Reply
There are many indications that the mutations at A·T (deoxyadenosine·deoxythymidine) pairs during somatic hypermutation (SHM) are introduced by mutagenic patch repair of activation-induced deaminase (AID)-generated U·G (deoxyuridine· deoxyguanosine) lesions. Franklin and Blanden rightly state that a popular model is that these mutations are caused by DNA-polymerase-catalysed incorporation of improperly paired bases during the patch-repair process (that is, polymerase error), whereas we have recently postulated that DNA-polymerase-catalysed incorporation of U residues, instead of T residues, opposite A residues might provide an alternative mechanism (that is, dUTP (deoxyuridine triphosphate) incorporation)1.
Franklin and Blanden propose that neither the polymerase-error model nor the dUTP-incorporation model can explain the pattern of nucleotide substitutions that is generated during SHM. We, however, do not see that either model is incompatible with published data, particularly considering that neither model is tightly defined.
Regarding polymerase error, we1 noted that reconciling a polymerase-error model with SHM spectra requires the postulated error-prone mechanism to be extremely biased, making errors with much higher frequency at A·T pairs than at G·C (G·deoxycytidine) pairs. With DNA polymerase-η already having been shown to result in biased mutation spectra in in vitro assays, it is clearly not inconceivable that DNA polymerase-η (or another DNA polymerase) might operate in phase 2 of SHM in vivo, showing an even more marked mutational bias2,3,4,5,6,7. So, we do not think that an analysis of SHM spectra allows one to jettison the polymerase-error model.
We were moved to propose the dUTP-incorporation model because the mechanism of phase 2 of SHM is unknown, and we think that dUTP incorporation is an attractive alternative to polymerase error that had not been put forward previously and that deserves serious consideration. However, we also think that current data are not sufficient to conclude or exclude the possibility that dUTP incorporation makes a considerable contribution to phase 2 of SHM.
Franklin and Blanden propose that the SHM spectra in UNG (uracil-DNA glycosylase)-deficient mice (in which transversions at C·G pairs form a very small proportion of the total mutations8) imply that phase 2 of SHM must create some mutations at C·G pairs (albeit far less than at A·T pairs) and that this is incompatible with the A·T-specific hypermutation process that is envisaged in the dUTP-incorporation model. However, it seems possible that the mutations at A·T pairs in UNG-deficient mice could result from dUTP incorporation, with the small number of transversions at C·G pairs being produced by replication over abasic sites that are generated by a low-frequency excision of the U residues in the U·G lesions by an UNG-back-up glycosylase, such as SMUG1 (single-strand selective monofunctional UNG).
Any model of phase 2 of SHM must explain the strand polarity of mutations at A·T pairs. Strand polarity in the SHM spectra could presumably be a consequence of strand discrimination during mutagenic patch DNA synthesis or during a possible subsequent repair stage. Franklin and Blanden consider a scheme in which the polarity is ascribed to a repair process that has the strand bias of conventional transcription-coupled repair: they conclude that this would generate a polarity that is the inverse of what is observed and that dUTP incorporation cannot therefore make a considerable contribution to phase 2 of SHM. However, this is only one of many hypotheses that can be put forward to explain the strand bias in A·T mutations in a dUTP-incorporation model: it cannot form a sufficient basis to generally exclude a marked contribution of dUTP incorporation to phase 2 of SHM.
So, we think that it remains entirely open whether either polymerase error or dUTP incorporation makes an important contribution to SHM at A·T pairs. We hope that further experimentation will help to shed light on the mechanism that is actually used in vivo.
References
Neuberger, M. S . et al. Somatic hypermutation at A·T pairs: polymerase error verus dUTP incorporation. Nature Rev. Immunol. 5, 171–178 (2005).
Zeng, X . et al. DNA polymerase η is an A–T mutator in somatic hypermutation of immunoglobulin variable genes. Nature Immunol. 2, 537–541 (2001).
Delbos, F., De Smet, A, Faili, A, Aoufouchi, S, Weill, J. C. & Reynaud, C. A. Contribution of DNA polymerase η to immunoglobulin gene hypermutation in the mouse. J. Exp. Med. 201, 1191–1196 (2005).
Martomo, S. A . et al. Different mutation signatures in DNA polymerase η- and MSH6-deficient mice suggest separate roles in antibody diversification. Proc. Natl Acad. Sci. USA 102, 8656–8661 (2005).
Matsuda, T, Bebenek, K, Masutani C., Hanaoka, F. & Kunkel, T. A. Low fidelity DNA synthesis by human DNA polymerase-η. Low fidelity DNA synthesis by human DNA polymerase-η. Nature 404, 1011–1013 (2000).
Matsuda, T . et al. Error rate and specificity of human and murine DNA polymerase-η. J. Mol. Biol. 312, 335–346 (2001).
Rogozin, I. B, Pavlov, Y. I, Bebenek, K, Matsuda, T. & Kunkel, T. A. Somatic mutation hotspots correlate with DNA polymerase η error spectrum. Nature Immunol. 2, 530–536 (2001).
Rada, C . et al. Immunoglobulin isotype switching is inhibited and somatic hypermutation perturbed in UNG-deficient mice. Curr. Biol. 12, 1748–1755 (2002).
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Neuberger, M., Di Noia, J., Beale, R. et al. The mechanism of somatic hypermutation at A·T pairs remains an open question. Nat Rev Immunol 5, 180 (2005). https://doi.org/10.1038/nri1553-c2
Issue Date:
DOI: https://doi.org/10.1038/nri1553-c2
