Transcriptional activation of germline (GL) promoters occurs through binding of NF-κB to three evolutionarily conserved sites within a CD40 response region in the human and mouse GL Iγ and Iε promoters. Here we identify and characterize a novel NF-κB binding site (κB6) within the human GL Iγ1 promoter that plays an essential role in basal- and CD40-induced transcription. This site is adjacent to identified CREB/activating transcription factor (ATF) sites, present in the Iγ1 but not the Iγ3 promoter, which are important for the amplification of transcription. Our data suggest a cohesive protein complex regulating Iγ1 promoter activity because disruption of any individual NF-κB or CREB/ATF site markedly lowers the overall inducible activity of the promoter. In addition, alteration of helical phasing within the promoter indicates spatial orientation of CREB/ATF and NF-κB, proteins contributes directly to promoter activity. We found that CREB and p50 transactivators, as well as coactivator p300, interact in vivo with the Iγ1 promoter in the presence and absence of CD40 signaling in Ramos and primary B cells. However, the level of CREB and p300 binding differs as a consequence of activation in primary B cells. Furthermore, overexpression of p300, and not a mutant lacking acetyltransferase activity, significantly increases Iγ1 construct-specific transcription. Together these data support a model whereby CREB and multiple NF-κB complexes bind to the Iγ1 promoter and recruit p300. CD40 signals induce p300-dependent changes that result in optimal Iγ1 promoter activity.
Alteration of chromatin structure is strongly correlated with gene expression in many facets of cellular function. Several classes of proteins participate in this overall process through direct binding to regulatory sequences, protein-protein interactions, and/or structural modifications of chromatin (reviewed in Refs.1, 2). An important class of chromatin modifying proteins that function as transcriptional coactivators includes the CREB binding protein (CBP)3 and its paralogue p300. These proteins participate in multiple signal-dependent transcription events by forming molecular bridges, providing physical scaffolds for stabilizing multicomponent regulatory complexes, modifying chromatin assembly via intrinsic histone acetyltransferase (HAT) activity, and modulating transactivator activity (reviewed in Ref.3). Furthermore, CBP/p300 proteins possess factor acetyltransferase activity that mediates transcriptional activation through direct acetylation of trans-acting inducible factors (reviewed in Refs.4, 5). CBP/p300 interactions with proteins such as CREB/activating transcription factor (ATF) (6, 7), NF-κB (8, 9), and c-Jun (10, 11, 12) provide factor modification and/or important physical bridges between the transcriptional regulatory complex and basal transcription factors TFIIB (7, 13), TATA-binding protein (13), and RNA polymerase II (14, 15, 16, 17). Finally, CBP/p300 proteins recruit additional coactivators with acetyltransferase activity, such as the p300/CBP-associating factor, into a transcription complex, and these proteins function by allowing direct modification of transcription factors (reviewed in Refs.3, 18).
Organization of enhancesomal complexes facilitating transcriptional activation and chromatin remodeling within regulatory regions has previously been identified in specific aspects of immune system control, including regulation of MHC class II and IFN-γ genes (19, 20, 21, 22, 23). The accessibility model of class switch recombination (CSR), which draws a direct link between transcription and recombination, supports the idea of such a regulatory protein complex mediating germline (GL) transcription from the intragenic or I region promoters. In addition to GL transcription, CSR requires association of the activation-induced cytidine deaminase (AID) protein with single-stranded S region DNA, which is facilitated via chromatin remodeling (reviewed in Ref.24). Recent work demonstrates that AID is bound to RNA polymerase II on actively transcribing loci, and chromatin remodeling in the absence of GL transcription is not sufficient for CSR to occur (25). These exciting findings suggest that GL transcription may facilitate AID recruitment through a promoter bound complex that provides accessibility as well as a necessary protein scaffold for CSR.
We have previously used the EBV-negative Burkitts lymphoma line, Ramos 2G6, as a model system to study early events associated with switch recombination in human B cells (26). This IL-4/CD40L-responsive cell line expresses Iγ transcripts and undergoes limited switch recombination to Cγ1 (26, 27). Although IL-4 and CD40L signaling induces transcription from all four Iγ loci, the response is strongly biased in favor of Iγ1. The preferential expression of Iγ1 transcripts and the commitment to CHγ1 recombination support experiments showing that a linear relationship exists between the efficiency of GL transcription and CSR and favors a model whereby a threshold of transcription is required for recombination (28). The mechanisms regulating the biased expression of Iγ transcripts are not entirely clear because there is an exceptionally high degree of homology between the four promoters. Importantly, a region of conservation identified in several human and murine Iγ promoters corresponds to a CD40 response region (CD40RR) within an evolutionarily conserved region comprised of three NF-κB consensus sites (29, 30). Previous work from our laboratory identified a region outside of the CD40RR that contributed to the difference in expression of Iγ1 and Iγ3 GL transcripts. Specifically, we identified a 36-bp region in the Iγ1 promoter that contained overlapping CREB and ATF-1/ATF-2 binding sites as well as a consensus NF-κB site (κB6) downstream and adjacent to the CREB site. High basal and IL-4/CD40L-induced responses corresponded to the binding of CREB/ATF proteins to motifs found in the Iγ1 but not the Iγ3 36-bp element (31).
In this work we extend our earlier findings by defining the function of the κB6 site in promoter activation and its relationship to the adjacent CREB/ATF binding sites under both basal and IL-4/CD40-induced conditions. We report that κB6 binding is required for optimal Iγ1 promoter activity and that the orientation of the CREB and κB6 sites relative to one another is critical for basal and induced expression. These findings suggest that CREB and NF-κB are communicating with each other on the Iγ1 promoter via direct or indirect interactions. We also present data demonstrating that the coactivator p300 is an essential component of NF-κB and CREB-specific complexes and that p300 HAT activity is necessary for optimal promoter activity. Finally, using chromatin immunoprecipitation (ChIP) we demonstrate an increase in p300 associated with the Iγ1 promoter in CD40-activated B cells.
Materials and Methods
Anti-p50 (sc-7178), anti-CREB (sc-58), and control Abs were obtained from Santa Cruz Biotechnology. Anti-p300 and anti-CREB Abs were purchased from BD Pharmingen and Upstate. The pCL.p300 and pCL.p300ΔHAT plasmids were generously provided by Dr. J. Boyes (Institute of Cancer Research, London). The E1A 12SWT (32) and E1A 12SRG2 (33) were gifts from Drs. E. White (Center for Advanced Biotechnology and Medicine, Rutgers University, Piscataway, NJ) and E. Moran (Fels Institute for Cancer Research and Molecular Biology, Temple School of Medicine, Philadelphia, PA), respectively.
Ramos 2G6.4CN3F10 (Ramos 2G6) cell line, an IgM+, non-EBV-transformed Burkitt’s lymphoma line was previously isolated by Siegel and Mostowski (34). Ramos 2G6 cells were grown in RPMI 1640 supplemented with 10% heat-inactivated FCS, 50 U/ml penicillin, 50 μg/ml streptomycin, and 1 mM l-glutamine at 37°C and 5% CO2. 293 cells are derived from adenovirus 5-transformed primary human embryonal kidney cells (American Type Culture Collection). The 293/CD40L line was constructed by the stable transfection of pCT-BAM into 293 cells as described previously (35). 293 cells were cultured in DMEM/F12 supplemented with 10% FCS, 50 U/ml penicillin, 50 μg/ml streptomycin, and 1 mM l-glutamine at 37°C and 5% CO2.
Variable length Iγ1 promoter deletions were created using the Advantage-HF 2 PCR Kit (BD Clontech). Primers were designed to flank both ends of the insert, while creating a HindIII site at the 5′ end to facilitate ligation into the pGL2E vector (Promega). The primer sequences were as follows: 21, 5′cta taa gct tag gta ttg aga ggc tga gat ca 3′; 138, 5′cta taa gct tac ctt ctc cca cga gta 3′; 228, 5′cta taa gct ttc gcc ctc tga ccc aga aac ca 3′; 274, 5′cta taa gct taa gta agt ggt gcc gcc ggt tt 3′; and GL-2, 5′ctt tat gtt ttt ggc gtc ttc c 3′. PCR parameters were 94°C for 1 min for one cycle, followed by 94°C for 15 s, 54°C for 45 s, and 68°C for 1 min for 30 cycles, and concluded with 68°C for 3 min. All of the PCR products were subsequently digested with HindIII, and purified using Qiagen QiaQuick PCR Purification Kit. Ligations were performed with each PCR product and the pGL2E vector, previously linearized with HindIII and treated with CIP for 1 h at 37°C.
Site-specific Iγ mutations were made using the Quikchange Site-Directed Mutagenesis kit (Stratagene) to disrupt individual NF-κB binding sites or introduce 5- and 10-bp additions within the minimal promoter region. The primer sequences were as follows, with induced mutations in bold: mutNF-κB3, 5′gca gcc tcg gct gtg tct gga ctg agt tt t acc ctg tga ccc 3′; mutNF-κB4, 5′ccc tcg ccc tct gac cca gaa gcg ctg aga aga aaa 3′; mutNF- κB5, 5′cca cca gaa gaa aag gct ggc tca gga agt aag tgg tgc cgc c 3′; mutNF-κB6, 5′gct gac ggc agg ggc ggg gga cgt gac atg tac ctc gcc ag 3′; Iγ1 + 5 bp, 5′cgg ctg acg gca ggg gcg cat ggg ggc tgc cca cat g 3′; and Iγ1 + 10 bp, 5′ ctg acg gca ggg gcg cat gct tca ggg gct gcc cac atg 3′. PCR parameters were 95°C for 30 s for one cycle and 95°C for 30 s, 65°C for 1 min, and 68°C for 12 min for 20 cycles. PCR products were incubated with DpnI to digest the parental DNA template, and used to transform JM109 competent bacteria.
Transient transfections and coculture conditions
Transfection and stimulation of Ramos 2G6 cells (5 × 106 per condition) were conducted exactly as described previously (31). After 48-h incubation at 37°C, cells were lysed and assayed for luciferase activity using the Dual-Luciferase assay kit (Promega). Firefly and Renilla luciferase activity was measured using a luminometer in relative light units (RLUs), and fold induction over basal expression was determined.
A total of 4 × 106 Ramos 26G B cells was plated in 4 ml/well in a 6-well plate. 293 or 293/CD40L membranes were prepared as described previously (31), and 1 × 107 cell equivalents were added with or without IL-4 (200 U/ml) to appropriate wells. Ramos cells were cultured for 24 h and harvested for nuclear extracts using a modification of Dignam’s method (36). Briefly, cells were harvested, washed in 1× PBS, and resuspended in ice-cold buffer A (10 mM HEPES (pH 7.9) 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 0.5 mM PMSF, 2 μM aprotinin, and 2 μM pepstatin A). After incubation on ice for 10 min, Nonidet P-40 was added to a final concentration of 0.6%, and nuclei were isolated by centrifugation at 14,000 rpm for 30 s. Nuclei were resuspended in ice-cold buffer C (20 mM HEPES (pH 7.9) 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 0.5 mM PMSF, 2 μM aprotinin, and 2 μM pepstatin A). Samples were kept on ice for 30 min, with brief vortexing every 5 min, and centrifuged at 14,000 rpm for 10 min. Protein concentration was determined using the Bradford Assay (Bio-Rad).
The following oligos, containing wild-type and mutated NF-κB binding sites, were used in EMSA reactions (base changes are denoted in bold): κB3, 5′gca gcc tcg gct gtg tgt gga ctc ccc ctc gcc ctc tga ccc 3′; mκB3, 5′gca gcc tcg gct gtg tgt gga ctg agt gtc gcc ctc tga ccc 3′; κB6, 5′gct gac ggc agg ggc ggg gct gcc cac atg tac ctc gcc ag 3′; mκB6, 5′gct gac ggc agg ggc ggg gga cgt gac atg tac ctc gcc ag 3′; CREB/κB6, 5′aaa tat cgg ctg acg gca ggg gcg ggg ctg ccc aca 3′; CREB/mκB6, 5′aaa tat cgg ctg acg gca ggg gcg ggg gac gtg aca 3′; and mCREB/κB6, 5′aaa tat cgg ctg tgg gca ggg gcg ggg ctg ccc aca 3′. Double-stranded NF-κB 5′agt tga ggg gac ttt ccc agg c 3′ competitor consensus DNA fragment was purchased from Promega. Complementary single-stranded oligonucleotides were annealed, gel purified, and end labeled with [γ-32P]ATP. Binding reactions were prepared using 4 μg of extract and 1 μg of poly(dI-dC) in binding buffer (10 mM Tris-Cl (pH 7.5), 50 mM NaCl, 1 mM DTT, 1 mM EDTA, 5% glycerol) in the presence or absence of competitor oligos for 10 min at 25°C. A total of 3 × 104 cpm was added to reactions and incubated for 20 min at 25°C. Samples were loaded on a 6% acrylamide gel and visualized by autoradiography.
Isolation of primary B cells
Buffy coats were prepared from peripheral blood samples from healthy donors by the New Brunswick Affiliated Hospitals Blood Center at Robert Wood Johnson University Medical Center. Mononuclear cells (PBMCs) were separated by Ficoll-Hypaque gradient centrifugation. B cells were removed from total PBMCs by biomagnetic separation using anti-CD19-conjugated superparamagnetic beads (Dynal Biotech). IgG+ cells were removed by plating onto anti-IgG-coated plates for 30 min at 25°C. To confirm isolation of specific B cell subsets, 1 × 105 cells were washed in 3% FCS/0.1% NaN3/1× PBS followed by incubation with 5 μg of heat-aggregated IgG to inhibit nonspecific binding. Cells were incubated for 45 min at 4°C with saturating amounts of FITC-conjugated mAbs against human CD20 (Ancell) or IgG (Southern Biotechnology Associates). Cells were washed and fixed with 1% paraformaldehyde in 1× PBS. Cells were analyzed using FACScan (BD Biosciences).
ChIP was performed essentially as reported previously with minor modifications (37, 38). Briefly, 5 × 107 Ramos 2G6 cells or peripheral IgM+IgG− B cells pooled from multiple donors were cultured in the presence or absence of 0.5 μg/ml sCD40L for 2 h (PeproTech). This reagent has been shown by size exclusion chromatography to trimerize in solution (PeproTech). The concentration of sCD40L was determined based on the induction of NF-κB binding activity in EMSA experiments. Cells were cross-linked by the addition of one-tenth volume of 11% formaldehyde in 0.1 M NaCl, 1 mM EDTA, 0.5 mM EGTA, and 50 mM HEPES (pH 8.0) in growth medium for 7 min at 37°C, before addition of glycine to a final concentration of 0.125 M. After washing two times with ice-cold PBS, cells were resuspended in cell lysis buffer (50 mM Tris-Cl (pH 8.1), 10 mM EDTA, 1% SDS, 1× protease inhibitor cocktail, and 1 mM PMSF) and incubated on ice 10 min. Chromatin was sonicated with glass beads, 10 rounds for 20 s using a 250 Branson Sonifier (30% output) alternating with 30-s incubations on ethanol/ice. Chromatin was centrifuged 10 min at 14,000 rpm, diluted 10-fold in dilution buffer (167 mM NaCl, 16.7 mM Tris-Cl (pH 8.1), 0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 1 mM PMSF, and 1× PIC), and filtered through a 0.45-μm filter to remove aggregates. Extract was divided into 5 × 106 cell equivalents/sample. Chromatin samples were precleared for 30 min at 4°C by adding 40 μl of Protein A/ssDNA-agarose beads (Upstate), followed by incubation with 2–4 μg of Ab at 4°C overnight. Immune complexes were recovered at 4°C for 1 h using 60 μl of Protein A/ssDNA-agarose beads. Chromatin input volume (10%) was reserved from the sample incubated without Ab. Samples performed in duplicate were combined, and complexes were washed five times with IP1 (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris (pH 8.0), and 150 mM NaCl), once with IP2 (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris (pH 8.0), and 500 mM NaCl), and twice with TE. Immunoprecipitation reactions and input chromatin were digested with 200 μg/ml RNase A for 1 h and 200 μg/ml proteinase K in TE with 0.5% SDS for 2 h at 55°C. Cross-links were reversed overnight at 65°C. Samples were extracted once with phenol/chloroform and once with chloroform/isoamyl alcohol, ethanol precipitated, and resuspended in TE.
PCR was performed using either 0.2 ng of purified HindIII-BamHI Iγ subclass promoter sequence, 1 μl of 1/200 dilution of input chromatin, 3 μl of undiluted or 10-fold serial diluted immunoprecipitant in a 50-μl reaction containing 10 mM Tris Cl (pH 8.1), 1.5 mM MgCl2, 50 mM KCl, 0.1% Triton X-100, 5% DMSO, 0.2 mM dNTPs, 2.5 U Taq (Promega), and 100 ng of 5′ and 3′ primers: Iγ nonspecific, 5′gcc ctc tga ccc aga aac c 3′ and 5′cct tcc tgt tct ggc gag gt 3′; Iγ1 specific, 5′ tcc atg tag tgc cgg aca cga ccc cat 3′; and IgHG1, 5′ctc cac caa ggg ccc atc ggt 3′ and 5′caa atc ttg tga caa aac tca cac at 3′. Amplification of all reactions was for 35 cycles, with a hot-start at 3 min at 92°C, 30 s at 92°C, 45 s at 56°C, 45 s at 72°C, and 3 min at 72°C. Final PCR amplifications were subjected to gel electrophoresis, and bands were quantified using Kodak imaging software.
Two distinct regions are important for regulating Iγ1 transcription
Before assessing the specific function of the putative NF-κB-6 site in both basal and inducible expression, we generated mutants to define the CD40 and IL-4 regulatory regions in the Iγ1 promoter (−310/+202). Six putative NF-κB sites have been previously identified in the Iγ3 region with sites 3, 4, and 5 lying within the evolutionarily conserved region (CD40RR, −95 to −30) shown to be important for CD40-inducible expression in both mouse and humans (Fig. 1,A) (29). Although the individual NF-κB elements in the human Iγ1 promoter have never been functionally analyzed, we made the assumption based on the homology of the CD40RR that κB sites 3–5 would be critical for inducible Iγ1 expression. Four promoter deletion mutants were generated from Iγ1-512 sequence that sequentially removed one NF-κB site in each promoter fragment. The corresponding constructs were termed as follows: mut21 (−289/+202) (minus κB site 1), mut138 (−172/+202) (minus sites 1 and 2), mut228 (−82/+202, minus sites 1–3), and mut274 (−36/+202, minus sites 1–5) (Fig. 1,A). These regions were subcloned upstream of the Firefly luciferase gene and introduced by transient transfection into Ramos 2G6 B cells with the Renilla luciferase-expressing plasmid, pRLSV40, as a control for transfection efficiency. Transfected cells were cultured with either 293 or 293-expressing CD40L (293/CD40L) cells in the presence or absence of IL-4, and promoter strength was measured 48 h later as a function of relative luciferase activity. In agreement with other studies, deletion of the two upstream κB sites (κB1 and κB2) had minimal effect on either the basal or inducible promoter activities. In contrast, the mut228 and mut274 constructs lacking the κB3 and the κB3–5 sites, respectively, corresponding directly to sites 1, 2, and 3 in the mouse Iγ1 locus (29), demonstrated a >15-fold decrease in both basal and inducible transcriptional responses (Fig. 1, B and C). These findings indicated that the κB3 site was critical for optimal activity; however, it was unclear whether κB4–6 sites also had distinct roles in promoter function.
To analyze the specific activity of the κB3-κB6 sites within the context of the complete promoter, each site was selectively mutated by PCR mutagenesis and assessed for luciferase expression after transient transfection. Mutation of the κB4 site resulted only in a marginal decrease in basal transcription, and disruption of the κB5 site reproducibly increased the basal expression ∼3-fold. In contrast, mutations in the κB3 and κB6 sites severely reduced the basal expression by >15-fold (Fig. 2,A). The effect of the κB3 mutation was consistent with previous reports showing the importance of the CD40RR in regulating the murine Iγ1and Iε and the human Iγ3 promoters (29, 39). However, the finding that mutations in the κB6 site had a profound effect on transcription was both novel and unexpected. This putative NF-κB site lies within a previously defined 36-bp element, shown to be functionally divergent between the Iγ1 and Iγ3 promoters, and is directly adjacent to multiple CREB/ATF binding sites implicated in optimal Iγ1 promoter activity (31). Analysis of the different binding sites with respect to CD40L and IL-4 stimulation was conducted by transfecting the mutant κB promoter constructs into Ramos B cells and culturing them with and without 293/CD40L cells in the presence and absence of IL-4 (Fig. 2, B and C). Surprisingly, disruption of the κB3 and κB6 sites failed to significantly change the overall inducibility of the promoter (Fig. 2,B). In direct contrast, mutations in the κB4 site markedly reduced both the IL-4- and the CD40-inducible response, and changes in the κB5 site severely affected the promoter’s ability to be induced over a basal level of activity (Fig. 2,B). Collectively, these data suggest that mutations within the CD40RR (κB3–κB5) and the κB6 site have profound effects on Iγ1 expression such that mutating any one site prevents CD40-inducible transcription to reach a threshold level that is consistent with that reached with the nonmutated promoter (Fig. 2 C). The inability to reach the targeted transcription threshold level is regulated either at the level of basal transcription (sites 3 and 6) or at the level of loss of inducible expression (site 5) or both (site 4). The fact that mutations in any one site directly affected transcription indicates that the individual sites are not functionally redundant and are in fact collectively required for maximal basal and inducible activities.
NF-κB/Rel subunits and CREB bind to the κB6 site
To further analyze the κB6 site with respect to the binding of specific NF-κB-dependent complexes, experiments were conducted using a probe that spans both the CREB/ATF1/2 sites and the adjacent κB6 site (CREB/κB6). As shown in Fig. 3 A, EMSA using extracts from CD40L− or CD40L+ IL-4-stimulated Ramos cells revealed four distinct complexes, with the third lowest complex being the strongest (termed B6). This complex was considerably increased relative to unstimulated extract (lane 2), suggesting that its activity was CD40-dependent and most likely contained NF-κB/Rel-family members (lanes 3 and 4). This finding was confirmed by competing the B6 complex with cold NF-κB consensus oligo (lane 6) but not with an oligo containing a mutant NF-κB site (lane 7). Competition of the B6 complex with cold NF-κB consensus oligo enhanced the intensity of the lower most complex (lane 6). The upper and lower complexes, termed CREB-L and CREB-U were specifically competed by unlabeled CREB consensus oligo (lane 8) but not with the mutated CREB/κB6 oligo (lane 9). Furthermore, addition of unlabeled consensus NF-κB and CREB oligos together successfully competed the CREB-U, CREB-L, and B6 complexes (lane 10). Together these results indicate that binding of NF-κB and CREB within the 36-bp element are independent events, and occupancy of the adjacent site is not required for binding of either CREB or NF-κB proteins.
To compare the pattern of binding at the κB6 site with the CD40RR, a similar analysis was conducted using the κB3 oligo (Fig. 3 B). With this probe we observed a single NF-κB-dependent complex that was present in unstimulated extracts at a low, but observable level (complex B3; lane 2). NF-κB binding slightly increased with IL-4 (lane 3) and significantly increased after stimulation with CD40L (lanes 4 and 5). Specificity of the inducible complex was shown by its ability to be fully competed with unlabeled NF-κB consensus oligo (lane 6) but not the κB3 probe containing a mutation in the NF-κB binding site (lane 7). Analysis of the NF-κB subunit composition of B3 and B6 revealed that p50, p65, and cRel were the primary components of both complexes upon CD40 stimulation (data not shown).
Iγ1 transcription is mediated in part through p300
The dependence on multiple sites for Iγ1 promoter function and the fact that disruption of any one site resulted in a marked decrease in transcription suggested that multiple signals are being integrated through interactions with coactivator proteins. The HAT-containing coactivators CBP and p300 are known to bridge signals between the basal transcriptional machinery and activation-induced trans-acting factors including CREB and NF-κB (reviewed in Ref.40). To explore a possible role for p300/CBP in GL Iγ1 transcription, initial binding experiments were conducted with the CREB/κB6 probe and Abs against p300 and CBP. As shown in Fig. 4 A anti-p300 Abs interfered with both CREB-U and B6 formation, indicating specific interactions among p300, CREB, and NF-κB at the CREB/κB6 site (lane 4). In contrast, anti-CBP Abs produced only a marginal loss of the B6 and CREB-U complexes (lane 6). To test whether p300 was also a component of the κB3-inducible complex, EMSA was conducted with the κB3 probe and anti-p300 or control Abs. Surprisingly, addition of anti-p300 Abs interfered with B3 complex formation but to a lesser extent than with B6 formation (compare lanes 10 and 4). We also observed a marked disruption of the κB4- and κB5-specific complexes with anti-p300 Abs (data not shown). Collectively these findings suggest that p300 is binding to CREB- and NF-κB-specific complexes that are bound to sites within the CD40RR (κB3 and κB) as well as in the 36-bp element (κB6 and the CREB sites).
If transcriptional activation requires either a direct (protein-protein) or indirect (p300-mediated) interaction between CREB and NF-κB, then introducing a spacer sequence that alters the helical phasing of the two consensus sites should result in a loss of promoter activity. To test this hypothesis a 5-bp or 10-bp spacer was inserted between the CREB and κB6 sites within the context of the Iγ1-512 promoter construct, and luciferase activity was analyzed upon transient transfection into Ramos B cells. Notably, insertion of 5 bp resulted in a dramatic decrease in expression relative to the unmodified promoter. In contrast, insertion of a 10-bp linker fully restored both basal and inducible expression to wild-type levels (Fig. 4 B). These results indicate that the binding of NF-κB and CREB proteins within the 36-bp element is orientation-dependent, and this restriction relates either to a direct requirement for a physical interaction between the CREB and NF-κB and/or to the positioning of these proteins relative to p300.
Expression of p300 up-regulates Iγ1 promoter activity
To demonstrate that p300 can modulate Iγ1 expression in vivo, the Iγ1-512 construct and pRLSV40 control plasmid were coexpressed with p300, p300ΔHAT (containing a mutation in the acetyltransferase transactivation domain), or an empty vector control. Overexpression of p300 reproducibly up-regulated both basal and CD40-regulated expression as measured by luciferase expression (Fig. 5 A). Notably, there was no significant difference in the basal or induced expression upon transfection with either the empty vector or p300ΔHAT vector. Because expression is being assayed from transiently transfected extrachromosomal constructs, the effect of chromatin modifications on transcriptional regulation should be negligible. Therefore, the finding that only the wild-type p300 and not the p300ΔHAT up-regulates expression, suggests that p300-specific acetyltransferase activity is required for Iγ1 transcription.
To further confirm that p300 directly regulates Iγ1 transcription, transfections were conducted with the Iγ1-512 plasmid and vectors expressing either the wild-type E1A 12S protein, a mutant E1A 12S protein lacking the p300 binding site (12SRG2), or a vector control. Adenovirus E1A 12S protein is known to be a dominant-negative regulator of p300 activity by virtue of its ability to bind to the transcriptional activation domain and inhibit coactivator activity (41, 42, 43, 44, 45). As shown in Fig. 5 B, transfection of the wild-type E1A 12S protein severely decreased both the basal and inducible transcription. In contrast, the 12SRG2 mutant had no effect on transcription relative to the empty vector control. These findings strongly support a role for p300 in Iγ1 transcription. In particular, the positive and negative effects of p300 and E1A, respectively, on the activity of the transfected promoter indicate that p300 is functioning either as a scaffold to integrate the different signals and/or as an acetylase of NF-κB/Rel and CREB proteins.
CREB, p300, and NF-κB bind to the Iγ1 promoter in vivo
To verify our transfection and binding results and show that NF-κB, CREB, and p300 are binding to the endogenous Iγ1 promoter region under conditions of transcriptional activation, ChIP assays were performed using unstimulated Ramos cells, and Ramos cells were stimulated for 2 h with sCD40L. We chose to assay p50 binding based on our in vitro data, showing that this NF-κB subunit was present in all NF-κB-specific complexes binding to the Iγ1 promoter (data not shown). The amplified target was ∼300 bp of the Iγ1 promoter that included the κB4–6 sites as well as the CREB/ATF binding site (Fig. 6,A). We designed a 3′ primer that was specific for the Iγ1 promoter by introducing four additional base pair changes into a sequence of 87% homology among the four subclasses. Specificity was confirmed by using recombinant DNA containing the Iγ1–Iγ4 promoters in separate PCR under conditions that amplified the promoter region of the Iγ1 subclass only (Fig. 6 B).
In unstimulated Ramos B cells we observed that p50, CREB, and p300 specifically associated with the Iγ1 promoter, as determined by increased levels of signal over the no Ab control and the absence of signal from a region within the Cγ1 gene (IgHG1) (Fig. 6 C). Surprisingly, levels of p50, CREB, and p300 appeared relatively unchanged upon 2-h CD40 activation, although from parallel binding studies it was clear that NF-κB was induced during this time period (data not shown). Overall, these results indicate that p50, CREB, and p300 are bound to the Iγ1 promoter before CD40-mediated transcriptional activation of the locus.
To demonstrate the biological relevance of complex formation to Iγ1 transcription, CD19+IgM+IgG− B cells were isolated from PBMCs and treated with or without sCD40L 2 h before isolation of chromatin. The percentage of IgM+ B cells in the isolated population was determined by FACS to be ∼96% (data not shown). As shown in Fig. 7, we observed p50, CREB, and p300 bound to the Iγ1 promoter in chromatin from resting and CD40-stimulated B cells as measured by an increase over controls. In particular, there were similar levels of p50 bound under both conditions. However, CREB binding was much more pronounced in unstimulated primary B cells compared with their stimulated counterparts. This observed difference may either reflect an influx of regulatory proteins binding to the CREB/NF-κB complex that interfere with Ab recognition or indicate that CREB is being displaced by other factors at the promoter. The first possibility is supported by the observation that p300 binding is enhanced with CD40 signaling, and its binding to CREB may obscure specific epitopes. Overall, these results indicate that p300 binding at the Iγ1 promoter increases in response to CD40 signaling and supports a role for this factor in regulating GL transcription in vivo.
Levels of GL transcription are important predictors of CSR to a particular CH region, and thus proteins that significantly enhance the transcriptional response of the I region promoters are critical players in the process of isotype selection. In this study, we have further elucidated the mechanism by which I region genes are induced in CD40-stimulated B cells and report a number of novel findings that extend our understanding of specific factors regulating class-directed CSR. Using the human Iγ1 promoter and the Ramos 2G6 B cell line, which is capable of undergoing CSR to CH1, we have demonstrated that in addition to the CD40RR containing three NF-κB sites critical for CD40-induced expression, the κB6 site is equally important for transcriptional induction. In fact, mutation of the κB6 site reduces overall promoter activity to a similar extent as mutating any of the three NF-κB sites in the CD40RR. This finding suggests that occupancy of the κB3–κB5 sites is necessary, but not sufficient for optimal activation, and that the κB6 site has an essential role in CD40-dependent activation of the Iγ1 promoter.
Our previous identification and functional analysis of CREB/ATF binding sites in the Iγ1 promoter revealed a role for regulating the magnitude of the transcriptional response and that the absence of this binding in the Iγ3 promoter corresponded to a significantly weaker response to stimulatory signals (31). Therefore, based on these findings we predicted that optimal promoter activity would require both binding of NF-κB and CREB/ATF to sequences within the 36-bp element. Our studies using competitor sequences demonstrated that CREB and NF-κB are capable of binding to the 36-bp element independently. However, insertion of a 5-bp spacer that introduced a half-helical turn between binding sites and shifted the orientation of the proteins relative to each other revealed a sharp drop-off in promoter function. The fact that activity was fully restored when a 10-bp insertion properly rephased the binding site strongly supports the idea that there is an orientation-dependent association between CREB and NF-κB that is required for activation. The physical constraints of this association may reflect a direct interaction between CREB and NF-κB and/or the coactivator, p300. In support of this idea, results from gel retardation assays showed that optimal formation of the B6 complex required proper phasing of the NF-κB site within the context of the 36-bp element (data not shown). This finding is highly reminiscent of the IFN-γ and the bfl-1 enhancesomal complexes and their functional dependence on the orientation of individual transcription factor binding sites (22, 46, 47). Changing the phasing or reversing the orientation of specific sites dramatically decreased the stability of the inducible complexes.
ChIP analysis of promoter occupancy in both Ramos and primary IgM+IgG− B cells revealed important clues regarding the association of trans-acting and coactivating factors with the quiescent and activated Iγ1 promoter. Previous in vitro and in vivo work points toward the importance of NF-κB in regulating GL transcription (48, 49, 50). Specifically, p50-deficient mice were unable to produce IgG3, IgE, and IgA isotypes, which was correlated with a decreased level of Iγ3 and Iε GL transcripts (48). Surprisingly, in experiments aimed at identifying the role of NF-κB in IL-4- and CD40L-mediated activation of the γ1 locus, a 17-kb transgene containing NF-κB mutations within the CD40RR appeared to be regulated similarly to an unmutated transgene, as well as the endogenous locus, in response to signaling (51). One possible explanation for these contradictory results is that NF-κB may be required for establishing an open configuration through its recruitment of additional factors such as p300. These factors may contribute to chromatin remodeling of the endogenous locus but may be dispensable with respect to the altered structure of the transgene. Alternatively, the 36-bp region is intact in the transgene and may compensate for the absence of activating sites within the CD40RR.
Our finding that there was little to no change in ChIP factor binding between unstimulated and CD40-stimulated Ramos cells may be explained by the fact that only a small number of cells are being induced to express endogenous GL transcripts in response to CD40 signals (26). Alternatively, recruitment of additional proteins or modifications to bound proteins would fail to give a quantitative change in binding. For example, the NF-κB p50/p50 homodimer is associated with transcriptional repression (52, 53, 54, 55), and the exchange of dimers can lead to a modulation of NF-κB activity (56) via the recruitment of distinct coactivators (57). The fact that CREB is bound to the promoter in nonactivated Ramos and primary B cells, possibly in association with p300, may reflect the tendency of the promoter to be either readily activated or in a partially activated state. The increased recruitment of p300 in primary B cells upon CD40 signaling is expected based on our in vitro data and reflects the induction of GL transcription associated with activation.
The results presented here support a role for p300 in GL transcription. The finding that overexpression of p300, but not a HAT mutant, enhanced transcription from the Iγ1-512 construct, which is in an open chromatin configuration, indicates that p300 is directly or indirectly modifying activation-induced transcription factors. For example, p300 may respond to activation signals by selectively acetylating transcription factors. This type of control has been shown to underlie TNF-α-induced binding of NF-κB to DNA, where modification of distinct lysine residues of p65 by CBP/p300 regulates NF-κB-specific functions including transcriptional activation, DNA binding affinity, IκBα assembly, and subcellular localization (reviewed in Ref.58). Another possibility is that p300 provides a scaffold for CREB/ATF and NF-κB binding at both the CD40RR and κB6 site. This physical bridge could link the basal transcription factors to the activation-induced factors. Finally, p300 may act to recruit other proteins to the scaffold that modify the trans-activating and/or coactivating factors. In support of this idea, it has been shown that CBP/p300 binds specifically to phosphorylated p65 modified by protein kinase A (59, 60). It is likely that multiple p300-directed mechanisms are functioning to integrate activation and basal signals to produce a specific level of transcription.
Thus, our current model of Iγ1 transcriptional regulation is that binding of CREB and NF-κB to the nonactivated Iγ promoter nucleates a transcriptional regulatory complex containing p300. CD40-mediated signals up-regulate the level of p300 potentially resulting in a threshold that allows for acetylation and recruitment of additional chromatin remodeling complexes. Also, optimal transcription requires full occupancy of κB3–κB6 sites as well as the CREB site. Future experiments will focus on identifying modifications of chromatin-bound transcription factors that result in the selective activation of the I region promoters in response to B cell-specific signals.
We thank Drs. Joan Boyes (Institute of Cancer Research, London), Elizabeth Moran (Temple School of Medicine, Philadelphia, PA), and Eileen White (Rutgers University, Piscataway, NJ) for the generous gifts of p300 and E1A 12S expression plasmids. The donation of Abs from Dr. Nancy Rice (National Cancer Institute, Bethesda, MD), as well as the technical advice from Dr. Celine Gelinas (University of Medicine and Dentistry of New Jersey, Piscataway, NJ; CBER, Food and Drug Administration (FDA), Bethesda, MD), Ralph Bernstein (CBER, FDA, Bethesda, MD), and Caroline Woo (Albert Einstein College of Medicine, Bronx, NY) were critical to the success of ChIP assays. We recognize Frank L. Sinquett for his technical contributions to this project, as well as all of the members of the Covey laboratory who provided valuable insight and helpful discussions.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by a National Institutes of Health Grant R01AI3708 (to L.R.C.) and a National Institutes of Health Training Fellowship T32AI07403 (to R.L.D.).
Abbreviations used in this paper: CBP, CREB binding protein; HAT, histone acetyltransferase; ATF, activating transcription factor; CSR, class switch recombination; GL, germline; AID, activation-induced cytidine deaminase; CD40RR, CD40 response region; ChIP, chromatin immunoprecipitation; RLU, relative light unit.