Cell Biology
Microtubule Dynamic Instability
The GTP cap, catastrophe and rescue — how a polymer that switches at random builds the mitotic spindle
Microtubule dynamic instability is the stochastic switching of a single microtubule between steady growth and abrupt, rapid shrinkage, powered not by reaching an equilibrium but by GTP hydrolysis on β-tubulin. Each subunit is an αβ-tubulin heterodimer; a cap of freshly-added GTP-tubulin at the plus end holds the lattice straight and lets the polymer grow at roughly 0.5 to 2 µm/min. When hydrolysis catches up and the cap is lost, the strained GDP lattice peels apart into curling protofilaments and the microtubule depolymerizes at up to ~20 µm/min — a catastrophe — after which it may stochastically rescue and grow again. Tim Mitchison and Marc Kirschner discovered this behavior in a 1984 Nature paper; it is the engine of mitotic search-and-capture and the target of paclitaxel (Taxol), colchicine, and the vinca alkaloids.
- Growth rate~0.5–2 µm/min
- Shrinkage rateup to ~20 µm/min
- Subunitαβ-tubulin + GTP on β
- Lattice13 protofilaments
- DiscoveredMitchison & Kirschner 1984
- DrugsTaxol, colchicine, vinca
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
Why dynamic instability matters
- It searches space without a map. A dividing cell must attach a spindle microtubule to the kinetochore of every chromosome, but the centrosome has no directions to any of them. Dynamic instability turns the problem into trial and error: microtubules grow out in every direction, and any that miss undergo catastrophe and are recycled. This search-and-capture is the cellular reason mitosis works at all.
- It is a non-equilibrium behavior. Unlike an actin filament near its critical concentration, a microtubule at a fixed tubulin concentration can have some polymers growing and others shrinking simultaneously. The energy that pays for this comes from GTP hydrolysis — roughly one GTP burned per dimer incorporated — making the microtubule a molecular machine, not a passive rope.
- It is the reason taxanes and vinca alkaloids are frontline chemotherapy. Paclitaxel (from the Pacific yew, Taxus brevifolia) and vincristine (from the Madagascar periwinkle) are among the most-used anticancer drugs in the world precisely because a cell that cannot rearrange its microtubules cannot build a spindle and dies in mitosis.
- It builds and remodels the whole cytoskeleton on demand. The interphase microtubule array turns over with a half-life of minutes, letting the cell re-route intracellular traffic, reposition organelles, and reshape itself far faster than a stable polymer ever could. Neurons exploit this to extend and retract growth cones while wiring the brain.
- It is tunable by dozens of regulators. Catastrophe factors (kinesin-13/MCAK, kinesin-8, stathmin/Op18), rescue and growth factors (XMAP215/ch-TOG, CLASPs, EB1 and the +TIP network) let the cell set microtubule lifetime locally — long and stable down an axon, short and exploratory near a leading edge.
- It converts chemical energy into mechanical force. A depolymerizing microtubule stores strain in its curved GDP protofilaments; when it peels apart at an attached kinetochore, that release can pull a chromosome. Dynamic instability is therefore not just a search algorithm but a force generator for chromosome segregation.
Common misconceptions
- "Dynamic instability is the same as treadmilling." No. Treadmilling is net addition at one end and net loss at the other, giving a filament of roughly constant length that appears to migrate — the dominant behavior of actin and of microtubules only under some conditions. Dynamic instability is a single end (usually the plus end) switching between growth and shrinkage. A treadmilling polymer never catastrophically depolymerizes; a dynamically unstable one does.
- "GTP is the fuel that pushes polymerization." GTP hydrolysis is not what drives subunit addition — addition is driven by tubulin concentration and favorable lattice contacts. Hydrolysis is a timer that destabilizes the lattice after the fact, storing the energy that later powers rapid disassembly. A non-hydrolyzable analog (GMPCPP) makes microtubules grow and stay stable, precisely because the destabilizing timer never fires.
- "Both α and β tubulin hydrolyze GTP." Only the exchangeable GTP on β-tubulin (the E-site) is hydrolyzed. The GTP on α-tubulin (the N-site) is buried at the intradimer interface and never exchanges or hydrolyzes — it is a structural cofactor. The E-site GTP sits at the interface between one dimer and the next, so its hydrolysis is triggered by the arrival of the following dimer.
- "Taxol works by making microtubules longer." At high concentrations paclitaxel does over-polymerize tubulin, but its therapeutic mechanism at nanomolar doses is to suppress dynamic instability — freezing the plus end so it neither grows nor shrinks. A frozen spindle cannot correct attachment errors, so the spindle-assembly checkpoint stays on and the cell dies. Kinetic suppression, not net stabilization, is the point.
- "Catastrophe happens when tubulin runs out." Falling tubulin concentration raises catastrophe frequency, but catastrophes occur even at high, constant tubulin — that was the whole surprise of the 1984 experiment. Catastrophe is a stochastic loss of the GTP cap, not the crossing of a concentration threshold.
- "The GTP cap is a large structure." Estimates vary widely, but many measurements put the stabilizing cap at only a few layers to a few dozen dimers under fast growth — sometimes argued to be as small as a single layer of GTP-tubulin. It is a razor-thin kinetic edge, which is exactly why loss of the cap and catastrophe can be so sudden.
How dynamic instability works, step by step
The polymer is a hollow tube, typically 25 nm across, built from 13 parallel protofilaments of head-to-tail αβ-tubulin heterodimers. Because every dimer points the same way, the tube is polar: the fast-growing plus end exposes β-tubulin, the slower minus end exposes α-tubulin and is usually anchored at a nucleating site such as the γ-tubulin ring complex of the centrosome. All the dynamics of interest happen at the plus end.
Growth and the GTP cap. A soluble αβ-tubulin dimer arrives carrying GTP on its β-subunit E-site. It docks onto the plus end and packs into a straight conformation that fits the lattice. Incorporation triggers hydrolysis of that E-site GTP to GDP, but hydrolysis lags a beat behind addition. As long as new GTP-dimers arrive faster than the tip's dimers are hydrolyzed, a layer of GTP- (and transient GDP·Pi-) tubulin sits at the very end. This GTP cap clamps the strained GDP lattice beneath it into a straight, stable tube. The microtubule elongates at roughly 0.5 to 2 µm/min in cells.
Catastrophe. Growth is a race between subunit addition and the advancing wave of hydrolysis. If the tip stalls, or hydrolysis simply catches up, the terminal dimers become GDP-bound and the cap is lost. GDP-tubulin prefers a curved conformation and has been holding elastic strain like a compressed spring; without the cap, that strain is released. The 13 protofilaments splay outward into ram's-horn curls, lateral bonds between them break, and the tube unzips. This is a catastrophe, and the microtubule now shrinks at up to ~20 µm/min — an order of magnitude faster than it grew.
Rescue. A shrinking microtubule is not doomed. It can stochastically stop depolymerizing and resume growth — a rescue — by re-establishing a GTP cap. Rescues are thought to occur at buried islands of GTP- or GDP·Pi-tubulin left in the aging lattice, or where rescue factors such as CLASP proteins bind. In a living cell, rescue is far more frequent than in purified tubulin, which is one reason cellular microtubules do not simply disassemble to nothing.
Regulation. The cell tunes all four parameters — growth rate, shrinkage rate, catastrophe frequency, rescue frequency — with a large cast of proteins. XMAP215/ch-TOG is a polymerase that speeds growth by feeding dimers to the tip. EB1 and the +TIP network recognize the GTP-cap conformation and recruit other regulators to growing ends. Kinesin-13 (MCAK) and kinesin-8 are depolymerases that actively promote catastrophe; stathmin/Op18 sequesters free dimers and raises catastrophe frequency. By deploying these locally, a neuron can keep an axonal microtubule stable for hours while a mitotic microtubule turns over in tens of seconds.
Dynamic instability vs treadmilling
| Feature | Dynamic instability | Treadmilling |
|---|---|---|
| Behavior | One end switches between growth and rapid shrinkage | Net addition at plus end, net loss at minus end |
| Length over time | Fluctuates sharply per polymer; population length distribution is broad | Roughly constant; polymer appears to migrate |
| Nucleotide basis | GTP cap at tip; catastrophe on cap loss | Difference in critical concentration between the two ends |
| Catastrophic collapse | Yes — signature of the mechanism | No |
| Dominant in | Interphase array, mitotic spindle (plus ends) | Some microtubule arrays, and classic for actin filaments |
| Energy source | GTP hydrolysis on β-tubulin | Nucleotide hydrolysis biasing on/off rates at each end |
| Discovered / framed by | Mitchison & Kirschner, 1984 | Wegner (actin), 1976; applied to microtubules later |
Anti-microtubule drugs at a glance
| Drug (class) | Binding site | Effect on microtubule | Net effect on dynamics | Clinical use |
|---|---|---|---|---|
| Paclitaxel / docetaxel (taxane) | β-tubulin taxane pocket, luminal wall | Stabilizes lattice in straight, GTP-like state | Suppresses dynamics (kinetic freeze) | Breast, ovarian, lung cancer |
| Colchicine | α–β intradimer interface on free tubulin | Poisons plus end once incorporated; blocks addition | Inhibits polymerization | Gout, familial Mediterranean fever (not oncology) |
| Vinblastine / vincristine (vinca) | Plus end, vinca domain on β-tubulin | Induces curved protofilaments; caps the end | Promotes depolymerization | Leukemia, lymphoma, solid tumors |
| Nocodazole (lab tool) | Colchicine site | Rapid, reversible depolymerization | Inhibits polymerization | Research reagent |
| Epothilones | Overlaps taxane pocket | Stabilizes lattice like taxanes | Suppresses dynamics | Taxane-resistant breast cancer (ixabepilone) |
Famous experiments and history
- Mitchison & Kirschner (1984). Working at UCSF, Tim Mitchison and Marc Kirschner measured the length distributions of microtubules assembled from purified tubulin and found they could not be explained by a single critical concentration: at a fixed tubulin level, some microtubules grew while others shrank, and individual polymers switched abruptly. Their paper, "Dynamic instability of microtubule growth" (Nature 312: 237–242), proposed the GTP-cap model and named the phenomenon. A companion paper the same year worked out the theory.
- Horio & Hotani (1986). Using dark-field microscopy, they watched single microtubules in real time and directly saw individual polymers growing, catastrophically shrinking, and rescuing — the first direct visual confirmation of Mitchison and Kirschner's prediction, published in Nature.
- Search-and-capture (Kirschner & Mitchison, 1986). The same duo proposed that dynamic instability lets centrosomal microtubules find kinetochores by random growth and selective stabilization on contact — the founding model for how the spindle assembles. Later work added chromosome-directed nucleation (the RanGTP gradient) and branching amplification (augmin), but the stochastic engine remained.
- GMPCPP and the cap concept. Microtubules grown with the slowly-hydrolyzable GTP analog GMPCPP are hyper-stable and rarely undergo catastrophe, and structural studies of GMPCPP versus GDP lattices revealed the ~conformational and lattice-spacing change that hydrolysis produces — direct physical evidence that the nucleotide state of the tip governs stability.
- Taxol's mechanism (Horwitz, 1979 onward). Susan Horwitz showed that paclitaxel, isolated from the Pacific yew by Wall and Wani, uniquely promotes tubulin assembly and stabilizes microtubules — the first drug known to work this way. Decades of follow-up established that its clinical action is suppression of dynamic instability, not bulk over-polymerization, and cryo-EM later mapped the taxane pocket on the microtubule luminal surface.
- EB1 comet imaging. Fluorescently tagged EB1 binds only the growing GTP-cap region of plus ends, appearing in live cells as bright comets that streak outward and vanish at each catastrophe — turning dynamic instability into something you can watch directly under a light microscope and quantify genome-wide.
Frequently asked questions
What is the GTP cap and why does it stabilize a microtubule?
Each microtubule subunit is an αβ-tubulin heterodimer. The β-tubulin subunit carries an exchangeable GTP at the interface between dimers; the α-tubulin GTP is buried and non-hydrolyzable. When a GTP-bound dimer adds to the growing plus end it packs into a straight, relaxed conformation that fits the tubular lattice. GTP hydrolysis to GDP is triggered by incorporation and lags slightly behind addition, so as long as dimers arrive faster than they are hydrolyzed, a layer of GTP- (or transient GDP·Pi-) tubulin persists at the very tip. That layer — the GTP cap, estimated at anywhere from a single layer up to a few dozen to a few hundred dimers depending on growth rate — clamps the underlying strained GDP lattice into a straight configuration. GDP-tubulin on its own prefers a curved conformation and stores mechanical strain like a compressed spring; the cap holds that spring shut. Lose the cap, and the strain is released.
What causes a microtubule catastrophe?
A catastrophe is the switch from growth to rapid shrinkage. It happens when the stabilizing GTP cap is lost — when GTP hydrolysis catches up to subunit addition and reaches the tip. Once the terminal dimers are GDP-bound, the lattice can no longer hold its straight, strained conformation: protofilaments splay outward into characteristic ram's-horn curls and peel away from one another, and the microtubule depolymerizes at up to roughly 20 µm/min, far faster than it grew. Catastrophe is stochastic rather than clock-like — the probability per unit time rises when growth slows, when the microtubule is longer (aged lattice), when free tubulin is scarce, and when catastrophe-promoting proteins such as kinesin-13 (MCAK) and kinesin-8 act on the tip. It is not triggered by hitting a critical concentration; it is a kinetic consequence of GTP hydrolysis outrunning the cap.
What is the difference between rescue and catastrophe?
They are the two transitions of dynamic instability, in opposite directions. A catastrophe is the switch from the growing phase to the shrinking phase, caused by loss of the GTP cap. A rescue is the reverse — a shrinking microtubule stochastically stops depolymerizing and resumes growth, re-establishing a cap. Rescue is thought to occur at islands of GTP- or GDP·Pi-tubulin that remain buried in the GDP lattice, or where rescue factors such as CLASP proteins and the +TIP network stabilize the tip. In a typical vertebrate interphase cell a microtubule spends most of its time either growing (~0.5 to 2 µm/min) or shrinking (~10 to 20 µm/min), with catastrophes every minute or two and rescues that are far more frequent inside the cell than in purified tubulin. The balance of catastrophe and rescue frequencies sets the average microtubule length and lifetime.
How does dynamic instability drive mitotic search-and-capture?
During mitosis, spindle microtubules must find and attach to the kinetochore of every chromosome — dozens of tiny targets scattered through the cell — without any map. Dynamic instability solves this by trial and error: microtubules nucleated at the centrosomes grow out in all directions, and any that fail to hit a kinetochore undergo catastrophe, shrink back, and are replaced by new probes growing in fresh directions. A microtubule that happens to contact a kinetochore is captured and selectively stabilized, converting a random search into a directed attachment. Mitchison and Kirschner proposed this 'search-and-capture' model in 1986. Modern work shows the search is accelerated by biased nucleation around chromosomes (the RanGTP gradient and augmin-mediated branching), but stochastic growth and catastrophe remain the engine. This is also why the spindle is exquisitely sensitive to drugs that freeze microtubule dynamics.
How do taxol and colchicine affect microtubules?
Both are anti-microtubule drugs, but they act at opposite ends of the polymer's life. Paclitaxel (Taxol) binds a pocket on β-tubulin on the inside of the microtubule wall and stabilizes lateral and longitudinal contacts, locking the lattice in a straight, GTP-like conformation even after hydrolysis. It does not simply make microtubules longer — at low, clinically relevant concentrations it suppresses dynamic instability, damping both growth and shrinkage so the microtubule becomes kinetically frozen. Colchicine and the vinca alkaloids (vinblastine, vincristine) do the opposite: colchicine binds free tubulin dimers at the α–β intradimer interface and, once incorporated at the tip, poisons the ends and blocks further addition; vinca alkaloids bind the plus end and induce curved protofilaments, promoting depolymerization. In every case the therapeutic effect is the same — a cell whose microtubules cannot dynamically rearrange cannot build a functional mitotic spindle, so it arrests at the spindle-assembly checkpoint and dies. This is why these drugs are frontline chemotherapeutics for breast, ovarian, lung, and hematologic cancers.
How were microtubule dynamics discovered?
Before 1984, microtubules were thought to be near-equilibrium polymers whose length was set by a single critical concentration of tubulin — grow above it, shrink below it, treadmill in between. Tim Mitchison and Marc Kirschner, then at UCSF, measured the length distributions of microtubules assembled from purified tubulin and found something a simple equilibrium could not explain: at a fixed tubulin concentration, some microtubules grew while others shrank at the same time, and individual polymers switched abruptly between the two. They published 'Dynamic instability of microtubule growth' in Nature in 1984, proposing that GTP hydrolysis, not a critical concentration, controlled the behavior, and that a GTP cap distinguished growing from shrinking ends. The prediction was confirmed directly in the late 1980s by dark-field microscopy of single microtubules (Horio and Hotani, 1986) and later by fluorescence at the single-molecule level. The model reshaped how biologists think about the cytoskeleton and mitosis.