News & Views | Published:

DNA replication

SOS polymerases

Naturevolume 409pages3334 (2001) | Download Citation


The standard DNA-replicating machinery cannot copy damaged DNA, so SOS enzymes come to the rescue. It now seems, at least in some bacteria, that different enzymes are required for different types of damage.

Imagine that an injured person is carried into a hospital's emergency department. The doctor needs to assess the situation quickly, and to decide on the most appropriate treatment — stitches, a drug or maybe an operation? A similar problem occurs during DNA replication in living cells, when the replication machinery — consisting of enzymes called DNA polymerases — encounters a damaged DNA site. Standard DNA polymerases cannot continue to copy DNA once they reach a lesion. But every species has several polymerases that can do so, in a process called translesion synthesis1. Why are there so many of these 'translesional' polymerases? And how does a cell decide which polymerase to use? Writing in EMBO Journal and EMBO Reports2,3, Fuchs and colleagues reveal some of the answers for the bacterium Escherichia coli.

There are five DNA polymerases in E. coli. Three of these — DNA polymerases II, IV and V — are induced as part of the cellular SOS response to DNA damage, and are involved in translesion synthesis. It was already known that DNA polymerase V (pol V)4,5 allows DNA replication to continue over bulky damage such as abasic lesions (when a base is lost from one strand of DNA) or errors produced by ultraviolet light6. Pol V is error-prone, and often leaves mutations in its wake within the newly synthesized strand opposite the damaged region. However, at certain types of lesion its activity is free from errors. It can also copy undamaged DNA; again, this is error-prone, although the mutations are 'untargeted' — they do not occur at a specific point in the DNA.

Pol IV, by contrast, is involved in getting replication started again after it stalls, and in inducing untargeted mutations when copying phage λ (ref. 7) — a small, circular DNA molecule (or plasmid) in E. coli that is replicated independently of the main chromosome. In fact, overexpression of pol IV results in a roughly 1,000-fold increase in mutations in plasmids8. Such mutations are likely to be useful in enabling a population to evolve quickly in response to environmental stress. Like pol V, pol IV is a bona fide DNA polymerase9. Little is known about the cellular function of E. coli pol II, however, even though it was discovered as long ago as 1970.

This is where Napolitano et al.2 come in. They have looked at the frequency of mutations in various strains of E. coli in response to different types of chemically induced damage. They confirm that pol V is capable of error-free, as well as error-prone (mutagenic), translesion synthesis. But the interesting thing is that so are pol II and pol IV, and that the choice of polymerase depends on the nature of the DNA injury and the sequence in which it is found.

For example, the chemical N-2-acetylaminofluorene binds to the fourth nucleotide in the NarI DNA sequence (GGCGCC). It increases the rate of frameshift mutations here — in effect, the middle two bases of this sequence are not copied during DNA replication. Napolitano et al. show that pol II replicates DNA to which this chemical is bound, and induces the frameshift. Pol V, however, can replicate the GGCGCC sequence correctly, without introducing mutations (Fig. 1).

Figure 1: Translesional polymerases continue to replicate DNA even when it contains errors that make other DNA polymerases stall.
Figure 1

There are several translesional polymerases in the bacterium Escherichia coli, and Napolitano et al.2 show that the choice of polymerase depends on the nature of the DNA injury. One type of injury is that induced by the chemical N-2-acetylaminofluorene (AAF), which forms a bond with the fourth nucleotide of the sequence GGCGCC. a, DNA polymerase V is needed to copy this DNA in an error-free way. Incorporation of a cytosine residue (C) opposite the lesion by pol V generates a key replication intermediate. b, The template DNA and the copy in progress can slip relative to each other (producing a slipped intermediate), meaning that two of the bases in the template are not copied (a '− 2 frameshift'). The presence of the lesion thermodynamically favours the formation of a slipped intermediate (as shown by the occurrence of isomerization). DNA polymerase II is required in this pathway.

Similarly, both pol IV and pol V are required for error-free and mutagenic translesion synthesis when benzo(a)pyrene — a major cancer-causing component of cigarette smoke and car exhaust fumes — is used. This chemical joins to the third guanosine of the sequence GGG. When N-2-acetylaminofluorene binds to a nucleotide in the same stretch of DNA, only pol V is required for replication, and it induces a frameshift mutation. Here, perhaps the structural distortion of the DNA template by two different lesions in the same sequence requires a different combination of polymerases to achieve translesion synthesis. In any case, it seems that E. coli chooses from a pool of these polymerases to bypass different DNA lesions.

How is the relevant translesional polymerase targeted to a DNA lesion? Last year, Tang et al.10 showed that the SOS protein RecA activates pol V while directing it to a damaged DNA site. Wagner et al.3 now have biochemical data that provide an answer for pol IV. This E. coli enzyme is strictly distributive in vitro — it can copy only one base each time it binds to the DNA template. The authors could not detect a stable in vitro complex containing pol IV and DNA. The implication is that a co-factor helps pol IV to be targeted to its DNA substrate.

Wagner et al.3 proposed that the β- subunit of pol III — the main enzyme that replicates undamaged DNA — might be such a targeting factor. This subunit acts as a clamp to connect pol III to DNA. It now seems that this β-subunit does the same for pol IV, too3. When Wagner et al. added this clamp, together with the γ-subunit of pol III as a clamp 'loader', to an in vitro DNA-replication reaction, pol IV and DNA formed a stable complex. Pol IV could then incorporate 300–400 nucleotides per DNA-binding event. This value contrasts with that of 6–7 nucleotides obtained previously for pol IV in the presence of the β- and γ-subunits10, but it is not clear why these authors obtained such different results. In any event, it seems that pol IV and pol V require the β/γ complex of pol III and the RecA protein, respectively, to be recruited to their substrate DNA. Pol V then also requires the β/γ complex to carry out translesion synthesis.

Translesion synthesis is less accurate than normal DNA replication, but organisms probably chose it over the more dangerous state of leaving an unreplicated portion of DNA in a single-stranded form. However, translesion synthesis might have evolutionary benefits. As predicted more than a quarter of a century ago11, the error-prone polymerases allow individual cells to mutate when their survival is threatened, increasing the genetic diversity and adaptability of an imperilled population.

On the other hand, one would not expect DNA polymerase η in humans to be as error-prone as its bacterial counterparts. This polymerase is important in preventing skin cancer — people suffering from xeroderma pigmentosum have a predisposition to skin cancers, and the gene mutated in these people is that encoding DNA polymerase η12,13,14. What is the difference between this enzyme and the E. coli translesional polymerases? To find out, we may have to wait until the structures of these enzymes in complex with their substrates have been determined. In the meantime, the work of Fuchs and colleagues2,3, as well as that of Tang et al .10, provides much-needed information about how cells cope with DNA damage.


  1. 1

    Friedberg, E. C., Feaver, W. J. & Gerlach, V. L. Proc. Natl Acad. Sci. USA 97, 5681–5683 (2000).

  2. 2

    Napolitano, R., Janel-Bintz, R., Wagner, J. & Fuchs, R. P. P. EMBO J. 19, 6259–6265 (2000).

  3. 3

    Wagner, J., Fujii, S., Gruz, P., Nohmi, T. & Fuchs, R. P. P. EMBO Rep. 1, 484 –488 (2000).

  4. 4

    Tang, M. et al. Proc. Natl Acad. Sci. USA 96, 8919– 8924 (1999).

  5. 5

    Reuven, N. B., Arad, G., Maor-Shoshani, A. & Livneh, Z. J. Biol. Chem. 274, 31763–31766 (1999).

  6. 6

    Friedberg, E. C., Walker, G. C. & Siede, W. DNA Repair and Mutagenesis (ASM, Washington DC, 1995).

  7. 7

    Brotcorne-Lannoye, A. & Maenhaut-Michel, G. Proc. Natl Acad. Sci. USA 83, 3904–3908 (1986).

  8. 8

    Kim, S.-R et al. Proc. Natl Acad. Sci. USA 94, 13792– 13797 (1997).

  9. 9

    Wagner, J. et al. Mol. Cell 4, 281–287 (1999).

  10. 10

    Tang, M. et al. Nature 404, 1014–1018 (2000).

  11. 11

    Radman, M. in Molecular and Environmental Aspects of Mutagenesis (eds Sherman, F., Miller, M., Lawrence, C. & Tabor, W. H.) 128–142 (Thomas, Springfield, IL, 1974).

  12. 12

    Masutani, C. et al. EMBO J. 18, 3491–3501 (1999).

  13. 13

    Masutani, C. et al. Nature 399, 700–704 (1999).

  14. 14

    Johnson, R. E., Kondratick, C. M., Prakash, S. & Prakash, L. Science 285, 263–265 ( 1999).

Download references

Author information


  1. Institute for Molecular and Cellular Biology, Osaka University, 1-3 Yamada-oka, Suita, 565-0871, Osaka, Japan

    • Fumio Hanaoka


  1. Search for Fumio Hanaoka in:

Corresponding author

Correspondence to Fumio Hanaoka.

About this article

Publication history

Issue Date



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