Kinetochore & Spindle Attachment

What the kinetochore is

Every time a cell divides, it has to solve a brutal logistics problem: 46 chromosomes have already been copied into 92 sister chromatids, and exactly one copy of each must end up in each of the two daughter cells. Get it wrong by even one chromosome and the daughter is aneuploid — a state that is usually lethal and, when it isn't, drives cancer and birth defects. The machine that solves this problem is the kinetochore: a disc-shaped protein assembly, only about 200–250 nanometers across, that is built on the centromere of each sister chromatid and acts as the handle by which the mitotic spindle grabs the chromosome.

The kinetochore is not just a passive hook. It does three jobs at once. First, it is a coupler: it attaches to the dynamic plus-ends of spindle microtubules and stays attached even as those microtubules grow and shrink underneath it. Second, it is a motor without a motor: it harvests the energy released when a microtubule depolymerizes and turns it into the force that drags the chromosome toward the pole. Third, and most remarkably, it is a sensor and a switch: it monitors its own attachment state and broadcasts a chemical "wait" signal — the spindle assembly checkpoint — that freezes the entire cell in metaphase until every last chromosome is correctly attached. No engineered machine does grip, force generation, and quality control in one ~100-protein package the size of a virus.

The layered architecture

The kinetochore is built in plates, like a sandwich anchored to DNA on one side and reaching for microtubules on the other.

The foundation — CENP-A and the centromere. The centromere is defined epigenetically, not by sequence. Its hallmark is the histone H3 variant CENP-A, which replaces canonical H3 in a subset of centromeric nucleosomes. CENP-A is the "you are here" mark that is faithfully copied each cell cycle and tells the cell where to build the kinetochore. This is why neocentromeres can form on chromosome arms that were never centromeric — and why the underlying alpha-satellite DNA (0.1–5 megabases of repeats in humans) is necessary scaffolding but not the actual address.

The inner plate — the CCAN. On top of CENP-A sits the constitutive centromere-associated network (CCAN), about 16 proteins (CENP-C, CENP-T, CENP-H/I/K/M, CENP-O/P/Q/U/R, CENP-N/L/W/S/X) that are present at the centromere throughout the cell cycle. CENP-C and CENP-T are the two main bridges that read the CENP-A foundation and recruit the outer plate.

The outer plate — the KMN network. This is the business end that touches microtubules, assembled only during mitosis. KMN stands for its three parts: KNL1 (the signaling scaffold), the Mis12 complex (the connector to the CCAN), and the Ndc80 complex (the load-bearing microtubule grip). The Ndc80 complex — a 57-nanometer dumbbell of Ndc80/Hec1, Nuf2, Spc24, and Spc25 — is the single most important coupling element; each human kinetochore packs roughly 240 of them, and they bind the microtubule lattice through a calponin-homology domain. Accessory couplers (the Ska complex in humans, the Dam1/DASH ring in budding yeast) help the kinetochore track depolymerizing ends.

How attachment and force generation work

A microtubule is a polar tube of α/β-tubulin dimers, growing and shrinking mostly at its plus-end through GTP-driven dynamic instability. The kinetochore captures these plus-ends — typically 20–40 microtubules per kinetochore in a human somatic cell, bundled into a "k-fiber."

The genius is in how it stays attached to a shrinking filament. When a microtubule depolymerizes, its 13 protofilaments lose the straightening force that GTP-tubulin gave them and peel outward into curved "ram's horns." That curl stores mechanical energy. The kinetochore's couplers — Ndc80 arrays plus the Ska complex or the yeast Dam1 ring — form a sleeve or fibril network around the microtubule that cannot fall off the end because the splaying protofilaments push against it. As the end recedes, the coupler is forced to follow, and the chromosome is dragged poleward. This "biased diffusion / forced walk" mechanism delivers up to about 5 piconewtons of force per microtubule, and crucially it spends no ATP on the pull itself — the energy was already paid in when GTP-tubulin was added to grow the filament. Multiplied across a k-fiber, that is tens of piconewtons hauling each chromosome to its pole.

Before anaphase, the chromosome doesn't move much — it oscillates and breathes around the spindle equator while sister kinetochores tug against each other. That tug is the signal the cell is waiting for.

Tension, error correction, and the three wrong ways to attach

Capturing microtubules is a stochastic search-and-capture process, so the kinetochore frequently grabs the wrong thing. There are four possible states for a pair of sister kinetochores:

• Amphitelic (correct, bi-oriented): each sister attaches to one opposite pole. Sisters get pulled apart; one copy goes to each daughter. This generates strong inter-kinetochore tension — sister centromeres stretch to about 1–1.5 µm apart, versus ~0.7 µm at rest.
• Syntelic (wrong): both sisters attach to the same pole. No tension. Both copies would go to one daughter.
• Merotelic (wrong, dangerous): a single kinetochore attaches to both poles at once. It is occupied and partly tense, so it can sneak past the checkpoint, then lag at anaphase and get cut or trapped in a micronucleus. Merotely is the main source of "silent" mis-segregation.
• Monotelic (incomplete): only one sister is attached; the other is bare and still screaming the checkpoint signal.

The cell fixes the wrong states with one beautiful trick: a tension sensor built around Aurora B kinase. Aurora B is the catalytic core of the chromosomal passenger complex (with INCENP, survivin, borealin) and sits at the inner centromere, midway between the two sister kinetochores. It phosphorylates the microtubule-grippers of the outer kinetochore — the Ndc80/Hec1 tail, the Ska complex, KNL1 — and phosphorylation loosens their grip. Under low tension (syntelic, merotelic), the outer kinetochore stays close to Aurora B, stays phosphorylated, and lets go to try again. Under correct amphitelic tension, the outer kinetochore is physically stretched away from inner-centromere Aurora B (the spatial-separation model); the grippers fall out of reach, PP1/PP2A phosphatases dephosphorylate them, and the grip locks in. In one sentence: low tension → keep letting go; high tension → hold on.

The spindle assembly checkpoint step by step

The kinetochore is also the cell's anaphase gatekeeper. The logic chains together like this:

1. An unattached kinetochore recruits MAD1–MAD2 and acts as a catalyst, converting soluble open-MAD2 into closed-MAD2.
2. Closed-MAD2, with BUBR1 and BUB3, assembles the mitotic checkpoint complex (MCC).
3. The MCC binds and inhibits CDC20, the activator of the anaphase-promoting complex/cyclosome (APC/C), an E3 ubiquitin ligase.
4. With APC/C off, two key targets — securin and cyclin B — survive. Securin keeps the protease separase inhibited.
5. Separase therefore cannot cleave cohesin, the ring complex gluing the sisters together. No cleavage → no separation → metaphase arrest.
6. When the last kinetochore attaches and comes under tension, it stops generating closed-MAD2. The MCC is actively disassembled (with help from p31-comet and the AAA-ATPase TRIP13).
7. Freed CDC20 switches APC/C on. APC/C ubiquitinates securin and cyclin B; the proteasome destroys them.
8. Separase, now uninhibited, cuts cohesin. The sister chromatids spring apart. Anaphase.

The sensitivity is what makes this remarkable: in a human cell with 46 chromosomes there are 92 kinetochores, and just one unattached kinetochore produces enough "wait anaphase" signal to hold the whole cell. It is amplification at the level of a single molecular machine gating the fate of the entire genome.

Quantified figures

Quantity	Value	Notes
Kinetochore diameter	~200–250 nm	Human; trilaminar disc in EM
Distinct proteins	~100 species	CCAN ~16, KMN ~10, plus SAC, motors, couplers
Ndc80 complexes per kinetochore	~240 (human)	Each is a 57 nm rod
Microtubules per kinetochore	~20–40 (human); 1 (budding yeast)	k-fiber bundle
Force per microtubule	up to ~5 pN	From depolymerization; no ATP for the pull
Inter-kinetochore stretch	~0.7 µm rest → ~1–1.5 µm under tension	The tension the cell reads
Centromere DNA span	~0.1–5 Mb (human)	Alpha-satellite repeats; defined by CENP-A
Anaphase chromosome speed	~1–2 µm/min	Poleward, in many cell types
Checkpoint threshold	1 unattached kinetochore	Out of 92 in a human cell

Correct vs incorrect attachment, side by side

Property	Amphitelic (correct)	Syntelic / Merotelic (incorrect)
Pole connectivity	Sisters → opposite poles	Syntelic: both → same pole. Merotelic: one kinetochore → both poles
Inter-kinetochore tension	High (~1–1.5 µm stretch)	Low (syntelic) or partial (merotelic)
Aurora B phosphorylation	Lost — grippers locked on	Maintained — grip released to retry
Passes the checkpoint?	Yes	Syntelic: no. Merotelic: often yes (silent error)
Anaphase outcome	One copy per daughter	Both copies to one daughter, or a lagging/cut chromosome
Genomic consequence	Euploid daughters	Aneuploidy, micronuclei, chromosome breakage
Disease link	—	Trisomy, miscarriage, tumor CIN

Where it shows up: organisms, disease, and drugs

• Meiosis in human eggs — the maternal-age effect. Oocytes arrest in prophase of meiosis I for years to decades. Over that time the cohesin holding sisters together slowly decays and is not replenished, so kinetochore-spindle errors rise sharply with maternal age. This is the dominant explanation for why aneuploid pregnancies — including Down syndrome (trisomy 21) — increase with age.
• Aneuploidy and miscarriage. Chromosome mis-segregation is the leading known cause of spontaneous first-trimester pregnancy loss; a large fraction of clinically recognized early miscarriages carry an aneuploidy, most of meiotic origin in the egg.
• Cancer and chromosomal instability (CIN). The majority of solid tumors show CIN — persistent gain and loss of whole chromosomes from chronic kinetochore-attachment errors (often merotely). CIN fuels the karyotypic churn that lets tumors evolve and resist therapy.
• Taxane chemotherapy. Paclitaxel (Taxol) and docetaxel stabilize microtubules and freeze their dynamics, so kinetochores can never satisfy the checkpoint; cells arrest in mitosis and die. These remain front-line drugs for breast, ovarian, and lung cancers.
• Targeted checkpoint drugs. Inhibitors of MPS1/TTK (the master checkpoint kinase) and Aurora kinases are in clinical trials to push CIN cells past a lethal level of aneuploidy.
• Model systems. Budding yeast (Saccharomyces cerevisiae) attaches just one microtubule per kinetochore and uses the Dam1/DASH ring, which made it the system where coupling was worked out; Caenorhabditis elegans has "holocentric" chromosomes whose kinetochores form along the entire chromosome length rather than at one spot.

Common misconceptions

• "The centromere and the kinetochore are the same thing." No. The centromere is the DNA locus (marked by CENP-A); the kinetochore is the protein machine assembled on it only during division. One is the address, the other is the building.
• "A motor protein pulls the chromosome to the pole." Motors (CENP-E, dynein) help capture and transport, but the main poleward pull at anaphase comes from microtubule depolymerization, not from an ATP-burning motor walking along the lattice. The kinetochore harvests the disassembly energy already stored in the lattice.
• "The checkpoint counts attachments." It doesn't tally — it integrates a diffusible inhibitory signal. Each unattached kinetochore emits inhibitor; the cell proceeds only when total production falls to near zero. That's why a single hold-out is enough.
• "Tension and attachment are the same signal." They're related but distinct. Attachment occupancy silences the MAD2-based checkpoint at a kinetochore; tension is what the Aurora B error-correction system reads to decide whether an attachment is correct. A syntelic kinetochore can be fully attached yet under low tension.
• "Merotelic attachments are caught by the checkpoint." Often they aren't — a merotelic kinetochore is occupied and partly tense, so it can satisfy the checkpoint and still cause a lagging chromosome at anaphase. This is precisely why merotely is the sneakiest source of aneuploidy.
• "Pulling chromosomes costs a lot of ATP." The pull itself is essentially free — the energy was spent earlier adding GTP-tubulin to grow the microtubule. The depolymerizing end releases that stored strain to do the work.