The Double-Edged Sword of Activation-Induced Cytidine Deaminase
http://www.100md.com
免疫学杂志 2005年第2期
Abstract
Activation-induced cytidine deaminase (AID) is required for Ig class switch recombination, a process that introduces DNA double-strand breaks in B cells. We show in this study that AID associates with the DNA-dependent protein kinase catalytic subunit (DNA-PKcs) promoting cell survival, presumably by resolving DNA double-strand breaks. Wild-type cells expressing AID mutants that fail to associate with DNA-PKcs or cells deficient in DNA-PKcs or 53BP1 expressing wild-type AID accumulate H2AX foci, indicative of heightened DNA damage response. Thus, AID has two independent functions. AID catalyzes cytidine deamination that originates DNA double-strand breaks needed for recombination, and it promotes DNA damage response and cell survival. Our results thus resolve the paradox of how B cells undergoing DNA cytidine deamination and recombination exhibit heightened survival and suggest a mechanism for hyperIgM type II syndrome associated with AID mutants deficient in DNA-PKcs binding.
Introduction
Activation-induced cytidine deaminase (AID)3 promotes somatic hypermutation (SHM) and class switch recombination (CSR) of Ig genes (1). AID cytidine deaminase activity was first proposed on the basis of its homology with the apolipoprotein B mRNA-editing catalytic polypeptide 1 (2). The cytidine deaminase property of AID led to two distinct hypotheses to explain the diversification of Ig genes. The first, the RNA-editing hypothesis, proposes that AID, like apolipoprotein B mRNA-editing catalytic polypeptide 1, modifies unknown RNA precursors that, in turn, originate endonucleases that cleave the DNA encoding the Ig genes. The finding by Begum et al. (3), showing that AID-dependent DNA cleavage in CSR requires de novo protein synthesis, is in agreement with this hypothesis. The second, the DNA deamination hypothesis, proposes that AID deaminates cytidine to uracil directly in the DNA encoding the Ig genes (4). Compatible with the second hypothesis are the results reported by Petersen-Marth et al. (4), showing that expression of AID in Escherichia coli originates a mutator phenotype that yields nucleotide transitions at dC/dG, and the findings of Dickerson et al. (5), Pham et al. (6), and Chaudhuri et al. (7), showing that AID deaminates DNA substrates in vitro.
Exactly how AID introduces point mutations or executes class switch recombination is not yet understood, but it is generally thought that cytidine deamination of DNA or RNA somehow generates double-strand breaks in Ig DNA (1, 8, 9, 10, 11). Ordinarily, cells respond to DNA double-strand breaks by undergoing cell cycle arrest to allow time for repair (12) and respond to persistent damage by inducing apoptosis, presumably as a protection against illegitimate recombination (13). However, B cells undergoing Ig class switch do not die, presumably because they efficiently repair DNA double-strand breaks. Although RAD54, RAD52, and RAD51 repair proteins are needed for AID-induced Ig gene conversion in chicken cell lines (14, 15), whether AID directly recruits repair factors to the locales of cytidine deamination is not known.
Materials and Methods
Generation of AID wild-type (AID-wt) and mutant constructs
RNA was obtained from C57BL/6 mouse lymph nodes using TRIzol reagent. AID cDNA was produced by RT using oligo(dT) primer and amplified by PCR using Turbo pfu polymerase (Stratagene) and primer set wu160/wu167. Full-length AID cDNA was cloned in-frame into pUHD10S vector downstream of Flag tag sequences. AID deletion mutants were generated by PCR using Turbo pfu. The primer sets used to generate the deletion mutants were: mutant F1, wu160/wu166; mutant F3, wu160/wu165; mutant F4, wu162/wu167; mutant F6, wu161/wu167; and mutant C, wu160/wu174. We used the QuikChange mutagenesis kit (Stratagene) to generate an AID dominant negative mutant and AID-R112H point mutation constructs with primer sets wu155/wu156 and wu157/wu158, respectively. PCR fragments were flanked by 5' NheI and 3' XbaI sites to allow subsequent cloning into the pUHD10S vector. The Flag-tagged AID fragment (EcoRI/XbaI) was subcloned into the pCI expression vector (Promega) for transient expression and into the pCI-neo expression vector (Promega) for stable transfections. For expression in 70Z/3 cells and splenocytes, Flag-AID or Flag-AID-C fragments were cloned into pIRES2GFP vector (BD Clontech) upstream of the internal ribosomal entry site sequence, followed by subcloning of the Flag-AID-(or Flag-AID-C)-IRES2-EGFP cassette into the pMSCV-puro retroviral vector (BD Clontech). To produce the GST-AID fusion protein, the full-length AID cDNA was cloned in-frame into the pGEX4T1 vector (Amersham Biosciences). All sequences were verified by DNA sequencing.
Cell culture, transfection, and retroviral transduction
Wild-type mouse embryonic fibroblasts (MEF) were generated from C57BL/6 mouse embryos at 14.5 days postcoitum and maintained in DMEM supplemented with 10% FCS, 1.0 IU/ml penicillin G, and 0.5 IU/ml streptomycin. HeLa cells and HEK293 cells were cultured in DMEM and supplemented with 10% FCS, 1.0 IU/ml penicillin G, and 0.5 IU/ml streptomycin. DNA-dependent protein kinase catalytic subunit null (DNA-PKcs–/–) MEF cells (PK33N) (16) were provided by Dr. D. J. Chen (Lawrence Berkeley National Laboratory) and maintained in -MEM supplemented with 10% FCS, 1.0 IU/ml penicillin G, and 0.5 IU/ml streptomycin. Mouse pre-B lines 70Z/3 and 18.81 cells were maintained in RPMI 1640 medium supplemented with 10% FCS, 1.0 IU/ml penicillin G, and 0.5 IU/ml streptomycin. LPS cultures were prepared by incubating B cells (95% purity, isolated with a MACS column; 2 x 105/well/100 μl) with 10 μg/ml LPS (Sigma-Aldrich) in RPMI 1640 medium supplemented with 10% FCS, 1.0 IU/ml penicillin G, and 0.5 IU/ml streptomycin. B cells were obtained from C57BL/6 mice as previously described (17).
Transient transfection of AID-wt or mutant construct into HeLa, HEK293, or MEF cells was performed with Lipofectamine (Invitrogen). The expression of AID-wt in 70Z/3 cells and splenocytes was performed by retroviral transduction. Linearized pMSCV-IRES2-EGFP-puro or pMSCV-Flag-AID-IRES2-EGFP-puro DNA was stably transfected into RetroPack PT67 (BD Clontech) packaging cells by electroporation, whereas pMSCV-Flag-AID-C-IRES2-EGFP-puro was transiently transfected into the same packaging line using Lipofectamine. The virus-containing supernatant of each kind (in DMEM) was collected every 24 h and additionally concentrated by centrifugation at 6000 x g for 4 h. Virus pellets were resuspended in a 1/50 volume of complete RPMI 1640. LPS cultures were transduced with proviruses at the time of seeding in 100 μl of virus-containing supernatant supplemented with 10 μg/ml LPS. AID- and AID-C-expressing splenocytes were analyzed with a FACSCalibur (BD Biosciences) by measuring GFP positivity and propidium iodide stain for cell death.
Subcellular fractionation
Nuclei and cytosol fractions were prepared according to published protocols (18). Briefly, 107 HEK293 cells expressing Flag-AID were harvested and washed three times with PBS. Cells were resuspended in 1 ml of lysis buffer (10 mM HEPES (pH 7.5), 10 mM KCl, 1.5 mM MgCl2, 1 mM DTT, and protease inhibitor mixture (Roche)) on ice for 20 min, followed by 20 gentle strokes for homogenization with a loose-fit Dounce homogenizer (Kontes). The homogenate was overlaid on top of 200 μl of 40% sucrose in lysis buffer and centrifuged at 800 x g for 15 min in a centrifuge with a swinging bucket rotor. The supernatant (corresponding to the cytosolic fraction) and the pellet (corresponding to the nuclear fraction) were collected and extracted with 10x lysis buffer (19).
Immunoprecipitation and Western blotting
Immunoprecipitation was performed as described previously (19). To identify AID-binding proteins, 2 x 108 HeLa cells expressing Flag-AID were used for a large-scale immunoprecipitation with EZview anti-Flag (M2) beads (Sigma-Aldrich). Otherwise, routine immunoprecipitation was conducted using 2 x 107 cells expressing Flag-AID. In each experiment, one-fourth of the precipitated proteins (equivalent to 5 x 106 cells) were resolved on a 6% SDS-PAGE (for DNA-PKcs and Ku80) or 12% SDS-PAGE (for AID). In the experiments testing DNA dependence of AID/DNA-PKcs binding, one-fourth of precipitated beads were incubated with 100 μl of PBS alone or 10 mM 3,3'-dithiobis-(sulfosuccinimidylpropionate) (STDDP) in PBS for 2 h on ice to cross-link protein complexes. After quenching the cross-linking reaction with 10 μl of 1.0 M Tris-HCl (pH 7.4) for 15 min and washed twice with PBS, the beads were then treated with 20 U of DNase I (as indicated in Fig. 3B) in 40 μl of PBS at room temperature for 30 min, followed by two washes before SDS-PAGE analysis under reducing condition (5% 2-ME) to cleave cross-linked complexes. Mouse anti-Flag (M2) mAb was purchased from Sigma-Aldrich. Rabbit anti-DNA-PKcs (specific to human DNA-PKcs, SC-9051) and anti-Ku80 (SC-5280) mAb were purchased from Santa Cruz Biotechnology. Anti-DNA-PKcs mAb (specific to mouse DNA-PKcs; NA57) was obtained from Oncogene, goat anti-GST Ab is a product of Amersham Biosciences (27-4577-01), anti--tubulin mAb was obtained from Santa Cruz Biotechnology (SC-5274), and rabbit anti-AID serum was a gift from Dr. F. W. Alt (Harvard Medical School, Boston, MA).
FIGURE 3. AID deamination and C-terminal domains are required for binding to DNA-PKcs. A, Schematic domain representation of full-length AID and deletion mutants. DNA-PKcs binding of wild-type or deletion mutants is noted on the right. Wild-type AID and mutants F1 and F6 associate with DNA-PKcs; mutants F1, F4, and C do not. B, Deamination and C-terminal domains of AID are necessary for binding to DNA-PKcs. Shown is a Western blot (WB) analysis of anti-Flag immunoprecipitates (IP) obtained from lysates of HEK293 cells transiently transfected with wt or deletion mutants (F1, F3, F4, F6, and C) of AID. Blots were probed with either anti-Flag or anti-DNA-PKcs Abs as indicated. C, Cytidine deamination-defective AID mutants do not bind to DNA-PKcs. Shown is a Western blot (WB) analysis of anti-Flag immunoprecipitates obtained from lysates of HEK293 cells stably expressing Flag-tagged AID-wt, dominant negative H56R/E58Q mutant (DN), or R112H mutant AID. Blots were probed with either anti-Flag or anti-DNA-PKcs Ab as indicated. The figures are representative of three independent experiments each.
DNA-PKcs in vitro pulldown by GST-AID
Recombinant GST and GST-AID fusion protein were purified from bacterial DL21 cells. Five hundred nanograms of GST or GST-AID fusion protein beads were incubated with 0.5 ml of cell extracts obtained from 2 x 107 HEK293 cells overnight at 4°C in the presence or the absence of 2 μg of ssDNA, 2 μg of dsDNA, or 20 U of DNase I. The ssDNA was a 59-base oligonucleotide containing RGYW repeats (AGCTGGCAGGCTAGCAAGTTGGTT-GGCAAGCAGGTAAGCAGG CAAGCTGGCTGAATTCC) (7). The dsDNA was an EcoRV-linearized pBluescript KS vector (Stratagene). Beads were washed and analyzed as described above.
Immunofluorescence staining and microscopy
Immunofluorescence staining and confocal microscopy were performed essentially as described previously (19). For microtubule staining in HeLa cells expressing Flag-AID, cells were fixed with methanol at –20°C for 10 min, air-dried, rehydrated, repermeabilized with 0.05% Triton X-100 in PBS for 3 min, and blocked for 90 min with blocking buffer (5% normal goat serum, 1% glycerol, 0.1% BSA, 0.1% fish skin gelatin, and 0.04% sodium azide), followed by staining with anti--tubulin Ab (Sigma-Aldrich; T5293). Cytoplasmic protein extraction was performed by incubating HeLa cells expressing GFP or Flag-AID with 50 μg/ml digitonin in PBS on ice for 5 min, followed by four washes with PBS and fixation with 4% paraformaldehyde. Affinity-purified rabbit anti-H2AX was provided by Dr. J. Chen (Mayo Clinic, Rochester, MN). Cell death of MEF was determined by TUNEL assay (Promega) and by the presence of condensed chromatin or fragmented nuclei in 4',6-diamido-2-phenylindole hydrochloride (DAPI) staining (20).
Identification of AID-associated proteins
The proteins coimmunoprecipitated with Flag-AID from HeLa cells were resolved on a SDS-PAGE (4–15% polyacrylamide gradient; Bio-Rad) and stained with Coomassie G-250 (Bio-Rad). Protein bands were excised and analyzed using MALDI-TOF mass spectrometry by the Rockefeller University Protein Resource Center (New York, NY).
Oligonucleotides used
The following oligonucleotides were used: Wu155, CTTCGCAACAAGTCTGGCTGCCGCGTGCAATTGTTGTTCCTACGCTACATC; Wu156, GATGTAGCGTAGGAACAACAATTGCACGCGGCAGCCAGACTTGTTGCGAAG; Wu157, CAGCCTGAGGATTTTCACCGCGCACCTCTACTTCTGTGAAGACCGC; Wu158, GCGGTCTTCACAGAAGTAGAGGTGCGCGGTGAAAATCCTCAGGCT; Wu160, GAATCAGCTAGCGACAGCCTTCTGATGAAGCAAAAG; Wu161, GAATCAGCTAGCGGCTGCCACGTGGAATTGTTGTTC; Wu162, GAATCAGCTAGCGAGGGGCTGCGGAGACTGCACC; Wu165, GAATCATCTAGATTAAGGCTCAGCCTTGCGGTCTTCAC; Wu166, GAATCATCTAGATTAATTTTCTACAAATGTATTCCAGCAG; Wu167, GAATCATCTAGATTAAAATCCCAACATACGAAATGCATC; and Wu174, GAATCATCTAGATTAGTCATCGACTTCGTACAAGGGCAAAAGG.
Results
AID binds to DNA-PKcs in the nucleus
We considered the possibility that association of AID with cofactors could promote survival of cells undergoing Ig gene diversification. To determine whether AID associates with other molecules, we analyzed AID protein complexes obtained from HeLa cells expressing AID. AID was immunoprecipitated from lysates, and the identity of any coprecipitates was sought by mass spectrometry. Fig. 1A shows a Coomassie-stained gel image showing proteins coimmunoprecipitated with AID. MALDI-TOF mass spectrometry identified the largest coprecipitated protein as DNA-PKcs with a Mr of 486 kDa based on 34 matched peptides. Other proteins identified were heat shock 70-kDa protein 8 isoform 1 (HSC70; 12 matched peptides), -tubulin (15 matched peptides), and a protein similar to mitochondrial solute carrier family 25 (four matched peptides; Fig. 1).
FIGURE 1. Identification of AID-binding proteins. Coomassie Blue staining of 4–15% gradient SDS-PAGE analysis of anti-Flag immunoprecipitates of lysates obtained from HeLa cells transiently transfected with Flag-AID or empty vector (mock) was performed. Flag-AID comigrates with the Ab L chain (25KD). AID-binding proteins were identified by MALTI-TOF mass spectrometry. Protein identities and accession numbers are indicated.
Because DNA-PKcs is required to efficiently resolve by nonhomologous end-joining the DNA double-strand breaks associated with class switch recombination (21), we questioned whether AID associated with DNA-PKcs in B-lineage cells and cells other than HeLa. Fig. 2A shows DNA-PKcs coimmunoprecipitated with AID and vice versa in extracts from HeLa cells, human embryonic kidney 293 (HEK293) cells, and murine B cells (70Z/3) transfected with Flag-tagged AID. We also show in Fig. 2B that endogenously expressed AID in 18.81 B cells (22) binds to DNA-PKcs (left panel). Our results indicate that the association of AID with DNA-PKcs may reflect the function of AID in B cells.
FIGURE 2. AID associates with DNA-PKcs of B and non-B cells. A, AID coprecipitates with DNA-PKcs in various AID-transfected cell lines. Shown are Western blot (WB) analyses of anti-Flag or anti-DNA-PKcs immunoprecipitates (IP) of cell lysates obtained from HeLa, HEK293, or 70Z/3 cells transfected with Flag-AID (+) or empty vector (–). Blots were probed with either anti-Flag (Flag) or anti-DNA-PKcs (DNA-PKcs) Abs as indicated. Mouse Ig L chain comigrating with 25-kDa Flag-AID is seen as a weak band in the controls. This figure is representative of three independent experiments. B, Endogenous AID associates with DNA-PKcs. Shown is a Western blot (WB) analysis (left panel) of anti-DNA-PKcs immunoprecipitates (IP) of cell lysates from mouse pre-B cell line 18.81 that constitutively expresses AID or from mouse pre-B cell line 70Z/3 that does not express endogenous AID. Blots were probed with either anti-DNA-PKcs Abs or rabbit polyclonal anti-AID (7 ) as indicated. The expression of endogenous AID in 18.81 and 70Z lines was verified by RT-PCR, as shown in the right panel. C, AID and DNA-PKcs associate in the nucleus. Shown is a Western blot analysis of Flag-AID or DNA-PKcs immunoprecipitates from whole cell extracts, cytoplasmic fraction, or nuclear fraction obtained from HEK293 cells expressing AID. The blots were probed with either anti-DNA-PKcs or anti-Flag Abs as indicated.
AID is mainly localized to the cytoplasm (23), whereas DNA-PKcs localizes predominantly in the nucleus (24), yet they associate with one another. We asked whether the association of DNA-PKcs and AID took place in the nucleus, where presumably AID and DNA-PKcs function. Fig. 2C shows that DNA-PKcs coprecipitates with AID in the nuclear fraction, but not in the cytoplasmic fraction of cell extracts (Fig. 2C). Our results indicate that the association of DNA-PKcs and AID takes place in the nucleus, representing the small fraction of total cellular AID (Fig. 2C).
Association of AID and DNA-PKcs requires the C-terminal and deamination domains of AID
Because DNA-PKcs is required for the repair of DNA double-strand breaks during class switch recombination (21), but is dispensable for somatic hypermutation (25), our findings showing that AID associates with DNA-PKcs predict that the complexes mediating class switch recombination and somatic hypermutation are distinct. Recent findings from two laboratories support this idea (26, 27). Barreto et al. (26) found that an AID mutant that lacks the C-terminal 10 aa retained cytidine deaminase activity, but failed to promote class switch recombination. Ta et al. (27) found that some subjects with type II hyperIgM syndrome have mutations in AID causing truncation or disruption of the C-terminal domain. These individuals have a severe defect in class switch, but normal somatic hypermutation (27). These findings suggested that the C-terminal region of AID is necessary for class switch recombination in B cells (26, 27).
We questioned whether the selective defect in class switch recombination of AID C terminus deletion mutants reflected defective association with DNA-PKcs. To test this idea, we generated a series of AID deletion mutants and tested the mutants for association with DNA-PKcs. C-terminal truncation mutants lost (mutant F1 (positions 1–154) and C (positions 1–189)) or markedly reduced (mutant F3 (positions 1–123)) their ability to bind to DNA-PKcs (Fig. 3, A and B), suggesting that the C-terminal domain of AID is necessary for the formation of AID/DNA-PKcs complexes. In contrast, mutant F6 (positions 54–198) lacking the N-terminal 53-aa segment of AID retained the ability to bind to DNA-PKcs. However, mutant F4 (positions 124–198), with a larger deletion encompassing the deamination domain, did not bind to DNA-PKcs (Fig. 3, A and B). These results led us to speculate that the binding of AID to DNA-PKcs also requires the deamination domain.
Deletion of the AID deamination domain could abrogate binding to DNA-PKcs because of alteration in the conformation of the AID C terminus or because of inactivation of the cytidine deaminase activity. To determine whether inactivation of cytidine deaminase activity abrogated binding to DNA-PKcs, we tested the ability of two deamination-defective AID mutants to associate with DNA-PKcs. An AID dominant negative mutant (H56R/E58Q) (7, 28) exhibited no binding, and an AID variant found in some patients with type II hyperIgM syndrome (AID-R112H) (27, 29) exhibited very little binding to DNA-PKcs (Fig. 3C). These results show that subtle mutations in the deamination domain of AID that impair cytidine deaminase activity abrogate binding to DNA-PKcs.
Association of AID and DNA-PKcs is DNA dependent
Some propose that AID promotes isotype class switch by deaminating cytidines in the DNA of switch regions (30); others suggest that AID edits RNA, originating a class switch-specific factor, such as an exonuclease or endonuclease, to resect DNA ends (31). Because our results indicate that the AID deamination domain is necessary for the recruitment of DNA-PKcs, we asked whether DNA is a cofactor for AID and DNA-PKcs complex formation. To answer this question, we tested whether a GST-AID fusion protein associates with DNA-PKcs in the presence or the absence of DNA. Fig. 4A shows that addition of exogenous ssDNA or dsDNA increases the efficiency of DNA-PKcs precipitation from HEK293 extracts by immobilized GST-AID (Fig. 4A, lanes 3 and 4). Precipitation reflected the specific properties of DNA, because adding DNase I disrupted GST-AID/DNA-PKcs complex formation (Fig. 4A, lanes 5–7). Likewise, DNase I treatment caused dissociation of coimmunoprecipitated AID/DNA-PKcs complexes (Fig. 4B, left panel). However, when the immunoprecipitates were cross-linked with primary amine-reactive and thiol-cleavable cross-linker STDDP that only cross-links proteins, DNase I treatment could no longer dissociate DNA-PKcs from AID (Fig. 4B, right panel). These results indicate that AID and DNA-PKcs form a stable complex through protein-protein interaction requiring DNA as a cofactor.
FIGURE 4. AID associates directly with DNA-PKcs, and association requires DNA as a cofactor. A, DNA is a cofactor for DNA-PKcs binding to AID because the complex is dissociated by DNase I treatment (lanes 5–7) and is enhanced by addition of dsDNA(lane 3) and ssDNA (lane 4). Shown is a Western blot (WB) analysis of GST or GST-AID bead pulldowns of HEK293 whole cell extracts in the presence or the absence of ssDNA or dsDNA and with or without DNase I. Blots were probed with either anti-DNA-PKcs or anti-GST Abs as indicated. B, DNA-PKcs and AID associate via protein-protein interaction, because cross-linking abolishes the sensitivity of the complex to DNase I. Shown is a Western blot (WB) analysis of anti-Flag immunoprecipitates obtained from whole cell lysates of HEK293 cells expressing Flag-tagged AID-wt treated with or without cross-linker STDDP and subsequently with or without 20 U of DNase I after immunoprecipitation. Cross-linking was disrupted with 2-ME before SDS-PAGE and Western blot analysis. Blots were probed with either anti-Flag or anti-DNA-PKcs Abs as indicated. C, DNA-PKcs association with AID is independent of Ku80. Shown is a Western blot (WB) analysis of immunoprecipitated AID/DNA-PKcs complexes probed with anti-DNA-PKcs, anti-Flag, or anti-Ku80 Ab as indicated. To show the presence of the tested proteins and the Ab reactivity, we included total cell extract from AID-expressing HEK293 cells (first lane in each panel) equivalent to 1% of the material used in immunoprecipitation as a positive control. D, Sequence alignments of the C-terminal domains of mouse (m) and human (h) AID and human (h) Ku80.
It is generally thought that high affinity binding of DNA-PKcs to DNA breaks requires the association with Ku70/Ku80 heterodimer (32). Hence, we asked whether Ku80 was also present in AID/DNA-PKcs complexes. Fig. 4C shows that Ku80 is not detectable in the AID/DNA-PKcs complexes, although it is clearly present in the cell extracts (Fig. 4C, control lanes). This result suggests that DNA-PKcs does not bind to Ku80, although it is associated with AID.
AID C-terminal deletion mutant causes cell death
How does the association of DNA-PKcs contribute to the function of AID? DNA-PKcs is thought to contribute to the generation of lymphocyte receptors by promoting the repair of double-strand breaks generated during V(D)J and class switch recombination by NHEJ (33). Because persistent DNA double-strand breaks cause cell death, repair mediated by DNA-PKcs may be critical for cell survival. Consistent with this idea is the finding that mice deficient in DNA-PKcs lack B and T lymphocytes (33) and also exhibit hypersensitivity to agents that cause double-strand breaks, such as ionizing radiation (33). Hence, we asked whether recruitment of DNA-PKcs by AID promotes the survival of cells undergoing DNA breaks associated with cytidine deamination (9). To test this idea, we measured DNA damage foci and death of cells expressing AID-wt or the C-terminal deletion mutant AID (AID-C). Fig. 5A shows that transient expression of AID-C that does not bind DNA-PKcs in MEF resulted in 18% dead cells, whereas transient expression of AID-wt led to only 5% dead cells, comparable to the cell death observed in nontransfected cells (4.5%). Our results indicate that recruitment of DNA-PKcs by AID promotes cell survival.
FIGURE 5. Association of AID with DNA-PKcs protects cells from death, and they do not form nuclear H2AX foci. A, Expression of an AID C-terminal deletion mutant (AID-C), but not AID-wt, induces cell death. Shown are representative confocal images (left panels) of wild-type MEF cells transiently transfected with Flag-tagged AID-wt or Flag-tagged AID-C. At 24 h post-transfection, apoptosis of Flag-positive cells (red) was scored for TUNEL positivity (green) and DAPI staining (blue). The chart (right panel) shows the proportion of apoptotic cells 24 h after transfection. Data were collected from four independent experiments by scoring at least 500 cells/experiment. NT, nontransfected cells. B, Expression of C-terminal-deleted mutant AID (AID-C), but not AID-wt, induces the formation of nuclear H2AX foci. Shown are representative confocal images (left panels) of wild-type MEF cells transiently transfected with Flag-AID-wt or Flag-AID-C, immunostained with anti-Flag (red) and anti-H2AX (green) Abs and counterstained with DAPI (blue). The chart (right panel) shows the proportion of nontransfected (left panel, white arrows) or transfected (red arrows) cells with large nuclear H2AX foci 24 h after transfection. Data were collected from three independent experiments by scoring at least 150 cells/experiment. C, Expression of AID-wt causes DNA damage foci in repair-deficient cells. Shown are representative confocal images (left panels) of DNA-PKcs –/– or 53BP1–/– MEF cells transiently transfected with Flag-AID-wt, immunostained with anti-Flag (red) and anti-H2AX (green) Abs, and counterstained with DAPI (blue). The chart (right panel) shows the proportion of nontransfected (left panels, white arrows) or transfected (red arrows) cells with large nuclear H2AX foci 24 h after transfection. Data were obtained from three independent experiments by scoring at least 150 cells/experiment.
AID C-terminal deletion mutant causes heightened DNA damage response
To test whether the expression of AID-C led to a DNA damage response, we stained MEF cells expressing AID-wt or AID-C proteins for phosphorylated H2A histone family member X (H2AX), which binds to DNA double-strand breaks, forming foci (9). Fig. 5B shows that only 1.1% of nontransfected MEF cells (Fig. 5B, white arrows) and 9.4% cells transfected with AID-wt (Fig. 5B, red arrow, upper panel) exhibited detectable nuclear H2AX foci. In contrast, 75.6%, upper panel MEF cells expressing AID-C had massive accumulation of nuclear H2AX foci (Fig. 5B, red arrow, lower panel). Our data are consistent with the concept that recruitment of DNA-PKcs by AID is needed to resolve DNA double-stranded breaks.
DNA-PKcs and 53BP1 are necessary to resolve AID-generated H2AX foci
Absence of H2AX nuclear foci in wild-type MEF cells expressing AID-wt could result from prompt DNA repair or, alternatively, from decreased DNA break formation. To determine the contribution of DNA repair to the lack of DNA damage foci in cells expressing AID-wt, we examined DNA damage foci in repair-deficient cells expressing AID-wt. Fig. 5C shows that expression of AID-wt induced nuclear H2AX foci in 80.5% DNA-PKcs–/– cells, whereas only 9.3% of nontransfected cells scored positive. Similarly, the expression of AID-wt also induced H2AX foci in 64% 53BP1–/– cells (34), whereas only 17.2% nontransfected cells were positive. These results indicate that AID generates DNA double-strand breaks and H2AX foci, which accumulate in the absence of DNA-PKcs or other DNA repair components, such as 53BP1.
AID C-terminal deletion mutant impairs survival of activated B cells
Our results indicating that AID deficient in DNA-PKcs binding causes accumulation of DNA double-strand breaks suggest a mechanism for the selective class-switching defect in some patients with hyperIgM type II syndrome. We tested whether the expression of AID mutants deficient in DNA-PKcs binding due to C-terminal deletion (AID-C) impaired the survival of B cells undergoing class switch recombination. To do this, we transduced LPS-activated B cells obtained from spleens of C57BL/6 mice with retroviral vectors encoding AID-wt, AID-C, or GFP (Fig. 6A). Fig. 6B shows that although 91.6% AID-C-transduced B cells (GFP-positive) were dead on day 3 of LPS culture, only 25.2% of the cells transduced with AID-wt and 19.2% of the cells expressing only GFP died (Fig. 6B). Our results indicate that AID-C causes the death of cells undergoing LPS stimulation. We observed that the surviving AID-C-expressing cells did not class switch, thus confirming the findings of Barreto et al. (26), who showed that AID-C does not promote class switch. Because AID-C transduced B cells undergo cell death, our results indicate that AID-C-associated defective class switch is due to the death of cells undergoing class switch recombination.
FIGURE 6. AID C-terminal deletion mutant (AID-C) induces cell death of LPS-stimulated B cells. A, Schematic representation of the retroviral constructs used to transduce B cells. B, C-terminal-deleted mutant AID (AID-C), but not AID-wt, induces death of LPS-stimulated B cells. Cell death was scored by propidium iodide (PI) staining of LPS-stimulated splenocytes transduced with proviruses encoding GFP, AID-wt, or C-terminal-deleted AID mutant (AID-C) on day 3. PI-positive cells were identified by flow cytometric analysis. The graph shows the fraction of transduced B cells (GFP-positive) that were PI positive (y-axis). Data were collected from three independent experiments and represent the mean and SD.
C-terminal deletion does not alter AID intracellular localization
McBride et al. (35) and Ito et al. (36) proposed an alternative mechanism for the impaired class switch recombination by AID-C mutants. These authors suggested that deficient class switch is a consequence of deletion of a leucine-rich nuclear export signal causing predominant nuclear localization of AID-C (35, 36). However, Fig. 5, A and B, show that Flag-tagged AID-C such as AID-wt localized predominantly in the cytoplasm of MEF. Our results showing that removal of the nuclear export signal (aa 189–198) in the Flag-tagged AID-C did not alter intracellular localization of the protein indicate that the prodeath effect of AID-C is not due to nuclear accumulation.
Intracellular localization of AID may not be primarily governed by nucleocytoplasmic shuttling. If AID shuttled between cytoplasm and nucleus, then addition of nuclear localization signal should drive AID to the nucleus. Fig. 7, A–F, shows that although tagging with three consecutive copies of SV40 nuclear localization signals (37) efficiently drives GFP into the nucleus (Fig. 7B), AID remains in the cytoplasm (Fig. 7E). These results suggest that AID is actively retained in the cytoplasm. Consistent with this possibility, cytoplasmic localization of AID-wt is resistant to digitonin treatment (Fig. 7F) that effectively depletes GFP from the cytoplasm of the cells (Fig. 7C). Our data indicate that AID is retained in the cytoplasm by digitonin-resistant cytoskeletal elements. One of these elements is -tubulin, because -tubulin specifically coprecipitated (Figs. 1 and 7G) and partially colocalized (Fig. 7, H–J) with AID.
FIGURE 7. AID is retained in the cytoplasm by cytoskeletal proteins. GFP distributes in the cytoplasm and nucleus of HeLa cells (A), whereas Flag-AID localizes predominantly to the cytoplasm. Addition of three SV40 nuclear localization signals to GFP (3xNLS-GFP) or to Flag-tagged AID (3xNLS-AID) causes nuclear localization of GFP (B), but not Flag-tagged AID (E). Digitonin treatment of cells removes GFP (C), but not Flag-tagged AID (F), from the cytoplasm, indicating cytoplasmic retention of AID. G, AID associates with -tubulin. Shown are Western blot (WB) analyses of anti-Flag immunoprecipitates (IP) of cell lysates obtained from HeLa, HEK293, or 70Z/3 cells transfected with Flag-AID (+) or empty vector (–). Blots were probed with either anti-Flag or anti--tubulin Abs, as indicated. The mouse Ig H chain, migrating slight above -tubulin, is seen as a weak band in the controls. H–J, AID partially colocalized with -tubulin. Shown are confocal images of HeLa cells expressing Flag-AID double stained with Flag (red)and anti--tubulin (green) Abs. Arrowheads show colocalization of AID and -tubulin.
It is possible that the properties of AID-GFP fusion proteins studied by McBride et al. (35) and Ito et al. (36) differ from the properties of Flag-tagged AID or native AID. Although the intracellular distribution of Flag-tagged AID-C is determined by cytoplasmic retention, the distribution of AID-C-GFP fusion proteins is not. Instead, intracellular distribution of AID-C-GFP fusion proteins is determined by nucleocytoplasminic shuttling.
Discussion
We show in this study that AID promotes cell survival by recruiting DNA-PKcs to the DNA, hence resolving double-strand breaks. Our findings explain how B cells survive while undergoing DNA double-strand breaks during class switch recombination. Consistent with this idea, compromised class switch recombination is a common phenotype in mice deficient in proteins necessary for DNA damage repair, such as H2AX (30), Ataxia Telangectasia Muntant (38), 53BP1 (39), and Mre11 (40).
Gell and Jackson (41) showed that DNA-PKcs associates with Ku70/Ku80 binding to the 12-aa C-terminal tail of the Ku80 subunit. Our data show that DNA-PKcs binding to AID requires the C terminus of AID (Fig. 3, A and B). Such a striking binding parallel made us wonder whether the C-terminal domain of Ku80 might be homologous in any way to the C-terminal domain of AID. Sequence alignment of the 14-aa C-terminal domains of AID and Ku80 revealed that both sequences possess a common (E/D)VDDL(X)D motif (Fig. 4D). The common motif in the C-terminal domains of AID and Ku80 suggest competition for the same binding site on DNA-PKcs. This mechanism may coordinate the formation of AID/DNA-PKcs and Ku70/Ku80/DNA-PKcs complexes for repair of double-strand breaks by nonhomologous end joining (NHEJ) after cytidine deamination of DNA.
We propose a working model to explain how AID promotes the survival of cells undergoing DNA double-strand breaks. AID binding to DNA through the DNA deamination domain undergoes a conformational change resulting in exposure of the C-terminal domain that, in turn, recruits DNA-PKcs to the DNA. Upon the generation of DNA breaks by cytidine deamination, DNA-PKcs initiates NHEJ by dissociating from AID and reassociating with Ku80 to assemble the NHEJ repair complex. Our results thus explain previous observations by Barreto et al. (26) and Ta et al. (27), who showed that C-terminal-deleted AID mutants fail to promote class switch recombination. Our results suggest that the mechanism underlying hyperIgM type II syndrome associated with AID mutations that truncate the C-terminal portion of the molecule is the selective death of B cells stimulated to undergo class switch recombination because of accumulation of DNA double-strand breaks. That C-terminal-deleted AID mutants may sustain somatic hypermutation (27) is also in agreement with our findings, because somatic hypermutation is less dependent on the recruitment of DNA-PKcs than class switch recombination, possibly due to engagement of alternative repair mechanisms (42, 43). Our results may also explain why lack of DNA-PKcs causes class switch deficiency for all isotypes except IgG1 (21), whereas the SCID mutation, which truncates DNA-PKcs, allows moderate class switching to all isotypes (44). Although AID in the absence of DNA-PKcs is unable to recruit repair factors to the sites of cytidine deamination, it may do so in the presence of DNA-PKcs with the SCID mutation.
Acknowledgments
We thank Drs. Zhenkun Lou and Junjie Chen for kindly providing us with 53BP1–/– MEF cells and Abs, Dr. Frederich W. Alt for generously providing us with rabbit anti-AID antiserum, and Dr. David Chen for kindly providing us with DNA-PKcs-deficient cell lines. We also thank Dr. Cristina Jo?o for valuable discussion, and Michelle Rebrovich for excellent technical assistance.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grants AI48602 (to M.C.) and HL46810 and HL52297 (to J.L.P.).
2 Address correspondence and reprint requests to Dr. Marilia Cascalho, Transplantation Biology, Mayo Clinic, 200 First Street SW, Medical Sciences 2-113, Rochester, MN 55905. E-mail address: cascalho.marilia@mayo.edu
3 Abbreviations used in this paper: AID, activation-induced cytidine deaminase; CSR, class switch recombination; DAPI, 4',6-diamido-2-phenylindole hydrochloride; DNA-PKcs, DNA-dependent protein kinase catalytic subunit; EGFP, enhanced GFP; MEF, mouse embryonic fibroblast; NHEJ, nonhomologous end joining; SHM, somatic hypermutation; STDDP, 3,3'-dithiobis-(sulfosuccinimidylpropionate).
Received for publication August 25, 2004. Accepted for publication October 25, 2004.
References
Durandy, A.. 2003. Activation-induced cytidine deaminase: a dual role in class-switch recombination and somatic hypermutation. Eur. J. Immunol. 33:2069.
Muramatsu, M., V. S. Sankaranand, S. Anant, M. Sugai, K. Kinoshita, N. O. Davidson, T. Honjo. 1999. 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. 274:18470.
Begum, N. A., K. Kinoshita, M. Muramatsu, H. Nagaoka, R. Shinkura, T. Honjo. 2004. De novo protein synthesis is required for activation-induced cytidine deamination-dependent DNA cleavage in immunoglobulin class switch recombination. Proc. Natl. Acad. Sci. USA 101:13003.
Petersen-Mahrt, S. K., R. S. Harris, M. S. Neuberger. 2002. AID mutates E. coli suggesting a DNA deamination mechanism for antibody diversification. Nature 418:99.
Dickerson, S. K., E. Market, E. Besmer, F. N. Papavasiliou. 2003. AID mediates hypermutation by deaminating single stranded DNA. J. Exp. Med. 197:1291
Pham, P., R. Bransteitter, J. Petruska, M. F. Goodman. 2003. Processive AID-catalysed cytosine deamination on single-stranded DNA simulates somatic hypermutation. Nature 424:103.
Chaudhuri, J., M. Tian, C. Khuong, K. Chua, E. Pinaud, F. W. Alt. 2003. Transcription-targeted DNA deamination by the AID antibody diversification enzyme. Nature 422:726.
Papavasiliou, F. N., D. G. Schatz. 2000. Cell-cycle-regulated DNA double-stranded breaks in somatic hypermutation of immunoglobulin genes. Nature 408:216.
Petersen, S., R. Casellas, B. Reina-San-Martin, H. T. Chen, M. J. Difilippantonio, P. C. Wilson, L. Hanitsch, A. Celeste, M. Muramatsu, D. R. Pilch, et al 2001. AID is required to initiate Nbs1/-H2AX focus formation and mutations at sites of class switching. Nature 414:660.
Zhang, K.. 2003. Accessibility control and machinery of immunoglobulin class switch recombination. J. Leukocyte Biol. 73:323.(Xiaosheng Wu, Pedro Geral)
Activation-induced cytidine deaminase (AID) is required for Ig class switch recombination, a process that introduces DNA double-strand breaks in B cells. We show in this study that AID associates with the DNA-dependent protein kinase catalytic subunit (DNA-PKcs) promoting cell survival, presumably by resolving DNA double-strand breaks. Wild-type cells expressing AID mutants that fail to associate with DNA-PKcs or cells deficient in DNA-PKcs or 53BP1 expressing wild-type AID accumulate H2AX foci, indicative of heightened DNA damage response. Thus, AID has two independent functions. AID catalyzes cytidine deamination that originates DNA double-strand breaks needed for recombination, and it promotes DNA damage response and cell survival. Our results thus resolve the paradox of how B cells undergoing DNA cytidine deamination and recombination exhibit heightened survival and suggest a mechanism for hyperIgM type II syndrome associated with AID mutants deficient in DNA-PKcs binding.
Introduction
Activation-induced cytidine deaminase (AID)3 promotes somatic hypermutation (SHM) and class switch recombination (CSR) of Ig genes (1). AID cytidine deaminase activity was first proposed on the basis of its homology with the apolipoprotein B mRNA-editing catalytic polypeptide 1 (2). The cytidine deaminase property of AID led to two distinct hypotheses to explain the diversification of Ig genes. The first, the RNA-editing hypothesis, proposes that AID, like apolipoprotein B mRNA-editing catalytic polypeptide 1, modifies unknown RNA precursors that, in turn, originate endonucleases that cleave the DNA encoding the Ig genes. The finding by Begum et al. (3), showing that AID-dependent DNA cleavage in CSR requires de novo protein synthesis, is in agreement with this hypothesis. The second, the DNA deamination hypothesis, proposes that AID deaminates cytidine to uracil directly in the DNA encoding the Ig genes (4). Compatible with the second hypothesis are the results reported by Petersen-Marth et al. (4), showing that expression of AID in Escherichia coli originates a mutator phenotype that yields nucleotide transitions at dC/dG, and the findings of Dickerson et al. (5), Pham et al. (6), and Chaudhuri et al. (7), showing that AID deaminates DNA substrates in vitro.
Exactly how AID introduces point mutations or executes class switch recombination is not yet understood, but it is generally thought that cytidine deamination of DNA or RNA somehow generates double-strand breaks in Ig DNA (1, 8, 9, 10, 11). Ordinarily, cells respond to DNA double-strand breaks by undergoing cell cycle arrest to allow time for repair (12) and respond to persistent damage by inducing apoptosis, presumably as a protection against illegitimate recombination (13). However, B cells undergoing Ig class switch do not die, presumably because they efficiently repair DNA double-strand breaks. Although RAD54, RAD52, and RAD51 repair proteins are needed for AID-induced Ig gene conversion in chicken cell lines (14, 15), whether AID directly recruits repair factors to the locales of cytidine deamination is not known.
Materials and Methods
Generation of AID wild-type (AID-wt) and mutant constructs
RNA was obtained from C57BL/6 mouse lymph nodes using TRIzol reagent. AID cDNA was produced by RT using oligo(dT) primer and amplified by PCR using Turbo pfu polymerase (Stratagene) and primer set wu160/wu167. Full-length AID cDNA was cloned in-frame into pUHD10S vector downstream of Flag tag sequences. AID deletion mutants were generated by PCR using Turbo pfu. The primer sets used to generate the deletion mutants were: mutant F1, wu160/wu166; mutant F3, wu160/wu165; mutant F4, wu162/wu167; mutant F6, wu161/wu167; and mutant C, wu160/wu174. We used the QuikChange mutagenesis kit (Stratagene) to generate an AID dominant negative mutant and AID-R112H point mutation constructs with primer sets wu155/wu156 and wu157/wu158, respectively. PCR fragments were flanked by 5' NheI and 3' XbaI sites to allow subsequent cloning into the pUHD10S vector. The Flag-tagged AID fragment (EcoRI/XbaI) was subcloned into the pCI expression vector (Promega) for transient expression and into the pCI-neo expression vector (Promega) for stable transfections. For expression in 70Z/3 cells and splenocytes, Flag-AID or Flag-AID-C fragments were cloned into pIRES2GFP vector (BD Clontech) upstream of the internal ribosomal entry site sequence, followed by subcloning of the Flag-AID-(or Flag-AID-C)-IRES2-EGFP cassette into the pMSCV-puro retroviral vector (BD Clontech). To produce the GST-AID fusion protein, the full-length AID cDNA was cloned in-frame into the pGEX4T1 vector (Amersham Biosciences). All sequences were verified by DNA sequencing.
Cell culture, transfection, and retroviral transduction
Wild-type mouse embryonic fibroblasts (MEF) were generated from C57BL/6 mouse embryos at 14.5 days postcoitum and maintained in DMEM supplemented with 10% FCS, 1.0 IU/ml penicillin G, and 0.5 IU/ml streptomycin. HeLa cells and HEK293 cells were cultured in DMEM and supplemented with 10% FCS, 1.0 IU/ml penicillin G, and 0.5 IU/ml streptomycin. DNA-dependent protein kinase catalytic subunit null (DNA-PKcs–/–) MEF cells (PK33N) (16) were provided by Dr. D. J. Chen (Lawrence Berkeley National Laboratory) and maintained in -MEM supplemented with 10% FCS, 1.0 IU/ml penicillin G, and 0.5 IU/ml streptomycin. Mouse pre-B lines 70Z/3 and 18.81 cells were maintained in RPMI 1640 medium supplemented with 10% FCS, 1.0 IU/ml penicillin G, and 0.5 IU/ml streptomycin. LPS cultures were prepared by incubating B cells (95% purity, isolated with a MACS column; 2 x 105/well/100 μl) with 10 μg/ml LPS (Sigma-Aldrich) in RPMI 1640 medium supplemented with 10% FCS, 1.0 IU/ml penicillin G, and 0.5 IU/ml streptomycin. B cells were obtained from C57BL/6 mice as previously described (17).
Transient transfection of AID-wt or mutant construct into HeLa, HEK293, or MEF cells was performed with Lipofectamine (Invitrogen). The expression of AID-wt in 70Z/3 cells and splenocytes was performed by retroviral transduction. Linearized pMSCV-IRES2-EGFP-puro or pMSCV-Flag-AID-IRES2-EGFP-puro DNA was stably transfected into RetroPack PT67 (BD Clontech) packaging cells by electroporation, whereas pMSCV-Flag-AID-C-IRES2-EGFP-puro was transiently transfected into the same packaging line using Lipofectamine. The virus-containing supernatant of each kind (in DMEM) was collected every 24 h and additionally concentrated by centrifugation at 6000 x g for 4 h. Virus pellets were resuspended in a 1/50 volume of complete RPMI 1640. LPS cultures were transduced with proviruses at the time of seeding in 100 μl of virus-containing supernatant supplemented with 10 μg/ml LPS. AID- and AID-C-expressing splenocytes were analyzed with a FACSCalibur (BD Biosciences) by measuring GFP positivity and propidium iodide stain for cell death.
Subcellular fractionation
Nuclei and cytosol fractions were prepared according to published protocols (18). Briefly, 107 HEK293 cells expressing Flag-AID were harvested and washed three times with PBS. Cells were resuspended in 1 ml of lysis buffer (10 mM HEPES (pH 7.5), 10 mM KCl, 1.5 mM MgCl2, 1 mM DTT, and protease inhibitor mixture (Roche)) on ice for 20 min, followed by 20 gentle strokes for homogenization with a loose-fit Dounce homogenizer (Kontes). The homogenate was overlaid on top of 200 μl of 40% sucrose in lysis buffer and centrifuged at 800 x g for 15 min in a centrifuge with a swinging bucket rotor. The supernatant (corresponding to the cytosolic fraction) and the pellet (corresponding to the nuclear fraction) were collected and extracted with 10x lysis buffer (19).
Immunoprecipitation and Western blotting
Immunoprecipitation was performed as described previously (19). To identify AID-binding proteins, 2 x 108 HeLa cells expressing Flag-AID were used for a large-scale immunoprecipitation with EZview anti-Flag (M2) beads (Sigma-Aldrich). Otherwise, routine immunoprecipitation was conducted using 2 x 107 cells expressing Flag-AID. In each experiment, one-fourth of the precipitated proteins (equivalent to 5 x 106 cells) were resolved on a 6% SDS-PAGE (for DNA-PKcs and Ku80) or 12% SDS-PAGE (for AID). In the experiments testing DNA dependence of AID/DNA-PKcs binding, one-fourth of precipitated beads were incubated with 100 μl of PBS alone or 10 mM 3,3'-dithiobis-(sulfosuccinimidylpropionate) (STDDP) in PBS for 2 h on ice to cross-link protein complexes. After quenching the cross-linking reaction with 10 μl of 1.0 M Tris-HCl (pH 7.4) for 15 min and washed twice with PBS, the beads were then treated with 20 U of DNase I (as indicated in Fig. 3B) in 40 μl of PBS at room temperature for 30 min, followed by two washes before SDS-PAGE analysis under reducing condition (5% 2-ME) to cleave cross-linked complexes. Mouse anti-Flag (M2) mAb was purchased from Sigma-Aldrich. Rabbit anti-DNA-PKcs (specific to human DNA-PKcs, SC-9051) and anti-Ku80 (SC-5280) mAb were purchased from Santa Cruz Biotechnology. Anti-DNA-PKcs mAb (specific to mouse DNA-PKcs; NA57) was obtained from Oncogene, goat anti-GST Ab is a product of Amersham Biosciences (27-4577-01), anti--tubulin mAb was obtained from Santa Cruz Biotechnology (SC-5274), and rabbit anti-AID serum was a gift from Dr. F. W. Alt (Harvard Medical School, Boston, MA).
FIGURE 3. AID deamination and C-terminal domains are required for binding to DNA-PKcs. A, Schematic domain representation of full-length AID and deletion mutants. DNA-PKcs binding of wild-type or deletion mutants is noted on the right. Wild-type AID and mutants F1 and F6 associate with DNA-PKcs; mutants F1, F4, and C do not. B, Deamination and C-terminal domains of AID are necessary for binding to DNA-PKcs. Shown is a Western blot (WB) analysis of anti-Flag immunoprecipitates (IP) obtained from lysates of HEK293 cells transiently transfected with wt or deletion mutants (F1, F3, F4, F6, and C) of AID. Blots were probed with either anti-Flag or anti-DNA-PKcs Abs as indicated. C, Cytidine deamination-defective AID mutants do not bind to DNA-PKcs. Shown is a Western blot (WB) analysis of anti-Flag immunoprecipitates obtained from lysates of HEK293 cells stably expressing Flag-tagged AID-wt, dominant negative H56R/E58Q mutant (DN), or R112H mutant AID. Blots were probed with either anti-Flag or anti-DNA-PKcs Ab as indicated. The figures are representative of three independent experiments each.
DNA-PKcs in vitro pulldown by GST-AID
Recombinant GST and GST-AID fusion protein were purified from bacterial DL21 cells. Five hundred nanograms of GST or GST-AID fusion protein beads were incubated with 0.5 ml of cell extracts obtained from 2 x 107 HEK293 cells overnight at 4°C in the presence or the absence of 2 μg of ssDNA, 2 μg of dsDNA, or 20 U of DNase I. The ssDNA was a 59-base oligonucleotide containing RGYW repeats (AGCTGGCAGGCTAGCAAGTTGGTT-GGCAAGCAGGTAAGCAGG CAAGCTGGCTGAATTCC) (7). The dsDNA was an EcoRV-linearized pBluescript KS vector (Stratagene). Beads were washed and analyzed as described above.
Immunofluorescence staining and microscopy
Immunofluorescence staining and confocal microscopy were performed essentially as described previously (19). For microtubule staining in HeLa cells expressing Flag-AID, cells were fixed with methanol at –20°C for 10 min, air-dried, rehydrated, repermeabilized with 0.05% Triton X-100 in PBS for 3 min, and blocked for 90 min with blocking buffer (5% normal goat serum, 1% glycerol, 0.1% BSA, 0.1% fish skin gelatin, and 0.04% sodium azide), followed by staining with anti--tubulin Ab (Sigma-Aldrich; T5293). Cytoplasmic protein extraction was performed by incubating HeLa cells expressing GFP or Flag-AID with 50 μg/ml digitonin in PBS on ice for 5 min, followed by four washes with PBS and fixation with 4% paraformaldehyde. Affinity-purified rabbit anti-H2AX was provided by Dr. J. Chen (Mayo Clinic, Rochester, MN). Cell death of MEF was determined by TUNEL assay (Promega) and by the presence of condensed chromatin or fragmented nuclei in 4',6-diamido-2-phenylindole hydrochloride (DAPI) staining (20).
Identification of AID-associated proteins
The proteins coimmunoprecipitated with Flag-AID from HeLa cells were resolved on a SDS-PAGE (4–15% polyacrylamide gradient; Bio-Rad) and stained with Coomassie G-250 (Bio-Rad). Protein bands were excised and analyzed using MALDI-TOF mass spectrometry by the Rockefeller University Protein Resource Center (New York, NY).
Oligonucleotides used
The following oligonucleotides were used: Wu155, CTTCGCAACAAGTCTGGCTGCCGCGTGCAATTGTTGTTCCTACGCTACATC; Wu156, GATGTAGCGTAGGAACAACAATTGCACGCGGCAGCCAGACTTGTTGCGAAG; Wu157, CAGCCTGAGGATTTTCACCGCGCACCTCTACTTCTGTGAAGACCGC; Wu158, GCGGTCTTCACAGAAGTAGAGGTGCGCGGTGAAAATCCTCAGGCT; Wu160, GAATCAGCTAGCGACAGCCTTCTGATGAAGCAAAAG; Wu161, GAATCAGCTAGCGGCTGCCACGTGGAATTGTTGTTC; Wu162, GAATCAGCTAGCGAGGGGCTGCGGAGACTGCACC; Wu165, GAATCATCTAGATTAAGGCTCAGCCTTGCGGTCTTCAC; Wu166, GAATCATCTAGATTAATTTTCTACAAATGTATTCCAGCAG; Wu167, GAATCATCTAGATTAAAATCCCAACATACGAAATGCATC; and Wu174, GAATCATCTAGATTAGTCATCGACTTCGTACAAGGGCAAAAGG.
Results
AID binds to DNA-PKcs in the nucleus
We considered the possibility that association of AID with cofactors could promote survival of cells undergoing Ig gene diversification. To determine whether AID associates with other molecules, we analyzed AID protein complexes obtained from HeLa cells expressing AID. AID was immunoprecipitated from lysates, and the identity of any coprecipitates was sought by mass spectrometry. Fig. 1A shows a Coomassie-stained gel image showing proteins coimmunoprecipitated with AID. MALDI-TOF mass spectrometry identified the largest coprecipitated protein as DNA-PKcs with a Mr of 486 kDa based on 34 matched peptides. Other proteins identified were heat shock 70-kDa protein 8 isoform 1 (HSC70; 12 matched peptides), -tubulin (15 matched peptides), and a protein similar to mitochondrial solute carrier family 25 (four matched peptides; Fig. 1).
FIGURE 1. Identification of AID-binding proteins. Coomassie Blue staining of 4–15% gradient SDS-PAGE analysis of anti-Flag immunoprecipitates of lysates obtained from HeLa cells transiently transfected with Flag-AID or empty vector (mock) was performed. Flag-AID comigrates with the Ab L chain (25KD). AID-binding proteins were identified by MALTI-TOF mass spectrometry. Protein identities and accession numbers are indicated.
Because DNA-PKcs is required to efficiently resolve by nonhomologous end-joining the DNA double-strand breaks associated with class switch recombination (21), we questioned whether AID associated with DNA-PKcs in B-lineage cells and cells other than HeLa. Fig. 2A shows DNA-PKcs coimmunoprecipitated with AID and vice versa in extracts from HeLa cells, human embryonic kidney 293 (HEK293) cells, and murine B cells (70Z/3) transfected with Flag-tagged AID. We also show in Fig. 2B that endogenously expressed AID in 18.81 B cells (22) binds to DNA-PKcs (left panel). Our results indicate that the association of AID with DNA-PKcs may reflect the function of AID in B cells.
FIGURE 2. AID associates with DNA-PKcs of B and non-B cells. A, AID coprecipitates with DNA-PKcs in various AID-transfected cell lines. Shown are Western blot (WB) analyses of anti-Flag or anti-DNA-PKcs immunoprecipitates (IP) of cell lysates obtained from HeLa, HEK293, or 70Z/3 cells transfected with Flag-AID (+) or empty vector (–). Blots were probed with either anti-Flag (Flag) or anti-DNA-PKcs (DNA-PKcs) Abs as indicated. Mouse Ig L chain comigrating with 25-kDa Flag-AID is seen as a weak band in the controls. This figure is representative of three independent experiments. B, Endogenous AID associates with DNA-PKcs. Shown is a Western blot (WB) analysis (left panel) of anti-DNA-PKcs immunoprecipitates (IP) of cell lysates from mouse pre-B cell line 18.81 that constitutively expresses AID or from mouse pre-B cell line 70Z/3 that does not express endogenous AID. Blots were probed with either anti-DNA-PKcs Abs or rabbit polyclonal anti-AID (7 ) as indicated. The expression of endogenous AID in 18.81 and 70Z lines was verified by RT-PCR, as shown in the right panel. C, AID and DNA-PKcs associate in the nucleus. Shown is a Western blot analysis of Flag-AID or DNA-PKcs immunoprecipitates from whole cell extracts, cytoplasmic fraction, or nuclear fraction obtained from HEK293 cells expressing AID. The blots were probed with either anti-DNA-PKcs or anti-Flag Abs as indicated.
AID is mainly localized to the cytoplasm (23), whereas DNA-PKcs localizes predominantly in the nucleus (24), yet they associate with one another. We asked whether the association of DNA-PKcs and AID took place in the nucleus, where presumably AID and DNA-PKcs function. Fig. 2C shows that DNA-PKcs coprecipitates with AID in the nuclear fraction, but not in the cytoplasmic fraction of cell extracts (Fig. 2C). Our results indicate that the association of DNA-PKcs and AID takes place in the nucleus, representing the small fraction of total cellular AID (Fig. 2C).
Association of AID and DNA-PKcs requires the C-terminal and deamination domains of AID
Because DNA-PKcs is required for the repair of DNA double-strand breaks during class switch recombination (21), but is dispensable for somatic hypermutation (25), our findings showing that AID associates with DNA-PKcs predict that the complexes mediating class switch recombination and somatic hypermutation are distinct. Recent findings from two laboratories support this idea (26, 27). Barreto et al. (26) found that an AID mutant that lacks the C-terminal 10 aa retained cytidine deaminase activity, but failed to promote class switch recombination. Ta et al. (27) found that some subjects with type II hyperIgM syndrome have mutations in AID causing truncation or disruption of the C-terminal domain. These individuals have a severe defect in class switch, but normal somatic hypermutation (27). These findings suggested that the C-terminal region of AID is necessary for class switch recombination in B cells (26, 27).
We questioned whether the selective defect in class switch recombination of AID C terminus deletion mutants reflected defective association with DNA-PKcs. To test this idea, we generated a series of AID deletion mutants and tested the mutants for association with DNA-PKcs. C-terminal truncation mutants lost (mutant F1 (positions 1–154) and C (positions 1–189)) or markedly reduced (mutant F3 (positions 1–123)) their ability to bind to DNA-PKcs (Fig. 3, A and B), suggesting that the C-terminal domain of AID is necessary for the formation of AID/DNA-PKcs complexes. In contrast, mutant F6 (positions 54–198) lacking the N-terminal 53-aa segment of AID retained the ability to bind to DNA-PKcs. However, mutant F4 (positions 124–198), with a larger deletion encompassing the deamination domain, did not bind to DNA-PKcs (Fig. 3, A and B). These results led us to speculate that the binding of AID to DNA-PKcs also requires the deamination domain.
Deletion of the AID deamination domain could abrogate binding to DNA-PKcs because of alteration in the conformation of the AID C terminus or because of inactivation of the cytidine deaminase activity. To determine whether inactivation of cytidine deaminase activity abrogated binding to DNA-PKcs, we tested the ability of two deamination-defective AID mutants to associate with DNA-PKcs. An AID dominant negative mutant (H56R/E58Q) (7, 28) exhibited no binding, and an AID variant found in some patients with type II hyperIgM syndrome (AID-R112H) (27, 29) exhibited very little binding to DNA-PKcs (Fig. 3C). These results show that subtle mutations in the deamination domain of AID that impair cytidine deaminase activity abrogate binding to DNA-PKcs.
Association of AID and DNA-PKcs is DNA dependent
Some propose that AID promotes isotype class switch by deaminating cytidines in the DNA of switch regions (30); others suggest that AID edits RNA, originating a class switch-specific factor, such as an exonuclease or endonuclease, to resect DNA ends (31). Because our results indicate that the AID deamination domain is necessary for the recruitment of DNA-PKcs, we asked whether DNA is a cofactor for AID and DNA-PKcs complex formation. To answer this question, we tested whether a GST-AID fusion protein associates with DNA-PKcs in the presence or the absence of DNA. Fig. 4A shows that addition of exogenous ssDNA or dsDNA increases the efficiency of DNA-PKcs precipitation from HEK293 extracts by immobilized GST-AID (Fig. 4A, lanes 3 and 4). Precipitation reflected the specific properties of DNA, because adding DNase I disrupted GST-AID/DNA-PKcs complex formation (Fig. 4A, lanes 5–7). Likewise, DNase I treatment caused dissociation of coimmunoprecipitated AID/DNA-PKcs complexes (Fig. 4B, left panel). However, when the immunoprecipitates were cross-linked with primary amine-reactive and thiol-cleavable cross-linker STDDP that only cross-links proteins, DNase I treatment could no longer dissociate DNA-PKcs from AID (Fig. 4B, right panel). These results indicate that AID and DNA-PKcs form a stable complex through protein-protein interaction requiring DNA as a cofactor.
FIGURE 4. AID associates directly with DNA-PKcs, and association requires DNA as a cofactor. A, DNA is a cofactor for DNA-PKcs binding to AID because the complex is dissociated by DNase I treatment (lanes 5–7) and is enhanced by addition of dsDNA(lane 3) and ssDNA (lane 4). Shown is a Western blot (WB) analysis of GST or GST-AID bead pulldowns of HEK293 whole cell extracts in the presence or the absence of ssDNA or dsDNA and with or without DNase I. Blots were probed with either anti-DNA-PKcs or anti-GST Abs as indicated. B, DNA-PKcs and AID associate via protein-protein interaction, because cross-linking abolishes the sensitivity of the complex to DNase I. Shown is a Western blot (WB) analysis of anti-Flag immunoprecipitates obtained from whole cell lysates of HEK293 cells expressing Flag-tagged AID-wt treated with or without cross-linker STDDP and subsequently with or without 20 U of DNase I after immunoprecipitation. Cross-linking was disrupted with 2-ME before SDS-PAGE and Western blot analysis. Blots were probed with either anti-Flag or anti-DNA-PKcs Abs as indicated. C, DNA-PKcs association with AID is independent of Ku80. Shown is a Western blot (WB) analysis of immunoprecipitated AID/DNA-PKcs complexes probed with anti-DNA-PKcs, anti-Flag, or anti-Ku80 Ab as indicated. To show the presence of the tested proteins and the Ab reactivity, we included total cell extract from AID-expressing HEK293 cells (first lane in each panel) equivalent to 1% of the material used in immunoprecipitation as a positive control. D, Sequence alignments of the C-terminal domains of mouse (m) and human (h) AID and human (h) Ku80.
It is generally thought that high affinity binding of DNA-PKcs to DNA breaks requires the association with Ku70/Ku80 heterodimer (32). Hence, we asked whether Ku80 was also present in AID/DNA-PKcs complexes. Fig. 4C shows that Ku80 is not detectable in the AID/DNA-PKcs complexes, although it is clearly present in the cell extracts (Fig. 4C, control lanes). This result suggests that DNA-PKcs does not bind to Ku80, although it is associated with AID.
AID C-terminal deletion mutant causes cell death
How does the association of DNA-PKcs contribute to the function of AID? DNA-PKcs is thought to contribute to the generation of lymphocyte receptors by promoting the repair of double-strand breaks generated during V(D)J and class switch recombination by NHEJ (33). Because persistent DNA double-strand breaks cause cell death, repair mediated by DNA-PKcs may be critical for cell survival. Consistent with this idea is the finding that mice deficient in DNA-PKcs lack B and T lymphocytes (33) and also exhibit hypersensitivity to agents that cause double-strand breaks, such as ionizing radiation (33). Hence, we asked whether recruitment of DNA-PKcs by AID promotes the survival of cells undergoing DNA breaks associated with cytidine deamination (9). To test this idea, we measured DNA damage foci and death of cells expressing AID-wt or the C-terminal deletion mutant AID (AID-C). Fig. 5A shows that transient expression of AID-C that does not bind DNA-PKcs in MEF resulted in 18% dead cells, whereas transient expression of AID-wt led to only 5% dead cells, comparable to the cell death observed in nontransfected cells (4.5%). Our results indicate that recruitment of DNA-PKcs by AID promotes cell survival.
FIGURE 5. Association of AID with DNA-PKcs protects cells from death, and they do not form nuclear H2AX foci. A, Expression of an AID C-terminal deletion mutant (AID-C), but not AID-wt, induces cell death. Shown are representative confocal images (left panels) of wild-type MEF cells transiently transfected with Flag-tagged AID-wt or Flag-tagged AID-C. At 24 h post-transfection, apoptosis of Flag-positive cells (red) was scored for TUNEL positivity (green) and DAPI staining (blue). The chart (right panel) shows the proportion of apoptotic cells 24 h after transfection. Data were collected from four independent experiments by scoring at least 500 cells/experiment. NT, nontransfected cells. B, Expression of C-terminal-deleted mutant AID (AID-C), but not AID-wt, induces the formation of nuclear H2AX foci. Shown are representative confocal images (left panels) of wild-type MEF cells transiently transfected with Flag-AID-wt or Flag-AID-C, immunostained with anti-Flag (red) and anti-H2AX (green) Abs and counterstained with DAPI (blue). The chart (right panel) shows the proportion of nontransfected (left panel, white arrows) or transfected (red arrows) cells with large nuclear H2AX foci 24 h after transfection. Data were collected from three independent experiments by scoring at least 150 cells/experiment. C, Expression of AID-wt causes DNA damage foci in repair-deficient cells. Shown are representative confocal images (left panels) of DNA-PKcs –/– or 53BP1–/– MEF cells transiently transfected with Flag-AID-wt, immunostained with anti-Flag (red) and anti-H2AX (green) Abs, and counterstained with DAPI (blue). The chart (right panel) shows the proportion of nontransfected (left panels, white arrows) or transfected (red arrows) cells with large nuclear H2AX foci 24 h after transfection. Data were obtained from three independent experiments by scoring at least 150 cells/experiment.
AID C-terminal deletion mutant causes heightened DNA damage response
To test whether the expression of AID-C led to a DNA damage response, we stained MEF cells expressing AID-wt or AID-C proteins for phosphorylated H2A histone family member X (H2AX), which binds to DNA double-strand breaks, forming foci (9). Fig. 5B shows that only 1.1% of nontransfected MEF cells (Fig. 5B, white arrows) and 9.4% cells transfected with AID-wt (Fig. 5B, red arrow, upper panel) exhibited detectable nuclear H2AX foci. In contrast, 75.6%, upper panel MEF cells expressing AID-C had massive accumulation of nuclear H2AX foci (Fig. 5B, red arrow, lower panel). Our data are consistent with the concept that recruitment of DNA-PKcs by AID is needed to resolve DNA double-stranded breaks.
DNA-PKcs and 53BP1 are necessary to resolve AID-generated H2AX foci
Absence of H2AX nuclear foci in wild-type MEF cells expressing AID-wt could result from prompt DNA repair or, alternatively, from decreased DNA break formation. To determine the contribution of DNA repair to the lack of DNA damage foci in cells expressing AID-wt, we examined DNA damage foci in repair-deficient cells expressing AID-wt. Fig. 5C shows that expression of AID-wt induced nuclear H2AX foci in 80.5% DNA-PKcs–/– cells, whereas only 9.3% of nontransfected cells scored positive. Similarly, the expression of AID-wt also induced H2AX foci in 64% 53BP1–/– cells (34), whereas only 17.2% nontransfected cells were positive. These results indicate that AID generates DNA double-strand breaks and H2AX foci, which accumulate in the absence of DNA-PKcs or other DNA repair components, such as 53BP1.
AID C-terminal deletion mutant impairs survival of activated B cells
Our results indicating that AID deficient in DNA-PKcs binding causes accumulation of DNA double-strand breaks suggest a mechanism for the selective class-switching defect in some patients with hyperIgM type II syndrome. We tested whether the expression of AID mutants deficient in DNA-PKcs binding due to C-terminal deletion (AID-C) impaired the survival of B cells undergoing class switch recombination. To do this, we transduced LPS-activated B cells obtained from spleens of C57BL/6 mice with retroviral vectors encoding AID-wt, AID-C, or GFP (Fig. 6A). Fig. 6B shows that although 91.6% AID-C-transduced B cells (GFP-positive) were dead on day 3 of LPS culture, only 25.2% of the cells transduced with AID-wt and 19.2% of the cells expressing only GFP died (Fig. 6B). Our results indicate that AID-C causes the death of cells undergoing LPS stimulation. We observed that the surviving AID-C-expressing cells did not class switch, thus confirming the findings of Barreto et al. (26), who showed that AID-C does not promote class switch. Because AID-C transduced B cells undergo cell death, our results indicate that AID-C-associated defective class switch is due to the death of cells undergoing class switch recombination.
FIGURE 6. AID C-terminal deletion mutant (AID-C) induces cell death of LPS-stimulated B cells. A, Schematic representation of the retroviral constructs used to transduce B cells. B, C-terminal-deleted mutant AID (AID-C), but not AID-wt, induces death of LPS-stimulated B cells. Cell death was scored by propidium iodide (PI) staining of LPS-stimulated splenocytes transduced with proviruses encoding GFP, AID-wt, or C-terminal-deleted AID mutant (AID-C) on day 3. PI-positive cells were identified by flow cytometric analysis. The graph shows the fraction of transduced B cells (GFP-positive) that were PI positive (y-axis). Data were collected from three independent experiments and represent the mean and SD.
C-terminal deletion does not alter AID intracellular localization
McBride et al. (35) and Ito et al. (36) proposed an alternative mechanism for the impaired class switch recombination by AID-C mutants. These authors suggested that deficient class switch is a consequence of deletion of a leucine-rich nuclear export signal causing predominant nuclear localization of AID-C (35, 36). However, Fig. 5, A and B, show that Flag-tagged AID-C such as AID-wt localized predominantly in the cytoplasm of MEF. Our results showing that removal of the nuclear export signal (aa 189–198) in the Flag-tagged AID-C did not alter intracellular localization of the protein indicate that the prodeath effect of AID-C is not due to nuclear accumulation.
Intracellular localization of AID may not be primarily governed by nucleocytoplasmic shuttling. If AID shuttled between cytoplasm and nucleus, then addition of nuclear localization signal should drive AID to the nucleus. Fig. 7, A–F, shows that although tagging with three consecutive copies of SV40 nuclear localization signals (37) efficiently drives GFP into the nucleus (Fig. 7B), AID remains in the cytoplasm (Fig. 7E). These results suggest that AID is actively retained in the cytoplasm. Consistent with this possibility, cytoplasmic localization of AID-wt is resistant to digitonin treatment (Fig. 7F) that effectively depletes GFP from the cytoplasm of the cells (Fig. 7C). Our data indicate that AID is retained in the cytoplasm by digitonin-resistant cytoskeletal elements. One of these elements is -tubulin, because -tubulin specifically coprecipitated (Figs. 1 and 7G) and partially colocalized (Fig. 7, H–J) with AID.
FIGURE 7. AID is retained in the cytoplasm by cytoskeletal proteins. GFP distributes in the cytoplasm and nucleus of HeLa cells (A), whereas Flag-AID localizes predominantly to the cytoplasm. Addition of three SV40 nuclear localization signals to GFP (3xNLS-GFP) or to Flag-tagged AID (3xNLS-AID) causes nuclear localization of GFP (B), but not Flag-tagged AID (E). Digitonin treatment of cells removes GFP (C), but not Flag-tagged AID (F), from the cytoplasm, indicating cytoplasmic retention of AID. G, AID associates with -tubulin. Shown are Western blot (WB) analyses of anti-Flag immunoprecipitates (IP) of cell lysates obtained from HeLa, HEK293, or 70Z/3 cells transfected with Flag-AID (+) or empty vector (–). Blots were probed with either anti-Flag or anti--tubulin Abs, as indicated. The mouse Ig H chain, migrating slight above -tubulin, is seen as a weak band in the controls. H–J, AID partially colocalized with -tubulin. Shown are confocal images of HeLa cells expressing Flag-AID double stained with Flag (red)and anti--tubulin (green) Abs. Arrowheads show colocalization of AID and -tubulin.
It is possible that the properties of AID-GFP fusion proteins studied by McBride et al. (35) and Ito et al. (36) differ from the properties of Flag-tagged AID or native AID. Although the intracellular distribution of Flag-tagged AID-C is determined by cytoplasmic retention, the distribution of AID-C-GFP fusion proteins is not. Instead, intracellular distribution of AID-C-GFP fusion proteins is determined by nucleocytoplasminic shuttling.
Discussion
We show in this study that AID promotes cell survival by recruiting DNA-PKcs to the DNA, hence resolving double-strand breaks. Our findings explain how B cells survive while undergoing DNA double-strand breaks during class switch recombination. Consistent with this idea, compromised class switch recombination is a common phenotype in mice deficient in proteins necessary for DNA damage repair, such as H2AX (30), Ataxia Telangectasia Muntant (38), 53BP1 (39), and Mre11 (40).
Gell and Jackson (41) showed that DNA-PKcs associates with Ku70/Ku80 binding to the 12-aa C-terminal tail of the Ku80 subunit. Our data show that DNA-PKcs binding to AID requires the C terminus of AID (Fig. 3, A and B). Such a striking binding parallel made us wonder whether the C-terminal domain of Ku80 might be homologous in any way to the C-terminal domain of AID. Sequence alignment of the 14-aa C-terminal domains of AID and Ku80 revealed that both sequences possess a common (E/D)VDDL(X)D motif (Fig. 4D). The common motif in the C-terminal domains of AID and Ku80 suggest competition for the same binding site on DNA-PKcs. This mechanism may coordinate the formation of AID/DNA-PKcs and Ku70/Ku80/DNA-PKcs complexes for repair of double-strand breaks by nonhomologous end joining (NHEJ) after cytidine deamination of DNA.
We propose a working model to explain how AID promotes the survival of cells undergoing DNA double-strand breaks. AID binding to DNA through the DNA deamination domain undergoes a conformational change resulting in exposure of the C-terminal domain that, in turn, recruits DNA-PKcs to the DNA. Upon the generation of DNA breaks by cytidine deamination, DNA-PKcs initiates NHEJ by dissociating from AID and reassociating with Ku80 to assemble the NHEJ repair complex. Our results thus explain previous observations by Barreto et al. (26) and Ta et al. (27), who showed that C-terminal-deleted AID mutants fail to promote class switch recombination. Our results suggest that the mechanism underlying hyperIgM type II syndrome associated with AID mutations that truncate the C-terminal portion of the molecule is the selective death of B cells stimulated to undergo class switch recombination because of accumulation of DNA double-strand breaks. That C-terminal-deleted AID mutants may sustain somatic hypermutation (27) is also in agreement with our findings, because somatic hypermutation is less dependent on the recruitment of DNA-PKcs than class switch recombination, possibly due to engagement of alternative repair mechanisms (42, 43). Our results may also explain why lack of DNA-PKcs causes class switch deficiency for all isotypes except IgG1 (21), whereas the SCID mutation, which truncates DNA-PKcs, allows moderate class switching to all isotypes (44). Although AID in the absence of DNA-PKcs is unable to recruit repair factors to the sites of cytidine deamination, it may do so in the presence of DNA-PKcs with the SCID mutation.
Acknowledgments
We thank Drs. Zhenkun Lou and Junjie Chen for kindly providing us with 53BP1–/– MEF cells and Abs, Dr. Frederich W. Alt for generously providing us with rabbit anti-AID antiserum, and Dr. David Chen for kindly providing us with DNA-PKcs-deficient cell lines. We also thank Dr. Cristina Jo?o for valuable discussion, and Michelle Rebrovich for excellent technical assistance.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grants AI48602 (to M.C.) and HL46810 and HL52297 (to J.L.P.).
2 Address correspondence and reprint requests to Dr. Marilia Cascalho, Transplantation Biology, Mayo Clinic, 200 First Street SW, Medical Sciences 2-113, Rochester, MN 55905. E-mail address: cascalho.marilia@mayo.edu
3 Abbreviations used in this paper: AID, activation-induced cytidine deaminase; CSR, class switch recombination; DAPI, 4',6-diamido-2-phenylindole hydrochloride; DNA-PKcs, DNA-dependent protein kinase catalytic subunit; EGFP, enhanced GFP; MEF, mouse embryonic fibroblast; NHEJ, nonhomologous end joining; SHM, somatic hypermutation; STDDP, 3,3'-dithiobis-(sulfosuccinimidylpropionate).
Received for publication August 25, 2004. Accepted for publication October 25, 2004.
References
Durandy, A.. 2003. Activation-induced cytidine deaminase: a dual role in class-switch recombination and somatic hypermutation. Eur. J. Immunol. 33:2069.
Muramatsu, M., V. S. Sankaranand, S. Anant, M. Sugai, K. Kinoshita, N. O. Davidson, T. Honjo. 1999. 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. 274:18470.
Begum, N. A., K. Kinoshita, M. Muramatsu, H. Nagaoka, R. Shinkura, T. Honjo. 2004. De novo protein synthesis is required for activation-induced cytidine deamination-dependent DNA cleavage in immunoglobulin class switch recombination. Proc. Natl. Acad. Sci. USA 101:13003.
Petersen-Mahrt, S. K., R. S. Harris, M. S. Neuberger. 2002. AID mutates E. coli suggesting a DNA deamination mechanism for antibody diversification. Nature 418:99.
Dickerson, S. K., E. Market, E. Besmer, F. N. Papavasiliou. 2003. AID mediates hypermutation by deaminating single stranded DNA. J. Exp. Med. 197:1291
Pham, P., R. Bransteitter, J. Petruska, M. F. Goodman. 2003. Processive AID-catalysed cytosine deamination on single-stranded DNA simulates somatic hypermutation. Nature 424:103.
Chaudhuri, J., M. Tian, C. Khuong, K. Chua, E. Pinaud, F. W. Alt. 2003. Transcription-targeted DNA deamination by the AID antibody diversification enzyme. Nature 422:726.
Papavasiliou, F. N., D. G. Schatz. 2000. Cell-cycle-regulated DNA double-stranded breaks in somatic hypermutation of immunoglobulin genes. Nature 408:216.
Petersen, S., R. Casellas, B. Reina-San-Martin, H. T. Chen, M. J. Difilippantonio, P. C. Wilson, L. Hanitsch, A. Celeste, M. Muramatsu, D. R. Pilch, et al 2001. AID is required to initiate Nbs1/-H2AX focus formation and mutations at sites of class switching. Nature 414:660.
Zhang, K.. 2003. Accessibility control and machinery of immunoglobulin class switch recombination. J. Leukocyte Biol. 73:323.(Xiaosheng Wu, Pedro Geral)