Cell Biology · Unit 4: Division and cell loss

Mitosis and Chromosome Segregation

How does a dividing cell give each daughter exactly one copy of every chromosome? This lesson builds the spindle from several microtubule sources, shows how kinetochores are attached, tested and corrected, and follows the one-way protease switch that releases sister chromatids. Afterwards you can measure anaphase A and B from a time-lapse movie and keep lagging chromosomes and micronuclei apart by their denominators.

Watch first The Cut That Cannot Be Undone · 2:53

In the dividing human HeLa and RPE1 cells measured in 2019, most spindle microtubules were not made at the centrosomes. This lesson builds the spindle, attaches and tests every chromosome, cuts the glue that holds sisters together, and then measures what happens from a movie.

You arrive at mitosis with each chromosome duplicated into two sister chromatids held together by cohesin, a protein ring that encircles DNA. CDK1-cyclin B has driven the cell into mitosis (see cell-cycle control) and the nuclear envelope has broken down. The task now is to give each daughter one sister of every chromosome. How chromosomes pair and separate in meiosis, and what happens to an organism when the count goes wrong, belong to chromosomes and meiosis and chromosome changes.

Building a bipolar spindle from several sources

The spindle is a bipolar array of microtubules, the dynamic polymers described in cytoskeletal polymers and dynamics. Their minus ends gather at the two poles and many plus ends point toward the chromosomes. A kinetochore is the protein machine built on each sister’s centromere that couples the chromosome to growing and shrinking microtubule tips. A k-fibre is the bundle of microtubules that connects one kinetochore to a pole.

The classic model, search-and-capture, has centrosomes fire microtubules in random directions until a plus end happens to hit a kinetochore and is stabilized. Like casting fishing lines at random, it works, but alone it would be slow. Other sources add microtubules:

  • Near chromatin and kinetochores. In human RPE1 cells, Sikirzhytski and colleagues used correlative light and electron microscopy and found short non-centrosomal microtubules near kinetochores from the start of spindle assembly. End-on attachments formed within the first three minutes of prometaphase. Inhibiting CENP-E, a kinetochore kinesin (see molecular motors), impeded the conversion of side-on contacts into end-on attachments.
  • Branching on existing microtubules. Augmin is a protein complex that recruits the gamma-tubulin ring complex (gamma-TuRC), the template that starts new microtubules, onto the side of an existing one, so a new microtubule branches off in the same direction. In HeLa and RPE1 cells, David and colleagues tracked growing ends and found that most spindle microtubules nucleated away from centrosomes; augmin-dependent amplification almost entirely explained where the growing ends were. In vitro, purified human augmin and gamma-TuRC were sufficient for branching, with NEDD1 bridging the two and phosphorylation by CDK1 and PLK1 needed for NEDD1 to bind augmin.

Keep the scope: these numbers come from two human cell lines and a test tube. Some cells, such as many oocytes, build spindles without centrosomes at all, and the share from each route differs between cell types.

Attaching and correcting: kinetochores, Aurora B and tension

The right outcome is biorientation: sister kinetochores attached to opposite poles. Three wrong outcomes are common in early mitosis. In a monotelic attachment one sister is attached and the other is free. In a syntelic attachment both sisters face the same pole. In a merotelic attachment a single kinetochore is bound by microtubules from both poles.

Biorientation, unlike syntelic or monotelic attachment, puts the sisters under tension, because the two poles pull against the cohesin that holds the sisters together. Two mechanisms use that difference.

Aurora B destabilizes. Aurora B is a kinase that sits at the inner centromere, between the sister kinetochores, as part of the chromosomal passenger complex. It phosphorylates outer kinetochore proteins, including the microtubule-binding Ndc80 complex, which lowers their grip on microtubules. Lampson and colleagues blocked Aurora kinase activity in living vertebrate cells with a reversible inhibitor, let wrong attachments pile up, then washed the inhibitor out. Reactivated Aurora kinase corrected the errors by selectively disassembling kinetochore-microtubule fibres. How tension turns Aurora B’s effect off is still argued. One popular model says stretching pulls the outer kinetochore out of reach of the kinase at the inner centromere; other evidence points to changes in kinetochore structure under force.

Tension stabilizes. Akiyoshi and colleagues removed the cell entirely. Purified budding-yeast kinetochore particles, held on microtubules in a laser trap, stayed attached longer at higher force: a mean of 21 plus or minus 5 minutes at 1 piconewton and 50 plus or minus 17 minutes at 5 piconewtons, with no detectable Aurora B and no ATP.

Keep the evidence separate: the in vitro result shows that force can stabilize attachment directly. It does not show that Aurora B is unnecessary in cells, where both mechanisms act together.

Correct attachment also satisfies the spindle-assembly checkpoint. An unattached kinetochore produces a signal that restrains APC/C-Cdc20 until every kinetochore is attached; the wiring of that signal belongs to checkpoints.

The one-way switch: securin, separase and cohesin

The APC/C, or anaphase-promoting complex/cyclosome, is an E3 ubiquitin ligase: it tags target proteins with chains of ubiquitin that send them to the proteasome (the chemistry is in ubiquitin and proteolysis). When the checkpoint releases it, APC/C bound to its coactivator Cdc20 tags two key targets.

  1. Securin. Securin binds separase and keeps it inactive. Once securin is destroyed, separase is free.
  2. Cyclin B. Its destruction inactivates CDK1, which lets the cell leave mitosis.

Separase is a cysteine protease. It cuts the kleisin subunit of cohesin, Scc1 (called Rad21 in vertebrates), and the ring opens. Note the division of labour: the APC/C and the proteasome destroy securin; separase cuts cohesin. Separase does not degrade securin.

Uhlmann and colleagues tested whether cutting Scc1 is enough. In budding yeast arrested at metaphase, where separase is held inactive, causing Scc1 to be cleaved was sufficient to separate and segregate the sisters. A cut protein cannot be rejoined, so anaphase cannot be reversed.

Vertebrates do not wait for separase to remove all their cohesin. During prophase and prometaphase the bulk of cohesin on chromosome arms leaves without Scc1 cleavage. Hauf and colleagues showed in human cultured cells, including HeLa, that phosphorylation of the cohesin subunit SA2 was essential for this prophase pathway. Centromeric cohesin is protected by shugoshin. When McGuinness and colleagues depleted shugoshin from HeLa cells, centromeric cohesin was lost early, and their results suggested that shugoshin works by preventing SA2 phosphorylation at centromeres. So in a vertebrate metaphase chromosome the sisters are held mainly at the centromere, and that is the cohesin separase cuts. In budding yeast, by contrast, most cohesin is cleaved at anaphase.

Anaphase A and B, read from a movie

Anaphase has two movements. Anaphase A is the movement of chromosomes toward their poles. Anaphase B is the separation of the poles themselves. In anaphase A, the k-fibres shorten. Gorbsky, Sammak and Borisy bleached a mark on kinetochore microtubules in cultured mammalian cells and found that the chromosomes moved toward the mark: the fibres lost subunits mainly at the kinetochore, though a small loss elsewhere was not excluded. In anaphase B, in many animal cells, motors slide apart the overlapping microtubules in the spindle midzone and dynein at the cell cortex pulls on astral microtubules.

Worked problem: two rates from one movie

The numbers here are illustrative. You film a cultured mammalian cell with labelled poles and kinetochores. At anaphase onset (t = 0) the poles are 12.0 um apart and each sister kinetochore is 5.5 um from its pole. At t = 4.0 min each kinetochore is 1.5 um from its pole and the poles are 16.0 um apart. Assume the spindle is straight and movement is steady.

Step 1, anaphase A. Kinetochore-to-pole distance fell from 5.5 to 1.5 um in 4.0 min.

anaphase A rate = (5.5 - 1.5) / 4.0 = 1.0 um/min

Step 2, anaphase B. Pole-to-pole distance rose from 12.0 to 16.0 um.

anaphase B rate = (16.0 - 12.0) / 4.0 = 1.0 um/min of pole separation

Each pole moved 0.5 um/min relative to the spindle centre.

Step 3, sister separation. The distance between sisters is the pole distance minus both kinetochore-to-pole distances. At t = 0 it is 12.0 - 2(5.5) = 1.0 um. At t = 4.0 min it is 16.0 - 2(1.5) = 13.0 um. The rate is 12.0 / 4.0 = 3.0 um/min.

Check. Each sister moves toward its pole and the poles move apart, so sister separation should equal twice anaphase A plus anaphase B: 2(1.0) + 1.0 = 3.0 um/min. It does.

A movie that tracks only chromosomes cannot tell you how much of their separation came from the poles.

Anaphase A and B: read the rates from a movie⚙ original · interactive

Anaphase A (kinetochore to pole)
Anaphase B (pole to pole)
Sister separation
Check: 2 × A + B

Sister separation is the pole distance minus both kinetochore-to-pole distances, so its rate is twice anaphase A plus anaphase B. A movie that tracks only chromosomes cannot say how much of that came from the poles. The picture assumes a straight spindle and steady movement, as the lesson does. All numbers here, the lesson’s included, are illustrative, not measurements, and frames between 0 and 4 min are interpolated by the steady-movement assumption.

When segregation fails: merotely, laggards and micronuclei

A merotelic kinetochore is attached, so it produces no checkpoint signal. Cimini and colleagues showed in PtK1 cells, a rat kangaroo kidney line, that merotelic attachment escaped the checkpoint. Lagging single chromatids appeared in 1.16% of untreated anaphases and 17.55% after recovery from a nocodazole block, and electron microscopy confirmed microtubules from both poles on the lagging kinetochores.

A lagging chromosome is a trajectory, not an outcome. It can reach the main nucleus late or end up in a micronucleus, a small separate nucleus. In human RPE1 and U2OS cells, Orr and colleagues found that most transient laggards rejoined the main nucleus; a midzone Aurora B mechanism helped correct errors in anaphase and delayed nuclear-envelope assembly on laggards. How persistent segregation errors feed chromosomal instability in tumours belongs to cancer genetics; how spindle orientation sets cell fate is in asymmetric division, and how mitotic defects in neural progenitors shrink the brain is in microcephaly and malformations. The final split of the cell is cytokinesis.

Worked problem, continued: watch the denominators. Orr and colleagues saw at least one transient laggard in 9% of RPE1 cells. That number is per cell. They found that 6% of lagging chromosomes formed micronuclei. That number is per chromosome. If every affected cell had exactly one laggard, the fraction of all cells with a laggard-derived micronucleus would be 0.09 x 0.06 = 0.0054, about 0.5%. If some cells had two or more laggards, the true figure is higher. Say the assumption out loud whenever you multiply rates with different denominators.

Design problem

Is anaphase held back by a signal from unattached kinetochores, or by lack of tension? Rieder and colleagues destroyed the last unattached kinetochore in PtK1 cells, and anaphase followed in a reported 20 plus or minus 2.5 minutes, without the damaged chromosome moving to the spindle middle. Li and Nicklas pulled on a mono-oriented chromosome with a microneedle in praying-mantis spermatocytes, and the delay ended. Do these results contradict each other, and what experiment would separate the two ideas?

One way to do it

They need not contradict each other. Tension stabilizes attachment, so in a cell, pulling on a kinetochore also tends to make it fully occupied, and a fully occupied kinetochore stops signalling. Occupancy and tension are coupled.

To separate them, change one while holding the other fixed. In vitro, reconstituted yeast kinetochores in a laser trap (Akiyoshi 2010) let you set force while the number of attached microtubules is known. In cells, you could ablate the last unattached kinetochore and compare it with a cell in which that kinetochore is held attached but under low tension, for example after a drug that dampens spindle forces.

Readout: time from last attachment, or from the lesion, to anaphase onset in single cells. Control: irradiate a spot of the same size near the kinetochore but off it, which should not release the delay. Limit: mantis spermatocytes in meiosis and PtK1 cells in mitosis are different systems, so a difference between the two papers may reflect the cell, not the mechanism.

How we measure it

Live-cell time-lapse microscopy

Label chromosomes, kinetochores or microtubules with fluorescent proteins and image the same cell every few seconds or minutes through division. It gives positions over time, so you can measure rates and follow the fate of a single chromosome; Orr and colleagues sampled human cells every 30 seconds.

Laser microsurgery

Focus a laser to destroy a chosen structure, such as one kinetochore, in a living cell and watch what happens next. It tests whether a structure is needed for an event, provided a lesion placed nearby but off the target is used as a control.

Laser-trap force clamp

Hold a bead attached to a purified kinetochore particle in a focused laser beam and keep the force on it constant while it tracks a dynamic microtubule end. It measures how long an attachment survives at each force, with no ATP present, so the particles cannot be phosphorylated.

Correlative light and electron microscopy

Film a cell live, fix it at a chosen moment, then section the same cell for electron microscopy. The light movie gives the stage and the history; the electron micrographs show which microtubules end at which kinetochore.

Finish line

Test yourself

8 questions

Practice and MCAT-style, with explained answers.

Next · Lesson 20

Cytokinesis and Organelle Inheritance

Push the spindle off-centre in a sea urchin blastomere, and the signal for the furrow moves with it before the furrow does.

Continue →

← Lesson 18 · Checkpoints and Irreversible Transitions

Ndc80 complex, engineered short version — the kinetochore's main microtubule grip
Ndc80 complex, engineered short version — the kinetochore's main microtubule grip ● LIVE · RCSB PDB 2VE7
open on RCSB PDB ↗
Loading live structure 2VE7 from RCSB PDB…

Look for the two globular heads at one end of a long coiled coil. Those heads contact the microtubule lattice; the disordered tail of Ndc80 that Aurora B phosphorylates is not part of this engineered construct, which spans Ndc80 residues 80 to 286.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Aurora B kinase bound to a fragment of INCENP — the error-correction kinase
Aurora B kinase bound to a fragment of INCENP — the error-correction kinase ● LIVE · RCSB PDB 2BFY
open on RCSB PDB ↗
Loading live structure 2BFY from RCSB PDB…

The INCENP fragment wraps around the small lobe of the kinase; that contact helps switch Aurora B on. Hesperadin sits in the ATP pocket between the two lobes.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Cohesin Smc1 head bound to the C-terminus of Scc1 — the ring's closure
Cohesin Smc1 head bound to the C-terminus of Scc1 — the ring's closure ● LIVE · RCSB PDB 1W1W
open on RCSB PDB ↗
Loading live structure 1W1W from RCSB PDB…

The end of the kleisin Scc1 binds the ATPase head of Smc1. Separase cuts Scc1 elsewhere in its length, which opens the ring.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Human anaphase-promoting complex — the E3 ubiquitin ligase that triggers anaphase
Human anaphase-promoting complex — the E3 ubiquitin ligase that triggers anaphase ● LIVE · RCSB PDB 4UI9
open on RCSB PDB ↗
Loading live structure 4UI9 from RCSB PDB…

This is the Cdh1-bound form, which acts later in mitosis and in G1. Cdc20, the coactivator that starts anaphase, binds the same region of the complex.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Nocodazole — a drug that depolymerizes microtubules
Nocodazole — a drug that depolymerizes microtubules ● LIVE · PubChem CID 4122
open on PubChem ↗
Loading live structure 4122 from PubChem…

Used by Cimini and colleagues to block mitosis and then release it; recovery from the block raised the frequency of merotelic attachments and lagging chromatids in PtK1 cells.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Conly L. Rieder

Conly L. Rieder

With R. W. Cole, Alexey Khodjakov and G. Sluder, used a laser to destroy the last unattached kinetochore in PtK1 cells and showed that anaphase then followed, locating the wait-anaphase signal at or near unattached kinetochores.

Rieder's tool was the PtK1 cell, a flat line from rat kangaroo kidney with few, large chromosomes that stay in view through division. A cell with one chromosome attached to only one pole waits before anaphase, and the question in the 1995 paper was what the cell is waiting for. Rieder, Cole, Khodjakov and Sluder combined video microscopy with a laser that could destroy a chosen part of a chromosome. When they destroyed the final unattached kinetochore, the cell entered anaphase about 20 minutes later, and the damaged chromosome never moved to the middle of the spindle. Destroying instead an attached kinetochore on another chromosome, while a different chromosome stayed unattached, did not release the delay. The experiment separated two things that normally go together, alignment and the signal to proceed, and showed that the signal comes from molecules in or near an unattached kinetochore. It did not name those molecules, and it did not rule out a role for tension. It made the unattached kinetochore the place where later work on the spindle-assembly checkpoint looked first.

✓ verbatim Rieder, Cole, Khodjakov & Sluder, J. Cell Biol. 130:941 (1995), doi:10.1083/jcb.130.4.941.
Frank Uhlmann

Frank Uhlmann

With Kim Nasmyth and colleagues, showed in budding yeast that separase (then called separin) is a cysteine protease that cuts the cohesin subunit Scc1, and that cutting Scc1 in cells held at metaphase is enough to separate and segregate sister chromatids.

By the late 1990s the cohesin ring was known to hold sister chromatids together, and its Scc1 subunit was known to be cut at anaphase in budding yeast. Cleavage was known to be needed, because a version of Scc1 that could not be cut blocked sister separation, but not whether it was enough. In the 2000 paper, Uhlmann, Dominik Wernic, Marc-André Poupart, Eugene Koonin and Kim Nasmyth asked two sharper questions. First, what kind of enzyme is separin? Sequence comparison placed it in the CD clan of cysteine proteases, and the purified protein cut Scc1 in vitro. Second, is cutting Scc1 enough? They arranged for Scc1 to be cleaved in yeast cells arrested at metaphase, when separase is normally held inactive, and the sisters separated and moved apart. That is a sufficiency test, which is stronger than watching a protein vanish. It does not say that cutting cohesin alone completes the exit from mitosis. The result made it reasonable to treat anaphase as a proteolytic event that the cell cannot take back.

✓ verbatim Uhlmann, Wernic, Poupart, Koonin & Nasmyth, Cell 103:375 (2000), doi:10.1016/s0092-8674(00)00130-6.
Daniela Cimini

Daniela Cimini

With colleagues, showed in PtK1 cells that a merotelic kinetochore, bound to microtubules from both poles, does not stop the cell from entering anaphase, and that merotely is a main source of lagging chromosomes.

A lagging chromosome is one left behind in the middle of the spindle while the others move to the poles. Cimini and her coauthors, Bonnie Howell, Paul Maddox, Alexey Khodjakov, Francesca Degrassi and E. D. Salmon, asked why laggards exist in a cell that has a checkpoint. They counted lagging single chromatids in 1.16% of untreated PtK1 anaphases. After a three-hour block in 2 micromolar nocodazole, a drug that depolymerizes microtubules, followed by an hour of recovery, the figure was 17.55%. Confocal imaging of fixed cells and electron microscopy showed why: the lagging kinetochores were bound by microtubule bundles running to both poles, the arrangement called merotelic. Such kinetochores are attached, so the checkpoint was satisfied and anaphase went ahead. The paper changed what a satisfied checkpoint means: it reports that kinetochores are attached, not that every attachment is correct. It set up the later search for how cells correct merotely, before and during anaphase.

✓ verbatim Cimini, Howell, Maddox, Khodjakov, Degrassi & Salmon, J. Cell Biol. 153:517 (2001), doi:10.1083/jcb.153.3.517.
R. Bruce Nicklas

R. Bruce Nicklas

With Xiaotong Li, pulled on a chromosome attached to only one pole with a glass microneedle in praying-mantis spermatocytes and found that the tension released the cell's delay before anaphase.

Nicklas spent much of his career touching chromosomes. Using fine glass microneedles in living insect spermatocytes, his laboratory could move a chromosome, detach it, or pull on it while watching under the microscope. In meiosis in praying-mantis spermatocytes, a chromosome attached to only one pole delays anaphase. In the 1995 paper, Xiaotong Li and Nicklas used a microneedle to pull on that mono-oriented chromosome, putting its attachment under tension as a correctly attached chromosome would be. The delay ended and the cell went into anaphase. The same year, Rieder's laser experiment in PtK1 cells pointed to unattached kinetochores as the source of the signal. The two papers are often set against each other, but they used different cells and a different kind of division. Together they showed that tension and attachment are tightly coupled, and that separating them needs an experiment that changes one while holding the other fixed.

✓ verbatim Li & Nicklas, Nature 373:630 (1995), doi:10.1038/373630a0.
Test yourself
Question 1 of 8 Score 0 / 0

Which arrangement describes a merotelic attachment?
Sources & Integrity

Every structure, slide, simulation, and clip on this page is listed below with the source it came from. None of the 3D structures is drawn by hand or by AI: each one loads by accession ID from RCSB PDB or PubChem. This panel is generated from the lesson's own data file, so what it lists is what the page loads.

Molecular structures live data

  • Ndc80 complex, engineered short version — the kinetochore's main microtubule grip — RCSB PDB 2VE7 · coordinates CC0 · Human Ndc80 'bonsai' complex, X-ray crystal structure (Ciferri et al., Cell, 2008)
  • Aurora B kinase bound to a fragment of INCENP — the error-correction kinase — RCSB PDB 2BFY · coordinates CC0 · Xenopus laevis Aurora B with an INCENP fragment and the inhibitor hesperadin, X-ray crystal structure (Sessa et al., Mol. Cell, 2005)
  • Cohesin Smc1 head bound to the C-terminus of Scc1 — the ring's closure — RCSB PDB 1W1W · coordinates CC0 · Budding yeast Smc1 head domain with the C-terminal domain of Scc1, ATP-gamma-S bound, X-ray crystal structure (Haering et al., Mol. Cell, 2004)
  • Human anaphase-promoting complex — the E3 ubiquitin ligase that triggers anaphase — RCSB PDB 4UI9 · coordinates CC0 · Human APC/C with the coactivator Cdh1 and the inhibitor Emi1, cryo-EM (Chang et al., Nature, 2015)
  • Nocodazole — a drug that depolymerizes microtubules — PubChem CID 4122 · PubChem compound record

Simulations

  • Anaphase A and B from one movie · Original interactive — Biochemistrypedia (Deppmann Lab).

Scientist profiles & quotes

  • Conly L. Rieder — Rieder, Cole, Khodjakov & Sluder, J. Cell Biol. 130:941 (1995), doi:10.1083/jcb.130.4.941.
  • Frank Uhlmann — Uhlmann, Wernic, Poupart, Koonin & Nasmyth, Cell 103:375 (2000), doi:10.1016/s0092-8674(00)00130-6.
  • Daniela Cimini — Cimini, Howell, Maddox, Khodjakov, Degrassi & Salmon, J. Cell Biol. 153:517 (2001), doi:10.1083/jcb.153.3.517.
  • R. Bruce Nicklas — Li & Nicklas, Nature 373:630 (1995), doi:10.1038/373630a0.

Quotes are verbatim from the cited source. Profile text is original to this site and was checked against the sources cited.

Animated explainer

The Cut That Cannot Be Undone · narration: an AI narrator voice (Gemini, Charon) reading narration written from this lesson's text · animation: AI-generated in code (Manim) by Claude, a schematic with no drawn molecular structures, plots computed from their equations · captions from the narration script.

Images & figures

This lesson carries no raster figures. Every structure is rendered live from RCSB PDB or PubChem, and every plot is drawn from its equation in the browser.

integrity Machines check the mechanics. The build confirms that every local file exists, that every question's answer key points at a real choice, and that any AI-generated scientific figure names the person who checked it; a weekly job re-resolves every accession ID against its database. Questions are original; MCAT-style items are written for this site and are not real AAMC questions. Banned sources (the Tymoczko textbook, publisher decks, copyrighted figures) are not used.

Like I'm five

Before a cell splits in two, it has two copies of every chromosome, stuck together in pairs. Thin ropes grab each pair from both sides. The cell checks that every pair is held from both sides, and fixes any that are held wrong. Then it cuts the glue all at once, and the ropes pull one copy to each side.
High school

In mitosis a cell divides its copied chromosomes between two new nuclei. A spindle made of microtubules forms, and each chromosome's two sister copies attach to fibres from opposite ends. Protein rings called cohesin hold the sisters together until every chromosome is attached correctly. Then an enzyme called separase cuts the rings, and the sisters move to opposite poles while the poles themselves move apart.
College

Spindle microtubules are nucleated at centrosomes, near chromatin and kinetochores, and by augmin-dependent branching from existing microtubules. Kinetochores capture microtubule plus ends; biorientation puts sister kinetochores under tension. Aurora B destabilizes attachments that lack tension, such as syntelic ones, and tension stabilizes correct ones. Once no kinetochore is unattached, the checkpoint releases APC/C-Cdc20, which ubiquitylates securin and cyclin B. Freed separase cuts the cohesin kleisin Scc1 (Rad21). In vertebrates most arm cohesin has already left by the prophase pathway, while shugoshin protects centromeric cohesin. Chromosomes then move poleward (anaphase A) and poles separate (anaphase B).
Graduate

In HeLa and RPE1 cells most spindle microtubules arise away from centrosomes, and augmin amplification of long-lived lattices almost entirely explains the distribution of growing ends; human augmin plus gamma-TuRC is sufficient for branching in vitro, with NEDD1 as the bridge and CDK1/PLK1 phosphorylation needed for binding. Error correction rests on two lines of evidence that should not be merged: reactivating Aurora kinase in living vertebrate cells selectively disassembled wrong k-fibres (Lampson 2004), and purified yeast kinetochore particles lived longer at 5 pN than at 1 pN (50 versus 21 min) with no detectable Aurora B and no ATP (Akiyoshi 2010).

Cohesion release is made irreversible by proteolysis: separase cleavage of Scc1 is sufficient for segregation in metaphase-arrested budding yeast. Merotely escapes the checkpoint (PtK1), and in RPE1 and U2OS cells a midzone Aurora B mechanism corrects many errors during anaphase, so only a minority of transient laggards become micronuclei. Open question: how the kinetochore converts force into a change in Aurora B phosphorylation in cells, whether by spatial separation from the kinase, by tension-dependent conformational change, or both.
Twenty-one minutes or fifty

To ask whether tension stabilizes attachment on its own, Akiyoshi and colleagues removed the cell. They purified kinetochore particles from budding yeast, bound them to beads, attached them to growing and shrinking microtubules, and used a servo-controlled laser trap to hold the force constant. At 1 piconewton the mean attachment lasted 21 plus or minus 5 minutes; at 5 piconewtons, 50 plus or minus 17 minutes. No Aurora B was needed for the effect. The result shows a direct mechanical effect; it does not show that Aurora B is dispensable in a living cell.

— Akiyoshi et al., Nature 468:576 (2010), doi:10.1038/nature09594.
A bleached stripe that stayed put

Where do kinetochore microtubules lose subunits as a chromosome moves to its pole? Gorbsky, Sammak and Borisy injected fluorescent tubulin into cultured mammalian cells and bleached a stripe across the anaphase spindle. If the fibres were reeled in at the pole, the stripe would travel poleward with the chromosomes. Instead, the chromosomes moved toward the stripe, and the stretch of fibre between kinetochore and stripe shrank. They concluded that the fibres shortened mainly by losing subunits at the kinetochore, and stated that a small loss elsewhere was not excluded.

— Gorbsky, Sammak & Borisy, J. Cell Biol. 106:1185 (1988), doi:10.1083/jcb.106.4.1185.
Most laggards go home

Orr and colleagues imaged human RPE1 and U2OS cells in four dimensions every 30 seconds and followed each lagging chromosome to the end of division. Transient laggards were common, seen in 9% of RPE1 cells and 44% of U2OS cells, yet only 6% and 9% of laggards ended up in micronuclei. Most rejoined the main nucleus. A midzone-based Aurora B mechanism helped correct errors during anaphase and delayed nuclear-envelope assembly on the laggards. A single frame showing a laggard does not tell you the daughter's final chromosome count.

— Orr et al., Cell Reports 37:109783 (2021), doi:10.1016/j.celrep.2021.109783.

 ↑ Top