Footprint Analysis of Recombination Signal Sequences in the 12/23 Synaptic Complex of V(D)J Recombination

Abstract

In V(D)J joining of antigen receptor genes, two recombination signal sequences (RSSs), 12-RSS and 23-RSS, are paired and complexed with the protein products of recombination-activating genes RAG1 and RAG2. Using magnetic beads, we purified the pre- and postcleavage complexes of V(D)J joining and analyzed them by DNase I footprinting. In the precleavage synaptic complex, strong protection was seen not only in the 9-mer and spacer regions but also near the coding border of the 7-mer. This is a sharp contrast to the single RSS-RAG complex where the 9-mer plays a major role in the interaction. We also analyzed the postcleavage signal end complex by footprinting. Unlike what was seen with the precleavage complex, the entire 7-mer and its neighboring spacer regions were protected. The present study indicates that the RAG-RSS interaction in the 7-mer region drastically changes once the synaptic complex is formed for cleavage.

V(D)J recombination plays key roles in the activation and diversification of the antigen receptor genes (39). Two pairs of recombination signal sequences (RSSs), 7-mer (CACAGTG) and 9-mer (ACAAAAACC), are required for this type of recombination (22, 30). Furthermore, the spacer separating the 7-mer and 9-mer is either 12 or 23 bp in length. V(D)J recombination takes place between two RSSs with different spacer lengths, one containing a 12-bp spacer (12-RSS) and the other containing a 23-bp spacer (23-RSS) (7, 31, 32). This is the so-called 12/23-bp spacer rule for V(D)J recombination (4).

In the initial step of V(D)J joining, each RSS is individually recognized by the protein products of the recombination-activating genes RAG1 and RAG2 (3, 15, 26). This RSS-RAG interaction primarily involves the Hin homeodomain of the RAG1 protein and the 9-mer region of RSS DNA (6, 26, 34, 35). The next step is the formation of the synaptic complex, where the two RSS-RAG complexes are brought together (8, 24, 25, 40) in the presence of a DNA-bending protein, HMG1 (18, 33, 41, 42, 43). When the 12/23 joining rule is satisfied, RSS DNA is then cleaved in two successive steps, nicking and hairpin formation (23). After the cleavage, the two coding ends and the two signal ends (SEs) recombine to form a coding joint and a signal joint, respectively, by DNA repair mechanisms (5, 10, 12, 13, 14, 21, 27, 28, 29, 38, 44).

In order to study the RSS-RAG interaction during the process of V(D)J joining, we purified the pre- and postcleavage complexes by using RSS DNA and RAG proteins in vitro and analyzed them by DNase I footprinting. The present results indicate that the mode of RAG-RSS interaction in the 7-mer drastically changes once the two RSS-RAG complexes are associated to form the synaptic complex.

MATERIALS AND METHODS

Preparation of proteins.

Truncated RAG1 and RAG2 fused with maltose binding protein were expressed in Sf.9 cells by using a baculovirus vector and were purified as previously described (40). Porcine HMG1 protein was prepared as previously described (1).

DNA substrates.

Oligonucleotides were chemically synthesized and were purified by high-pressure liquid chromatography or by electrophoresis in a denaturing polyacrylamide gel. The top strand was labeled by filling in the 3′ end with either ³²P-labeled nucleoside triphosphates or biotin-14-dCTP (BRL) by using a Klenow fragment (3′-5′ exonuclease free; New England Biolabs). The bottom strand was labeled by 5′ phosphorylation with T4 polynucleotide kinase (New England Biolabs). The labeled bottom strand was then annealed to the top strand oligonucleotide. All substrates were separated 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 reversed-phase column chromatography (Elutip-d; Schleicher & Schuell).

Oligonucleotides.

RSS substrates were prepared as follows. The 12-RSS top strand (68-mer), 5′-CCTGCGCTGAATTCGTCTTACACAGTGCTCCAGGGCTGAACAAAAACCTCCTAGGGTTGCAGCTGACT-3′, was annealed to the 12-RSS bottom strand (69-mer), 5′-GAGTCAGCTGCAACCCTAGGAGGTTTTTGTTCAGCCCTGGAGCACTGTGTAAGACGAATTCAGCGCAGG-3′ (7-mer and 9-mer signal sequences are underlined). Similarly, the 23-RSS top strand (79-mer), 5′-CCTCCTGGGAATTCGTCTTACACAGTGGTAGTACTCCACTGTCTGGGTGTACAAAAACCTCCTAGGGTTGCCATGGACT-3′, and the 23-RSS bottom strand (80-mer), 5′-GAGTCCATGGCAACCCTAGGAGGTTTTTGTACACCCAGACAGTGGAGTACTACCACTGTGTAAGACGAATTCCCAGGAGG-3′, were annealed. Wild-type SE sequences are as follows: 12-SE top strand (48-mer), 5′-CACAGTGCTCCAGGGCTGAACAAAAACCTCCTAGGGTTGCAGCTGACT-3′; 12-SE bottom strand (49-mer),5′-GAGTCAGCTGCAACCCTAGGAGGTTTTTGTTCAGCCCTGGAGCACTGTG-3′; 23-SE top strand (59-mer), 5′-CACAGTGGTAGTACTCCACTGTCTGGGTGTACAAAAACCTCCTAGGGTTGCCATGGACT-3′; and 23-SE bottom strand (60-mer), 5′-GAGTCCATGGCAACCCTAGGAGGTTTTTGTACACCCAGACAGTGGAGTACTACCACTGTG-3′ (7-mer and 9-mer sequences are underlined). In the 7-mer variant of 12-SE (7-mer-less SE), the 7-mer (5′-CACAGTG-3′) was changed to 5′-GCTGAAC-3′. In the 12-SE mutant lacking the 9-mer (9-mer-less SE), the 9-mer (5′-ACAAAAACC-3′) was changed to 5′-CACCCCCAA-3′. In the double-9-mer SE, the 5′ half of the 23-SE sequence containing the 7-mer (28 bp) was replaced with the 5′ half of the 12-SE sequence (28 bp) containing its 7-mer and 9-mer.

Isolation of pre- and postcleavage complexes.

End-labeled DNA (about 0.04 pmol; 4 × 10⁴ to 8 × 10⁴ cpm) was incubated with biotinylated DNA (0.2 pmol) and RAG1 and RAG2 proteins (400 ng each) with or without HMG1 protein (160 ng) at 37°C for 90 to 120 min in 20 μl of binding buffer containing 25 mM MOPS (morpholinepropanesulfonic acid)-KOH (pH 7.0), 5 mM Tris-HCl (pH 8.0), 2.4 mM dithiothreitol, 90 mM potassium acetate, 30 mM KCl, a 2 μM concentration of double-stranded oligonucleotide not related to the RSS sequence (5′ GGCCGCCAACCCCATGGAGCT 3′ and 5′ CCATGGGGTTGGC 3′), 10 mM CaCl₂ or MgCl₂, 0.1 mg of bovine serum albumin per ml, and 2% glycerol. To isolate the complexes, streptavidin-coated magnetic beads (Dynabeads M-280; 100 μg in 10 μl of binding buffer) were then added, and the reaction mixture was incubated for 30 min at 37°C. After incubation, magnetic beads and supernatant were separated by using a magnet stand. The beads were washed three to four times at room temperature with 50 μl of binding buffer and then resuspended in 30 μl of binding buffer. The washed magnetic beads and the supernatant were subjected to DNase I footprinting.

DNase I footprinting.

DNA samples were preheated at 37°C for 6 min and digested with 5 U of DNase I (Stratagene) for 2 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 on an 8% polyacrylamide DNA-sequencing gel.

RESULTS

Isolation and footprint analysis of SE complexes.

After double-strand breakage of the 12- and 23-RSSs, the two SEs (12-SE and 23-SE) stay together with RAG proteins in the form of an SE complex (2, 17). This complex can be reconstituted in vitro with two types of synthetic DNAs, 12-SE and 23-SE, and RAG1 and RAG2 proteins (16). For the isolation of the SE complex, we incubated ³²P-labeled SE DNA with biotinylated partner SE DNA in the presence of RAG1 and RAG2 proteins (Fig. 1A). Figure 1B shows the formation of SE complexes with four different combinations of RSS spacers, 12/12, 12/23, 23/12, and 23/23. The SE complexes were then isolated with streptavidin-coated magnetic beads. Formation of the SE complex was facilitated by the DNA-bending protein HMG1 and occurs more efficiently in the heterologous combinations of RSS spacers following the 12/23 rule (Fig. 1B). While the 9-mer mutant did not completely abolish the formation of the complexes, no complex was formed when the 7-mer was mutated (Fig. 1C).

FIG. 1.

Isolation of the reconstituted SE complex. (A) Schematic diagram of the purification procedure for the reconstituted SE complex. Triangles indicate RSSs. ³²P-labeled SE DNA was paired with biotinylated partner SE DNA and complexed with the RAG1 and RAG2 proteins in the presence of the DNA-bending protein HMG1. The complex was purified with streptavidin-coated magnetic beads. (B) Formation of the SE complex. ³²P-labeled SE DNA was paired with biotinylated partner SE DNA in four different combinations of RSS spacers, 12/12, 23/12, 12/23, and 23/23. Counts of ³²P incorporated into the complex are divided by the input counts (% bound). (C) Formation of the SE complex with variant 12-SE substrates. In the 9-mer-less SE (9-less), the nonamer signal (ACAAAAACC) was changed to CACCCCCAA. In the 7-mer-less SE (7-less), the heptamer signal (CACAGTG) was changed to GCTGAAC. The double-9-mer SE contains two nonamer signals 12 and 23 bp downstream from the heptamer signal. ³²P-labeled variant or wild-type SE DNA was paired with biotinylated partner SE DNA. The RAG1 mutant for the DDE motif (D708N) was also tested. Counts of ³²P incorporated into the complex are divided by the input counts (% bound).

The isolated complexes were then subjected to DNase I footprinting. Figure 2A shows the footprint patterns of the 12-SE DNA containing a 12-bp spacer. The ³²P-labeled 12-SE was paired with a biotinylated 23-SE in the presence of RAG1 and RAG2 proteins with or without HMG1 protein. The RAG-bound 12-SE gave strong protection in the 7-mer, 9-mer, and neighboring spacer regions on both the bottom and top strands. One residue in the spacer of the bottom strand (10 bases downstream from the 7-mer) became highly susceptible to DNase I digestion.

FIG. 2.

Footprint analysis of the reconstituted SE complex. (A) ³²P-labeled 12-SE DNA was paired with the biotinylated partner 23-SE, complexed with RAG1/RAG2 and HMG1 proteins, and analyzed by DNase I footprinting. The complex (Complexed) was purified with streptavidin-coated magnetic beads, digested partially with DNase I, and electrophoresed in DNA-sequencing gels. The input DNA (Input) and the SE DNA fraction that did not bind to magnetic beads (Unbound) were also analyzed as negative controls. The locations of 7-mer and 9-mer signals are indicated, and asterisks indicate ³²P labels. Both top and bottom strands were analyzed. (B) ³²P-labeled 23-SE DNA was paired with the biotinylated partner 12-SE, complexed with RAG1/RAG2 and HMG1 proteins, and analyzed by DNase I footprinting.

The labeled 23-SE paired with a biotinylated 12-SE was also analyzed by footprinting (Fig. 2B). Three protected regions were evident in the 7-mer, spacer, and 9-mer regions on both the top and bottom strands. As in the 12-SE, one residue in the bottom-strand spacer (10 bases from the 7-mer) became sensitive to DNase I digestion. Unlike the 12-SE, a residue in the top-strand spacer (14 bases downstream from the 7-mer) also became susceptible. It is interesting that two residues approximately one helical turn downstream from the susceptible residue remain sensitive to DNase I digestion, indicating that these two sensitive sites are located on the same side of double-stranded DNA.

Since HMG1 was shown to facilitate the formation of the SE complex with RAG1 and RAG2 proteins (Fig. 1B), we examined the effect of this protein on the footprint patterns. To our surprise, HMG1 did not significantly change the footprint patterns of 12-SE in the complex, but in the 23-SE, the enhancement of DNase I sensitivity within the spacer was reduced on the top strand when HMG1 was present (Fig. 2B). These results are consistent with previous observations of the single RAG-RSS interaction (37, 43): only the interaction of 23-RSS is affected by HMG1 protein, and HMG1 interacts with the side of the RSS DNA opposite the one contacting the RAG proteins.

The 7-mer is essential to the formation of the SE complex.

In order to study how the SE complex is formed following the 12/23 rule, we analyzed the double-9-mer SE, a variant SE containing two 9-mer sequences that are 12 and 23 bp downstream from the 7-mer. When this SE was paired with 12-SE, three regions, the 7-mer, spacer, and 9-mer regions, were protected in the footprint pattern (Fig. 3A), which was quite similar to what was seen when the wild-type 23-SE was paired with 12-SE (Fig. 2B). When this variant SE was paired with 23-SE, the protection pattern was indistinguishable from that obtained for the wild-type 12-SE paired with 23-SE (Fig. 2A). We then analyzed SE complexes that violate the 12/23 joining rule. In the 12/12 combination, only the 7-mer and its neighboring spacer regions were protected, while protection in the 9-mer region was much reduced (Fig. 3B). A similar observation was made with the 23/23 combination (data not shown).

FIG. 3.

Footprint analysis of the SE complexes containing variant SEs. (A) Double-9-mer SE was labeled with ³²P and paired with the biotinylated partner 12-SE or 23-SE. This variant SE (double-9) contains two 9-mer sequences (ACAAAAACC) 12 and 23 bp downstream from the 7-mer. Samples were ³²P-labeled at the 3′ end of the top strand. The locations of 7-mer and 9-mer signals are indicated, and asterisks indicate ³²P labels. (B) ³²P-labeled 12-SE DNA was paired with the biotinylated partner 12-SE, violating the 12/23 rule. The 12/23 combination was analyzed in parallel as a positive control. (C) ³²P-labeled 9-mer-less SE (9-less) was paired with the biotinylated partner 12-SE or 23-SE. The 9-mer signal (ACAAAAACC) was changed to CACCCCCAA.

We also examined two variant SE substrates. One was the 7-mer-less SE, and the other was the 9-mer-less SE. As already mentioned, when the 7-mer-less SE was paired with the wild-type 23-SE, no SE complex was detected with the magnetic-bead method (Fig. 1C), suggesting that the 7-mer sequence is essential in the formation of the complex. In contrast, when the 9-mer-less SE was paired with the wild-type 23-SE, a reduced amount of the complex was formed (Fig. 1C). Figure 3C shows the footprint patterns of the 9-mer-less SE paired with 12-SE or with 23-SE. In either case, the 7-mer and its neighboring spacer regions were normally protected even without the 9-mer. The mutated 9-mer region was not protected; however, footprint patterns in the spacer regions were basically the same as those found for the double-9-mer SE paired with 12-SE or with 23-SE (Fig. 3A). Our footprint study indicated that even without the 9-mer, the 9-mer-less SE interacts with the partner SE (12- or 23-SE) to follow the 12/23 rule; the 9-mer-less SE behaves as if it were the 23- or 12-SE. As a result, the 9-mer-less SE, when paired with the 12-SE, shows a spacer footprint similar to that of the 23-SE, where the hypersensitive region divides the protected region into two parts. Protection of the distal part is not evident in the 9-mer-less SE, probably due to the absence of the 9-mer. Since the 12-SE has no hypersensitive region in the protected spacer, the 9-mer-less SE showed more extensive protection in the spacer when it was paired with the 23-SE than when it was paired with the 12-SE.

These results indicate that RAG interaction with the 9-mer depends not only on the sequence but also the location of the 9-mer. It appears that when the SE complex is formed, the 7-mer is recognized by the RAG proteins first and the 9-mer is recognized following the 12/23 joining rule.

Footprint analysis of the precleavage complexes.

We also analyzed the precleavage complex formed with the RSS DNA that contains flanking sequences outside of the 7-mer. To prevent RSS cleavage, we used Ca²⁺ instead of Mg²⁺ in the buffer. The ³²P-labeled 12-RSS was paired with the biotinylated 23-RSS and incubated with RAG1/RAG2 and HMG1 proteins to form the precleavage complex. The complex was isolated with magnetic beads and analyzed by DNase I footprinting (Fig. 4A). Protection patterns in the 9-mer and spacer regions were similar to those for the reconstituted postcleavage SE complex (Fig. 2A), but the patterns were different in the 7-mer region. Although the entire 7-mer is significantly less protected in both the top and bottom strands, strong protection was seen around the coding border but not on the distal side of the 7-mer (Fig. 4A). We also found enhanced bands in the coding flank of the top strand (12 and 13 bases upstream from the 7-mer). It is quite remarkable that RAG protection reaches so far into the flanking region. When the complex was incubated with Mg²⁺ (Fig. 4B), RSS was cleaved and showed a footprint pattern similar to that of the SE in the reconstituted postcleavage complex (Fig. 2A). In our system, about two-thirds of the substrate was cleaved and the rest was mostly nicked at the top strand. Therefore, the footprint patterns of the precleavage complex with Mg²⁺ looked similar to those of the reconstituted SE, particularly for the nicked strand.

FIG. 4.

Footprint analysis of precleavage and postcleavage RSSs in the synaptic complex. (A) ³²P-labeled 12-RSS DNA was paired with biotinylated partner 23-RSS and complexed with RAG1/RAG2 and HMG1 proteins. The complex was isolated with streptavidin-coated magnetic beads in the presence of Ca²⁺ and analyzed by DNase I footprinting. (B) The precleavage synaptic complex was isolated in the presence of Mg²⁺ and analyzed by DNase I footprinting. Under the Mg²⁺ condition, RSS DNA was cleaved, yielding the SE. The input DNA (Input) and the DNA fraction that did not bind to magnetic beads (Unbound) were analyzed as negative controls. The locations of 7-mer and 9-mer signals are indicated, and asterisks indicate ³²P labels.

In order to study the effect of DDE mutations of RAG1 on the RSS-RAG interaction, we then analyzed the precleavage complex formed with the D708N mutant protein (11, 19, 20). When the labeled 12-RSS was paired with the biotinylated wild-type 23-RSS (Fig. 5A), the protection patterns in the 9-mer and spacer regions were basically the same as that seen with the wild-type RAG1. However, the protection of the 7-mer was not evident with the mutant RAG1. Furthermore, unlike the wild-type complex, the mutant complex did not cleave the bottom strand of RSS in the presence of Mg²⁺ nor did it show protection in the 7-mer region (Fig. 5B). We also tried to study the footprint of the SE complex with the D708N mutant. However, we could not recover a sufficient amount of the SE complex, probably due to the poor 7-mer interaction with the mutant RAG1 (Fig. 1C).

FIG. 5.

Footprint analysis of the synaptic complex with the mutant RAG1 protein. ³²P-labeled 12-RSS DNA was paired with the biotinylated partner 23-RSS and complexed with RAG1/RAG2 and HMG1 proteins. For the RAG1 protein, either the DDE mutant (D708N) or the wild type (WT) was used. The complex was isolated with streptavidin-coated magnetic beads in the presence of Ca²⁺ (A) or Mg²⁺ (B). The input DNA (Input) and the DNA fraction that did not bind to magnetic beads (Unbound) were analyzed as negative controls. The locations of 7-mer and 9-mer signals are indicated, and asterisks indicate ³²P labels.

DISCUSSION

In the present study, we purified an SE complex with magnetic beads and analyzed it by DNase I footprinting. Our data as well as that of others (2, 16) demonstrate that the SE complex is formed efficiently in the heterologous combinations of two RSS spacers, 12/23 or 23/12 (Fig. 1B). If combinations of spacers violate the 12/23 rule, the formation of the SE complex is reduced by three- to fivefold and DNase I protection in the 9-mer region is greatly reduced in both SE DNAs (Fig. 3B). It appears that in the SE complex, the RAG-9-mer interaction occurs only when the 12/23 rule is satisfied (Fig. 3A).

As for the 12/23 specificity in the formation of the SE complex, the preference for the 12/23 SE pair was only threefold more than that for the 12/12 SE pair. As reported previously (16), the cleavage preference was quite high: 20-fold more for the 12/23 pair than for the 12/12 pair. It is possible that the 12/23 rule is more strictly followed for the RSS cleavage than for the formation of the synaptic complex. It is also possible that the difference is due to the strong interaction of the cleaved 7-mer with the RAG proteins in the SE complex. The SE complex can be formed even with the 9-mer-less SE when it is paired with the wild-type SE, while such a complex cannot be detected with the uncleaved RSSs, i.e., 9-mer-less RSS and wild-type RSS (our unpublished observations).

Unlike the 9-mer, the 7-mer region was strongly protected in all types of SE complexes. Even with the 9-mer-less SE, protection was found in the 7-mer region (Fig. 3C), while the 7-mer-less SE failed to form the SE complex (Fig. 1C). This is in sharp contrast to the single RSS-RAG interaction, where the 9-mer plays a major role (Fig. 6). When the SE complex is formed, the interaction in the 9-mer seems to take place based on the location of the 9-mer, satisfying the 12/23 joining rule. In contrast, the 7-mer interaction appears to occur regardless of the 9-mer.

FIG. 6.

Schematic diagrams of the protection sites. The 12-RSS in the single RSS-RAG complex, precleavage synaptic complex, and postcleavage SE complex was analyzed by DNase I footprinting. Protected phosphates are shown in red, and hypersensitive sites are in green. The 7-mer and 9-mer sequences are shaded in light purple. Both sides (Front and Back) of the double-stranded RSS are shown. The footprint analysis of the single RSS-RAG complex has been described previously (26).

Although the 9-mer interaction is essential to the single RSS-RAG complex, weak interaction has been detected in the 7-mer region by DNase I footprinting (26), UV cross-linking (9, 24, 36), and the modification interference assay (26, 35, 37). In the present study, we analyzed the synaptic complex by DNase I footprinting and detected strong protection in the 7-mer region. The pairing of two RSSs may be necessary to correctly place the 7-mer in the active site of the RAG1 protein. Even when the 12/23 rule is violated, 7-mer protection is normally found in the SE complex. Interestingly, however, the 9-mer region is not protected if the 12/23 rule is not satisfied. The 9-mer interaction as well as the 7-mer interaction may be needed for the RSS cleavage by RAG proteins. Using the DDE motif mutant (11, 19, 20), we have shown that the coding border of the 7-mer is not protected by the mutant RAG1 in the precleavage synaptic complex (Fig. 5A). To solve the smearing problem in the mutant footprint, we performed the experiment several times. However, we repeatedly got smearing, possibly due to salts and other impurities contained in the sample as a result of the low recovery of the mutant complex. The formation of the precleavage complex with the mutant RAG1 was quite inefficient and represented no more than one-tenth of the SE complex formation with the wild-type RAG1. The smearing problem was not solved when Mg²⁺ was used (Fig. 5B) nor when the top strand was analyzed (our unpublished observations). We also tried to study the SE complex with the D708N mutant, because the SE complex is easier to isolate and gives better footprint patterns than the precleavage complex. However, we could not obtain enough SE complex with the mutant RAG1. As we have shown in the present study, the 7-mer-RAG interaction is essential for the formation of the SE complex. The inefficient formation of the mutant SE complex also supports this conclusion. Footprint analyses with other RAG mutants are needed to further study the effect of the DDE motif in the 7-mer-RAG interaction.

After the cleavage of the 7-mer, the SE is retained in the postcleavage SE complex, while the coding end is released (2, 17). Our footprint analysis demonstrated that the entire 7-mer region is protected in the SE complex, while the 7-mer-RAG interaction was found in the coding-proximal region in the precleavage complex (Fig. 6). It appears that the mode of 7-mer interaction is different before and after the RSS cleavage in the synaptic complex.

So far, the footprint studies have been quite helpful in identifying the interaction sites in the RSS-RAG complex. However, it is yet to be clarified exactly how the RSSs and RAG proteins are assembled and how the 12/23 rule is ensured in the synaptic complex. If the RSS-RAG complex is crystallized, three-dimensional studies of it will shed light on not only the RSS-RAG interaction but also the structural features of the 12/23 joining rule.