Alkyladenine DNA glycosylase (Aag) in somatic hypermutation and class switch recombination
Introduction
Somatic hypermutation (SHM) in mammals is a process of secondary diversification of immunoglobulin (Ig) genes in activated B cells during which point mutations, and occasional insertions and deletions, are introduced into DNA encoding the antibody variable (V) regions. SHM thereby alters the affinity of antibodies for cognate antigens, a hallmark of adaptive immunity. SHM requires AID, the most likely function of which is to deaminate cytosines in the promoter proximal region DNA of Ig genes (reviewed in [1]). Deamination of cytosine (C) in DNA creates a uracil (U) base mispaired with guanine (G) that, when present in DNA, is a substrate for DNA repair. During SHM, some of the DNA repair systems that would normally faithfully repair such U:G mismatches paradoxically appear to be co-opted to generate mutations. Proteins of two major systems co-opted for SHM are the uracil glycosylase Ung, and the mismatch repair proteins Msh2, Msh6 (which comprise the Msh2/6 heterodimer), Mlh1, Pms2 (reviewed in [1], [2]) and Mlh3 [3], [4].
Mice and humans deficient for Ung are proficient for SHM, but the pattern of SHM is altered characteristically in that mutations at C and G are mainly transitions, as if uracils left unrepaired serve as a template for DNA replication [5], [6]. As a glycosylase, the function of Ung is to remove U from the DNA backbone, leaving an abasic (AP) site that is processed by concerted action of an AP endonuclease (APE1), and a deoxyribophosphodiesterase (dRPase activity of polymerase (pol) β) to produce a single-strand gap [2]. Error-free filling-in of the gap is accomplished by pol β or the high-fidelity polymerase, δ [7]. However, to explain the SHM pattern shift due to Ung-deficiency, AID/Ung-mediated single-strand gaps are likely to be filled-in using one or more translesion polymerases to generate mutations from all four nucleotides. Of the many recently identified translesion polymerases, particularly polymerase η, ι and θ have been implicated in SHM (reviewed by [8]).
Importantly, Ung-deficiency affects the SHM pattern at C (and G), but has less effect on mutations at A (and T). Mutations at A are at most 52% reduced in Ung-deficient mice [9], presumably, because Ung-dependent mutations at A can arise during long-patch base excision repair (BER) in which polymerases β or δ are replaced by translesion polymerases. In the absence of Ung, these mutations in part seem to depend on functional mismatch recognition by Msh2/6, because, although the pattern of mutations at A is not altered, their frequency is decreased from ∼50% in wildtype mice to between 26% and as little as 2% of total mutations in both Msh2−/− and Msh6−/− mice (reviewed in [1], [10]). This decrease in frequency suggests that recognition of the U:G mismatch by MutSα (the Msh2/6 heterodimer) initiates a process that ultimately leads to mutations at adenines, perhaps via recruitment of the error-prone pol η [11], [12]. Polη deficiency in mice and humans also leads to a SHM pattern characterized by a low frequency of mutations at A [12], [13], [14], [15], [16].
Currently available data suggest that mutations at A can be explained by the activities of Ung and mismatch repair, and errors created by translesion DNA polymerases. However, the precise mechanism by which these mutations arise are not yet known and possibly there are unknown factors that are involved in DNA repair during SHM that could influence mutations at A. Like mutations at C, transition mutations at A are predominant in SHM. A:T → G:C transitions could result from deamination of A to hypoxanthine (Hx), which codes as a G during DNA synthesis [17], [18]. Adenosine deaminases exist, although currently are known to act only on RNA [19]. Yet, by analogy to AID, we questioned whether a DNA adenosine deaminase might be involved in SHM. The major DNA repair enzyme for Hx is alkyladenine DNA glycosylase (Aag) [20]. In order to determine whether adenosine deamination plays a role during SHM we examined the SHM pattern in Aag−/− mice. We report here that, while the mutational pattern at A is not changed in the Aag-deficient animal, activation of B cells leads to a significant induction of Hx glycosylase activity in wild type mice. Furthermore, we see a small yet significant bias towards T:A → C:G transition mutations in the absence of Aag leading us to suggest that Aag glycosylase activity plays a role during SHM.
Section snippets
Aag−/− mice
The generation of the Aag−/− mice was previously described [20]. Aag−/− animals were backcrossed to a pure C57Bl/6J background (at least 12 backcrosses) and were 6 to 8 months old when received at the University of Chicago. Mice were analyzed soon after arrival, due to quarantine space constraints and institutional animal shipment regulations. Mice were genotyped as described [20]; genomic template DNA for PCR was derived from either PNA-low, sorted B cells from Peyer's patches, or from kidney.
Ung−/− mice
Expression of Aag in mutating and activated murine B cells
Aag is ubiquitously expressed in all tissues (NCBI Entrez Gene, GeneID: 26839), albeit at very different levels. However, in postulating a role for Aag in SHM that occurs specifically in a small subset of activated B cells during an immune response, expression of Aag was analyzed in germinal center activated B cells of immunized wildtype C57BL/6 mice. By RT-PCR, Aag is expressed in PNA-high, GL7+ (germinal center) B cells as well as other B220+ B cells, CD3+ cells (mainly T lymphocytes) and
Discussion
The findings in Aag−/− mice challenge the idea that an adenine DNA deaminase exists that operates in parallel with AID during somatic hypermutation. Aag possesses unusually broad substrate specificity for damaged DNA bases, in particular for alkylated purines, although Hx is a preferred substrate at least in vitro [32]. Moreover, Aag is reported to be the major mammalian glycosylase that functions to remove Hx from DNA in liver, testes, kidney and lung [20]. So, if deaminated A (Hx) is present
Acknowledgements
We are grateful to D. Nicolae for statistical analysis and G. Bozek for excellent technical assistance. We thank Drs. D. Barnes and T. Lindahl for Ung−/− mice. Research was funded by grants AI047380 (NIH-NIAID) and AI053130 (NIH-NIAID) to US, and P30-ES02109 (NIH/NIEHS) and 7-RO1-CA75576 (NIH/NCI) to LDS. SL was supported by an American Association of University Women International Ph.D. Fellowship. LDS is an American Cancer Society Research Professor.
References (42)
- et al.
AID in somatic hypermutation and class switch recombination
Curr. Opin. Immunol.
(2006) - et al.
A role for Mlh3 in somatic hypermutation
DNA Repair (Amst.)
(2006) - et al.
Immunoglobulin isotype switching is inhibited and somatic hypermutation perturbed in UNG-deficient mice
Curr. Biol.
(2002) - et al.
Site-specific mutagenesis in vivo by single methylated or deaminated purine bases
Mutat. Res.
(1986) - et al.
A-to-I RNA editing: recent news and residual mysteries
J. Biol. Chem.
(2003) - et al.
Uracil-DNA glycosylase (UNG)-deficient mice reveal a primary role of the enzyme during DNA replication
Mol. Cell.
(2000) - et al.
Specific expression of activation-induced cytidine deaminase (AID), a novel member of the RNA-editing deaminase family in germinal center B cells
J. Biol. Chem.
(1999) - et al.
The zab domain of the human RNA editing enzyme ADAR1 recognizes Z-DNA when surrounded by B-DNA
J. Biol. Chem.
(2000) - et al.
Non-B DNA structure-induced genetic instability
Mutat. Res.
(2006) - et al.
Dissecting the broad substrate specificity of human 3-methyladenine-DNA glycosylase
J. Biol. Chem.
(2004)
Endonuclease V protects Escherichia coli against specific mutations caused by nitrous acid
Mutat. Res.
Functions of human DNA polymerases eta, kappa, and iota
DNA Rep.
Base excision repair and nucleotide excision repair contribute to the removal of N-methylpurines from active genes
DNA Rep. (Amst.)
DNA Repair and Mutagenesis
A role for the MutL mismatch repair Mlh3 protein in immunoglobulin class switch DNA recombination and somatic hypermutation
J. Immunol.
Human uracil-DNA glycosylase deficiency associated with profoundly impaired immunoglobulin class-switch recombination
Nat. Immunol.
Second pathway for completion of human DNA base excision-repair: reconstitution with purified proteins and requirement for DNase IV (FEN1)
EMBO J.
DNA polymerases and somatic hypermutation of immunoglobulin genes
EMBO Rep.
Somatic hypermutation and class switch recombination in Msh6−/−Ung−/− double-knockout mice
J. Immunol.
Examination of Msh6- and Msh3-deficient mice in class switching reveals overlapping distinct roles of MutS homologues in antibody diversification
J. Exp. Med.
Somatic hypermutation at A-T pairs: polymerase error versus dUTP incorporation
Nat. Rev. Immunol.
Cited by (8)
RNA-directed DNA repair and antibody somatic hypermutation
2022, Trends in GeneticsCitation Excerpt :By analogy, although several glycosylases can excise U lesions from DNA, the activity of UNG is favoured at AID-induced U lesions [71]. Although it is conceivable that the generation of mutations at A/T pairs could somehow result entirely from the deamination of A residues in DNA (and the direct removal of resultant noncanonical Hx lesions from the template strand by Hx-DNA glycosylase activity to create AP lesions) [70], it is unclear to us why such a mechanism should depend on polymerase η. The mechanism by which polymerase η facilitates the generation of substitutions almost exclusively at A/T pairs during SHM remains unknown (see Outstanding questions).
ADAR deaminase A-to-I editing of DNA and RNA moieties of RNA:DNA hybrids has implications for the mechanism of Ig somatic hypermutation
2017, DNA RepairCitation Excerpt :If significant numbers of mutations at A were due to the direct action of an adenosine deaminase on the ssDNA of the displaced non-transcribed strand at Transcription Bubbles as proposed and then followed by alkyladenine DNA glycosylase (Aag) repair, we would find a preponderance of A:T > G:C transition mutations during SHM in an Aag deleted background i.e. A-to-G mutations exceeding reverse complement T-to-C mutations. This was not observed, and the frequencies of SHM and CSR were not significantly altered in Aag−/− mice [3]. However there was one exception which has direct relevance to the analysis of the ADAR-mediated A-to-I editing of RNA:DNA hybrids studied by Zheng et al. (2017).
Somatic hypermutation in immunity and cancer: Critical analysis of strand-biased and codon-context mutation signatures
2016, DNA RepairCitation Excerpt :In one series of experiments a version of a proposed DNA-based model that actually predicts A>>T strand bias was tested. It was hypothesized that direct DNA deamination of deoxyadenosines in ssDNA would preferentially result in A-to-G like lesions on the Top (NTS) strand at the Transcription Bubble viz. adenosine-to-hypoxanthine conversion by alkyladenine DNA glycosylase [79]. In that study it is unclear what to make of the statement “… statistically significant, albeit low increase of T:A > C:G transition mutations in Aag (−/−) animals...”.
Activation-induced cytidine deaminase structure and functions: A species comparative view
2011, Developmental and Comparative ImmunologyCitation Excerpt :Neuberger and collaborators have introduced the idea of the incorporation of dUTPs due to imbalances in the pool of dNTPs, which would explain some features of the mutational spectrum (Neuberger et al., 2005), but the A/T bias was not specifically addressed and the model has not been validated by recent experiments (Sharbeen et al., 2010). Deamination of adenines followed by the creation of an abasic site by alkyladenine DNA glycosylase (Aag) has also been proposed to contribute to the transitions found at adenines, but this model also does not address the strand asymmetry and in mice the impact of the Aag deficiency in the frequency and pattern of SHM is none or minor, respectively (Longerich et al., 2007). A complex model involving a reverse transcriptase and RNA editing activities that lead to strand asymmetry via an RNA intermediate from a mutated strand does address the strand bias, but there is no conclusive experimental data to support it (Steele and Pollard, 1987).
Reflections on the state of play in somatic hypermutation
2008, Molecular Immunology