Portal de Periódicos da CAPES

Identification of a Specific Domain Required for Dimerization of Activation-induced Cytidine Deaminase

2006; Elsevier BV; Volume: 281; Issue: 28 Linguagem: Inglês

10.1074/jbc.m601645200

1083-351X

Jishu Wang, Reiko Shinkura, Masamichi Muramatsu, Hitoshi Nagaoka, Kazuo Kinoshita, Tasuku Honjo,

Cytomegalovirus and herpesvirus research

Activation-induced cytidine deaminase (AID) is essential to all three genetic alterations required for generation of antigen-specific immunoglobulin: class switch recombination, somatic hypermutation, and gene conversion. Here we demonstrate that AID molecules form a homodimer autonomously in the absence of RNA, DNA, other cofactors, or post-translational modifications. Studies on serial deletion mutants revealed the minimum region between Thr27 and His56 responsible for dimerization. Analyses of point mutations within this region revealed that the residues between Gly47 and Gly54 are most important for the dimer formation. Functional analyses of these mutations indicate that all mutations impairing the dimer formation are inefficient for class switching, suggesting that dimer formation is required for class switching activity. Dimer formation and its requirement for the function of AID are features that AID shares with APOBEC-1, an RNA editing enzyme of apolipoprotein B100 mRNA. Activation-induced cytidine deaminase (AID) is essential to all three genetic alterations required for generation of antigen-specific immunoglobulin: class switch recombination, somatic hypermutation, and gene conversion. Here we demonstrate that AID molecules form a homodimer autonomously in the absence of RNA, DNA, other cofactors, or post-translational modifications. Studies on serial deletion mutants revealed the minimum region between Thr27 and His56 responsible for dimerization. Analyses of point mutations within this region revealed that the residues between Gly47 and Gly54 are most important for the dimer formation. Functional analyses of these mutations indicate that all mutations impairing the dimer formation are inefficient for class switching, suggesting that dimer formation is required for class switching activity. Dimer formation and its requirement for the function of AID are features that AID shares with APOBEC-1, an RNA editing enzyme of apolipoprotein B100 mRNA. Withdrawal: Identification of a specific domain required for dimerization of activation-induced cytidine deaminase. VOLUME 281 (2006) PAGES 19115-19123Journal of Biological ChemistryVol. 283Issue 1PreviewThis article has been withdrawn by the authors. Full-Text PDF Open Access Activation-induced cytidine deaminase (AID) 2The abbreviations used are: AID, activation-induced cytidine deaminase; mAID, mouse AID; hAID, human AID; CSR, class switch recombination; SHM, somatic hypermutation; apo, apolipoprotein; SUMO, small ubiquitin-related modifier; 293T, human embryonic kidney 293 with large T antigen expression; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate; ER, estrogen receptor; GFP, green fluorescent protein; NTA, nitrilotriacetic acid; mAb, monoclonal antibody; IRES, internal ribosome entry site. is essential to class switch recombination (CSR) and somatic hypermutation (SHM) in human and mice (1Muramatsu M. Kinoshita K. Fagarasan S. Yamada S. Shinkai Y. Honjo H. Cell. 2000; 102: 553-563Abstract Full Text Full Text PDF PubMed Scopus (2701) Google Scholar, 2Revy P. Muto T. Levy Y. Geissmann F. Plebani A. Sanal O. Catalan N. Forveille M. Dufourcq-Labelouse R. Gennery A. Tezcan I. Ersoy F. Kayserili H. Ugazio A.G. Brousse N. Muramatsu M. Notarangelo L.D. Kinoshita K. Honjo T. Fischer A. Durandy A. Cell. 2000; 102: 565-575Abstract Full Text Full Text PDF PubMed Scopus (1338) Google Scholar) and gene conversion in chickens (3Arakawa H. Hauschild J. Buerstedde J.M. Science. 2002; 295: 1301-1306Crossref PubMed Scopus (385) Google Scholar, 4Harris R.S. Sale J.E. Petersen-Mahrt S.K. Neuberger M.S. Curr. Biol. 2002; 12: 435-438Abstract Full Text Full Text PDF PubMed Scopus (195) Google Scholar), the three genetic alterations that occur in mature B cells. Expression of AID is not only essential but also sufficient to the three events described above, although they are different in their mechanisms as well as reaction products (5Honjo T. Muramatsu M. Fagarasan S. Immunity. 2004; 20: 659-668Abstract Full Text Full Text PDF PubMed Scopus (174) Google Scholar, 6Yoshikawa K. Okazaki I.M. Eto T. Kinoshita K. Muramatsu M. Nagaoka H. Honjo T. Science. 2002; 296: 2033-2036Crossref PubMed Scopus (327) Google Scholar, 7Okazaki I.M. Kinoshita K. Muramatsu M. Yoshikawa K. Honjo T. Nature. 2002; 416: 340-345Crossref PubMed Scopus (229) Google Scholar, 8Martin A. Bardwell P.D. Woo C.J. Fan M. Shulman M.J. Scharff M.D. Nature. 2002; 415: 802-806Crossref PubMed Scopus (229) Google Scholar). AID is thus the key molecule for generation of the antigen-specific immunoglobulin (Ig) with various isotypes. CSR takes place between two switch (S) regions located upstream of each constant region of the Ig heavy chain (CH) gene except for Cδ. CSR causes intrachromosomal deletion of the intervening DNA segment including C genes, resulting in replacement of the C regions and alteration of effector functions of Ig. SHM generates innumerable variations in the antigen-binding site of Ig by a high rate of mutagenesis focused on the Ig V region. In gene conversion, the V region is diversified by homologous recombination utilizing pseudo-V genes as template. AID has been shown to be involved in the DNA cleavage step of CSR and SHM, but its detailed molecular mechanism is elusive and intensively debated (9Petersen-Mahrt S.K. Harris R.S. Neuberger M.S. Nature. 2002; 418: 99-103Crossref PubMed Scopus (741) Google Scholar, 10Durandy A. Eur. J. Immunol. 2003; 33: 2069-7203Crossref PubMed Scopus (60) Google Scholar, 11Barreto V.M. Ramiro A.R. Nussenzweig M.C. Trends Immunol. 2005; 26: 90-96Abstract Full Text Full Text PDF PubMed Scopus (50) Google Scholar, 12Honjo T. Nagaoka H. Shinkura R. Muramatsu M. Nat. Immunol. 2005; 6: 655-661Crossref PubMed Scopus (89) Google Scholar). The AID protein contains 198 amino acid residues with several structural and functional domains. It is proven that the N terminus and C terminus of AID are responsible for SHM and CSR, respectively, because mutations in the N terminus of AID resulted in severe reduction of SHM, whereas CSR is intact (13Barreto V. Reina-San-Martin B. Ramiro A.R. McBride K.M. Nussenzweig M.C. Mol. Cell. 2003; 12: 501-508Abstract Full Text Full Text PDF PubMed Scopus (230) Google Scholar, 14Shinkura R. Ito S. Begum N.A. Nagaoka H. Muramatsu M. Kinoshita K. Sakakibara Y. Hijikata H. Honjo T. Nat. Immunol. 2004; 5: 707-712Crossref PubMed Scopus (191) Google Scholar) and vice versa (15Ta V.T. Nagaoka H. Catalan N. Durandy A. Fischer A. Imai K. Nonoyama S. Tashiro J. Ikegawa M. Ito S. Kinoshita K. Muramatsu M. Honjo T. Nat. Immunol. 2003; 4: 843-848Crossref PubMed Scopus (281) Google Scholar). Among the functionally characterized genes, the most homologous to AID is an RNA-editing enzyme, APOBEC-1, the catalytic subunit of the apolipoprotein (apo) B100 mRNA-editing enzyme (16Teng B. Burant C.F. Davidson N.O. Science. 1993; 260: 1816-1819Crossref PubMed Scopus (496) Google Scholar, 17Navaratnam N. Morrison J.R. Bhattacharya S. Patel D. Funahashi T. Giannoni F. Teng B.B. Davidson N.O. Scott J. J. Biol. Chem. 1993; 268: 20709-20712Abstract Full Text PDF PubMed Google Scholar), which deaminates cytidine to uridine at 6666 of apoB100 mRNA, generating apoB48 mRNA by introducing the UAA stop codon in place of the CAA glutamine codon (18Navaratnam N. Bhattacharya S. Fujino T. Patel D. Jarmus A.L. Scott J. Cell. 1995; 81: 187-195Abstract Full Text PDF PubMed Scopus (157) Google Scholar). Translation of apoB100 and apoB48 mRNAs produces the protein components of low density lipoprotein and chylomicron, respectively. Genetically the genes encoding AID and APOBEC-1 are located in a close proximity on the same chromosome of human and mouse (19Muto T. Muramatsu M. Taniwaki M. Kinoshita K. Honjo T. Genomics. 2000; 68: 85-88Crossref PubMed Scopus (126) Google Scholar, 20Conticello S.G. Thomas C.J. Petersen-Mahrt S.K. Neuberger M.S. Mol. Biol. Evol. 2005; 22: 367-377Crossref PubMed Scopus (394) Google Scholar), implying that they are derived by a recent duplication event. Functionally both proteins possess a nuclear localization signal and a nuclear export signal at their N terminus and C terminus, respectively (21Ito S. Nagaoka H. Shinkura R. Begum N. Muramatsu M. Nakata M. Honjo T. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 1975-19805Crossref PubMed Scopus (251) Google Scholar, 22McBride K.M. Barreto V. Ramiro A.R. Stavropoulos P. Nussenzweig M.C. J. Exp. Med. 2004; 199: 1235-1244Crossref PubMed Scopus (192) Google Scholar), which endue them with shuttling between the nucleus and cytoplasm, a requisite of the RNA-editing enzymes. Similar three-dimensional structures of AID and APOBEC-1 were predicted based on the crystallographic analyses of yeast (Saccharomyces cerevisiae) cytosine deaminase D1 (CDD1) (23Xie K. Sowden M.P. Dance G.S. Torelli A.T. Smith H.C. Wedekind J.E. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 8114-8119Crossref PubMed Scopus (81) Google Scholar). APOBEC-1 has been shown to function as a dimer (24Navaratnam N. Fujino T. Bayliss J. Jarmuz A. How A. Richardson N. Somasekaram A. Bhattacharya S. Carter C. Scott J. J. Mol. Biol. 1998; 275: 695-714Crossref PubMed Scopus (126) Google Scholar, 25Teng B.B. Ochsner S. Zhang Q. Soman K.V. Lau P.P. Chan L. J. Lipid Res. 1999; 40: 623-635Abstract Full Text Full Text PDF PubMed Google Scholar). The dimerization of APOBEC-1 creates an active structure that is essential for its RNA binding and deamination activity of apoB mRNA (25Teng B.B. Ochsner S. Zhang Q. Soman K.V. Lau P.P. Chan L. J. Lipid Res. 1999; 40: 623-635Abstract Full Text Full Text PDF PubMed Google Scholar). Mutations abolishing dimerization of APOBEC-1 also destroy its RNA binding and editing activities (24Navaratnam N. Fujino T. Bayliss J. Jarmuz A. How A. Richardson N. Somasekaram A. Bhattacharya S. Carter C. Scott J. J. Mol. Biol. 1998; 275: 695-714Crossref PubMed Scopus (126) Google Scholar), although the dimerization motif, the RNA-binding region, and the catalytic site do not completely overlap in APOBEC-1. Deletion of either 7 residues from the N terminus or 5 residues from the C terminus disrupts the dimer formation of APOBEC-1 (24Navaratnam N. Fujino T. Bayliss J. Jarmuz A. How A. Richardson N. Somasekaram A. Bhattacharya S. Carter C. Scott J. J. Mol. Biol. 1998; 275: 695-714Crossref PubMed Scopus (126) Google Scholar), suggesting that both N terminus and C terminus might be vital for the dimerization of APOBEC-1. On the other hand, Teng et al. (25Teng B.B. Ochsner S. Zhang Q. Soman K.V. Lau P.P. Chan L. J. Lipid Res. 1999; 40: 623-635Abstract Full Text Full Text PDF PubMed Google Scholar) reported that two C-terminal regions (from Leu196 to Leu210 and from Leu221 to Lys229) of APOBEC-1 are equally critical for dimerization. AID is 31 residues shorter than APOBEC-1 with 9 and 24 residues missing in the N terminus and C terminus, respectively, that were reported to be critical for the dimerization in APOBEC-1 (24Navaratnam N. Fujino T. Bayliss J. Jarmuz A. How A. Richardson N. Somasekaram A. Bhattacharya S. Carter C. Scott J. J. Mol. Biol. 1998; 275: 695-714Crossref PubMed Scopus (126) Google Scholar, 25Teng B.B. Ochsner S. Zhang Q. Soman K.V. Lau P.P. Chan L. J. Lipid Res. 1999; 40: 623-635Abstract Full Text Full Text PDF PubMed Google Scholar). However, the C-terminal deletion mutants of AID, known as JP8B, P20, and JP41, still maintain the SHM activity (15Ta V.T. Nagaoka H. Catalan N. Durandy A. Fischer A. Imai K. Nonoyama S. Tashiro J. Ikegawa M. Ito S. Kinoshita K. Muramatsu M. Honjo T. Nat. Immunol. 2003; 4: 843-848Crossref PubMed Scopus (281) Google Scholar). Moreover the regions of APOBEC-1 (Leu196–Leu210 and Leu221–Lys229) that were reported to be important for its dimerization (25Teng B.B. Ochsner S. Zhang Q. Soman K.V. Lau P.P. Chan L. J. Lipid Res. 1999; 40: 623-635Abstract Full Text Full Text PDF PubMed Google Scholar) are not conserved in AID. Although AID has been shown to form a multimeric complex (15Ta V.T. Nagaoka H. Catalan N. Durandy A. Fischer A. Imai K. Nonoyama S. Tashiro J. Ikegawa M. Ito S. Kinoshita K. Muramatsu M. Honjo T. Nat. Immunol. 2003; 4: 843-848Crossref PubMed Scopus (281) Google Scholar), the exact nature of the AID multimer is not known. Therefore, it is important to examine whether AID exists as a dimer and if so to identify the motif(s) responsible for its dimerization. To answer these questions, we constructed AID and its mutants with different tags and co-expressed and immunoprecipitated them by different antibodies. These studies clearly showed that the dimer is the major species of AID multimer in cells, and its formation is dependent on the residues between Gly47 and Gly54. Furthermore we demonstrated that the AID mutations that affect dimerization also impair CSR activity, indicating that the dimer formation is required for AID function. Constructs—Deletion and truncation mutants of AID were produced by PCR with wild-type AID cDNA as a template. Mouse AID was used except for the experiments in Fig. 3 in which truncation and deletion mutants of human (h) AID were used. For N-terminal truncation constructs of hAID, a Kozak sequence was added to the 5′-end by PCR to facilitate protein expression. The amplified fragments were digested with EcoRI and BamHI and inserted into pEGFP-N1 (Clontech) restricted with the same enzymes, generating vectors expressing mutant hAID fused with GFP at the C terminus. Alternatively the mutant hAID cDNAs were inserted into pCMV-Flag-ER (15Ta V.T. Nagaoka H. Catalan N. Durandy A. Fischer A. Imai K. Nonoyama S. Tashiro J. Ikegawa M. Ito S. Kinoshita K. Muramatsu M. Honjo T. Nat. Immunol. 2003; 4: 843-848Crossref PubMed Scopus (281) Google Scholar) to generate mutant hAID fused with FLAG-ER at C terminus. The internal deletion mutant (hAIDdel26–80) was generated by ligating PCR-amplified N- and C-terminal parts of AID with a GGSGG linker (5′-ggaggtagcggaggt-3′). Full-length AID with different tags at C termini were generated by PCR and were cloned into pcDNA5 (Invitrogen) to construct expression vectors. The AID-Myc and AID-Myc·His contained 33 and 21 extra residues, which were encoded by the sequences aatggagaacagaaattgatcagtgaggaagacctcaacggtgagcagaagttaatatccgaggaggatcttaatagttgtctagagggccctattcta (NGEQKLISEEDLNGEQKLISEEDLNSCLEGPIL) and aatggagaacagaaattgatcagtgaggaagacctcaacggtgagcagaagttaatatccgag (NGEQKLISEEDLNGEQKLISE), respectively. All constructs were confirmed by DNA sequencing. Mutants with amino acid replacements were generated by PCR. Cell Culture and Transfection—The 293T cells were maintained in Dulbecco's modified Eagle's medium (Sigma) containing 10% fetal calf serum, 2 mm l-glutamine, 100 units/ml penicillin, and 100 mg/ml streptomycin. Cells were transfected with plasmid DNA using liposomes (Lipofectamine 2000, Invitrogen) according to the manufacturer's instructions. Immunoprecipitation and Western Blotting—Cells (293T) were lysed 60 h after transfection in an ice-cold lysis buffer (1% Nonidet P-40, 150 mm NaCl, 50 mm Tris-HCl, pH 7.5). Cell lysates were centrifuged at 12,000 × g at 4 °C. For immunoprecipitation, cell lysates with an equal amount of total proteins were incubated with antibodies (Sigma, anti-c-Myc-agarose and anti-FLAG M2 affinity gel). After washing with the lysis buffer, the co-precipitated proteins were fractionated on an SDS, 12% polyacrylamide gel and transferred to the nitrocellulose membrane. The membrane was probed with antibodies as indicated in Fig. 1 and visualized using the ECL reagents (Amersham Biosciences). Anti-AID antibodies were described previously (14Shinkura R. Ito S. Begum N.A. Nagaoka H. Muramatsu M. Kinoshita K. Sakakibara Y. Hijikata H. Honjo T. Nat. Immunol. 2004; 5: 707-712Crossref PubMed Scopus (191) Google Scholar, 15Ta V.T. Nagaoka H. Catalan N. Durandy A. Fischer A. Imai K. Nonoyama S. Tashiro J. Ikegawa M. Ito S. Kinoshita K. Muramatsu M. Honjo T. Nat. Immunol. 2003; 4: 843-848Crossref PubMed Scopus (281) Google Scholar). Purification of SUMO-AID—Mouse AID cDNA was cloned into pSUMO vector (Life Sensors, Malvern, PA) to generate an in-frame fusion to the C terminus of SUMO, and the construct was transformed into BL21-CodonPlus™ (Stratagene). A 1-liter culture was induced with 0.1 mm isopropyl β-d-thiogalactoside and incubated for 24 h at 25 °C. The bacteria were lysed and sonicated in the lysis buffer (40 mm Tris, pH 8.0, 40 mm KCl, 50 mm NaCl, 10% glycerol, 1% CHAPS). The lysate was cleared by centrifugation at 100,000 × g for 1 h and followed by incubation with Ni2+-NTA beads (Qiagen) with 20 mm imidazole for 2 h at 4°C. The beads were washed three times with the lysis buffer containing 80 mm imidazole and 500 mm NaCl, and bound protein was eluted twice in 2.5 ml of elution buffer (40 mm KCl, 50 mm NaCl, 10% glycerol, 1% CHAPS, 500 mm NaCl, 500 mm imidazole) at 4 °C. Eluted proteins were concentrated and dialyzed in 45 mm Tris, pH 7.4, and the purity of SUMO-AID proteins was more than 90% by Coomassie Bluestained PAGE analysis. Mice—AID-deficient mice on a C57BL/6 background were maintained in our animal facility and were used at 2–3 months of age (1Muramatsu M. Kinoshita K. Fagarasan S. Yamada S. Shinkai Y. Honjo H. Cell. 2000; 102: 553-563Abstract Full Text Full Text PDF PubMed Scopus (2701) Google Scholar). All mouse protocols were approved by the Institute of Laboratory Animals, Faculty of Medicine, Kyoto University (Kyoto, Japan). Bacterial Two-hybrid Assay—Mutant and wild-type AIDs were subcloned into pBT and pTRG vectors, respectively (BacterioMatch II two-hybrid system, Invitrogen). pBT- and pTRG-derived constructs were co-transformed into reporter competent cells following the instruction manual. Briefly 1 μg of pBT and 1 μg of pTRG construct transformed 100 μl of competent cells; after heat shock the 100 μl of competent cells were diluted into 5 ml of SOC culture medium (0.5% yeast extract, 2% tryptone, 10 mm NaCl, 2.5 mm KCL, 10 mm MgCl2, 20mm MgSO4, 20 mm glucose). After 2-h induction with isopropyl β-d-thiogalactoside, 5 μl of 5 ml were spread onto the non-selective Petri dish to calculate transformation efficiency. The rest of the cells were spun down and spread onto the selective Petri dish containing 5 mm 3-amino-1,2,4-triazole. The culture Petri dishes were incubated at 30 °C for 2.5 days, and colonies were counted. In Vitro Assays for CSR—For retrovirus infection, cDNAs of mouse wild-type AID and mutants were inserted into pMSCV-IRES-GFP (14Shinkura R. Ito S. Begum N.A. Nagaoka H. Muramatsu M. Kinoshita K. Sakakibara Y. Hijikata H. Honjo T. Nat. Immunol. 2004; 5: 707-712Crossref PubMed Scopus (191) Google Scholar). The preparation and infection of retroviruses were described previously (1Muramatsu M. Kinoshita K. Fagarasan S. Yamada S. Shinkai Y. Honjo H. Cell. 2000; 102: 553-563Abstract Full Text Full Text PDF PubMed Scopus (2701) Google Scholar, 14Shinkura R. Ito S. Begum N.A. Nagaoka H. Muramatsu M. Kinoshita K. Sakakibara Y. Hijikata H. Honjo T. Nat. Immunol. 2004; 5: 707-712Crossref PubMed Scopus (191) Google Scholar, 15Ta V.T. Nagaoka H. Catalan N. Durandy A. Fischer A. Imai K. Nonoyama S. Tashiro J. Ikegawa M. Ito S. Kinoshita K. Muramatsu M. Honjo T. Nat. Immunol. 2003; 4: 843-848Crossref PubMed Scopus (281) Google Scholar). For CSR, stimulated and infected AID-deficient spleen B cells positive for GFP were analyzed for surface IgG1 expression by flow cytometry 3 days after infection. Cells were stained with biotinylated antibody to IgG1 (anti-IgG1, Pharmingen) followed by incubation with allophycocyanin-labeled streptavidin. The same population of cells was analyzed by Western blot with anti-AID antibodies (14Shinkura R. Ito S. Begum N.A. Nagaoka H. Muramatsu M. Kinoshita K. Sakakibara Y. Hijikata H. Honjo T. Nat. Immunol. 2004; 5: 707-712Crossref PubMed Scopus (191) Google Scholar). Autologous Multimerization of AID—Although the size exclusion chromatography and glycerol gradient sedimentation revealed the presence of AID in a complex larger than 200 kDa, 3J. Wang, R. Shinkura, M. Muramatsu, H. Nagaoka, K. Kinoshita, and T. Honjo, unpublished data. it is not clear whether AID itself forms multimers or many other cellular components are associated with AID. It has been reported that AID associates with nucleic acids in various cells (26Chaudhuri J. Tian M. Khuong C. Chua K. Pinaud E. Alt F.W. Nature. 2003; 422: 726-730Crossref PubMed Scopus (613) Google Scholar, 27Nambu Y. Sugai M. Gonda H. Lee C.G. Katakai T. Agata Y. Yokota Y. Shimizu A. Science. 2003; 302: 2137-2140Crossref PubMed Scopus (222) Google Scholar). We therefore first examined whether AID forms a stable multimeric complex in the absence of nucleic acids. For this purpose, we used FLAG- and Myc-tagged mouse AID and expressed them in 293T cells by transfection of an equal amount of two plasmids. An additional 33 irrelevant residues were fused to the Myc epitope to distinguish AID-FLAG and AID-Myc by their molecular sizes. Cell lysates were immunoprecipitated with anti-FLAG monoclonal antibody (mAb)-conjugated agarose (M2 beads) and then analyzed by Western blot with specific antibodies to AID. Both FLAG- and Myc-tagged AIDs were detected in the precipitates in approximately similar amounts in agreement with the previous report (15Ta V.T. Nagaoka H. Catalan N. Durandy A. Fischer A. Imai K. Nonoyama S. Tashiro J. Ikegawa M. Ito S. Kinoshita K. Muramatsu M. Honjo T. Nat. Immunol. 2003; 4: 843-848Crossref PubMed Scopus (281) Google Scholar). The multimeric complex formed in vivo was resistant to the treatment with RNase, DNase, or EDTA before immunoprecipitation (Fig. 1A). Neither AID itself nor Myc-tagged AID can interact with M2 beads directly (data not shown). The heteromeric multimer formation of AID was confirmed by reciprocal experiments in which two recombinant AID proteins were immunoprecipitated with anti-Myc mAb and then detected by anti-AID antibodies (data not shown). When we combined separately expressed AID-Myc and AID-FLAG proteins, an almost equal amount of two AID monomers was precipitated by the anti-FLAG mAb indicating that the monomers in the AID multimeric complex can be exchanged (Fig. 1B). We then treated both cellular extracts with RNase or DNase before mixing and examined the heteromultimer formation in vitro. As shown in Fig. 1C, the treatment with neither RNase nor DNase inhibited the multimer formation of AID derived from different extracts. These results suggest that the AID multimer formation is independent of binding to nucleic acids. To further confirm dispensability of cofactors such as oligonucleotides for the AID multimer formation, we carried out denaturation and renaturation of AID tagged with either FLAG-His or Myc-His, which was expressed separately in 293T cells. Cell lysates were denatured in 8 m urea after RNase A and DNase I treatments, and then the AID recombinant proteins were purified by the Ni2+ affinity resin (Fig. 1D). This treatment would remove all potentially remaining fragments of DNA or RNA from AID. Two denatured and purified recombinant AID proteins were mixed to form the multimeric complex while refolding by gradient dialysis. After 48-h dialysis, co-immunoprecipitation was performed separately or tandemly by the anti-FLAG and anti-Myc mAbs. Obviously dialysis could not refold all AID molecules. The unrefolded proteins lost the ability to associate with their partners because one-step precipitation with either the anti-FLAG or anti-Myc mAb caused a strong bias of two protein ratios in favor of that with the epitope of the mAb used (Fig. 1D, lanes 6 and 7). However, the multimer still formed in vitro by the refolded AID because approximately an equal ratio of two-epitope AID was obtained by tandem immunoprecipitation with anti-FLAG and anti-Myc mAbs (Fig. 1D, lane 8). Thus we concluded that DNA or RNA molecules are not required for AID multimer formation. In addition it is unlikely that AID multimerization depends on other non-covalently associated cofactors because 8 m urea denaturation should strip off such cofactors. To examine whether specific modifications are required for multimerization of AID, the recombinant SUMO-AID purified from Escherichia coli was mixed with AID-FLAG produced in 293T cells, and the complex was immunoprecipitated with anti-FLAG mAb. As shown in Fig. 1E, E. coli SUMO-AID was co-immunoprecipitated with 293T-produced AID-FLAG. Because E. coli is unlikely to have post-translational modifications of AID similar to those in eukaryotic cells, AID multimerization appears to be independent of the post-translational modification. Homodimer Is the Major Species of the AID Multimer—We then used a sequential precipitation approach to determine the exact oligomeric state of AID. AID molecules fused with three different epitopes, FLAG, His, or GFP at the C terminus, were co-expressed in 293T cells using approximately equal amounts of plasmid DNA. The cell lysates were first precipitated by the Ni2+ affinity resin and then by anti-FLAG M2-agarose as depicted in Fig. 2A. Each precipitation step was monitored by Western blot using anti-AID antibodies. After the first precipitation, all three isoforms of AID should be detected if all types of AID can freely interact (Fig. 2A). However, the subsequent precipitation with anti-FLAG mAb would distinguish the dimer from the other oligomeric forms. Only His- and FLAG-tagged AID heterodimer would remain in the precipitates in case of the dimeric form, whereas additional AID-GFP should co-exist in case of the larger oligomers. Precipitation with Ni2+ affinity or anti-FLAG-agarose showed that GFP-, His-, and FLAG-tagged AID can interact with each other equally and randomly to form oligomeric complexes regardless of the tag difference (Fig. 2B). However, only AID-FLAG and AID-His were precipitated after the tandem treatment with the Ni2+ affinity resin and anti-FLAG-agarose, indicating that AID formed predominantly the dimer. AID-GFP in the sequential precipitates was quantitated by densitometry and shown to be less than 1% of AID-FLAG, indicating that the trimer, tetramer, or the other larger oligomeric states are unlikely because the trimer and tetramer formation should co-precipitate AID-GFP in the amount equivalent to 30 and 40% of AID-FLAG, respectively (Fig. 2A). The reciprocal procedure using first anti-FLAG M2-agarose and then Ni2+ affinity resin yielded the same result (data not shown), indicating that our immunoprecipitation scheme did not selectively enrich one form of the tagged AID. Thus we concluded that the dimeric state is the dominant structure of AID in mammalian cells. Specific Region Required for Dimerization—To define the regions that are required for dimerization, we constructed a series of C- and N-terminal deletion mutants of hAID with the GFP tag (Fig. 3A) and co-expressed them with wild-type mouse (m) AID-FLAG in 293T cells. Because hAID and mAID have only a 7-residue difference in 198 residues, this pair should represent homologous interactions; we confirmed this below. All C-terminal deletion mutants of hAID we generated were expressed in comparable amounts, although some of them had a slightly reduced expression compared with the wild type (Fig. 3B). All of them associated with wild-type mAID. However, a C-terminal deletion mutant H56X (where X represents stop codon) associated with the wild-type mAID much more weakly than the other truncation mutants including W80X (Fig. 3B). N-terminal deletion of 30 and 40 residues abolished the expression of hAID (data not shown), whereas deletion of 20 and 26 residues (Δ20 and Δ26, respectively) did not appear to affect stability of hAID drastically (Fig. 3B). Because Δ20 and Δ26 could form the dimer, the N- and C-terminal deletion experiments indicate that the region between residue 27 (Thr27) and residue 79 (Thr79) appears to be required for dimerization. When we measured interaction between the identical mutant with different tags, the results were similar to those of interaction of each mutant AID with wild-type mAID (Fig. 3A, Interaction with itself). To confirm this we tested the truncation mutants Glu26– Trp80, Glu26–His56, and His56–Phe108 for their ability to form the dimer with mAID. Glu26–Trp80 formed the dimer despite its low expression. Glu26–His56 interacted with wild-type AID very weakly but significantly. His56–Phe108 failed to form the dimer. However, truncation experiments cannot exclude the possibility that another alternative region also contributes to dimerization. To exclude this possibility, the residues between Glu26 and Trp80 or between Glu58 and Cys93 were deleted, and the remaining N- and C-terminal regions were ligated again by a GGSGG linker to make AIDdel26–80 and AIDdel58–93, respectively. AIDdel26–80 mutant, although expressed at a level similar to Glu26–Trp80, showed no interaction with the mAID or itself. By contrast AIDdel58–93 interacted with mAID or itself weakly but significantly. These results taken together indicated that the region between Glu26 and His56 is necessary and sufficient for dimerization of AID. We confirmed the ability of dimer formation of mAIDdel26–80 with wild-type mAID (Fig. 4A). We then analyzed many mutants in the region between Glu26 and Leu60 for interaction between mutants and the wild-type AID by co-immunoprecipitation to find out the critical residue(s) for dimerization (Table 1). Each mutant and the wild-type AID were fused to FLAG and Myc epitope tags, respectively, and the mutants were further tagged with GFP to distinguish them from the wild-type molecule by size. Most mutants interacted with the wild-type AID with efficiencies comparable to that between two wild-type AID molecules except for R50A, N51A, K52A, and G54A, which showed a significant (to less than 70%) reduction of the interaction (Fig. 4A and Ta