In Vitro Processing of the 3′-Overhanging DNA in the Postcleavage Complex Involved in V(D)J Joining

Abstract

The postcleavage complex involved in V(D)J joining is known to possess a transpositional strand transfer activity, whose physiological role is yet to be clarified. Here we report that RAG1 and RAG2 proteins in the signal end (SE) complex cleave the 3′-overhanging structure of the synthetic coding-end (CE) DNA in two successive steps in vitro. The 3′-overhanging structure is attacked by the SE complex imprecisely, near the double-stranded/single-stranded (ds/ss) junction, and transferred to the SE. The transferred overhang is then resolved and cleaved precisely at the ds/ss junction, generating either the linear or the circular cleavage products. Thus, the blunt-end structure is restored for the SE and variably processed ends are generated for the synthetic CE. This 3′-processing activity is observed not only with the core RAG2 but also with the full-length protein.

V(D)J joining plays key roles in activating and diversifying the antigen receptor genes. In the initial phase of V(D)J joining, the protein products of recombination-activating genes (RAG1 and RAG2) (39, 48) recognize the recombination signal sequences (RSS), each consisting of a conserved 7-mer (CACAGTG) and a conserved 9-mer (ACAAAAACC), separated by a spacer of constant length of either 12 or 23 bp (6, 17, 35, 44, 45, 46, 52, 53). For the coordinate cleavage of RSSs, synaptic complex formation of the 12- and 23-RSSs is required (8, 18, 59). RSS DNA is cleaved by RAG proteins in two successive steps, nicking and hairpin formation (31, 58). A nick is first introduced at the coding and 7-mer border on the top strand. The resulting 3′-hydroxyl group (3′-OH) then attacks the bottom strand to form a hairpin structure at the coding end (CE) and a blunt end at the signal end (SE). After the cleavage of RSSs, the SEs stably stay with the RAG proteins in vitro (2, 18, 21, 36). Physical association of the CEs with the SE complex has been shown to occur in vitro (18, 56). This association in the presence of other repair proteins appears to be necessary for the CE processing in vivo (18, 20, 41, 56, 60).

In the joining phase of V(D)J recombination, CEs are processed and ligated to form a coding joint. Several factors are required in DNA end joining, including the Ku heterodimer (Ku70/80), the catalytic subunit of DNA-dependent protein kinase (DNA-PKcs), Artemis, XRCC4, and DNA ligase IV (4, 11, 14, 15, 27, 28, 34, 38, 42, 43, 54, 61). It has been shown that the Artemis/DNA-PKcs complex opens the hairpin a few nucleotides from the tip, generating the 3′-overhanging structure (30, 50). After the hairpin opening, the CEs are modified by nucleotide deletions and additions for junctional diversification (50). For nucleotide additions, terminal deoxynucleotidyltransferase (TdT) has been reported to be responsible for the non-germ line encoded nucleotides (N) (13, 24). In contrast, the exact mechanism for the nucleotide deletions remains largely unknown. The RAG proteins are known to mediate the strand transfer of the SE to the hairpin CE or to the double-stranded (ds) DNA, resulting in the aberrant joining products in vitro (3, 19, 26, 32). The 3′ flap endonuclease activity was also reported for the RAG proteins and was suggested to be a candidate for the CE processing activity (47). However, this activity was SE independent and requires Mn²⁺.

In the present study, we examined the SE complex for the 3′-processing activity using 3′-overhanging DNA. This is because the RAG proteins remain bound to the SEs in vivo even after the RSS has been cleaved (2, 40). Here, we report that the RAG proteins in the SE complex can process and cleave the 3′ overhang of the ds DNA in vitro in two successive steps.

MATERIALS AND METHODS

Preparation of proteins.

The glutathione S-transferase (GST)-tagged truncated RAG1 protein (amino acids 384 to 1040) was coexpressed either with the GST-tagged truncated RAG2 protein or with the GST-tagged full-length RAG2 protein (amino acids 1 to 383 or 1 to 527, respectively) in HEK-293T cells (56), purified with glutathione-agarose affinity chromatography (52), and dialyzed against 25 mM Tris-HCl (pH 8.0)-2 mM dithiothreitol (DTT)-150 mM KCl-10% glycerol. The proteins used in the experiments illustrated in Fig. 6 were dialyzed against a different buffer containing 25 mM Tris-HCl (pH 8.0), 2 mM DTT, 150 mM NaCl, and 20% glycerol. These two buffers showed little difference in the 3′-processing activity of the RAG proteins. Porcine high-mobility-group 1 (HMG1) protein was prepared as described previously (1). The RSS cleavage activity of the full-length RAG2 was about 60% of that of the truncated core RAG2.

FIG. 6.

Full-length RAG2 has a 3′-processing activity comparable to that of the truncated RAG2. (A) Detection of the transposition activity. The reaction was performed with the labeled 23-SE and target plasmid as described in Materials and Methods. Transposed bands represent a two-ended insertion product schematically shown at the top. (B) Detection of the 3′-processing activity. The substrate DNA was labeled in the 3′ overhang. The SE complex was reconstituted with the biotinylated 12-SE in the presence of 1 mM Mg²⁺ by using either the full-length RAG2 (FL) or the truncated RAG2 (core). The C-terminal region is missing in the core RAG2. The target or substrate DNA was allowed to react with the bead-isolated SE complex in the buffer containing 3 mM Mg²⁺ and 75 mM K⁺.

DNA substrates.

Oligonucleotides were synthesized and purified with high-pressure liquid chromatography and/or by electrophoresis in a denaturing polyacrylamide gel. The strands and their sequences were as follows, with 7-mer and 9-mer signal sequences underlined: 12-SE top strand (49-mer), 5′-CACAGTGCTCCAGGGCTGAACAAAAACCTCCTAGGGTTGCAGCTGACTC-3′; 23-SE top strand (60-mer), 5′-CACAGTGGTAGTACTCCACTGTCTGGGTGTACAAAAACCTCCTAGGGTTGCCATGGACTC-3′; top strand of the 12-ST, 5′-CACAGTGCTCCAGGGCTGAACAAAAACCTCCTAGGGTTGCAGCTGACT-3′; bottom strand of the 12-ST, 5′-AGTCAGCTGCAACCCTAGGAGGTTTTTGTTCAGCCCTGGAGCACTGTGGACCTAATAC-3′; 59-nucleotide (nt) top strand of the ds DNA, 5′-CTTAATACGACTCACTATAGGGCTATGTACTACCCGAACCACCAACCTAATACGACGAA-3′. Oligonucleotide substrates were 5′ end labeled with [γ-³²P]ATP (Amersham Biosciences) by using T4 polynucleotide kinase (New England Biolabs) or 3′ end labeled by annealing with appropriate complementary oligonucleotide and filling in 1 nt with [α-³²P] dGTP or [α-³²P]dCTP (Amersham Biosciences) by using a Klenow fragment (3′ exo⁻; New England Biolabs). To prepare the 3′-dideoxy oligonucleotide, the oligonucleotide was extended with ddCTP or ddGTP (Roche Diagnostics) by using terminal deoxyribonucleotide transferase (Roche Diagnostics). An internally labeled strand was prepared by ligating the 5′-³²P-labeled 3′ half (5′-GAGTCAGCTGCAACCCTAGGAGGTTTTTGTTCAGCCCTGGAGCACTGTG-3′) to the unlabeled 5′ half (5′-GACCTAATAC-3′) in the presence of the complementary strand by using T4 DNA ligase (New England Biolabs). The labeled oligonucleotide was purified from a denaturing polyacrylamide gel and reannealed to the indicated complementary strand. Annealed DNA was further purified by electrophoresis in an 8% polyacrylamide gel, eluted from gel slices with an elution buffer (0.2 M NaCl, 1 mM EDTA, and 20 mM Tris-HCl [pH 7.5]), and purified with reversed-phase column chromatography (Elutip-d; Schleicher & Schuell).

3′-Processing reactions.

To examine the SE requirements, 12-SE DNA (8 nM) and 23-SE DNA (8 nM) were incubated with RAG1 (10 μg/ml), RAG2 (10 μg/ml), and HMG1 (8 μg/ml) proteins at 37°C for 120 min in binding buffer (25 mM MOPS [morpholinepropanesulfonic acid]-KOH [pH 7.0], 5 mM Tris-HCl [pH 8.0], 2.4 mM DTT, 90 mM potassium acetate, 30 mM KCl, 0.1 mg of bovine serum albumin [BSA]/ml, and 2% glycerol) containing 10 mM MgCl₂. The labeled ds DNA (200 cpm/μl) was added and incubated at 37°C for 60 min. In all other experiments, the unlabeled or labeled 12-SE DNA (20 nM or 200 cpm/μl, respectively) and its partner biotinylated 23-SE DNA (20 or 8 nM, respectively) were incubated with RAG1 (10 μg/ml), RAG2 (10 μg/ml), and HMG1 (8 μg/ml) proteins at 37°C for 100 min in 20 μl of the binding buffer containing 10 mM CaCl₂. Streptavidin-coated magnetic beads (Dynabeads M-280; 10 μg/μl) were added, and the reaction mixture was incubated at 37°C for 20 min. After incubation, magnetic beads and supernatant were separated by use of a magnet stand. The beads were washed four times at room temperature with 30 μl of binding buffer containing 10 mM CaCl₂ and resuspended in the binding buffer containing Mg²⁺ (10 mM if not otherwise specified). The labeled or unlabeled substrate ds DNA (200 cpm/μl or 5 nM, respectively) was added to the SE complex and incubated at 37°C. To examine the effect of the Mg²⁺ concentration, the SE complex was reconstituted and isolated with 1 mM CaCl₂ instead of 10 mM CaCl₂. DNA was extracted from the reaction mixture by using phenol-chloroform-isoamyl alcohol (25:24:1), precipitated with ethanol, washed with 70% ethanol, dissolved in formamide dye mix, and electrophoresed in a denaturing polyacrylamide gel.

Transposition.

The SE complex was reconstituted by using the biotinylated 12-SE DNA (16 nM) and the labeled 23-SE DNA (12 nM) in a buffer containing 1 mM MgCl₂, 25 mM MOPS-KOH (pH 7.0), 5 mM Tris-HCl (pH 8.0), 2.8 mM DTT, 60 mM potassium acetate, 60 mM NaCl, 1 mg of BSA/ml, and 8% glycerol. The bead-bound SE complex was washed three times with wash buffer (25 mM MOPS-KOH [pH 7.0], 2.4 mM DTT, 90 mM potassium acetate, 210 mM KCl, 1 mg of BSA/ml, and 2% glycerol) and once with a Tn buffer (25 mM MOPS-KOH [pH 7.0], 2.4 mM DTT, 45 mM potassium acetate, 30 mM KCl, 0.1 mg of BSA/ml, and 2% glycerol). The amount of the isolated SE complex was estimated by a count of the ³²P incorporated into the complex. The target plasmid (pBluescript II SK [Stratagene]) (25 ng/μl) was incubated with the SE complex (0.1 nM) at 37°C in the Tn buffer containing 3 mM MgCl₂. After the transposition reaction, sodium dodecyl sulfate (0.5%) and proteinase K (1 mg/ml) (Roche Diagnostics) were added and incubated at 55°C for 60 min. DNA was extracted from the reaction mixture by using phenol-chloroform-isoamyl alcohol (25:24:1), precipitated with ethanol, resuspended in a Tris-EDTA buffer, electrophoresed in a 1% Tris-acetate-EDTA agarose gel, transferred to a nylon membrane filter, and autoradiographed. The SE complex was reconstituted with the unlabeled 23-SE DNA in parallel and assayed for the 3′-processing activity in the Tn buffer containing 3 mM MgCl₂.

Isolation of the complexes containing ST intermediates.

After the 3′-processing reaction, the complex was reisolated with magnetic beads from the reaction mixture and washed once with the wash buffer containing 1 mM CaCl₂ and four times with the Ca²⁺-free wash buffer. The isolated complex was resuspended in the binding buffer containing 10 mM MgCl₂.

Enzymatic treatments of reaction products.

For the calf intestine alkaline phosphatase (CIP) treatment, DNA extracted from the reaction mixture was incubated at 37°C for 30 min with 0.5 U of CIP (Roche Diagnostics) in the manufacturer's buffer. For the T4 polynucleotide kinase treatment, DNA was incubated with 2 U of polynucleotide kinase in the manufacturer's buffer at 37°C for 30 min. For restriction enzyme treatment, DNA was incubated at 37°C for 120 min with 1 U of PvuII or BsrGI (TaKaRa Biochemicals) in the manufacturer's buffer. For the exonuclease V (exoV) treatment, DNA was incubated at 37°C for 120 min with 2 U of exoV (Amersham Biosciences) in 67 mM glycine-NaOH (pH 9.4)-10 mM β-mercaptoethanol-6.7 mM MgCl₂.

Purification and endonuclease treatment of each cleavage product.

Each cleavage product was separated by electrophoresis in a 15% denaturing polyacrylamide gel and eluted from the gel slice with the elution buffer. DNA was precipitated with ethanol, washed with 70% ethanol, dissolved in a Tris-EDTA buffer, and treated at 25°C for 10 min with mung bean nuclease (3.56, 10.7, or 32 U/ml) (TaKaRa Biochemicals) in 50 mM Tris-HCl (pH 5.2)-30 mM NaCl-10 mM ZnSO₄. DNA was extracted by phenol-chloroform-isoamyl alcohol (25:24:1), concentrated by 1-butanol extraction, dissolved in formamide dye mix, and electrophoresed in a 16.5% denaturing polyacrylamide gel.

RESULTS

The SE complex processes the 3′-overhanging structure.

To examine the CE processing activity of the SE complex, the complex was reconstituted with 12-SE, 23-SE, and RAG1/RAG2 proteins in the presence of HMG1, and incubated with various ds DNA samples labeled at the 5′ end of the top strand. As shown in Fig. 1A, ds DNA with the 3′ overhang (4 nt or longer) was cleaved by the SE complex in the presence of Mg²⁺. No cleavage was found to occur near the blunt end or the recessed 3′ end of ds DNA. In the ds DNA with a 15-nt 3′ overhang, three major cleavage sites were detected in the ss DNA region, at or near the ds/ss junction (Fig. 1B). The reaction was much reduced when either the 12- or the 23-SE was omitted, and it was totally abolished without the SEs (Fig. 1B). In the ds DNA region, cleavages were weakly detectable, even in the absence of the SEs, at or near the ds/ss junction. These cleavages were probably due to the 3′ flap endonuclease activity, previously reported for the RAG proteins (47), because the reaction was enhanced by Mn²⁺ and did not require the presence of SEs (Fig. 1B). We also examined the 3′ CE processing by using uncleaved RSSs. However, since our reaction did not include repair factors such as Artemis, the 3′-overhanging CEs were not generated and unprocessed hairpins accumulated (data not shown).

FIG. 1.

3′-Processing activity of the SE complex. (A) The SE complex cleaves the 3′-overhanging ds DNA near the ds/ss junction. Five ds DNAs with different end structures were labeled with ³²P (indicated by the asterisk) at the 5′ end of the 59-nt top strand. The SE complex was reconstituted with HMG1, copurified RAG1/RAG2 proteins, and the 12-SE and biotinylated 23-SE DNAs; immobilized on the streptavidin-coated magnetic beads; and isolated by using a magnet stand (see Materials and Methods for details). The labeled substrates were incubated with the isolated SE complex in the Mg²⁺-containing buffer (see Materials and Methods) at 37°C for 60 min. DNA was extracted, electrophoresed in a 10% denaturing polyacrylamide gel, and autoradiographed. The 3′-overhanging ds DNA was cleaved near the ds/ss junction, while no cleavage was seen for the 5′-overhanging or the blunt-ended ds DNA. (B) The 3′-processing activity requires the 12- and 23-SEs. The SE complex was assembled with HMG1, RAG1/RAG2 proteins, and the 12-SE and 23-SE DNAs (complete) in the Mg²⁺-containing buffer (see Materials and Methods for details). The ds DNA with a 15-nt 3′ overhang (labeled with ³²P at the 5′ end of the top strand) was allowed to react with the SE complex in the presence of 10 mM Mg²⁺ at 37°C for 60 min. Omission (w/o) of either the 12- or 23-SE greatly reduced the cleavage of the substrate (▾). When Mg²⁺ (10 mM) was replaced with Mn²⁺ (2.5 mM) in the absence of the SEs, the 3′ flap endonucleolytic cleavage (▿) previously reported for the RAG proteins was enhanced.

The cleaved 3′ overhang is transferred to the SE.

To study the fate of the cleaved 3′ overhang, the 3′-terminal nucleotide was labeled in the overhang (Fig. 2A) and the labeled ds DNA was incubated in the presence of Mg²⁺ (Fig. 2B). Three types of reaction products were detected: one (Z) was a slow migrating species, and two (X and Y) were fast migrating species compared with the uncleaved substrate (Fig. 2A). Since none of these species were detected when the 5′ end of the top strand was labeled, they should represent the reaction products cleaved from the 3′ overhang (data not shown). Formation of species X, Y, and Z was totally abolished when either RAG1 or RAG2 was omitted (Fig. 2A). It should be mentioned that the SE complex containing the mutant RAG1 (D708A) (12, 23) failed to generate any of these reaction products (Fig. 2A, right). The substrate was efficiently cleaved at higher Mg²⁺ concentrations, e.g., 10 mM (Fig. 2B). With Ca²⁺, the subspecies of Z that are a few nucleotides longer than those generated with Mg²⁺ were found after prolonged incubation (>2 h) (data not shown). At all Mg²⁺ concentrations tested, species Z appeared first, peaked at an early time, and then slowly diminished, while species X and Y appeared later and continued to accumulate with time. The three bands of species Y, migrating at 13, 14, and 15 nt (designated Y1, Y2, and Y3, respectively), were the cleaved 3′-overhanging structure. The faster-migrating species X was assumed to be the circular form of the species Y, because it was resistant to the exonucleolytic cleavage, as will be discussed later. The third species, Z1 and Z2, each consisting of a cluster of three bands, migrated slower than the 12- and 23-SEs. Since the species Z appeared transiently during the cleavage, we assumed these bands to be the reaction intermediates. To study the structural nature of the species Z and its possible relationship to the SEs, we labeled the 12-SE at the 3′-terminal nucleotide of the bottom strand and reconstituted the SE complex. The complex was incubated with the unlabeled 3′-overhanging ds DNA in the Mg²⁺-containing buffer. After the reaction, DNA was extracted, electrophoresed, and autoradiographed. On the autoradiogram, slower migrating bands newly appeared behind the 12-SE, whose mobilities were identical to that of the Z1 band (Fig. 3A). These new bands, as well as the Z1 band, were cleaved into three smaller fragments of 55, 56, and 57 nt by PvuII digestion. Since the 12-SE DNA contained a PvuII cleavage site 42 nt downstream from the signal end, we assumed that the Z1 bands were generated by the transfer of the 3′ overhang from the ds DNA to the 12-SE (Fig. 3A). Similar results were obtained for Z2 by using the 3′-labeled 23-SE (data not shown). We assume that transfer of the 3′ overhang to the SE is due to the attack by the SE on the 3′ overhang. This is because the experiments with the 3′-dideoxy SE revealed that the 3′-OH group of the 12-SE is required for the generation of 12-STs (Z1), but not 23-STs (Z2), and vice versa (Fig. 2A). We also found that formation of the cleavage products Xs and Ys requires the 3′-OH of the 12- or 23-SE (Fig. 2A). Henceforth, the strand transfer intermediate (Z) will be referred to as ST DNA.

FIG. 2.

Cleavage of the 3′-overhanging DNA. (A) The substrate ds DNA was labeled with ³²P at the 3′-terminal nucleotide in the overhang and incubated with the SE complex in the presence of Mg²⁺ at 37°C as described for Fig. 1B. After the reaction, DNA was extracted, electrophoresed, and autoradiographed. Two sets of faster migrating products (X and Y) were detected. Behind the substrate band, two sets of faint bands (Z1 and Z2) appeared. (Left panel) When the 3′-terminal nucleotide of either the 12- or 23-SE, deoxyguanidine, was replaced with dideoxyguanidine (ddG), formation of species Z1 or Z2 was abolished. When both the 12- and 23-SEs had dideoxyguanidine at the end (ddG SEs), X, Y, and Z were all abolished. (Middle panel) Formation of X, Y, and Z are dependent on the presence of both RAG1 and RAG2. Reactions were performed with the separately purified RAG1 and RAG2. (Right panel) Reactions with the mutant RAG1. In contrast to the wild-type control (WT), the SE complex containing the mutant RAG1 (D708A) failed to generate any reaction products. The purified SE complex was used in these experiments. (B) Kinetics of the 3′ overhang cleavage. The relative amounts of the three different reaction products, species X, Y, and Z, as well as of the unreacted substrate, were quantified by autoradiography. Reactions were performed for various lengths of time (1 to 100 min) by using the purified SE complex in three different Mg²⁺ concentrations (1, 3, and 10 mM).

FIG. 3.

Fate of the cleaved 3′ overhang. (A) The 3′ overhang is transferred to the SE. The substrate ds DNA with the 15-nt 3′ overhang was incubated with the isolated SE complex in the presence of 10 mM Mg²⁺ for 10 min. Either the 12-SE or the 3′-overhanging ds DNA was labeled with ³²P at the 3′-terminal nucleotide. After the reaction, an aliquot of extracted DNA was digested with PvuII. Samples were electrophoresed ina 10% denaturing polyacrylamide gel and autoradiographed. A possible mechanism of the transfer of the 3′ overhang is schematically shown on the right. A triangle indicates the RSS. (B) The transferred 3′ overhang is cleaved from the SE. The SE complex was isolated and immobilized on avidin beads. The 3′-overhanging substrate DNA was allowed to react with the SE complex in the presence of 1 mM Mg²⁺ for 30 min. The reaction mix was separated into the bead-bound complex and unbound supernatant (sup). The isolated SE complex was extensively washed with the Mg²⁺-free buffer and incubated again with 10 mM Mg²⁺. Either the 3′-overhanging substrate DNA or the 12-SE was labeled with ³²P as indicated above for panel A. When the 3′ overhang was labeled in the substrate, the ³²P label transferred to the STs and then appeared in the species X and Y. These pathways are schematically shown on the right.

The 3′ overhang sequence transferred to the SE is precisely cleaved off.

To examine whether the ST DNA is contained in the SE complex, we isolated the SE complex after the incubation with the 3′-overhanging ds DNA (Fig. 3B). The SE complex was immobilized on streptavidin-coated magnetic beads with biotinylated 23-SE. The ds DNA with the ³²P-labeled 3′ overhang was allowed to react with the SE complex in the presence of 1 mM Mg²⁺ for 30 min. The bead-bound SE complex was washed with the Mg²⁺-free buffer, and DNA was extracted from the complex, electrophoresed, and autoradiographed. Almost all of the unreacted substrate and species X and Y were washed out from the complex. However, the 12- and 23-STs remained with the SE complex even after the washing (Fig. 3B). When the washed complex was reincubated in the Mg²⁺-containing buffer, the 12- and 23-ST bands disappeared, yielding the species X and Y bands (Fig. 3B). We performed a similar experiment with the unlabeled 3′-overhanging ds DNA and the 12-SE labeled at the 3′ end of the bottom strand. Again, most of the 12-ST DNA remained in the complex after washing. When the complex was subsequently incubated in the Mg²⁺-containing buffer, the 12-STs gradually disappeared. No cleavage products (species X and Y) were seen in this autoradiogram, because the 3′ overhang of ds DNA was not labeled in this experiment. These results support the idea that the species X and Y are generated from the 12- and 23-STs by the resolution reaction of the strand transfer intermediates.

To demonstrate the resolution process more directly, we reconstituted the SE complex with the 12-ST instead of the 12-SE. The 12-ST DNA was labeled with ³²P at the 5′ end of the bottom strand, paired with the 23-SE, and incubated with the RAG1, RAG2, and HMG1 proteins in the Mg²⁺-containing buffer. The ST band of 58 nt (bottom strand) was converted to the 12-SE of 48 nt (Fig. 4A). Even STs with overhangs as short as 1 nt were cleaved (data not shown). We then labeled the 12-ST at the 3′-terminal nucleotide of the bottom strand and performed the same cleavage reaction. This time, two faster migrating cleavage products (X and Y) appeared (Fig. 4A). The band Y showed the electrophoretic mobility expected for the cleavage product of the 12-ST nicked at the ds/ss junction. The other product, band X, was assumed to be the circular form of the linear cleavage product, band Y (see below). Similar cleavage products were also generated from the labeled 23-ST when it was paired with the 12-SE (data not shown). The cleavage of the 12-ST was significantly reduced when the 12/23 rule was violated by pairing with the 12-SE (Fig. 4A). These results indicate that the 3′ overhang in the ST, transferred from ds DNA to the SE, is indeed cleaved in the SE complex. We examined whether the resolution reaction is HMG dependent. It was found that the HMG protein is dispensable in the ST cleavage reaction (data not shown). Although HMG slightly stimulates the ST cleavage, this could be due to the stimulating effect of the HMG on the SE complex reconstitution rather than on the catalysis.

FIG. 4.

The 3′ overhang sequence is precisely cleaved from the ST. (A) Cleavage of the 12-ST. The 12-ST was labeled either at the 5′ end or at the 3′-terminal nucleotide of the bottom strand. The labeled ST was paired with unlabeled partner SE and incubated with HMG1, RAG1, and RAG2 proteins in the presence of 10 mM Mg²⁺ at 37°C for 120 min. To block the circular DNA formation, the 3′-terminal nucleotide of the overhang, deoxycytidine (dC), was replaced with dideoxycytidine (ddC). The 12-ST was labeled at the ds/ss junction, paired with the 23-SE, and incubated with HMG1, RAG1, and RAG2 proteins. (B) Species X DNA is in a circular form. The 12-ST was labeled at the 3′-terminal nucleotide of the overhang and allowed to react as indicated above for panel A. Bands X and Y were isolated separately from a 15% denaturing polyacrylamide gel. Aliquots of X and Y samples were partially digested with mung bean nuclease (MBN) and electrophoresed in a 16.5% denaturing polyacrylamide gel. Digestion of X and Y are schematically shown.

Linear and circular DNA fragments as cleavage products.

We then studied the structures of species X DNA by using various enzymes that attack the DNA ends, e.g., alkaline phosphatase (which removes the terminal phosphates), T4 polynucleotide kinase (which phosphorylates the 5′-OH end), and exonuclease V (which removes linear DNA from both ends). None of these enzymes altered the mobilities or the intensities of the bands for X DNA (data not shown). These results indicate that the species X DNA is in the circular form. To confirm this notion further, both X and Y bands were purified from a denaturing polyacrylamide gel and digested partially with mung bean endonuclease (Fig. 4B). Band Y was degraded quickly, generating a ladder of small nucleotides. In contrast, band X was converted to a slower-migrating band, whose mobility was the same as that of the band Y of 10 nt (Fig. 4B). At higher concentrations of the mung bean nuclease, a ladder of degradation products appeared (Fig. 4B). When the phosphate at the ds/ss junction was labeled in the 3′ overhang of the 12-ST, bands X and Y were also generated after the incubation of the SE complex in the Mg²⁺-containing buffer (Fig. 4A). These results indicate that the 3′ overhang of 10 nt was precisely cleaved at the ds/ss junction in the 12-ST, resulting in the circular (X) or linear (Y) product. We examined the length effect of the 3′ overhang on the formation of circular DNA product (X). We found that only linear products (Y) were generated with short 3′ overhangs (2, 4, and 6 nt) (data not shown). With longer 3′ overhangs (8 and 10 nt), both linear (Y) and circular (X) products were generated. The terminal 3′-OH in the overhang appeared to be essential in generating the circular product (X). When the 3′ end nucleotide, deoxycytosine, was replaced with a dideoxycytosine, band Y was produced without the generation of band X from the 12-ST (Fig. 4A). Similar results were also obtained with the 3′-overhanging ds DNA when the end nucleotide in the overhang was replaced with a dideoxynucleotide (data not shown). These results indicate that the ST is resolved in two alternative pathways: one generating the circular product (X), and the other generating the linear form (Y). Neither reaction requires an external energy source, such as ATP. The Y band is probably generated by hydrolysis, while the X band appears to be produced by intrastrand transesterification: a water molecule or the terminal 3′-OH of the overhang attacks the phosphate at the ds/ss junction of the ST. To examine whether the resolution of the ST molecule can occur with other nucleophiles, we tested 1,2-ethanediol and glycerol in the 12-ST cleavage reaction. In a 20% denaturing polyacrylamide gel, the alcoholyzed Y migrated more slowly than the hydrolyzed Y (Fig. 5). The shifted mobilities correlated with the sizes of the alcohol molecules used in the reaction.

FIG. 5.

The ST can be resolved by alcoholysis. The SE complex was reconstituted with the labeled 12-ST and biotinylated 23-SE, isolated, resuspended in the Mg²⁺-containing buffer without glycerol, and divided into aliquots. The indicated amount (12.5 or 25%) of glycerol or 1,2-ethanediol was added, and the mixture was incubated at 37°C for 60 min. An aliquot of DNA extracted from the alcohol-free reaction mixture was treated with CIP. DNA samples were electrophoresed on a 20% denaturing polyacrylamide gel. An autoradiogram near the Y band is shown. Alcoholyzed products as well as the hydrolyzed Y band are indicated. A possible mechanism of alcoholysis is schematically shown on the bottom. HO-R indicates the alcohol.

Full-length RAG2 has a 3′-processing activity comparable to that of the truncated RAG2.

It has been reported that the C-terminal region of the full-length RAG2 inhibits the RAG-mediated transposition activity (9, 57). Since the 3′-processing and the transposition reactions are mechanistically similar, we examined whether the C-terminal region of the RAG2 also affects the 3′-processing activity. We reconstituted the SE complex either with the full-length RAG2 or with the truncated RAG2 lacking the C-terminal region. As shown in Fig. 6A, the transposition activity was significantly reduced with the full-length RAG2 as reported previously (57). The double-ended insertion products predominate over the single-ended ones under the Mg²⁺ condition without polyethylene glycol (Fig. 6A and unpublished observation by T. Nishihara). The predominant generation of double-ended insertion products was also reported by Neiditch et al. (37). In our experiments, the amount of transposition products peaked before 10 min in 3 mM Mg²⁺ and then diminished. Transposition products did not accumulate even after 120 min with either core RAG2 or full-length RAG2 (data not shown). Gellert and colleagues reported that the transposition products increased and then diminished in 25 mM Mg²⁺ but accumulated in 5 mM Mg²⁺ (33). This discrepancy is probably due to the difference in the reaction procedures. According to their study, RSS cleavage and SE complex formation can proceed during the transposition reaction (33). This would result in the accumulation of transposition products during the reaction and mask the effect of disintegration. In our study, the SE complex was first purified and then used for the transposition assay. Thus, the transposition products did not accumulate. Unlike the transposition activity, the 3′-processing activity under identical buffer conditions was not affected by the presence of the C-terminal region of the full-length RAG2 (Fig. 6B). It should be mentioned that the full-length RAG2 also supported the cleavage of the 4-nt 3′ overhang (data not shown).

DISCUSSION

Mechanism of the 3′ processing of the overhanging DNA with the SE complex.

In the present study, we have found that the SE complex cleaves the 3′-overhanging structure of the ds DNA under the physiological metal condition. The 3′-overhanging structure is attacked by the SE complex imprecisely near the ds/ss junction and transferred to the SE. The transferred overhang is then resolved and cleaved precisely at the ds/ss junction. Thus, the blunt-end structure is restored for the SE and variably processed ends are generated for the CE DNA. The cleavage and transfer of the 3′ overhang is likely to occur by transesterification: the 3′-OH group of the SE attacks the substrate DNA near the ds/ss junction, resulting in a covalent linkage of the SE to the 3′ overhang (26). However, it is still possible that the 3′ overhang is cleaved by a mechanism other than the strand transfer. Further study, e.g., on the stereochemical course of reaction, is needed to clarify this issue. Unlike the RAG-mediated transposition (33), reversal of the strand transfer is not favored, and the 3′-processed ds DNA is thus released from the SE complex (data not shown). The transferred 3′ overhang is then cleaved from the SE by hydrolysis, alcoholysis, or circular DNA formation. This resolution is strikingly similar to the reaction mediated by the human immunodeficiency virus type 1 integrase, which removes two nucleotides from the 3′ end of the viral DNA via hydrolysis, alcoholysis, or circular-dinucleotide formation in vitro (10). A catalytic DDE motif, found in the transposases and retroviral integrases, is also found in the RAG1 protein (12, 23, 25). We have found that the SE complex reconstituted with the mutant RAG1 (D708A), which has a mutation in the DDE motif (12, 23), failed to mediate both steps in the 3′-processing reaction (data not shown).

The 3′-processing reaction is similar to but distinct from the transposition and disintegration reaction.

The first step in the 3′-processing reaction can be regarded as the initial transposition reaction targeting the distorted DNA structure (26, 55). However, the 3′-processing and transposition reactions are different in the subsequent steps. The transpositional strand transfer to the ds DNA is followed by the reversal reaction that removes the SE from the transposition product and reseals the target DNA (33). This reaction was proposed to explain at least in part why the RAG-mediated transposition does not usually occur in cells (33). In contrast, strand transfer to the 3′ overhang is followed by the release of the 3′-processed ds DNA and by the hydrolysis-mediated resolution of the ST intermediate, accumulating the cleavage products.

It should be noted that the C-terminal region of the full-length RAG2 has different effects on the 3′-processing and the transposition and disintegration reactions: the 3′-processing activity was not affected by the presence of the C-terminal region of RAG2, whereas the transposition activity was significantly reduced (Fig. 6) (57). We suppose that the presence of the C-terminal region of RAG2 may interfere with the accessibility of the target ds DNA (9, 57), but not with the 3′-overhanging end. It is possible that the C-terminal region of the RAG2 may play a role in preventing the harmful transposition in cells (9, 51, 57) while maintaining the 3′-processing activity.

Implication for the junctional diversification in V(D)J joining.

It is important to determine that the 3′-processing activity detected in vitro in the present study plays a role in processing the hairpin-opened CEs in vivo. It has been reported that the hairpin opening yields mainly the 3′-overhanging structure in the CE (7, 29, 30, 49). Since the 3′ overhang can be cleaved at variable sites near the ds/ss junction in vitro (Fig. 1), this cleavage, if it also occurs in vivo, may contribute to the diversity at the junctional sequences of antigen receptor genes. In contrast, the cleavage on the ST occurs precisely at the ds/ss junction (Fig. 4). Restoration of the blunt-end structure at the SE is consistent with the observation that most signal joints do not contain any nucleotide addition or deletion (50).

In the double-strand break repair, nucleotide removal is usually necessary to tailor the DNA ends suitable for ligation. It has been reported that the Artemis/DNA-PKcs complex possesses the 3′ and 5′ overhang endonuclease activities that may be involved in the repair and processing of the CEs (30). Variable CE processing may require other nucleases with different cleavage site preferences. Since the SE complex tends to cleave at sites in the 3′ overhang that are more 5′ than those of Artemis/DNA-PKcs (30), it could contribute more junctional diversity than the Artemis/DNA-PKcs alone, if the SE complex is actually involved in the CE processing in vivo.

To examine the CE processing by the SE complex in vivo, mutant RAGs defective in the 3′-processing activity will be helpful. Reconstitution of the postcleavage processes in vitro, which depend on repair factors such as Ku, will also be necessary. Although these attempts are now under way, some recent findings are encouraging. First, studies with the mutant RAGs have suggested that the RAG proteins may play an architectural role in the postcleavage phase of V(D)J recombination (20, 56). If this is the case, the RAG proteins are likely to associate with or stay near the CEs in the SE complex (2, 18, 40). Second, unlike the 3′ flap endonuclease, the 3′-processing activity of the SE complex is high when Mg²⁺ is present in vitro (Fig. 1B). Third, among the known strand transfer activities of the SE complex, only the 3′ processing could have a physiologic role, preventing a potentially harmful transposition in cells. Finally, the 3′-processing activity is not affected by the presence of the C-terminal region of the RAG2 (Fig. 6), while the transposition activity is significantly reduced (57). The C-terminal region of the RAG2 may be important in preventing the harmful transposition (9, 51, 57) while maintaining the CE processing activity, although it has yet to be verified that this occurs in cells.

The RAG-mediated transposition can be found in Saccharomyces cerevisiae cells (5) but rarely occurs in the mammalian cells. However, the SE complex appears to possess the strand transfer activity in the mammalian cells, because the RAG-mediated hybrid joining has been reported for the extrachromosomal recombination with the truncated RAG proteins (16, 51). A similar observation was also made for the antigen receptor loci in the Ku gene knockout mice where the hairpin CEs accumulate (4, 61). The hybrid joining may result from the unregulated strand transfer of the SE on the CEs before hairpin opening.

V(D)J recombination may be a reversal of an accidental insertion of a transposable element into a primordial V gene, which was exploited later by the vertebrate immune system during evolution (45). In the initial phase of V(D)J joining, the RAG proteins cleave the RSS by a mechanism similar to that of transposition or that of retroviral integration (22, 58). The SE complex has been reported to possess the transposition activity in vitro, while such harmful transpositional strand transfer is rarely seen in vivo. It is tempting to assume that the strand transfer activity of RAG proteins is utilized to diversify the antigen receptor genes in V(D)J joining by removing the 3′ overhang from the hairpin-opened CEs. However, this, of course, must be verified in the in vivo system in the future.