The Immunoglobulin Heavy Chain Locus Control Region Increases Histone Acetylation along Linked c-myc Genes

Abstract

In chromosome translocations characteristic of Burkitt lymphomas (BL) and murine plasmacytomas, c-myc genes become juxtaposed to immunoglobulin heavy-chain (IgH) sequences, resulting in aberrant c-myc transcription. Translocated c-myc alleles that retain the first exon exhibit increased transcription from the normally minor c-myc promoter, P1, and increased transcriptional elongation through inherent pause sites proximal to the major c-myc promoter, P2. We recently demonstrated that a cassette derived from four DNase I-hypersensitive sites (HS1234) in the 3′Cα region of the IgH locus functions as an enhancer-locus control region (LCR) and directs a similar pattern of deregulated expression of linked c-myc genes in BL and plasmacytoma cell lines. Here, we report that the HS1234 enhancer-LCR mediates a widespread increase in histone acetylation along linked c-myc genes in Raji BL cells. Significantly, the increase in acetylation was not restricted to nucleosomes within the promoter region but also was apparent upstream and downstream of the transcription start sites as well as along vector sequences. Histone hyperacetylation of control c-myc genes, which was induced by the deacetylase inhibitor trichostatin A, mimics the effect of the HS1234 enhancer on expression from the c-myc P2 promoter, but not that from the P1 promoter. These results suggest that the HS1234 enhancer stimulates transcription of c-myc by a combination of mechanisms. Whereas HS1234 activates expression from the P2 promoter through a mechanism that includes increased histone acetylation, a general increase in histone acetylation is not sufficient to explain the HS1234-mediated activation of transcription from P1.

In the B-cell tumors Burkitt lymphoma (BL) and murine plasmacytoma, one c-myc allele becomes juxtaposed to immunoglobulin (Ig) sequences through a reciprocal chromosomal translocation (11, 29, 50). As a result of this recombination, the translocated c-myc gene remains transcriptionally active, whereas the unrearranged allele is silenced, as c-myc genes normally are during B-cell differentiation (3, 35). In translocation events that leave the first c-myc exon intact, expression from the translocated gene is further distinguished by increased initiation from the normally minor c-myc promoter, P1, and an increased ability of RNA polymerase II complexes to elongate through pause sites proximal to the major c-myc promoter, P2 (6, 23, 52, 53, 56). These findings support a model in which sequences present in the Ig locus deregulate expression from cis-linked c-myc alleles both by maintaining the gene in a chromatin structure permissive for transcription and by promoting interactions between c-myc and Ig regulatory elements that affect c-myc initiation and elongation.

The region 3′ of the Cα gene within the murine Ig heavy-chain (IgH) locus and c-myc segregate to the same chromosome following recombination in t(15;12) plasmacytomas (11). A series of four B-cell-specific and cell-stage-dependent DNase I-hypersensitive sites (HSs), which were denoted HS1 to HS4, map 10 to 30 kb 3′ of the Cα gene and constitute lymphoid cell-specific and developmentally regulated enhancer elements in transient transfection assays (14, 28, 32). We demonstrated previously that a 6.5-kb DNA fragment comprising sequences within these HSs, HS1234, is sufficient to affect c-myc transcription in stable transfections in BL and plasmacytoma cell lines, when linked 2.3 kb upstream of the c-myc promoter region (28). HS1234 c-myc genes showed increased initiation from the P1 promoter and increased transcriptional elongation past exon 1. In addition, HS1234-linked genes were expressed in a position-independent, copy-number-dependent manner when stably integrated into a plasmacytoma cell line. We interpreted these findings to indicate that the HS1234 fragment could regulate both chromatin structure and aspects of transcription over a distance and thus could function as a locus control region (LCR) in these cultured cells.

The deregulated expression of c-myc genes linked to the HS1234 fragment could arise through a number of different processes. For example, an early model proposed that the assembly of transcription complexes highly efficient in elongation occurs preferentially at the P1 promoter compared to the P2 promoter (48, 49). Accordingly, the primary effect of elements bound to cis-linked sequences within the IgH locus (or the HS1234 fragment) might be the specific activation of P1-initiated transcription. However, characterization of plasmacytoma lines has indicated that constitutive c-myc expression in these cell types is not strictly dependent on the P2-to-P1 promoter switch (66). Thus, the altered expression of c-myc genes linked to the HS1234 fragment might be directed by a mechanism more general than promoter switching.

Recent studies have emphasized the essential role that chromatin structure plays in the regulation of gene transcription. By restricting access of transactivators to DNA, nucleosomes and linker histones generally repress promoter function (22, 26, 39, 55, 59). Evidence from a number of systems indicates that a primary aspect of enhancer function is to counteract this repressive chromatin structure, thereby enabling transcriptional initiation (13, 30, 63). Factors bound to or associated with enhancers may contribute to gene activation by directing the reorganization or displacement of nucleosomes from crucial regulatory sequences and the transcriptional start site (reviewed in reference 55). Furthermore, findings that some transcriptional cofactors, including GCN5, PCAF, CBP/p300, and TAF₁₁230/250, possess histone acetyltransferase (HAT) activity suggest that enhancer regions could specifically recruit HAT-containing molecules to chromatin domains and promoter regions (4, 10, 33, 36, 67). In vitro, histone hyperacetylation facilitates the binding of some trans factors to sites within nucleosomal DNA as well as stimulates transcriptional initiation along chromatinized templates (27, 34, 62).

Chromatin structure has also been implicated in regulating the promoter clearance and/or elongation steps of transcription. For example, the duration of transcriptional pausing at inherent sites along a template in vitro is increased when the template is assembled into chromatin compared to naked DNA (21). Enhancer-bound activators and cofactors may direct the recruitment of an elongation factor such as TFIIF, TFIIS, elongin, or P-TEF or a chromatin remodeling complex such as SWI/SNF to RNA polymerase to facilitate processive transcription, as has been observed in vitro (31, 38, 44, 51, 57). Alternatively, enhancer-associated proteins might directly modify the underlying chromatin to expedite transcription through pause sites. For example, the in vivo release of pausing along the human hsp70 gene by heat shock is accompanied by an alteration in nucleosome structure in the gene’s promoter-proximal region, as assessed by restriction enzyme accessibility; this remodeling event occurs independently of transcription (9). The efficiency of transcriptional elongation might also be influenced by the acetylation state of nucleosomes downstream of the start site, since histone hyperacetylation has been postulated to induce chromatin unwinding (60). In this regard, the direct recruitment of CBP or PCAF to transiently expressed reporter genes has been shown to increase both transcriptional initiation and elongation (24).

In this report, we investigate the mechanism by which the HS1234 enhancer deregulates transcription of linked c-myc genes. We have conducted our studies of c-myc alleles stably maintained in the Raji BL cell line on Epstein-Barr virus (EBV)-derived episomal vectors (pHEBO) (54). Episomally maintained genes are not subject to variable position effects that often prohibit the expression of integrated control c-myc alleles. This allows us to compare directly aspects of transcription and chromatin structure along control and enhancer-containing templates. As described below, the HS1234 fragment stimulates elongation-efficient transcription from both the P1 and the P2 promoters; thus, the promoter switch observed in this cell type is not required for high-level c-myc expression. Analyses of chromatin organization along control and HS1234-linked c-myc templates reveal that enhancer-linked alleles undergo limited structural remodeling upstream of the P1 promoter. Notably, however, chromatin immunoprecipitation assays demonstrate increased histone acetylation of nucleosomes associated with HS1234-linked c-myc genes, within both regulatory and transcribed regions, compared to control alleles. Furthermore, general histone hyperacetylation induced by the deacetylase inhibitor trichostatin A (TSA) differentially activates transcription from the P2 promoter of control compared to enhancer-linked c-myc genes and inhibits P1 transcription from HS1234-linked templates. These results suggest that the c-myc P1 and P2 promoters are activated through different mechanisms mediated by the HS1234 enhancer-LCR.

MATERIALS AND METHODS

Plasmid constructs and Raji cell transfections.

The c-myc control and HS1234 c-myc episomal vectors have been described previously (28). Briefly, the 8.1-kb HindIII-EcoRI human c-myc genomic fragment was cloned into HindIII-BamHI sites in pHEBO, an episomal vector containing the EBV origin of latent replication (oriP) and a thymidine kinase (TK)-driven hygromycin B resistance gene (54). A 6.5-kb fragment of genomic DNA comprising the murine IgH 3′Cα HS1 to HS4 was cloned immediately upstream of the c-myc HindIII site in a 5′-to-5′ orientation. Construction of the ΔP1 promoter mutant construct has been described previously; 29 bp of sequence from the TATA box to +1 was deleted by oligonucleotide-mediated mutagenesis (49).

Stably transfected pools of Raji BL cells were generated by electroporating 10⁷ cells with 40 μg of episomal DNA and then selecting with 1 mg of hygromycin B per ml for 2 weeks in RPMI 1640 medium containing 10% fetal bovine serum. During passage, transfected Raji cell pools were maintained at 400 to 500 μg of hygromycin B per ml.

DNA and RNA analyses.

Total cellular DNA was isolated by standard techniques, and episomal copy numbers were determined by Southern blot analysis. During extended passage (>6 months), we found the episome copy numbers of some pools to be unstable, and Southern analyses were repeated at various time points. Episome copy numbers of Raji-transfected pools varied between 50 and 120 for c-myc and ΔP1 c-myc and between 10 and 50 for HS1234 c-myc and HS1234 ΔP1 c-myc.

Steady-state expression of human c-myc and GAPDH in Raji cell pools was determined by S1 protection analyses of total RNA isolated with RNAzol B (Tel-Test, Inc.). The translocated c-myc allele in Raji BL contains a deletion of sequences at the end of the first exon that allows for the specific detection of expression from wild-type c-myc genes in these cells (43). c-myc S1 probes were generated by unidirectional PCR as follows. A 22-base oligonucleotide complementary to c-myc sequences +505 to +527, relative to P1, was phosphorylated with [γ-³²P]ATP, combined with 5 μg of SmaI-digested c-myc plasmid, heated at 94°C for 10 min, and used in a 10-cycle PCR. Probes used to detect ΔP1 mutant transcripts were similarly made, with the ΔP1 template for PCR amplification. End-labeled single-stranded PCR products were purified on 6% acrylamide–urea gels. The human GAPDH S1 probe was an end-labeled 71-base oligonucleotide that had a 55-base identity to the human GAPDH antisense strand. S1 assays were performed as described elsewhere (28), with 25 μg of total RNA and 1 × 10⁴ to 5 × 10⁴ cpm of each probe.

Nuclear run-on assays.

Raji cells were collected, washed once in phosphate-buffered saline (PBS), resuspended in 5 ml of RSB (10 mM Tris-Cl [pH 7.4], 10 mM NaCl, 5 mM MgCl₂), inverted several times after addition of 45 ml of RSB plus 0.25% Nonidet P-40 (NP-40), and then spun at 1,000 rpm for 5 min to collect nuclei. Buffers and nuclei were maintained at 4°C throughout. Nuclei were resuspended in freezing buffer (50 mM Tris-Cl [pH 8], 5 mM MgCl₂, 40% glycerol, 0.5 mM dithiothreitol [DTT]) at 2 × 10⁷ to 5 × 10⁷ per 210 μl. Run-on reactions were performed with [α-³²P]CTP in 150 mM KCl as described elsewhere (23), with the following modifications. After the run-on reaction, the nuclei were treated with DNase I (Worthington Biochemicals) for 15 min at 37°C and then with proteinase K for 45 to 60 min at 55°C. Labeled RNA was purified by three phenol-chloroform (1:1) extractions and passage through a Sephadex G50 column. Labeled RNAs were hybridized to excess c-myc single-stranded probes bound to GeneScreen Plus membranes; following hybridization, the filters were treated with RNase A and washed as described elsewhere (28). The positions, relative to the P1 promoter, and the cytidine contents of the c-myc run-on probes were as follows: PO, −672 to −104 with 175 Cs; SR, −104 to +153 with 76 Cs; NS, +208 to +330 with 40 Cs; RS, +330 to +510 with 66 Cs; and SA, +936 to +1072 with 43 Cs.

Chromatin assays.

Nuclei for chromatin assays were prepared as follows. Raji cells were washed once in PBS, resuspended in 10 ml of RSB, inverted after the addition of 4 ml of RSB plus 0.25% NP-40, pelleted, resuspended in 5 ml of RSB, and spun through a 30% sucrose cushion in RSB. Buffers and nuclei were maintained at 4°C throughout, and buffers contained either 5 mM sodium butyrate or 500 ng of TSA per ml.

For HS mapping, aliquots of approximately 10⁷ nuclei were treated at 37°C for 15 min with DNase I at final concentrations of 0.1 to 5 mg/ml in RSB. Nuclei were lysed and treated with proteinase K in 2× stop buffer (0.6 M NaCl, 20 mM Tris-Cl [pH 7.6], 10 mM EDTA, 1% sodium dodecyl sulfate) at 37°C for 16 h. DNA was precipitated, treated with RNase A, phenol-chloroform extracted, and reprecipitated prior to restriction enzyme digestion. Southern analysis was performed with 20 μg of Bgl2-digested DNA by using the c-myc 721-bp Bgl2-EcoRV fragment as a hybridization probe.

For micrococcal nuclease (MNase) mapping of nucleosomal boundaries, aliquots of approximately 10⁷ nuclei were treated with 1.5 to 50 U of MNase (Pharmacia) at room temperature for 5 min in 1 ml of digestion buffer (10 mM Tris-Cl [pH 7.5], 10 mM NaCl, 3 mM MgCl₂, 1 mM CaCl₂). Digestions were terminated by the addition of 0.1 ml of MNase stop (100 mM EDTA, 10 mM EGTA), and DNA was isolated and purified as described above. The extent of MNase digestion was evaluated with agarose gels, and DNA fractions that appeared to be equally digested were used for positioning analyses. For Southern blots, 20 μg of DNA was digested with either StyI or XbaI and separated on 1.5% agarose gels, and the blots were hybridized to probes of the 200-bp StyI-SpeI or the 239-bp XbaI-SacI c-myc fragment.

For restriction enzyme accessibility assays, 10⁶ nuclei were digested for 30 min with 40 U of enzyme in 0.1 ml at either 27°C (SmaI) or 37°C. Digestions were performed with New England Biolab buffers 2 or 4 as appropriate and in the presence of bovine serum albumin. In the TSA experiment, both TSA-treated and untreated samples were digested in the presence of 500 ng of TSA per ml; all other nucleus digestions were done in the absence of TSA. After nucleus digestion, DNA was isolated and purified as described above. Southern analyses were performed with 10 μg of DNA that had been digested to completion as follows. Those that hybridized to probe d, a 3.5-kb HindIII-XbaI fragment, were digested with HindIII and XbaI; those that hybridized to probe e, a 1,113-bp XhoI-XbaI fragment, were digested with XhoI and SalI; and those that hybridized to probe f, a 1,878-bp XbaI-HpaI fragment, were digested with XbaI and SalI. Hybridization signals were quantitated with a PhosphorImager, and percent digestion was calculated as the ratio of signal in detected product(s) to total signal.

Immunoselection of chromatin.

Immunoprecipitations with the antiacetyllysine antibody were performed as described elsewhere (17, 18), with the following modifications: 1 × 10⁸ to 2 × 10⁸ nuclei were digested with 225 U of MNase at room temperature for 10 min in 1 ml of digestion buffer (above) supplemented with 0.1 mM each sodium butyrate and phenylmethylsulfonyl fluoride (PMSF). Digestions were terminated with 10 mM EDTA, and chromatin was extracted as described; no steps were taken to deplete chromatin preparations of histone H1. Combined chromatin fractions were concentrated on Centricon C30 (Amicon) columns and were spun through 5 to 25% linear sucrose gradients with an SW41 rotor at 36,000 rpm for 14 h at 4°C. Fractions containing mononucleosomal chromatin were combined and concentrated, and 50 μg of this input chromatin was incubated with 20 μg of affinity-purified acetyllysine antibody in incubation buffer (50 mM NaCl, 10 mM Tris-Cl [pH 7.5], 1 mM EDTA, 10 mM sodium butyrate, 0.1 mM PMSF) for 16 h at 4°C. Immunocomplexes were collected with protein A-Sepharose, and DNA was purified from antibody-bound and -unbound fractions as described elsewhere (17, 18).

Slot blots of input, antibody-unbound, and antibody-bound DNAs were made with 500 ng of each fraction by the protocol supplied with the GeneScreen Plus membrane. The filters were hybridized with the c-myc fragment probes indicated; HindIII-XbaI was 3,507 bp, HindIII-SmaI was 2,226 bp, AccI-NaeI was 818 bp, SmaI-NaeI was 312 bp, and XbaI-Bgl2 was 1,878 bp. The pHEBO vector probe was a 1.5-kb EcoRI-FspI fragment containing plasmid sequence. After hybridization to the various c-myc probes, slot blot filters were hybridized to an end-labeled telomere repeat-recognizing oligonucleotide, (GGGTAA)_n, at 37°C in 50% formamide. Signals were quantitated by PhosphorImager analysis, and signal higher than that of wild-type Raji was used in the bound-input calculations.

RESULTS

HS1234 activates the c-myc P2 promoter independently of transcription from P1.

As described in the introduction, stable transfection of Raji BL cells with a vector containing human c-myc linked 2.3 kb downstream of the HS1234 fragment results in a 50-fold increase in c-myc expression per DNA copy, relatively efficient transcriptional elongation, and an increase in the P1/P2 promoter use ratio (28). To determine whether the observed P2-to-P1 shift in promoter usage is required for the HS1234 effect on c-myc transcription, we deleted the P1 promoter region of our c-myc gene and assayed transcription from P2 in the absence of P1. The ΔP1 template used in these studies contained a 29-bp deletion from the TATA box to the P1 CAP site (Fig. 1A and B).

FIG. 1.

The IgH HS1234 enhancer fragment activates the c-myc P2 promoter independently of P1 promoter activity. (A) Schematic diagram of the murine IgH locus depicting the relative location of the IgH intronic enhancer, Eμ, the 3′ Cα regulatory region, containing HS1 to HS4, and representative chromosomal breakpoints mapped in various t(15;12) plasmacytomas. Only selected restriction sites are shown. E, EcoRI; X, XbaI; H, HindIII; S, Sau3A; P, PstI. The Cα 3′ enhancer is represented by an open circle. DNA fragments comprising HS1 to HS4 were cloned upstream, in a 5′-to-5′ orientation, of human c-myc in the episomal vector pHEBO. (B) S1 protection analysis of steady-state expression from wild-type and ΔP1 mutant constructs. The translocated c-myc allele in Raji BL contains a deletion at the end of exon 1 (open box) that allows the end-labeled probe to be specific for transcription from unrearranged c-myc genes. Transcripts initiating from the c-myc P1 and P2 promoters are indicated, along with control transcription from the human GAPDH gene. (C) Nuclear run-on analyses of untransfected (untx) Raji BL cells and of cell pools containing control or enhancer-linked c-myc episomes. (D) Distribution of polymerase complexes along the c-myc templates. Hybridization signals in panel C were quantitated by PhosphorImager analysis, and GAPDH normalized values above those from untransfected Raji cells were corrected for the cytidine content of each probe. Graph indicates the self-distribution of polymerase density along each template. Cytidine contents were as follows: PO, 175; SR, 76; NS, 40; RS, 66; and SA, 43.

As we had observed previously, transcription from the unrearranged c-myc allele in our Raji BL line is generally not detectable. Similarly, the level of expression from the enhancerless wild-type and ΔP1 c-myc genes was also quite low. In contrast, linkage of the HS1234 fragment to both templates resulted in a 50- to 100-fold increase in expression per episome copy. To distinguish HS1234-mediated effects on transcriptional initiation from elongation, nuclear run-on analyses were performed on these Raji pools; the results from an average experiment are shown in Fig. 1C and D.

As an indication of elongational attenuation, we found a disproportionately high density of polymerase complexes within the P2 proximal region (NS probe) along both the wild-type c-myc and the ΔP1 c-myc templates, compared with a more 3′ region of the gene (SA probe). In contrast, transcripts initiated from HS1234-linked genes demonstrated more efficient elongation, with a higher percentage of transcribing complexes proceeding further 3′ into the gene; the ratio of SA to NS polymerase density increased 5-fold along the control c-myc template and 10-fold along the ΔP1 promoter mutant. The HS1234 enhancer had a relatively smaller effect on initiation along these two templates; the NS signal per episome copy increased by two- to threefold along both.

The above results indicate that although the c-myc P1 promoter is highly activated by the HS1234 enhancer, a P2-to-P1 promoter switch is not required to achieve high-level expression from c-myc genes. Hence, the basic mechanism underlying the deregulated expression of c-myc alleles linked to the HS1234 fragment may reflect a more general change along the c-myc template.

HS1234 induces limited chromatin remodeling along linked c-myc genes.

Alterations in chromatin organization are typically detected in assays of DNase I HS formation, nucleosome positioning, and the accessibility of particular DNA sequences, when in chromatin, to restriction enzymes. Therefore, we used these assays to determine if the activation of c-myc transcription by the HS1234 enhancer is associated with chromatin remodeling of the c-myc promoter region.

(i) Analysis of DNase I HSs.

A characteristic pattern of DNase I HSs is present in the upstream region of actively transcribing c-myc genes (12, 46, 47). To determine whether the HS1234 enhancer stimulated the formation of DNase I HSs within the c-myc promoter region in conjunction with its effect on transcription, we analyzed the formation of these sites along control and HS1234-containing templates. Nuclei from untransfected Raji cells and Raji cells containing either control c-myc or HS1234-linked c-myc episomes were treated with increasing amounts of DNase I, purified genomic DNAs were digested with Bgl2, and HSs present in the c-myc upstream regions were detected by a Southern blot with an EcoRV-Bgl2 probe (probe a in Fig. 2B).

FIG. 2.

The HS1234 enhancer has little effect on the chromatin organization of linked c-myc genes on episomal templates. (A) DNase I mapping of HSs in the upstream and promoter regions of control and enhancer-linked templates. Genomic DNAs purified from DNase I-treated nuclei were digested with Bgl2 and used in a Southern analysis with probe a, a Bgl2-EcoRV fragment indicated on the map shown in panel B. Major HSs I, II₂, III₁, III₂, and V are indicated, and minor sites within the first intron are denoted by open circles. Hybridization fragments larger than 5,385 bp originate from vector or HS1234 sequences upstream of the c-myc HindIII site. (B) Map of restriction sites and probes used in these analyses. (C and D) MNase mapping of nucleosome positioning along control and HS1234-linked c-myc episomes. Genomic DNAs purified from MNase-treated nuclei were digested with either Sty1 (C) or XbaI (D) and used in Southern analyses with probes b and c, as indicated below the blots and on the map in panel B. The locations of upstream regulatory sequences, the P1 and P2 promoters, and the exon 1-intron 1 junction (Int1/Ex1) are indicated alongside the blots. Region of increased MNase sensitivity along HS1234-linked templates within the c-myc promoter and downstream of the transcriptional start sites is indicated by the asterisk in panel D. WT, wild type.

As can be seen in Fig. 2A, a very similar pattern of DNase I HSs forms along the control and enhancer-linked c-myc episomes in Raji cells (indicated on the map in Fig. 2B). We did observe subtle differences in the relative intensities of particular HSs along the two templates that are consistent with the observed expression from these templates in the BL cell line. For example, HS III₁, which is frequently associated with high P1 promoter activity, formed more prominently along HS1234-linked c-myc genes than along control templates. In contrast, several minor HSs, mapping to intron 1 and correlating with down-regulated c-myc transcription, are more prevalent along control c-myc templates (7). Overall, however, our analysis indicates that the differential expression of HS1234-linked and control c-myc genes in the BL cell line is not a result of gross changes in chromatin structure, as assessed by DNase I hypersensitivity.

(ii) Nucleosome positioning.

Nucleosome positioning along control and HS1234-linked c-myc genes was assessed by MNase digestion and Southern analyses with indirect end labeling. Figure 2C and D shows representative MNase digestion profiles of chromatin encompassing the c-myc promoter region and first exon. Nucleosomes upstream of the promoter region were mapped by StyI digestion, which cuts 1.6 kb 5′ of P1, and hybridization to probe b (Fig. 2B), whereas nucleosomes within the promoter region and exon 1 were mapped by XbaI digestion, which cuts in intron 1, and hybridization to probe c (Fig. 2B). The relative positions of the c-myc P1 and P2 promoters and the exon 1-intron 1 boundary are indicated along the Southern blots.

As shown in these figures, the chromatin surrounding the c-myc P1 and P2 promoters is packaged in an array of nucleosomes, regardless of the presence of the HS1234 enhancer. Our analyses revealed no evidence of nucleosome displacement or large-scale reorganization along HS1234-linked c-myc genes compared to control genes. Enhancer-linked c-myc templates appeared slightly more sensitive to MNase digestion within the promoter region and downstream of the P2 initiation site than control templates (denoted by asterisk in Fig. 2D). Major hybridization bands within these regions were relatively less intense and more diffuse along HS1234 c-myc templates, and multiple minor cleavage sites were detected as well. This subtle difference in MNase susceptibility might arise from a relaxation of DNA-histone contacts or from less rigid nucleosome positioning or displacement on a population of HS1234 c-myc templates; both events could facilitate assembly of transcription complexes at the P1 and P2 promoters.

(iii) Restriction enzyme accessibility.

As part of our investigation into the effect of the HS1234 enhancer on the chromatin structures of linked c-myc genes, we compared the accessibility of restriction sites along the two chromatinized episomal templates. A representative example of our analysis is shown in Fig. 3B, and the locations of the evaluated sites and probes used in hybridizations are shown on the map in Fig. 3A.

FIG. 3.

The HS1234 enhancer has a limited effect on restriction enzyme accessibility of linked c-myc genes on episomal templates. (A) Partial map of the c-myc episomal constructs indicating the locations of restriction sites and probes used to evaluate enzyme accessibility along c-myc templates. (B) Comparison of restriction enzyme accessibility in control and HS1234-linked c-myc nuclei. Nuclei from untransfected Raji cells and from control c-myc or HS1234 c-myc transfected cell pools were digested with various restriction enzymes and used in Southern analyses with the probes listed above the gels. Shown below each panel are the percentages of hybridization signal in detected digestion products. Representative samples from wild-type (WT) Raji cells are shown without detail. Specific restriction digests and probe fragments are described in Materials and Methods.

As shown in Fig. 3B, we observed a trend of slightly increased enzyme accessibility along HS1234-linked c-myc genes at some sites between the enhancer and the c-myc promoters, but never more than a twofold difference compared to our control c-myc templates. The lack of a significant change in accessibility at these distal sites is consistent with the MNase results of Fig. 2C, which also indicated no major alteration of the nucleosome array within this region.

A more significant increase in enzyme accessibility along the HS1234 templates was detected at two restriction sites very close to the P1 and P2 promoters. A Sma site located 102 bp 5′ of P1 reproducibly digested two- to fivefold more extensively in the presence of the HS1234 fragment; this increase does not appear to result from differential methylation at this site (data not shown). The increased accessibility at this Sma site along HS1234 c-myc is consistent with the more prominent formation of the closely located HS III₁ and coincides with increased transcription from the P1 promoter of HS1234-linked c-myc genes. Accessibility at the Xho site, which was located between P1 and P2, was also consistently higher along HS1234-containing templates than on control c-myc genes. Enzyme accessibility at sites further 3′ in the gene did not appear to be influenced by linkage to the HS1234 enhancer (Fig. 3B).

The above assays indicate that the chromatin organizations of control and HS1234-linked c-myc episomal genes in Raji BL cells are quite similar, despite the significant difference in the transcriptional activities of these templates. Subtle variations in structure immediately 5′ of the P1 promoter along HS1234-linked c-myc templates coincide with high-level expression from P1 and may indicate a role for chromatin remodeling in P1 activation.

HS1234-linked c-myc genes are hyperacetylated compared to control c-myc genes.

Nucleosomes within transcriptionally competent genes often contain histones that are highly acetylated at specific amino-terminal lysine residues. Studies indicate that histone hyperacetylation facilitates the binding of some transcription factors to sites within nucleosomes and may alter histone H1 association with chromatin as well (15, 27, 62, 65). These combined effects likely influence higher-order structure within a region and contribute to the transcriptional competence of particular genes (58, 60). Given the potential of enhancer-recruited cofactors to possess HAT activity, we compared the levels of histone acetylation along control and HS1234 c-myc templates by analyzing the relative abundance of c-myc sequences in the bound and unbound fractions of chromatin following immunoselection with an antibody to acetylated histones (17).

Nuclei from untransfected Raji cells and from cells stably transfected with either c-myc or HS1234 c-myc templates were digested with MNase under conditions that reduced approximately 10% of chromatin to mononucleosome length (Fig. 4A). Under these relatively mild digestion conditions, exogenous c-myc DNA does not appear to be selectively lost (smaller than mononucleosome length) from either the HS1234-linked or control gene preparations as determined by Southern hybridization (data not shown). Mono- and dinucleosomes were isolated by sucrose gradients and then incubated with an antibody that recognizes ε-acetyllysine; previous characterization of this antibody has shown that it complexes preferentially with highly acetylated histones (H2A, H2B, H3, and H4) (17, 18). DNA was prepared from antibody-bound and unbound fractions and 500 ng of input, unbound, and antibody-bound DNA was slot blotted and then sequentially probed with various c-myc or vector sequences and to a telomere repeat-recognizing oligonucleotide. Hybridization was quantitated by PhosphorImager analysis, and signal above Raji background was used to calculate a bound-to-input fraction ratio for each probe along the HS1234 c-myc and control c-myc templates (Fig. 4B). This bound-to-input fraction ratio represents enrichment of a particular sequence, due to immunoselection, over its abundance in the input fraction. (Note that the relative input signals for control and HS1234 c-myc samples are not equal due to the difference in episome copy numbers in these pools.)

FIG. 4.

HS1234-linked c-myc genes show increased histone acetylation compared to control c-myc alleles. (A) Ethidium bromide-stained agarose gel showing MNase-digested genomic DNAs, purified either before (lanes 1 to 3) or after (lanes 4 to 6) sucrose gradient separation. Lanes: 1 and 4, untransfected Raji cells; 2 and 5, control c-myc transfectants; 3 and 6, HS1234 c-myc transfectants. (B) Equal amounts of DNA purified from input (In) and from the antiacetyllysine antibody-bound (Bound) and unbound (Unb) fractions were slot blotted onto nylon membranes and then hybridized to the c-myc fragment probes indicated above the filters. Restriction sites used to generate these c-myc probes are indicated on the map at the top. Acetylation along the pHEBO vector was evaluated with a DNA probe containing only plasmid sequences. The specificity of the immunoselection reactions and the equal loading of DNA on the slot blots were confirmed by hybridization to a telomeric repeat-recognizing oligonucleotide. The episome copy number in control c-myc pools was approximately fivefold higher than that in HS1234 c-myc pools (approximately 50 and 10 copies, respectively). WT, wild type.

As shown in Fig. 4B, immunoselection with the antiacetyllysine antibody resulted in strong enrichment (3.8- to 10.6-fold) for c-myc sequences when linked to the HS1234 enhancer. We found 3- to 4-fold less enrichment for these sequences when analyzing chromatin from the unlinked c-myc transfectants. Interestingly, the increase in histone acetylation that we observed along HS1234-linked c-myc genes was not limited to the transcription unit, as revealed by hybridization to the HindIII-Sma probe. In addition, the c-myc promoter region of HS1234-linked genes showed an increase in acetylation comparable to but not higher than that of upstream sequences as revealed by the Sma-Nae probe. In fact, the magnitude of acetylation increase observed along the HS1234-linked templates appeared constant throughout the c-myc gene; exon 2 sequences detected by hybridization to the Xba-Bgl2 probe also were enriched by three- to fourfold. In this regard, hybridization of our filters to an episomal vector probe indicated that the presence of the HS1234 fragment was sufficient to result in higher levels of acetylation throughout the entire template compared to that found on the control c-myc vector. Consistent with the results of others, the telomere repeat-recognizing probe indicated hypoacetylation of telomeric sequences, thereby verifying the specificity of our results with the c-myc probes (37).

Thus, using immunoselection with an antiacetyllysine antibody, we have uncovered a pattern of increased histone acetylation that is directly linked to the presence of the HS1234 enhancer fragment. The increase in acetylation is found over the entire episomal vector and is not targeted to the c-myc promoter region or transcription unit.

TSA significantly activates the P2 promoter of control but not HS1234-linked c-myc genes.

To determine whether hyperacetylation of nucleosomes is sufficient to affect transcription of c-myc alleles in our Raji episome system, we treated wild-type and transfected cells with TSA, a specific inhibitor of histone deacetylases and assayed c-myc transcription. TSA has been demonstrated to increase the acetylation state of histones within treated cell lines (5, 61). Similarly, we observed increased acetylation of histone H4 following TSA treatment of our Raji cells as indicated by Triton-acetic acid-urea gel analyses (data not shown).

Consistent with the findings of others, we observed that TSA treatment of untransfected Raji cells increases expression from the unrearranged c-myc allele; however, expression remains low relative to that from our episomal genes. As shown in Fig. 5A, steady-state expression of c-myc from control episomal templates increased significantly with 9 and 18 h of TSA treatment. Although increases in transcription were apparent from both promoters, P2-initiated transcription was particularly activated, to a level 50-fold high than that of untreated cells. TSA induced P1 expression to a much lower degree, indicating that the P1 and P2 promoters may be differentially regulated by changes in acetylation states in this cell type. TSA treatment of HS1234 c-myc genes further emphasized the differential response of the two c-myc promoters to histone hyperacetylation; the high level of P1 expression induced by HS1234 initially was inhibited by TSA treatment, whereas P2 expression increased slightly over time (Fig. 5A). Actinomycin D studies indicate that TSA-induced changes in steady-state c-myc expression are not attributable to mRNA stability (data not shown).

FIG. 5.

Treatment of c-myc episomes with the deacetylase inhibitor TSA preferentially activates the P2 promoter on control genes and represses P1 transcription from HS1234-linked c-myc genes. (A) Untransfected Raji cells and pools containing control and HS1234-linked c-myc episomes were treated with 500 ng of TSA per ml for 9 or 18 h before mRNA collection. Un, untreated. Steady-state expression from the c-myc P1 and P2 promoters and from the control GAPDH gene were assayed by S1 protection as indicated in the legend to Fig. 1B. (B) Nuclear run-on analyses of Raji cell pools before and after a 9-h TSA treatment. The probes used are those detailed in the legend to Fig. 1. (C and D) S1 protection and nuclear run-on analyses, respectively, of transcription from untreated and TSA-treated ΔP1 c-myc- and HS1234 ΔP1 c-myc-containing Raji cell pools.

The HS1234 enhancer activates transcription from the P2 promoter largely by increasing the efficiency of transcriptional elongation through pause sites proximal to P2 (Fig. 1C and D). To determine whether histone hyperacetylation is sufficient to similarly increase elongation efficiency, we analyzed TSA-induced changes in transcription of our c-myc genes by nuclear run-on assays. As shown in Fig. 5B, TSA treatment increases both the level of initiation and the elongation competence of polymerase complexes transcribing control c-myc genes; NS/copy increased by three- to fivefold, and SA/NS increased by four- to sixfold following 9 h of TSA treatment. In agreement with the S1 results in Fig. 5A, transcription along HS1234-linked c-myc genes was less affected by TSA treatment; elongation efficiency increased by two- to threefold (SA/NS), whereas initiation (NS/copy) increased by less than twofold after TSA treatment.

The ability of TSA to increase initiation and elongation from the c-myc P2 promoter was confirmed by treatment and analysis of our ΔP1 mutant alleles. Consistent with our findings with wild-type c-myc alleles, P2-initiated mRNA expression from ΔP1 constructs increased by 50- to 200-fold with 9 and 18 h of TSA treatment (Fig. 5C). Nuclear run-on analyses of ΔP1 pools treated with TSA for 9 h revealed a 5- to 7-fold increase in elongation and a 3- to 5-fold increase in initiation (Fig. 5D). In contrast, we observed a much smaller effect of TSA treatment on the already activated P2 transcription of HS1234-linked ΔP1 alleles, similar to the effect of TSA on the activated P2 promoter of HS1234 c-myc genes. Steady-state expression from HS1234 ΔP1 c-myc alleles increased by 2- to 4-fold, and readthrough transcription increased by 2- to 3-fold with TSA treatment; initiation was unaffected.

In summary, these results indicate that the effect of the HS1234 enhancer on transcription from the c-myc P2 promoter can be mimicked on control templates by treatment of cells with an inhibitor of deacetylase activity, i.e., TSA. In contrast, the high-level induction of P1 transcription observed from HS1234-linked c-myc genes cannot be induced by TSA, suggesting that the HS1234-mediated activation of P1 requires a process other than increased general histone acetylation.

TSA has little effect on the chromatin structure of episomal c-myc templates.

Treatment of cells with a general inhibitor of deacetylase activity potentially affects the acetylation (and function) of a number of proteins in addition to nucleosomal histones, including transcription factors and HMG proteins (16, 40). In addition, TSA treatment of cells has been shown to alter chromatin structure along some genes in conjunction with its effect on transcription (5, 61). To determine whether, similarly to the HS1234 enhancer, TSA increases initiation and elongation of control c-myc genes in the absence of gross chromatin reorganization, we repeated our chromatin analyses on TSA-treated cell pools. Evaluations of DNase I HSs, MNase-sensitive sites, and restriction enzyme accessibility were performed as described for Fig. 2 and 3, and representative results are presented in Fig. 6 (a map of restriction sites and hybridization probes used is given in Fig. 6C).

FIG. 6.

Treatment of c-myc episomes with TSA has limited effect on chromatin structure. The locations of restriction sites and probes used in these analyses are described in the legends to Fig. 2 and 3. (A) DNase I HS mapping in the upstream and promoter regions of control and HS1234-linked c-myc templates following an 18-h treatment with 500 ng of TSA per ml. Included in each series is one lane from the TSA-untreated analysis shown in Fig. 2A for comparison. (B) MNase mapping of nucleosomes within the first exon (Ex1) and intron (Int1) of c-myc templates following an 18-h TSA treatment. Included in each series is one lane from the TSA-untreated analysis shown in Fig. 2D for comparison. (C) Map of restriction sites and probes used in these analyses. (D) Comparison of enzyme accessibility at the Sma site, 102 bp 5′ of P1, on control and HS1234 c-myc chromatin templates before and after an 18-h treatment of cells with TSA. The change in Sma accessibility attributable to TSA treatment is represented by the ratio of percent digestion in TSA treated to that in untreated samples.

As shown in Fig. 6A, the formation of DNase I HSs in the upstream c-myc region was not significantly affected by an 18-h TSA treatment of Raji cells containing either c-myc construct. (A single lane from the nontreated DNase I series in Fig. 2A is included for each template as comparison.) As before, subtle variations in the intensities of particular HSs were observed; the relevance of these minor differences remains unclear. Similarly, MNase analysis of nucleosome positioning within the first exon and over the promoter region of c-myc following TSA treatment revealed no evidence of nucleosome displacement from either template (Fig. 6B). However, as shown in Fig. 6B, both the HS1234-linked and control c-myc templates appeared somewhat more sensitive to MNase digestion following extended TSA treatment, as suggested by the greater predominance of multiple bands flanking the more compact major hybridization bands. Interestingly, this pattern of slightly increased MNase sensitivity of the c-myc template following TSA treatment is similar to that observed along untreated HS1234 c-myc episomal templates.

The accessibility of restriction enzymes to various sites within the upstream and promoter regions of c-myc templates following TSA treatment was also determined. In general, we observed that TSA treatment had less than a twofold effect on the level of digestion at each site assayed and that TSA treatment affected both c-myc templates comparably, i.e., there was a slight increase or decrease in digestion (data not shown). One exception to these generalizations, shown in Fig. 6D, occurred at the Sma site 102 bp 5′ of P1. Enzyme accessibility at this site increased by twofold along control c-myc genes induced by TSA treatment but decreased 50% along TSA-treated HS1234 c-myc templates. Thus, the relative accessibility at this Sma site correlates well with the relative activity of the P1 promoter on both TSA-treated and untreated templates.

In summary, the above studies indicate that TSA stimulates transcriptional initiation and elongation along control c-myc templates without inducing significant changes in chromatin structure. While we cannot discount the role highly acetylated nonhistone proteins may play in the TSA activation of control c-myc transcription, the above results are consistent with our conclusion that increased histone acetylation along c-myc genes, which is mediated by a linked HS1234 enhancer, is sufficient to affect aspects of c-myc transcription in the absence of pronounced chromatin reorganization.

DISCUSSION

We have previously reported that a 6.5-kb DNA fragment derived from sequences 3′ of the murine IgH Cα gene, HS1234, is sufficient to affect transcriptional initiation and elongation from linked c-myc genes in BL and plasmacytoma cell lines. In the present work, we have investigated the mechanism by which the HS1234 enhancer deregulates transcription of linked c-myc genes stably maintained on episomes in the Raji BL cell line.

HS1234 and TSA activate the c-myc P2 promoter in conjunction with increased histone acetylation.

Immunoselection of mononucleosomal chromatin with an antibody against acetyllysine revealed increased acetylation of HS1234-linked c-myc genes compared to control genes. Importantly, the increase in acetylation was not restricted to sequences within the c-myc promoter region; both the c-myc 2.3-kb upstream region and the episomal vector itself were more abundant in antibody-bound DNA prepared from HS1234 c-myc chromatin than that from control c-myc chromatin. The relevance of the HS1234-mediated increase in nucleosomal acetylation is suggested by our studies with the histone deacetylase inhibitor TSA. Treatment of Raji cell pools containing control and ΔP1 c-myc templates with TSA resulted in a pattern of activated transcription from the P2 promoter very similar to that observed from HS1234-linked genes. Both HS1234 linkage and TSA treatment induced large increases in P2-initiated mRNA, in conjunction with increased transcriptional initiation and efficiency of elongation. TSA treatment had a relatively smaller effect on the already highly expressing P2 promoters of HS1234-linked alleles, consistent with a model in which the HS1234 enhancer mediates activation of P2 through a process involving histone acetylation. Interestingly, in contrast to its effect on transcription from P2, TSA induced a significantly smaller increase in expression from P1 that did not approach the degree of activation conferred by the HS1234 enhancer. Thus, the c-myc P2 promoter can be activated by increased histone acetylation associated with the presence of the HS1234 enhancer, whereas a general change in histone acetylation is not sufficient to induce high-level P1 expression.

HS1234 and TSA induce high-level c-myc transcription without significant alteration of chromatin structure.

Gene induction is frequently accompanied by chromatin reorganization both within regulatory regions and downstream of the start site (51, 55, 64). Yet despite the significant effect of the HS1234 enhancer on initiation and elongation of linked c-myc transcription, we found many aspects of chromatin organization along the two templates to be similar. DNase I HSs characteristic of active c-myc genes were present in the upstream regions of both templates, and nucleosomes were detected over the promoter and proximal transcribed regions, regardless of HS1234 presence. Consistent with these findings, the accessibility of the c-myc upstream region to restriction enzymes was quite similar along control and HS1234 c-myc templates.

The lack of pronounced chromatin remodeling along c-myc templates linked to the HS1234 enhancer was somewhat surprising, given the large impact that HS1234 has on transcriptional initiation and elongation. Interestingly, chromatin analyses of templates activated by TSA treatment also revealed only minor variations in chromatin structure. These findings are in agreement with other studies reporting little difference in the chromatin structure of nonexpressing, uninduced c-myc genes and that of highly expressing alleles, induced by sodium butyrate (1, 42). In combination, our results implicate histone acetylation as a primary regulator of c-myc transcription in the Raji cell line. Notably, these results do not rule out the possibility that increased acetylation of a nonhistone protein(s) by TSA and factors associated with the HS1234 enhancer contribute to c-myc activation in this BL. The role of acetylated transcription factors and/or components of the RNA polymerase complex in regulating c-myc initiation and elongation requires further investigation.

Interestingly, all three of our chromatin assays revealed a subtle change in structure immediately 5′ of the P1 promoter that was associated with the presence of the HS1234 enhancer. The formation of HS III₁, the accessibility to Sma digestion 102 bp 5′ of P1, and the sensitivity of the promoter region to MNase digestion were all slightly increased along HS1234 c-myc genes. Although the significance of these subtle differences remains unclear, the opposing effect of TSA treatment on enzyme accessibility at the Sma site along the two templates suggests that P1 promoter activity may be regulated by acetylation of specific lysines and/or by the chromatin structure immediately upstream.

The role of acetylation in transcription.

Numerous studies have indicated a relationship between histone acetylation and the transcriptional activity of specific genes and chromatin domains (8, 37; for a review, see reference 41). Increased histone acetylation has been shown to facilitate the binding of some transcription factors to sites within nucleosomal DNA and may influence the composition and stability of higher-order chromatin structure. Recent characterization of transcription factors and associated proteins that possess intrinsic acetylase-deacetylase activity further suggests a mechanism by which changes in acetylation state may be targeted to specific promoters in order to regulate transcriptional activity. In vitro, the binding of enhancers to sites immediately upstream of the human immunodeficiency virus (HIV) promoter has been shown to increase acetylation of H4 along the template, concomitantly with increased HIV transcription (45). Although it was not determined in that study whether the increase in acetylation was limited to the promoter region, more recent work describing the requirement of HAT activity for the activation function of yeast Gcn5p in vivo did assess the localization of factor-mediated hyperacetylation along reporter genes (25). Interestingly, in that study, Kuo et al. found that overexpression of wild-type Gcn5p resulted in increased levels of acetylated H3 specifically within the promoters of target genes; increased H3 acetylation was not observed in more promoter-distal gene regions. These results support a model in which the modification of specific nucleosomes contributes to transcriptional activity.

However, despite the apparent promoter specificity for targeting by hyperacetylation suggested by the above studies, other evidence indicates that hyperacetylation affects entire chromatin domains as part of a mechanism to establish or maintain genes in a transcriptionally competent state. For example, in chicken embryo erythrocytes, the entire β-globin locus is hyperacetylated, regardless of transcriptional activity, and the physical boundaries of increased acetylation map closely to those defined by DNase I sensitivity (18). Furthermore, fluorescence in situ hybridization analysis of quiescent NIH 3T3 cells stimulated by external agents revealed induction of global H4 hyperacetylation apparently unrelated to specific gene activation; the widespread response in hyperacetylation was postulated to represent a modification that may serve as a prerequisite for transcription (2). Consistent with these examples of wide-scale changes in histone acetylation, we have found that increased histone acetylation mediated by the HS1234 enhancer is not limited to either nucleosomes within the promoter region or even to coding sequences of our episomal vector. Clearly, however, the antiacetyllysine antibody used in our studies immunoprecipitates all acetylated histones; thus, changes in acetylation of particular histones or lysine residues along HS1234-linked templates may actually localize to specific gene regions. An additional distinguishing aspect of our studies of HS1234-mediated hyperacetylation is the ability of HS1234 to activate transcription from a distance and to function as an LCR in cultured cell lines. It is possible that one component of LCR function is the regulation of histone acetylation throughout a domain, as opposed to the more targeted acetylation directed by promoter proximally bound (or RNA polymerase-associated) HAT activator complexes. Alternatively, it remains possible that widespread acetylation along HS1234-linked templates results from the episomal context of these genes; transcription from genes maintained on episomes may be subject to controls slightly different from those within a highly structured chromosome environment.

HS1234-mediated histone hyperacetylation.

The mechanism by which the HS1234 fragment mediates increased histone acetylation on linked genes is unknown. Recent findings that some transcriptional cofactors such as GCN5, CBP/p300, and PCAF possess HAT activity suggest that enhancer regions such as HS1234 could specifically recruit factors to achieve this modification. Indeed, the direct recruitment of CBP or PCAF has been found to stimulate transcriptional initiation and elongation along a reporter construct in transient assays (24). Although the physiological targets of the HAT activity of these molecules is unclear, in vitro assays indicate that H3 and H4 can be acetylated by PCAF and p300 (20). The recruitment of one or more HAT-containing factors to the HS1234 fragment might induce local histone acetylation that could be propagated throughout the template by H1 displacement and subsequent binding of additional factors to newly accessible sites. Alternatively, the increase in acetylation along HS1234 templates may be mediated through a less-direct mechanism. For example, proteins bound to the HS1234 fragment may transport the episome to a particular nuclear compartment rich in acetyltransferases. The nuclear matrix has been postulated as a site for active transcription and histone acetyltransferase, and deacetylase activities have been purified from matrix preparations (19). Clearly, further studies are required to determine the specific mechanisms by which the HS1234 enhancer mediates histone acetylation as well as activates the P1 promoter of linked c-myc genes.