Joining Mutants of RAG1 and RAG2 that Demonstrate Impaired Interactions with the Coding-end DNA*

V(D)J recombination plays key roles in activating and diversifying the antigen receptor genes in lymphocytes (1). Two sets of recombination signal sequences (RSSs)¹ are required, each consisting of a conserved 7-mer (CACAGTG) and a conserved 9-mer (ACAAAAACC). V(D)J recombination takes place only between two RSSs with different spacer lengths, one containing a 12-bp spacer (12-RSS) and the other, a 23-bp spacer (23-RSS) (2).

At the initial step of V(D)J joining, one RSS is bound to the protein products of recombination-activating genes, RAG1 and RAG2 (3–12). This RSS-RAG interaction primarily involves the Hin homeodomain of the RAG1 protein and the 9-mer region of RSS DNA (3, 4, 6). The next step is the synaptic complex formation, bringing the two RSSs together (13–15) in the presence of a DNA-bending protein, HMG1 (11, 16–19). Although a nick can be formed on the top strand in the single RSS-RAG complex, the double strand cleavage, which generates a hairpin at the coding end (CE) and a blunt end at the signal end (SE), requires the 12/23 pairing of the two RSSs (20–23). After the 12- and 23-RSSs are cleaved, the hairpin CEs are opened and processed, and then the two CEs and the two SEs recombine to form a coding joint and a signal joint, respectively, by DNA repair mechanisms (24–38). It has been suggested that the RAG proteins also participate in the joining phase of V(D)J recombination (28, 30, 39–47). Although this process is largely unknown, an architectural role has been suggested for the RAG proteins, in holding the two sets of cleaved ends for further processing and ligation in the post cleavage complex (39, 40, 43, 44, 47). In the in vitro reaction, the coding ends are partly retained in the synaptic complex after the double strand cleavage (39, 40, 43).

Using DNaseI footprinting, we have been studying the interaction between the RAG proteins and recombination signals in both primary and synaptic complexes (6, 48). It was found that RAG-RSS interaction in the 7-mer region drastically changes once the synaptic complex is formed for cleavage (48). Here we report the footprint analysis of the coding region in the synaptic complex of V(D)J joining. We have found that the RAG proteins interact with the RSS DNA not only in the signal region, but also in the coding region in the both pre-cleavage and post-cleavage type synaptic complexes. The RAG mutants, K118A/K119A (41) and S723C (43), defective in the post-cleavage process of V(D)J joining, demonstrated the reduced interaction with the coding sequence.

EXPERIMENTAL PROCEDURES

Preparation of Proteins—Glutathione S-transferase-tagged, truncated RAG proteins (amino acids 384–1040 of RAG1 and 1–383 of RAG2) were coexpressed in HEK-293T cells, purified with glutathioneagarose affinity chromatography (4), and dialyzed against 25 mm Tris-HCl (pH 8.0), 2 mm dithiothreitol, 150 mm KCl, and 10% glycerol. It has been reported that the cleavage and joining activities of the truncated RAG proteins are comparable with those of the full-length proteins (16–23, 41–47). Purities and concentrations of the RAG proteins used in our studies were determined in the SYPRO Ruby (Invitrogen)-stained gels with a fluorescent imaging analyzer, FLA-2000 (Fujifilm). The purified RAG proteins were found to be >90% pure. The expression plasmid for the mutant RAG2 (K118A/K119A) was provided by Dr. D. B. Roth. Plasmids for the RAG1 mutants (S723C and E962A) were generated by in vitro mutagenesis.

DNA Substrates—Oligonucleotides were synthesized and purified as described previously (48). The following are the strand sequences with the 7-mer and 9-mer signals underlined: 12-RSS top strand (89-mer): 5′-CTTCAAACCATCCAATAAACCCTGCGCTGAATTCGTCTTACACAGTGCTCCAGGGCTGAACAAAAACCTCCTAGGGTTGCAGCTGACTC-3′; 23-RSS top strand (100-mer): 5′-CTTCAAACCATCCAATAAACCCTGCGCTGAATTCGTCTTACACAGTGGTAGTACTCCACTGTCTGGGTGTACAAAAACCTCCTAGGGTTGCCATGGACTC-3′; 12-SE top strand (49-mer): 5′-CACAGTGCTCCAGGGCTGAACAAAAACCTCCTAGGGTTGCAGCTGACTC-3′; 23-SE top strand (60-mer): 5′-CTTCAAACCATCCAATAAACCCTGCGCTGAATTCGTCTTACACAGTGGTAGTACTCCACTGTCTGGGTGTACAAAAACCTCCTAGGGTTGCCATGGACTC-3′; and the coding top strand (40-mer): 5′-CTTCAAACCATCCAATAAACCCTGCGCTGAATTCGTCTTA-3′. The oligonucleotides containing the 3′-phosphate or the 3′-biotin were synthesized. For the 5′-end labeling, oligonucleotide substrates were incubated with [γ-³²P]ATP, using the wild-type T4 polynucleotide kinase (New England Biolabs) or the 3′-phosphatase-free mutant (Roche Diagnostics). For the 3′-end labeling, the substrates were annealed to appropriate complementary oligonucleotides and filled in one nucleotide with [α-³²P]dGTP or dCTP using Klenow Fragment (3′ exo-; New England Biolabs). To prepare the 3′-dideoxy oligonucleotide, the oligonucleotide was extended with ddATP (Amersham Bioscience) using terminal deoxyribonucleotide transferase (Roche Diagnostics). The labeled oligonucleotides were purified from the denaturing polyacrylamide gel and reannealed to the indicated complementary strands. The nicked RSS was constructed by annealing three oligonucleotides: an RSS bottom strand, an SE top strand, and a coding top strand with a 3′-dideoxy end. Annealed DNA was purified by electrophoresis in an 8% polyacrylamide gel, eluted from gel slices with an elution buffer containing 0.2 m NaCl, 1 mm EDTA, and 20 mm Tris-HCl (pH 7.5), and purified with Elutip-d (Schleicher & Schuell).

Isolation of Pre- and Post-cleavage Complexes—Biotinylated DNA (8 nm) was incubated with RAG1 and RAG2 proteins (4 μg/ml) and human HMG1 protein (Sigma) (8 μg/ml) at 37 °C for 10 min in 20 μl of binding buffer containing 25 mm MOPS-KOH (pH 7.0), 5 mm Tris-HCl (pH 8.0), 2.4 mm dithiothreitol, 90 mm KOAc, 30 mm KCl, 10 mm CaCl₂, 0.1 mg/ml bovine serum albumin, and 2% glycerol. As carrier DNA, 2 μm double-stranded oligonucleotides, unrelated to the RSS sequence (the top strand is 5′-CTTCAAACCATCCAATAAACCCTGCGCTGAATTCGTCTTAGCTGAACCTCCAGGGCTGACACCCCCAATCCTAGGGTTGCAGCTGACT-3′), were added. End-labeled DNA (∼8 nm, 0.6–1.2 × 10¹⁸ cpm/mol) was then added and further incubated for 10 min. To isolate the complex, streptavidin-coated magnetic beads (Dynabeads M-280, 10 μg/μl) were added, and the reaction mixture was incubated for 15min at 37 °C with occasional mixing. After incubation, magnetic beads and supernatant were separated using a magnet stand. The beads were washed six times at room temperature with 60 μl of binding buffer and then resuspended in 30 μl of binding buffer. The washed magnetic beads and the supernatant were subjected to DNaseI footprinting.

DNaseI Footprinting—DNA samples were pre-heated at 37 °C for 4 min and digested with 5 units of DNaseI (Stratagene) for 4 min. DNA was extracted once with phenol/chloroform/isoamyl alcohol (25:24:1), precipitated with ethanol, and washed with 70% ethanol. Samples were dissolved in formamide dye mix and electrophoresed in an 8% denaturing polyacrylamide gel.

Incorporation of dsDNA into the SE Complex—The SE complex was reconstituted with unlabeled (8 nm) or ³²P-labeled 12-SE DNA (∼8 nm, 0.6–1.2 × 10¹⁸ cpm/mol) and biotinylated 23-SE DNA (10 nm) as described above. The SE complex was purified with magnetic beads and then incubated with end-labeled dsDNA of various ends (∼8 nm, 0.6–1.2 × 10¹⁸ cpm/mol) in the presence of Ca²⁺ for 30 min at 37 °C. The dsDNA was constructed by annealing the 40-nucleotide top strand of the coding region with the complementary strand. The top strand is labeled at the 5′-end and has either an OH or PO₄ at the 3′-end. After incubation, the bead-bound SE complex was washed several times. DNA was extracted from the bead-bound SE complex and supernatant, separately. Samples were electrophoresed in an 8% denaturing polyacrylamide gel.

Incorporation of dsDNA into the Hybrid Complex—Biotinylated RSS DNA (8 nm) was incubated with RAG1 and RAG2 proteins (4 μg/ml) and human HMG1 protein (Sigma) (8 μg/ml) at 37 °C for 10 min in 20 μl of the binding buffer, as described above. The resulting complex was isolated with streptavidin-coated magnetic beads (Dynabeads M-280, 10 μg/μl). End-labeled dsDNA (∼8 nm, 0.6–1.2 × 10¹⁸ cpm/mol) and partner SE or RSS DNA (∼8 nm, 0.6–1.2 × 10¹⁸ cpm/mol) were added, and the incubation was continued for another 10 min. After incubation, magnetic beads and supernatant were separated using a magnet stand. The beads were washed with 60 μl of binding buffer, six times at room temperature, and then resuspended in 30 μl of binding buffer. DNA was extracted from the bead-bound complex, and electrophoresed in an 8% denaturing polyacrylamide gel.

RESULTS

Protection of the Coding Region in the Pre-cleavage Synaptic Complex—We reconstituted three types of pre-cleavage complexes with un-nicked, hemi-nicked, or nicked pairs of synthetic 12- and 23-RSSs, both containing identical coding sequences of 40 nucleotides (Fig. 1A). Each nick was introduced into the coding/7-mer boundary on the top strand of the RSS DNA. Each complex contained ³²P-labeled 12-RSS, biotinylated 23-RSS, and RAG1, RAG2, and HMG1 proteins. To prevent the double strand cleavage of nicked substrates, the complex was formed in the presence of Ca²⁺, and the 3′-OH at the nick was reduced to the 3′-deoxy form. The synaptic complex was isolated with streptavidin-coated magnetic beads as described previously (48). In Fig. 1B, proportions of the ³²P-labeled 12-RSS incorporated into the synaptic complex are compared. Interestingly, the complex formation was facilitated by nicking the top strand, but not the bottom strand. It should be noted that such DNA nicked on the top strand is representative of the RSS cleavage intermediate during V(D)J joining (20). No enhancement was observed by the nick on the bottom strand (data not shown).

We then analyzed the bead-isolated synaptic complex by footprinting (Fig. 2). In 12-RSS, protection was not restricted to the signal region, but was extended into the coding region up to the 12th nucleotide from the 7-mer (Fig. 2A). When the 12-RSS was nicked at the coding/7-mer border on the top strand, the protection was much stronger over all, and it extended 16 nucleotides into the coding region. In the signal region, the footprint pattern was more similar to that of 12-SE in the SE complex. Basically the same results were obtained with 23-RSS (Fig. 2B). It is quite striking that the coding region protection patterns of the 12- and 23-RSSs are virtually indistinguishable from each other. This observation suggests that the two coding regions within the synaptic complex interact with the RAG proteins in essentially the same manner. This is in great contrast to the protection patterns of the signal regions, where the interactions with the RAG proteins are asymmetrical between the 12- and 23-RSSs (48). We also analyzed the hemi-nicked complex, in which either the 12- or the 23-RSS is nicked. Fig. 2C shows footprint patterns of two types of hemi-nicked complexes: one where the substrate 12-RSS is nicked, and the other where the partner 23-RSS is nicked. The effect of nicking on protection appears to act in cis but not in trans. Nicking on the bottom strand did not show any enhancement in protection (data not shown). We also analyzed three other variations of coding sequences for both 12- and 23-RSSs. The protected areas were roughly the same, despite the change in nucleotide sequences (data not shown).

Synaptic Complex Formation with the RSSs Containing the Shorter Coding Regions—We then studied whether the length of the coding region affects the synaptic complex formation. The ³²P-labeled 12-RSSs with coding sequences of various lengths (Fig. 3A) were paired with the biotinylated 23-RSS (nicked) containing the 40-bp coding region. As shown in Fig. 3B, when the coding regions were longer than 15 bp, the 12-RSS was incorporated into the synaptic complex at a level comparable to that of the original coding sequence of 40 bp. When the coding region was shortened to 10 bp, the complex formation was lowered and even further decreased when it was 6 bp (Fig. 3B). These results indicate that at least 15 bp are necessary for the coding region to be incorporated efficiently into the hemi-nicked synaptic complex. This is consistent with the observation that ∼12 bp of coding region, proximal to the 7-mer, is protected in the DNaseI footprinting (Fig. 2). Curiously, however, when the coding region was shorter than 6 bp, incorporation of the 12-RSS increased (Fig. 3B). We speculate that the 12-RSS with an extremely short coding sequence behaves like an SE DNA in the SE complex where the 7-mer end appears to facilitate the complex formation (48).

Protection of dsDNA Incorporated into the Post-cleavage SE Complex—The double strand cleavage in V(D)J recombination generates a hairpin structure at the CE and a blunt end at the SE in the synaptic complex. After the cleavage of the two RSSs, the SEs as well as the CEs appear to remain, at least transiently, in the post-cleavage complex (39, 40, 43, 44) for further processing. To study how the RAG proteins would interact with the coding region after the cleavage, we examined the reconstitution of various post-cleavage complexes (Fig. 4A). The bead-isolated SE complex was incubated under the Ca²⁺ condition with ³²P-labeled dsDNA (as artificial CE DNA) with different right-end structures: hairpin, 3′-OH, and 3′-P (Fig. 4B). The radioactivity of the SE complex after washing was measured. The dsDNA with a hairpin structure or 3′-OH end was retained in the complex, but at low levels. In contrast, dsDNA with 3′-phosphorylated end is incorporated into the SE complex efficiently (Fig. 4B).

Because the 3′-phosphorylated dsDNA was shown to bind to the SE complex, we performed footprinting to examine whether the interaction is specific and how the dsDNA is bound to the complex. As shown in Fig. 4C, each strand was analyzed in both directions. About 16 nucleotides proximal to the right-side end (3′-phosphorylated side) of the dsDNA were protected on both the bottom and top strands. It should be noted that no protection was seen on the other side of the dsDNA, where the 3′-end has an OH group. The 3′-phosphorylated dsDNA may occupy the same site as the coding region of the RSS DNA does in the pre-cleavage synaptic complex. Supporting this idea, the 3′-phosphorylated dsDNA was incorporated into the hybrid complex with the SE DNA, but not into the pre-cleavage complex with the uncleaved RSS DNA (Fig. 5, A and B). This observation supports the idea that the 3′-phosphorylated dsDNA occupies the same site in the synaptic complex as does the coding region of the uncleaved RSS DNA. This notion was confirmed by DNaseI footprinting. Protection patterns of the 3′-phosphorylated dsDNA in the SE and hybrid complexes resembles that of the coding region in the RSS DNA (Fig. 5C). Both the 12-coding and the 23-coding regions were analyzed for interactions with RAG protein using the 12- and 23-SEs, respectively. These results indicate that the 3′-phosphorylated dsDNA retained in the SE or in the hybrid complex interacts with the RAG proteins in a manner similar to that of the RSS coding region in the pre-cleavage synaptic complex.

RAG Mutants That Alter the Coding Interaction in the Synaptic Complex—It has been proposed that the RAG proteins may be involved in the post-cleavage phase of V(D)J recombination (28, 30, 39, 41–44, 45–47). RAG mutants capable of cleaving the substrate, but defective in the joining process have been isolated for each RAG1 and RAG2 protein (41–43, 45). To examine whether such mutants show any alterations in their ability to interact with RSS DNA, we reconstituted the synaptic complex with the mutant RAGs and examined the protection of the signal and coding regions (Fig. 6). Footprint analysis revealed that the SE DNA is protected normally in the SE complex containing either the mutant RAG1 (S723C) (43) or the mutant RAG2 (K118A/K119A) (41) (Fig. 6A). However, when the pre-cleavage complex was analyzed, protection in the coding region was significantly reduced with the joining mutants (Fig. 6B). In contrast, the cleavage mutants of RAG1, e.g. E962A (49), showed no reduction of coding region protection (Fig. 6B). We then studied the reconstitution of the post-cleavage type complex, using the joining mutant proteins, because their ability to interact with the coding region was shown to be impaired by footprinting (Fig. 6C). The SE complex was formed first, bead-isolated, and then incubated with the 3′-phosphorylated dsDNA as an artificial CE. As shown in Fig. 6C, the SE complex containing the joining mutant S723C or K118A/K119A, but not the cleavage mutant E962A, failed to incorporate the 3′-phosphorylated dsDNA. The wild-type SE complex retained roughly the same amount of dsDNA as SE, and the SE DNA stably remained in the complex even with the joining mutant proteins. The loss of the coding region interaction may explain, at least in part, the defective phenotype of the joining mutants S723C and K118A/K119A in the post-cleavage phase of recombination.

DISCUSSION

In the present study, we have demonstrated that RSS DNA is protected not only in the signal region but also in the adjacent coding region in the synaptic complex. In contrast to the signal regions, protection patterns in the coding regions are indistinguishable between the 12- and 23-RSSs (Figs. 2A, 2B, and 5C), suggesting that the mode of interaction with RAG proteins in the coding region are basically identical for the 12- and 23-RSSs.

Footprint analysis revealed that the nick on the top strand, but not on the bottom, facilitates the interaction of RSS DNA with the RAG proteins (Fig. 2). It appears that nicking of the RSSs leads to a more stable structure of the synaptic complex. Such a structural change may help activate the next cleavage reaction, i.e. double strand cleavage/hairpin formation. It is possible that nicking at the 7-mer end provides more flexibility for the substrate DNA to interact with the RAG proteins in the signal and coding regions on the same molecule. It is also possible that the nicked 7-mer on the top strand, which represents a natural cleavage intermediate, may resemble the SE in structure and interaction with RAGs: The SE shows higher affinity to the complex than the 7-mer in the intact RSS.

In the single RSS-RAG complex, before the synaptic complex is formed, only a few nucleotides adjacent to the 7-mer are protected by the RAG proteins. Our present study demonstrated that a much larger coding region, of at least 12 bp, is protected in the synaptic complex. What causes such a drastic change in the interaction, and what would be a possible role of the coding-RAG interaction? When we analyzed the synaptic complex containing the shorter coding sequence for the coupled cleavage, we found that the RSS DNA with a 6-bp coding region was efficiently cleaved (data not shown). This is consistent with the previous observation by others (22). In contrast, the longer (>12 bp) coding sequence was necessary for the stable formation of the synaptic complex (Fig. 3B), and at least the 12-bp coding region was protected in the pre-cleavage synaptic complex (Fig. 2). As mentioned above, when the RSS DNA contained the coding region of 6 bp, the complex formation was at the lowest level, but the cleavage took place in the isolated complex at more or less the same level as the control (40-bp coding). Curiously, when the coding sequence was further shortened, RSS DNA was not nicked,² and the synaptic complex formation was increased (Fig. 3B). There seems to be two types of interactions for the RSS DNA in the synaptic complex: one is the SE type and the other is the pre-cleavage type. The pre-cleavage type complex requires the coding region, whereas the SE type complex requires the 7-mer sequence near the end. The SE type complex is inactive for the nicking. We also examined the effect of the coding sequence on the complex formation. Four different coding sequences were analyzed by footprinting for protection in both the 12- and 23-RSSs. In all eight samples, the same extent of the coding region was protected regardless of their sequences. It has been reported that not only the recombination signals but also some coding sequences are involved in the regulation of V(D)J recombination (50). Our present results support such a notion.

It has been proposed that the RAG proteins play an architectural role in holding the substrate DNA or cleaved ends in the synaptic complex (39, 40, 43). This notion was supported in our present study not only by the footprint analysis but also by the binding experiment with dsDNA. We found that the SE complex efficiently incorporates the dsDNA when its 3′-end is phosphorylated (Fig. 4B). Because the footprint pattern of dsDNA is analogous to that of the coding region of the pre-cleavage complex, it is unlikely that the binding of dsDNA to the post-cleavage complex is due to nonspecific interaction. The transposition or the target capture activity does not explain the incorporation of the 3′-phosphorylated dsDNA into the SE complex, because the incorporation is so efficient: by these activities, the dsDNA with a hairpin structure or 3′-OH end could be incorporated into the complex, but with low efficiency (Fig. 4B). Why is the dsDNA incorporated so efficiently into the SE complex only when the 3′-end is phosphorylated? Is the 3′-end of the processed CE phosphorylated in vivo? Although some nucleases are known to generate the 3′-phosphorylated end (51), involvement of such enzymes in V(D)J joining is not yet clear. Both Artemis and RAG proteins may be involved in the CE processing. However, these proteins produce the 3′-OH end after the 3′-end processing (36, 46, 52). After the double strand cleavage, repair factors such as DNA-PKcs, Artemis, Ku70/80, LigaseIV, and XRCC4 are recruited to the cleaved ends. These repair factors may help to maintain the cleaved OH ends in the post-cleavage complex. In our in vitro system, it is possible that the 3′-phosphate of dsDNA may mimic such in vivo interactions, at least in part, although further studies are needed to clarify this issue.

The RAG1 mutant S723C, defective in the joining process, is known to have a reduced hairpin retention after the coupled cleavage (43). Our present results demonstrated that the SE complex containing the S723C protein failed to incorporate the 3′-phosphorylated dsDNA. In the footprint analysis, protection of the coding region was reduced in the pre-cleavage complex containing the S723C RAG1. It is possible that the low hairpin retention by the mutant protein is due to the reduced interaction with the coding region DNA. Similar observations were made with the RAG2 joining mutant, K118A/K119A. It has been reported that the amino acid residue Lys-119 in the RAG2 protein may contact with the substrate DNA, using the phosphate backbone rather than specific bases (53). This is consistent with our observation that the protection of the coding region occurs in a sequence-independent manner. Although the exact nature of the defect of the mutants is still unknown, the failure to retain the coding sequence in the post-cleavage complex may be the reason for the impaired joining observed with mutant proteins. Our present results suggest that the coding region interaction with the RAG proteins is important not only in the pre-cleavage, but also in the post-cleavage process of V(D)J recombination.