The DNA Deamination Model of Somatic Antibody Diversification

Classical work had established that Ab diversity in mice and humans arises at two distinct developmental stages. First, a primary Ab repertoire is set up in developing B cells through an ordered program of Ab V region gene rearrangements by V(D)J recombination, including sequential rearrangements during so-called receptor editing. This repertoire is subsequently modified in Ag-driven Ab responses through somatic hypermutation (SHM) and class switch recombination (CSR), resulting in a secondary Ab repertoire adapted to environmental needs and persisting as immunological memory. A third mechanism of somatic Ab diversification, operating in other species at certain developmental stages, is gene conversion, in which repertoire diversification is achieved through the conversion of recipient Ab V region genes by upstream pseudogenes.

Despite extensive work during several decades, the molecular mechanisms of SHM, CSR, and gene conversion had long remained a mystery, although all three processes had been studied in much detail at the level of DNA sequence and the stages of B cell differentiation at which they operated. In the case of CSR, environmental signals inducing switches to particular Ab isotypes had been identified (reviewed in Ref. 1). From the early 1990s, with the availability of gene targeting approaches in mice, the search was on for the direct identification of molecules involved in the control of SHM and CSR. Furthermore, in the chicken, where gene conversion diversifies Ab V region genes, a B cell line called DT40 became available at around the same time in which genes could be efficiently targeted owing to a high frequency of homologous recombination between genomic genes and artificially introduced DNA (2). In the case of SHM, much of the thinking in the field was guided by the early Brenner and Milstein (3) model of Ab diversification through error-prone DNA break repair. Because DNA repair and error-prone DNA polymerases were the subjects of a rapidly expanding field of research, extensive work was done to study the impact of the ablation of genes involved in (error-prone) repair on the processes of secondary Ab diversification. This went hand in hand with the identification of cis-regulatory genetic elements in the Ab gene loci that control SHM and CSR.

Leaders in this area with regard to SHM were César Milstein and Michael Neuberger, who worked closely together at the Medical Research Council Laboratory of Molecular Biology in Cambridge, U.K. Over the years, they and their collaborators made major contributions to defining the intrinsic features of SHM and gained extensive knowledge of enhancer and promoter elements controlling SHM. With Cristina Rada, and in parallel with others, they had also studied the role of DNA mismatch repair enzymes in the control of SHM (reviewed in Ref. 4) and interpreted changes of the mutational signature observed in MutS homolog 2 knockout mice as an indication that SHM may proceed in two phases, the initial one preferentially targeting C:G base pairs, often in the context of C:G-centered mutational hot spots (5). They also suggested a mechanistic similarity between SHM and CSR owing to a similar effect that MutS homolog 2 deficiency has on the two processes in terms of efficiency and the targeting of DNA sequence–defined hot spots of somatic mutation or recombination breakpoints, respectively (6). This notion was surprising at the time, because SHM and CSR had mostly been considered entirely distinct processes.

Like a bombshell, in 2000 came the discovery by Tasuku Honjo, Alain Fischer, and Ann Durandy that activation-induced (cytidine) deaminase (AID), identified as a germinal center B cell–specific gene by Honjo and colleagues (7), is critical for both SHM and CSR in mice and humans (8, 9). That this is also true for Ab diversification through gene conversion was subsequently shown by Jean-Marie Buerstedde’s and Neuberger’s groups (10, 11). AID is thus central for the control of all secondary Ab diversification. Although the homology between AID and the RNA editing enzyme APOBEC1 raised the possibility of RNA editing being involved in the control of these processes (7−9), Neuberger’s mind was immediately drawn, on the basis of his and Milstein’s earlier results, to the possibility that AID might directly deaminate cytidine in DNA to uracil. This crystallized in a DNA deamination model for SHM, CSR, and gene conversion being famously put forward, and to an initial test in the Petersen-Mahrt et al. (12) Nature article reprinted here in the Pillars of Immunology series. (Note that Matthew Scharff and colleagues independently proposed a DNA deamination–based model of somatic Ab diversification; see Refs. 13, 14.) In the Petersen-Mahrt et al. article, the model was tested in Escherichia coli by induced expression of AID. This yielded a mutator phenotype: AID targeted C:G base pairs in a context-dependent manner, and its mutator activity was enhanced by uracil DNA glycosylase deficiency, exactly as predicted by the model. Two additional articles from Neuberger and colleagues (15, 16), published in short sequence thereafter in the same year, provided additional support for their model of C:G-targeted SHM, both in the DT40 chicken B cell line and in mice. Evidence from other laboratories suggested that in mice and humans, following detection of AID-generated lesions in DNA (17−20), nucleotide misincorporation by DNA polymerase η is the major contributor to SHM at A:T pairs in vivo (21, 22).

Since this time of dramatic progress, the DNA deamination model of secondary Ab diversification has guided much of the experimental work in the field and has received ample support from a wealth of biochemical and other data, which are beyond the scope of this brief introduction. More importantly, DNA deamination by AID and other members of the APOBEC family has emerged as a mechanism operating beyond the Ab system, mediating intracellular defense against retroviruses such as HIV (23−25), as well as mutagenic processes associated with tumorigenesis (26, 27). The Petersen-Mahrt et al. article has thus played a founding role in a broad new area of biomedical research.