Activation of 12/23-RSS-dependent RAG cleavage by hSWI/SNF complex in the absence of transcription

Abstract

Maintenance of genomic integrity during antigen receptor gene rearrangements requires (1) regulated access of the V(D)J recombinase to specific loci and (2) generation of double-strand DNA breaks only after recognition of a pair of matched recombination signal sequences (RSSs). Here we recapitulate both key aspects of regulated recombinase accessibility in a cell-free system using plasmid substrates assembled into chromatin. We show that recruitment of the SWI/SNF chromatin-remodeling complex to both RSSs increases coupled cleavage by RAG1 and RAG2 proteins. SWI/SNF functions by altering local chromatin structure in the absence of RNA polymerase II-dependent transcription or histone modifications. These observations demonstrate a direct role for cis-sequence-regulated local chromatin remodeling in RAG1/2-dependent initiation of V(D)J recombination.

Figure 1. 12- and 23-RSS-Dependent RAG Cleavage of Nucleosome-Assembled Plasmids

(A) Plasmids that contain 12- and 23-RSS (indicated by solid and open triangles, respectively) in deletion orientation. pLPd-Δ has a shorter spacer between the two RSSs and was used as an internal control in RAG cleavage assays. pLPd-G4 and pLPd-(G4)₂ contain one or two sets of five tandem GAL4 sites located close to the 12- or 23-RSS, respectively. RAG1/2-induced coupled cleavage results in cuts at the vertical sides of the RSS triangles and release of the intervening DNA of sizes as indicated. This fragment was detected by Southern blots using a HincII-HindIII (392 bp) fragment as a probe. (B) Chromatin assembly in Drosophila embryo S190 extracts was followed by transcription factor binding in the presence or absence of hSWI/SNF for 30 min. A portion of the reaction was used for RAG cleavage by supplementing the assembly reaction with purified MBP-RAG1, GST-RAG2 core, and HMG2 proteins obtained as described in the Experimental Procedures. After RAG1/2 cleavage, the DNA was purified, fractionated through 1.1% agarose gels, and transferred to nylon membranes, and the cleavage product was detected by hybridization to the probe described in (A). The remainder of the assembly reaction was treated with micrococcal nuclease to confirm chromatin assembly, or with DNase I to analyze transcription factor DNA binding or the generation of DNase I-hypersensitive sites as described. (C) DNA purified from micrococcal nuclease treated assembly reactions was fractionated through 1.1% agarose gels and visualized by ethidium bromide staining. Lane 1, 100 bp DNA ladder; lanes 2–3, chromatin samples digested with decreasing concentrations of micrococcal nuclease. (D) Deletion substrates indicated above the gel were used for RAG cleavage before (lanes 1, 4, 7, and 10) or after chromatin assembly (lanes 2, 5, 8, and 11). Gal4-VP16 was added to a subset of reactions for 30 min prior to RAG cleavage (lanes 3, 6, 9, and 12). The numbers below the lanes show the level of RAG cleavage of assembled DNA in the presence of Gal4-VP16 normalized to that in its absence for each plasmid substrate. A typical experiment of several independent experiments is shown.

Figure 2. hSWI/SNF Enhances Gal4-VP16-Dependent RAG1/2 Cleavage In Vitro

(A) pLPd-(G4)₂ and pLPd-Δ (internal control) were assembled into nucleosomes using S190 extracts. Separate aliquots of the assembly reaction were incubated with Gal4-VP16 or Gal4-DBD in the presence or absence of hSWI/SNF as indicated above the figure, followed by RAG1/2 cleavage. The DNA fragments released after RAG cleavage were detected by Southern blotting and are labeled to the left of the figure. (B) Signal intensities of RAG cleavage products of test plasmids were normalized to the internal control (pLPd-Δ) after phosphorimager quantitation. The efficiency of RAG cleavage (y axis) between conditions is represented after normalization to that of pLPd-(G4)₂ alone (lane 1), which is arbitrarily set to 1. The numbers below each bar corresponds to lane numbers in (A). Error bars represent the standard deviation between three independent experiments derived using TTEST (Microsoft Office Excel, 2003). (C) DNase I footprinting analyses of assembly reactions that contain different combinations of transcription factors and hSWI/SNF complex. An aliquots of each sample was digested with DNase I, and the purified DNA was analyzed by primer extension. The first four lanes are di-deoxy sequencing reactions. The bracket shows the region protected from DNase I digestion by Gal4-derivative binding near the 12-RSS; similar results were obtained with Gal4 binding sites near the 23-RSS. Representative data from one of five experiments is shown. (D) Induction of DNase I-hypersensitive sites by Gal4 fusion proteins. Chromatin assembly was carried out under conditions indicated above the gel, followed by DNase I digestion. Purified DNA was digested with DraI located 3′ of the 23-RSS and at two additional sites within 1 kb. The products were separated by agarose gel electrophoresis and transferred to nylon membranes, which were hybridized to an oligonucleotide probe between the 23-RSS and the first DraI site. Hypersensitive sites at the 12-RSSs (top) and 23-RSSs (bottom) are indicated by triangles. Data shown are representative of five independent experiments.

Figure 3. Gal4-VP16 Increases Restriction Endonuclease Access to RSS

(A) Chromatin assembly was carried out using S190 extracts followed by incubation with Gal4-DBD (lanes 4–6) or Gal4-VP16 (lanes 7–9). After transcription factor binding, aliquots were treated with increasing concentrations of BamHI, which cuts at a site between the Gal4-binding sites and the 12-RSS as indicated on the top line. Purified DNA was then cut to completion with AlwNI and fractionated by agarose gel electrophoresis, and the BamHI-AlwNI fragment was detected by Southern blotting using the indicated probe. Data shown are representative of three independent experiments. (B) The ratio of BamHI cut to uncut DNA (y axis) was averaged for three experiments at the enzyme concentration corresponding to lanes 3, 6, and 9 of (A) (indicated on the x axis). Error bars represent the standard deviation between experiments derived using TTEST (Microsoft Office Excel, 2003).

Figure 4. Analysis of V(D)J Recombination Inversion Substrates In Vitro

(A) Inversion substrates were generated as described in the Experimental Procedures. Except for the orientation of 12-RSS, all other parts of the substrate plasmids are the same as the deletion substrates. The sizes of the fragments released after RAG cleavage change slightly compared to plasmids described in Figure 1 as noted. (B) Inversion substrates and internal control (pLPi-Δ) were assembled into chromatin using S190 extracts, incubated with Gal4 derivatives in the presence or absence of hSWI/SNF, and subjected to RAG cleavage. The released fragment was detected by Southern blotting as described in Figures 1 and 2. Data shown are representative of three independent experiments.

Figure 5. 12- and 23-RSS-Dependent RAG1/2 Cleavage Using Recombinant Proteins

(A) pLPd-(G4)₂ and pLPd-Δ were assembled into chromatin using ACF purified from SF9 cells, NAP1 and topoisomerase I (expressed in bacteria), and Drosphila histones. Chromatin assembly was verified by micrococcal nuclease digestion (left panel). After assembly the chromatin was incubated with transcription factors and hSWI/SNF as required. Induction of DNase I-hypersensitive sites (right panel) at the 12- and 23-RSSs (top and bottom triangles, respectively) was assessed as described in the Figure 2 legend. Data shown are representative of three independent experiments. (B and C) Following transcription factor binding, the substrates were treated with RAG1/2, and cleavage products were visualized by Southern blotting. Data shown are representative of three independent experiments whose average is shown in (C). Quantitation was carried out as described in Figure 2B. Numbers below the bars correspond to lane numbers in (B). Error bars represent the standard deviation between three independent experiments derived using TTEST (Microsoft Office Excel, 2003). (D–F) Deletion substrates were assembled into chromatin using recombinant histone octamer and S190 extracts. Assembly was verified by micrococcal nuclease (Mnase) digestion (D), and after incubation with transcription factors and hSWI/SNF as indicated, RAG1/2 cleavage assays were performed (E). The fragment released after 12/23 coupled cleavage was visualized by Southern blotting as described in Figure 2B. pLPd-Δ served as the normalizing control. A representative experiment out of four is shown. All four are averaged in (F) with error bars representing standard deviation between experiments derived using TTEST (Microsoft Office Excel, 2003).

Figure 6. Transcription-Coupled Cleavage of 12- and 23-RSS by RAG1/2 In Vitro

(A) pLPd-(G4)₂ plasmid was assembled into chromatin using S190 extracts followed by incubation with Gal4-VP16 and SWI/SNF, as indicated in (B) and (C). Nuclear extracts from BJAB B lymphoma cells were used to initiate transcription for 30 min followed by addition of RAG1, RAG2, and HMG2 for additional 2 hr. RAG1/2 cleavage was assayed by Southern blotting and RNA purified for RT-PCR analysis (Figure S4). (B) Unassembled plasmids (lanes 1, 2, 6, 7, and 11) or chromatin-assembled plasmids (all other lanes) were used in RAG1/2 cleavage assays according to the experimental design outlined in (A). The control plasmid pLPd-Δ was included just prior to addition of RAG1/2 and was therefore not assembled into chromatin. BJAB extracts were pretreated with 40 μg/ml α-amanitin for 10 min prior to use in the indicated reactions (lanes 7–10). RAG1/2 cleavage products were detected by Southern hybridization (indicated by arrows). Data shown are representative of three independent experiments. (C) Signal intensities of RAG cleavage products from pLPd-(G4)₂ were normalized to that of pLP-Δ after phosphorimager quantitation. The efficiency of RAG cleavage (y axis) between conditions is represented after normalization to the value in lane 3, which is set to 1. Numbers below each bar correspond to the lane numbers of the autoradiogram in (B). Error bars represent the standard deviation between three independent experiments derived using TTEST (Microsoft Office Excel, 2003).