News & Views | Published:

Cell biology

The checkpoint brake relieved

Nature volume 446, pages 868869 (19 April 2007) | Download Citation

When a cell divides, each daughter cell inherits a complete set of chromosomes. A sophisticated inhibitory mechanism delays chromosome segregation and cell division until everything is in its place.

It is equally important for a cell to do things at the right times as it is for an organism. This is particularly true during the final phase of cell division — mitosis — when chromosomes segregate. Chromosome segregation must not be initiated too early, because otherwise the daughter cells will carry an abnormal number of chromosomes and, in rare cases, tumour cells or diseases such as Down's syndrome might develop. So, a special surveillance system known as the spindle-assembly checkpoint inhibits chromosome segregation until this process can be carried out properly. How this system works is one of the big mysteries of the cell cycle. On pages 876 and 921 of this issue, two studies from the laboratories of Elledge1 and Kirschner2 propose an intriguing model for how the activity of the spindle-assembly checkpoint may be maintained, and eventually relieved, when the cell is ready to separate its chromosomes.

Long before a cell divides, it generates an extra copy of its chromosomes by DNA replication, so that every chromosome contains two identical copies of DNA, packaged into two halves of the chromosome — the sister chromatids. During mitosis, sister chromatids are separated and are transported towards opposite poles of the cell so that, after division, each daughter cell carries an identical set of sister chromatids. For this to occur, it is essential that the two sister chromatids become attached to the opposite halves of the mitotic spindle before they are separated from each other. The spindle-assembly checkpoint ensures this by 'sensing' the presence of chromosomes whose two kinetochores (spindle-attachment sites) are not properly attached to the spindle3. As long as a dividing cell contains such unattached kinetochores, the spindle-assembly checkpoint is active and prevents the separation of sister chromatids4 (Fig. 1a).

Figure 1: Regulation of cell division by the spindle-assembly checkpoint.
Figure 1

a, Before a cell divides, the sister chromatids from all of its chromosomes have to attach to the two poles of the mitotic spindle. As long as chromosomes are present that do not obey this rule, a cellular surveillance system known as the spindle-assembly checkpoint is active. This inhibits the continuation of cell division and therefore prevents an abnormal number of sister chromatids reaching each daughter cell. b, At the molecular level, the spindle-assembly checkpoint proteins Mad2 and BubR1 associate with Cdc20 and inhibit its ability to activate the ubiquitin ligase APC/C, which is required for the progression of cell division. According to a model proposed by Elledge et al.1 and Kirschner et al.2, APC/C and the ubiquitin-conjugating enzyme UbcH10 can overcome this inhibition by ubiquitinating Cdc20, which causes Mad2 and BubR1 to become dissociated from it. Cdc20 ubiquitination is, in turn, reversed by the de-ubiquitinating enzyme USP44. A balance between ubiquitination and de-ubiquitination reactions may therefore control the activity of Cdc20. Once all chromosomes have been properly attached to the mitotic spindle, this balance would tip towards ubiquitination of Cdc20, with Cdc20 and the APC/C subsequently being able to initiate the separation of sister chromatids.

The spindle-assembly checkpoint halts cell division by inhibiting the anaphase-promoting complex/cyclosome (APC/C)5,6,7 — an enzyme that adds chains of the small protein ubiquitin to specific substrate proteins, thereby targeting them for destruction8. During cell division, these reactions are mediated by APC/C and two other proteins — Cdc20, which helps substrate recruitment to APC/C, and UbcH10, which is a ubiquitin-conjugating enzyme that transfers ubiquitin residues onto the substrate proteins.

These ubiquitinating reactions are inhibited when unattached kinetochores keep the spindle-assembly checkpoint active. The kinetochores promote the association of Cdc20 with at least two checkpoint proteins known as Mad2 and BubR1, which prevent Cdc20 from activating the APC/C9. But this inhibition can only persist until all the chromosomes are attached to both poles of the spindle, because soon after this the APC/C becomes active and sister chromatids begin to separate.

How, then, is Cdc20 relieved from its inhibition? It has often been assumed that the Cdc20–Mad2–BubR1 complex is simply short-lived, and that the inactivation of the spindle-assembly checkpoint might therefore passively release Cdc20 from its inhibitors. However, structural studies of the Mad2–Cdc20 interaction have indicated that Mad2, at least, holds on to Cdc20 very tightly, like the way that a safety belt secures a passenger in a car seat10,11. Thus it is plausible that the release of Cdc20 from Mad2 is an active process.

This is exactly what the two papers from the Kirschner and Elledge groups propose. More surprisingly, these authors provide evidence to suggest that the APC/C itself might be the molecule that liberates Cdc20 from inhibition by Mad2 and BubR1 (Fig. 1b).

The Kirschner2 lab came to this conclusion by studying extracts from cells with an active spindle-assembly checkpoint, which, not surprisingly, contained little or no APC/C activity. However, when UbcH10 was added to these extracts, APC/C became active, coincident with the ubiquitination of Cdc20 and the dissociation of Mad2 and BubR1 from Cdc20. Ubiquitination often targets proteins for destruction by the 26S proteasome enzyme complex, but Kirschner's team found that proteasome inhibition did not prevent the activation of APC/C by UbcH10. On the basis of these and other observations, the authors propose that the addition of a ubiquitin chain to Cdc20 by APC/C and UbcH10 does not necessarily involve protein degradation, but leads to the dissociation of Mad2 and BubR1 from Cdc20. One implication of this model is that APC/C would constantly antagonize its inhibition by the spindle-assembly checkpoint. If so, how could Mad2 and BubR1 ever inhibit the APC/C in cells with an active checkpoint?

A possible answer comes from a study carried out by Elledge's team1. In a search for proteins that are required for the activity of the spindle-assembly checkpoint, these authors identified a de-ubiquitinating enzyme known as USP44. Enzymes of this type disassemble ubiquitin chains by cleaving the bonds that connect the ubiquitin residues in the chain. Interestingly, USP44 differs from other known spindle-assembly checkpoint proteins in that it is not required to recruit Mad2 to unattached kinetochores, where Mad2 is believed to form complexes with Cdc20 and BubR1. So how else could USP44 function at the checkpoint? It turns out that, in vitro, USP44 can inhibit the ability of UbcH10 to activate checkpoint-inhibited APC/C, leading Elledge and colleagues to propose that USP44 might stabilize Cdc20–Mad2–BubR1 complexes by destroying the ubiquitin chains that APC/C adds to Cdc20 (Fig. 1b). Consistent with this argument, depletion of USP44 prematurely inactivates the spindle-assembly checkpoint in mitotic cells and leads to defects in chromosome segregation.

The model proposing that the stability of Cdc20–Mad2–BubR1 complexes is controlled by a fine balance between ubiquitination, mediated by the APC/C, and de-ubiquitination, catalysed by USP44, makes a number of predictions. Testing these will be an essential goal for the future. For example, could a mutant of Cdc20 be created that couldn't be ubiquitinated but would otherwise be functional? If so, such a mutant would be predicted to assemble into unusually stable checkpoint complexes from which Mad2 and BubR1 could not easily dissociate.

These studies1,2 also raise a number of other questions. Is de-ubiquitinating Cdc20 the main role of USP44 in maintaining the spindle-assembly checkpoint, or does it also antagonize APC/C more directly by disassembling ubiquitin chains on its other protein substrates such as securin and cyclin B? How does ubiquitination dissociate Mad2 and BubR1 from Cdc20 — by inducing conformational changes in these proteins, or by recruiting enzymes (such as the p97/Cdc48–ATPase) that would catalyse the dissociation process? Finally, how is the balance between de-ubiquitination and ubiquitination reactions tipped once all chromosomes have become attached to both poles of the mitotic spindle? Elledge et al. found that USP44 itself is degraded at the end of cell division. Could this be the primary switch for checkpoint inactivation, or is it merely a consequence of APC/C activation once the checkpoint has been silenced? Answering these questions will keep researchers busy for some time to come.


  1. 1.

    et al. Nature 446, 876–880 (2007).

  2. 2.

    , , & Nature 446, 921–925 (2007).

  3. 3.

    , , & J. Cell Biol. 127, 1301–1310 (1994).

  4. 4.

    Cell 120, 739–746 (2005).

  5. 5.

    , , , & Proc. Natl Acad. Sci. USA 94, 12431–12436 (1997).

  6. 6.

    et al. Science 279, 1041–1044 (1998).

  7. 7.

    , , , & Science 279, 1045–1047 (1998).

  8. 8.

    Nature Rev. Mol. Cell Biol. 7, 644–656 (2006).

  9. 9.

    , & J. Cell Biol. 154, 925–936 (2001).

  10. 10.

    , , & Mol. Cell 9, 59–71 (2002).

  11. 11.

    et al. EMBO J. 21, 2496–2506 (2002).

Download references

Author information


  1. Jan-Michael Peters is at the Research Institute of Molecular Pathology (IMP), Dr. Bohr-Gasse 7, A-1030 Vienna, Austria.

    • Jan-Michael Peters


  1. Search for Jan-Michael Peters in:

About this article

Publication history



Further reading


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.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing