Somatic hypermutation (SHM) introduces point mutations into immunoglobulin variable-region genes in germinal-center B cells and is the engine of antibody affinity maturation. How these mutations are produced mechanistically was contested for decades between a DNA-based model, centered on cytidine deamination of DNA by activation-induced cytidine deaminase (AID) followed by error-prone repair, and a reverse-transcriptase (RT) model, in which mutation information passes through RNA before being written back into immunoglobulin genes. The initiating mechanism of mutations at G:C pairs - AID-mediated deamination of cytosine to uracil in DNA - is now established and is conceded even by the RT model's proponents. What remains unresolved is the surviving core of the 40-year-old dispute: whether mutations at A:T pairs, and the characteristic strand-biased mutation signatures observed at immunoglobulin and non-immunoglobulin loci, arise exclusively from DNA-templated error-prone repair synthesis by DNA polymerase eta, or whether ADAR-edited immunoglobulin transcripts are copied back into DNA by polymerase eta acting as a reverse transcriptase through RNA-directed DNA repair. The only direct in vivo test of reverse-transcription involvement (2001, retroviral RT inhibitors) predates the modern polymerase-eta-based form of the model and its interpretation is disputed, while independent biochemistry has since confirmed that repair polymerases can reverse-transcribe RNA in cells.

Background

After infection or vaccination, B lymphocytes carrying rearranged immunoglobulin (Ig) variable-region genes - assembled in earlier development from V, D and J gene segments - proliferate in germinal centers, specialized microanatomical structures within lymphoid follicles. There, point mutations accumulate in the Ig variable regions at roughly a million times the spontaneous genomic rate, and B cells whose receptors bind antigen more tightly are selectively expanded. This two-step process of mutation plus selection, called affinity maturation, is the basis of improved antibody responses; the mutational process itself is somatic hypermutation (SHM).

The consensus mechanism of SHM, consolidated between about 2002 and 2007, is DNA-based. The enzyme activation-induced cytidine deaminase (AID) converts deoxycytidine to deoxyuridine in single-stranded DNA that becomes exposed when the Ig gene is transcribed. The resulting uracil-guanine mismatch is then processed along two mutagenic routes. In one, the repair enzyme uracil-DNA glycosylase (UNG) excises the uracil, leaving an abasic site that is bypassed by error-prone polymerases, producing most mutations at G:C pairs. In the other, the mismatch-recognition proteins MSH2 and MSH6 initiate excision and resynthesis in which DNA polymerase eta (Pol eta), a low-fidelity member of the Y-family of translesion synthesis polymerases (polymerases that copy damaged or undamaged DNA at the cost of fidelity), misincorporates bases and thereby generates most mutations at A:T pairs, with a preference for WA motifs (W denotes A or T). The evidence for this framework is direct: mice and humans lacking AID have no SHM or class-switch recombination (the AID-dependent DNA rearrangement that switches an antibody's constant-region isotype); AID expression alone mutates bacterial DNA; AID-dependent uracils have been physically detected in the DNA of Ig variable and switch regions; and humans or mice lacking Pol eta lose most A:T mutations while retaining G:C mutations.

A minority program, initiated by Steele and Pollard in 1987, has long advocated an alternative: the reverse transcriptase (RT) model, in which the mutations are generated through an error-prone DNA-to-RNA-to-DNA information loop - an Ig transcript is copied by a cellular reverse transcriptase (an RNA-dependent DNA polymerase) and the error-containing complementary DNA is recombined back into the chromosomal gene. In its modern form this model no longer denies that AID deaminates DNA: it accepts AID-generated uracils in DNA as the initiating lesion, but proposes that mutations at A:T pairs arise when ADAR enzymes ('adenosine deaminases acting on RNA', which convert adenosine to inosine in double-stranded RNA and, at lower efficiency, in RNA:DNA hybrids) edit the nascent Ig transcript, and the edited RNA is then copied back into the gene by Pol eta operating in reverse-transcriptase mode via RNA-directed DNA repair (the use of an RNA molecule as the template for repairing the DNA gene it was transcribed from, a process demonstrated in yeast and, for Pol theta, in mammalian cells). A key empirical claim of this program is that SHM shows strand-biased signatures - mutations at A sites exceed those at T sites by about 2.9-fold and mutations at G sites exceed those at C sites by about 1.7-fold - which the group argues are expected consequences of the RT route and inconsistent with pure DNA-based repair; they attribute conflicting published analyses to polymerase-chain-reaction recombination artifacts. The same group extends the model to cancer, interpreting strand-asymmetric single-base-substitution signatures in tumor genomes as off-target 'Ig-like' deaminase-plus-reverse-transcription mutagenesis.

The question is genuinely unresolved at its core because the two candidate routes for the A:T phase use the same enzyme: Pol eta is the mutator in both the DNA-templated repair model and the RT model, so deleting Pol eta cannot discriminate them. The only direct in vivo test of reverse-transcription involvement, performed in 2001 with the retroviral reverse-transcriptase inhibitors AZT and ddC, found SHM unchanged, but the RT camp argues those inhibitors were not designed to block a Pol-eta-type activity and the mainstream interpretation of that experiment is disputed rather than settled by a repeat. Meanwhile, biochemistry independent of the advocacy group has established that human Pol eta can extend primers on RNA templates and exhibits reverse-transcriptase activity in cellular environments, and that another repair polymerase, Pol theta, promotes RNA-templated DNA repair in mammalian cells - making the proposed route physically plausible even though it has never been demonstrated at Ig loci. A further complication is that ADAR1 is required for germinal-center B-cell survival through suppression of double-stranded-RNA sensing (involving the innate-immune receptor MDA5), so simple ADAR knockout collapses the germinal-center reaction before SHM can be measured; the editing-specific question therefore requires editing-defective but RNA-binding-competent knock-in alleles, which exist but have not been used to measure SHM spectra.

Problem Statement

Determine the molecular mechanism by which mutations at A:T base pairs are produced during immunoglobulin somatic hypermutation in germinal-center B cells, and correspondingly the origin of the strand-biased mutation signatures observed at immunoglobulin and non-immunoglobulin loci. Specifically: do these A:T mutations arise exclusively from DNA-templated error-prone repair synthesis by DNA polymerase eta following AID-initiated lesions, or does an RNA-mediated route contribute, in which ADAR-edited immunoglobulin transcripts are copied back into the gene by polymerase eta acting as a reverse transcriptase through RNA-directed DNA repair? Equivalently, in the generality posed by the source literature: does reverse information flow (RNA to DNA) participate in writing somatic mutations into immunoglobulin genes, or is the mutational information path entirely DNA-templated? Since the initiating cytidine-deamination step is now common ground between the two models, the unresolved parts are: (i) which route operates in authentic germinal-center somatic hypermutation; (ii) how the operating route or routes account quantitatively for the observed spectra, including the A:T versus G:C distribution, WA-motif targeting, and the reported strand biases (mutations at A sites exceeding those at T sites by about 2.9-fold, and mutations at G sites exceeding those at C sites by about 1.7-fold); and (iii) whether the analogous strand-biased and codon-context signatures at non-immunoglobulin (cancer) loci do or do not require the proposed deaminase-plus-reverse-transcription process. The RNA route could contribute a fraction of A:T mutations rather than all of them, so resolving the two routes requires distinguishing a partial contribution from a complete absence.

References

  1. Steele EJ, Pollard JW (1987). Hypothesis: somatic hypermutation by gene conversion via the error prone DNA-to-RNA-to-DNA information loop. Molecular Immunology 24(6):667-673. doi:10.1016/0161-5890(87)90049-6. https://doi.org/10.1016/0161-5890(87)90049-6
  2. Muramatsu M, Kinoshita K, Fagarasan S, Yamada S, Shinkai Y, Honjo T (2000). Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell 102(5):553-563. doi:10.1016/S0092-8674(00)00078-7. https://doi.org/10.1016/S0092-8674(00)00078-7
  3. Sack SZ, Bardwell PD, Scharff MD (2001). Testing the reverse transcriptase model of somatic mutation. Molecular Immunology 38(4):303-311. doi:10.1016/S0161-5890(01)00058-X. https://doi.org/10.1016/S0161-5890(01)00058-X
  4. Petersen-Mahrt SK, Harris RS, Neuberger MS (2002). AID mutates E. coli suggesting a DNA deamination mechanism for antibody diversification. Nature 418(6893):99-104. doi:10.1038/nature00862. https://doi.org/10.1038/nature00862
  5. Imai K, Slupphaug G, Lee WI, Revy P, Nonoyama S, Catalan N, Yel L, Forveille M, Kavli B, Krokan HE, Ochs HD, Fischer A, Durandy A (2003). Human uracil-DNA glycosylase deficiency associated with profoundly impaired immunoglobulin class-switch recombination. Nature Immunology 4(10):1023-1028. doi:10.1038/ni974. https://doi.org/10.1038/ni974
  6. Franklin A, Milburn PJ, Blanden RV, Steele EJ (2004). Human DNA polymerase-eta, an A-T mutator in somatic hypermutation of rearranged immunoglobulin genes, is a reverse transcriptase. Immunology and Cell Biology 82(2):219-225. doi:10.1046/j.0818-9641.2004.01221.x. https://doi.org/10.1046/j.0818-9641.2004.01221.x
  7. Wilson TM, Vaisman A, Martomo SA, Sullivan P, Lan L, Hanaoka F, Yasui A, Woodgate R, Gearhart PJ (2005). MSH2-MSH6 stimulates DNA polymerase eta, suggesting a role for A:T mutations in antibody genes. Journal of Experimental Medicine 201(4):637-645. doi:10.1084/jem.20042066. https://doi.org/10.1084/jem.20042066
  8. Delbos F, Aoufouchi S, Faili A, Weill JC, Reynaud CA (2007). DNA polymerase eta is the sole contributor of A/T modifications during immunoglobulin gene hypermutation in the mouse. Journal of Experimental Medicine 204(1):17-23. doi:10.1084/jem.20062131. https://doi.org/10.1084/jem.20062131
  9. Neuberger MS, Rada C (2007). Somatic hypermutation: activation-induced deaminase for C/G followed by polymerase eta for A/T. Journal of Experimental Medicine 204(1):7-10. doi:10.1084/jem.20062409. https://doi.org/10.1084/jem.20062409
  10. Di Noia JM, Neuberger MS (2007). Molecular mechanisms of antibody somatic hypermutation. Annual Review of Biochemistry 76:1-22. doi:10.1146/annurev.biochem.76.061705.090740. https://doi.org/10.1146/annurev.biochem.76.061705.090740
  11. Basu U, Meng FL, Keim C, Grinstein V, Pefanis E, Eccleston J, Zhang T, Myers D, Wasserman CR, Wesemann DR, Januszyk K, Gregory RI, Deng H, Lima CD, Alt FW (2011). The RNA exosome targets the AID cytidine deaminase to both strands of transcribed duplex DNA substrates. Cell 144(3):353-363. doi:10.1016/j.cell.2011.01.001. https://doi.org/10.1016/j.cell.2011.01.001
  12. Maul RW, Saribasak H, Martomo SA, McClure RL, Yang W, Vaisman A, Gramlich HS, Schatz DG, Woodgate R, Wilson DM 3rd, Gearhart PJ (2011). Uracil residues dependent on the deaminase AID in immunoglobulin gene variable and switch regions. Nature Immunology 12(1):70-76. doi:10.1038/ni.1970. https://doi.org/10.1038/ni.1970
  13. Zhao Y, Gregory MT, Biertumpfel C, Hua YJ, Hanaoka F, Yang W (2013). Mechanism of somatic hypermutation at the WA motif by human DNA polymerase eta. Proceedings of the National Academy of Sciences USA 110(20):8146-8151. doi:10.1073/pnas.1303126110. https://doi.org/10.1073/pnas.1303126110
  14. Steele EJ (2016). Somatic hypermutation in immunity and cancer: critical analysis of strand-biased and codon-context mutation signatures. DNA Repair 45:1-24. doi:10.1016/j.dnarep.2016.07.001. https://doi.org/10.1016/j.dnarep.2016.07.001
  15. Steele EJ (2017). Reverse transcriptase mechanism of somatic hypermutation: 60 years of clonal selection theory. Frontiers in Immunology 8:1611. doi:10.3389/fimmu.2017.01611. https://doi.org/10.3389/fimmu.2017.01611
  16. Steele EJ, Lindley RA (2017). ADAR deaminase A-to-I editing of DNA and RNA moieties of RNA:DNA hybrids has implications for the mechanism of Ig somatic hypermutation. DNA Repair 55:1-6. doi:10.1016/j.dnarep.2017.04.004. https://doi.org/10.1016/j.dnarep.2017.04.004
  17. Methot SP, Di Noia JM (2017). Molecular mechanisms of somatic hypermutation and class switch recombination. Advances in Immunology 133:37-87. doi:10.1016/bs.ai.2016.11.002. https://doi.org/10.1016/bs.ai.2016.11.002
  18. Su Y, Egli M, Guengerich FP (2017). Human DNA polymerase eta accommodates RNA for strand extension. Journal of Biological Chemistry 292(44):18044-18051. doi:10.1074/jbc.M117.809723. https://doi.org/10.1074/jbc.M117.809723
  19. Su Y, Ghodke PP, Egli M, Li L, Wang Y, Guengerich FP (2019). Human DNA polymerase eta has reverse transcriptase activity in cellular environments. Journal of Biological Chemistry 294(15):6073-6081. doi:10.1074/jbc.RA119.007925. https://doi.org/10.1074/jbc.RA119.007925
  20. Meers C, Keskin H, Banyai G, Mazina O, Yang T, Gombolay AL, Mukherjee K, Kaparos EI, Newnam G, Mazin A, Storici F (2020). Genetic characterization of three distinct mechanisms supporting RNA-driven DNA repair and modification reveals major role of DNA polymerase ζ. Molecular Cell 79(6):1037-1050.e5. doi:10.1016/j.molcel.2020.08.011. https://doi.org/10.1016/j.molcel.2020.08.011
  21. Chandramouly G, Zhao J, McDevitt S, Rusanov T, Hoang T, Borisonnik N, Treddinick T, Lopezcolorado FW, Kent T, Siddique LA, Mallon J, Huhn J, Shoda Z, Kashkina E, Brambati A, Stark JM, Chen XS, Pomerantz RT (2021). Polθ reverse transcribes RNA and promotes RNA-templated DNA repair. Science Advances 7(24):eabf1771. doi:10.1126/sciadv.abf1771. https://doi.org/10.1126/sciadv.abf1771
  22. Franklin A, Steele EJ (2022). RNA-directed DNA repair and antibody somatic hypermutation. Trends in Genetics 38(5):426-436. doi:10.1016/j.tig.2021.10.005. https://doi.org/10.1016/j.tig.2021.10.005
  23. Li Y, Ruan GX, Chen W, Huang H, Zhang R, Wang J, Li Y, Xu S, Ou X (2022). RNA-editing enzyme ADAR1 p150 isoform is critical for germinal center B cell response. Journal of Immunology 209(6):1071-1082. doi:10.4049/jimmunol.2200149. https://doi.org/10.4049/jimmunol.2200149
  24. Yu K (2022). AID function in somatic hypermutation and class switch recombination. Acta Biochimica et Biophysica Sinica 54(6):759-766. doi:10.3724/abbs.2022070. https://doi.org/10.3724/abbs.2022070
  25. Miglierina E, Ordanoska D, Le Noir S, Laffleur B (2024). RNA processing mechanisms contribute to genome organization and stability in B cells. Oncogene 43(9):615-623. doi:10.1038/s41388-024-02952-2. https://doi.org/10.1038/s41388-024-02952-2
  26. Hao Q, Li J, Yeap LS (2024). Molecular mechanisms of DNA lesion and repair during antibody somatic hypermutation. Science China Life Sciences 67(11):2344-2353. doi:10.1007/s11427-024-2615-1. https://doi.org/10.1007/s11427-024-2615-1
  27. Lauring MC, Basu U (2024). Somatic hypermutation mechanisms during lymphomagenesis and transformation. Current Opinion in Genetics and Development 85:102165. doi:10.1016/j.gde.2024.102165. https://doi.org/10.1016/j.gde.2024.102165
  28. Steele EJ, Franklin A, Lindley RA (2024). Somatic mutation patterns at Ig and non-Ig loci. DNA Repair 133:103607. doi:10.1016/j.dnarep.2023.103607. https://doi.org/10.1016/j.dnarep.2023.103607
  29. Steele EJ, Lindley RA (2025). Deaminase-driven reverse transcription mutagenesis in oncogenesis: critical analysis of transcriptional strand asymmetries of single base substitution signatures. International Journal of Molecular Sciences 26(3):989. doi:10.3390/ijms26030989. https://doi.org/10.3390/ijms26030989
  30. Mariani MGE (2026). RNA-templated diversity generation via reverse transcription and homologous recombination: an evolutionary hypothesis. Theory in Biosciences 145(2):18. doi:10.1007/s12064-026-00469-8. https://doi.org/10.1007/s12064-026-00469-8
  31. Lindley RA, Gorczynski RM, Steele EJ (2026). Epigenetic-genetic coupling and understanding the molecular and cellular basis of Lamarckian inheritance. International Journal of Molecular Sciences 27(4):2003. doi:10.3390/ijms27042003. https://doi.org/10.3390/ijms27042003