The RAG1 homeodomain recruits HMG1 and HMG2 to facilitate recombination signal sequence binding and to enhance the intrinsic DNA-bending activity of RAG1-RAG2

Abstract

V(D)J recombination is initiated by the specific binding of the RAG1-RAG2 (RAG1/2) complex to the heptamer-nonamer recombination signal sequences (RSS). Several steps of the V(D)J recombination reaction can be reconstituted in vitro with only RAG1/2 plus the high-mobility-group protein HMG1 or HMG2. Here we show that the RAG1 homeodomain directly interacts with both HMG boxes of HMG1 and HMG2 (HMG1,2). This interaction facilitates the binding of RAG1/2 to the RSS, mainly by promoting high-affinity binding to the nonamer motif. Using circular-permutation assays, we found that the RAG1/2 complex bends the RSS DNA between the heptamer and nonamer motifs. HMG1,2 significantly enhance the binding and bending of the 23RSS but are not essential for the formation of a bent DNA intermediate on the 12RSS. A transient increase of HMG1,2 concentration in transfected cells increases the production of the final V(D)J recombinants in vivo.

FIG. 1

RAG1 interacts with HMG1. Purified GST-RAG1 (about 0.2 μg in 160 μl) was incubated with Sepharose beads bearing immobilized, bacterially expressed tailless HMG1 (M1-V176) (A) or control beads bearing immobilized cytochrome c (B). Conversely, RAG1ΔN380, a GST fusion derivative of RAG1 that retains enzymatic activity, was bound to Sepharose-glutathione beads and used to pull down soluble HMG-I(Y), an HMG protein structurally diverse from HMG1,2, and HMG1 (C). Input (I), bound (B), and free (F) RAG1 and HMG proteins were detected by Coomassie blue staining. The protein in lane 8 is GST, as demonstrated by the appearance of the same protein in lane 10 (control [C]), where RAG1ΔN380 beads were directly boiled in loading buffer without prior exposure to HMG-I(Y). The RAG1ΔN380 protein itself migrates much higher and is not shown in the gel.

FIG. 2

HMG1,2 directly interact with RAG1 through its HD. (A) Purified, eukaryotically expressed, GST-fused RAG1ΔN330, RAG1ΔN456, or RAG2ΔC (aa 1 to 388) was incubated with Sepharose beads bearing immobilized, bacterially expressed tailless HMG1 (M1-V176) or full-length HMG2, as described in Materials and Methods. The input (I), bound (B), and free (F) materials were immunoblotted with an anti-GST antibody, following SDS-PAGE. (B) Schematic representation of full-length RAG1 and derivatives. RF, ring finger (aa 288 to 330); hdh, homodimerization helices (aa 340 to 351); ZFA, zinc finger A (aa 353 to 374); HD, homeodomain (aa 389 to 446); ZFB, zinc finger B (aa 727 to 750) (3, 41).

FIG. 3

The RAG1 HD directly interacts with HMG boxes A and B of HMG1 and -2. In vitro-translated HMG1 (A) and HMG2 (B) derivatives were incubated with Sepharose beads bearing immobilized, eukaryotically expressed RAG1 HD (Fig. 2B). The input (I), bound (B), and free (F) materials were visualized by autoradiography following SDS-PAGE. Schematic representations of full-length HMG1 and -2 are shown below panels A and B, respectively. The HMG boxes are stippled, and the acidic tails are hatched. The derivatives are identified by their first and last amino acids.

FIG. 4

HMG1 stimulates RAG1/2 binding through the HD of RAG1. (A) EMSAs with a radiolabelled 23RSS oligonucleotide probe. RAG1 deletion derivatives and RAG2ΔC (50 ng each) and 40 ng of HMG1 were added as indicated (+). Lanes 11 to 14 represent a longer exposure of lanes 7 to 10. (B) In vitro cleavage assays. A 23RSS radiolabelled oligonucleotide probe was 5′ end labelled on the upper strand and incubated with 50 ng of each RAG1 derivative and 50 ng of RAG2ΔC; 40 ng of HMG1 was added where indicated (+). The products of the reaction, analyzed in a 16% polyacrylamide-6 M urea gel, are indicated. HP, hairpin; N, nick.

FIG. 5

HMG1 enhances nonamer-dependent binding to the RSS. EMSAs with radiolabelled oligonucleotide probes carrying either wild-type (wt) or mutated RSS sequences. (A) Schematic representation of the RSS and the positions of the mutations employed. The mutations are further described in Materials and Methods. (B and C) RAG1ΔN380 and RAG2ΔC (50 ng each) were incubated (+) with 23RSS (B) or 12RSS (C) radiolabelled substrates carrying the indicated mutations. Forty nanograms of HMG1 was added (+) where indicated.

FIG. 6

Bending of the RSS DNA by RAG1/2 and HMG1. (A) Schematic representation of the isomeric probes used in the circular-permutation analysis. (B and D) EMSAs with equimolar amounts of the isomeric probes shown in panel A carrying 12RSS (B) or 23RSS (D) sequences (the asterisk indicates a probe artifact). The probes were 5′ end labelled to the same specific activity and incubated with 50 ng each of RAG1ΔN380 and RAG2ΔC. Forty nanograms of HMG1 was added where indicated. (C) The locus and extent of bending were estimated as described in Materials and Methods. We used several gels for each estimate; shown is one example of the data obtained for 12RSS. Rb/Rf (vertical axis) is the relative mobility of bound versus free DNA. D/L (horizontal axis) is the fractional distance of the center of the RSS from the 5′ end of the probe. M, MluI; N, NheI; X, XhoI; E, EcoRV; S, StuI; B, BamHI.

FIG. 7

In vivo stimulation of RAG1 and RAG1/2 binding by HMG1,2. The binding of RAG proteins to a multimerized RSS in vivo can be measured by means of a mammalian one-hybrid assay (13). RAG1 and RAG2 proteins were transformed into transcriptional activators by fusing them to the VP16 transactivation domain. The occupancy of the binding site by the RAG-VP16 fusion proteins is proportional to the expression in 293 cells of a reporter luciferase gene driven by a cytomegalovirus minimal promoter. The binding site contains either 8 copies of the consensus 12RSS or 10 copies of the consensus 23RSS or mutated forms of the RSS where indicated. Comparable expression of proteins encoded by transfected plasmids was ascertained by Western blotting (see Fig. 8 for examples). The values are normalized to the expression of the reporter construct in the absence of RAG-VP16 fusion proteins. The plotted values represent the means of 8 to 10 individual experiments. The standard deviation was <8% of the mean and is not indicated. (A) Transactivation of the luciferase gene (12RSS construct) by RAG1-VP16, RAG2-VP16, or a combination of RAG1-VP16 and RAG2-VP16 as indicated below the diagram, either alone or in combination with cotransfected HMG1, HMG2, HMG-I(Y), or TCF-1b. (B) Transactivation of the reporter construct indicated below the diagram by RAG1-VP16 alone or RAG1-VP16 plus HMG1 or HMG2. The mutant 7-mer contains the sequence CGACGTC; the mutant 9-mer contains the sequence ACACTGGTA. wt, wild type. (C) Effect of HMG1, HMG2, HMG-I(Y), and TCF-1b on transactivation by RAG1/2-VP16. The transactivation by RAG1-VP16 alone or RAG2-VP16 alone is also reported for comparison. The reporter constructs are indicated below the corresponding groups of bars. (D) Transactivation of the luciferase gene (12RSS construct) by different RAG1-VP16 fusion proteins, either alone, in combination with RAG2-VP16, or in combination with RAG2-VP16 and in the presence of HMG1. (E) Schematic representation of the reporter construct. RAG1-VP16 is indicated as R1, and RAG2-VP16 is indicated as R2. The grey oval represents the VP16 transactivation domain.

FIG. 8

HMG1,2 increase the yield of V(D)J recombination products in vivo. 293T cells were cotransfected with the recombination substrate pJH299 and the expression constructs for the indicated proteins. The RAGs were GST tagged, while HMG1, HMG2, HMG-I(Y), and TCF-1b were HA tagged. (A) Recombination efficiency was measured by PCR analysis (20 cycles) of the recovered plasmid, using primers that amplify only the recombined sequences (signal joints [SJ]), as described in Materials and Methods. (B) Loading control of panel A. A fragment of the CAT gene, present in the pJH299 recombination substrate, was amplified under the same conditions as for panel A. (B′) Templates as in panel B were diluted 10-fold prior to PCR. The amount of PCR product obtained was proportional to the template input. (C) Expression control of panel A. Total cellular extract was immunoblotted with anti-GST and anti-HA antibodies, following SDS-PAGE. (C′) The same membrane shown in panel C was stripped and immunoblotted with anti-HMG1 antibodies. (C") The same membrane shown in panel C was stripped once more and immunoblotted with anti-HMG2 antibodies. (D) The autoradiogram shown in panel A and those from three more experiments were scanned, and the amounts of recombination products were normalized to the efficiency of V(D)J recombination with RAG1/2 alone, which was set to 1. The error bars indicate standard deviations.