Chromosomal instability: stretching the role of checkpoint silencing

Abstract

Microtubule attachments to kinetochores cause their deformation — a murky phenomenon known as intra-kinetochore stretching. A new study proposes that intra-kinetochore stretching is independent of microtubule-pulling forces and mediates efficient spindle assembly checkpoint silencing to prevent chromosomal instability.

During cell division, two copies of the genome must be equally segregated between daughter cells, as failure to do so may cause chromosomal instability, a hallmark of human cancers¹. To separate the chromosomes, the cell assembles a mitotic spindle composed of microtubules and microtubule associated proteins. Faithful chromosome segregation requires the interaction between spindle microtubules and kinetochores, multi-protein platforms that assemble on the centromeres of chromosomes at the start of mitosis. Kinetochores are also the hub of the spindle assembly checkpoint (SAC) that delays sister chromatid separation until all kinetochores are attached to the spindle². The mechanism underlying the generation and silencing of this ‘wait-anaphase’ signal has been a key question in the field since the early 1990s. The microtubule attachment status of kinetochores seemed the most obvious explanation, since single unattached kinetochores can hold cells in mitosis for several hours³ and laser-mediated ablation of the last unattached kinetochore triggers anaphase within minutes⁴. Moreover, SAC proteins such as MAD2 are released from kinetochores upon microtubule attachment, allowing the activation of the anaphase promoting complex/cyclosome required for anaphase onset². However, a key experiment by Bruce Nicklas in the mid-1990s added an extra layer of complexity: using calibrated microneedles, the act of pulling on a syntellically attached chromosome (that is, one in which both kinetochores attach to a common spindle pole) could satisfy the SAC and trigger anaphase⁵. This provided evidence that the SAC was sensitive to attachment, but also to tension by spindle forces acting on centromeres. This was readily measurable by the distance between sister kinetochores, which increases as chromosomes become attached to the spindle and bi-orient. Because attachments on bi-oriented chromosomes generate tension, and tension promotes attachment stability at kinetochores⁶, it created a vicious cycle (and endless discussions) in which dissecting the relative contributions of attachment versus tension for SAC silencing became virtually impossible.

In 2009, in two back-to-back studies using Drosophila ⁷ and human cell lines⁸, the labs of E. Salmon and T. Hirota, respectively, attempted to break this cycle by revealing a second type of tension. Using differential fluorescence labelling of both inner and outer kinetochore proteins to assess their separation over the course of mitosis, both groups surprisingly found that, upon microtubule attachment, kinetochores themselves could stretch around 100 nm in a cyclic manner. This kinetochore deformation (also referred to as ‘delta’) was largely independent of microtubule pulling by dynein at kinetochores⁷ and correlated well with SAC silencing under conditions in which centromere stretch was negligible (for example, when low doses of taxol or nocodazole were applied^7,8; Figure 1A). It was also found that sister kinetochores undergo stretching independently, suggesting that this phenomenon is intrinsic to each kinetochore⁸. Together with another study in cells undergoing mitosis with unreplicated genomes, in which single kinetochores were able to satisfy the SAC⁹, these studies made clear that centromere stretching was not a requirement for SAC silencing. However, the debate over whether microtubule attachments or intra-kinetochore stretching was responsible for SAC silencing only got more lively¹⁰. Moreover, whether intra-kinetochore stretching was caused by tension or by an internal conformational change caused by microtubule attachments remained unclear. Now, a decade later, a new study¹¹ by the Hirota lab published in this issue of Current Biology sheds light on the topic.

Figure 1. The different force models for intra-kinetochore stretching and its role in preventing chromosomal instability.

A) The various models of intra-kinetochore stretching observed on monopolar spindles that have been suggested and tested throughout the literature. Intra-kinetochore stretching appears to be independent of tensile forces and may occur in the absence of polar ejection forces. B) Inhibiting intra-kinetochore stretching slows spindle assembly checkpoint silencing, which prevents an abrupt activation of separase at anaphase onset²⁰. This leads to an increase in chromosome bridges caused by inefficient cohesin removal and may account for chromosomal instability¹¹.

Using monopolar spindles, where only one kinetochore of a pair is transiently attached (that is, mono-oriented), Uchida et al.¹¹ show that intra-kinetochore stretching still takes place and this depends on the establishment of end-on kinetochore–microtubule attachments. To discount the possible contribution of polar ejection forces on chromosome arms in the generation of intra-kinetochore stretching, the authors studied kinetochore behaviour in cells depleted of the chromokinesin KIF22A/Kid, and on human artificial chromosomes (HAC), which lack chromosome arms. Neither condition produced any tangible changes in intra-kinetochore stretching. Although it could be argued that other chromokinesins (for example, KIF4A) or the dynamic instability of astral microtubules might still generate enough polar ejection force on mono-oriented chromosomes, the HAC experiment elegantly demonstrates that intra-kinetochore stretching may occur independently of polar ejection forces on chromosome arms (Figure 1A). Hirota and co-workers¹¹ favour the idea that intra-kinetochore stretching is independent of microtubule pulling forces at kinetochores, however, classic laser microsurgery experiments in which mono-oriented chromosomes were cut free of chromosome arms have shown that the resulting kinetochore-containing fragment continued to oscillate toward and away from the pole¹², indicating that it still experiences both pulling and pushing forces (Figure 1A). Interestingly, in fixed-cell analyses HACs seem to be positioned closer to the center of monopolar spindles when compared with normal arm-containing chromosomes¹¹, suggesting that attached kinetochores in HACs may experience pulling forces at some point. Because it is unlikely that cytoplasmic resistance is able to cause kinetochore deformation, and the stability of end-on kinetochore-microtubule attachments, including of mono-oriented chromosomes, depends on polar ejection forces^13,14 (Figure 1A), how kinetochores in mono-oriented HACs experience intra-kinetochore stretching while remaining attached to microtubules remains puzzling.

Maresca and Salmon⁷ previously showed that taxol treatment strongly reduced centromeric tension whilst having little effect on intra-kinetochore stretching, suggesting that microtubule plus-end dynamics play a role in generating a conformational change within the kinetochore. In their current study, Uchida et al.¹¹ found a strong correlation between growing microtubule plus-ends (indicated by enrichment of fluorescent EB3) and stretched kinetochores, supporting the idea of a microtubule-driven conformational change upon attachment, and they propose a provocative model whereby intra-kinetochore stretching is created by polymerisation at the microtubule tip within the kinetochore (Figure 1A). That growing microtubules can exert pushing forces on chromosomes has been known since the pioneering works of Shinya Inoue (reviewed in¹⁵), and taxol treatment in plants can even cause chromosomes to move backward during anaphase¹⁶. However, a study in 2016 by the Khodjakov lab used correlative light and electron microscopy to show that kinetochores undergo compaction upon microtubule attachment, and acute taxol treatment causes the outer kinetochore to expand outward due to microtubule detachment from kinetochores¹⁷, a condition that prevents SAC silencing. This finding raised caution about the interpretation of ‘delta’ measurements and the role played by intra-kinetochore stretching in the control of mitotic progression.

To overcome the limitations introduced by the use of microtubule drugs, Uchida and colleagues¹¹ devised a clever tool to investigate the role of intra-kinetochore stretch. Using the FKBP-FRB dimerization system, they designed a mutant version of CENP-T, which links the inner and outer plates of kinetochores¹⁸, that could be inducibly forced to fold upon itself by adding rapamycin. This folding prevented intra-kinetochore stretching and produced a metaphase delay by triggering MAD1 recruitment to a subset of kinetochores. This did not appear to be caused by lack of microtubule attachment, as these cells were able to form apparently normal kinetochore fibers (K-fibers) and showed unperturbed chromosome congression. The authors suggest that intra-kinetochore stretching is required for efficient recruitment of PP1 phosphatase necessary for the removal of checkpoint proteins at kinetochores and consequent checkpoint silencing¹⁹. This indicates that kinetochore stretching is somehow required for SAC silencing following the establishment of end-on attachments, yet much remains to be learned about how this new CENP-T mutant impacts kinetochore function. Intriguingly, Uchida and colleagues¹¹ found that only 40% of the kinetochores in metaphase experience intra-kinetochore stretch, with a similar observation in anaphase kinetochores when attached microtubules undergo net depolymerization, putting into doubt a model where SAC is silenced by kinetochore conformational changes induced by microtubule polymerisation forces. A similar correlative light and electron microscopy approach to that used by Khodjakov and colleagues¹⁷ to investigate microtubule attachments at high resolution on those MAD1-positive kinetochores appearing after induction of the CENP-T mutant would be the gold standard to unequivocally validate the model proposed by Uchida et al.¹¹ Likewise, it would be interesting to investigate whether microtubule dynamics at kinetochores are altered using photoactivation of tubulin and poleward flux measurements, as well as to investigate the dynamics of kinetochore oscillations and SAC protein removal upon reduction of intra-kinetochore stretching in live cells.

Finally, Uchida et al.¹¹ propose an interesting twist in the relationship between SAC function and cancer. Accordingly, they propose that rather than due to a defective SAC, chromosomal instability commonly observed in cancer cells might be related to defective SAC silencing (Figure 1B). Indeed, cells where intra-kinetochore stretching was prevented by the mutant CENP-T construct showed a metaphase delay and a two-fold increase in chromosome bridges (but not lagging chromosomes, suggesting kinetochore-microtubule attachments were unaffected) during anaphase and this might be due to inefficient cohesin removal²⁰ (Figure 1B). In agreement, some cancer cells with chromosomal instability showed lower frequency of kinetochore stretching and slower checkpoint silencing¹¹. Although future work is necessary to fully understand the cause and function of intra-kinetochore stretching and its implications for mitotic fidelity, this new study offers new clues and ingenious new tools for its manipulation in living cells.

In Brief.

Microtubule attachments to kinetochores cause their deformation — a murky phenomenon known as intra-kinetochore stretching. A new study proposes that intra-kinetochore stretching is independent of microtubule-pulling forces and mediates efficient spindle assembly checkpoint silencing to prevent chromosomal instability.