In human cells, DNA is arranged into structures called chromosomes. Before a cell divides it copies its entire set of chromosomes to make paired chomosomes known as sister chromatids. Then, the sister chromatids are separated to ensure that each new daughter cell contains a full set of chromosomes. A structure called the spindle is responsible for separating the sister chromatids. It is made of long filaments called microtubules that grow out from 'poles' at opposite sides of a cell. The two sister chromatids in each pair attach to microtubules that originate from opposite ends of the cell. This attachment is achieved by a protein machine called the kinetochore, which can move along microtubules as they grow or shrink.

Prior to the separation of the chromatids, the paired sister chromatids are moved into positions so that they are approximately an equal distance from the two poles. For the majority of the time, one sister kinetochore moves towards the pole it is attached to (called the lead sister), while the other sister moves away from the pole it is attached to (the trailing sister). Then, the kinetochores swap roles and move in the opposite direction. In most cells, the pairs of sister kinetochores repeatedly switch between moving backwards and forwards giving rise to oscillations in the positions of the sister chromatids.

It is not clear how the two sister kinetochores are able to communicate with each other so that they can co-ordinate their backwards and forwards movement. It is currently thought that the lead kinetochore changes direction first because of an increase in the distance between the two sisters, thereby increasing the tension between the sisters (like a spring being stretched).

Burroughs, Harry and McAinsh revisit this idea by looking at living human cells and tracking the movement of the kinetochores in great detail. Using computational techniques to analyse kinetochore movements in living cells, the experiments reveal that the trailing sister kinetochore can sometimes change direction before the lead sister. When this happens the sisters start to move apart until the previously leading sister switches direction, so that the sisters then movingve together in the same direction. The distance between the sisters remains high until the next time the sisters change direction, which means that 'tension' cannot solely be responsible for the communication between sister kinetochores.

Burroughs, Harry and McAinsh’s findings suggest instead that sister kinetochores contain a 'clock' that decides when they will change direction. The tension between the two sister chromatids is still important, and acts to change the time of the clocks. The next challenge is to understand how these clocks work and which parts of the kinetochore are involved.

The accurate segregation of chromosomes during anaphase requires that all sister kinetochores bi-orientate, an attachment state in which sisters form stable attachments to the plus-ends of microtubules that originate at opposite spindle poles. Bi-orientation begins immediately after nuclear envelope breakdown during prometaphase when scattered chromosomes engage the nascent mitotic spindle, and concludes with the formation of the metaphase plate – a state where all sister kinetochores are bi-orientated and aligned on the equator of a bipolar spindle (McIntosh et al., 2012). To achieve this bi-orientation, sister kinetochores must be able to undergo directed movements to the equator (this is termed congression) and then maintain their position prior to anaphase onset – a feature of this latter phase is oscillations of the chromosomes along the spindle axis. Directed motility is possible because one sister adopts a poleward (P) moving state (the lead sister) while the other is in an away-from-the-pole (AP) moving state (the trailing sister). These two movement states reflect the balance of microtubule polymerisation/depolymerisation within the kinetochore-fibre (K-fibre), which is typically 20–25 microtubules in human cells (Compton, 2000; Rieder, 2005; Wendell et al., 1993). While such K-fibres are rarely coherent, there is a small polymerisation bias between sister kinetochores (Armond et al., 2015; VandenBeldt et al., 2006) and they can be thought of as being in either a net polymerising or depolymerising state. The adaptive switching between these AP and P states then defines the directionality of chromosome motion and can give rise to the quasi-periodic oscillations that are observed in the majority of vertebrate cells (Skibbens et al., 1993). An outstanding question is to understand the mechanisms by which the two sister kinetochores are able to communicate in order to coordinate their P/AP states and thereby generate chromosome movements.

Initial investigations into the control of chromosome movement utilised time-lapse imaging in newt lung cells using video-enhanced differential interference contrast microscopy (Skibbens et al., 1993). Kinetochores were shown to undergo periods of relatively constant velocity separated by abrupt changes in direction – a behaviour termed ‘directional instability’. Subsequent experiments demonstrated that weakening the centromeric chromatin which links the sisters (with a laser) uncoupled the normally coordinated motility of sister kinetochores (Skibbens et al., 1995). These experiments led to a model (also see Rieder and Salmon, 1994) in which tension in the centromeric chromatin triggers a lead sister switch (P-to-AP) at a certain threshold, the loss of tension then triggering a directional switch in the second sister (AP-to-P). More recent kinetochore-tracking experiments in PtK1 cells are consistent with this model and show that switching initiates at maximum inter-kinetochore stretch (Wan et al., 2012), schematic shown in Figure 1. The polar ejection force, which increases with proximity to the pole, pushes the chromosomes towards the metaphase plate. When chromosomes stray far from the equator, this anti-poleward force increases the load on the leading (P) kinetochore and promotes switching – an idea supported by experiments in newt and human cells (Ke et al., 2009; Stumpff et al., 2012). The standard tension model thus predicts a fixed sequence of sister kinetochore-switching events during a directional reversal – lead switch first, followed by trail. However, the timeframe for these events is short (several seconds), requiring a sampling rate that avoids temporal aliasing. Existing kinetochore-tracking assays have a frame rate in the 7.5–15 s range (Dumont et al., 2012; Jaqaman et al., 2010; Wan et al., 2012), meaning that detailed analysis of the switching mechanism has not been possible to date.

Figure 1

Download asset Open asset

We developed a high-resolution kinetochore-tracking procedure and used a switching point detection algorithm to examine the fine detail of paired sister kinetochore trajectory data. Our previous kinetochore-tracking assay (Jaqaman et al., 2010) with a 7.5-s sampling time lacks the time resolution to resolve the relative sister switching order. To improve resolution, we used spinning disk confocal microscopy to capture 3D image stacks every 2 s over 150 time steps in HeLa-K cells expressing a marker for the kinetochores (either eGFP-CENP-A or eGFP-CENP-A eGFP-Centrin1; Figure 2B; Video 1, Video 2). Phototoxicity was minimal, as >90% of cells successfully underwent anaphase. Sister kinetochores were tracked as previously described (Jaqaman et al., 2010), except that we implemented 3D Gaussian mixture model fitting for determining sub-pixel spot locations (Thomann et al., 2002 see Figure 2A, C; Videos 3–5) – important here because our faster time sampling results in a smaller inter-frame spot displacement requiring higher localisation accuracy. This sub-pixel (super-resolution) tracking gives high theoretical positional accuracy (x,y = ± 2.8 nm; z = ± 5.7 nm; see ‘Materials and methods’) and reveals kinetochore dynamics in exquisite detail (Figure 2D). Consistent with previous work (Jaqaman et al., 2010; Vladimirou et al., 2013), sister kinetochores had a mean inter-kinetochore distance of ~910 nm and underwent quasi-periodic oscillations normal to the metaphase plate with a half-period of 35 s (Figure 2—figure supplement 1A,B). Finally, we constructed a Bayesian switching point inference algorithm that estimates from an observed sister pair trajectory the switching times for each sister (most probable frame) and the directional switching events by assignment of a direction of movement to each sister (see ‘Materials and methods’). Here, we focus on coherent runs (periods when the sisters are moving in the same direction) and the switching events that end runs. We tested this algorithm on simulated data where the true switch time is known giving accuracies of 94% (see ‘Materials and methods’; Figure 3A and Figure 3—figure supplement 1). This switching point algorithm determined whether the leading or trailing sister switches first in a directional reversal of the sister pair and by how many frames.

Figure 2 with 1 supplement see all

Download asset Open asset

Video 1

Download asset

Video 2

Download asset

Video 3

Download asset

Video 4

Download asset

Video 5

Download asset

Figure 3 with 1 supplement see all

Download asset Open asset

Next, we identified switch events in the trajectories of 1529 sister pairs across 55 eGFP-CENP-A cells and calculated the frequency that lead or trailing sisters switch first (n=9022 events ending a coherent run). We note that these trajectories were distributed throughout the metaphase plate (Figure 3B). We defined a directional switching event as a lead initiated directional switch (LIDS) if the lead sister switched at least one frame before the trailing sister and a trail initiated directional switch (TIDS) similarly. The remaining events correspond to sisters switching within the same frame and are denoted joint directional switches (JDS, e.g. see Figure 3C,D). These criteria illustrate that there is a strong lead bias with the fraction of LIDS and TIDS being 54.3% and 34.8%, respectively, the remaining fraction (10.9%) of events being joint. These data demonstrate that trailing sister switching is suppressed in human cells, leaving a lead-to-trail bias of 1.56:1.

Current directional switching models propose that switching of the lead kinetochore is initiated when the inter-sister distance (centromere spring) reaches a maximum stretch (tension) (Rieder and Salmon, 1994; Skibbens et al., 1995; Skibbens et al., 1993; Wan et al., 2012). Moreover, no mechanism has been proposed to explain trail first switching; trailing sister initiated switching has also been reported for PtK1 cells at a 15% frequency (Dumont et al., 2012). However, as we show here, the existence of trail first switching has ramifications for both the sister-sister coupling and the possible switching control mechanisms. By aligning profiles of the inter-sister distance – which reflects tension in the centromeric chromatin, 40 s before and 40 s after the first sister switching event (Figure 4A), we demonstrate clearly that LIDS and TIDS both have strong pre-event and post-event inter-sister distance signatures (Figure 4A; compare red [TIDS] and black [LIDS] traces). These stereotypical dynamic-tension signatures at LIDS/TIDS events (discussed below) demonstrate that our assignment of a LIDS or TIDS (Figure 3) are robust and meaningful.

Figure 4 with 2 supplements see all

Download asset Open asset

For a LIDS, the inter-sister distance increases corresponding to an average increase of spring stretch from 14% to 20% over the 20 s prior to the first sister switch, stabilises 4 s prior to that switch event and then decreases rapidly during the 6 s following the first sister switch to a minimum average stretch of 5% (Figure 4A; black trace). Here, we used a baseline rest length of 788 nm determined from nocodazole-treated cells (see ‘Materials and methods’). The LIDS event leaves the two K-fibres in a polymerising state, which decreases the inter-sister distance; this is in line with the standard model where the inter-sister distance will then undergo extension over the following coherent run (Figure 1). The relaxation of the spring immediately prior to the first sister switch appears to be the result of the lead sister slowing down (data not shown). This would be in line with experiments in Ptk1 cells in which the velocity slows at maximal inter-sister stretch (Wan et al., 2012). Finally, the switching of the second (trailing) sister correlates with the reduction in spring extension to near zero (Figure 4A), possibly suggesting that switching is the result of a loss of tension in accord with the standard tension model (Rieder et al., 1994).

However, the profile for a TIDS is distinct: the inter-sister distance is much lower prior to the first sister switch (Figure 4A; red trace). The maximum stretch (25%) is only reached 4 s after the switch has occurred. This indicates that the TIDS itself is necessary to build up tension in the centromeric chromatin while the maximum inter-sister stretch is higher than that seen during LIDS (average 200 nm spring extension). The LIDS and TIDS signatures are also present if we condition on the previous event type (Figure 4B), showing that although the previous event type does affect the inter-sister distance prior to the next event, the qualitative form is similar close to the switching event. The high overstretch of the centromeric chromatin during a TIDS raises a fundamental challenge to the standard tension model of kinetochore oscillations which states that the spring tension rises during a run of the sister pair (to the left or right) triggering a leading sister switch and relaxes during a directional switching event which triggers switching of the second sister (Figure 1). As illustrated above, this describes a repeated LIDS choreography. However, following a TIDS event, the centromeric spring tension escalates during the directional switch and starts to relax 4 s after the directional switch (Figure 4A), suggesting that the tension remains high over the following run.

To dissect this further, we examined the inter-sister distance over averaged coherent runs, (n=6339) by aligning the start and the end of the intervening run (rescaling time of each run to a standard length of 1) (Figure 4C,D). For runs with a preceding switch that was either a LIDS or a TIDS, the inter-sister distance rose (Figure 4C) or relaxed (Figure 4D) over the following run, respectively. At the end of the run, the spring profile had the LIDS or TIDS signature depending on the next event type (Figure 4C,D). Crucially, for runs starting after a TIDS the subsequent directional switch still has a LIDS bias (1.53 compared to a bias of 1.64 for a preceding LIDS), while the (median) time of the coherent run is identical to that following a LIDS (27.1 ± 0.21 s for TIDS, 26.7 ± 0.29 s for LIDS, p=5.5%). However, TIDS take a shorter amount of time to complete (second sister switches) than LIDS, median times 4.05 ± 0.15 s versus 4.31 ± 0.13 s, respectively (p=1.3 x 10^–4). In essence, kinetochore oscillations are robust to sister switching order and the dynamics of the kinetochores is nearly identical after a TIDS and LIDS, except that the inter-sister distance decreases instead of increases, respectively (Figure 4C,D). Our observations suggest that the classic choreography (Figure 1) represents half of the dynamic with the inter-sister distance relaxing under a LIDS then increasing over the subsequent coherent run – depolymerising K-fibres pulling kinetochores with greater force than polymerising fibres push. However, under a TIDS the inter-sister distance is overstretched and relaxes during the subsequent coherent run – the centromeric spring force increasing the velocity of the trailing sister so that it exceeds that of the lead sister. This TIDS choreography implies that directional switching cannot be triggered by a threshold on the spring tension as overstretching and subsequent relaxation of the spring tension prior to a directional switch are a natural part of metaphase kinetochore oscillations.

Kinetochores move in 3D with the kinetochore sister axis being compliant to twist away from the metaphase plate normal (Figure 2D,E). The twist and inter-sister distance are in fact inversely correlated (r=–0.13, significant at p<10^–200), with twist showing inverted profiles over switching events relative to that for the inter-sister distance (Figure 5). Specifically, the twist of the sisters falls prior to a LIDS and increases after the switch event, while a TIDS demonstrates the opposite (Figure 5A). The average twist increase seen at a LIDS relaxes over the following coherent run (Figure 5C), similar to the relaxation in the spring extension seen after a TIDS. This can be explained mechanistically since a high inter-sister stretch, indicative of a high inter-sister tension, aligns the sisters, reducing the twist, while a low stretch, with low tension, allows the twist angle to increase under thermal and mechanical fluctuations. Thus, a negative correlation between twist and inter-sister distance is predicted if there is mechanical compliance in the attachment of the kinetochore to the K-fibres and the centromeric spring.

Figure 5

Download asset Open asset

Examination of the inter-sister distance profiles between consecutive events reveals a key pattern; on average, a TIDS is associated with a lower inter-sister distance than a LIDS 4 s prior to the event regardless of a preceding LIDS or TIDS (Figure 4B), a LIDS event having a mean separation above 920 nm (dark blue and grey traces), a TIDS event being below 920 nm (orange and light blue traces). Coupled with the fact that the run is invariant to the nature of the preceding switching event, particularly with regard to timing (the coherent run time to the next directional switch is identical, p=5.5%), this suggests that switching of the lead and trail sisters are governed by clocks. This could be a mechanical timing mechanism associated with the attached K-fibre, and thus force sensitive. In this way, the trailing sister is kept in a polymerising state by the pulling force from the stretched centromeric chromatin (see model in Figure 6). This may reflect the ability of the kinetochore to inhibit catastrophe when under tension – as shown by biophysical experiments using purified budding yeast kinetochores (Akiyoshi et al., 2010). If that tension falls too low (or fails to build up), a TIDS occurs (as microtubules in the K-fibre undergo catastrophe; Figure 6). The same stabilisation mechanism can be invoked to explain switching resolution – during a LIDS the second sister (previously trailing), switches under the loss of the tension in the centromeric chromatin. When tension remains high enough to stabilise the trailing sister, then the lead sister switches because of its clock. The escalating force during a TIDS (after the trailing sister switches) could accelerate the lead sister clock triggering switching of the second sister. Again, this reflects the increase in the rescue rate of in vitro microtubules attached to kinetochores as shown in Akiyoshi et al., 2010, while it emphasises that the two sisters are not symmetrical and the time since their last switch event determines which sister switches.

Figure 6

Download asset Open asset

Although we have detected clear signatures in the switching choreographies (Figure 4), these reflect regulatory and mechanical processes of a highly stochastic system. This stochasticity is evident on many scales. First, kinetochores are known to display a range of stochasticity in their trajectories, from near deterministic oscillations to the near random (Jaqaman et al., 2010; Magidson et al., 2011). Second, the switching times are stochastic; the duration of a coherent run has a large variability with a coefficient of variation (standard deviation [SD]/mean) of 0.45, similar for both runs terminated by an LIDS (mean time and SD 29.6 ± 12.7 s) or a TIDS (29.2 ± 13.6 s, Figure 4—figure supplement 1). Thus, although we have invoked a clock mechanism as a switching time regulator, it is inherently stochastic. This stochasticity could stem from both the number and depolymerisation/polymerisation state of individual microtubules that make up the K-fibre. The fraction of microtubules that are in a polymerising state within a K-fibre is highly variable among kinetochores (Armond et al., 2015), indicating that growing and shrinking K-fibres are unlikely to be composed of fully coherent microtubules. Third, the signatures in Figure 4 are a mean behaviour, while variability in the inter-sister stretch throughout the dynamics is in fact large. This can be seen at the population level of trajectories, where the inter-sister distance distributions for LIDS and TIDS only show marginal separation before the switching event (–4 s), are hardly separated at the event, while separation increases after the event (+4 s) (Figure 4—figure supplement 2A–C). The inter-sister distance distribution over a switching event is in fact far from bimodal; a mixture of two Gaussian distributions requires the respective means to be separated by at least 2 SDs to be bimodal, the largest we observe is 70% at 4 s post event (Figure 4—figure supplement 2A–C). The separation of these distributions does not improve even on further categorising by the prior event (i.e. prior LIDS or TIDS; Figure 4—figure supplement 2D–F). Therefore, we have to conclude that the signatures shown in Figure 4 are not a universal behaviour but only detectable on averaging; that is, the actual switching process is highly stochastic. It may be that analysis of the most deterministic trajectories will reduce this stochasticity in the switching dynamics and signatures. However, directional switching may be a composite process that integrates over multiple signals, that is, tension may not be the only determinant; the stochasticity in our signatures would then be due to measuring only one of these determinants. It remains unknown to what extent the stochasticity in switching time and switching type explains the observed diversity in kinetochore trajectory dynamics, or whether other sources of variability exist.

This paper demonstrates how dynamic and mechanistic insight can be extracted from high-resolution tracking data. Although the leading sister typically initiates directional switching, the reverse switching order is also frequently observed (Figure 3D), with directional switching biases, timings and subsequent oscillations remaining robust to such events. By classifying switching order events, we were able to demonstrate clear stereotypical behaviour associated with these events, with both prior (potentially causal) and post-event signatures in the inter-sister distance dynamics. This confirms that event classification is physical and not due to noisy fluctuations in switching times coming from localisation measurement noise.

Our data support a new model of kinetochore oscillations comprising mechanical clocks on both the lead and trailing sister, which likely reflect the time- and force-dependent rescue and catastrophe of the K-fibre microtubules. Our analysis suggests that the degree of stretch of the inter-sister centromeric chromatin is a major determinant in orchestrating this switching; first, a mechanism based on stabilisation of the trailing sister polymerisation state through the centromere tension, effectively slowing the trailing sister clock, and second, a clock on the lead sister that is accelerated under high tension (Figure 6). Our data indicate that the standard tension model of sister kinetochore switching (Figure 1) is only able to explain part of the dynamics, while it is incompatible with the overstretch and subsequent relaxation of the inter-sister distance following a TIDS. Thus, kinetochores utilise multiple sensory and fail-safe mechanisms that ensure high-fidelity chromosome organisation within the spindle, despite high levels of stochasticity.

1. 1. Akiyoshi B
   2. Sarangapani KK
   3. Powers AF
   4. Nelson CR
   5. Reichow SL
   6. Arellano-Santoyo H
   7. Gonen T
   8. Ranish JA
   9. Asbury CL
   10. Biggins S

   (2010) Tension directly stabilizes reconstituted kinetochore-microtubule attachments

   Nature 468:576–579.

   https://doi.org/10.1038/nature09594

   • Google Scholar
2. 1. Armond JW
   2. Vladimirou E
   3. Erent M
   4. Mcainsh AD
   5. Burroughs NJ

   (2015) Probing microtubule polymerisation state at single kinetochores during metaphase chromosome motion

   Journal of Cell Science 128:1991–2001.

   https://doi.org/10.1242/jcs.168682

   • Google Scholar
3. 1. Compton DA

   (2000) Spindle assembly in animal cells

   Annual Review of Biochemistry 69:95–114.

   https://doi.org/10.1146/annurev.biochem.69.1.95

   • Google Scholar
4. 1. Dumont S
   2. Salmon ED
   3. Mitchison TJ

   (2012) Deformations within moving kinetochores reveal different sites of active and passive force generation

   Science 337:355–358.

   https://doi.org/10.1126/science.1221886

   • Google Scholar
5. 1. Jaqaman K
   2. King EM
   3. Amaro AC
   4. Winter JR
   5. Dorn JF
   6. Elliott HL
   7. Mchedlishvili N
   8. Mcclelland SE
   9. Porter IM
   10. Posch M
   11. Toso A
   12. Danuser G
   13. Mcainsh AD
   14. Meraldi P
   15. Swedlow JR

   (2010) Kinetochore alignment within the metaphase plate is regulated by centromere stiffness and microtubule depolymerases

   The Journal of Cell Biology 188:665–679.

   https://doi.org/10.1083/jcb.200909005.dv

   • Google Scholar
6. 1. Ke K
   2. Cheng J
   3. Hunt AJ

   (2009) The distribution of polar ejection forces determines the amplitude of chromosome directional instability

   Current Biology 19:807–815.

   https://doi.org/10.1016/j.cub.2009.04.036

   • Google Scholar
7. 1. Magidson V
   2. O'Connell Christopher B.
   3. Lončarek J
   4. Paul R
   5. Mogilner A
   6. Khodjakov A

   (2011) The spatial arrangement of chromosomes during prometaphase facilitates spindle assembly

   Cell 146:555–567.

   https://doi.org/10.1016/j.cell.2011.07.012

   • Google Scholar
8. 1. Mcintosh JR
   2. Molodtsov MI
   3. Ataullakhanov FI

   (2012) Biophysics of mitosis

   Quarterly Reviews of Biophysics 45:147–207.

   https://doi.org/10.1017/S0033583512000017

   • Google Scholar
9. 1. Rieder CL

   (1994) Motile kinetochores and polar ejection forces dictate chromosome position on the vertebrate mitotic spindle

   The Journal of Cell Biology 124:223–233.

   https://doi.org/10.1083/jcb.124.3.223

   • Google Scholar
10. 1. Rieder CL

    (2005) Kinetochore fiber formation in animal somatic cells: dueling mechanisms come to a draw

    Chromosoma 114:310–318.

    https://doi.org/10.1007/s00412-005-0028-2

    • Google Scholar
11. 1. Skibbens RV
    2. Rieder CL
    3. Salmon ED

    (1995)

    Kinetochore motility after severing between sister centromeres using laser microsurgery: evidence that kinetochore directional instability and position is regulated by tension

    Journal of Cell Science 108 (Pt 7):2537–2548.

    • Google Scholar
12. 1. Skibbens RV
    2. Skeen VP
    3. Salmon ED

    (1993) Directional instability of kinetochore motility during chromosome congression and segregation in mitotic newt lung cells: a push-pull mechanism

    The Journal of Cell Biology 122:859–875.

    https://doi.org/10.1083/jcb.122.4.859

    • Google Scholar
13. 1. Stumpff J
    2. Wagenbach M
    3. Franck A
    4. Asbury Charles L.
    5. Wordeman L

    (2012) Kif18A and chromokinesins confine centromere movements via microtubule growth suppression and spatial control of kinetochore tension

    Developmental Cell 22:1017–1029.

    https://doi.org/10.1016/j.devcel.2012.02.013

    • Google Scholar
14. 1. Thomann D
    2. Rines DR
    3. Sorger PK
    4. Danuser G

    (2002) Automatic fluorescent tag detection in 3D with super-resolution: application to the analysis of chromosome movement

    Journal of Microscopy 208:49–64.

    https://doi.org/10.1046/j.1365-2818.2002.01066.x

    • Google Scholar
15. 1. Thompson RE
    2. Larson DR
    3. Webb WW

    (2002) Precise nanometer localization analysis for individual fluorescent probes

    Biophysical Journal 82:2775–2783.

    https://doi.org/10.1016/S0006-3495(02)75618-X

    • Google Scholar
16. 1. Vandenbeldt KJ
    2. Barnard RM
    3. Hergert PJ
    4. Meng X
    5. Maiato H
    6. Mcewen BF

    (2006) Kinetochores use a novel mechanism for coordinating the dynamics of individual microtubules

    Current Biology 16:1217–1223.

    https://doi.org/10.1016/j.cub.2006.04.046

    • Google Scholar
17. 1. Vladimirou E
    2. Mchedlishvili N
    3. Gasic I
    4. Armond Jonathan W
    5. Samora Catarina P
    6. Meraldi P
    7. Mcainsh Andrew D

    (2013) Nonautonomous movement of chromosomes in mitosis

    Developmental Cell 27:60–71.

    https://doi.org/10.1016/j.devcel.2013.08.004

    • Google Scholar
18. 1. Wan X
    2. Cimini D
    3. Cameron LA
    4. Salmon ED

    (2012) The coupling between sister kinetochore directional instability and oscillations in centromere stretch in metaphase PtK1 cells

    Molecular Biology of the Cell 23:1035–1046.

    https://doi.org/10.1091/mbc.E11-09-0767

    • Google Scholar
19. 1. Weber SC
    2. Spakowitz AJ
    3. Theriot JA

    (2012) Nonthermal ATP-dependent fluctuations contribute to the in vivo motion of chromosomal loci

    Proceedings of the National Academy of Sciences 109:7338–7343.

    https://doi.org/10.1073/pnas.1119505109

    • Google Scholar
20. 1. Wendell KL
    2. Wilson L
    3. Jordan MA

    (1993)

    Mitotic block in HeLa cells by vinblastine: ultrastructural changes in kinetochore-microtubule attachment and in centrosomes

    Journal of Cell Science 104:261–274.

    • Google Scholar

Author details

1. Nigel J Burroughs

   Warwick Systems Biology Centre, Warwick Mathematics Institute, University of Warwick, Coventry, United Kingdom

   Contribution

   NJB, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   Contributed equally with

   Edward F Harry

   For correspondence

   N.J.Burroughs@warwick.ac.uk

   Competing interests

   The authors declare that no competing interests exist.

2. Edward F Harry

   Warwick Molecular Organisation and Assembly in Cells, University of Warwick, Coventry, United Kingdom

   Contribution

   EFH, Acquisition of data, Analysis and interpretation of data

   Contributed equally with

   Nigel J Burroughs

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0003-1914-6432

3. Andrew D McAinsh

   Centre for Mechanochemical Cell Biology, Division of Biomedical Cell Biology, Warwick Medical School, University of Warwick, Coventry, United Kingdom

   Contribution

   ADM, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   A.D.McAinsh@warwick.ac.uk

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0001-6808-0711

• 1,947

  views

• 612

  downloads

• 17

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.