Tension on kinetochore substrates is insufficient to prevent Aurora-triggered detachment

Background:

When proliferating, cells divide their genetic material by employing a highly complex and dynamic machine – the mitotic spindle. Microtubules emanating from opposite poles of the spindle contact duplicated chromosomes at specialised macromolecular structures called kinetochores. Microtubule-kinetochore interactions are a vital aspect of cell division and have incidentally been the focus of many preprints quite recently (1-3).

To faithfully pull apart sister chromatids, i.e. to provide every new cell a full complement of genetic material, each kinetochore in a pair must attach to a microtubule from opposite poles. Crucially, cells must be able determine whether this has been achieved before segregating chromosomes. This complex task is performed by Aurora B kinase. Proper attachments involve kinetochore pairs pulled on from opposite poles, thereby generating tension on kinetochores. Aurora B selectively phosphorylates its substrate proteins on kinetochores under low tension (4), where presumably both kinetochores attach to one pole. This phosphorylation reduces their affinity for microtubules, destabilising these erroneous attachments. How does Aurora B sense the level of tension on kinetochores, and is that the only cue underlying kinase-triggered detachment of microtubules?

Kinetochores are known to undergo dramatic structural changes when under tension (5-7). An appealing hypothesis is that this occludes key Aurora B substrates from being phosphorylated, thus retaining correct microtubule attachments. It is a considerable challenge to test such models in vivo. Aurora B performs several roles in cell division and isolating its function at kinetochores, particularly in the presence of opposing phosphatase activity, is not trivial.

Key findings:

The authors in this study distil this complex intracellular phenomenon into a reconstituted in vitro assay. They purify yeast kinetochores such that endogenous kinase and phosphatase activity has been quenched. Next, they design and purify an active recombinant Aurora B kinase (AurB*) that can phosphorylate key microtubule-interacting substrates (particularly the Ndc80 protein) on isolated kinetochores.They show that AurB* phosphorylation impairs the ability of these kinetochores to bind fluorescently labelled, stabilised microtubules in vitro. This effect is partially Ndc80 dependent.

What is the effect of AurB* on physiologically relevant microtubule attachments that are load-bearing, end-on and dynamic? The authors modified an optical trap assay (8) to generate tension on microtubule-attached kinetochores where AurB* can be flowed in. In such experiments, kinetochores linked to polystyrene beads are added to coverslips on which anchored microtubules are grown. The position and force experienced by these beads can be controlled via a tightly focused laser beam (optical trap). The kinetochores are then indirectly manipulated by the laser to attach to tips of dynamic microtubules and held at a desired force (to generate tension). Exogenous AurB* is added once the attachment is made.

While these manipulations are technically challenging, the authors were able to observe that phosphorylation by AurB* releases tips of microtubules from kinetochores under low tension (~1 piconewton (pN)), as would be predicted for an improper attachment in vivo. A kinase-dead mutant of AurB* is 3-fold less effective in performing this “correction”.

Importantly, the prevailing model also predicts that a properly attached kinetochore under high tension should not be detached from microtubules by Aurora B. However, quite surprisingly, the authors find that AurB* activity releases microtubule attachments even at high tension regimes (5 and 8 pN). This suggests that tension across kinetochores is not sufficient to block Aurora B phosphorylation on kinetochore substrates.

 

What I like about this preprint:

It is very exciting to see a reconstituted system that draws on decades of work and has the potential to test several dogmas in the field of chromosome segregation. To quote the authors, “Rigorously testing such models requires independent control of tension, attachment, and enzyme activity”. This study (a) tackles all three of these challenges, (b) reveals unexpected complexities in the system and (c) sets the stage to further probe elusive mechanisms of how our genomes are faithfully divided.

Future directions and questions for the authors:

In the light of their data, the authors state that alternate models for error correction bear consideration. For example, we think of tension as primarily modifying the kinetochore but it may instead affect kinase activity or even that of a phosphatase. These would be very exciting new directions for the field. However, to delve further in the current experiments, I would love to ask the authors a few questions:

1. How do detachment rates of non-phosphorylatable kinetochores (Ndc80-7A) compare with wild-type under low and high tension? In other words, could detachment under high tension be occurring via an unexpected mechanism?
2. Could the response to kinetochore tension depend on strength of AurB activity? That is, if AurB* activity is too high, perhaps it phosphorylates its kinetochore substrates regardless of tension. In vivo AurB activity at kinetochores might be tightly regulated to respond to changes in tension. Were a range of AurB* concentrations tested, other than 0.5uM and 5uM, as even 0.5uM of exogenous, chimeric AurB* may exceed a threshold?
3. It seems surprising that not only does AurB* function persist at high tension, but the detachment rates increase with increasing force. How would you explain this positive correlation?

References:

1. Kuhn and Dumont, bioRxiv 2018; https://doi.org/10.1101/463471
2. Roy et al., bioRxiv 2018; https://doi.org/10.1101/459594
3. Doodhi et al., bioRxiv 2018; https://doi.org/10.1101/455873
4. Kelly and Funabiki., Curr. Op. in Cell Biol 2009
5. Wan et al., Cell 2009
6. Joglekar et al., Curr. Biol 2009
7. Tien et al., Genetics 2014
8. Akiyoshi et al., Nature 2010

 

doi: https://doi.org/10.1242/prelights.5522

Read preprint (No Ratings Yet)