News & Views | Published:

Cell division

Feeling tense enough?

Accurately distributing half of each replicated chromosome to both daughters is a major challenge for dividing cells. The mechanisms used to achieve this are becoming apparent, thanks to studies old and new.

At the beginning of mitosis, the process of cell division, chromosomes are organized randomly — like jigsaw puzzle pieces spread out on the floor. Their constituent two ‘sister chromatids’, each of which contains one of the two identical DNA molecules produced by replication, must be oriented such that they will be pulled in opposite directions into the two newly forming cells. Like a jigsaw, the solution for correctly orienting all chromosomes comes partly through trial and error. Mechanisms must exist to eliminate wrong configurations while selecting the right ones. The nature of such mechanisms was first suggested by micromanipulation experiments begun more than 30 years ago1. Recent molecular analyses in budding yeast, including a new paper from Dewar et al. on page 93 of this issue2, extend this earlier work by showing that mechanical tension is crucial to solving the chromosome orientation puzzle.

To segregate chromosomes, eukaryotic cells assemble spindles, bipolar arrays of microtubule polymers formed from the protein tubulin. Spindle microtubules attach to a single site on each chromatid where a large multi-protein complex, the kinetochore, is formed3. When mitosis begins, microtubules start a ‘search and capture’ process to find these kinetochores. This process is random, so some kinetochores fail to attach efficiently to spindle microtubules. Others generate incorrect attachments — for example, where both sister kinetochores attach to the same pole (Fig. 1). Such ‘syntelic’ attachments must be corrected so that all chromosomes are bi-oriented — that is, sister kinetochores are stably attached to microtubules from opposite spindle poles. Only then can the glue holding the sister chromatids together be dissolved so that the two spindle poles and their associated chromatids are distributed to the daughter cells.

Figure 1: Bi-orientation is required for accurate division of sister chromatids.

a, Microtubules of the spindle attach to sister chromatids through multi-protein complexes known as kinetochores. If the microtubules are connected to the opposing spindle poles, the chromosomes are said to be bi-oriented, and tension is generated because the sister chromatids are held together by a cohesin glue. If, however, the kinetochores of sister chromatids attach to microtubules coming from the same pole, the chromosomes are said to be syntelic, and no tension is generated. The Aurora B kinase recognizes and corrects the lack of tension in syntelic chromosomes. b, Dewar et al.2 made a circular minichromosome containing two kinetochores and showed that tension could still be generated in the absence of cohesin because the kinetochores were still physically connected.

Micromanipulation studies1 first provided evidence that tension between two kinetochores — which is generated only in the bi-oriented state (Fig. 1a) — discriminates between bi-oriented and syntelic attachments. However, these studies were conducted on cells undergoing the first phase of meiosis, the reductive division used to generate gametes. During this phase, two homologous chromosomes, rather than sister chromatids, are connected to opposite spindle poles, so the relevance of this work in mitosis was unclear. Studies of mitotic chromosomes have shown that sister kinetochores face in opposite directions regardless of whether they are attached to spindle microtubules4. So when one kinetochore attaches to the spindle it might geometrically force its sister to face the opposite spindle pole and thereby prevent syntely. But recent work demonstrating frequent syntelic attachments during mitosis suggests that additional mechanisms also function to ensure bi-orientation. Consistent with this expectation is the observation that syntelic attachments are corrected by a highly conserved enzyme, Aurora B kinase, in both yeast and vertebrates5,6.

Now, Dewar et al.2 have elegantly manipulated budding yeast chromosomes to show that the tension generated in the physical linkage between bi-oriented kinetochores and the activity of Aurora B, rather than any specific chromosomal architecture, are sufficient to ensure that the sister chromatids are appropriately aligned. They reached this striking conclusion in two ways, both of which relied on microscopic imaging of fluorescently tagged chromosomes to study their bi-orientation in living cells.

Under normal circumstances, protein complexes termed cohesins hold sister chromatids together until all chromosomes are properly aligned on the spindle. In the bi-oriented state, microtubules from opposite spindle poles attach to sister kinetochores, and poleward forces act on the kinetochores to generate tension in the connection between them. Using yeast cells that contained defective cohesin and inactive topoisomerase II, an enzyme that resolves physically intertwined DNA, Dewar et al.2 showed that sister chromatids with physically linked DNA molecules can efficiently bi-orient without cohesin. A similar finding was also reported recently in vertebrate cells7. So cohesin complexes and the specific connection they form between sister chromatids are not required for bi-orientation — a different type of physical connection will suffice.

In a technical tour de force, the authors engineered a circular minichromosome with two kinetochores on a single DNA molecule. Remarkably, these artificial minichromosomes bi-oriented with similar efficiency and kinetics to those of normal chromosomes (Fig. 1b). Thus, any physical connection between two kinetochores that supports the development of tension can facilitate bi-orientation.

Dewar et al.2 also show that Aurora B kinase is required to bi-orient the artificial minichromosomes with two kinetochores. Therefore, Aurora B must recognize incorrect attachments simply by virtue of their lack of tension, regardless of the precise structure that links the two kinetochores. Finding out how Aurora B identifies and corrects them is an obvious next step. In budding yeast, where kinetochores assemble on a short (120-base-pair) piece of DNA and bind to just a single microtubule, Aurora B must periodically detach kinetochores from microtubules until the tense, bi-oriented state is achieved. The authors2 provide direct evidence for this by generating yeast cells that have more than two spindle poles. In such aberrant cells, an engineered minichromosome with just one kinetochore, which can only form a single attachment and thus lacks tension, moves repeatedly between the different spindle poles as though it is searching for a stable conformation. If Aurora B function is inhibited, the minichromosome remains linked to just one spindle pole, indicating that Aurora B detaches microtubules from kinetochores in the absence of tension.

In contrast to budding yeast, kinetochores of other eukaryotes bind multiple microtubules (about 20 in humans)8. These larger kinetochores must coordinate all these microtubules and also deal with incorrect attachments in which microtubules from opposite spindle poles connect to a single kinetochore (termed ‘merotely’)9. Another study10, in this month's Nature Cell Biology, found that Aurora B does not merely detach syntelic kinetochores from microtubules in vertebrates — it orchestrates the coordinated disassembly of all the microtubules that are bound to each kinetochore, so that the syntelically oriented chromosomes move towards the spindle poles before they are bi-oriented.

Although sister kinetochore geometry seems to be dispensable in budding yeasts with their single-microtubule-connected kinetochores, it could contribute to reducing merotely, as implied by the conservation of this aspect of chromosome architecture throughout eukaryotic evolution. Tackling the extra dimension that the multiplicity of microtubule-binding sites at kinetochores introduces will undoubtedly be another brain-teaser — and a particularly important one, too, because the loss of a single chromosome can be lethal, and aberrant numbers of chromosomes can contribute to birth defects and cancer.


  1. 1

    Nicklas, R. B. & Koch, C. A. J. Cell Biol. 43, 40–50 (1969).

  2. 2

    Dewar, H., Tanaka, K., Nasmyth, K. & Tanaka, T. U. Nature 428, 93–97 (2004).

  3. 3

    Cleveland, D. W., Mao, Y. & Sullivan, K. F. Cell 112, 407–421 (2003).

  4. 4

    Rieder, C. L. Int. Rev. Cytol. 79, 1–58 (1982).

  5. 5

    Tanaka, T. U. et al. Cell 108, 317–329 (2002).

  6. 6

    Hauf, S. et al. J. Cell Biol. 161, 281–294 (2003).

  7. 7

    Vagnarelli, P. et al. EMBO Rep. 5, 167–171 (2004).

  8. 8

    Bloom, K. Cell 73, 621–624 (1993).

  9. 9

    Cimini, D. et al. J. Cell Biol. 153, 517–527 (2001).

  10. 10

    Lampson, M. A., Renduchitala, K., Khodjakov, A. & Kapoor, T. M. Nature Cell Biol. 6, 232–237 (2004).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.