Distinct requirements for Ku in N nucleotide addition at V(D)J- and no
http://www.100md.com
《核酸研究医学期刊》
Verna and Marrs McLean Department of Biochemistry and Molecular Biology, Baylor College of Medicine, Houston, TX 77030, USA, 1 Department of Immunology, Baylor College of Medicine, Houston, TX 77030, USA and 2 Skirball Institute of Biomolecular Medicine and Department of Pathology, New York University School of Medicine, New York, NY 10016, USA
*To whom correspondence should be addressed. Tel: +1 713 798 5760; Fax +1 713 796 9438; Email: jwilson@bcm.tmc.edu Correspondence may also be addressed to David B. Roth. Tel: +1 212 263 0945; Fax: +1 212 937 2433; Email: roth@saturn.med.nyu.edu
ABSTRACT
Loss or addition of nucleotides at junctions generated by V(D)J recombination significantly expands the antigen-receptor repertoire. Addition of nontemplated (N) nucleotides is carried out by terminal deoxynucleotidyl transferase (TdT), whose only known physiological role is to create diversity at V(D)J junctions during lymphocyte development. Although purified TdT can act at free DNA ends, its ability to add nucleotides (i.e. form N regions) at coding joints appears to depend on the nonhomologous end-joining factor Ku80. Because the DNA ends generated during V(D)J rearrangements remain associated with the RAG proteins after cleavage, TdT might be targeted for N region addition through interactions with RAG proteins or with Ku80 during remodeling of the post-cleavage complex. Such regulated access would help to prevent TdT from acting at other types of broken ends and degrading the fidelity of end joining. To test this hypothesis, we measured TdT’s ability to add nucleotides to endonuclease-induced chromosomal and extrachromosomal breaks. In both cases TdT added nucleotides efficiently to the cleaved DNA ends. Strikingly, the frequency of N regions at non-V(D)J-generated ends was not dependent on Ku80. Thus our results suggest that Ku80 is required to allow TdT access to RAG post-cleavage complexes, providing support for the hypothesis that Ku is involved in disassembling or remodeling the post-cleavage complex. We also found that N regions were abnormally long in the absence of Ku80, indicating that Ku80 may regulate TdT’s activity at DNA ends in vivo.
INTRODUCTION
During their development, B and T lymphocytes generate an astonishing variety of antigen receptors by employing a combinatorial gene assembly strategy called V(D)J recombination. This site-specific DNA recombination reaction brings together gene segments (termed V, D and J) to form a rearranged antigen receptor variable region gene (1). The V, D and J coding segments are flanked by recombination signal sequences (RSSs) that are recognized and bound by the recombinase, which consists of the RAG-1 and RAG-2 proteins. RAG-1 and RAG-2 act in concert to assemble an appropriate pair of RSS into a synaptic complex and introduce double-strand breaks (DSBs) into the DNA to form two types of DNA ends: blunt signal ends and covalently sealed (hairpin) coding ends. The RAG proteins remain associated with the four ends in the form of a post-cleavage complex in vitro (2–4) and in vivo (5–9), but end joining, i.e. formation of coding joints (which encode the rearranged antigen receptor gene) and signal joints, is accomplished with the help of nonhomologous end-joining (NHEJ) proteins (Ku70, Ku80, DNA-PK, XRCC4, DNA Ligase IV, Artemis and at least one other unidentified factor) (1).
Although the combinatorial diversity created by the rearrangement process contributes substantially to the size of the antigen receptor repertoire, the lion’s share of the heterogeneity in antigen-binding sites is actually generated by imprecision in the joining of coding segments (10). This so-called junctional diversity is produced by loss and/or addition of nucleotides to the broken DNA ends generated during V(D)J recombination. One of the major mechanisms responsible for addition of nucleotides to V(D)J junctions involves the enzyme terminal deoxynucleotidyl transferase (TdT). TdT is expressed specifically in developing lymphocytes at the time of V(D)J recombination and adds nontemplated (N) nucleotides to free 3' ends (11). By this mechanism, TdT increases diversity in T-cell receptor and immunoglobulin molecules and reduces the formation of repetitive homology-driven junctions (12–14). More than 70% of coding joints from adult mice contain N nucleotide additions, with an average insert size of 3–5 nt (15). A similar situation is observed in Chinese hamster ovary (CHO) cell lines transfected with plasmid substrates and RAG and TdT expression vectors; in this system, efficient N nucleotide addition is observed at both coding and signal joints (16).
Yet the mechanism by which TdT adds N nucleotides to broken DNA ends remains murky, and even the form of the substrate with which TdT interacts is undefined. Early observations that TdT efficiently adds nucleotides to free DNA ends in vitro in the absence of other proteins initially suggested that TdT might act independently of the complexes involved in V(D)J recombination (such as the RAG post-cleavage complex), perhaps by direct interaction with exposed DNA ends. This view was thoroughly undermined by the surprising observation that N nucleotides are virtually absent from V(D)J joints that form in Ku80-deficient mice, which have a defect comparable to that seen in TdT-deficient mice (17). Analysis of N nucleotide addition in Ku80-deficient fibroblasts transfected with TdT expression vectors confirmed that Ku80 is required for N nucleotide addition at V(D)J joints (16). Indeed, Ku and TdT can be coimmunoprecipitated from cell extracts (18). These data suggest that Ku might normally recruit TdT to the broken DNA ends during V(D)J recombination (16).
What is the mechanistic basis for the requirement of Ku in N nucleotide addition? We considered two possibilities. We have previously proposed that Ku might help disassemble or remodel the highly stable RAG post-cleavage complex (19), much like the chaperone ClpX removes the bacteriophage Mu transposase (20). In this case, we would predict that the requirement for Ku in N nucleotide addition would be specific to V(D)J recombination-generated DSBs. Alternatively, Ku could help recruit TdT to broken DNA ends, a function that need not be restricted to V(D)J recombination intermediates. To test these possibilities, we set out to determine whether TdT can efficiently add N nucleotides to ends generated by non-V(D)J means and, if so, whether this activity is Ku dependent. We find that Ku is not necessary for TdT to add nucleotides to non-V(D)J-generated ends, but that N regions are abnormally long in its absence. These data support the notion that Ku involvement in N region addition in V(D)J recombination involves remodeling of the post-cleavage complex to make the DNA ends accessible to the NHEJ machinery. Furthermore, the abnormal length of junctional inserts formed in the absence of Ku80 indicates that Ku regulates the activity of TdT at DNA ends in vivo.
MATERIALS AND METHODS
Vectors for chromosomal experiments
The vector pMHAd is a derivative of pHPRTminigene (21) in which exons 1–2 and 3–9 of the human HPRT gene are fused together to form a 3.5 kb gene that contains two exons with the 1.7 kb intron 2 of the HPRT gene between them (Fig. 1). Into the XbaI site near the middle of the intron, we cloned a 181 bp fragment containing an exon and the surrounding splice regions of the major late mRNA leader sequence of adenovirus type 2 (Genbank locus: ADRCG, region 7026–7188). The original adenovirus exon was extended by 22 bp to form a 94 bp exon (22) and the pyrimidine run of the splice acceptor region was replaced with a triple-helix-forming sequence for the sake of experiments not described here. These modifications do not impair the ability of the adenovirus exon to disrupt function of the HPRT minigene, presumably by efficient incorporation into the mRNA (22).
Figure 1. Structure of the killer exon construct. The phosphoglycerol kinase (PGK) promoter is shown upstream of the two exons of the HPRT minigene (fused together from original exons 1–2 and 3–9 of the human HPRT gene). The shaded box indicates the adenovirus killer exon with the surrounding adenovirus intron sequences. The killer exon is flanked by I-SceI endonuclease sites (small rectangles), which are in opposite orientations. Cleavage by I-SceI thus generates ends with noncomplementary 3' overhangs. The final junction is shown as a hatched box, which may contain both deletion and addition of nucleotides to the ends. Rejoining without excessive deletion generates a functional HPRT gene, which confers the ability to grow in selective media.
On either side of the adenovirus fragment we placed recognition sites for the rare cutting endonuclease I-SceI (5'-TAGGGATAACAGGGTAAT) in opposing orientations, so that the 3' single-stranded extensions produced by I-SceI cleavage are noncomplementary. The vector also contains a hygromycin-resistance gene cloned from pInd-Hygro (Invitrogen, Carlsbad, CA), which allows selection for the integrated vector in mammalian cells. The expression vector pCMVISceI (23) was used to express the I-SceI endonuclease transiently in CHO cells, and pSV40TdTs (16) was used to transiently express the short form of TdT.
Chromosomal end-joining experiments
GS22–23 CHO cells, which carry a large deletion at the APRT locus, were derived from HPRT– GS19–43 cells (24). The cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal calf serum, nonessential amino acids and antibiotics. Stable cell lines with the integrated pMHAd vector were constructed by transfecting 1 μg of linearized plasmid into 106 cells with Fugene 6 transfection reagent (Roche) according to the supplier’s protocol. Forty-eight individual colonies, obtained by selection with 200 μg/ml hygromycin, were checked for spontaneous HPRT+ background, good induction of HPRT+ colonies with I-SceI treatment and low copy number by Southern analysis. Clone DSB1/20, which carries a single integrated copy of the construct, was used for the experiments described here.
HPRT+ colonies were obtained by cotransfecting cell line DSB1/20 with 2 μg of the I-SceI expression vector pCMVISceI and 4 μg of the TdT expression vector pSV40TdTs. A similar sized vector that lacks TdT was used in control experiments. Cells were replated the following day at 104–105 cells per plate and selection for the HPRT+ phenotype was applied 1 day later using Hat medium (100 μM hypoxanthine, 0.4 μM aminopterine, 16 μM thymidine). After 14 days of selection individual clones were isolated and grown up from six independent transfections.
Analysis of chromosomal end-joining junctions
Chromosomal DNA was isolated from HPRT+ colonies according to standard protocols, and a 962 bp region around the adenovirus exon was amplified by PCR. Oligonucleotides 5'-CAG TCC TAA ACA GGG TAA TGGA and 5'-GCA ACA GAG ACC TTG TCT CAAA were used as sense and antisense primers with binding sites at 381 bp upstream and 318 bp downstream, respectively, of the I-SceI cut sites. The PCR amplified DNA from each HPRT+ colony was cloned using the TOPO-TA PCR cloning kit (Invitrogen), transformed into Escherichia coli and sequenced. About 10% of HPRT+ colonies initially yielded no detectable PCR product, suggesting that one of the primer sites may have been deleted; those colonies were not analyzed further. HPRT+ colonies with identical intron sequences that arose from different transfections were counted as independent events, whereas identical clones from the same transfection were considered siblings and only one of them was counted.
Plasmid end-joining experiments
The plasmid pDR3 was created from the parent plasmid pJH452 (25) by removal of a ClaI fragment containing the bacteriophage oop transcription terminator, allowing expression of the chloramphenicol acetyl transferase gene. Thus pDR3, which also bears a ?-lactamase (bla) gene, confers resistance to both chloramphenicol and ampicillin. To create a substrate for end joining that contains a pair of 3' extensions, pDR3 was linearized by digestion with PvuI (which cuts twice within the bla gene). After digestion, the DNA was gel purified and quantified prior to transfection.
Cells were transiently transfected with 1.5 μg of linear pDR3 with or without the addition of 1.5 μg of TdT expression vector. For complementation experiments, xrs-6 cells were transfected with 1.5 μg hamster Ku86 cDNA in pcDNA3 (16) (a kind gift from Dr Penny Jeggo). Transfections were performed using Fugene 6 (Boehringer Mannheim, USA) according to the manufacturer’s instructions. DNA was harvested after 48 h according to the method of Hirt (28). Transfection efficiency was determined to be >90% using a plasmid that encodes ?-galactosidase followed by a ?-galactosidase assay (29).
Assay for N-nucleotide addition
N-nucleotide addition was measured by transforming the DNA recovered from CHO cell transfections into competent E.coli DH5 cells. Transformants arising from rejoined circular plasmids were selected on LB agar plates containing chloramphenicol (11 μg/ml). Since PvuI cuts twice within the bla gene, the rare transformants resulting from undigested parental plasmid DNA were readily excluded by screening for the presence of the intervening segment.
PCR products were obtained from bacterial colonies using either MLC2 (5'-GGC GTT ACC CAA CTT AAT CGC C-3') and MLC3 (5'-GTT CCC AAC GAT CAA GGC GAG-3') or MLC4 (5'-CGG TCA ATT CAC TGG CCG TCG-3') and MLC5 (5'-CAT CGT GGT GTC ACG CTC GTC-3') as primers. Products were sequenced using ABI Prism BigDye Terminator Cycle Sequencing Kit (PE Biosystems, USA). Sequencing primers were bla23 (5'-TGC GCA ACG TTG TTG CCA-3'), MLC2, MLC4 or MLC5.
Statistical analysis
Means were compared by the two-tailed t test. Distributions were compared by the chi-square test. For all comparisons a P value of 0.05 was used to accept or reject the null hypothesis, i.e. that the means or distributions were the same. All calculations for the statistical tests were performed by using the PHStat add-in for Excel.
RESULTS
The ‘killer exon’ system
To generate a system that would allow us to induce chromosomal DSBs and analyze their repair, we inserted an adenovirus exon with the surrounding RNA splicing signals into the middle of the intron in the human HPRT minigene (Fig. 1). This ‘killer exon’ blocks the function of the HPRT minigene, presumably by incorporation into the spliced mRNA (22), which would shift the translation reading frame for the protein. For the experiments described here, we used the cell line DSB1/20, which carries a single integrated copy of the disabled HPRT minigene.
The killer exon in the HPRT minigene in DSB1/20 is flanked by recognition sites for the rare cutting endonuclease I-SceI, which has no other identified sites in the CHO genome. Cleavage by I-SceI removes the killer exon, and subsequent joining of the broken ends allows HPRT to be expressed. The I-SceI sites were placed in opposite orientations to ensure that the two 4 nt single-stranded extensions produced by cleavage are noncomplementary and cannot be directly ligated. Rejoining of the broken ends restores the function of the HPRT gene, allowing clones of cells that have undergone repair of the site-specific DSBs to be identified by growth in HAT medium, which selects for HPRT function. As shown in Table 1, expression of I-SceI in DSB1/20 cells increased the frequency of HPRT+ colonies by more than three orders of magnitude.
Table 1. Summary of analysis of end-joining junctions arising in the presence and absence of TdT
TdT efficiently adds N nucleotides to chromosomal DSBs
The killer exon system was used to determine whether TdT can add nucleotides at the site of a chromosomal DSB. Cell line DSB1/20 was cotransfected with an I-SceI expression vector and either a TdT expression vector or an equivalent amount of a control plasmid. HPRT+ colonies was selected in HAT medium and counted (Table 1). The frequencies of I-SceI-induced HPRT+ colonies were similar in the presence and absence of TdT expression, indicating that TdT expression does not have a dramatic effect on the efficiency of end joining.
To examine the characteristics of the repaired DSBs, DNA from individual HPRT+ colonies was amplified by PCR and sequenced (Table 2). As expected, the killer exon was deleted from the intron in the minigene in the HPRT+ clones. Only one of the 28 junctions (4%) analyzed from HPRT+ colonies generated in the absence of TdT treatment had N nucleotides, whereas 17 of 24 HPRT+ colonies (71%) generated in the presence of TdT had N nucleotides at the junction. The difference in frequencies of N nucleotides in the presence and absence of TdT is highly significant (P = 4 x 10–7). Moreover, the frequency of N nucleotides observed here in the presence of TdT is similar to the frequency of N nucleotides at junctions produced by V(D)J recombination in CHO cells transfected with the same TdT expression vector (16). In addition, the GC content of the extra nucleotides in these experiments was 68% (26/38), which is in agreement with previous observations that TdT preferentially adds G nucleotides to ends in vitro (30) and also with the high GC content of N additions observed at V(D)J junctions formed in lymphocytes (31,32) and in CHO cells transfected with the same TdT expression vector (16).
Table 2. Junctions in GS22–23 CHO cells with and without transfected TdT
Variable numbers of nucleotides were lost from the I-SceI-digested ends before the breaks were repaired, but the losses were not extensive (Table 2). Indeed, 19/27 junctions in the absence of TdT and 17/24 in its presence had retained nucleotides from at least one of the original four base overhangs. On average, the number of nucleotides removed from an end in the absence of TdT was 11.9 ± 13.9, and in the presence of TdT it was 8.0 ± 8.6. These means are not significantly different (P = 0.24). Thus the presence of TdT apparently does not affect the processes responsible for removal of nucleotides from broken DNA ends. It should be noted that the short form of TdT was used in these experiments. In contrast to the long form of TdT, which is a 3' to 5' exonuclease (33), the short form lacks nuclease activity and was not expected to contribute to nucleotide removal at I-SceI-induced DNA breaks.
In the absence of TdT, 23 of 28 junctions exhibited a 1 to 4 nucleotide microhomology at the junction (Table 2), in agreement with previous observations on end joining in mammalian cells . Microhomology usage at the five of seven junctions that lacked N nucleotides in the TdT treatment group was not significantly different (P = 0.85). Assuming that TdT was present in the cells where junctions without N nucleotides were generated, these results suggest that the mere presence of TdT does not affect the processes responsible for microhomology usage.
Addition of normal-length N regions during end joining requires Ku
To determine whether N nucleotide addition to broken DNA ends not created by V(D)J recombination depends on Ku, we turned to a plasmid system that has been used previously for the study of DNA end joining in cells (25). Wild-type or Ku80-deficient CHO fibroblasts were transfected with a linear substrate plasmid (pDR3) digested with PvuI, which generates 3' overhangs. In experiments involving TdT addition, the TdT expression vector was transfected first and the linearized plasmid substrate was added 24 h later, by which time TdT expression was at its peak as determined by western blotting (data not shown). DNA was harvested 48 h after the last transfection, and circular plasmids produced by end joining were identified by bacterial transformation with selection on media containing the appropriate antibiotics. We then determined junctional sequences.
As expected, the frequency of N nucleotides at junctions produced in the absence of TdT was low (5/41 junctions; 12%) (Tables 1 and 3). Extra nucleotides are observed in 10% of junctions produced by recircularization of transfected DNA molecules (34,35), and several TdT-independent mechanisms have been proposed to account for them (35–37). In the presence of TdT, however, N regions were seen at 9/27 junctions (33%), which is significantly higher (P = 0.03) (Table 1).
Table 3. Junctions with inserted nucleotides derived from plasmid end-joining experiments
To examine the dependence of N nucleotide addition on Ku80, we employed the Ku80-deficient CHO cell line xrs-6, which lacks the ability to add N regions to junctions produced by V(D)J recombination (16). None of the 15 junctions (<7%) contained extra nucleotides in the absence of TdT (Tables 1 and 3). In the presence of TdT, 8/28 junctions (29%) contained inserted nucleotides (Table 1), but six of these eight inserts were unusually long (11–27 nt) (Table 3). These abnormally long inserts contrast sharply with those seen in wild-type cells: 17 of 17 inserts at chromosomal breaks were 10 nt in wild-type cells (Table 2), and the longest N addition observed in plasmid end-joining experiments in wild-type cells in this study was 7 nt (Table 3).
Short N regions are consistent with observations at endogenous V(D)J rearrangements. The average length of N regions at endogenous antigen receptor gene rearrangements in mice is <5 nt (15); N regions >10 nt in length are quite rare (17,38,39), even in the absence of selection for a productive receptor (40). Likewise, N regions >10 nt are very rarely observed in junctions formed by V(D)J rearrangement of artificial substrates (41), even when TdT is ectopically expressed at high levels (42). Indeed, analysis of a large number of N regions formed by V(D)J recombination in wild-type CHO cell lines transfected with the TdT expression vector used in this study failed to detect N regions >10 nt (16). These data strongly suggest that N regions >10 nt are abnormal.
If these junctions with abnormal inserts are removed from the dataset obtained in the presence of TdT in xrs-6 cells, only 2/28 junctions (7%) contain normal N regions, which is not significantly different from the results in the absence of TdT (P = 0.3). In other words, addition of normal-length N regions is Ku80-dependent. The presence of extraordinarily long N regions at junctions produced in Ku80-deficient cells is reminiscent of previous in vitro experiments which found that the DNA-dependent protein kinase, of which Ku is a component, modulates TdT activity by limiting the length of N nucleotide insertions (43).
To verify that the abnormal-length N regions in xrs-6 cells indeed resulted from a lack of Ku80, we performed a complementation experiment in which 1.5 μg of a Ku80 expression vector, an amount sufficient to normalize V(D)J junctions formed in xrs-6 cells (16), was transiently transfected into xrs-6 cells. Six of 13 junctions (46%) isolated from the complemented cells displayed N regions, and all were of normal length. These data demonstrate that addition of normal-length N regions by TdT to free DNA ends requires Ku80.
DISCUSSION
Both chromosomal and plasmid NHEJ pathways are accessible to TdT
The observation that TdT is present only in cells actively undergoing V(D)J recombination, along with its extremely restricted physiologic function (diversification of rearranging antigen receptor genes), suggested that N nucleotide addition might be specific to V(D)J recombination junctions and targeted through interactions with the RAG proteins themselves or with Ku80 during remodeling of the RAG post-cleavage complex (16,17). In view of the potential deleterious effects of N nucleotide addition to non-V(D)J-related ends, one might imagine that cells would not allow TdT free access to chromosomal DSBs. We found, however, that TdT efficiently adds N nucleotides to chromosomal and extrachromosomal DSBs. Indeed, joining of I-SceI-induced chromosomal DSB in the presence of TdT generated a proportion of junctions with N regions (71%) similar to that typically produced by V(D)J recombination in CHO cells transfected with the same TdT expression vector (16) and in lymphoid cells expressing endogenous TdT (17,43–45). These data provide conclusive evidence that TdT has ready access to broken DNA intermediates joined by NHEJ and is not specifically targeted to V(D)J recombination intermediates.
Regulation of TdT activity by DNA-PK
Our observation that most of the N regions formed in the absence of Ku80 are abnormally long (six out of eight are >10 nt) agrees with our previous discovery of rare extremely long N inserts formed during V(D)J recombination in Ku80-deficient mice and cell lines (16,17). Our findings indicate that, although TdT can access broken DNA ends in cells on its own, it behaves differently in the presence of Ku80. It is interesting to note that DNA-PK modulates the activity of purified TdT in the test tube, limiting the length of N additions to broken DNA ends (43). Because Ku is a component of the DNA-PK holoenzyme and is thought to help recruit the DNA-PK complex to DNA ends (46), our data converge nicely with this previous finding. Thus, although Ku is not absolutely required for N addition to non-V(D)J-mediated breaks, it nevertheless plays some role in modulating N nucleotide addition. This is consistent with the observation that Ku and TdT colocalize to etoposide-induced nuclear foci, which presumably represent DSBs (18).
Ku80 is not required for addition of N regions during plasmid end joining: implications for end processing of V(D)J recombination intermediates
We previously suggested that the requirement for Ku80 for the addition of N nucleotides during V(D)J recombination might reflect a need for Ku-mediated disassembly or remodeling of stable RAG-DNA end complexes to make ends available for TdT (16,17). That possibility is supported by our discovery that non-V(D)J-generated ends are accessible to TdT even in the absence of Ku80. This result suggests that Ku80 either increases the accessibility of the coding and signal ends (e.g. by disassembling the complex) or that it recruits TdT to the complex (but is not required for targeting TdT to other broken DNA ends).
Our data raise an interesting question. At what point during the V(D)J joining reaction does TdT access the ends? Biochemical experiments provide strong evidence that the signal ends remain associated with the RAG proteins in an extremely stable post-cleavage complex whose disruption requires harsh treatments (high temperatures, high salt, phenol extraction) (4,9,47,48). In addition, experiments in cells with a number of joining-deficient RAG mutants provide the best evidence to date that a RAG post-cleavage complex is essential for both signal and coding joint formation (5–7,9). We have suggested that the end processing and joining steps, including N nucleotide addition, might occur in the context of this post-cleavage complex (16,19). Our observation that N nucleotides are added with similar efficiency to both I-SceI-generated DSB and to ends generated by the V(D)J recombinase supports an alternative possibility: end processing (and presumably the subsequent joining step) might occur after the coding and signal ends have been ‘handed off’ to a general (NHEJ) end-joining complex. Indeed, we have recently obtained evidence that one function of the post-cleavage complex is to shepherd the broken ends to the classical NHEJ pathway (9). Understanding precisely how the NHEJ apparatus coordinates end processing (including addition of N nucleotides by TdT) and the subsequent joining steps will require further investigation.
ACKNOWLEDGEMENTS
We thank Vicky Brandt for editorial assistance and Mary Purugganan for helpful discussions. Work in the Roth Laboratory is supported by the Irene Diamond Foundation. This work was supported by grants from the NIH to D.B.R. (AI-36420) and J.H.W. (GM38219 and EY11731).
REFERENCES
Roth,D.B. (2003) Restraining the V(D)J recombinase. Nat. Rev. Immunol., 3, 656–666.
Agrawal,A. and Schatz,D.G. (1997) RAG1 and RAG2 form a stable postcleavage synaptic complex with DNA containing signal ends in V(D)J recombination. Cell, 89, 43–53.
Hiom,K. and Gellert,M. (1998) Assembly of a 12/23 paired signal complex: a critical control point in V(D)J recombination. Mol. Cell, 1, 1011–1019.
Jones,J.M. and Gellert,M. (2001) Intermediates in V(D)J recombination: a stable RAG1/2 complex sequesters cleaved RSS ends. Proc. Natl Acad. Sci. USA, 98, 12926–12931.
Qiu,J.X., Kale,S.B., Yarnall Schultz,H. and Roth,D.B. (2001) Separation-of-function mutants reveal critical roles for RAG2 in both the cleavage and joining steps of V(D)J recombination. Mol. Cell, 7, 77–87.
Yarnall Schultz,H., Landree,M.A., Qiu,J.X., Kale,S.B. and Roth,D.B. (2001) Joining-deficient RAG1 mutants block V(D)J recombination in vivo and hairpin opening in vitro. Mol. Cell, 7, 65–75.
Huye,L.E., Purugganan,M.M., Jiang,M.M. and Roth,D.B. (2002) Mutational analysis of all conserved basic amino acids in RAG-1 reveals catalytic, step arrest and joining-deficient mutants in the V(D)J recombinase. Mol. Cell Biol., 22, 3460–3473.
Tsai,C.L., Drejer,A.H. and Schatz,D.G. (2002) Evidence of a critical architectural function for the RAG proteins in end processing, protection and joining in V(D)J recombination. Genes Dev., 16, 1934–1949.
Lee,G.S., Neiditch,M.B., Salus,S.S. and Roth,D.B. (2004) The RAG post-cleavage complex shepherds double-strand breaks to a specific repair pathway, but RAG nicking stimulates homologous recombination: implications for oncogenesis. Cell, in press.
Brandt,V.L. and Roth,D.B. (2004) Generating antigen receptor diversity. Immunology, www.ergito.com.
Kung,P.C., Siverstone,A.E., McCaffrey,R.P. and Baltimore,D. (1975) Murine terminal deoxynucleotidyl transferase: cellular distribution and response to cortisone. J. Exp. Med., 141, 855–865.
Gerstein,R.M. and Lieber,M.R. (1993) Extent to which homology can constrain coding exon junctional diversity in V(D)J recombination. Nature, 363, 625–627.
Zhang,Y., Cado,D., Asarnow,D.M., Komori,T., Alt,F.W., Raulet,D.H. and Allison,J.P. (1995) The role of short homology repeats and TdT in generation of the invariant gamma delta antigen receptor repertoire in the fetal thymus. Immunity, 3, 439–447.
Komori,T., Pricop,L., Hatakeyama,A., Bona,C.A. and Alt,F.W. (1996) Repertoires of antigen receptors in Tdt congenitally deficient mice. Int. Rev. Immunol., 13, 317–325.
Gilfillan,S., Benoist,C. and Mathis,D. (1995) Mice lacking terminal deoxynucleotidyl transferase: adult mice with a fetal antigen receptor repertoire. Immunol. Rev., 148, 201–219.
Purugganan,M.M., Shah,S., Kearney,J.F. and Roth,D.B. (2001) Ku80 is required for addition of N nucleotides to V(D)J recombination junctions by terminal deoxynucleotidyl transferase. Nucleic Acids Res., 29, 1638–1646.
Bogue,M.A., Wang,C., Zhu,C. and Roth,D.B. (1997) V(D)J recombination in Ku86-deficient mice: distinct effects on coding, signal and hybrid joint formation. Immunity, 7, 37–47.
Mahajan,K.N., Gangi-Peterson,L., Sorscher,D.H., Wang,J., Gathy,K.N., Mahajan,N.P., Reeves,W.H. and Mitchell,B.S. (1999) Association of terminal deoxynucleotidyl transferase with Ku. Proc. Natl Acad. Sci. USA, 96, 13926–13931.
Zhu,C., Bogue,M.A., Lim,D.-S., Hasty,P. and Roth,D.B. (1996) Ku86-deficient mice exhibit severe combined immunodeficiency and defective processing of V(D)J recombination intermediates. Cell, 86, 379–389.
Levchenko,I., Luo,L. and Baker,T.A. (1995) Disassembly of the Mu transposase tetramer by the ClpX chaperone. Genes Dev., 9, 2399–2408.
Matzuk,M.M., Finegold,M.J., Su,J.G., Hsueh,A.J. and Bradley,A. (1992) Alpha-inhibin is a tumour-suppressor gene with gonadal specificity in mice. Nature, 360, 313–319.
Sterner,D.A., Carlo,T. and Berget,S.M. (1996) Architectural limits on split genes. Proc. Natl Acad. Sci. USA, 93, 15081–15085.
Rouet,P., Smih,F. and Jasin,M. (1994) Expression of a site-specific endonuclease stimulates homologous recombination in mammalian cells. Proc. Natl Acad. Sci. USA, 91, 6064–6068.
Sargent,R.G., Brenneman,M.A. and Wilson,J.H. (1997) Repair of site-specific double-strand breaks in a mammalian chromosome by homologous and illegitimate recombination. Mol. Cell. Biol., 17, 267–277.
Kabotyanski,E.B., Gomelsky,L., Han,J.-O., Stamato,T.D. and Roth,D.B. (1998) Double-strand break repair in Ku86- and XRCC4-deficient cells. Nucleic Acids Res., 26, 5333–5342.
Merrihew,R.V., Marburger,K., Pennington,S.L., Roth,D.B. and Wilson,J.H. (1996) High-frequency illegitimate integration of transfected DNA at preintegrated target sites in a mammalian genome. Mol. Cell. Biol., 16, 10–18.
Jeggo,P.A. (1985) X-ray sensitive mutants of Chinese hamster ovary cell line: radio-sensitivity of DNA synthesis. Mutat. Res., 145, 171–176.
Hirt,B. (1967) Selective extraction of polyoma DNA from infected mouse cell cultures. J. Mol. Biol., 26, 365–369.
Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: A Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Press, Plainview, NY.
Robbins,D.J. and Coleman,M.S. (1988) Initiator role of double stranded DNA in terminal transferase catalyzed polymerization reactions. Nucleic Acids Res., 16, 2943–2957.
Feeney,A.J. (1990) Lack of N regions in fetal and neonatal mouse immunoglobulin V-D-J junctional sequences. J. Exp. Med., 172, 1377–1390.
Bangs,L.A., Sanz,I.E. and Teale,J.M. (1991) Comparison of D, JH and junctional diversity in the fetal, adult and aged B cell repertoires. J. Immunol., 146, 1996–2004.
Thai,T.H., Purugganan,M.M., Roth,D.B. and Kearney,J.F. (2002) Distinct and opposite diversifying activities of terminal transferase splice variants. Nat. Immunol., 3, 457–462.
Roth,D.B. and Wilson,J.H. (1986) Nonhomologous recombination in mammalian cells: role for short sequence homologies in the joining reaction. Mol. Cell. Biol., 6, 4295–4304.
Roth,D.B., Chang,X.-B. and Wilson,J.H. (1989) Comparison of filler DNA at immune, nonimmune and oncogenic rearrangements suggests multiple mechanisms of formation. Mol. Cell. Biol., 9, 3049–3057.
Roth,D.B., Porter,T.N. and Wilson,J.H. (1985) Mechanisms of nonhomologous recombination in mammalian cells. Mol. Cell. Biol., 5, 2599–2607.
Roth,D.B., Proctor,G.N., Stewart,L.K. and Wilson,J.H. (1991) Oligonucleotide capture during end joining in mammalian cells. Nucleic Acids Res., 19, 7201–7205.
Desiderio,S.V., Yancopoulos,G., Paskind,M., Thomas,E., Boss,M.A., Landau,N., Alt,F.W. and Baltimore,D. (1984) Insertion of N regions into heavy-chain genes is correlated with expression of terminal deoxytransferase in B cells. Nature, 311, 752–755.
Bentolila,L.A., Wu,G.E., Nourrit,F., D’Andon,M.F., Rougeon,F. and Doyen,N. (1997) Constitutive expression of terminal deoxynucleotidyl transferase in transgenic mice is sufficient for N region diversity to occur at any Ig locus throughout B cell differentiation. J. Immunol., 158, 715–723.
Kepler,T.B., Borrero,M., Rugerio,B., McCray,S.K. and Clarke,S.H. (1996) Interdependence of N nucleotide addition and recombination site choice in V(D)J rearrangement. J. Immunol., 157, 4451–4457.
Lieber,M.R., Hesse,J.E., Mizuuchi,K. and Gellert,M. (1988) Lymphoid V(D)J recombination: nucleotide insertion at signal joints as well as coding joints. Proc. Natl Acad. Sci. USA, 85, 8588–8592.
Kallenbach,S., Doyen,N., D’Andon,M.F. and Rougeon,F. (1992) Three lymphoid-specific factors account for all junctional diversity characteristic of somatic assembly of T-cell receptor and immunoglobulin genes. Proc. Natl Acad. Sci. USA, 89, 2799–2803.
Mickelsen,S., Snyder,C., Trujillo,K., Roth,D.B. and Meek,K. (1999) Modulation of terminal deoxynucleotidyl transferase activity by the DNA dependent protein kinase J. Immunol., 163, 834–843.
Shimizu,T. and Yamagishi,H. (1992) Biased reading frames of pre-existing DH–JH coding joints and preferential nucleotide insertions at VH–DJH signal joints of excision products of immunoglobulin heavy chain gene rearrangements. EMBO J., 11, 4869–4875.
Iwasoto,T. and Yamagishi,H. (1992) Novel excision products of T cell receptor gamma gene rearrangements and developmental stage specificity implied by the frequency of nucleotide insertions at signal joints. Eur. J. Immunol., 22, 101–106.
Lieber,M.R., Ma,Y., Pannicke,U. and Schwarz,K. (2003) Mechanism and regulation of human non-homologous DNA end-joining. Nat. Rev. Mol. Cell. Biol., 4, 712–720.
Leu,T.M., Eastman,Q.M. and Schatz,D.G. (1997) Coding joint formation in a cell-free V(D)J recombination system. Immunity, 7, 303–314.
Ramsden,D.A., Paull,T.T. and Gellert,M. (1997) Cell-free V(D)J recombination. Nature, 388, 488–491.(Zoltan Sandor, Monica L. Calicchio1, R. )
*To whom correspondence should be addressed. Tel: +1 713 798 5760; Fax +1 713 796 9438; Email: jwilson@bcm.tmc.edu Correspondence may also be addressed to David B. Roth. Tel: +1 212 263 0945; Fax: +1 212 937 2433; Email: roth@saturn.med.nyu.edu
ABSTRACT
Loss or addition of nucleotides at junctions generated by V(D)J recombination significantly expands the antigen-receptor repertoire. Addition of nontemplated (N) nucleotides is carried out by terminal deoxynucleotidyl transferase (TdT), whose only known physiological role is to create diversity at V(D)J junctions during lymphocyte development. Although purified TdT can act at free DNA ends, its ability to add nucleotides (i.e. form N regions) at coding joints appears to depend on the nonhomologous end-joining factor Ku80. Because the DNA ends generated during V(D)J rearrangements remain associated with the RAG proteins after cleavage, TdT might be targeted for N region addition through interactions with RAG proteins or with Ku80 during remodeling of the post-cleavage complex. Such regulated access would help to prevent TdT from acting at other types of broken ends and degrading the fidelity of end joining. To test this hypothesis, we measured TdT’s ability to add nucleotides to endonuclease-induced chromosomal and extrachromosomal breaks. In both cases TdT added nucleotides efficiently to the cleaved DNA ends. Strikingly, the frequency of N regions at non-V(D)J-generated ends was not dependent on Ku80. Thus our results suggest that Ku80 is required to allow TdT access to RAG post-cleavage complexes, providing support for the hypothesis that Ku is involved in disassembling or remodeling the post-cleavage complex. We also found that N regions were abnormally long in the absence of Ku80, indicating that Ku80 may regulate TdT’s activity at DNA ends in vivo.
INTRODUCTION
During their development, B and T lymphocytes generate an astonishing variety of antigen receptors by employing a combinatorial gene assembly strategy called V(D)J recombination. This site-specific DNA recombination reaction brings together gene segments (termed V, D and J) to form a rearranged antigen receptor variable region gene (1). The V, D and J coding segments are flanked by recombination signal sequences (RSSs) that are recognized and bound by the recombinase, which consists of the RAG-1 and RAG-2 proteins. RAG-1 and RAG-2 act in concert to assemble an appropriate pair of RSS into a synaptic complex and introduce double-strand breaks (DSBs) into the DNA to form two types of DNA ends: blunt signal ends and covalently sealed (hairpin) coding ends. The RAG proteins remain associated with the four ends in the form of a post-cleavage complex in vitro (2–4) and in vivo (5–9), but end joining, i.e. formation of coding joints (which encode the rearranged antigen receptor gene) and signal joints, is accomplished with the help of nonhomologous end-joining (NHEJ) proteins (Ku70, Ku80, DNA-PK, XRCC4, DNA Ligase IV, Artemis and at least one other unidentified factor) (1).
Although the combinatorial diversity created by the rearrangement process contributes substantially to the size of the antigen receptor repertoire, the lion’s share of the heterogeneity in antigen-binding sites is actually generated by imprecision in the joining of coding segments (10). This so-called junctional diversity is produced by loss and/or addition of nucleotides to the broken DNA ends generated during V(D)J recombination. One of the major mechanisms responsible for addition of nucleotides to V(D)J junctions involves the enzyme terminal deoxynucleotidyl transferase (TdT). TdT is expressed specifically in developing lymphocytes at the time of V(D)J recombination and adds nontemplated (N) nucleotides to free 3' ends (11). By this mechanism, TdT increases diversity in T-cell receptor and immunoglobulin molecules and reduces the formation of repetitive homology-driven junctions (12–14). More than 70% of coding joints from adult mice contain N nucleotide additions, with an average insert size of 3–5 nt (15). A similar situation is observed in Chinese hamster ovary (CHO) cell lines transfected with plasmid substrates and RAG and TdT expression vectors; in this system, efficient N nucleotide addition is observed at both coding and signal joints (16).
Yet the mechanism by which TdT adds N nucleotides to broken DNA ends remains murky, and even the form of the substrate with which TdT interacts is undefined. Early observations that TdT efficiently adds nucleotides to free DNA ends in vitro in the absence of other proteins initially suggested that TdT might act independently of the complexes involved in V(D)J recombination (such as the RAG post-cleavage complex), perhaps by direct interaction with exposed DNA ends. This view was thoroughly undermined by the surprising observation that N nucleotides are virtually absent from V(D)J joints that form in Ku80-deficient mice, which have a defect comparable to that seen in TdT-deficient mice (17). Analysis of N nucleotide addition in Ku80-deficient fibroblasts transfected with TdT expression vectors confirmed that Ku80 is required for N nucleotide addition at V(D)J joints (16). Indeed, Ku and TdT can be coimmunoprecipitated from cell extracts (18). These data suggest that Ku might normally recruit TdT to the broken DNA ends during V(D)J recombination (16).
What is the mechanistic basis for the requirement of Ku in N nucleotide addition? We considered two possibilities. We have previously proposed that Ku might help disassemble or remodel the highly stable RAG post-cleavage complex (19), much like the chaperone ClpX removes the bacteriophage Mu transposase (20). In this case, we would predict that the requirement for Ku in N nucleotide addition would be specific to V(D)J recombination-generated DSBs. Alternatively, Ku could help recruit TdT to broken DNA ends, a function that need not be restricted to V(D)J recombination intermediates. To test these possibilities, we set out to determine whether TdT can efficiently add N nucleotides to ends generated by non-V(D)J means and, if so, whether this activity is Ku dependent. We find that Ku is not necessary for TdT to add nucleotides to non-V(D)J-generated ends, but that N regions are abnormally long in its absence. These data support the notion that Ku involvement in N region addition in V(D)J recombination involves remodeling of the post-cleavage complex to make the DNA ends accessible to the NHEJ machinery. Furthermore, the abnormal length of junctional inserts formed in the absence of Ku80 indicates that Ku regulates the activity of TdT at DNA ends in vivo.
MATERIALS AND METHODS
Vectors for chromosomal experiments
The vector pMHAd is a derivative of pHPRTminigene (21) in which exons 1–2 and 3–9 of the human HPRT gene are fused together to form a 3.5 kb gene that contains two exons with the 1.7 kb intron 2 of the HPRT gene between them (Fig. 1). Into the XbaI site near the middle of the intron, we cloned a 181 bp fragment containing an exon and the surrounding splice regions of the major late mRNA leader sequence of adenovirus type 2 (Genbank locus: ADRCG, region 7026–7188). The original adenovirus exon was extended by 22 bp to form a 94 bp exon (22) and the pyrimidine run of the splice acceptor region was replaced with a triple-helix-forming sequence for the sake of experiments not described here. These modifications do not impair the ability of the adenovirus exon to disrupt function of the HPRT minigene, presumably by efficient incorporation into the mRNA (22).
Figure 1. Structure of the killer exon construct. The phosphoglycerol kinase (PGK) promoter is shown upstream of the two exons of the HPRT minigene (fused together from original exons 1–2 and 3–9 of the human HPRT gene). The shaded box indicates the adenovirus killer exon with the surrounding adenovirus intron sequences. The killer exon is flanked by I-SceI endonuclease sites (small rectangles), which are in opposite orientations. Cleavage by I-SceI thus generates ends with noncomplementary 3' overhangs. The final junction is shown as a hatched box, which may contain both deletion and addition of nucleotides to the ends. Rejoining without excessive deletion generates a functional HPRT gene, which confers the ability to grow in selective media.
On either side of the adenovirus fragment we placed recognition sites for the rare cutting endonuclease I-SceI (5'-TAGGGATAACAGGGTAAT) in opposing orientations, so that the 3' single-stranded extensions produced by I-SceI cleavage are noncomplementary. The vector also contains a hygromycin-resistance gene cloned from pInd-Hygro (Invitrogen, Carlsbad, CA), which allows selection for the integrated vector in mammalian cells. The expression vector pCMVISceI (23) was used to express the I-SceI endonuclease transiently in CHO cells, and pSV40TdTs (16) was used to transiently express the short form of TdT.
Chromosomal end-joining experiments
GS22–23 CHO cells, which carry a large deletion at the APRT locus, were derived from HPRT– GS19–43 cells (24). The cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal calf serum, nonessential amino acids and antibiotics. Stable cell lines with the integrated pMHAd vector were constructed by transfecting 1 μg of linearized plasmid into 106 cells with Fugene 6 transfection reagent (Roche) according to the supplier’s protocol. Forty-eight individual colonies, obtained by selection with 200 μg/ml hygromycin, were checked for spontaneous HPRT+ background, good induction of HPRT+ colonies with I-SceI treatment and low copy number by Southern analysis. Clone DSB1/20, which carries a single integrated copy of the construct, was used for the experiments described here.
HPRT+ colonies were obtained by cotransfecting cell line DSB1/20 with 2 μg of the I-SceI expression vector pCMVISceI and 4 μg of the TdT expression vector pSV40TdTs. A similar sized vector that lacks TdT was used in control experiments. Cells were replated the following day at 104–105 cells per plate and selection for the HPRT+ phenotype was applied 1 day later using Hat medium (100 μM hypoxanthine, 0.4 μM aminopterine, 16 μM thymidine). After 14 days of selection individual clones were isolated and grown up from six independent transfections.
Analysis of chromosomal end-joining junctions
Chromosomal DNA was isolated from HPRT+ colonies according to standard protocols, and a 962 bp region around the adenovirus exon was amplified by PCR. Oligonucleotides 5'-CAG TCC TAA ACA GGG TAA TGGA and 5'-GCA ACA GAG ACC TTG TCT CAAA were used as sense and antisense primers with binding sites at 381 bp upstream and 318 bp downstream, respectively, of the I-SceI cut sites. The PCR amplified DNA from each HPRT+ colony was cloned using the TOPO-TA PCR cloning kit (Invitrogen), transformed into Escherichia coli and sequenced. About 10% of HPRT+ colonies initially yielded no detectable PCR product, suggesting that one of the primer sites may have been deleted; those colonies were not analyzed further. HPRT+ colonies with identical intron sequences that arose from different transfections were counted as independent events, whereas identical clones from the same transfection were considered siblings and only one of them was counted.
Plasmid end-joining experiments
The plasmid pDR3 was created from the parent plasmid pJH452 (25) by removal of a ClaI fragment containing the bacteriophage oop transcription terminator, allowing expression of the chloramphenicol acetyl transferase gene. Thus pDR3, which also bears a ?-lactamase (bla) gene, confers resistance to both chloramphenicol and ampicillin. To create a substrate for end joining that contains a pair of 3' extensions, pDR3 was linearized by digestion with PvuI (which cuts twice within the bla gene). After digestion, the DNA was gel purified and quantified prior to transfection.
Cells were transiently transfected with 1.5 μg of linear pDR3 with or without the addition of 1.5 μg of TdT expression vector. For complementation experiments, xrs-6 cells were transfected with 1.5 μg hamster Ku86 cDNA in pcDNA3 (16) (a kind gift from Dr Penny Jeggo). Transfections were performed using Fugene 6 (Boehringer Mannheim, USA) according to the manufacturer’s instructions. DNA was harvested after 48 h according to the method of Hirt (28). Transfection efficiency was determined to be >90% using a plasmid that encodes ?-galactosidase followed by a ?-galactosidase assay (29).
Assay for N-nucleotide addition
N-nucleotide addition was measured by transforming the DNA recovered from CHO cell transfections into competent E.coli DH5 cells. Transformants arising from rejoined circular plasmids were selected on LB agar plates containing chloramphenicol (11 μg/ml). Since PvuI cuts twice within the bla gene, the rare transformants resulting from undigested parental plasmid DNA were readily excluded by screening for the presence of the intervening segment.
PCR products were obtained from bacterial colonies using either MLC2 (5'-GGC GTT ACC CAA CTT AAT CGC C-3') and MLC3 (5'-GTT CCC AAC GAT CAA GGC GAG-3') or MLC4 (5'-CGG TCA ATT CAC TGG CCG TCG-3') and MLC5 (5'-CAT CGT GGT GTC ACG CTC GTC-3') as primers. Products were sequenced using ABI Prism BigDye Terminator Cycle Sequencing Kit (PE Biosystems, USA). Sequencing primers were bla23 (5'-TGC GCA ACG TTG TTG CCA-3'), MLC2, MLC4 or MLC5.
Statistical analysis
Means were compared by the two-tailed t test. Distributions were compared by the chi-square test. For all comparisons a P value of 0.05 was used to accept or reject the null hypothesis, i.e. that the means or distributions were the same. All calculations for the statistical tests were performed by using the PHStat add-in for Excel.
RESULTS
The ‘killer exon’ system
To generate a system that would allow us to induce chromosomal DSBs and analyze their repair, we inserted an adenovirus exon with the surrounding RNA splicing signals into the middle of the intron in the human HPRT minigene (Fig. 1). This ‘killer exon’ blocks the function of the HPRT minigene, presumably by incorporation into the spliced mRNA (22), which would shift the translation reading frame for the protein. For the experiments described here, we used the cell line DSB1/20, which carries a single integrated copy of the disabled HPRT minigene.
The killer exon in the HPRT minigene in DSB1/20 is flanked by recognition sites for the rare cutting endonuclease I-SceI, which has no other identified sites in the CHO genome. Cleavage by I-SceI removes the killer exon, and subsequent joining of the broken ends allows HPRT to be expressed. The I-SceI sites were placed in opposite orientations to ensure that the two 4 nt single-stranded extensions produced by cleavage are noncomplementary and cannot be directly ligated. Rejoining of the broken ends restores the function of the HPRT gene, allowing clones of cells that have undergone repair of the site-specific DSBs to be identified by growth in HAT medium, which selects for HPRT function. As shown in Table 1, expression of I-SceI in DSB1/20 cells increased the frequency of HPRT+ colonies by more than three orders of magnitude.
Table 1. Summary of analysis of end-joining junctions arising in the presence and absence of TdT
TdT efficiently adds N nucleotides to chromosomal DSBs
The killer exon system was used to determine whether TdT can add nucleotides at the site of a chromosomal DSB. Cell line DSB1/20 was cotransfected with an I-SceI expression vector and either a TdT expression vector or an equivalent amount of a control plasmid. HPRT+ colonies was selected in HAT medium and counted (Table 1). The frequencies of I-SceI-induced HPRT+ colonies were similar in the presence and absence of TdT expression, indicating that TdT expression does not have a dramatic effect on the efficiency of end joining.
To examine the characteristics of the repaired DSBs, DNA from individual HPRT+ colonies was amplified by PCR and sequenced (Table 2). As expected, the killer exon was deleted from the intron in the minigene in the HPRT+ clones. Only one of the 28 junctions (4%) analyzed from HPRT+ colonies generated in the absence of TdT treatment had N nucleotides, whereas 17 of 24 HPRT+ colonies (71%) generated in the presence of TdT had N nucleotides at the junction. The difference in frequencies of N nucleotides in the presence and absence of TdT is highly significant (P = 4 x 10–7). Moreover, the frequency of N nucleotides observed here in the presence of TdT is similar to the frequency of N nucleotides at junctions produced by V(D)J recombination in CHO cells transfected with the same TdT expression vector (16). In addition, the GC content of the extra nucleotides in these experiments was 68% (26/38), which is in agreement with previous observations that TdT preferentially adds G nucleotides to ends in vitro (30) and also with the high GC content of N additions observed at V(D)J junctions formed in lymphocytes (31,32) and in CHO cells transfected with the same TdT expression vector (16).
Table 2. Junctions in GS22–23 CHO cells with and without transfected TdT
Variable numbers of nucleotides were lost from the I-SceI-digested ends before the breaks were repaired, but the losses were not extensive (Table 2). Indeed, 19/27 junctions in the absence of TdT and 17/24 in its presence had retained nucleotides from at least one of the original four base overhangs. On average, the number of nucleotides removed from an end in the absence of TdT was 11.9 ± 13.9, and in the presence of TdT it was 8.0 ± 8.6. These means are not significantly different (P = 0.24). Thus the presence of TdT apparently does not affect the processes responsible for removal of nucleotides from broken DNA ends. It should be noted that the short form of TdT was used in these experiments. In contrast to the long form of TdT, which is a 3' to 5' exonuclease (33), the short form lacks nuclease activity and was not expected to contribute to nucleotide removal at I-SceI-induced DNA breaks.
In the absence of TdT, 23 of 28 junctions exhibited a 1 to 4 nucleotide microhomology at the junction (Table 2), in agreement with previous observations on end joining in mammalian cells . Microhomology usage at the five of seven junctions that lacked N nucleotides in the TdT treatment group was not significantly different (P = 0.85). Assuming that TdT was present in the cells where junctions without N nucleotides were generated, these results suggest that the mere presence of TdT does not affect the processes responsible for microhomology usage.
Addition of normal-length N regions during end joining requires Ku
To determine whether N nucleotide addition to broken DNA ends not created by V(D)J recombination depends on Ku, we turned to a plasmid system that has been used previously for the study of DNA end joining in cells (25). Wild-type or Ku80-deficient CHO fibroblasts were transfected with a linear substrate plasmid (pDR3) digested with PvuI, which generates 3' overhangs. In experiments involving TdT addition, the TdT expression vector was transfected first and the linearized plasmid substrate was added 24 h later, by which time TdT expression was at its peak as determined by western blotting (data not shown). DNA was harvested 48 h after the last transfection, and circular plasmids produced by end joining were identified by bacterial transformation with selection on media containing the appropriate antibiotics. We then determined junctional sequences.
As expected, the frequency of N nucleotides at junctions produced in the absence of TdT was low (5/41 junctions; 12%) (Tables 1 and 3). Extra nucleotides are observed in 10% of junctions produced by recircularization of transfected DNA molecules (34,35), and several TdT-independent mechanisms have been proposed to account for them (35–37). In the presence of TdT, however, N regions were seen at 9/27 junctions (33%), which is significantly higher (P = 0.03) (Table 1).
Table 3. Junctions with inserted nucleotides derived from plasmid end-joining experiments
To examine the dependence of N nucleotide addition on Ku80, we employed the Ku80-deficient CHO cell line xrs-6, which lacks the ability to add N regions to junctions produced by V(D)J recombination (16). None of the 15 junctions (<7%) contained extra nucleotides in the absence of TdT (Tables 1 and 3). In the presence of TdT, 8/28 junctions (29%) contained inserted nucleotides (Table 1), but six of these eight inserts were unusually long (11–27 nt) (Table 3). These abnormally long inserts contrast sharply with those seen in wild-type cells: 17 of 17 inserts at chromosomal breaks were 10 nt in wild-type cells (Table 2), and the longest N addition observed in plasmid end-joining experiments in wild-type cells in this study was 7 nt (Table 3).
Short N regions are consistent with observations at endogenous V(D)J rearrangements. The average length of N regions at endogenous antigen receptor gene rearrangements in mice is <5 nt (15); N regions >10 nt in length are quite rare (17,38,39), even in the absence of selection for a productive receptor (40). Likewise, N regions >10 nt are very rarely observed in junctions formed by V(D)J rearrangement of artificial substrates (41), even when TdT is ectopically expressed at high levels (42). Indeed, analysis of a large number of N regions formed by V(D)J recombination in wild-type CHO cell lines transfected with the TdT expression vector used in this study failed to detect N regions >10 nt (16). These data strongly suggest that N regions >10 nt are abnormal.
If these junctions with abnormal inserts are removed from the dataset obtained in the presence of TdT in xrs-6 cells, only 2/28 junctions (7%) contain normal N regions, which is not significantly different from the results in the absence of TdT (P = 0.3). In other words, addition of normal-length N regions is Ku80-dependent. The presence of extraordinarily long N regions at junctions produced in Ku80-deficient cells is reminiscent of previous in vitro experiments which found that the DNA-dependent protein kinase, of which Ku is a component, modulates TdT activity by limiting the length of N nucleotide insertions (43).
To verify that the abnormal-length N regions in xrs-6 cells indeed resulted from a lack of Ku80, we performed a complementation experiment in which 1.5 μg of a Ku80 expression vector, an amount sufficient to normalize V(D)J junctions formed in xrs-6 cells (16), was transiently transfected into xrs-6 cells. Six of 13 junctions (46%) isolated from the complemented cells displayed N regions, and all were of normal length. These data demonstrate that addition of normal-length N regions by TdT to free DNA ends requires Ku80.
DISCUSSION
Both chromosomal and plasmid NHEJ pathways are accessible to TdT
The observation that TdT is present only in cells actively undergoing V(D)J recombination, along with its extremely restricted physiologic function (diversification of rearranging antigen receptor genes), suggested that N nucleotide addition might be specific to V(D)J recombination junctions and targeted through interactions with the RAG proteins themselves or with Ku80 during remodeling of the RAG post-cleavage complex (16,17). In view of the potential deleterious effects of N nucleotide addition to non-V(D)J-related ends, one might imagine that cells would not allow TdT free access to chromosomal DSBs. We found, however, that TdT efficiently adds N nucleotides to chromosomal and extrachromosomal DSBs. Indeed, joining of I-SceI-induced chromosomal DSB in the presence of TdT generated a proportion of junctions with N regions (71%) similar to that typically produced by V(D)J recombination in CHO cells transfected with the same TdT expression vector (16) and in lymphoid cells expressing endogenous TdT (17,43–45). These data provide conclusive evidence that TdT has ready access to broken DNA intermediates joined by NHEJ and is not specifically targeted to V(D)J recombination intermediates.
Regulation of TdT activity by DNA-PK
Our observation that most of the N regions formed in the absence of Ku80 are abnormally long (six out of eight are >10 nt) agrees with our previous discovery of rare extremely long N inserts formed during V(D)J recombination in Ku80-deficient mice and cell lines (16,17). Our findings indicate that, although TdT can access broken DNA ends in cells on its own, it behaves differently in the presence of Ku80. It is interesting to note that DNA-PK modulates the activity of purified TdT in the test tube, limiting the length of N additions to broken DNA ends (43). Because Ku is a component of the DNA-PK holoenzyme and is thought to help recruit the DNA-PK complex to DNA ends (46), our data converge nicely with this previous finding. Thus, although Ku is not absolutely required for N addition to non-V(D)J-mediated breaks, it nevertheless plays some role in modulating N nucleotide addition. This is consistent with the observation that Ku and TdT colocalize to etoposide-induced nuclear foci, which presumably represent DSBs (18).
Ku80 is not required for addition of N regions during plasmid end joining: implications for end processing of V(D)J recombination intermediates
We previously suggested that the requirement for Ku80 for the addition of N nucleotides during V(D)J recombination might reflect a need for Ku-mediated disassembly or remodeling of stable RAG-DNA end complexes to make ends available for TdT (16,17). That possibility is supported by our discovery that non-V(D)J-generated ends are accessible to TdT even in the absence of Ku80. This result suggests that Ku80 either increases the accessibility of the coding and signal ends (e.g. by disassembling the complex) or that it recruits TdT to the complex (but is not required for targeting TdT to other broken DNA ends).
Our data raise an interesting question. At what point during the V(D)J joining reaction does TdT access the ends? Biochemical experiments provide strong evidence that the signal ends remain associated with the RAG proteins in an extremely stable post-cleavage complex whose disruption requires harsh treatments (high temperatures, high salt, phenol extraction) (4,9,47,48). In addition, experiments in cells with a number of joining-deficient RAG mutants provide the best evidence to date that a RAG post-cleavage complex is essential for both signal and coding joint formation (5–7,9). We have suggested that the end processing and joining steps, including N nucleotide addition, might occur in the context of this post-cleavage complex (16,19). Our observation that N nucleotides are added with similar efficiency to both I-SceI-generated DSB and to ends generated by the V(D)J recombinase supports an alternative possibility: end processing (and presumably the subsequent joining step) might occur after the coding and signal ends have been ‘handed off’ to a general (NHEJ) end-joining complex. Indeed, we have recently obtained evidence that one function of the post-cleavage complex is to shepherd the broken ends to the classical NHEJ pathway (9). Understanding precisely how the NHEJ apparatus coordinates end processing (including addition of N nucleotides by TdT) and the subsequent joining steps will require further investigation.
ACKNOWLEDGEMENTS
We thank Vicky Brandt for editorial assistance and Mary Purugganan for helpful discussions. Work in the Roth Laboratory is supported by the Irene Diamond Foundation. This work was supported by grants from the NIH to D.B.R. (AI-36420) and J.H.W. (GM38219 and EY11731).
REFERENCES
Roth,D.B. (2003) Restraining the V(D)J recombinase. Nat. Rev. Immunol., 3, 656–666.
Agrawal,A. and Schatz,D.G. (1997) RAG1 and RAG2 form a stable postcleavage synaptic complex with DNA containing signal ends in V(D)J recombination. Cell, 89, 43–53.
Hiom,K. and Gellert,M. (1998) Assembly of a 12/23 paired signal complex: a critical control point in V(D)J recombination. Mol. Cell, 1, 1011–1019.
Jones,J.M. and Gellert,M. (2001) Intermediates in V(D)J recombination: a stable RAG1/2 complex sequesters cleaved RSS ends. Proc. Natl Acad. Sci. USA, 98, 12926–12931.
Qiu,J.X., Kale,S.B., Yarnall Schultz,H. and Roth,D.B. (2001) Separation-of-function mutants reveal critical roles for RAG2 in both the cleavage and joining steps of V(D)J recombination. Mol. Cell, 7, 77–87.
Yarnall Schultz,H., Landree,M.A., Qiu,J.X., Kale,S.B. and Roth,D.B. (2001) Joining-deficient RAG1 mutants block V(D)J recombination in vivo and hairpin opening in vitro. Mol. Cell, 7, 65–75.
Huye,L.E., Purugganan,M.M., Jiang,M.M. and Roth,D.B. (2002) Mutational analysis of all conserved basic amino acids in RAG-1 reveals catalytic, step arrest and joining-deficient mutants in the V(D)J recombinase. Mol. Cell Biol., 22, 3460–3473.
Tsai,C.L., Drejer,A.H. and Schatz,D.G. (2002) Evidence of a critical architectural function for the RAG proteins in end processing, protection and joining in V(D)J recombination. Genes Dev., 16, 1934–1949.
Lee,G.S., Neiditch,M.B., Salus,S.S. and Roth,D.B. (2004) The RAG post-cleavage complex shepherds double-strand breaks to a specific repair pathway, but RAG nicking stimulates homologous recombination: implications for oncogenesis. Cell, in press.
Brandt,V.L. and Roth,D.B. (2004) Generating antigen receptor diversity. Immunology, www.ergito.com.
Kung,P.C., Siverstone,A.E., McCaffrey,R.P. and Baltimore,D. (1975) Murine terminal deoxynucleotidyl transferase: cellular distribution and response to cortisone. J. Exp. Med., 141, 855–865.
Gerstein,R.M. and Lieber,M.R. (1993) Extent to which homology can constrain coding exon junctional diversity in V(D)J recombination. Nature, 363, 625–627.
Zhang,Y., Cado,D., Asarnow,D.M., Komori,T., Alt,F.W., Raulet,D.H. and Allison,J.P. (1995) The role of short homology repeats and TdT in generation of the invariant gamma delta antigen receptor repertoire in the fetal thymus. Immunity, 3, 439–447.
Komori,T., Pricop,L., Hatakeyama,A., Bona,C.A. and Alt,F.W. (1996) Repertoires of antigen receptors in Tdt congenitally deficient mice. Int. Rev. Immunol., 13, 317–325.
Gilfillan,S., Benoist,C. and Mathis,D. (1995) Mice lacking terminal deoxynucleotidyl transferase: adult mice with a fetal antigen receptor repertoire. Immunol. Rev., 148, 201–219.
Purugganan,M.M., Shah,S., Kearney,J.F. and Roth,D.B. (2001) Ku80 is required for addition of N nucleotides to V(D)J recombination junctions by terminal deoxynucleotidyl transferase. Nucleic Acids Res., 29, 1638–1646.
Bogue,M.A., Wang,C., Zhu,C. and Roth,D.B. (1997) V(D)J recombination in Ku86-deficient mice: distinct effects on coding, signal and hybrid joint formation. Immunity, 7, 37–47.
Mahajan,K.N., Gangi-Peterson,L., Sorscher,D.H., Wang,J., Gathy,K.N., Mahajan,N.P., Reeves,W.H. and Mitchell,B.S. (1999) Association of terminal deoxynucleotidyl transferase with Ku. Proc. Natl Acad. Sci. USA, 96, 13926–13931.
Zhu,C., Bogue,M.A., Lim,D.-S., Hasty,P. and Roth,D.B. (1996) Ku86-deficient mice exhibit severe combined immunodeficiency and defective processing of V(D)J recombination intermediates. Cell, 86, 379–389.
Levchenko,I., Luo,L. and Baker,T.A. (1995) Disassembly of the Mu transposase tetramer by the ClpX chaperone. Genes Dev., 9, 2399–2408.
Matzuk,M.M., Finegold,M.J., Su,J.G., Hsueh,A.J. and Bradley,A. (1992) Alpha-inhibin is a tumour-suppressor gene with gonadal specificity in mice. Nature, 360, 313–319.
Sterner,D.A., Carlo,T. and Berget,S.M. (1996) Architectural limits on split genes. Proc. Natl Acad. Sci. USA, 93, 15081–15085.
Rouet,P., Smih,F. and Jasin,M. (1994) Expression of a site-specific endonuclease stimulates homologous recombination in mammalian cells. Proc. Natl Acad. Sci. USA, 91, 6064–6068.
Sargent,R.G., Brenneman,M.A. and Wilson,J.H. (1997) Repair of site-specific double-strand breaks in a mammalian chromosome by homologous and illegitimate recombination. Mol. Cell. Biol., 17, 267–277.
Kabotyanski,E.B., Gomelsky,L., Han,J.-O., Stamato,T.D. and Roth,D.B. (1998) Double-strand break repair in Ku86- and XRCC4-deficient cells. Nucleic Acids Res., 26, 5333–5342.
Merrihew,R.V., Marburger,K., Pennington,S.L., Roth,D.B. and Wilson,J.H. (1996) High-frequency illegitimate integration of transfected DNA at preintegrated target sites in a mammalian genome. Mol. Cell. Biol., 16, 10–18.
Jeggo,P.A. (1985) X-ray sensitive mutants of Chinese hamster ovary cell line: radio-sensitivity of DNA synthesis. Mutat. Res., 145, 171–176.
Hirt,B. (1967) Selective extraction of polyoma DNA from infected mouse cell cultures. J. Mol. Biol., 26, 365–369.
Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: A Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Press, Plainview, NY.
Robbins,D.J. and Coleman,M.S. (1988) Initiator role of double stranded DNA in terminal transferase catalyzed polymerization reactions. Nucleic Acids Res., 16, 2943–2957.
Feeney,A.J. (1990) Lack of N regions in fetal and neonatal mouse immunoglobulin V-D-J junctional sequences. J. Exp. Med., 172, 1377–1390.
Bangs,L.A., Sanz,I.E. and Teale,J.M. (1991) Comparison of D, JH and junctional diversity in the fetal, adult and aged B cell repertoires. J. Immunol., 146, 1996–2004.
Thai,T.H., Purugganan,M.M., Roth,D.B. and Kearney,J.F. (2002) Distinct and opposite diversifying activities of terminal transferase splice variants. Nat. Immunol., 3, 457–462.
Roth,D.B. and Wilson,J.H. (1986) Nonhomologous recombination in mammalian cells: role for short sequence homologies in the joining reaction. Mol. Cell. Biol., 6, 4295–4304.
Roth,D.B., Chang,X.-B. and Wilson,J.H. (1989) Comparison of filler DNA at immune, nonimmune and oncogenic rearrangements suggests multiple mechanisms of formation. Mol. Cell. Biol., 9, 3049–3057.
Roth,D.B., Porter,T.N. and Wilson,J.H. (1985) Mechanisms of nonhomologous recombination in mammalian cells. Mol. Cell. Biol., 5, 2599–2607.
Roth,D.B., Proctor,G.N., Stewart,L.K. and Wilson,J.H. (1991) Oligonucleotide capture during end joining in mammalian cells. Nucleic Acids Res., 19, 7201–7205.
Desiderio,S.V., Yancopoulos,G., Paskind,M., Thomas,E., Boss,M.A., Landau,N., Alt,F.W. and Baltimore,D. (1984) Insertion of N regions into heavy-chain genes is correlated with expression of terminal deoxytransferase in B cells. Nature, 311, 752–755.
Bentolila,L.A., Wu,G.E., Nourrit,F., D’Andon,M.F., Rougeon,F. and Doyen,N. (1997) Constitutive expression of terminal deoxynucleotidyl transferase in transgenic mice is sufficient for N region diversity to occur at any Ig locus throughout B cell differentiation. J. Immunol., 158, 715–723.
Kepler,T.B., Borrero,M., Rugerio,B., McCray,S.K. and Clarke,S.H. (1996) Interdependence of N nucleotide addition and recombination site choice in V(D)J rearrangement. J. Immunol., 157, 4451–4457.
Lieber,M.R., Hesse,J.E., Mizuuchi,K. and Gellert,M. (1988) Lymphoid V(D)J recombination: nucleotide insertion at signal joints as well as coding joints. Proc. Natl Acad. Sci. USA, 85, 8588–8592.
Kallenbach,S., Doyen,N., D’Andon,M.F. and Rougeon,F. (1992) Three lymphoid-specific factors account for all junctional diversity characteristic of somatic assembly of T-cell receptor and immunoglobulin genes. Proc. Natl Acad. Sci. USA, 89, 2799–2803.
Mickelsen,S., Snyder,C., Trujillo,K., Roth,D.B. and Meek,K. (1999) Modulation of terminal deoxynucleotidyl transferase activity by the DNA dependent protein kinase J. Immunol., 163, 834–843.
Shimizu,T. and Yamagishi,H. (1992) Biased reading frames of pre-existing DH–JH coding joints and preferential nucleotide insertions at VH–DJH signal joints of excision products of immunoglobulin heavy chain gene rearrangements. EMBO J., 11, 4869–4875.
Iwasoto,T. and Yamagishi,H. (1992) Novel excision products of T cell receptor gamma gene rearrangements and developmental stage specificity implied by the frequency of nucleotide insertions at signal joints. Eur. J. Immunol., 22, 101–106.
Lieber,M.R., Ma,Y., Pannicke,U. and Schwarz,K. (2003) Mechanism and regulation of human non-homologous DNA end-joining. Nat. Rev. Mol. Cell. Biol., 4, 712–720.
Leu,T.M., Eastman,Q.M. and Schatz,D.G. (1997) Coding joint formation in a cell-free V(D)J recombination system. Immunity, 7, 303–314.
Ramsden,D.A., Paull,T.T. and Gellert,M. (1997) Cell-free V(D)J recombination. Nature, 388, 488–491.(Zoltan Sandor, Monica L. Calicchio1, R. )