and answer within minutes?

Continue →

← Lesson 26 · Hedgehog and the Primary Cilium

Integrin αVβ3, outer part — the bent form
Integrin αVβ3, outer part — the bent form ● LIVE · RCSB PDB 1JV2
open on RCSB PDB ↗
Loading live structure 1JV2 from RCSB PDB…

The outer parts of the α and β chains without the membrane segments or tails. The long legs fold at a knee, so the head that binds the matrix tips back toward where the membrane would be.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Integrin αVβ3 with an RGD peptide — the handle in its site
Integrin αVβ3 with an RGD peptide — the handle in its site ● LIVE · RCSB PDB 1L5G
open on RCSB PDB ↗
Loading live structure 1L5G from RCSB PDB…

A small ring-shaped RGD peptide sits in the cleft between the two chains. Its arginine reaches the α chain, and its aspartate holds a metal ion that is bound to the β chain.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Focal adhesion kinase — folded shut
Focal adhesion kinase — folded shut ● LIVE · RCSB PDB 2J0J
open on RCSB PDB ↗
Loading live structure 2J0J from RCSB PDB…

A fragment of FAK holding its FERM domain and its kinase domain. The FERM domain lies against the kinase and blocks it, which is the off state before an adhesion forms.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
RhoA with a GTP analogue — the switch turned on
RhoA with a GTP analogue — the switch turned on ● LIVE · RCSB PDB 1A2B
open on RCSB PDB ↗
Loading live structure 1A2B from RCSB PDB…

A dominantly active mutant of RhoA holding GTPγS, a form of GTP that is hydrolysed very slowly. Compare its fold and its two switch regions with the Ras structures in the lesson on GTPase switches.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Cdc42 with a piece of WASP — a switch holding its effector
Cdc42 with a piece of WASP — a switch holding its effector ● LIVE · RCSB PDB 1CEE
open on RCSB PDB ↗
Loading live structure 1CEE from RCSB PDB…

Cdc42 in its on state, bound to the segment of WASP that recognises it. WASP is named for Wiskott-Aldrich syndrome, an inherited immune deficiency in which it is mutated.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Michael Sheetz

Michael Sheetz 1946–2025

His laboratory showed that pulling on a single talin rod uncovers buried binding sites for vinculin, which gives a focal adhesion a physical way to turn force into binding.

Michael Sheetz asked where, in a cell that is being pulled, force turns into chemistry. A living cell has too many moving parts to answer that directly, so his laboratory took parts away in stages. In 1997, at Duke University, Daniel Choquet, Daniel Felsenfeld and Sheetz held fibronectin-coated beads on fibroblasts with a laser trap. The harder the trap held a bead back, the more firmly the cell tied that patch of integrins to its actin. In 2006, at Columbia University, Yasuhiro Sawada and colleagues stretched a purified piece of the adhesion protein p130Cas and found that Src-family kinases phosphorylated it faster when it was extended. In 2009 Armando del Rio, working with Sheetz and with Julio Fernandez's laboratory, pulled on single talin rods with a magnet. An unpulled rod bound at most one vinculin, and a rod under about 12 piconewtons bound up to three. The work gave the field a mechanism for sensing force that can be tested one molecule at a time.

✓ verbatim del Rio et al., Science 323:638 (2009); Choquet, Felsenfeld & Sheetz, Cell 88:39 (1997).
Richard Hynes

Richard Hynes

His laboratory, with collaborators, cloned one chain of the receptor that links fibronectin outside the cell to actin inside, and proposed the name integrin for it.

Richard Hynes came to integrins from the outside of the cell. In 1973 he reported that a large protein on the surface of normal cultured cells was missing from cells that a tumour virus had transformed. Several laboratories found the same protein at about that time, and it was later named fibronectin. It formed fibres outside the cell that lined up with actin bundles inside, so something had to cross the membrane and join the two. In 1986 his laboratory at the Massachusetts Institute of Technology, with the laboratories of Clayton Buck and Alan Horwitz, cloned one chain of a membrane protein from chick cells that did this job. The paper proposed the name integrin, for an integral membrane complex that ties the matrix to the cytoskeleton. As more sequences came in, the fibronectin receptor, the fibrinogen receptor of platelets and the adhesion proteins of white blood cells proved to be one family. One name let people working on clotting, immunity, development and cancer see that they were studying the same kind of receptor.

✓ verbatim Tamkun et al., Cell 46:271 (1986); Hynes, Proc. Natl. Acad. Sci. USA 70:3170 (1973).
Anne Ridley and Alan Hall

Anne Ridley and Alan Hall

By injecting proteins into single fibroblasts they showed that Rho controls stress fibres and focal adhesions and that Rac controls membrane ruffles in cells responding to growth factors.

In 1992 Anne Ridley was a postdoctoral researcher in Alan Hall's laboratory at the Institute of Cancer Research in London. Rho and Rac were relatives of Ras whose jobs in the cell were unclear, although there were hints that Rho affected cell shape. Ridley used Swiss 3T3 fibroblasts that had been starved of serum, which leaves them with very little organised actin, and injected purified proteins into single cells through a fine glass needle. Rho locked in its active form filled the cells with stress fibres and focal adhesions, and an enzyme that inactivates Rho stopped serum from doing the same. Active Rac made the cell edge ruffle, and a blocking form of Rac stopped growth factors from causing ruffles. The two papers appeared back to back in the journal Cell. In 1995 Catherine Nobes and Hall added Cdc42 and filopodia, as did the laboratory of Louis Lim. The work gave everyone who studies cell movement named switches to test, each tied to a structure that can be seen in a microscope.

✓ verbatim Ridley & Hall, Cell 70:389 (1992); Ridley et al., Cell 70:401 (1992).
Test yourself
Question 1 of 8 Score 0 / 0

Which description fits an integrin?
Sources & Integrity

Every structure, slide, simulation, and clip on this page is listed below with the source it came from. None of the 3D structures is drawn by hand or by AI: each one loads by accession ID from RCSB PDB or PubChem. This panel is generated from the lesson's own data file, so what it lists is what the page loads.

Molecular structures live data

  • Integrin αVβ3, outer part — the bent form — RCSB PDB 1JV2 · coordinates CC0 · Xiong et al., Science 294:339 (2001); X-ray crystal structure
  • Integrin αVβ3 with an RGD peptide — the handle in its site — RCSB PDB 1L5G · coordinates CC0 · Xiong et al., Science 296:151 (2002); X-ray crystal structure
  • Focal adhesion kinase — folded shut — RCSB PDB 2J0J · coordinates CC0 · Lietha et al., Cell 129:1177 (2007); X-ray crystal structure
  • RhoA with a GTP analogue — the switch turned on — RCSB PDB 1A2B · coordinates CC0 · Ihara et al., J. Biol. Chem. 273:9656 (1998); X-ray crystal structure
  • Cdc42 with a piece of WASP — a switch holding its effector — RCSB PDB 1CEE · coordinates CC0 · Abdul-Manan et al., Nature 399:379 (1999); NMR structure

Scientist profiles & quotes

  • Michael Sheetz — del Rio et al., Science 323:638 (2009); Choquet, Felsenfeld & Sheetz, Cell 88:39 (1997).
  • Richard Hynes — Tamkun et al., Cell 46:271 (1986); Hynes, Proc. Natl. Acad. Sci. USA 70:3170 (1973).
  • Anne Ridley and Alan Hall — Ridley & Hall, Cell 70:389 (1992); Ridley et al., Cell 70:401 (1992).

Quotes are verbatim from the cited source. Profile text is original to this site and was checked against the sources cited.

Animated explainer

A Grip Switched On From Inside · narration: an AI narrator voice (Gemini, Charon) reading narration written from this lesson's text · animation: AI-generated in code (Manim) by Claude, a schematic with no drawn molecular structures, plots computed from their equations · captions from the narration script.

Section films

  • Three Switches, Three Shapes · narration: an AI narrator voice (Gemini, Charon) reading narration written from this lesson's text · animation AI-generated in code (Manim) by Claude, a schematic with no drawn molecular structures · captions from the narration script.

Images & figures

This lesson carries no raster figures. Every structure is rendered live from RCSB PDB or PubChem, and every plot is drawn from its equation in the browser.

integrity Machines check the mechanics. The build confirms that every local file exists, that every question's answer key points at a real choice, and that any AI-generated scientific figure names the person who checked it; a weekly job re-resolves every accession ID against its database. Questions are original; MCAT-style items are written for this site and are not real AAMC questions. Banned sources (the Tymoczko textbook, publisher decks, copyrighted figures) are not used.

Like I'm five

Your cells have tiny hands that hold on to the stuff around them. A cell can make its hands grip hard or let go, and when something tugs on a hand the cell feels it and holds on tighter. To move, a cell reaches out in front, grabs on, pulls itself along and lets go at the back. Little switches inside tell it where to reach and where to pull.
High school

Cells are anchored to a mesh of protein outside them by receptors called integrins. An integrin works like a clamp with two ends: the outside grips the mesh and the inside is tied to the cell's internal skeleton of actin filaments. The cell can tighten or loosen the clamp from the inside, which is how a white blood cell stops at an infection and how a platelet starts a clot. Where many integrins grip together, enzymes gather and send signals into the cell. Three switch proteins called Rho, Rac and Cdc42 then rearrange the actin so that the cell pushes out in front, grips, pulls its body forward and releases at the back.
College

Integrins are heterodimers of an α and a β chain with a large extracellular head, one membrane-spanning segment per chain and short cytoplasmic tails that have no enzyme activity. Signalling is bidirectional. Inside-out, talin (with kindlin) binds the β tail and shifts the integrin from a bent, low-affinity form to an extended, high-affinity form. Outside-in, ligand-bound integrins cluster and assemble a focal adhesion that contains talin, vinculin, paxillin and focal adhesion kinase. FAK phosphorylates itself on Tyr397, Src docks there through its SH2 domain, and the pair phosphorylates neighbouring proteins and can feed Ras and MAPK signalling held on the scaffold paxillin. Tension on talin exposes buried vinculin-binding sites, so the adhesion strengthens under load. The Rho-family GTPases organise actin: Rho drives stress fibres and contraction through Rho kinase and myosin II, Rac drives lamellipodia through the WAVE complex and Arp2/3, and Cdc42 drives filopodia and activates WASP. Migration is a cycle of protrusion, adhesion, contraction and release.
Graduate

Integrins are αβ heterodimers (18 α and 8 β chains giving 24 pairs in humans) that exchange among bent-closed, extended-closed and extended-open states. The FERM-domain head of talin binds the membrane-proximal NPxY motif of the β tail, breaks the association between the α and β membrane and tail segments and, with kindlin, shifts the balance toward the extended-open, high-affinity state. Rap1-GTP is a major upstream input in blood cells. Outside-in, clustered integrins assemble talin, vinculin, paxillin and FAK. Release of the FERM domain from the FAK kinase domain permits phosphorylation of Tyr397, Src docks by its SH2 domain and phosphorylates the FAK activation loop (Tyr576 and Tyr577) and Tyr925, a Grb2 site, and the FAK-Src pair phosphorylates paxillin and p130Cas. Rho, Rac and Cdc42 couple to Rho kinase and formins, to the WAVE complex and Arp2/3, and to WASP-family proteins. Force enters at talin. In del Rio et al. (2009) a rod fragment, residues 482 to 889, bound at most one vinculin head when slack and up to three at about 12 pN. The open question is how this maps onto a living adhesion. In a separate atomic force microscope assay, the same paper's five unfolding peaks averaged 29 to 51 pN, though the authors argued that a lower force held for longer could be as effective, while a tension sensor placed in vinculin reported about 2.5 pN in stable adhesions and the weakest talin bundle, R3, has been reported to open near 5 pN. Which bundles open in a cell, how often, and whether stretch-uncovered binding sites are common outside talin are not settled.
A drug from a rattlesnake

Many viper venoms contain small proteins that carry an RGD sequence and jam the platelet integrin, which helps to keep a bitten animal's blood from clotting. They are called disintegrins. In 1991 Robert Scarborough and colleagues described one from the southeastern pygmy rattlesnake, named barbourin, that has a lysine where the arginine would be. With KGD in place of RGD it blocked the platelet integrin and largely spared its relatives. The drug eptifibatide is a small ring-shaped peptide modelled on that sequence.

— Scarborough et al., J. Biol. Chem. 266:9359 (1991).
Counting molecules by the steps

To count how many vinculin molecules sat on one stretched talin rod, del Rio and colleagues gave each vinculin head a single fluorescent dye. One dye molecule does not fade gradually. It goes dark in one step, so the number of steps in which a spot went dark was the number of vinculins bound. A control protein, α-actinin, which has a vinculin site that is already exposed, bound one or none whether the magnet was on or off. Computer simulations from other laboratories had predicted the buried-site mechanism before the experiment was done.

— del Rio et al., Science 323:638 (2009).
Why it is called an integrin

Before 1986 the protein went by several names, each from the antibody or the cell type used to find it. When the laboratories of Richard Hynes, Clayton Buck and Alan Horwitz worked out the sequence of one chain from chick cells, they proposed a single name, integrin, for an integral membrane complex that ties the matrix outside a cell to the cytoskeleton inside. As more members were sequenced, receptors studied separately by people working on platelets, white blood cells and fibroblasts turned out to be relatives, and the name spread to all of them.

— Tamkun et al., Cell 46:271 (1986).

 ↑ Top