Possible contribution of DNase γ to immunoglobulin V gene diversification

Immunology Letters 125 (2009) 22–30

Contents lists available at ScienceDirect

Immunology Letters journal homepage: www.elsevier.com/locate/

Possible contribution of DNase ␥ to immunoglobulin V gene diversification Noriaki Okamoto a , Mariko Okamoto a , Shinsuke Araki a,b , Hiroshi Arakawa c , Ryushin Mizuta a , Daisuke Kitamura a,b,∗ a b c

Division of Molecular Biology, Research Institute for Biological Sciences, Tokyo University of Science, 2669 Yamazaki, Noda, Chiba 278-0022, Japan Faculty of Pharmaceutical Science, Tokyo University of Science, Noda, Chiba, Japan Helmholtz Center Munich, German Research Center for Environmental Health, Institute for Molecular Radiobiology, Neuherberg-Munich, Germany

a r t i c l e

i n f o

Article history: Received 21 March 2009 Received in revised form 11 May 2009 Accepted 24 May 2009 Available online 6 June 2009 Keywords: Somatic hypermutation Immunoglobulin variable region Double-strand DNA breaks Endonucleases DNase ␥

a b s t r a c t Somatic hypermutation (SHM) diversifies the rearranged immunoglobulin variable (V) region gene in B cells, contributing to affinity maturation of antibodies. It is believed that SHM is generated either by direct replication or by error-prone repair systems resolving V region DNA lesions caused directly or indirectly by cytidine deaminase AID. In accord with a part of these mechanisms, it was reported that SHM is associated with staggered double-strand DNA breaks (DSBs) occurring in the rearranged V regions. However, endonucleases responsible for the DSBs remain elusive. Here we show that DNase ␥, a member of DNase I family endonucleases, contributes to the generation of SHM including point mutation, and nucleotide insertion and deletion in chicken DT40 B cell line. DNase ␥ also contributes to the generation of staggered DSBs in the rearranged V region. These results raise a possibility that DNase ␥ is involved in the V gene mutation machinery. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Antigen receptors on B lymphocytes are encoded by immunoglobulin (Ig) heavy chain (H) and light chain (L) genes, which is highly diversified through V(D)J rearrangements during B cell development and later by somatic hypermutation (SHM) in the rearranged variable (V) region and class switch recombination (CSR) at mature B cell stage. SHM introduces non-templated single base substitutions and occasional deletions or insertions into rearranged V regions and their immediate 3 -flanking sequences, but not in constant (C) region, of Ig genes at a rate of 10−3 to 10−4 per base per cell generation [1]. Some species such as chickens predominantly use pseudogene-templated gene conversion (GC) to diversify their rearranged V region genes. For all three processes of Ig gene diversification, SHM, GC and CSR, the enzyme activation-induced cytidine deaminase (AID) is essential [2]. AID is expressed specifically in proliferating B cells in the germinal center in the peripheral lymphoid tissues and in the bursa of Fabricius in chicken, where SHM, CSR and GC take place [3,4], and appears to be the sole B-cell specific factor required for these events [5,6]. Two hypotheses for AID action have been postulated [2]. (1) RNA hypothesis: similarly to its paralogue APOBEC-1, AID deaminates C to U in the mRNA precursor encoding crucial factor(s) involved in SHM, GC and CSR, possibly including an

∗ Corresponding author. Tel.: +81 4 7121 4071; fax: +81 4 7121 4079. E-mail address: [email protected] (D. Kitamura). 0165-2478/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.imlet.2009.05.005

endonuclease or its guiding factor, and thereby converts it to an active form. The endonuclease is supposed to cleave DNA in Ig V and Switch (S) regions [6]. The DNA strand breaks in the V region would be resolved by error-prone repair systems to generate SHM. (2) DNA hypothesis: AID deaminates C in single-stranded DNA of transcribed V and S regions, producing U:G mismatches. For the generation of SHM, the U:G mismatches are (a) directly replicated (generating C/G to T/A transition), (b) resolved by mismatch repair (MMR) system involving MutS homologues (MSH2-MSH6), exonuclease 1, and low-fidelity DNA polymerases (such as Pol ␩), the latter generating base substitutions predominantly at A/T near the initiating C/G lesion, or (c) resolved by base-excision repair (BER) system: uracil-DNA glycosylase (UNG) removes uracil, producing an abasic site, which is either replicated by translesion DNA synthesis (TLS) polymerases, or cleaved and repaired by low-fidelity DNA polymerases. Either process generates both transitions and transversions [7–10]. Mechanisms accounting for the deletions and insertions remain elusive [11,12]. Besides accumulating evidences supporting above mechanisms for SHM, it has been reported that SHM is associated with double-stranded DNA breaks (DSBs) occurring in Ig V regions [13–15]. The breaks are locally associated with mutations, coupled to transcription, Ig enhancer-dependent, and present in cells in the late S/G2 phase of the cell cycle [14,15]. Zan et al. demonstrated that DSBs staggered to yield free 5 - and 3 protruding ends specifically occur at rearranged V regions upon induction of SHM depending on AID, although the DSBs with blunt ends occur also in non-rearranged V gene irrespective of SHM even in non-B cells [16]. In addition, Honjo and colleagues reported that

N. Okamoto et al. / Immunology Letters 125 (2009) 22–30

SHM-associated, AID-dependent DSBs at IgH locus, as assessed by ␥-H2AX focus formation, depend on de novo protein synthesis [17]. The staggered DSBs are likely generated when single-strand breaks (SSBs) occur in proximity on both strands. Thus, the SSBs may be induced by a yet unidentified endonuclease, an active form of which is produced depending on AID (RNA hypothesis), or by so-far uncertain endonucleases involved in BER or MMR machinery generating SHM (DNA hypothesis). It remains elusive whether the DSBs are simply a by-product of the SSBs or involved in an unknown pathway that generates mutations. In any case, such DSBs may be responsible for chromosomal translocations and deletion/insertion that are associated with SHM [18]. We previously demonstrated that endonuclease DNase ␥ promotes the generation of staggered DSBs in the V region gene, but not in C region, in Ramos B-lymphoma cell line known to undergo SHM at a low frequency [19]. DNase ␥ (also known as DNase1L3, DHP2 or LS-DNase) is a Ca2+ /Mg2+ -dependent neutral endonuclease belonging to the DNase I family, but uniquely possesses nuclear localization signals [20]. DNase ␥ is highly expressed in lymphoid organs and selectively in germinal center B cells, and its expression markedly increases in B cells upon CD40 stimulation in parallel with AID expression [19]. DNase ␥ is responsible for nucleosomal DNA fragmentation in some instances of apoptosis [21,22], although it preferably cuts a single-strand portion of naked DNA in vitro at physiological ionic strength [23]. Despite its well-defined enzymatic nature, physiological role of DNase ␥ in vivo remains unknown. To clarify the DNase ␥ contribution to Ig V region gene diversity, we have generated and analyzed DNase ␥-deficient DT40 cell lines. DT40 is a chicken bursal B-cell lymphoma line known to undergo AID-dependent GC and SHM in culture [4,24]. We demonstrate here that DNase ␥ activity is correlated with the rate of SHM, including point mutation and nucleotide insertion/deletion, and also with the generation of the staggered DSB in V region gene.

23

2.4. Generation of DNase -defecient DT40 cell line A DNase  genomic fragment of approximately 3.5 kb was amplified with the primers, c-DNase F1 (5 -GGCTCGAGAGATGTTGCTCTTTGCCTTGTTTTCG-3 ) and c-DNase R2 (5 -GCCCGCGGCTGTGAGCGTGTCAGCACCGC-3 ) from DT40 genomic DNA library, cloned into pCR-XL-TOPO vector (Invitrogen) and its nucleotide sequence was verified. With this fragment as a template, 5 and 3 arms of homology in a gene-targeting vector were amplified with primers, c-DNase F1 and c-DNase R1 (5 or c-DNase GCGGATCCGAGGGACTATAGCAAACTCTTTGAC-3 ) F2 (5 -GCGGATCCGGAGATGGCAGTACGGGAGATC-3 ) and c-DNase R3 (5 -GGGAATTCGTCGCTCACCCCCAGAGC-3 ), digested with Xho I and Bam HI, or Bam HI and Eco RI, respectively, and simultaneously cloned in between Xho I and Eco RI sites of pSP73 vector (Promega). Into the resultant construct, a neomycin- or a histidinol-resistance cassette [25] was inserted at a Bam HI site to generate pKO-chDNase␥-neo and pKO-chDNase␥-his, respectively. 1 × 107 DT40 cells were electroporated with 20 ␮g Pvu I-linearized pKO-chDNase␥-neo vector at 250 V and 960 ␮F using Gene Pulser II (BIORAD), cultured for 24 h, then seeded into 96-well plates at 3 × 104 cells/well, and selected with G418 (2 mg/ml). Drugresistant clones were verified for homologous-recombination event by Southern blot analysis using two kinds of probes outside the homologous regions (Fig. 1). 16 out of 23 drug-resistant clones were thus identified as heterozygous mutants. One of them was next transfected with pKO-chDNase␥-his as above, selected with histidiol (1 mg/ml), and homologous-recombination was verified as above. 13 of 24 drug-resistant clones were identified as homozygous mutants (DNase ␥−/− ). An expression vector carrying coding sequences of chicken AID and EGFP separated by IRES, pAidExpressPuro2, was transfected into DT40 or one of the DNase ␥−/− clones by electroporation as described above to generate AID/DT40 and AID/DNase ␥−/− clones.

2. Materials and methods 2.1. Cell culture DT40 and its derivatives were cultured at 40 ◦ C in RPMI-1640 medium (Sigma–Aldrich) supplemented with 10% fetal calf serum, penicillin and streptomycin, 50 ␮M 2-mercaptoethanol and 1% chicken serum (JRH Biosciences). Cos7 cells were cultured at 37 ◦ C in DMEM medium (Sigma–Aldrich) supplemented with 10% fetal calf serum, penicillin and streptomycin. 2.2. Cloning of chicken DNase  BBSRC ChickEST Database (http://www.chick.umist.ac.uk/) was searched using the tblastn program with the amino acid sequence of human DNase ␥, and several chicken EST clones were identified. Based on a consensus sequence among them, the chicken DNase ␥ coding sequence was amplified with the 5 -primer (chDNG5; GGCTCGAGAACCATGTTGCTCTTTGCCTTGTTTTCG) and the 3 primer (chDN-G3; GGGGCGCCGCGGGCACGCTGAGCCCGAGCGGCG) from an oligo-dT-primed cDNA library made from DT40 cells, using high-fidelity KOD plus polymerase (Toyobo). The PCR products were subcloned into pCR2.1 vector (Invitrogen) and sequenced. The DNase ␥ coding sequence was subcloned into an expression vector pcDNA3.1-Myc-His B (Invitrogen) to generate pcDNA-chDNase ␥. 2.3. Nuclease activity gel assay Endonuclease activities present in the extracts from Cos7 cells transfected with pcDNA-chDNase ␥ and DT40-derived cell lines were identified using a DNA-PAGE activity assay adapted to an optimal condition for DNase ␥ as described previously [21].

2.5. RT-PCR Total RNA was extracted from cells using TRI reagent (Sigma–Aldrich). First-strand cDNA was synthesized using SuperScript III Reverse Transcriptase (Invitrogen) in a 20 ␮l reaction buffer. Chicken DNase  cDNA was amplified with the primers c-DNase F1 and c-DNase R3 (see Section 2.4) in the condition: 94 ◦ C for 30 s, 60 ◦ C for 30 s, 72 ◦ C for 30 s, 40 cycles, whereas HPRT with the primers, cHPRT-F (CCATGGCGACTCACAGCCCCTGC) and cHPRT-R (GTGCTTTCATGCTTTGTACTTCTGC), in the condition: 94 ◦ C for 30 s, 59 ◦ C for 30 s, 72 ◦ C for 60 s, 35 cycles. Chicken AID cDNA was amplified as described previously [4]. Chicken GAPDH cDNA was amplified with the primers chGAPDH5 (ATTTGGCCGTATTGGCCGCC) and chGAPDH3 (CATAAGACCCTCCACAATGCC), in the condition: 94 ◦ C for 15 s, 68 ◦ C for 45 s, 25 cycles. 2.6. Quantitative real-time PCR Expression of chicken AID and chicken GAPDH was quantified by real-time PCR using SYBR Premix Ex Taq (TAKARA) in a Smart Cycler II System (TAKARA) using one-tenth of the synthesized cDNA and primers as described above with the following cycling condition: 95 ◦ C for 5 s, 60 ◦ C for 15 s, 72 ◦ C for 20 s, 40 cycles. Fluorescence emitted by the SYBR Green I dye bound to DNA was measured at the end of each PCR cycle. The cycle threshold (Ct), defined from the cycle number at which the reaction begins the exponential phase was used to create the standard curve (Ct values against the serial 10-fold diluted standard concentrations give a linear relationship) and to evaluate AID and GAPDH concentration in the samples. Each value of AID mRNA amount was normalized to that of GAPDH.

24

N. Okamoto et al. / Immunology Letters 125 (2009) 22–30

Fig. 1. DNase ␥ gene-targeting in DT40 chicken B cell line. (A) Nuclease activity gel assay for chicken DNase ␥. The empty vector pcDNA3.1-Myc-His B (mock) and the same vector carrying coding sequence of chicken DNase ␥ (DNase ␥) were transfected into Cos7 cells, the cell lysates were subjected to SDS-PAGE for nuclease activity gel assay. Arrowhead indicates the band corresponding to DNase ␥ protein. (B) Schematic representation of the chicken DNase  allele (top), gene-targeting construct (middle) and the mutated allele (bottom). In the targeting construct, catalytic residues in Exon 5 of the DNase ␥ gene were replaced with Neo or HisD cassette. The length of BamHI fragments in wild-type and the mutated alleles are denoted. Open boxes represent exons, and thick bars probes for Southern blot analysis. (C) Southern blot analysis. Genomic DNA from selected drug-resistant DT40 clones after transfection of the targeting construct was digested with Bam HI and probed with the 5 - and 3 -probes denoted in panel B. (D) RT-PCR analysis of DNase  and HPRT (as a control) mRNAs from each DT40 cell clones. (E) Equivalent expression level of AID-IRES-GFP gene stably transfected in parental and DNase ␥−/− DT40 cells (AID/DT40 and AID/DNase ␥−/− , respectively). Histograms show GFP fluorescence profile analyzed by flow cytometry of the representative stable transfectants as indicated. (F) Quantitative analysis of AID mRNA expression in AID/DT40 and AID/DNase ␥−/− clones (indicated in panel E) by real-time PCR. AID mRNA values relative to those of GAPDH mRNA (as an internal control) are presented, with the value of the clone 1–2 set as 10. (G) Equivalent growth rate in culture of AID-IRES-GFP transfected DT40 clones. Live cell number was counted by trypan blue dye exclusion method at the indicated time points.

2.7. Flow cytometry DT40 cells and their derivatives were stained with PEconjugated goat anti-chicken IgM (Southern Biotechnologies Associates) and analyzed using FACSCalibur and CellQuest software (Becton Dikinson). IgM-negative fraction was sorted by FACSvantage (Becton Dickenson).

each clone, cultured for 4 days, and genomic DNA was extracted from these cells. The rearranged V␭ gene was amplified by KOD plus polymerase using primers, CVLF6: CAGGAGCTCGCGGGGCCGTCACTGATTGCCG and CVLR3: GCGCAAGCTTCCCCAGCCTGCCGCCAAGTCCAAG. The PCR products were cloned into pCR2.1 vector and sequenced. 2.9. Analysis of DNA DSBs by LM-PCR

2.8. Analysis of V region sequences Derivatives of DT40 cells were cultured for 1.5 months after transfection and cloning, then sIgM− cells were sorted from

A BW linker was made by annealing oligonucleotides BW1 and BW2 (BW1: GCGGTGACCCGGGAGATCTGAATTC; BW2: GAATTCAGATC). Genomic DNA was prepared using an anion-exchange

N. Okamoto et al. / Immunology Letters 125 (2009) 22–30

resin (QIAGEN Genomic-tips, Qiagen Sciences), incubated with or without T4 polymerase (Toyobo) in the presence of 200 ␮M each dNTP at 37 ◦ C for 20 min, and then incubated with a BW linker and T4 DNA ligase. The DNA was serially 2-fold diluted with untreated genomic DNA and used as a template for LMPCR, or serially diluted with water and used for PCR of C1 region of Ig␭ gene. For the LM-PCR, the diluted samples were amplified (10 cycles with a 50 ◦ C annealing temperature) with BW1 and a V␭ -specific primer (CVLR3: GCGCAAGCTTCCCCAGCCTGCCGCCAAGTCCAAG). One-25th of the first round reaction was used for the second PCR (30 cycles with a 60 ◦ C annealing temperature) with the same BW1 primer and a V␭ -specific primer nested to the CVLR3 (CHLVR1: CCAAATCACCAAAAATCGACAAAATGTCAC). Amplified DNA was fractionated through 2% agarose, blotted onto Biodyne B membranes (Pall Corporation), and hybridized to [␥-32 P] ATP-labeled V␭ -specific oligonucleotide probes CHLVR4 (GAAGAAAGACCGAGACGAGGTCAGCG). 2.10. Reconstitution of DNase -defecient DT40 cell line with DNase  One of the AID/DNase ␥−/− clones (2-2) was transfected with the expression vectors for wild-type chicken DNase ␥ or its mutant (His153 to Ala) fused with EGFP, constructed as follows: cDNAs for the wild-type and the mutant DNase ␥ were first subcloned into pEGFP-N1 (Clontech), and each cDNA fused with the EGFP coding sequence was next subcloned into the pExpress vector, then transfered to blasticidin-selectable vector pLoxBsr as described previously [26]. The transfected cells were selected with blasticidin S hydrochloride (50 ␮g/ml). 3. Results 3.1. Generation of DNase -deficient DT40 cells To address whether DNase ␥ is involved in diversification of V regions, we sought to disrupt DNase ␥ genes in bursal B cell line, DT40, where Ig H and ␭L chain genes are spontaneously accumulating mutations at their V regions through GC as well as point mutations and occasional deletion or insertion of nucleotides. We first identified a chicken ortholog of DNase ␥ by the homology search in the chicken EST database, and cloned a cDNA by PCR (GenBank accession number: DQ151449). The deduced amino acid sequence showed 60% identity and 78% similarity to human DNase ␥. Active site residues His153 and His272 were conserved among species, and a specific feature of DNase ␥ among the DNase I family, namely a nuclear localization signal at the C-terminus [20], was also conserved. The protein encoded by the chicken DNase ␥ cDNA possessed a characteristic DNase activity at neutral pH (Fig. 1A). Based on the cDNA sequence information, we then cloned a genomic clone encoding entire DNase ␥ gene, and constructed gene-targeting vectors (Fig. 2B). In the vectors, the fifth exon of DNase ␥ genes was disrupted by insertion of neomycin- or histidinol-resistant gene cassettes into the site one codon upstream of the His153, deleting the following 4 codons. The DNase ␥ genes on two homologous chromosomes were sequentially targeted by the vectors, which was verified by Southern blot analysis (Fig. 1C). DNase ␥ mRNA transcribed beyond the deleted codons could not be detected by RT-PCR in the homozygous mutant (DNase ␥−/− ) DT40 cells, probably due to a transcriptional termination at the poly A additional signal in the drug-resistance gene cassettes (Fig. 1D). The parental line and DNase ␥−/− cells proliferated at the same rate and did not undergo apoptosis in the normal culture condition (data not shown).

25

3.2. DNase -deficient DT40 cells exhibit lower rate of V gene non-templated mutations We first sought to assess the mutation rate of Ig genes by measuring the appearance of IgM-loss variants in culture, which is known to arise from occasional generation of a frame-shift or a stop codon in rearranged V region genes through the GC as well as the non-templated mutations [27]. However, the fraction of the IgM-loss variants after culture for 1.5 months was almost negligible (∼0.2%) by flow cytometry in our parental DT40 cells, which might be attributable to a very low expression of AID in our stock of DT40 cells (data not shown). Therefore we transfected each of the parental and DNase ␥−/− DT40 cell lines with the vector expressing an mRNA encoding chicken AID and EGFP separated by IRES, and selected stable clones expressing a similar level of EGFP that should parallel the level of AID protein (Fig. 1E). Quantitative analysis of relative mRNA expression revealed similar levels of AID mRNA among the clones except for one (AID/DNase␥−/− 2-2) expressing about 2-fold more AID mRNA (Fig. 1F). Proliferation rates of these clones were equivalent (Fig. 1G), and no spontaneous apoptosis was obvious (data not shown). The AID-transfected DT40 clones (AID/DT40) accumulated significant fraction of IgM-loss variants ranging from 1% to 4% (2.1% on average) of total cells in each clone 1.5 months after the initiation of cloning (Fig. 2A). On the other hand, the AID-transfected DNase ␥−/− DT40 clones (AID/DNase ␥−/− ) showed a significantly lower rate of appearance of the IgM-loss variants (0.6%, 3.5-fold less than AID/DT40 clones, on average). These results suggest that DNase ␥ is positively involved in GC and/or mutations of Ig V regions induced by AID. To verify the actual rate and nature of the V-gene mutations, we cloned and sequenced the rearranged V␭ J␭ gene from sIgM-loss variants that were sorted from AID/DT40 and AID/DNase ␥−/− clones after clonal expansion for 1.5 months, in the same way as reported previously [28,29]. As shown in Fig. 2B and C, almost all V␭ J␭ sequences from AID/DT40 clones contained one or more mutations attributable either to a GC templated by an upstream V␭ pseudogene, to a non-templated point mutation, or to a deletion or an insertion of nucleotides. Point mutations that are also attributable to GC were classified as ‘ambiguous’ mutations [28]. Identical mutations among sequences from the same clone were counted as a single event since they are likely originated from one ‘ancestral’ mutation. The overexpression of AID appeared to strikingly increase the rate of the non-templated mutations (point mutation and nucleotide insertion/deletion) as compared to the previous data of V␭ J␭ sequences from DT40 cells [28]. The sequences from AID/DNase ␥−/− clones showed significantly lower rate of total point mutations and ‘ambiguous’ mutations compared to AID/DT40 clones (2.3-fold and 2.0-fold, respectively; Fig. 2B, C, E). The former contained apparently less point mutations per sequence on average than the latter (Fig. 2C). More strikingly, the former showed a markedly lower incidence of nucleotide deletions and insertions (4.1-fold and 12-fold, respectively; Fig. 2B, D, E). Together, the rate of total non-templated (unambiguous) mutations was 3.3-fold lower in the DNase ␥-deficient DT40 cells. The point mutations were dominantly targeted at C/G showed no obvious bias to transition or transversion in both AID/DT40 and AID/DNase ␥−/− clones as reported for DT40 cells (data not shown [28,29]). The rate of GC in AID/DNase ␥−/− clones was not significantly reduced (Fig. 2E). We did not find any type of mutations in C␭ sequences from AID/DT40 or AID/DNase ␥−/− clones (data not shown). Together with the reduced incidence of sIgM-loss variants by the absence of DNase ␥ (Fig. 2A), these results clearly indicate the significant contribution of DNase ␥ to the generation of point mutations and nucleotide deletion/insertion, in the V region gene.

26

N. Okamoto et al. / Immunology Letters 125 (2009) 22–30

Fig. 2. Decreased V-region diversification and loss of staggered DSBs in DNase ␥-deficient DT40 cells. (A) The frequency (%) of sIgM-loss variants in each clonal sIgM+ AID-transfectant (collectively named AID/DT40 or AID/DNase ␥−/− as indicated) was determined by flow cytometry after clonal expansion for 1.5 months. Seven clones from each transfection were analyzed and depicted as dots. Horizontal lines indicate averages. Statistics was analyzed by Student’s t-test. (B) Mutation analysis of V␭ sequences from sIgM-loss cells sorted from each representative transfectant clone (indicated on top) after clonal expansion for 1.5 months. Each horizontal thin line represents the rearranged V␭ J␭ (427 bp) with mutations indicated as follows; underline: gene conversion; closed triangle with vertical bar: non-templated point mutation; open circle with vertical bar: ambiguous mutation (either gene conversion or non-templated point mutation); open and closed arrowhead: nucleotide deletion and insertion, respectively (with the number of deleted/inserted nucleotides). A succession of non-templated base substitutions, which may include 25% or less unchanged bases (that would happen as

N. Okamoto et al. / Immunology Letters 125 (2009) 22–30

3.3. DNase  is required for the generation of the staggered DSBs in the V gene It has been proposed that the somatic mutation in Ig V region genes coincides with the generation of DSBs, especially in the form of staggered broken ends [14–16], although it has not been examined in chicken B cells. Therefore we investigated whether the staggered DSBs in the rearranged V␭ region gene are generated in the AID-transfected DT40 clones by linker-mediated-PCR (LMPCR) method. In this method, the staggered broken ends cannot be ligated with the linkers and detected by PCR using primers for the linker and V␭ gene without being polished with T4 DNA polymerase, whereas the blunt broken ends can be ligated with the linkers and thus detected by PCR independent of T4 DNA polymerase treatment. As shown in Fig. 2F, AID/DT40 clones carried dominantly the staggered DSBs in the rearranged V␭ gene. By contrast, the V␭ gene in AID/DNase ␥−/− clones carried dominantly blunt DSBs, which may be generated irrespective of SHM in this gene locus as reported in human B cells [16]. Since the frequency of overall DSBs was similar between AID/DT40 and AID/DNase ␥−/− clones, DNase ␥ may generate the staggered DSBs at a relatively low frequency in DT40 cells, just replacing the background blunt-ended DSBs. This result is in line with our previous data from Ramos cells [19], and that the appearance of staggered DSBs coincides with the generation of the non-templated mutations. 3.4. Enzyme activity of DNase  is necessary for acceleration of V gene mutations To confirm the above results and also to examine whether endonuclease activity of DNase ␥ is required for the promotion of non-templated mutations in DT40 cells, we transfected the AID/DNase ␥−/− clone with expression vectors coding for wildtype chicken DNase ␥ (wt) or the same protein with a substitution of a putative active site His153 to alanine (H153A), each tagged with a GFP, and established stable clones expressing either protein (wt/AID/␥−/− or H153A/AID/␥−/− , respectively). We also made the clones transfected with the empty GFP vector as a control (mock/AID/␥−/− ). sIgM expression level and growth rate did not significantly differ among each other (Fig. 3A and B). DNase activity was detectable in wt/AID/␥−/− but not in other transfectants (Fig. 3C). Expression level of exogenous AID mRNA was equivalent among the transfectants (Fig. 3D). LM-PCR analysis showed a strong augmentation of the linker-V␭ ligation/amplification by T4polymerase treatment of substrate genomic DNA in wt/AID/␥−/− , but not in mock/AID/␥−/− , indicating that the frequency of staggered DSBs in the rearranged V␭ region genes was markedly increased by overexpression of wild-type DNase ␥ (Fig. 3E). On the other hand, the level of overall DSBs was suppressed in H153A/AID/␥−/− (Fig. 3E). We amplified the rearranged V␭ J␭ gene by PCR from sIgM-loss variants that were sorted from the three transfectants after clonal expansion for 1.5 months and analyzed the sequences of the amplified clones. As shown in Fig. 3F, the rate of point mutations in the V␭ sequences was significantly increased

27

(2.4-fold) in wt/AID/␥−/− compared to mock/AID/␥−/− , but not in H153A/AID/␥−/− , indicating that the endonuclease activity of the overexpressed DNase ␥ was correlated with the acceleration of point mutations. The rate of nucleotide insertion was also increased in wt/AID/␥−/− . Unexpectedly, the rate of nucleotide deletion was not significantly changed in wt/AID/␥−/− , but was increased in H153A/AID/␥−/− compared to mock/AID/␥−/− , as so was that of GC, a possible reason for which we discuss in the Section 4. GC of Ig V region and the gene-targeting with an exogenous DNA vector are achieved through homologous-recombination that requires RAD54 [30]. In line with the above data indicating no correlation between the expression level of DNase ␥ and the frequency of GC in DT40 cells, the frequency of homologous-recombination of an ovalbumin gene-targeting construct was equivalent between DNase ␥−/− and the parental DT40 cells (Fig. 3G). Thus DNase ␥ does not seem to play a critical role in the generation of homologousrecombination. 4. Discussion We have demonstrated here that DNase ␥ contributes significantly to the generation of point mutations in the rearranged V region genes in DT40 B cell lines and more noticeably of nucleotide insertions in DT40 cells. The rate of these mutations was correlated with the incidence of staggered DSBs in the same region and the level of DNase ␥ expression. On the other hand, the generation of GC was independent of DNase ␥ expression. Given the staggered DSBs are raised by single-strand breaks occurring in proximity on both strands, DNase ␥, preferably cutting single-strand DNA in a physiological ionic strength in vitro [23], may be directly involved in the generation of such DSBs. These results together suggest that DNase ␥-mediated staggered DSBs in the V region gene is at least partly involved in the machinery of the non-templated mutations of this region. This was confirmed by the DNase ␥-reconstitution experiment in DT40 cells for point mutations and nucleotide insertions. Unlike these mutations, the rate of nucleotide deletion was not recovered by the reconstitution with wild-type DNase ␥, but increased with the mutant H153A. This data could be explained as follows. Although the nuclease activity of the H153A mutant was undetectable in our in-gel assay, this assay is not very sensitive and indeed unable to detect the endogenous DNase ␥ activity of DT40 cells (data not shown). Therefore the H153A mutant may still retain a weaker endonuclease activity, possibly through the second active site (H272). Thus, when overexpressed, the residual endonuclease activity of the H153A mutant might be enough for accelerating nucleotide deletions, but not for point mutations and nucleotide insertions. On the other hand, strong wild-type DNase ␥ activity causes exaggerated DNA DSBs, which is lethal for cells unless they are repaired in a way that generates point mutations or nucleotide insertions. This view leads us to propose a hypothesis that the frequency of the staggered DSBs by DNase ␥ determines the way the lesion is resolved: the frequent DSBs are repaired by errorprone machinery that generates point mutations and nucleotide

a result of random base selection), is counted as a single insertion event. The same mutations among different V␭ J␭ sequences obtained from a single transfectant clone (here depicted in gray symbols except for one representative) are likely derived from the same ancestral mutation induced during the culture, and thus scored only once in the mutation frequency analysis in panel E. (C) Proportion of V␭ J␭ sequences carrying different numbers (indicated at the periphery of each pie chart) of gene conversions (GC), non-templated point mutations (PM) or mutations of ambiguous origin (Amb) in the total sequences obtained from sorted sIgM-loss fractions. The number of the analyzed V␭ J␭ sequences taken from two clones for each transfectant (AID/DT40 1–2 and 3–9, or AID/DNase ␥−/− 2-2 and 2–15) is indicated in the center of each chart. (D) Nucleotide length (vertical bar) of each insertion (top) or deletion (bottom) found in the same set of V␭ J␭ sequences as in panel C. (E) Summary of mutation frequency analysis of the V␭ J␭ sequences shown in panels B–D. Total number of the analyzed sequences from each clone, total number of incidence of each type of mutation in such sequences, and the relative frequency of the latter (in parentheses; a value of the latter divided by the former) are indicated in the table. The ‘non-templated mutation’ is the sum of the number of ‘point mutation’, ‘deletions’ and ‘insertions’. (F) Loss of staggered DSBs in the rearranged V␭ region genes in the DNase ␥-deficient DT40 cells. Genomic DNAs from the indicated clones were treated with T4 DNA polymerase (+) or not (−), linker-ligated, and then serially 2-fold diluted into unligated homologous genomic DNA and used as template in LM-PCR for detecting V␭ -region DSBs (each top panel). To compare the total DNA used in the samples, the linker-ligated DNA samples were serially 2-fold diluted into water and used as templates for PCR to amplify C␭ region gene (each bottom panel).

28

N. Okamoto et al. / Immunology Letters 125 (2009) 22–30

Fig. 3. Endonuclease activity of DNase ␥ is correlated with the generation of staggered DSBs, point mutations and nucleotide insertions in V␭ region gene. AID/DNase ␥−/− 2-2 cells were stably transfected with empty vector (mock/AID/␥−/− ), or either vector expressing wild-type chicken DNase ␥-GFP fusion protein (wt/AID/␥−/− ) or mutant chicken DNase ␥ (H153A)-GFP fusion protein (H153A/AID/␥−/− ). (A) Flow cytometric analysis of surface sIgM expression. (B) Growth rate in culture of the stable transfectants indicated in panel A. Live cell number was counted as in Fig. 1G. (C) Expression levels of DNase ␥-EGFP fusion protein (top) and endogenous ␤-actin protein (middle) in lysates of the indicated cells were analyzed by Western blotting using anti-GFP (Molecular Probes) and anti-␤-actin (Santa Cruz Biotechnology) antibodies, respectively. Endonuclease activity in the cell lysates was assayed as in Fig. 1B (bottom). (D) mRNA levels of AID and GAPDH were analyzed by RT-PCR with 3-fold serially diluted cDNA prepared from the indicated cells. (E) Analysis of the DSBs in genomic DNA from each transfectants by LM-PCR as in Fig. 2F. (F) Summary of mutation frequency analysis of the rearranged V␭ J␭ sequences obtained from sorted sIgM-loss variants in each transfectant. The mutations in these sequences were analyzed as in Fig. 2B and E. (G) Homologous recombination in parental and DNase ␥−/− DT40 cells. These cells were transfected with ovalbumin gene-targeting vector and selected by puromycin as described previously [39]. Twenty-four of the drug-resistant clones were propagated and subjected to Southern blot analysis. Genomic DNA digested with Sca I was analyzed with a probe derived from a 5 region of the ovalbumin gene. Eighteen clones each in either the parental or the DNase ␥−/− transfectants carried one allele targeted through homologous-recombination.

N. Okamoto et al. / Immunology Letters 125 (2009) 22–30

insertions, whereas the infrequent DSBs can be resolved through trimming of the broken ends by some exonuclease and joining to generate nucleotide deletions. Moderate expression of the endogenous DNase ␥ in AID/DT40 cells may have caused the DSBs with various frequencies along the time course of culture period, contributing to the generation of the three types of non-templated mutations that were all reduced in AID/DNase ␥−/− cells. On the other hand, constitutive strong expression of exogenous wild-type or the mutant DNase ␥ may have caused constantly frequent or infrequent DSBs, resulting in the dominant generation of point mutations/insertions or deletions, respectively. Although the incidence of GC was independent of DNase ␥ (Fig. 2E), it was increased in the H153A/AID/␥−/− cells (Fig. 3F), implying that GC can be initiated by the blunt DSBs but more efficiently by the infrequent staggered DSBs. Obviously further studies are necessary to prove this hypothesis. Suppressed gene conversion and marked increase in point mutations in DT40 cells lacking homologous-recombination repair proteins such as Rad51-paralogues or Brca2 have led to propose that the majority of the initial lesions for SHM are DSBs that are normally repaired through homologous-recombination with the sister chromatid (faithful repair) or upstream V-pseudogenes (GC) [28,31]. This was supported by the finding that deletion of all Vpseudogenes similarly abolished GC and instead activated point mutations in DT40 cells [24]. These results suggest that the point mutations are likely generated when the DSBs are not resolved by such homologous recombination repair machinery [28,31]. The requirement of DSBs for the point mutations is also supported by the fact that the supposed mechanisms involved in SHM, namely BER and MMR, are also required for CSR that is unequivocally mediated through DSBs in S regions [32]. Although DSBs are likely involved in the generation of the point mutations, our data suggest that the nature of the DSBs initiating the point mutations is distinct from that initiating GC as stated in the previous paragraph. Thus we suppose that GC machinery may suppress the generation of frequent staggered DSBs, and thus of point mutations, perhaps by diminishing the frequency of DNase ␥-engagement to the V-region DNA. According to the DNA hypothesis of AID, SSBs are introduced at UNG-induced abasic sites or during MMR, which could develop into DSBs. In faithful BER pathway, the endonuclease cleaving the abasic site is AP-endonuclease (APE1 or APE2). It was reported that, in B cells from APE1+/− APE2Y/− mice (APE−/− causes lethality), DSBs in S␮ region were markedly reduced but isotype switching was only modestly affected, while V region was not examined [33]. Alternatively, MRE11-RAD50-NBS1 (MRN) complex involved in general DNA break repair, meiotic recombination, etc. was proposed to be responsible for cleaving the AID/UNG-mediated abasic sites [34]. MRE11, but not APE1, was shown to be associated with the rearranged VH region gene in hypermutating human B cells [34]. It was also shown that overexpression of NBS1 in Ramos and DT40 cells accelerates SHM and GC, respectively, although the data also indicated that NBS1 lacking a domain necessary for interaction with MRE11 significantly accelerated SHM [35]. The analysis of VH gene sequences in human patients with Mre11- or NBS1-deficiency, however, revealed that the frequency of SHM was comparable to that of controls [36]. Thus neither APEs nor MRN have been proved to be actually involved in the generation of SHM in V region. DNA strand breaks in V region gene could also be induced by MMR machinery. Endonuclease involved in general MMR has been unknown, but recently MutL␣ (MLH1-PMS2 heterdimer) was shown to be the one [37]. However, unlike MSH2- or MSH6deficient mice, reduction of A/T mutations in SHM was not so remarkable in MLH1- or PMS2-deficient mice. Therefore it remains unclear whether MutL␣ is involved in the MMR during SHM [38]. RNA hypothesis postulates that a putative endonuclease is

29

critical to initiate an unknown error-prone repair machinery generating SHM. However, the endonuclease is yet to be identified [2]. Taken together, although DNA strand breaks are involved in a major part of the supposed SHM machinery, endonuclease responsible for it remains elusive. Our results suggest that DNase ␥ may be involved in the mechanisms that generate SHM at least partly. According to the DNA-hypothesis, DNase ␥ might be one of the endonucleases that generates strand breaks at the UNG-induced abasic sites during BER or at the sites of MMR. According to the RNA-hypothesis, mRNA encoding DNase ␥ might be edited by AID to produce an endonuclease that is specifically recruited to V region genes. Alternatively, DNase ␥ might be recruited there through an unidentified guiding factor that is produced depending on AID. Of course these possibilities should need to be verified. Acknowledgments We thank Drs. T. Kurosaki for providing the Neo and HisD cassettes; S. Takeda for the ovalbumin gene-targeting vector; R. Goitsuka for cDNA and genomic libraries and helpful discussion; D. Shiokawa and S. Tanuma for information on DNase ␥; N. Nakamura for technical help, and Y. Hara for cell sorting. This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology in Japan, and Japan Society for the Promotion of Science (D.K., R.M.). References [1] Rajewsky K, Forster I, Cumano A. Evolutionary and somatic selection of the antibody repertoire in the mouse. Science 1987;238:1088–94. [2] Honjo T, Nagaoka H, Shinkura R, Muramatsu M. AID to overcome the limitations of genomic information. Nat Immunol 2005;6:655–61. [3] Muramatsu M, Sankaranand VS, Anant S, Sugai M, Kinoshita K, Davidson NO, et al. Specific expression of activation-induced cytidine deaminase (AID), a novel member of the RNA-editing deaminase family in germinal center B cells. J Biol Chem 1999;274:18470–6. [4] Arakawa H, Hauschild J, Buerstedde JM. Requirement of the activationinduced deaminase (AID) gene for immunoglobulin gene conversion. Science 2002;295:1301–6. [5] Muramatsu M, Kinoshita K, Fagarasan S, Yamada S, Shinkai Y, Honjo T. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell 2000;102:553–63. [6] Honjo T, Muramatsu M, Fagarasan S. AID: how does it aid antibody diversity? Immunity 2004;20:659–68. [7] Chaudhuri J, Tian M, Khuong C, Chua K, Pinaud E, Alt FW. Transcriptiontargeted DNA deamination by the AID antibody diversification enzyme. Nature 2003;422:726–30. [8] Bransteitter R, Pham P, Scharff MD, Goodman MF. Activation-induced cytidine deaminase deaminates deoxycytidine on single-stranded DNA but requires the action of RNase. Proc Natl Acad Sci USA 2003;100:4102–7. [9] Dickerson SK, Market E, Besmer E, Papavasiliou FN. AID mediates hypermutation by deaminating single stranded DNA. J Exp Med 2003;197:1291–6. [10] Ramiro AR, Stavropoulos P, Jankovic M, Nussenzweig MC. Transcription enhances AID-mediated cytidine deamination by exposing single-stranded DNA on the nontemplate strand. Nat Immunol 2003;4:452–6. [11] Goossens T, Klein U, Küppers R. Frequent occurrence of deletions and duplications during somatic hypermutation: implications for oncogene translocations and heavy chain disease. Proc Natl Acad Sci USA 1998;95:2463–8. [12] Wilson PC, de Bouteiller O, Liu YJ, Potter K, Banchereau J, Capra JD, et al. Somatic hypermutation introduces insertions and deletions into immunoglobulin V genes. J Exp Med 1998;187:59–70. [13] Sale JE, Neuberger MS. TdT-accessible breaks are scattered over the immunoglobulin V domain in a constitutively hypermutating B cell line. Immunity 1998;9:859–69. [14] Bross L, Fukita Y, McBlane F, Demolliere C, Rajewsky K, Jacobs H. DNA doublestrand breaks in immunoglobulin genes undergoing somatic hypermutation. Immunity 2000;13:589–97. [15] Papavasiliou FN, Schatz DG. Cell-cycle-regulated DNA double-stranded breaks in somatic hypermutation of immunoglobulin genes. Nature 2000;408:216–21. [16] Zan H, Wu X, Komori A, Holloman WK, Casali P. AID-dependent generation of resected double-strand DNA breaks and recruitment of Rad52/Rad51 in somatic hypermutation. Immunity 2003;18:727–38. [17] Nagaoka H, Ito S, Muramatsu M, Nakata M, Honjo T. DNA cleavage in immunoglobulin somatic hypermutation depends on de novo protein synthesis but not on uracil DNA glycosylase. Proc Natl Acad Sci USA 2005;102:2022–7.

30

N. Okamoto et al. / Immunology Letters 125 (2009) 22–30

[18] Dorsett Y, Robbiani DF, Jankovic M, Reina-San-Martin B, Eisenreich TR, Nussenzweig MC. A role for AID in chromosome translocations between c-myc and the IgH variable region. J Exp Med 2007;204:2225–32. [19] Okamoto M, Okamoto N, Yashiro H, Shiokawa D, Sunaga S, Yoshimori A, et al. Involvement of DNase ␥ in the resected double-strand DNA breaks in immunoglobulin genes. Biochem Biophys Res Commun 2005;327:76–83. [20] Shiokawa D, Shika Y, Tanuma S. Identification of two functional nuclear localization signals in DNase ␥ and their roles in its apoptotic DNase activity. Biochem J 2003;376:377–81. [21] Shiokawa D, Tanuma S. Molecular cloning and expression of a cDNA encoding an apoptotic endonuclease DNase ␥. Biochem J 1998;332:713–20. [22] Yakovlev AG, Wang G, Stoica BA, Simbulan-Rosenthal CM, Yoshihara K, Smulson ME. Role of DNAS1L3 in Ca2+ - and Mg2+ -dependent cleavage of DNA into oligonucleosomal and high molecular mass fragments. Nucleic Acids Res 1999;27:1999–2005. [23] Mizuta R, Mizuta M, Araki S, Shiokawa D, Tanuma S, Kitamura D. Action of apoptotic endonuclease DNase ␥ on naked DNA and chromatin substrates. Biochem Biophys Res Commun 2006;345:560–7. [24] Arakawa H, Saribasak H, Buerstedde JM. Activation-induced cytidine deaminase initiates immunoglobulin gene conversion and hypermutation by a common intermediate. PLoS Biol 2004;2:967–74. [25] Takata M, Sabe H, Hata A, Inazu T, Homma Y, Nukada T, et al. Tyrosine kinases Lyn and Syk regulate B cell receptor-coupled Ca2+ mobilization through distinct pathways. EMBO J 1994;13:1341–9. [26] Arakawa H, Lodygin D, Buerstedde JM. Mutant loxP vectors for selectable marker recycle and conditional knock-outs. BMC Biotechnol 2001;1:7. [27] Buerstedde JM, Reynaud CA, Humphries EH, Olson W, Ewert DL, Weill JC. Light chain gene conversion continues at high rate in an ALV-induced cell line. EMBO J 1990;9:921–7. [28] Sale JE, Calandrini DM, Takata M, Takeda S, Neuberger MS. Ablation of XRCC2/3 transforms immunoglobulin V gene conversion into somatic hypermutation. Nature 2001;412:921–6.

[29] Di Noia J, Neuberger MS. Altering the pathway of immunoglobulin hypermutation by inhibiting uracil-DNA glycosylase. Nature 2002;419:43–8. [30] Bezzubova O, Silbergleit A, Yamaguchi-Iwai Y, Takeda S, Buerstedde JM. Reduced X-ray resistance and homologous recombination frequencies in a RAD54-/mutant of the chicken DT40 cell line. Cell 1997;89:185–93. [31] Hatanaka A, Yamazoe M, Sale JE, Takata M, Yamamoto K, Kitao H, et al. Similar effects of Brca2 truncation and Rad51 paralog deficiency on immunoglobulin V gene diversification in DT40 cells support an early role for Rad51 paralogs in homologous recombination. Mol Cell Biol 2005;25:1124–34. [32] Lee GS, Brandt VL, Roth DB. B cell development leads off with a base hit: dU:dG mismatches in class switching and hypermutation. Mol Cell 2004;19: 505–8. [33] Guikema JE, Linehan EK, Tsuchimoto D, Nakabeppu Y, Strauss PR, Stavnezer J, et al. APE1- and APE2-dependent DNA breaks in immunoglobulin class switch recombination. J Exp Med 2007;204:3017–26. [34] Larson ED, Cummings WJ, Bednarski DW, Maizels N. MRE11/RAD50 cleaves DNA in the AID/UNG-dependent pathway of immunoglobulin gene diversification. Mol Cell 2005;20:367–75. [35] Yabuki M, Fujii MM, Maizels N. The MRE11-RAD50-NBS1 complex accelerates somatic hypermutation and gene conversion of immunoglobulin variable regions. Nat Immunol 2005;7:730–6. [36] Du L, Dunn-Walters DK, Chrzanowska KH, Stankovic T, Kotnis A, Li X, et al. A regulatory role for NBS1 in strand-specific mutagenesis during somatic hypermutation. PLoS One 2008;3:2482. [37] Kadyrov FA, Dzantiev L, Constantin N, Modrich P. Endonucleolytic function of MutLalpha in human mismatch repair. Cell 2006;126:297–308. [38] Li Z, Woo CJ, Iglesias-Ussel MD, Ronai D, Scharff MD. The generation of antibody diversity through somatic hypermutation and class switch recombination. Genes Dev 2004;18:1–11. [39] Buerstedde JM, Takeda S. Increased ratio of targeted to random integration after transfection of chicken B cell lines. Cell 1991;67:179–88.