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.

Molecular biology

Ensuring error-free DNA repair

Damaged DNA must be removed with the utmost precision, as mistakes are costly. The structure of a repair enzyme bound to its substrate provides a welcome clue to how this is achieved.

Certain forms of oxygen, known as reactive oxygen species, can be deleterious for living organisms. Although most cells can tolerate and even exploit them, these oxygen forms may contribute to cancer, tissue degeneration and ageing. In the nucleus of cells, active oxygen damages DNA, so it is important to repair an altered DNA base before it is copied during DNA replication. Failing this, the newly incorporated base opposite the original lesion may need to be removed and corrected.

One example of such repair kicks in after the conversion of guanine (G) — one of the four bases — to 8-hydroxyguanine (also known as 8-oxoguanine; oxoG) by active oxygen1 and misincorporation of the normal base adenine (A) opposite the oxoG. A specific enzyme, called MutY in bacteria and MYH in higher organisms, breaks the link between such a misplaced adenine and the DNA sugar–phosphate backbone to initiate repair2. But the enzyme leaves well alone when oxoG is correctly paired with the base cytosine (C). On page 652, Fromme et al.3 describe an ingenious way by which this specificity is achieved (see Fig. 1 of their paper for a scheme of the repair cycle).

The catalytic domain of the MutY protein had already been crystallized, so the three-dimensional structure of the protein was largely known4. But attempts failed to co-crystallize MutY and a short piece of double-stranded DNA containing an oxoG residue and its mismatched partner, adenine, apparently because the complex was unstable.

Fromme et al. have overcome this technical problem by trapping the protein and the DNA together by means of a covalent bond, a crosslink. The strategy used was to generate a short stretch of DNA that had a base containing a sulphur (thiol) residue located near the oxoG. This allowed a disulphide bond to form between the oligonucleotide and MutY. Ideas for where to position the thiol residue came from information gleaned from preliminary structural data and the conformation of the protein4. Fromme et al. show that MutY interacts at several sites within the A·oxoG pair. But it cannot ‘flip out’ the oxoG residue from the double-stranded DNA helix. Instead, the unmodified adenine is flipped out and excised by MutY.

MutY belongs to a group of enzymes known as DNA glycosylases, which recognize altered bases in DNA and help to remove them. Like other DNA glycosylases, it generates a sharp bend in the DNA at the site of the mismatch. The new structural data provide a suitable explanation for why — as is desired — MutY doesn't recognize and remove an adenine opposite its normal base partner, thymine (T): the extensive and precise contacts between MutY and an A·oxoG pair are entirely absent in a normal A·T pair. Similarly, the enzyme's active site does not accommodate a cytosine opposite an oxoG; for coding reasons, it is important that the oxidized base rather than the normal base is repaired in this partnership.

In humans, different forms (polymorphisms) of the MYH gene have been detected that result in the production of enzymes with a reduced ability to specifically and efficiently recognize these rare A·oxoG pairs. Given that reactive oxygen species are cancer-causing, and that mutations in MYH are risk factors for colorectal cancer5,6, the results of Fromme et al.3 will help in providing a detailed molecular picture of the consequences of such mutations6,7. Better relative risk estimates for the development of colorectal cancer associated with a malfunctioning MYH enzyme should also gradually become available for defects that affect different sites in the protein.


  1. Kasai, H. & Nishimura, S. Nucleic Acids Res. 12, 2137–2145 (1984).

    CAS  Article  Google Scholar 

  2. Michaels, M. L., Cruz, C., Grollman, A. P. & Miller, J. H. Proc. Natl Acad. Sci. USA 89, 7022–7025 (1992).

    ADS  CAS  Article  Google Scholar 

  3. Fromme, J. C. et al. Nature 427, 652–656 (2004).

    ADS  CAS  Article  Google Scholar 

  4. Guan, Y. et al. Nature Struct. Biol. 5, 1058–1064 (1998).

    CAS  Article  Google Scholar 

  5. Al-Tassan, N. et al. Nature Genet. 30, 227–232 (2002).

    CAS  Article  Google Scholar 

  6. Sieber, O. M. et al. N. Engl. J. Med. 348, 791–799 (2003).

    Article  Google Scholar 

  7. Chmiel, N. H., Livingston, A. L. & David, S. S. J. Mol. Biol. 327, 431–443 (2003).

    CAS  Article  Google Scholar 

Download references

Author information

Authors and Affiliations


Rights and permissions

Reprints and Permissions

About this article

Cite this article

Lindahl, T. Ensuring error-free DNA repair. Nature 427, 598 (2004).

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