Cell Biology · Unit 3: Traffic and movement

Cell Shape, Migration and Cilia

How does a cell built of polymers and motors move itself, or move fluid past itself? This lesson follows a crawling cell through protrusion, adhesion, contraction and release, uses the clutch idea to compute how fast a leading edge advances, and then shows how dynein sliding inside a cilium becomes a bend. You will be able to predict the cellular phenotype of a broken migration or ciliary part.

Watch first Sliding Becomes Bending · 2:45

In 1976 Björn Afzelius described four men whose sperm did not swim. Three of them also could not clear mucus from their airways, and under the electron microscope their cilia lacked the small projections called dynein arms.

One missing part, two failures in two organs. That only makes sense if a sperm tail and an airway cilium share a machine, and they do. This lesson takes apart two ways a cell turns chemical energy into movement. A crawling cell pushes, grips, pulls and lets go, using actin, adhesion receptors and myosin II. A cilium bends because dynein motors slide microtubules past each other inside a structure that resists the sliding. Both build on the polymers of Cytoskeletal Polymers and Dynamics and the motors of Molecular Motors and Cargo.

How a cell crawls: protrude, grip, pull, release

Picture a fibroblast or a fish skin cell (a keratocyte) gliding over a dish. Its front is a lamellipodium, a sheet of cytoplasm a fraction of a micrometre thick, packed with actin filaments. The cycle has four steps.

Protrude. Actin filaments push the membrane forward by adding subunits at their barbed ends, the fast-growing ends. In extracted fish keratocytes, decorating filaments with myosin fragments showed that barbed ends predominantly face forward. The network is branched. Svitkina and Borisy imaged extracted lamellipodia of Xenopus keratocytes and fibroblasts by electron microscopy and found extensive branching, with the Arp2/3 complex at the Y-shaped branch points. The Arp2/3 complex is a seven-protein assembly that starts a new filament off the side of an old one. In keratocytes thinned with cytochalasin D, the branches met their mother filaments at 67 ± 12 degrees. Which signals switch Arp2/3 on at the front, through Rac and its partners, belongs to the signaling lesson on integrins and Rho GTPases.

Grip. Integrins are transmembrane receptors that bind the extracellular matrix outside and, through linker proteins, actin inside. Clusters of them with their linkers form focal adhesions, the cell’s anchor points on the substrate (more in Adhesion and the Extracellular Matrix).

Pull. Myosin II, the two-headed motor of muscle, assembles into small filaments that contract the actin network behind the front. In fish keratocytes the evidence supported a model in which actomyosin contraction where the lamellipodium meets the cell body moves the body forward.

Release. The rear must let go. In CHO cells on matrix-coated glass, rear retraction often tore integrins out of the cell, leaving them on the substrate, and at intermediate and high adhesiveness the rate of rear retraction appeared to limit migration speed. Disassembled filaments at the back supply monomers for the front.

Not every cell crawls this way. Some push out blebs, round bulges of membrane driven by pressure from contraction of the cortex, the actin layer under the membrane. And leukocytes squeezed in confined spaces have been shown to move without transmembrane force coupling at all, using actin flow that follows the texture of their surroundings (Reversat and colleagues, 2020). The neurons and neural crest cells of the developing brain use versions of these modes; see the growth cone, neural crest, radial migration and tangential migration.

Retrograde flow and the clutch

Watch fluorescent actin speckles in a lamellipodium and you see the network moving backward, away from the edge, even as the edge moves forward. This is retrograde flow. Two forces drive it: the membrane pushes back on the growing filaments, and myosin II pulls the network rearward.

Think of a car’s clutch. With the clutch in, the engine turns the wheels. With it slipping, the engine races and the car barely moves. Polymerization is the engine. Adhesions are the clutch: where actin is linked to integrins gripping the substrate, the network cannot slide back as freely, so more of the polymerization becomes forward motion.

edge speed (relative to substrate) = polymerization speed (relative to the network) - retrograde flow speed

A worked problem

This is a one-dimensional model, and only the flow value is measured. In PtK1 cells (a rat-kangaroo epithelial line) on compliant fibronectin-coated gels, Gardel and colleagues measured distal lamellipodial actin flowing rearward at about 25 nm/s. Suppose, as an illustrative assumption, that filaments at the edge extend at 40 nm/s relative to the network, and that both rates stay constant.

  1. Edge speed. With forward positive, edge speed = 40 + (-25) = 15 nm/s.
  2. Distance in 120 s. 15 nm/s x 120 s = 1,800 nm = 1.8 µm. Check: polymerization alone gives 40 x 120 = 4,800 nm, flow takes back 25 x 120 = 3,000 nm, and 4,800 - 3,000 = 1,800 nm.
  3. Engage the clutch. Suppose a change raises adhesion engagement so rearward flow falls to 10 nm/s, with polymerization unchanged. Edge speed = 40 - 10 = 30 nm/s, and in 120 s the edge advances 30 x 120 = 3,600 nm = 3.6 µm. Cutting the flow by more than half doubled the advance.
  4. What the model leaves out. Edge advance is not movement of the cell body. Traction is not a simple function of flow either: in the PtK1 cells, traction stress and actin speed were related biphasically, switching near 8 to 10 nm/s. And adhesion cannot rise without cost, since in the CHO cells strong adhesion slowed migration because the rear could not detach.

So faster flow does not always mean more traction, and stronger adhesion does not always mean faster migration. The clutch is a starting model, not the whole story.

Lamellipodium clutch — edge speed = polymerization − flow⚙ original · interactive
Edge speed = polymerization − flow
40 − 25 = 15 nm/s
Distance in 120 s
15 × 120 = 1,800 nm = 1.8 µm
Check: forward gain − rearward loss
4,800 − 3,000 = 1,800 nm

Starting case.

Rule to keep: edge speed = polymerization speed − retrograde flow speed. One-dimensional sketch: only the 25 nm/s flow is measured (PtK1 cells); the 40 nm/s polymerization is an illustrative assumption, as are the slider ranges and any flow other than 25 nm/s (10 nm/s is the lesson’s supposed case). Stronger adhesion is not always faster movement: in PtK1 cells traction stress and actin speed were related biphasically, switching near 8 to 10 nm/s, and in CHO cells strong adhesion slowed migration because the rear could not detach. Edge advance is not movement of the cell body. Speckle positions are arbitrary, and 120 s of model time plays in 8 s, so this is a sketch of the logic.

The axoneme: sliding turned into bending

A motile cilium is built on an axoneme, a cylinder of microtubules. In the classic 9+2 pattern, nine outer doublets surround a central pair of single microtubules. Each doublet is a complete A-tubule with an incomplete B-tubule fused to it. The axoneme grows from a basal body, which anchors it in the cell.

Along each A-tubule sit two rows of axonemal dynein, the outer and inner arms. Their motor heads act on the B-tubule of the next doublet and walk toward its minus end, at the base. So each doublet pushes its neighbour toward the tip. Radial spokes reach from the doublets toward the central-pair complex. Between neighbouring doublets runs the nexin link, which cryo-electron tomography of Chlamydomonas axonemes identified as the dynein regulatory complex, hence the name nexin-dynein regulatory complex. In Chlamydomonas the whole set of arms, spokes and links repeats every 96 nm along the doublet.

Why does sliding make a bend? Hold two strips of card side by side, glued together at one end and tied loosely along their length. Push one strip toward the far end and the pair cannot slide apart, so it curves. In the axoneme the base is anchored and the links resist, so shear between doublets becomes a bend. Summers and Gibbons tested this by removing the resistance. They briefly digested sea-urchin sperm axonemes with trypsin, added ATP and watched by dark-field microscopy: the doublets slid apart. Dynein does not shorten microtubules; it slides them.

To get a wave, the arms on one side must be more active than those on the other, and the sides must take turns. Cryo-electron tomography of rapidly frozen beating sea-urchin sperm flagella found most dyneins in an active conformation, with asymmetric states across bends, which the authors suggested favours switching by inhibiting one set of motors. How the switch is signalled is still open.

Not every cilium is 9+2 or motile. Primary cilia are single, non-motile 9+0 cilia found on many vertebrate cell types; they lack dynein arms and serve as signalling antennae (see hedgehog at the primary cilium). Cilia of the mouse embryonic node are mostly 9+0, yet they move, rotating rather than waving.

When cilia fail

Each part of the axoneme predicts a phenotype. Airway cilia that do not beat cannot carry mucus up out of the lungs, as in Afzelius’s patients. Sperm whose dynein arms are missing do not swim. Nodal cilia that do not rotate cannot drive the leftward flow over the mouse node. In mice lacking the motor KIF3B the node had no cilia and organ placement was randomized, and some people with immotile cilia have mirror-image organs. This set of conditions is called primary ciliary dyskinesia. Here we stay with mechanism.

One mutation need not give one phenotype. In mice lacking Cfap53, nodal cilia were completely immotile, while tracheal and brain-ventricle cilia still beat, abnormally. The instrument matters too: in one study, conventional electron microscopy of nasal cilia from patients with DNAH11 mutations looked normal, but electron tomography found a deficit of outer-arm volume near the base.

Migration defects follow the same logic. Leukocytes that cannot make a functional integrin of the β2 family cannot grip the vessel wall firmly enough to leave the blood, and so fail to reach sites of infection.

Design problem

A Chlamydomonas mutant, or a patient’s sperm, has paralysed flagella. Design a test of whether the dynein motors themselves work.

One way to do it

Isolate axonemes, so the membrane and the cell’s own ATP supply are out of the picture. Digest briefly with trypsin to cut the links that turn sliding into bending, then add ATP (Summers and Gibbons used 0.1 mM with sea-urchin sperm axonemes) and watch by dark-field microscopy.

If doublets slide apart, the motors work. The defect then lies in regulating them or in converting sliding to bending: the radial spokes, the central pair or the nexin-dynein regulatory complex. If doublets do not slide, suspect missing or inactive dynein arms.

Confirm by electron microscopy, and by electron tomography if conventional images look normal, as they did in some DNAH11 cases.

Controls: wild-type axonemes treated the same way (sliding expected), mutant and wild-type axonemes without ATP (no sliding expected), and axonemes given ATP without trypsin (wild-type should bend or beat rather than fall apart).

How we measure it

Sliding-disintegration assay

Isolate axonemes, digest them briefly with a protease to cut the links between doublets, add ATP and watch by dark-field microscopy. If doublets slide apart, the dynein arms work; if not, the motors are missing or inactive.

Fluorescent speckle microscopy

Label a small fraction of actin so the network appears as speckles, then track the speckles over time. The result is a map of how fast and in which direction the actin network moves at each point, such as the rearward flow in a lamellipodium.

Traction-force microscopy

Grow cells on a soft gel containing fluorescent beads and measure how far the cell displaces the beads. From the displacements and the gel's stiffness you calculate the stresses the cell exerts on the substrate.

Cryo-electron tomography

Freeze cilia or flagella rapidly, tilt them in the electron microscope and reconstruct them in three dimensions; averaging the 96-nm repeat shows dynein arms, spokes and links, and their shapes in a beating flagellum.

Finish line

Test yourself

8 questions

Practice and MCAT-style, with explained answers.

Next · Lesson 17

Cell-Cycle States and Control Logic

Fuse a cell that is copying its DNA with one that has finished. Why does the finished nucleus refuse to copy again?

Continue →

← Lesson 15 · Molecular Motors and Cargo

Arp2/3 complex — the branch-maker at the leading edge
Arp2/3 complex — the branch-maker at the leading edge ● LIVE · RCSB PDB 1K8K
open on RCSB PDB ↗
Loading live structure 1K8K from RCSB PDB…

Seven subunits, two of them actin-related proteins (Arp2 and Arp3). When activated on the side of an existing filament, Arp2 and Arp3 act like the first two subunits of a new filament that grows off at an angle.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Actin monomer — the subunit added at the barbed end
Actin monomer — the subunit added at the barbed end ● LIVE · RCSB PDB 1ATN
open on RCSB PDB ↗
Loading live structure 1ATN from RCSB PDB…

Find the nucleotide in the cleft between the two halves of the actin molecule. Monomers carrying ATP add at the barbed ends that face the membrane; filaments age and come apart further back.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Tubulin dimer — the subunit of every doublet
Tubulin dimer — the subunit of every doublet ● LIVE · RCSB PDB 1JFF
open on RCSB PDB ↗
Loading live structure 1JFF from RCSB PDB…

The alpha-beta dimer is the building block of the microtubules in the axoneme. Doublets are made of a complete A-tubule and an incomplete B-tubule fused to it.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Ian Gibbons

Ian Gibbons

With A. J. Rowe he isolated the ATPase of cilia and named it dynein (1965); with Keith Summers he showed that ATP makes the outer doublets of a protease-treated sperm axoneme slide past one another (1971).

Gibbons and A. J. Rowe extracted a protein from the cilia of the protozoan Tetrahymena that split ATP. In their 1965 Science paper they called it dynein, after the dyne, a unit of force. A protein that splits ATP in cilia suggested a motor, but not what the motor did. If cilia bend, do the microtubules shorten on one side, or do they slide? In 1971 Keith Summers and Gibbons took isolated axonemes from the sperm of the sea urchin Tripneustes gratilla, digested them briefly with trypsin, and added 0.1 mM ATP. Watched by dark-field microscopy, the axonemes did not bend. They fell apart, with groups of doublets sliding out of the bundle. Their reading was that the links trypsin had cut normally resist sliding and turn it into bending. That assay, sliding apart in a dish, became a standard way to ask whether a flagellum's motors work.

✓ verbatim Gibbons & Rowe, Science 149:424–426 (1965); Summers & Gibbons, PNAS 68:3092–3096 (1971).
Björn Afzelius

Björn Afzelius

Described four men with immotile sperm, three of whom also lacked mucociliary transport, and found by electron microscopy that their cilia lacked dynein arms (1976).

Afzelius was an electron microscopist who studied the fine structure of sperm tails. His 1976 report in Science described four men whose sperm did not move. Three of them had frequent bronchitis and sinusitis, and when their tracheobronchial clearance was measured, no mucociliary transport could be detected: the cilia lining their airways were not moving mucus. Electron microscopy indicated that the dynein arms, the small projections on each outer doublet, were absent from their cilia. One structural defect thus accounted for two complaints in two organs, because sperm tails and airway cilia are built on the same axoneme. In the same report he suggested that ciliary motion might bear on which side of the body the internal organs develop, an idea the mouse node later made concrete. The work helped found the clinical study of what is now called primary ciliary dyskinesia and made the axoneme a target for human genetics.

✓ verbatim Afzelius, Science 193:317–319 (1976), 'A human syndrome caused by immotile cilia'.
Tatyana Svitkina

Tatyana Svitkina

With Gary Borisy, imaged the leading-edge actin network of Xenopus keratocytes and fibroblasts by electron microscopy, found it extensively branched, and localized the Arp2/3 complex to the branch points (1999).

A lamellipodium is so thin that individual actin filaments in it cannot be resolved by ordinary light microscopy. Svitkina, working with Gary Borisy at the University of Wisconsin in Madison, took the electron microscope route: extract the membrane with detergent, keep the cytoskeleton, and image it. In a 1997 paper on fish keratocytes the group decorated filaments with myosin fragments to read their polarity and found barbed ends predominantly facing forward. In 1999 Svitkina and Borisy looked at Xenopus keratocytes and fibroblasts and saw a network that branched extensively. Thinning the network with cytochalasin D let them measure single junctions, which averaged 67 ± 12 degrees across 212 of them, and immunogold labelling placed the Arp2/3 complex at the Y-shaped branch points. Location is not mechanism, and the authors framed it as suggesting a role in branch formation. Combined with biochemistry of purified Arp2/3, it gave the field a picture of the leading edge as a branched, forward-facing array that could push.

✓ verbatim Svitkina & Borisy, J. Cell Biol. 145:1009–1026 (1999); Svitkina, Verkhovsky, McQuade & Borisy, J. Cell Biol. 139:397–415 (1997).
Nobutaka Hirokawa

Nobutaka Hirokawa

His laboratory (Nonaka et al., 1998) reported that cilia of the mouse embryonic node move in a rotating pattern and drive a leftward flow, and that mice lacking the kinesin KIF3B lack nodal cilia and show randomized left-right asymmetry.

Hirokawa's laboratory in Tokyo studied kinesins, the plus-end motors that carry cargo along microtubules. One of them, KIF3B, turned out to matter for building cilia. In a 1998 Cell paper by Shigenori Nonaka and colleagues from the laboratory, mouse embryos lacking KIF3B had no cilia on the node, a small pit on the embryo's underside at the stage when the body's left and right are first set. In normal embryos those cilia moved, and their motion drove the fluid over the node toward the left. In the mutants the flow was absent and the left-right placement of organs such as the heart was randomized. Nodal cilia are mostly 9+0, without a central pair, yet they move. The paper tied an old clinical puzzle, why some people with immotile cilia have mirror-image organs, to a fluid flow that could be watched under a microscope.

✓ verbatim Nonaka et al., Cell 95:829–837 (1998).
Test yourself
Question 1 of 8 Score 0 / 0

In a 9+2 motile axoneme, where are the dynein arms anchored, and what do their motor heads act on?
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

  • Arp2/3 complex — the branch-maker at the leading edge — RCSB PDB 1K8K · coordinates CC0 · Bovine Arp2/3 complex, X-ray crystal structure (Robinson et al., Science, 2001)
  • Actin monomer — the subunit added at the barbed end — RCSB PDB 1ATN · coordinates CC0 · Rabbit skeletal muscle actin in complex with DNase I, X-ray crystal structure (Kabsch et al., Nature, 1990)
  • Tubulin dimer — the subunit of every doublet — RCSB PDB 1JFF · coordinates CC0 · Alpha-beta tubulin with taxol, refined electron crystallography structure (Löwe et al., J. Mol. Biol., 2001)

Simulations

  • Polymerization minus flow · Original interactive — Biochemistrypedia (Deppmann Lab).

Scientist profiles & quotes

  • Ian Gibbons — Gibbons & Rowe, Science 149:424–426 (1965); Summers & Gibbons, PNAS 68:3092–3096 (1971).
  • Björn Afzelius — Afzelius, Science 193:317–319 (1976), 'A human syndrome caused by immotile cilia'.
  • Tatyana Svitkina — Svitkina & Borisy, J. Cell Biol. 145:1009–1026 (1999); Svitkina, Verkhovsky, McQuade & Borisy, J. Cell Biol. 139:397–415 (1997).
  • Nobutaka Hirokawa — Nonaka et al., Cell 95:829–837 (1998).

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

Animated explainer

Sliding Becomes Bending · 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.

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

A crawling cell is like a person in a sleeping bag inching forward. It pushes its front out by building a scaffold, grabs the floor, and pulls its back along. A cilium is a tiny hair that waves. Inside, little motors try to slide the hair's rods past each other, and because the rods are tied together, the hair bends instead.
High school

Cells move in two main ways. Many crawl: actin protein builds a branched mesh that pushes the front edge forward, adhesion proteins called integrins grab the surface, myosin motors contract, and the back lets go. Others beat cilia or flagella, hair-like projections with nine pairs of microtubules around a central two. Dynein motors try to slide neighbouring pairs past each other. Links between the pairs stop them sliding freely, so the cilium bends. If the dynein is missing, sperm cannot swim and airway cilia cannot sweep away mucus.
College

Lamellipodial crawling cycles through protrusion, adhesion, contraction and release. The Arp2/3 complex nucleates branches off existing filaments, building a dendritic network with barbed ends toward the membrane; polymerization pushes the edge. Membrane tension and myosin II pull the network rearward (retrograde flow). Integrins in focal adhesions act as a clutch: the more the network is coupled to the substrate, the less polymerization is lost to flow, so edge speed is polymerization minus flow. Myosin II contraction and adhesion disassembly let the rear detach. In a 9+2 axoneme, axonemal dyneins anchored on each A-tubule walk toward the minus end along the neighbouring B-tubule. With radial spokes, the central pair and the nexin-dynein regulatory complex constraining sliding, and the base anchored, shear becomes a bend. Primary cilia are non-motile 9+0 sensory organelles; mouse nodal cilia are 9+0 yet rotate.
Graduate

The clutch picture treats edge velocity as polymerization velocity minus retrograde flow, with adhesion engagement setting how much flow is converted to traction. Measurements in PtK1 cells (Gardel et al., 2008) show the relationship is not monotonic: traction stress and local F-actin speed were biphasically related, with a switch near 8 to 10 nm/s, so neither fast flow nor strong adhesion predicts traction or speed on its own. In CHO cells, rear release limited speed at intermediate and high adhesiveness. Confined leukocytes can migrate without transmembrane force coupling, using flow that follows substrate texture (Reversat et al., 2020). In the axoneme, sliding was shown directly in protease-treated sea-urchin sperm axonemes (Summers and Gibbons, 1971). How motors on opposite sides take turns is still argued. Cryo-ET of rapidly frozen beating sea-urchin flagella found asymmetric dynein states across bends, with most dyneins in an active conformation, which the authors read as favouring switching by selective inhibition. Open question: what signal, mechanical or chemical, tells each set of arms to switch off, and is it the same in a nodal cilium that rotates as in a sperm tail that waves?
A motor named after a unit of force

When Ian Gibbons and A. J. Rowe purified an ATP-splitting protein from Tetrahymena cilia, they needed a name for something whose job they could guess but not yet prove. They chose dynein, from the dyne, the old unit of force. The name was a bet that the protein produced force on the outer doublets. Six years later the sliding-disintegration assay with sea-urchin sperm axonemes showed that ATP makes doublets slide, and the bet was paid. The same word now covers the cytoplasmic dyneins that carry cargo toward microtubule minus ends.

— Gibbons & Rowe, Science 149:424–426 (1965).
One mutation, two kinds of cilia

Mice lacking the protein Cfap53 make a useful warning. In their embryonic node the cilia were completely immotile and had lost their outer dynein arms. Yet in the trachea and the brain ventricles, cilia of the same mice still moved, only with an abnormal beat. The same missing protein gave a stopped cilium in one tissue and a misbeating one in another. The lesson for anyone reading a ciliary mutant is to ask which tissue was looked at before generalizing the phenotype to every cilium in the body.

— Ide et al., PLoS Genetics 16:e1009232 (2020).

 ↑ Top