# Cell Shape, Migration and Cilia

> In a crawling cell and in a beating cilium, polymers and motors make internal motion, and constraints such as adhesion and the links between doublets help shape what visible movement it becomes. 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.

Course: Cell Biology (Traffic and movement). Level: core university (third year). About 18 minutes.
Lesson page (link this one when citing): https://stemhop.com/lessons/migration-and-cilia/

Builds on: Cytoskeletal Polymers and Dynamics · Molecular Motors and Cargo

## What you'll learn

- Describe the crawling cycle (Arp2/3-driven protrusion, integrin adhesion, myosin II contraction, rear release) and name at least one mode of crawling that works differently.
- Calculate leading-edge speed from a polymerization rate and a retrograde flow rate, and explain why stronger adhesion or faster flow does not always mean faster movement.
- Describe the 9+2 axoneme and explain how dynein sliding between doublets, resisted by links, becomes bending; distinguish motile, primary and nodal cilia.
- Connect a ciliary or migration defect to the cellular phenotype it produces.

The first film built both machines in their simplest form. It left out the numbers, the rest of a crawling cell, and what happens when a part breaks; this page adds them in that order.

> **The one idea to carry through this lesson.** In these two systems, polymers and motors create internal relative motion, while mechanical constraints and control of which motors are active shape its visible output. The first film used the simplest case of each; the sections below give the fuller picture. The rule does not cover every kind of cell movement.

For background, review the underlying [cytoskeletal polymers](https://stemhop.com/lessons/cytoskeleton-dynamics/) and [molecular motors](https://stemhop.com/lessons/molecular-motors/).

## A crawling edge: measure the hidden motion

Put numbers on the balance, with polymerization measured against the actin network and retrograde flow and the edge against the substrate:

> **edge speed = polymerization speed − retrograde-flow speed**

This one-dimensional relation predicts the leading edge, not automatically the cell body.

**Predict.** Filaments extend at 40 nm/s relative to the network while the network flows rearward at 25 nm/s. How fast does the edge advance, and how far in 120 s?

> **Check your prediction.** Take forward as positive. Edge speed = 40 + (−25) = 15 nm/s. In 120 s the edge advances 15 × 120 = 1,800 nm = 1.8 µm. Check the same result from the two contributions: polymerization adds 4,800 nm and flow subtracts 3,000 nm.

Only the flow value comes from the experiment behind this example. In PtK1 cells—a rat-kangaroo epithelial line—on flexible, fibronectin-coated gels, Gardel and colleagues measured distal actin flow at about 25 nm/s. The 40 nm/s polymerization rate is an illustrative assumption, and the calculation holds both rates constant.

Clusters of integrins and linker proteins form **focal adhesions**, substrate anchors that [the adhesion lesson](https://stemhop.com/lessons/adhesion-and-ecm/) develops. In the first film’s analogy, polymerization is the engine and adhesion is the clutch.

**Predict.** Keep polymerization at 40 nm/s, but let stronger clutch engagement reduce rearward flow from 25 to 10 nm/s. By what factor does the edge’s 120-second advance change?

> **Check your prediction.** Edge speed becomes 40 − 10 = 30 nm/s. The edge advances 3,600 nm, or 3.6 µm, in 120 s. Here, cutting flow from 25 to 10 nm/s doubles edge advance, from 1.8 to 3.6 µm.

*Interactive on the lesson page:* Lamellipodium clutch — edge speed = polymerization − flow

## From edge advance to whole-cell crawling

The first film stopped at edge advance; a whole crawling cell also has to move its body and release its rear.

The actin-rich lamellipodium is a sheet of cytoplasm a fraction of a micrometre thick. At its front, actin subunits add at **barbed ends**, the fast-growing filament ends, and push the membrane. The **Arp2/3 complex**, a seven-protein assembly, starts a branch from an existing filament. Rac, a small signalling switch, helps activate Arp2/3 there; [integrins and Rho GTPases](https://stemhop.com/lessons/integrins-and-rho-gtpases/) develops that control. Behind the front, non-muscle myosin II—a two-headed relative of muscle myosin—forms small filaments and contracts actin. In fish keratocytes, evidence supported contraction where the lamellipodium meets the cell body as a driver of body movement.

The membrane pushing back on growing filaments and myosin II pulling the network rearward both contribute to retrograde flow.

Together, protrusion, grip, contraction and release form the crawling cycle. These four processes run at the same time in a smoothly gliding keratocyte; in cells such as fibroblasts they are more cyclic. At the rear, adhesions must release. Separately, filaments that come apart a short way behind the leading edge return monomers for growth at the front.

*Figure (AI-generated illustration).* A crawling cell coordinates four concurrent processes: actin pushes the front, integrins grip, myosin II contracts the network, and the rear releases. The inset shows Arp2/3 at a branch. Image: side view of a crawling cell moving right, with a flat lamellipodium of branched actin at the front, integrin adhesions on the substrate, myosin II behind, and torn integrins left at the rear; an inset shows an Arp2/3 complex at a Y-shaped branch.

The clutch remains a starting model. In the PtK1 cells, traction stress—the force per area exerted on the substrate—and local actin speed had a two-part relationship that changed near 8 to 10 nm/s, so faster flow does not always mean more traction. Stronger adhesion can also slow migration by resisting rear release.

**Predict.** CHO cells engineered to express integrins locked in a high-affinity state grip fibronectin strongly but migrate slowly. Which crawling process is most likely limiting?

> **Check your prediction.** Rear release. In CHO cells on matrix-coated glass, rear retraction often tore integrins from the cell and left them behind. At intermediate and high adhesiveness, rear retraction appeared to limit migration speed. Strong coupling at the front cannot help if the back cannot let go.

That cell failed because one part worked too well. The second film picks up here: what protease treatment or a missing part reveals, first in a dish and then in the body.

**Film transcript: Reading a Broken Machine**

### Film transcript: Reading a Broken Machine

In chapter 1, we cut the links in a drawing, and the doublets slid apart instead of bending. Now turn the question around. When a real cilium stops, which part broke, and where in the body does it show? Summers and Gibbons tested that idea in 1971.

They took axonemes from sea urchin sperm. They briefly treated them with trypsin, a protein-cutting enzyme, then added 0.1 mM ATP. Under dark-field microscopy, the axonemes did not beat. Groups of doublets slid out of the bundle. That dish test became a standard way to ask whether a flagellum's motors work. Now the body. In 1976, Björn Afzelius described four men whose sperm did not swim.

Electron microscopy indicated that the dynein arms, small projections on each outer doublet, were missing from their cilia. Sperm tails and the cilia lining the airways are built on the same basic axoneme. So predict: if the dynein arms are missing, what else should fail?

Three of the four men also had frequent bronchitis and sinusitis. Their airway cilia were not moving mucus, a sweep called mucociliary transport. One missing part disrupted motion in two organs. The match is not perfect: sperm and airway cilia differ in some dynein parts. And one mutation need not act the same everywhere. In mice lacking a protein called Cfap53, cilia on the embryonic node stopped completely. Windpipe and brain-ventricle cilia still beat, but abnormally. The instrument matters too.

Patients with DNAH11 mutations had normal-looking nasal cilia by conventional electron microscopy. Three-dimensional electron tomography found less outer-arm volume near the base. Crawling cells break the same way. Chapter 1's clutch needed integrins to grip. Neutrophils without a working integrin of the beta β2 family grip the vessel wall poorly. They largely fail to leave the blood, so infections are poorly controlled. Read a phenotype backward, tissue by tissue. Ask which machine failed, which part is defective, which tissue you examined, and with what instrument. Motors and constraints from chapter 1 become the places to look.

From the film

- After brief trypsin treatment, sperm axonemes given ATP slid apart instead of beating.
- Afzelius's four patients had immotile sperm; three had no detectable mucociliary transport.
- Sperm tails and airway cilia share a basic axoneme, though some dynein parts differ.
- One mutation can stop cilia in one tissue and only disturb the beat in another.
- Neutrophils without a working β2-family integrin largely fail to leave the blood.

## The axoneme: what the model predicts

The dish test in the second film disrupted the axoneme’s resistance; here is the fuller machine it acted on.

In the classic 9+2 layout, nine outer doublets, each a complete A-tubule fused to an incomplete B-tubule, surround a central pair of single microtubules. The ring grows from a **basal body**, which anchors the axoneme in the cell. Dynein arms sit in outer and inner rows on each A-tubule; their motor heads grip the neighbouring B-tubule and walk toward its minus end, at the base, so each doublet pushes its neighbour toward the tip. Protein assemblies called **radial spokes** reach toward the central-pair complex and help regulate dynein. Between neighbouring doublets, the **nexin-dynein regulatory complex** links the pair, resists sliding and helps regulate the motors.

To feel how resisted shear can make curvature, hold two card strips side by side, joined at one end and tied loosely along their length. Push one strip lengthwise: because the pair cannot freely slide apart, it curves. In the axoneme, interdoublet resistance and base anchoring similarly turn dynein-driven shear into bending, while regulated motor activity helps determine where the bend forms. This is a model, not a full waveform.

Summers and Gibbons tested whether dynein changes doublets’ relative positions or shortens microtubules. They briefly treated sea-urchin sperm axonemes with trypsin, a protein-cutting enzyme that disrupted structures normally resisting sliding, then added 0.1 mM ATP. Under dark-field microscopy, groups of doublets slid out of the bundle. Disrupting the resistance exposed ATP-driven sliding as disintegration rather than bending.

### Design problem

A *Chlamydomonas* mutant, or a patient’s sperm, has paralysed flagella. Design a test of whether its dynein motors themselves work. Isolate the axonemes first, taking the membrane and the cell’s own ATP supply out of the experiment.

**Predict.** After brief trypsin treatment and ATP, the mutant doublets slide apart like wild type. What has the result ruled in, and what has it not ruled out?

> **One way to reason through it.** Sliding apart shows that at least some dynein arms can bind the neighbouring B-tubule, use ATP and generate motion. It does not show that both arm rows are intact; sliding can survive loss of one row. If both rows are present, a paralysed but sliding axoneme probably has a defect in motor regulation or in converting sliding into bending, such as in the radial spokes, central pair or nexin-dynein regulatory complex.
>
> If doublets do not slide apart, suspect missing or inactive dynein arms. Confirm structures by electron microscopy, using three-dimensional electron tomography if conventional images look normal, as they did in some DNAH11 cases.
>
> Controls are wild-type axonemes treated the same way, mutant and wild-type axonemes without ATP, and axonemes given ATP without trypsin. Sliding apart is expected only with ATP after the resistance has been cut. Intact wild-type axonemes given ATP should still slide internally but bend or beat rather than fall apart.

Even a single bend needs dyneins on one side to be more active than those on the other; to make a wave, the sides must take turns.

## From a broken part to a cellular phenotype

Constraints and motors shape the output, so a broken part changes it. In 1976 Björn Afzelius described four men whose sperm could not swim. Three also had frequent bronchitis and sinusitis and no detectable mucociliary transport, and [electron microscopy](https://stemhop.com/methods/electron-microscopy/) indicated that their cilia lacked dynein arms. One missing part disrupted motion in two organs because sperm tails and airway cilia share the same basic axoneme, though some of their dynein parts differ. **Mucociliary transport** is the sweep of mucus out of the airways, and airway cilia that do not beat severely impair it. The **node** is a small pit on the early mouse embryo, and nodal cilia that do not rotate cannot drive the leftward flow over it.

**Predict.** Mice lack the kinesin motor KIF3B and therefore have no cilia on the embryonic node. What happens to left-right organ placement?

> **Check your prediction.** It becomes randomized because the node cannot make its normal leftward flow. Some people with immotile but present cilia have mirror-image organs; inherited disorders of motile cilia like theirs are called **primary ciliary dyskinesia**. KIF3B loss is different: it removes cilia altogether.

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

Migration defects follow the same logic. In leukocyte adhesion deficiency type I, neutrophils without functional β2-family integrins grip the vessel wall poorly and largely fail to leave the blood, so infections are poorly controlled.

## Going deeper

Going deeper adds exceptions and methods.

### Other ways to migrate

Not every cell uses lamellipodial clutch-based crawling. Some form **blebs**, rounded membrane bulges driven by pressure from contraction of the actin cortex under the membrane. Confined leukocytes can move without transmembrane force coupling, using actin flow that follows the texture of their surroundings (Reversat and colleagues, 2020). Developing neurons and neural crest cells use versions with their own features; see [the growth cone](https://stemhop.com/lessons/growth-cone/), [neural crest](https://stemhop.com/lessons/neural-crest/), [radial migration](https://stemhop.com/lessons/radial-migration/) and [tangential migration](https://stemhop.com/lessons/tangential-migration/).

### What the leading-edge images showed

In extracted fish keratocytes, myosin-fragment decoration showed that barbed ends predominantly face forward. Svitkina and Borisy imaged extracted lamellipodia of *Xenopus* keratocytes and fibroblasts by [electron microscopy](https://stemhop.com/methods/electron-microscopy/) and found extensive branching, with Arp2/3 at Y-shaped branch points. In keratocytes thinned with cytochalasin D, branches met their mother filaments at 67 ± 12 degrees.

### Cilia that bend the simple categories

Not every cilium is 9+2 or motile. **Primary cilia** are single, non-motile 9+0 sensory organelles on many vertebrate cell types; they lack dynein arms and act as signalling antennae (see [hedgehog at the primary cilium](https://stemhop.com/lessons/hedgehog-and-cilium/)). Cilia of the mouse embryonic node are mostly 9+0. Those near its centre move by rotating rather than waving; those at its edge are thought to be non-motile and to sense the flow.

### Repeats and the switching proposal

[Cryo-electron tomography](https://stemhop.com/methods/cryo-em/) of *Chlamydomonas* axonemes identified the nexin link as the dynein regulatory complex, which supplied the combined name used above. In *Chlamydomonas*, the set of arms, spokes and links repeats every 96 nm along a doublet.

Cryo-electron tomography of rapidly frozen, beating sea-urchin sperm flagella found asymmetric dynein states across bends, with most dyneins in an active conformation. The authors suggested that the sides switch by selectively inhibiting one set of motors. This is a proposal; the switching signal remains open.

---

From STEMhop (https://stemhop.com), free lessons by Christopher Deppmann. Lesson text © the authors. Practice questions: https://stemhop.com/lessons/migration-and-cilia/
