Cutting Edge: SWI/SNF Mediates Antisense Igh Transcription and Locus-Wide Accessibility in B Cell Precursors

Abstract

The stepwise process of Ag receptor gene assembly, termed V(D)J recombination, is coordinated during lymphocyte development by sweeping changes in chromatin that permit or deny access to a single recombinase enzyme. We now show that switching/sucrose nonfermenting (SWI/SNF) chromatin remodeling complexes are recruited to the Igh locus by an enhancer-dependent process and that these complexes are essential for generating recombinase accessibility throughout the locus. Depletion of SWI/SNF in pro-B cells also inhibits antisense transcription through all clusters of Igh gene segments, a pioneering process that has been implicated in the initial opening of chromatin. We conclude that SWI/SNF complexes play multiple roles in Igh gene assembly, ranging from initial locus activation to the spreading and maintenance of chromatin accessibility over large V_H, D_H, and J_H domains.

Antigen receptor gene assembly occurs via a genetic reorganization, termed V(D)J recombination, which generates the enormous diversity of the Ig and Tcr variable regions required for adaptive immunity (1). V(D)J recombination is mediated at all Ag receptor loci by a single enzymatic complex consisting of the recombinases RAG1 and RAG2, which recognize recombination signal sequences flanking all variable (V), diversity (D), and joining (J) gene segments. Prior studies have shown that RAG complexes are targeted to specific regions within Ag receptor loci by changes in chromatin accessibility that either enhance or diminish the availability of recombination signal sequences to recombinase (1, 2). Chromatin accessibility at Ig and Tcr loci is regulated by cis-acting elements (promoters and enhancers) that communicate to drive the transcription of intervening gene segments in sense and antisense directions (3, 4).

These cis elements also sculpt regional patterns of histone modifications on chromatin associated with each cluster of gene segments. For example, deletion of Tcr enhancer elements alters histone modifications from active (e.g., histone 3 lysine 9 (H3K9)⁴ acetylation) to inaccessible patterns (e.g., H3K9 methylation) over neighboring chromatin (5, 6). Histone modification patterns constitute regional codes that are recognized by other nuclear factors, including nucleosome remodeling complexes (7). These multisubunit enzymes use ATP to render neighboring chromatin either accessible or inaccessible for transcription, recombination, and DNA repair (1). We recently showed that recruitment of switching/sucrose nonfermenting (SWI/SNF) remodeling complexes to a promoter in the Tcrb locus is required to activate transcription and recombination of neighboring DβJβ gene segments (8). However, the generality of SWI/SNF requirements in the mechanisms that regulate Ag receptor gene assembly remains to be established.

For this purpose, we explored SWI/SNF-dependent activation of Igh loci in pro-B cells. Prior studies have shown that the intronic enhancer Eμ directs sense and antisense transcription (αST) of the D_HJ_H cluster in pro-B cells (9, 10). Eμ-directed transcription may be crucial for opening chromatin throughout the D_HJ_H region (3), making it accessible to recombinase and recruiting RAG complexes via transcription-coupled histone modifications (2). Indeed, D_HJ_H rearrangement is significantly impaired in pro-B cells harboring a targeted deletion of Eμ (9, 11). In contrast, the large V_H cluster becomes accessible via Eμ-independent mechanisms and is opened in a regional manner, with D_H-proximal gene segments opening before distal V_H segments (12). Emerging evidence is consistent with a role for αST in converting inaccessible chromatin to a more accessible state over the 2-Mb V_H cluster (3). We now show that SWI/SNF complexes are essential for initiation of αSTs throughout the Igh locus. Consequently, pro-B cells lacking SWI/SNF fail to activate sense Igh transcription and assembly of V_H, D_H, and J_H gene segments. These findings underscore a central role for SWI/SNF complexes in the genetic-epigenetic crosstalk that directs locus-wide opening of Ag receptor genes for tissue- and stage-specific assembly during lymphocyte development.

Materials and Methods

Cell lines and mice

The RAG2-deficient cell lines P5424 (pro-T) and 63−12 (pro-B) as well as their derivative clones were described previously (4, 6). Mice lacking Eμ or both PDQ52 and Eμ were reported previously (9). All experiments were conducted according to guidelines of the Animal Care and Use Committee at Vanderbilt University (Nashville, TN). Data shown in the manuscript are all representative of at least two independent experiments (three for Fig. 2A and Fig. 4).

FIGURE 2.

SWI/SNF depletion in pro-B cell lines inhibits Igh germline transcription. A, RAG2-deficient pro-B cell lines (63−12, 2B9, and 2C10) were transfected transiently with an RNAi vector specific for BRG1 and BRM transcripts or with an empty control vector. 63−12 was transiently transfected with a Rag2 expression vector to generate subclones 2B9 and 2C10, which harbor stable Igh rearrangements (see Results and Discussion). Total RNA from transfected cells (GFP⁺) was analyzed by RT-PCR for germline Igh transcripts, specific upstream D_HJ_H transcripts (D_HLCμ), and mRNAs corresponding to SWI/SNF-dependent (c-myc) or -independent genes (Gapdh). Assay linearity was confirmed using sample titrations (far right). B, RT-PCR assays described in A were performed on pro-B cell cultures from RAG-deficient bone marrows supplemented with IL-7. Cells were infected with either control or RNAi retroviral vectors specific for BRG1/BRM. Quantified data were normalized for GAPDH and the relative inhibition (% Inh) resulting from SWI/SNF depletion is shown below each panel for two independent experiments.

FIGURE 4.

Igh rearrangement requires SWI/SNF. A, D_HJ_H coding joins in RAG2-deficient 63−12 cells cotransfected with a RAG2 expression vector and the episomal recombination substrate pJ_H200. Transfected cells were sorted on the basis of GFP expression and harvested for DNA. PCR assays were used to measure levels of D_HJ_H rearrangements (DFL16 and all DSP elements), recombinase activity (pJ_H200), and total DNA (Cλ). B, The 2C10 pro-B cell line harboring a DFL16.1J_H1 rearrangement was transfected and analyzed as in A except that PCR assays were used to detect V_H→D_HJ_H joins (V_H558 family).

Retrovirus production

BRG1/BRM-specific short hairpin RNA sequences (13) were cloned into the pGIPZ lentiviral vector (OpenBiosystems) modified as described (14). For viral production, the EcoPack2 packaging cell line (Clontech) was transfected with retroviral plasmid using Fugene6 (Roche) according to manufacturer's protocols. Viral supernatant was collected 48h later and filtered before use.

Cell line transfections

The pro-B clones 63−12, 2B9, and 2C10 were transfected with a vector expressing short hairpin RNA for BRG1/BRM (RNA interference (RNAi)) or a control GFP-only vector as described (8). Live GFP⁺ cells were sorted 72 h post-transfection for molecular analysis (BD FACSAria). For recombination assays, cells were cotransfected with PGK-Rag2 and pJ_H200 plasmids in addition to the RNAi vector (8).

Primary pro-B cell cultures and retroviral infections

Bone marrow cells from RAG2-deficient mice (6−8 wk) were cultured in RPMI plus IL-7 as described (14). Cultures were infected with a high-titer viral supernatant corresponding to RNAi or GFP control vectors, as described for transfections, containing Polybrene and IL-7 in 12-well plates. Fresh viral supernatant was substituted after 24 h of incubation and cultures maintained an additional 24 h. Cells were harvested for RNA 48 h postinfection.

PCR assays

Total RNA was treated with RNase-free DNase I and reverse transcribed using random hexanucleotides or strand-specific primers. PCR assays to measure μ0, Iμ, upstream D_HJ_H, V_H, and antisense transcripts were described previously (4, 9, 10). Total input cDNA in each reaction was measured with an RT-PCR assay specific for Gapdh transcripts (8). To measure Igh rearrangement, DNA extracts were prepared from cells and analyzed by PCR for D_HJ_H and V_HD_HJ_H joins as described (9).

Chromatin immunoprecipitation (ChIP) assay

IL-7-cultured pro-B cells were crosslinked with formaldehyde and chromatin was sheared using a Bioruptor sonicator (Diagenode) according to manufacturer's protocol. Chromatin from 10⁶ cells (average size 200−700 bp) was immunoprecipitated with Abs specific for BRG1 (07−478; Upstate Biotechnology) or with rabbit IgG (AB-105-C; R&D Systems). Immunocomplexes were washed and crosslinks were reversed as described (6). DNA from bound and input samples was analyzed by PCR using primers I plus K (Fig. 1A), which are specific for the Jh4-Eμ region of Igh (10). Dilutions of each sample confirmed assay linearity (input, 0.08 and 0.02% of total collected chromatin; bound, 5 and 1.25%).

FIGURE 1.

BRG1 recruitment requires Eμ. A, Schematic of mouse Igh showing relative positions of gene segment clusters and cis-elements. Initiation regions for major sense (μ0, Iμ) and antisense transcripts (αST) within the D_HJ_H cluster are shown. Lettered arrows indicate primers used in various PCR assays (μ0: G plus M; upstream D_HJ_H transcripts: D_HL plus M; Iμ: L plus M; DQ52 rearrangement: O plus N; V_H558 rearrangement: B plus N; αST-J_H:H (RT) and I plus K (PCR); αST-D_H: D (RT) and E plus F (PCR); αST-V_H:A (RT) and B plus C (PCR). B, BRG1-specific ChIP assays detect SWI/SNF association within the J_H cluster. Chromatin was prepared from IL-7 cultures of bone marrow cells derived from the indicated mouse genotypes and analyzed by PCR using primers I plus K. Signals were quantified using the Fuji Phosphor-Imager MultiGauge program and presented on the right as bound to input ratios for two independent experiments (separated by semicolons).

Results and Discussion

Eμ-dependent recruitment of SWI/SNF to the J_H cluster

Prior studies have shown that Eμ is essential for germline transcription and recombination throughout the D_HJ_H cluster (9, 11). In what may be a related finding, accessible Igh loci in transformed pro-B cells are hyperacetylated at H3K9 and associate with BRG1, an essential ATPase subunit of many SWI/SNF complexes (12, 15). These findings suggest that Eμ may recruit SWI/SNF remodeling complexes to generate chromatin accessibility at some or all D_HJ_H gene segments. As an initial test of this hypothesis, we analyzed whether Igh cis-acting elements are required for localized recruitment of BRG1 to the J_H cluster (Fig. 1A). Accordingly, ChIP assays were performed on primary pro-B cells from Rag-deficient mice harboring targeted deletions of the DQ52 germline promoter (P⁻E⁺) or a dual deletion of the promoter and Eμ (P⁻E⁻) (9). Consistent with reported effects on D_H→J_H recombination, tissue-specific recruitment of BRG1 to J_H segments was impaired at the P⁻E⁻ allele when compared with either wild-type (P⁺E⁺) or promoter-deficient Igh loci (Fig. 1B). These findings indicate that Eμ is required for efficient recruitment and/or continued occupancy of BRG1-containing complexes at the J_H cluster.

Classic germline Igh transcription requires SWI/SNF

To establish a causal relationship between Eμ-dependent recruitment of remodeling complexes and accessibility control at Igh, we used RNAi to simultaneously deplete BRG1 and BRM, the two known catalytic subunits of SWI/SNF. Depletion of SWI/SNF activity in a RAG2-deficient pro-B cell line (63−12) inhibits both μ0 and Iμ transcription (see Fig. 1A) through the germline DQ52-J_H-Cμ region (Fig. 2A, left column). In contrast to DQ52, other germline D_H elements are embedded in semirepressive chromatin but switch to a highly transcribed state upon rearrangement with a J_H segment, presumably because they are brought into the Eμ regulatory domain (4). To determine whether upstream D_H promoters also require SWI/SNF, we examined RAG-deficient cell lines harboring an upstream D_HJ_H rearrangement on one allele (2B9: DSP2.2-J_H1; 2C10: DFL16-J_H1) and a V_HD_HJ_H rearrangement on the second Igh allele (2B9: V_H7183DJ_H2; 2C10: V_H7183DJ_H1) (Ref. 4) and our unpublished data). As shown in Fig. 2A (top two panels), transcription of the J_HCμ region on both Igh alleles is inhibited upon depletion of BRG/BRM1. To specifically detect transcripts originating from the rearranged D_H promoters, we used a PCR primer (D_HL) that recognizes a region within all upstream D_H segments but not in the J_H-proximal DQ52 gene segment (Fig. 1A and Ref. 9). Consistent with this primer specificity, the new PCR assay fails to detect transcripts originating from the germline DQ52J_H region in the 63−12 cell line (Fig. 2A, third panel from top). However, BRG1/BRM depletion again led to a significant decrease in the level of rearranged D_HJ_H transcripts in 2B9 and 2C10 cells. These data suggest that Eμ-directed transcription from any of the D_H promoters requires continual SWI/SNF activity.

In contrast to the D_HJ_H cluster, transcription of unrearranged V_H gene segments is independent of Eμ function (9, 11). As shown in Fig. 2A, depletion of SWI/SNF inhibits germline V_H transcription at both the 5′ (V_H558) and 3′ ends (V_H7183) of this large cluster. Thus, SWI/SNF controls Eμ-dependent and -independent expression of germline gene segments positioned throughout the Igh locus.

To confirm and extend these findings from transformed cells, we knocked down BRG1/BRM in primary pro-B cell cultures using retroviral infection of the requisite RNAi vector (14). Bone marrow-derived pro-B cells from RAG2-deficient animals were expanded in medium containing IL-7, and cycling cells were infected with either RNAi or control retroviral vectors (>85% infection). As expected, RNAi-mediated depletion of Brg1 and Brm transcripts led to a dramatic reduction in the level of SWI/SNF-dependent c-myc mRNAs (Fig. 2B). Consistent with results from transformed cell lines, inhibition of SWI/SNF activity significantly impaired locus-wide expression of Igh as measured by steady-state levels of V_H, D_H, and J_H germline transcripts. We conclude that SWI/SNF remodeling complexes are essential for establishing the accessible chromatin at Igh promoters that drive classical germline V_H, D_H, and J_H transcription.

SWI/SNF is required for antisense Igh transcription

In addition to conventional germline transcripts that traverse Igh gene segments in a 5′ to 3′ (sense) orientation, recent studies showed that primary activation of Igh in pro-B cells is accompanied by locus-wide αST (3, 4). This type of intergenic αST has been implicated as a pioneering process in the opening of chromatin at several genetic loci that serves as a prelude to gene expression (16). In the J_H region, αST and D_H→J_H recombination are both Eμ dependent, with αST initiating within or near this enhancer (10). The cis-elements and initiation sites for αST of V_H segments remain unknown. These observations are consistent with the hypothesis that αST plays an important role in establishing chromatin accessibility at the D_HJ_H cluster and perhaps throughout the entire Igh locus.

To determine whether αST within distinct Igh domains requires SWI/SNF, we primed total RNA from pro-B cells (SWI/SNF depleted or control) with strand-specific oligonucleotides to generate cDNAs corresponding to αSTs (10). Consistent with prior studies, PCR analysis of the resultant cDNAs showed that αST of Igh is specific to pro-B cells (Fig. 3A and data not shown). Importantly, αST throughout the V_H, D_H, and J_H clusters is significantly impaired in primary pro-B cell cultures upon depletion of BRG1/BRM (Fig. 3B). These data indicate that pioneering αST throughout the entire Igh locus is dependent upon the nucleosome remodeling activity of SWI/SNF complexes.

FIGURE 3.

Pioneering αST is SWI/SNF-dependent. A, Pro-B cells (63−12) transfected with RNAi or control vectors were analyzed using a strand-specific RT-PCR assay for αST (AS) of the J_H region. RNA from mouse embryonic stem (ES) cells or pro-B cell mRNA untreated with reverse transcriptase (no RT) are shown as negative controls. B, Strand-specific RT-PCR assays were performed as in A on RNA from primary pro-B cell cultures using primers for αST (AS) of the J_H, V_H, and D_H regions indicated. An assay for Gapdh transcripts is shown as a control for RNA levels.

Recombinational accessibility of Igh requires SWI/SNF activity

The Igh locus is assembled in a stepwise fashion, with the opening of D_HJ_H chromatin and its recombination preceding the final V_H to D_HJ_H joining events. Although the transcription of gene segments can often reflect their accessibility to V(D)J recombinase, these two properties do not always correlate (1). To directly test whether SWI/SNF is required for the recombination of D_HJ_H clusters, we transiently transfected RAG2-null pro-B cells with a RAG2 expression construct, creating a readout for recombinase accessibility at Igh. As shown in Fig. 4A, germline Igh alleles were readily converted to D_HJ_H rearrangements in pro-B cells that received both RAG2 and control vectors. In contrast, RAG2 expression failed to induce D_HJ_H recombination in cells that received a RNAi vector targeting BRG1/BRM (the observed inhibition of rearrangement is >9-fold). However, this was not due to an absence of recombinase activity, because an episomal substrate (pJ_H200) was rearranged at similar levels in all samples (8).

For probing recombinase accessibility within the V_H cluster, we performed similar experiments with RAG2-deficient pro-B cells that harbored a preformed D_HJ_H junction on one allele (2C10; DFL16.1J_H1) (4). SWI/SNF depletion also inhibited V_H to D_HJ_H recombination in this pro-B cell clone (Fig. 4B; consistent >3-fold inhibition in three independent experiments). Similar results were obtained with other clones harboring distinct D_HJ_H joins (data not shown).

Together, our results demonstrate that SWI/SNF remodeling complexes have a global function in Igh locus activation rather than a localized effect restricted to the Eμ-proximal DQ52J_H cluster. SWI/SNF remodeling activity is required for αST as well as conventional germline transcription, the latter of which initiates from well-defined Igh promoters. Our studies support the following stepwise model for accessibility control at Igh. First, Eμ recruits SWI/SNF to activate αST through inaccessible chromatin associated with upstream D_HJ_H gene segments, resulting in a partial opening of this chromatin domain (3, 10). Second, Eμ interacts with promoters in partially accessible D_HJ_H chromatin, activating conventional (sense) germline transcription in a SWI/SNF-dependent manner. Both sense and antisense transcription may be required to generate full chromatin accessibility at D_HJ_H gene segments or to recruit recombinase via transcription-coupled histone modifications (2, 9, 11). Ultimately, SWI/SNF action at Igh must culminate in a recombinase-accessible configuration of chromatin throughout the locus that triggers complete VDJ exon assembly and B cell development. Important future studies will focus on the mechanisms of SWI/SNF recruitment to Eμ and to novel cis-elements that regulate the accessibility of the large V_H cluster.