The Mitotic Spindle: How a Cell Pulls Its Chromosomes Apart

What the spindle is

The mitotic spindle is a transient, self-organizing structure built almost entirely from microtubules — hollow, 25-nm-wide tubes assembled from αβ-tubulin dimers — plus hundreds of associated proteins. In a typical animal cell it is a bipolar array: two spindle poles (in animals, organized by centrosomes containing a pair of centrioles) with microtubules radiating between and outward from them. Three functional classes of microtubule make up the array:

• Kinetochore microtubules (kMTs) — their plus ends attach to the kinetochore, a protein plate on each chromosome's centromere. In human cells a single kinetochore captures roughly 20–30 of them; in budding yeast, exactly one.
• Interpolar (overlap) microtubules — grow from opposite poles and interdigitate at the spindle midzone, where motors and crosslinkers push the poles apart.
• Astral microtubules — radiate from the poles to the cell cortex, anchoring and positioning the whole spindle and orienting the future division plane.

The entire apparatus in a human cell is only about 10–15 µm from pole to pole and lasts perhaps 20–30 minutes before it is disassembled and recycled.

Dynamic instability: the engine of assembly

The spindle's defining trick is that its microtubules are never at rest. Each one stochastically switches between growing and shrinking — a behavior called dynamic instability, discovered by Tim Mitchison and Marc Kirschner in 1984. Growth happens when GTP-bound tubulin dimers add to the plus end faster than the GTP hydrolyzes, leaving a stabilizing GTP cap. When the cap is lost, the tube undergoes a catastrophe: its protofilaments curl outward like ram's horns and it depolymerizes explosively. A rescue can recap it and restart growth.

Because a microtubule that hasn't found a chromosome quickly shrinks back, the cell can 'search and capture' — Kirschner and Mitchison's 1986 model — trying countless microtubule geometries per minute until plus ends stumble onto kinetochores and are stabilized. The chemical cost is real: each tubulin subunit hydrolyzes one GTP, so the spindle burns nucleotide continuously to stay searchable. Drugs that freeze this dynamism — taxol (stabilizes) and vinca alkaloids or nocodazole (destabilize) — arrest cells in mitosis and are frontline chemotherapeutics precisely because dividing cells cannot build a functional spindle without it.

Building a bipolar machine

Bipolarity is not guaranteed — it is assembled and enforced. In late prophase the duplicated centrosomes separate to opposite sides of the nucleus, driven partly by the tetrameric motor Eg5 (kinesin-5), which crosslinks antiparallel interpolar microtubules and slides them apart. When the nuclear envelope breaks down at prophase/prometaphase, microtubules invade the former nuclear space.

Two additional assembly pathways feed the array. The chromatin-mediated pathway, centered on the small GTPase Ran, generates a gradient of Ran-GTP highest near chromosomes; this locally releases spindle-assembly factors (like TPX2) so microtubules nucleate right where the chromosomes are. The augmin/HAUS pathway nucleates new microtubules off the sides of existing ones, amplifying density within the spindle body. Nucleation itself depends on the γ-tubulin ring complex (γ-TuRC), which templates the 13-protofilament lattice and caps minus ends at the poles. Even plant and oocyte cells that lack centrosomes build perfectly bipolar spindles this way — proof that the poles are an emergent property of motors and crosslinkers, not a rigid scaffold.

Attachment, correction, and the checkpoint

Capture is only the beginning; the attachment must be bi-oriented (amphitelic) — the two sister kinetochores linked to opposite poles. Wrong geometries occur constantly: syntelic (both sisters to one pole) and merotelic (one kinetochore to both poles) attachments would cause missegregation. The kinase Aurora B, sitting at the inner centromere, senses the low tension of incorrect attachments and phosphorylates kinetochore proteins (the Ndc80/Hec1 complex) to release the errant microtubules for another try — an error-correction loop that runs until only tension-bearing, correct attachments survive.

Meanwhile the spindle assembly checkpoint (SAC), discovered genetically by Li & Murray and Hoyt et al. in 1991, watches for any unattached kinetochore. A single naked kinetochore is enough to generate the mitotic checkpoint complex (Mad2, BubR1, Bub3, Cdc20), which inhibits the anaphase-promoting complex/cyclosome (APC/C) and halts the cell cycle. Only when every kinetochore is attached and under tension is the last checkpoint signal silenced. This surveillance is why chromosome missegregation happens fewer than once per 100,000 human cell divisions.

Anaphase: pulling the chromosomes apart

The trigger is proteolytic, not mechanical. Once the SAC is satisfied, APC/C tags securin and cyclin B with ubiquitin for destruction by the proteasome. Destroying securin frees the protease separase, which cleaves the cohesin rings holding the sister chromatids together. The instant cohesin is cut, the sisters are pulled toward opposite poles. Segregation uses two additive mechanisms:

• Anaphase A — kinetochore microtubules shorten, reeling chromosomes toward the poles. Depolymerization occurs at both ends; the kinetochore stays coupled to the shrinking tube via the depolymerization-tracking Ska complex and the load-bearing Ndc80 fibrils, letting the disassembling lattice do the pulling. Chromosomes move at roughly 1–2 µm per minute.
• Anaphase B — the poles themselves separate. Eg5-type and kinesin-6 (MKLP) motors slide antiparallel interpolar microtubules apart at the midzone, while cortical dynein pulls astral microtubules outward, elongating the whole spindle.

Bruce Nicklas's classic micromanipulation experiments in grasshopper spermatocytes (1983) showed the force needed to move one chromosome is only about 0.1 pN, yet a kinetochore can withstand tens to hundreds of piconewtons — the cell operates with an enormous safety margin.

Why it matters, and common misconceptions

The spindle is where the fidelity of inheritance is physically enforced, so its failures are consequential. Aneuploidy — the wrong chromosome number — is found in over 90% of solid tumors, drives resistance and heterogeneity, and in germ cells causes trisomy 21 (Down syndrome), most first-trimester miscarriages, and age-related infertility (human oocytes are especially error-prone because they sit arrested for decades with weakened cohesion). Because rapidly dividing cells depend so heavily on spindle dynamics, spindle poisons — taxanes, vinca alkaloids — remain among the most-used cancer drugs, and newer agents target Eg5 or the mitotic kinases directly.

Several intuitions are wrong. First, the spindle does not push chromosomes apart like pistons; the dominant force in anaphase A is a pull generated by microtubule depolymerization, which can do work without any motor consuming ATP at that site. Second, motors are not the whole story — a shrinking lattice, coupled through the Ndc80/Ska interface, moves cargo on its own. Third, spindle poles are not required to be centrosomes: acentrosomal spindles in oocytes and plants prove that self-organization by motors, crosslinkers, and Ran-GTP suffices. Finally, the checkpoint counts attachment and tension, not chromosome position — chromosomes can sit off the metaphase plate and still satisfy the SAC if properly bi-oriented.

Microtubule dynamic instability: the two states a spindle fiber switches between

Property	Growing (polymerizing)	Shrinking (depolymerizing)
Plus-end rate (vitro/in cell)	~1–30 µm·min⁻¹	~10–50 µm·min⁻¹
GTP-tubulin cap	present at plus end	lost (catastrophe)
Protofilament shape	straight, sheet-like	curling outward ('ram's horns')
Triggering transition	rescue (resume growth)	catastrophe (abrupt collapse)
Energy source	GTP bound to β-tubulin	stored strain from GTP hydrolysis