Centrosome and Spindle Apparatus

How the centrosome organizes microtubules

Microtubules are stiff hollow tubes built from a single repeating unit: a heterodimer of α-tubulin and β-tubulin. In a normal somatic cell most of these tubes radiate outward from a single, tiny structure near the nucleus — the centrosome. The centrosome itself doesn't move. It anchors the minus ends of microtubules and lets the plus ends explore the cytoplasm by repeatedly polymerizing and depolymerizing.

At the heart of the centrosome are two centrioles — barrel-shaped towers of nine microtubule triplets. They sit perpendicular to each other, one mother and one daughter, like a tiny letter T. Around them is a cloud of pericentriolar material (PCM), a dense matrix of scaffolding proteins (pericentrin, CEP192, CDK5RAP2). The PCM is where the actual microtubule nucleation happens. Embedded throughout the PCM are γ-tubulin ring complexes, or γ-TuRCs — 13-membered rings of γ-tubulin that act like a thirteen-fingered hand templating each new microtubule.

            ── PCM ──            ── PCM ──
       γ-TuRC \  | /  γ-TuRC                         (+) plus end:
               \ | /                                 dimers add/leave;
       ▶ ─── M T ─────────────────────────── ▶       dynamic instability
       ▶ ─── M T ─────────────────────────── ▶
       ▶ ─── M T ─────────────────────────── ▶       (−) minus end:
              ⌽⌽                                     anchored at γ-TuRC
       ┌──── centrioles ────┐
       │  9 × triplet MT    │  ← mother
       │  9 × triplet MT    │  ← daughter (perpendicular)
       └────────────────────┘

An α/β-tubulin dimer carries a GTP on its β subunit. Once added to a plus end, GTP slowly hydrolyzes to GDP. While the tip stays GTP-loaded, the tube grows; when hydrolysis catches up, the tip "catastrophes" and shrinks rapidly. This dynamic instability — discovered by Mitchison and Kirschner in 1984 — is the cell's search engine. Microtubules that bump into something useful (kinetochore, cortical anchor) are stabilized; ones that find nothing collapse.

Centrosome duplication — once and only once

The centrosome cycle runs in lockstep with the cell cycle. In G1, one centrosome. During S phase — same time DNA is being copied — each centriole templates a new daughter at its base, leaving G2 with two centrosomes (two centrioles each). Cyclin E-CDK2 triggers Plk4 kinase, which nucleates SAS-6 oligomers into the cartwheel scaffold of the new centriole.

Like DNA replication, duplication must happen exactly once. Extra centrosomes mean multipolar spindles, which mean chromosome mis-segregation. Cells enforce single duplication by physically separating mother and daughter centrioles ("disengagement") only after M phase — engaged pairs cannot template a new round.

Building the bipolar spindle

At G2/M, two duplicated centrosomes sit beside the nucleus. As cyclin B-CDK1 activates and the nuclear envelope dissolves, the centrosomes are pushed apart by motor proteins — primarily kinesin-5 (Eg5/KIF11), which forms tetramers that walk toward the plus ends of two antiparallel microtubules and slide them apart. This produces the bipolar geometry.

Three populations of microtubules then emerge from each pole:

• Astral microtubules reach out to the cell cortex. They orient and position the spindle by pulling on dynein motors anchored at the membrane.
• Kinetochore microtubules attach to chromosomes. These are the load-bearing fibers that pull sisters to the poles.
• Interpolar microtubules overlap from opposite poles in the spindle midzone. Kinesin-5 sliding here pushes the poles further apart; in anaphase, kinesin-4 and PRC1 organize a stable midzone that sets up cytokinesis.

Chromosomes connect via their kinetochores — multilayer protein platforms (~100 components, with the Ndc80 complex as the load-bearing coupler) assembled on centromeric DNA. The search-and-capture model says microtubule plus ends grow and shrink randomly; when one happens to bump into a kinetochore, it's stabilized. Each chromosome has two sister kinetochores, which must be captured by microtubules from opposite poles. Tension between them is the cell's signal that biorientation is correct. Tensionless attachments are detected by Aurora B kinase, which phosphorylates and detaches them, giving the chromosome another chance.

Spindle architecture across cell types

	Animal somatic	Animal oocyte	Plant cell	S. cerevisiae	S. pombe	E. coli
Centrosomes	Yes (2 centrioles)	None	None	Spindle pole body (no centrioles)	Spindle pole body	None — no spindle
Nuclear envelope	Breaks down (open mitosis)	Breaks down	Breaks down	Stays intact (closed mitosis)	Stays intact	n/a
Spindle nucleation	Centrosomal	Chromatin-driven (Ran-GTP)	Chromatin + nuclear envelope	SPB-embedded	SPB-embedded	n/a
γ-tubulin needed	Yes	Yes	Yes	Yes (Tub4)	Yes	No (uses FtsZ)
Cytokinesis machinery	Actomyosin furrow	Polar body extrusion	Phragmoplast → cell plate	Bud neck septum	Septum	FtsZ Z-ring
Chromosome segregation	Sister chromatids to opposite poles	Bivalents (meiosis I)	Sisters to opposite poles	Sisters to opposite poles	Sisters to opposite poles	OriC tethered to poles by ParAB

Real numbers

• Human metaphase spindle: ~1500 microtubules; 20–40 attach to each kinetochore.
• Plus-end polymerization 5–20 μm/min; catastrophe shrinkage 15–35 μm/min. Individual microtubule lifetime: seconds.
• Tubulin concentration ~20 μM — above the critical concentration for plus-end growth (~10 μM), below it for minus-end growth (~30 μM); minus ends rarely polymerize without a γ-TuRC template.
• Kinesin-5 steps at ~50 nm/sec, generates piconewton-scale sliding forces.
• First chromosome locates the spindle in ~5 minutes; the last can take 10–30 minutes if mis-attached.
• Centrosome amplification (>2 per cell) is detectable in ~80% of high-grade tumors.

Variants and drugs

• Stabilizers: paclitaxel (taxol), docetaxel — bind β-tubulin, block depolymerization. Frozen spindle trips SAC. Mainstay for breast, ovarian, lung cancer.
• Destabilizers: vincristine, vinblastine, colchicine — bind dimers, prevent polymerization. Spindle never forms.
• Kinesin-5 inhibitors: monastrol, ispinesib — block Eg5. Centrosomes don't separate; cell makes a monopolar "monaster" and arrests.
• Aurora kinase inhibitors: alisertib — disrupt kinetochore error correction.
• Centrosome amplification: many cancers overexpress PLK4 and accumulate extra centrosomes; tumors cluster them into pseudo-bipolar spindles to survive. KIFC1 inhibitors block clustering and selectively kill these cells.
• Acentriolar mitosis: human oocytes lack centrosomes; egg spindles self-assemble via Ran-GTP gradient — explaining steep age-related aneuploidy rise.

Common misconceptions

• "Centrosomes are organelles." No membrane, no genome — proteinaceous structures, two per cell, not a population.
• "Centrioles nucleate microtubules." The PCM does, via γ-TuRCs. Centrioles are scaffolds; PCM is the foundry.
• "All dividing cells need centrosomes." Plant cells, fungi, and animal oocytes all lack centrioles. Acentrosomal spindle assembly is widespread.
• "The spindle is static." Microtubules turn over with second-scale half-lives even at metaphase — continuous flux, replacing itself faster than chromosomes move.
• "Centrosome abnormalities are just a cancer consequence." Both directions — amplification can be a driver. Boveri proposed the link in 1914 and was right a century early.
• "γ-tubulin is part of the microtubule wall." α and β are the wall; γ caps the minus end. Related but functionally distinct.