The coactivators CBP/p300 and the histone chaperone NAP1 promote transcription-independent nucleosome eviction at the HTLV-1 promoter

Abstract

The human T cell leukemia virus type 1 (HTLV-1) is the causative agent of adult T cell leukemia/lymphoma. The multifunctional virally encoded oncoprotein Tax is responsible for malignant transformation and potent activation of HTLV-1 transcription. Tax, in complex with phosphorylated cAMP response element binding protein (pCREB), strongly recruits the cellular coactivators CREB binding protein (CBP)/p300 to the viral promoter concomitant with transcriptional activation. Although the mechanism of activator/coactivator-mediated transcriptional activation is poorly understood, the recruitment of CBP/p300 by regulatory factors appears to function, in part, by promoting changes in chromatin architecture that are permissive to transcriptional activation. Here, we show that CBP/p300 recruitment promotes histone acetylation and eviction of the histone octamer from the chromatin-assembled HTLV-1 promoter template in vitro. Nucleosome disassembly is strictly acetyl-CoA dependent and is not linked to ATP utilization. We find that the histone chaperone, nucleosome assembly protein 1 (NAP1), cooperates with CBP/p300 in eviction of the acetylated histones from the chromatin template. These findings reveal a unique mechanism in which the DNA-bound Tax/pCREB complex recruits CBP/p300, and together with NAP1, the coactivators cooperate to dramatically reduce nucleosome occupancy at the viral promoter in an acetylation-dependent and transcription-independent fashion.

The relationship between chromatin dynamics and transcriptional activation in metazoans is poorly understood. As a potent activator of transcription, the human T cell leukemia virus type 1 (HTLV-1)-encoded Tax protein serves as an outstanding model for studies on transcriptional regulation in a chromatin context in higher eukaryotes (1–5). Tax and Ser¹³³-phosphorylated cAMP response element binding protein (pCREB) together bind the HTLV-1 promoter at three conserved cAMP response elements (CREs) called viral CREs (vCREs). CREB binds to the vCRE octanucleotide core, whereas Tax interacts with CREB and the adjacent minor groove DNA (6, 7). Although pCREB alone inefficiently recruits the cellular coactivators CREB binding protein (CBP)/p300, Tax and pCREB together strongly recruit both CBP and p300, resulting in high-level viral transcription (8).

CBP and p300 are highly homologous coactivator proteins that regulate essentially all known pathways of gene expression in multicellular organisms via interaction with DNA-bound transcriptional regulatory proteins (9, 10). CBP/p300 stimulate transcription, in part, through acetylation of both histone and nonhistone substrates. Histone hyperacetylation is strongly associated with actively transcribed genes (11). Therefore, recruitment of CBP/p300 by transcription factors appears to promote changes in chromatin architecture that are permissive to transcriptional activation. We recently demonstrated that Tax binding to the chromosomally integrated HTLV-1 promoter correlates with p300 recruitment and strong transcriptional activation (12). Unexpectedly, we detected a decrease in nucleosome occupancy within both the HTLV-1 promoter and transcribed region coincident with p300 recruitment in vivo. Although nucleosome depletion has been widely characterized in yeast (13), very little is known about the formation of nucleosome-free regions at transcriptionally active genes in higher eukaryotes.

In this study, we find that Tax/pCREB recruitment of both CBP and p300 to the HTLV-1 promoter promotes nucleosome eviction in vitro. Disassembly of the octamer from the promoter DNA is independent of ATP utilization and transcription, but requires the acetyltransferase activity of CBP/p300, which leads to hyperacetylation of the four core histones. We further find that the histone chaperone NAP1 facilitates the acetylation-dependent eviction of nucleosome from promoter DNA. These observations are notable as only a very few examples of nucleosome eviction have been reported in higher eukaryotes and the mechanism is unknown.

Results

To explore the molecular basis for the observed in vivo decrease in nucleosome binding at the HTLV-1 promoter, we performed a biochemical analysis of nucleosomes assembled on the promoter after the binding of Tax and pCREB. A biotinylated 643-bp promoter fragment carrying the full HTLV-1 promoter linked to a G-less cassette was immobilized on magnetic streptavidin-agarose beads. The bound fragment was assembled into chromatin by using the recombinant Drosophila assembly proteins Acf1/ISWI, nucleosome assembly protein 1 (NAP1), and purified core histones (2, 14). Chromatin assembly was verified by micrococcal nuclease analysis of the immobilized template (Fig. 1A). This assay revealed that three to four nucleosomes occupy the 643-bp promoter fragment, consistent with the nucleosomal repeat interval in the absence of histone H1. The immobilized nucleosome-assembled promoter was preincubated with highly purified, recombinant pCREB and Tax (Fig. 1B) followed by incubation with CEM nuclear extract. The templates were extensively washed. Bound transcriptional regulatory proteins were analyzed by Western blot and bound histones were analyzed by Coomassie staining. As expected, p300 from the nuclear extract was efficiently recruited to the promoter by the Tax/pCREB complex (8) (Fig. 1C Top). Acetyl-CoA and ATP were added to individual binding reactions to assess their potential role in modifying chromatin architecture. Unexpectedly, addition of acetyl-CoA, but not ATP, correlated with significant nucleosome loss from the promoter template (Fig. 1C, lanes 5 and 6). Quantification of bound histones revealed that the majority of nucleosomes were displaced from the promoter template in the presence of acetyl-CoA. In parallel, we added nucleoside triphosphates to a small portion of each binding reaction and performed in vitro transcription assays. We found that the presence of acetyl-CoA was required for both transcription-independent nucleosome eviction from the promoter template, and strong transcriptional activation (Fig. 1C Bottom). Together, these data corroborate our previous in vivo findings (12) and demonstrate that nucleosome octamers are displaced from the HTLV-1 promoter in a transcription-independent manner.

Fig. 1.

The Tax and pCREB complex promotes nucleosome eviction from the HTLV-1 promoter in an acetyl-CoA dependent manner. (A) Microccocal nuclease (MNase) digestion of the immobilized, chromatin-assembled 643-bp HTLV-1 promoter fragment. Native Drosophila core histones were assembled onto the promoter template by using recombinant dAcf-1/ISWI and dNAP1, as described (14). Time of MNase treatment is indicated. (B) Purified, recombinant Ser¹³³-pCREB and Tax visualized by Coomassie staining. (C) DNA pull-down and in vitro transcription assays were performed by using chromatin templates, CEM nuclear extract, and each of the indicated components. Lanes showing input protein (50% nuclear extract and 10% Tax and pCREB), relative to starting material, are demarcated by a dashed line. A portion of each binding reaction (10%) was fractionated on a 4–20% gradient SDS/PAGE. (Top) Template-bound proteins were detected by immunoblot using antibodies against p300, Tax, and pCREB. (Middle) A second portion of each binding reaction (85%) was separated on 18% SDS/PAGE and histones were visualized by Coomassie staining. Streptavidin (from the magnetic Dynabeads) comigrates with histone H4. (Bottom) To the remaining 5% of each binding reaction, nucleoside triphosphates were added to initiate RNA synthesis (see Materials and Methods), and the transcripts were analyzed by denaturing PAGE and autoradiography. The transcript and recovery standard (RS) are indicated.

The observation that nucleosome displacement required acetyl-CoA and correlated with p300 recruitment led us to examine whether the intrinsic CBP/p300 acetyltransferase activity played a role in the eviction reaction. The immobilized template assays were performed as shown in Fig. 1C; however, the nuclear extract was preincubated with Lys-CoA or curcumin, selective inhibitors of CBP/p300 acetyltransferase activity (15, 16). Each inhibitor blocks the acetyltransferase activity of CBP/p300 by a distinct mechanism. Both Lys-CoA and curcumin prevented nucleosome eviction from the HTLV-1 promoter template (Fig. 2A). These data indicate that the acetyltransferase activity of CBP/p300 is required for displacement of the octamer from the template DNA. Based on these results, we next tested whether purified recombinant p300 (Fig. 2B) could substitute for nuclear extract in the nucleosome eviction reaction. Fig. 2C shows that p300 supported nucleosome loss from the HTLV-1 promoter template comparable to nuclear extract, suggesting that coactivators present in the nuclear extract play a prominent role in the disassembly of nucleosomes (Fig. 2C, lanes 3 and 4). Histone H3 was detected in the unbound fraction, indicating that intact histones were displaced from the DNA and not lost through degradation (Fig. 2C). Purified recombinant CBP behaved indistinguishably from p300 in the nucleosome eviction reaction (Fig. 2D). These data indicate that the intrinsic acetyltransferase activity of both CBP and p300 facilitate nucleosome disassembly from the HTLV-1 promoter. Facilitation of nucleosome displacement has not previously been associated with these important coactivators.

Fig. 2.

The acetyltransferase activity of CBP/p300 is required for nucleosome disassembly from the HTLV-1 promoter. (A) The addition of the CBP/p300-specific acetyltransferase inhibitors Lys-CoA and curcumin to CEM nuclear extract blocks nucleosome eviction. DNA pull-down reactions were performed as described in Fig. 1C. (B) Purified, recombinant p300 visualized by Coomassie staining. (C) DNA pull-down assays were performed with chromatin templates and each of the indicated components. (Top) Template bound proteins (10% of the reaction) were detected by immunoblot with their respective antibodies. (Middle) Histone H3 released into the supernatant of the binding reaction was detected by anti-H3 antibody (40% of supernatant). Lanes showing input protein (10% recombinant p300, Tax, pCREB; 25% H3) are demarcated by a dashed line. (Lower) The remaining 90% of each binding reaction was analyzed as described in Fig. 1C to visualize template bound histones. Solid line denotes where the gel was cropped to move relevant lanes adjacent to one another. (D) DNA pull-down assay was performed as in C, except recombinant, purified CBP was used in place of p300.

The Drosophila chromatin assembly proteins Acf1/ISWI and NAP1 were used to assemble nucleosomes onto the HTLV-1 promoter template in the experiments shown in Figs. 1 and 2. The histone chaperone NAP1 has previously been shown to play a role in nucleosome assembly, exchange, and disassembly of the H2A/H2B dimer (17, 18). Furthermore, NAP1 functions in an ATP-independent manner. We therefore considered whether NAP1 plays a role in nucleosome eviction from the HTLV-1 promoter. To explore this possibility, we assembled chromatin templates in the absence of assembly proteins by salt deposition (19). This method produces chromatin that is indistinguishable from that formed by using the assembly factors, as measured by micrococcal nuclease assays and response to Tax/pCREB activation in an in vitro transcription assay (Figs. 3 A and B). Analysis of these templates by DNA pull-down assay, however, revealed that the presence of p300 and acetyl-CoA was no longer sufficient for nucleosome eviction (Fig. 3D, lane 2). However, the addition of highly purified NAP1 (Fig. 3C) to the binding reactions, together with purified Tax, pCREB and p300, restored acetyl-CoA-dependent octamer disassembly from the template (Fig. 3D, lane 4). These data reveal that NAP1 is required in the nucleosome disassembly reaction. Finally, we systematically examined the contribution of individual components of the binding reaction to determine their individual contribution to nucleosome eviction. The promoter template was incubated with various combinations of the proteins in the absence or presence of acetyl-CoA. To provide an unbiased view of the binding of each protein in the reaction, the washed templates were analyzed by Coomassie staining. Fig. 3E shows that the Tax/pCREB complex, p300, NAP1, and acetyl-CoA were each required for disassembly of nucleosomes from the HTLV-1 promoter in vitro (Fig. 3E). Interestingly, the binding of NAP1 to the template remained relatively unchanged in the absence or presence of acetyl-CoA (Fig. 3C, compare lanes 1 and 2). Furthermore, we did not observe enrichment of NAP1 in the supernatant fractions containing the evicted histones (data not shown). Therefore, the fate of NAP1 in the eviction reaction is unclear. NAP1 has previously been shown to physically interact with p300 and has been implicated in p300-mediated transcriptional activation (20, 21). The histone acceptor in the supernatant is likely the DNA, which was added to all reactions as a nonspecific competitor (see Materials and Methods).

Fig. 3.

NAP1 is required for nucleosome eviction from the HTLV-1 promoter. (A) Micrococcal nuclease digestion assay of chromatin assembled on the promoter template by salt deposition (19). Time of MNase treatment is indicated. (B) An in vitro transcription assay was performed to confirm the quality of the chromatin assembled by salt deposition as described (8). Transcription reactions were performed in the presence of acetyl-CoA. (C) Recombinant, purified Drosophila NAP1 protein visualized by Coomassie staining. (D) DNA pull-down assays were performed with chromatin templates assembled by salt deposition in the presence of the indicated components. p300 and histone protein binding was assessed as described above. (E) DNA pull-down assays were performed with sequential omission of each regulatory protein (Tax/pCREB, p300, NAP). Template bound proteins were separated on 18% SDS/PAGE and visualized by Coomassie staining.

We define a critical role for both p300 histone acetyltransferase activity in the disassembly of nucleosomes from the promoter template. We were therefore interested in identifying the relevant targets of acetylation in the eviction reaction. Because p300 has previously been shown to undergo autoacetylation, we first tested whether p300 acetylation was sufficient for nucleosome eviction. We acetylated purified p300 (and removed unincorporated acetyl-CoA) before incubation with the chromatin template, Tax/pCREB, and NAP1. Fig. 4A shows that preacetylated p300 was insufficient for nucleosome disassembly, and that histone eviction required the addition of exogenous acetyl-CoA (lanes 3 and 4). These data point to another (or additional) p300 acetylation target that is functionally relevant in the disassembly reaction. To identify this target, we performed DNA pull-down reactions in the presence of ¹⁴C-labeled acetyl-CoA. In this experiment, we analyzed both template-associated (bound) histones and the histones evicted into the supernatant (unbound). Both fractions were visualized by Coomassie staining and autoradiography. Fig. 4B (lanes 1–4) shows that the majority of the four core histones were evicted into the supernatant in the presence of [¹⁴C] acetyl-CoA, and that these evicted histones were highly acetylated (lanes 5–8). p300 was the only other acetylated protein detected in the assay (data not shown). Mass spectrometry revealed the acquisition of four acetyl groups on histone H3 in the DNA pull-down reaction in the presence of acetyl-CoA (data not shown). These data indicate that the four core histones are the major targets of p300 acetylation and strongly support a role for histone hyperacetylation in NAP1-dependent nucleosome eviction.

Fig. 4.

The histone tails are the relevant sites of p300 acetylation. (A) Acetylated p300 is not sufficient for nucleosome disassembly from the HTLV-1 promoter. Purified p300 was preacetylated (via autoacetylation), and excess acetyl-CoA was removed by gel filtration. Autoacetylated p300 was then used in the DNA pull-down assays, in parallel with unacetylatead p300, using purified components and chromatin templates assembled by salt deposition. Nucleosome eviction was analyzed as described above. (B) The evicted histones are highly acetylated. ¹⁴C-labeled acetyl-CoA was used in place of unlabeled acetyl-CoA to visualize the targets of p300 acetyltransferase activity. The bound and unbound proteins were separated by 18% SDS/PAGE, and histones were visualized by Coomassie staining. The gel was dried, and histone acetylation was visualized by PhosphorImager analysis. Histone H2B appears to be preferentially evicted; however, acetylation shifts the protein such that it migrates coincident with H2A.

Discussion

Understanding the interplay between DNA-binding transcription factors and coactivators in nucleosome dynamics is essential to furthering our knowledge of transcriptional activation in higher eukaryotes. Although CBP/p300 are well established as crucial players in metazoan transcription, relatively little is known regarding their precise role in chromatin dynamics and transcriptional activation. The HTLV-encoded oncoprotein Tax serves as an outstanding model activator for studies on eukaryotic transcription. Paradoxically, Tax confers CBP/p300 recruitment and transcriptional activation properties to pCREB, as pCREB alone only weakly binds the full-length coactivators (8). Consistent with this observation, Tax/pCREB recruitment of CBP/p300 to the viral promoter strongly correlates with transcriptional activation (8).

In the studies described herein, we find that Tax/pCREB recruitment of CBP/p300 to the HTLV-1 promoter is required for eviction of nucleosomes from the template DNA. Importantly, nucleosome disassembly depends on the acetyltransferase activity of CBP/p300, and the four core histones serve as the major targets of p300 acetylation. Disassembly of the octamer from the promoter DNA is independent of ATP utilization and transcriptional activity. This work defines additional roles for both CBP/p300 and the histone chaperone NAP1 in the facilitation of acetylation-dependent nucleosome eviction from promoter DNA. These observations are particularly notable as only a very few examples of nucleosome eviction have been reported in higher eukaryotes (12, 22–24).

A model depicting nucleosome eviction at the HTLV-1 promoter is shown in Fig. 5. It is currently not known whether histone loss from the HTLV-1 promoter results from a process of nucleosome sliding or nucleosome disassembly. However, the requirement for NAP1 in our system strongly supports a mechanism of acetylation-dependent nucleosome disassembly analogous to that observed with the histone chaperone antisilencing factor 1 (Asf1) in yeast (25, 26). This mechanism has likely evolved to clear promoter nucleosomes, exposing the DNA to enable binding of the general transcription machinery and subsequent activation of transcription.

Fig. 5.

Schematic showing Tax/pCREB recruitment of p300 to the HTLV-1 promoter and nucleosome eviction in the presence of NAP1. For simplicity, only one of the three vCREs is shown. We depict NAP physically associated with p300 based on published studies (20, 21) and data not shown.

Materials and Methods

Protein Expression and Purification.

Recombinant CREB₃₂₇ (27) and Tax-His₆ (28) were purified as described. CREB was phosphorylated at Ser-133 using the catalytic subunit of PKA. Full-length His₆-tagged p300, FLAG-tagged CBP, and His₆-tagged Drosophila NAP (dNAP1) were expressed from recombinant baculovirus in Sf9 cells and purified as described (29, 30). FLAG-tagged Acf-1 and ISWI were coexpressed from baculovirus in Sf9 cells and purified as described (31). Nuclear extract from CEM cells, an HTLV-1 negative human T cell line, was prepared as described (32). To autoacetylate p300, the purified protein was incubated with acetyl-CoA (100 μM) in a buffer containing 50 mM Tris·HCl (pH 7.9), 20% (vol/vol) glycerol, 12.5 mM MgCl₂, 100 mM KCl, and 2 mM DTT at 30°C for 60 min. Unincorporated acetyl-CoA was removed by passing the reaction mixture through a Bio-Spin gel filtration column. p300 autoacetylation was confirmed in parallel reactions using ¹⁴C-labeled acetyl CoA (data not shown).

Chromatin Assembly.

A 643-bp fragment carrying the full HTLV-1 promoter linked to a G-less cassette was amplified by PCR, incorporating a biotin group at the upstream end of the fragment, and immobilized on streptavidin-coated magnetic beads (8). The template was assembled into chromatin by using Drosophila core histones at a histone/DNA ratio of 0.6:1 (wt/wt). Chromatin assembly was performed by using purified Acf-1/ISWI and dNAP1 as described (31, 33). For salt deposition, the immobilized template DNA and histones were mixed together in a reaction containing 1 M NaCl, 10 mM Tris·HCl (pH 8.0), and 1 mM EDTA. Nucleosome deposition was performed as described (19) by stepwise dilution of the NaCl in the sample with 10 mM Tris·HCl (pH 8.0) containing 1 mM EDTA to a final NaCl concentration of 0.1 M. Each dilution step was incubated for 45 min at 4°C with shaking. The assembled chromatin was stored in a buffer containing 0.1 M NaCl, 0.1% Nonidet P-40, 10 mM Tris-HCl (pH 7.5), 1 mM EDTA, 5 mM DTT, and 20% glycerol at 4°C. Micrococcal nuclease assays were performed as described (34).

DNA Pull-Down and in Vitro Transcription Assays.

Immobilized template assays were performed as described (8). Briefly, 1.5 pmol of the chromatin-assembled template was preincubated with Tax and pCREB (12 pmol each, 2-fold molar excess) for 15 min at 30°C. The binding reactions (50 μl final volume) were supplemented with purified, recombinant p300 or CBP (6 pmol each), dNAP (133 pmol), ATP (100 μM), and/or acetyl-CoA (100 μM), and 40 ng/μl of poly(dA-dT)-poly(dA-dT), as indicated in each experiment. Reactions were then incubated at 30°C for 40 min. In binding reactions supplemented with a human T cell nuclear extract, the samples were incubated for 60 min at 4°C. After incubation, the reactions were magnetically isolated, washed twice, resuspended in 8 μl of SDS-sample dyes, subjected to SDS/PAGE, and analyzed by either immunoblot or Coomassie staining, as indicated in each experiment. Nuclear extract was preincubated with either curcumin (300 μM) (15) or Lys-CoA (100 μM) (16) at 30°C for 5 min to inhibit the CBP/p300 acetyltransferase activity present in the nuclear extract. To determine the target/s of p300 acetylation, 100 μM acetyl-CoA (51 mCi/mmol) was used in place of unlabeled acetyl-CoA in the binding reactions. Supernatants from the binding reactions were removed, template-bound proteins were washed, and both the bound and supernatant (unbound) fractions were analyzed in parallel by Coomassie staining followed by PhosphorImager analysis. In vitro transcription reactions were performed as described (8). Briefly, the promoter template carried a G-less cassette for transcription analysis. In the experiment shown in Fig. 1C, a portion of each binding reaction was assayed after the addition of 250 μM ATP, 250 μM CTP, 12 μM UTP, and 0.8 μM [α-³²P] UTP (3,000 Ci/mmol). Molecular weight markers (radiolabeled HpaII-digested pBR322) were used to estimate the size of the RNA products. A labeled 622-bp DNA fragment was added to each reaction mixture as a recovery standard. The experiment shown in Fig. 3B was performed as described for Fig. 1C, except 100 μM acetyl CoA was added to each reaction mixture.

Immunoblot.

Immunoblot assays were performed as described (9) with the following antibodies from Santa Cruz Biotechnology: anti-phospho Ser¹³³CREB (sc-7978), anti-p300 (sc-584), and anti-CBP (sc-583). Alexa Fluor IR700 and IR800 goat anti-mouse and goat anti-rabbit secondary antibodies were purchased from Molecular Probes. The monoclonal Tax antibody (Hybridoma 168B17-46-92) was obtained from the National Institutes of Health AIDS Research and Reference Reagent Program.