Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Cell division

Genome maintenance

Early fruitfly embryos have an unusual means of halting the division of any nuclei containing damaged DNA. A key component of this mechanism has now been identified, and might have implications for cancer.

Maintaining the integrity of the genome is a crucial task for any cell. Two proteins, called checkpoint kinases 1 (Chk1) and 2 (Chk2), help to achieve this in many species, and mutations in the genes encoding these proteins have been linked to the generation of cancer in humans. The proteins are activated by DNA damage, and help to initiate cellular defence responses that include the stimulation of DNA-repair pathways and the slowing down of the cell-division cycle to allow time for repair1,2. In multicellular organisms, if the DNA is not successfully mended, the damaged cells usually kill themselves — thereby eliminating the defective genome. As they describe in Cell, Theurkauf and colleagues3 have discovered that Chk2 is also involved in a rather different defence mechanism that is triggered by DNA damage in early fruitfly embryos.

This particular defence response is especially well suited to the early fruitfly (Drosophila) embryo, in which the cell nuclei undergo a series of 13 rapid divisions within a common cytoplasm4. These swift nuclear divisions occur synchronously, and consist entirely of alternating phases of DNA synthesis (S phase) and DNA segregation (mitosis or M phase), with none of the intervening gap phases that separate S and M in more typical cell cycles. Because there is no gap phase, if one of the embryonic nuclei is damaged during S phase, it does not have the usual option of stopping the division cycle before mitosis. Instead, it will be driven into mitosis, in synchrony with the surrounding undamaged nuclei in the common cytoplasm.

During nuclear-division cycles 9 and 10, most nuclei migrate to the outer rim of the embryo — the cortex (Fig. 1a). A few remain in the interior, but these will not contribute to the adult fly. It has been known for some time that, during cycles 10 to 13, any cortical nucleus that suffers DNA damage eventually drops into the interior of the embryo and is thereby effectively eliminated from the organism5. It has been a mystery, however, how such damaged nuclei are recognized and how they are then discharged into the interior.

Figure 1: Dealing with DNA damage in Drosophila.
figure 1

a, An early Drosophila embryo that has undergone 10–13 nuclear divisions. Most nuclei (green) are aligned around the outside, with only a few in the interior. These internal nuclei — 'yolk nuclei' — will not contribute to adult tissues. b–d, The fate of a nucleus whose DNA has been damaged in interphase (the period between one nuclear division and the next). b, At the time of DNA damage the centrosome is unaffected, and microtubules (black) are still organized by γ-tubulin ring complexes (γ-TURCs; orange) that are concentrated around a core centrosomal structure. c, As all the undamaged nuclei enter mitosis, so too does the damaged nucleus, but the γ-TURCs appear to be released from the centrosomes and the mitotic spindle (black) does not form properly. (Note that, at this point, the nuclear membrane has disintegrated, allowing the spindle to attach to chromosomes.) Chromosome segregation fails and, in the following interphase (d), the defective nucleus falls into the interior of the embryo, leaving the centrosomes behind. Theurkauf and colleagues3 have found that the Chk2 protein is crucial for this nuclear defence mechanism.

Three years ago, Theurkauf and colleagues6 described the phenomenon of centrosome inactivation in Drosophila embryos. Centrosomes are structures that are needed to efficiently segregate DNA during mitosis. They contain so-called γ-tubulin ring complexes (γ-TURCs), which organize the long filaments, or microtubules, that make up the mitotic spindle — a bipolar apparatus on which chromosomes are segregated. Theurkauf and co-workers6 noticed that nuclei that failed to complete DNA synthesis or suffered DNA damage during S phase formed abnormal spindles when they entered mitosis. This was apparently because the DNA damage triggered the displacement of the γ-TURCs from the centrosomes during mitosis. Intriguingly, the γ-TURCs reappeared at centrosomes after mitosis was complete, and several other centrosomal proteins remained concentrated at centrosomes during mitosis, hinting that a core centrosome structure remained intact throughout the division cycle. The abnormal nuclei that reformed after the aberrant mitosis then rapidly dropped into the interior of the embryo, and so were effectively eliminated (Fig. 1b–d).

Theurkauf and colleagues3 have now shown that double-stranded DNA breaks are responsible for triggering this centrosome inactivation, and that Chk2 is essential for the process. In vertebrate cells, many Chk2-dependent responses to DNA damage are induced via the activation of the p53 protein, but the authors found that this is not the case for centrosome inactivation. Moreover, they discovered that Chk2 itself becomes concentrated at centrosomes, and that DNA damage seems to enhance its accumulation there. So, given that Chk2 is a kinase — it modifies proteins by phosphorylating them — perhaps it inactivates centrosomes by directly phosphorylating one or more of their protein components.

Might Chk2 also induce centrosome inactivation in other cell types? As mentioned above, in most cells, DNA damage that occurs in the phases between mitoses (these phases are collectively known as interphase) causes a cell-cycle arrest prior to mitosis1,2. But this cannot occur in early Drosophila embryos, so a unique Chk2-dependent mechanism may have evolved to eliminate defective nuclei during mitosis instead. In support of this possibility, DNA damage in older, cellularized fruitfly embryos (when the nuclei no longer share a common cytoplasm, and there is a gap phase between S and M phases) does not appear to lead to centrosome inactivation, but instead causes a delay in both the entry into and exit from mitosis7,8. And in vertebrate cells in culture, DNA damage during mitosis does not induce centrosome inactivation9 —although, even in early Drosophila embryos, DNA damage that occurs during mitosis does not appear to trigger centrosome inactivation, implying that Chk2 may need to be activated in interphase to cause centrosome inactivation in mitosis3,6.

It seems unlikely, then, that centrosome inactivation is a major response to DNA damage in most normal cells, as DNA damage during interphase will usually lead to cell-cycle arrest before the cell enters mitosis. If, however, a cell manages to enter mitosis with unrepaired DNA damage, it might then become important to trigger centrosome inactivation in order to eliminate the defective DNA. In fact, there is some evidence that vertebrate cells that enter mitosis carrying damaged DNA die by a 'mitotic catastrophe'10, although it is not clear whether this mechanism requires Chk2 or is indeed caused by centrosome inactivation. If such a process does exist, however, it might help to protect against cancer. The mechanisms that monitor DNA damage are often impaired in pre-cancerous cells, and so it may be relatively common for such cells to enter mitosis carrying defective DNA. If centrosome inactivation proves to be more than just a specialization of flies, the race will be on to understand how Chk2 brings it about, and to test whether it is involved in preventing human cancer.


  1. Dasika, G. K. et al. Oncogene 18, 7883–7899 (1999).

    CAS  Article  Google Scholar 

  2. Taylor, W. R. & Stark, G. R. Oncogene 20, 1803–1815 (2001).

    CAS  Article  Google Scholar 

  3. Takada, S., Kelkar, A. & Theurkauf, W. E. Cell 113, 87–99 (2003).

    CAS  Article  Google Scholar 

  4. Foe, V. E. & Alberts, B. M. J. Cell Sci. 61, 31–70 (1983).

    CAS  PubMed  Google Scholar 

  5. Sullivan, W., Daily, D. R., Fogarty, P., Yook, K. J. & Pimpinelli, S. Mol. Biol. Cell 4, 885–896 (1993).

    CAS  Article  Google Scholar 

  6. Sibon, O. C., Kelkar, A., Lemstra, W. & Theurkauf, W. E. Nature Cell Biol. 2, 90–95 (2000).

    CAS  Article  Google Scholar 

  7. Su, T. T., Walker, J. & Stumpff, J. Curr. Biol. 10, 119–126 (2000).

    CAS  Article  Google Scholar 

  8. Su, T. T. & Jaklevic, B. Curr. Biol. 11, 8–17 (2001).

    CAS  Article  Google Scholar 

  9. Mikhailov, A., Cole, R. W. & Rieder, C. L. Curr. Biol. 12, 1797–1806 (2002).

    CAS  Article  Google Scholar 

  10. Roninson, I. B., Broude, E. V. & Chang, B. D. Drug Resist. Updates 4, 303–313 (2001).

    CAS  Article  Google Scholar 

Download references

Author information

Authors and Affiliations


Corresponding author

Correspondence to Jordan W. Raff.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Raff, J. Genome maintenance. Nature 423, 493–495 (2003).

Download citation

  • Issue Date:

  • DOI:

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

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