The Mitotic Spindle and Its Assembly

The mitotic spindle is a bipolar, football-shaped array of microtubules that forms during mitosis (and meiosis) to segregate chromosomes into two daughter cells. "Bipolar" means the spindle has two poles, usually organized by centrosomes in animal cells, with microtubule minus ends anchored near each pole and plus ends extending toward the equator. The spindle is built from three functionally distinct microtubule populations. Kinetochore microtubules (kMTs) attach directly to the kinetochore, a proteinaceous structure assembled on the centromeric DNA of each sister chromatid; bundles of 15–20 kMTs form the "kinetochore fiber" in vertebrate cells (far fewer, often just one, in yeast). Interpolar (polar) microtubules extend from each pole toward the equator without attaching to kinetochores, overlapping with interpolar microtubules from the opposite pole in an antiparallel array at the spindle midzone. Astral microtubules radiate outward from the poles toward the cell cortex, positioning the spindle and orienting the eventual cleavage plane. Assembly of this structure is best understood not as directed construction but as self-organization: no single component contains a blueprint of the final spindle. Instead, local rules — dynamic instability of individual microtubules, motor-protein-driven sliding and crosslinking, and selective stabilization of correctly attached microtubules — combine to produce a robust, reproducible bipolar structure. This idea, crystallized in the "search-and-capture" model of Marc Kirschner and Tim Mitchison, remains the conceptual backbone of spindle biology, though it has been substantially extended by later discoveries of chromatin-based and microtubule-based (augmin) nucleation pathways that supplement centrosomal nucleation. Central to the mechanism is dynamic instability, the property (established for microtubules generally, a prerequisite covered under microtubule dynamics) that individual microtubules stochastically switch between phases of growth and rapid shrinkage. In the spindle, this instability is not a nuisance but the engine of search: it lets microtubules sample the cytoplasm rapidly, and capture by a kinetochore or by a motor-crosslinker converts a randomly exploring filament into a stably anchored, load-bearing structural element.

Step 1 — Nucleation of microtubules from multiple sources. In most animal cells, the two centrosomes (each with a pair of centrioles surrounded by pericentriolar material) separate and migrate to opposite sides of the nucleus as the cell enters mitosis, driven largely by the plus-end-directed motor kinesin-5 (Eg5) crosslinking and sliding apart antiparallel microtubules nucleated from each centrosome. Each centrosome nucleates microtubules via γ-tubulin ring complexes (γ-TuRCs), which template thirteen-protofilament microtubules with defined polarity: minus end anchored at the γ-TuRC, plus end free to grow and shrink. But centrosomal nucleation is only part of the story. Two other pathways add substantially to spindle microtubule mass:

• Chromatin-mediated nucleation, driven by a gradient of the small GTPase Ran. Chromatin-bound RCC1 (a guanine nucleotide exchange factor) generates a local high concentration of RanGTP around chromosomes. RanGTP releases spindle assembly factors (such as TPX2 and NuMA) from inhibitory importin-α/β complexes specifically near chromatin, locally licensing microtubule nucleation and stabilization where it is most needed — near the kinetochores that must be captured.
• Augmin-dependent amplification, in which the augmin complex binds the sides of existing microtubules and recruits γ-TuRC to nucleate new microtubules at a shallow angle, autocatalytically increasing microtubule density within the spindle body itself, independent of new input from the poles.

Step 2 — Dynamic instability drives search. Newly nucleated microtubules grow and shrink stochastically. A microtubule plus end growing toward a kinetochore has some probability of being captured before catastrophe (spontaneous switch to shrinkage); the kinetochore acts as a "trap." Once end-on attachment occurs, the kinetochore stabilizes the microtubule against depolymerization — an example of end-on capture converting a transient encounter into a persistent, load-bearing connection.

Step 3 — Initial lateral capture and error correction. Kinetochores are frequently captured first by the sides of microtubules (lateral attachment), then converted to end-on attachment via motor-driven transport along the microtubule (dynein-mediated poleward sliding) followed by maturation into a stable, force-generating end-on attachment. Many initial attachments are wrong — a kinetochore may attach to microtubules from both poles (merotelic) or both sister kinetochores may attach to the same pole (syntelic). These errors are corrected by Aurora B kinase, which phosphorylates kinetochore-microtubule interface proteins under conditions of low tension, destabilizing incorrect attachments and giving the kinetochore another chance to be captured correctly — tension itself (generated only by correct, bioriented attachment) is the signal that stabilizes the right geometry.

Step 4 — Motor proteins convert microtubule populations into a coherent bipolar structure. Several conserved motors organize the raw microtubule population:

Spindle bipolarity emerges from a force balance: kinesin-5 pushes poles apart along interpolar microtubules; kinesin-14 and cortical dynein pull poles together or anchor them; the resulting steady-state spindle length reflects the equilibrium of these opposing motor activities, not a fixed structural template. This is dramatically demonstrated by the fact that inhibiting kinesin-5 (e.g., with the drug monastrol) collapses the bipolar spindle into a monopolar aster, because the unopposed inward forces of kinesin-14/dynein win.

Step 5 — Congression and metaphase alignment. Once each sister kinetochore is stably attached end-on to microtubules from opposite poles (biorientation), the interplay of poleward pulling forces (dynein/depolymerization at kinetochore microtubule plus ends) and polar ejection forces (chromokinesins pushing chromosome arms away from poles) positions all chromosomes at the spindle equator, forming the metaphase plate. This alignment sets the stage for the events covered in Chromosome Segregation and the Metaphase-to-Anaphase Transition, and is monitored by The Spindle Assembly Checkpoint, which delays anaphase until every kinetochore is properly attached and under tension.

Mazia and Dan's isolation of the mitotic apparatus (1952). Daniel Mazia and Katsuma Dan isolated intact mitotic spindles from dividing sea urchin eggs by rapidly chilling cells (to stabilize microtubules against the cold-lability that normally causes their disassembly) and then extracting them in a digitonin/detergent solution. This demonstrated for the first time that the spindle is a discrete, physically isolable structure with defined biochemical composition, not merely an optical artifact of the light microscope, and it opened the door to biochemical study of spindle proteins.

Inoué's polarization microscopy (1950s–1960s). Shinya Inoué used polarized light microscopy to visualize the birefringence of the living, unstained mitotic spindle, showing that spindle fibers are ordered arrays of oriented filaments (later identified as microtubules) that could be made to depolymerize reversibly by cold or by the drug colchicine, and to reform on rewarming. This directly demonstrated the dynamic, reversible nature of spindle microtubules — the physical basis for dynamic instability being later shown at the single-filament level by Mitchison and Kirschner (1984) using in vitro assays of pure tubulin.

Kirschner and Mitchison's "search-and-capture" model (1986). Building on the discovery of dynamic instability, Mitchison and Kirschner proposed that spindle microtubules find kinetochores essentially by random search: microtubules nucleated from centrosomes grow and shrink stochastically in all directions, and those that happen to encounter a kinetochore are selectively stabilized (rescued from catastrophe) by the capture event, while unattached ones continue to fluctuate and eventually retract. This explained, quantitatively, how a small number of centrosomal nucleation sites could reliably find widely dispersed kinetochores within the observed timescale of prometaphase (minutes), given measured rates of microtubule growth and catastrophe. Later refinements (chromatin-based nucleation, augmin amplification) showed that pure random search from centrosomes alone is somewhat too slow to explain observed capture times in large cells, motivating the additional chromatin-proximal nucleation pathways described above — an important correction that search-and-capture, while essentially correct, is supplemented rather than exhaustive.

Rieder's laser microsurgery experiments (1990s). Conly Rieder and colleagues used a focused laser microbeam to sever kinetochore fibers or to ablate individual kinetochores in living PtK1 (rat kangaroo) cells during mitosis, then tracked chromosome behavior by video light microscopy. Severing a kinetochore fiber caused the chromosome to lose poleward movement until a new microtubule connection was captured, directly demonstrating that kinetochore fibers, not some diffuse force, are necessary for chromosome-to-pole movement. Ablating one kinetochore of an unattached chromosome delayed anaphase onset in the whole cell, providing direct physical proof — years before the checkpoint's molecular components were known — that unattached kinetochores generate an inhibitory "wait" signal, the physiological basis of The Spindle Assembly Checkpoint.

Heald, Karsenti and colleagues' Xenopus egg extract experiments (1996). Rebecca Heald and coworkers showed that bipolar, spindle-like structures could self-assemble in cell-free Xenopus laevis egg extracts around DNA-coated beads, in the complete absence of centrosomes or kinetochores. This was a landmark demonstration that centrosomes are not strictly required for bipolar spindle formation — chromatin itself, via the RanGTP pathway, can nucleate and organize a bipolar array — proving that spindle bipolarity is a self-organizing property of microtubules and motors acting on any local source of nucleation, rather than something imposed solely by centrosome number or position. This finding also explains how female meiotic spindles form normally in oocytes of many species that lack centrioles entirely.

RNAi and small-molecule motor perturbation (2000s). Systematic depletion or inhibition of individual motors — kinesin-5 by RNAi or by the specific inhibitor monastrol (discovered in a chemical screen by Mitchison's lab), kinesin-14 by RNAi, dynein by dynamitin overexpression — produced predictable, motor-specific spindle phenotypes (monopolar spindles on kinesin-5 loss, over-elongated or splayed spindles on kinesin-14/dynein loss) that quantitatively confirmed the force-balance model of bipolarity described above.

Example 1 — Quantifying search-and-capture timing. Suppose a centrosome nucleates microtubules that grow at a rate of 10 μm/min, undergo catastrophe (switch to shrinkage) with a characteristic mean lifetime of 30 seconds if unattached, and the nearest unattached kinetochore lies 5 μm away. Estimate whether a single microtubule is likely to reach the kinetochore before catastrophe, and reason about why a cell doesn't just rely on one microtubule per kinetochore. Growth rate 10 μm/min = 1 μm per 6 seconds. To travel 5 μm requires 30 seconds of continuous growth. Given a mean catastrophe lifetime of 30 seconds, roughly half of microtubules growing in that direction would be expected to undergo catastrophe before reaching the target (assuming roughly exponential catastrophe statistics, the probability of surviving at least 30 s is about e⁻¹ ≈ 37%). Because each centrosome nucleates dozens to hundreds of microtubules simultaneously, radiating in essentially all directions, only a small fraction need to be aimed anywhere near the correct direction and survive long enough — with ~100 microtubules探索ing per centrosome, even low per-microtubule success probability yields near-certain capture within a few rounds of nucleation-catastrophe-renucleation, occurring on the timescale of tens of seconds to a few minutes. This matches the observed duration of prometaphase capture (a few minutes for all ~92 human kinetochores) and illustrates why redundant, stochastic search massively outperforms a single "aimed" filament — and also why chromatin-proximal nucleation (shortening the effective search distance) meaningfully speeds up capture in large cells, since search time scales steeply with distance.

Example 2 — Diagnosing a spindle defect from phenotype. A researcher treats cultured human cells with a novel small molecule and observes, by immunofluorescence, that cells arrest in mitosis with two clearly separated centrosomes but with all chromosomes clustered near one pole in a rosette, rather than aligned at a metaphase plate; astral microtubules appear normal in number, but no distinct spindle midzone of overlapping antiparallel microtubules is visible, and cold-treatment shows an unusually short distance between the two spindle poles compared to untreated controls. Which motor or process is the most likely target, and why not kinesin-5? Reasoning through the phenotype: if kinesin-5 (Eg5) were inhibited, the classic and well-documented phenotype is a monopolar spindle — a single microtubule aster with unseparated centrosomes collapsed together (as with monastrol), because kinesin-5's outward sliding force is completely lost and the unopposed inward force of kinesin-14/dynein pulls the two centrosomes into one mass. Here, however, the two centrosomes are separated (ruling out kinesin-5 loss), but the spindle is abnormally short and lacks a proper antiparallel midzone, with chromosomes failing to reach the equator. This pattern — bipolar but short, with a poor midzone and chromosome misalignment — is consistent with a defect in the balance of interpolar microtubule crosslinking/sliding rather than nucleation itself: for instance, a molecule that partially inhibits kinesin-5's sliding activity without abolishing pole separation entirely, or one that hyperactivates kinesin-14/dynein-mediated inward force, would produce exactly this intermediate, force-imbalanced phenotype. The correct experimental follow-up is to measure interpolar microtubule overlap length directly and to test whether co-depleting kinesin-14 restores normal spindle length and chromosome alignment — a rescue would confirm that the drug tips the kinesin-5/kinesin-14 force balance rather than blocking nucleation or kinetochore attachment outright.

Marc Kirschner trained as a biochemist and became one of the central figures in reconstructing complex cellular processes using purified components and cell-free extracts, an approach that transformed cell biology from a largely descriptive, microscopy-based discipline into a quantitative, mechanistic one. Working with Timothy Mitchison at UC San Francisco in the early 1980s, Kirschner helped establish that individual microtubules assembled from pure tubulin in vitro alternate stochastically between growth and rapid shrinkage — the phenomenon of dynamic instability — using video-enhanced differential interference contrast microscopy to watch single filaments in real time. This discovery, published in 1984, provided the missing physical mechanism that made search-and-capture a testable, quantitative model rather than a qualitative metaphor. Kirschner later pioneered the use of Xenopus laevis egg extracts as a cell-free system for reconstituting complex processes including spindle assembly, DNA replication, and the cell cycle's biochemical oscillator — extracts that his former student Rebecca Heald used for the DNA-bead spindle-assembly experiments described above.

Cancer chemotherapy. Taxanes (paclitaxel, docetaxel) stabilize microtubules and suppress dynamic instability; vinca alkaloids (vincristine, vinblastine) destabilize microtubules. Both classes arrest rapidly dividing cancer cells in mitosis by preventing proper spindle function, ultimately triggering cell death via prolonged Spindle Assembly Checkpoint activation. Kinesin-5 (Eg5) inhibitors were developed as more selective anti-mitotic agents intended to avoid the peripheral neuropathy caused by taxane/vinca action on non-dividing neuronal microtubules, though clinical efficacy has been more limited than hoped.

Reproductive medicine and aneuploidy screening. Human oocytes assemble meiotic spindles without centrioles, relying heavily on chromatin-based and augmin-based nucleation; this acentrosomal pathway is more error-prone with maternal age, contributing to the well-documented age-related rise in aneuploid eggs and conditions such as trisomy 21. Understanding spindle assembly fidelity informs preimplantation genetic testing and assisted reproduction protocols.

Cancer genomics and chromosomal instability. Tumors frequently show merotelic kinetochore attachments and lagging chromosomes at anaphase, driven by defects in kinesin-13/MCAK-mediated error correction or centrosome amplification (extra centrosomes nucleating multipolar spindles that are subsequently clustered into a pseudo-bipolar array); this chromosomal instability generates the karyotypic diversity that fuels tumor evolution and drug resistance.

Agricultural and veterinary biotechnology. Colchicine, a classic spindle-poisoning alkaloid, is used deliberately to induce polyploidy in crop breeding (e.g., doubling chromosome number in watermelon or wheat lines) by blocking spindle function during a mitotic division, causing chromosome doubling without cell division.

Ask the tutor: How does the spindle assembly checkpoint molecularly sense the difference between an unattached kinetochore and one under tension?