Nucleosome eviction and activated transcription require p300 acetylation of histone H3 lysine 14

Abstract

Histone posttranslational modifications and chromatin dynamics are inextricably linked to eukaryotic gene expression. Among the many modifications that have been characterized, histone tail acetylation is most strongly correlated with transcriptional activation. In Metazoa, promoters of transcriptionally active genes are generally devoid of physically repressive nucleosomes, consistent with the contemporaneous binding of the large RNA polymerase II transcription machinery. The histone acetyltransferase p300 is also detected at active gene promoters, flanked by regions of histone hyperacetylation. Although the correlation between histone tail acetylation and gene activation is firmly established, the mechanisms by which acetylation facilitates this fundamental biological process remain poorly understood. To explore the role of acetylation in nucleosome dynamics, we utilized an immobilized template carrying a natural promoter reconstituted with various combinations of wild-type and mutant histones. We find that the histone H3 N-terminal tail is indispensable for activator, p300, and acetyl-CoA-dependent nucleosome eviction mediated by the histone chaperone Nap1. Significantly, we identify H3 lysine 14 as the essential p300 acetylation substrate required for dissociation of the histone octamer from the promoter DNA. Together, a total of 11 unique mutant octamer sets corroborated these observations and revealed a striking correlation between nucleosome eviction and strong activator and acetyl-CoA-dependent transcriptional activation. These novel findings uncover an exclusive role for H3 lysine 14 acetylation in facilitating the ATP-independent and transcription-independent disassembly of promoter nucleosomes by Nap1. Furthermore, these studies directly couple nucleosome disassembly with strong, activator-dependent transcription.

Chromatin facilitates the extraordinary compaction of genomic DNA in all eukaryotes. Nucleosomes, the basic repeating unit of chromatin, are composed of two copies each of the core histone proteins H2A, H2B, H3, and H4 with 147 bp of DNA wrapped around the histone octamer (1, 2). Nucleosomes present an intrinsic paradox for the cell: They must form stable contacts with the DNA to compact the genetic material, and they must allow local disassembly to expose the underlying DNA for the chromosomal transactions required for life. Eukaryotes have evolved multiple strategies to negotiate the inherently repressive chromatin to access their genetic material, including posttranslational modifications of the nucleosomal histones (3).

Histone modifications occur primarily on the N-terminal tails that extend beyond the nucleosome core particle. The first identified and most intensively studied histone modification is acetylation (4, 5). Gene-associated histone H3 and H4 tail acetylation is one of the strongest predictors of transcriptional activity in eukaryotes (5, 6). In native chromatin, acetylation of specific lysine residues on the H3 and H4 N-terminal tails is nonrandom and highly ordered (7), a finding in agreement with the patterns of acetylation associated with multiple histone acetyltransferases (HATs) throughout eukaryotes (8–11). Histone tail acetylation has multiple effects on chromatin structure and fluidity. These include H4 acetylation-dependent relaxation of the chromatin fiber and recruitment of ATP-dependent chromatin-remodeling complexes via acetyl-lysine binding bromodomains (12–17). However, the mechanism by which histone tail acetylation elicits chromatin reconfiguration and coupled transcriptional activation is unknown (6).

In recent years, numerous high profile mapping studies of nucleosomes and their modifications identified nucleosome-free regions (NFRs) at the promoters of transcriptionally active genes (18–24). The core promoter must be devoid of nucleosomes to enable binding of the RNA polymerase holoenzyme complex. Several studies have correlated the formation of promoter NFRs with the acetylation of specific lysines on the H3 tail (20, 22, 23, 25). Furthermore, the binding of the essential coactivator and HAT, p300, correlates with NFRs and adjacent regions of H3 hyperacetylation (20, 26). These observations suggest an underlying functional connection between histone N-terminal tail acetylation and the disassembly of promoter-associated nucleosomes. However, the link between histone tail acetylation and NFR formation remains obscure.

Studies in our laboratory recently uncovered a biochemical pathway of activator and p300-dependent promoter nucleosome disassembly mediated by the histone chaperone Nap1 (27). Nucleosome eviction from our model retroviral promoter in vitro recapitulated activator-dependent nucleosome loss from this promoter in vivo (28). Nucleosome disassembly in vitro required purified activators, catalytically active p300, Nap1, and acetyl CoA (Ac-CoA). The four core histones served as the presumed p300 acetylation substrates. Notably, promoter nucleosome disassembly was independent of ATP, chromatin-remodeling factors, and transcription per se, but required acceptor DNA and Nap1 (27). These observations highlight a unique function for Nap1 in acetylation-dependent octamer disassembly and expand our understanding of the multiple mechanisms utilized by eukaryotes to mobilize repressive nucleosomes during gene activation.

In the present study, we explore the role of histone acetylation by p300 during nucleosome disassembly and define the interconnectivities among histone acetylation, promoter nucleosome disassembly, and transcriptional activation. We utilized an in vitro immobilized template assay to investigate activator and Ac-CoA-dependent disassembly of nucleosomes from a native promoter. We compared chromatin templates composed of a mixture of wild-type histones and histones carrying lysine to arginine (K → R) N-terminal tail mutations. This analysis identified the histone H3 tail as the essential, functionally relevant acetylation target in nucleosome eviction by Nap1. Chromatin templates bearing these point mutations were also tested in parallel in vitro transcription assays. We observed a significant correlation between activator/Ac-CoA-dependent nucleosome eviction and activator/Ac-CoA-dependent transcriptional activation, establishing a critical functional link between H3 tail acetylation, NFR formation, and potent activation of transcription. We next introduced specific K → R point mutations in the H3 tail and tested chromatin bearing these mutations in both nucleosome eviction and transcriptional activation. Lysine 14 (H3 K14) emerged as the essential functional target of p300 acetylation. Remarkably, a single point mutation throughout the entire nucleosome, H3 K14 → R, was sufficient to block acetylation-dependent nucleosome disassembly. Likewise, chromatin reconstituted with the reciprocal octamer (WT H3 K14, with K → R tail mutations throughout the remainder of the nucleosome), supported acetylation-dependent nucleosome eviction comparable to that of wild-type chromatin. These data identify H3 K14 acetylation as the pivotal event in Nap1-mediated promoter nucleosome disassembly. Together, these unique findings provide a molecular mechanism linking histone acetylation with gene activation and thus significantly advance our understanding of the biological outcome of this critical epigenetic modification.

Results

p300 Targets the Histone Amino-Terminal Tails for Nucleosome Disassembly.

To identify the specific histone targets of acetylation by p300, we used a chromatin-based immobilized template assay comprised of the model human T-cell leukemia virus (HTLV-1) promoter linked to a G-less cassette. This system has been previously described (27) and is illustrated in Fig. S1A. The natural HTLV-1 promoter carries three highly conserved 21-bp enhancer elements, called viral CREs (vCREs), located at -100, -200, and -250 bp upstream of the transcriptional start site (see ref. 29). The vCREs serve as binding sites for the virally encoded transcriptional activator protein Tax, which binds together with the phosphorylated form of the cellular transcription factor cAMP response element binding protein, or CREB (pCREB). Together, the Tax/pCREB complex recruits p300 to promote strong HTLV-1 transcriptional activation in vivo and in vitro (29).

To begin characterization of acetylation-dependent nucleosome disassembly, we first performed a HAT assay to examine the pattern of p300 acetylation on both the template-bound histones and the histones liberated into the supernatant fraction (evicted). The 588-bp HTLV-1 promoter fragment was assembled into chromatin using highly purified, recombinant Xenopus histones by the method of salt deposition (Fig. S1B). Highly purified Tax, pCREB, p300, and human Nap1L1 (hNap1) (see Fig. S1B) were preincubated with the chromatin template, and nucleosome eviction was assayed in the presence of [¹⁴C] Ac-CoA. Fig. 1A shows that all four core histones were released from the immobilized promoter template in the presence of [¹⁴C] Ac-CoA and that the evicted histones were highly acetylated. This observation is further corroborated in an eviction time course shown in Fig. S2A.

Fig. 1.

Histone amino-terminal tail acetylation is required for nucleosome disassembly. (A) HAT assay reveals that H3 and H4 are highly acetylated in the evicted (supernatant) fraction. The nucleosome eviction assay was performed as shown schematically in Fig. S1A. Briefly, the promoter template was assembled with wild-type core histones, and nucleosome eviction was analyzed following the addition of highly purified Tax, pCREB, p300, hNap1, and acceptor DNA; in the absence or presence of [¹⁴C] Ac-CoA. The purified proteins used in these assays are shown in Fig. S1B. The same gel was analyzed by Coomassie staining and PhosphorImage analysis. Molecular weight markers (M) are indicated. (B) Schematic of the four core histone N-terminal tails showing the positions of the lysine to arginine (K → R) point mutations (43). (C) Lysines in the N-terminal core histone tails are required for Ac-CoA-dependent nucleosome eviction. Chromatin was assembled with either wild-type or K → R histone octamers. Nucleosome disassembly was assayed as described in A and Fig. S1A (using nonradiolabeled Ac-CoA). Dashed lines demarcate the lanes. Asterisk denotes contaminating streptavidin (from the Dynabeads®) that comigrates with H4.

Because the histone N-terminal tails are the primary targets of p300 (9), we next sought to determine if histone tail acetylation was required for nucleosome eviction. We prepared histone octamers carrying K → R point mutations at the known in vivo targets of acetylation on all four histones (7) (Fig. 1B). The K → R mutations retain the positive charge at each position, but are no longer substrates for acetylation. Octamers composed of either wild-type histones, or histones carrying the K → R point mutations, were assembled in parallel into chromatin on our model promoter template, incubated with purified Tax, pCREB, p300, and Nap1, and assayed for nucleosome eviction in the absence or presence of Ac-CoA. Fig. 1C shows that chromatin carrying the K → R tail mutations does not support Ac-CoA-dependent nucleosome disassembly (lanes 3 and 4), indicating that the histone N-terminal tails serve as the functionally relevant p300 acetylation targets. From the data presented in Fig. 1, we conclude that p300 acetylation of lysine residues present in one or more of the four core histone tails is required for Nap1-mediated disassembly of promoter-associated nucleosomes.

H3 N-terminal Tail Lysines Are Required for Nucleosome Eviction and Transcriptional Activation.

To identify the specific core histone tail (or tails) that functionally support nucleosome disassembly, we prepared four unique sets of histone octamers (mixed octamers) using combinations of wild-type histones and histones carrying the N-terminal K → R mutations (Fig. S1B). Each of the mixed octamer sets were assembled into chromatin on the immobilized HTLV-1 promoter and assayed for acetylation-dependent nucleosome disassembly using purified components, exactly as described above. Remarkably, these experiments revealed that the lysines present on the N-terminal tails of H2A, H2B, and H4 were dispensable and that wild-type histone H3 alone was sufficient for acetylation-dependent nucleosome disassembly (Fig. 2A). An eviction HAT assay performed with these mixed octamers is shown in Fig. S2B. To further confirm that the H3 N-terminal tail served as the functionally relevant p300 acetylation target, we compared nucleosome eviction on chromatin composed of K → R H3, assembled with wild-type H2A, H2B, H4, in parallel with the reciprocal chromatin, wild-type H3, assembled with K → R H2A, H2B, H4. Fig. 2B establishes that acetylation-dependent nucleosome eviction exclusively requires lysine residues present on the H3 N-terminal tail.

Fig. 2.

The H3 amino-terminal tail is required for nucleosome disassembly and transcriptional activation. (A) Mixed octamers were prepared using combinations of wild-type histones and histones carrying K → R tail mutations (see Fig. 1B). The mixed octamers were assembled into chromatin and tested for their ability to support activator/Ac-CoA-dependent nucleosome disassembly, exactly as described in the Fig. 1A legend (and Fig. S1A). (B) Direct comparison of chromatin templates carrying the wild-type H3 tail assembled together with K → R tail mutations in H2A, H2B, and H4, with the reciprocal combination of core histones. (C and D) Using the chromatin templates shown in A and B, in vitro transcription reactions were performed using T-cell nuclear extract in the absence or presence of activators (Tax/pCREB) and/or Ac-CoA. The transcript (RNA) and recovery standard (RS) are indicated. The chromatin samples in 2C (lanes 1–9) and 2D are slightly underassembled, enabling visualization of “basal” transcription. A more detailed description of the in vitro transcription assay is provided in the legend to Fig. S1A.

The disassembly of nucleosome octamers from our model promoter fragment must, by definition, reduce nucleosome density on the template. Because the fragment is 588 bp in length and assembles ∼3 nucleosomes (27), it is likely that the core promoter is nucleosome-free on a significant fraction of the templates. We therefore hypothesized that chromatin templates permissive to nucleosome eviction should serve as excellent substrates for preinitiation complex formation and activated transcription. Similarly, templates reconstituted with chromatin carrying K → R point mutations that block nucleosome eviction should also block activated transcription. As such, the chromatin templates used in Fig. 2 A and B provided us with powerful tools to directly test whether acetylation-dependent nucleosome disassembly is a prerequisite for transcriptional activation. Of note, we previously established that transcription from our chromatin-assembled model HTLV-1 promoter is strongly dependent upon Tax/CREB, p300, and Ac-CoA (27, 30, 31). To test transcriptional activation from the same wild-type and K → R mutant chromatin templates described in Fig. 2 A and B above, we performed in vitro transcription assays in the absence or presence of activators (Tax/pCREB) and/or Ac-CoA. The experiments were carried out similar to the protocol shown in Fig. S1A, except that following Tax/pCREB binding, nuclear extract was incubated with the templates to enable preinitiation complex formation. The T-cell nuclear extract supplies the RNA polymerase holoenzyme complex, p300, and nucleosome disassembly activity indistinguishable from Nap1 (27). Remarkably, we observe a precise correlation between activator/Ac-CoA-dependent nucleosome eviction and strong activator/Ac-CoA-dependent transcriptional activation (Fig. 2). Specifically, the three octamers carrying wild-type H3 supported both eviction and transcription, whereas the three octamers carrying K → R mutations in the H3 tail failed to support eviction and transcription (Fig. 2 C and D; summarized in Table S1). Together, the findings presented in Fig. 2 and Table S1 demonstrate that the histone H3 tail alone is sufficient for acetylation-dependent nucleosome disassembly and transcriptional activation and implicate one (or more) of the H3 tail lysine residues as the functionally relevant p300 acetylation target (e.g., K9, K14, K18, and/or K23). Further, the precise correspondence between nucleosome eviction and transcriptional activation directly links these two fundamental biological processes (Table S1).

Lysine 14 on the H3 N-Terminal Tail Is Necessary and Sufficient for Nucleosome Eviction and Transcriptional Activation.

Based on the data presented above, we next focused on identifying the specific H3 N-terminal tail lysine(s) responsible for supporting acetylation-dependent nucleosome eviction. We first prepared two H3 tail mutants carrying wild-type K9 (WT K9) or wild-type K14 (WT K14), each against a background of K → R mutations at the other three positions in the H3 tail (Fig. 3A). Octamers composed of H3 tail mutants, together with wild-type core histones H2A, H2B, and H4 (Fig. S1B), were assembled into chromatin and tested for their ability to support activator-dependent and Ac-CoA-dependent nucleosome disassembly and transcriptional activation, as described in Fig. 2. As shown in Fig. 3 B and C, nucleosomes carrying wild-type H3 K9, with K → R mutations at positions 14, 18, and 23, failed to support nucleosome eviction or transcriptional activation. In marked contrast, nucleosomes carrying wild-type H3 K14, with K → R mutations at 9, 18, and 23, supported both eviction and transcription, comparable to that observed with wild-type chromatin. We verified K14 acetylation with this mutant by mass spectrometry. As a control, a HAT assay time course demonstrates that chromatin assembled with both of these mutants support p300 acetylation throughout the nucleosome (Fig. S2C). These results establish that acetylation of H3 K14 is sufficient for both Nap1-mediated nucleosome disassembly and activator-dependent in vitro transcription.

Fig. 3.

Point mutations in the H3 tail implicate K14, and not K9, as the functional target of acetylation in activator-dependent nucleosome disassembly and transcriptional activation. (A) Schematic of the H3 tail showing the K → R point mutations. (B and C) The histone H3 tail mutants were reconstituted with wild-type core histones H2A, H2B, and H4. The chromatin templates were assayed for nucleosome disassembly and transcriptional activation as described above.

To further establish the functional relevance of K14 acetylation, we introduced a single K → R point mutation at this position in the H3 tail and reconstituted chromatin templates with WT H2A, H2B, and H4 (Fig. 4A). Fig. 4B shows that this single K14 → R point mutation within the nucleosome blocks acetylation-dependent nucleosome eviction. We next prepared chromatin carrying K → R point mutations at every major N-terminal tail lysine residue in the nucleosome, except position H3 K14. Fig. 4C shows that wild-type H3 K14, present as the only major acetylation site, supports disassembly of the nucleosome octamer from the promoter DNA. These experiments establish that H3 K14 is both necessary and sufficient for activator, p300, and Ac-CoA-dependent promoter nucleosome disassembly by Nap1.

Fig. 4.

H3 K14 is necessary and sufficient for acetylation-dependent nucleosome disassembly. (A) Schematic of the H3 tail showing the K → R point mutations. Relevant wild-type lysines are shown in red. (B) Each of the indicated H3 tail mutant was reconstituted with wild-type H2A, H2B, and H4 and tested in activator/Ac-CoA-dependent nucleosome eviction assays. (C) The “WT K14” histone H3 tail mutant (K9, 18, 23 → R) was reconstituted with core histones carrying K → R mutations throughout the H2A, H2B, and H4 N-terminal tails (see Fig. 1B). Importantly, this chromatin (lanes 5 and 6) carries H3 K14 as the only major site of acetylation throughout the nucleosome. This chromatin was tested in parallel with H3 WT K14 reconstituted with wild type H2A, H2B, and H4 (from 4B, lanes 3 and 4) in activator/Ac-CoA-dependent nucleosome eviction.

Discussion

We provide compelling evidence that p300 acetylation of histone H3 lysine 14 is essential for activator-dependent promoter nucleosome disassembly by the histone chaperone Nap1. Histone H3 K14 is also required for robust activator and acetyl-CoA-dependent transcription in vitro. Remarkably, introduction of only a single arginine point mutation at position H3 K14 was sufficient to block promoter nucleosome disassembly, demonstrating that acetylation at any other lysine throughout the nucleosome cannot support eviction. In the reciprocal experiment, chromatin carrying wild-type H3 K14, with K → R tail mutations at every major acetylation target throughout the nucleosome, fully supported acetylation-dependent nucleosome eviction. These data convincingly demonstrate that H3 K14 is a highly specific, nonredundant acetylation substrate that serves an essential role in nucleosome disassembly. As summarized in Table S1, a total of 13 unique octamer sets were reconstituted into chromatin on our model promoter template and tested for their ability to support activator (Tax/pCREB) and Ac-CoA-dependent nucleosome disassembly by Nap1. Eight of these chromatin templates were tested in parallel in activator and Ac-CoA-dependent in vitro transcription assays. Together, these studies revealed a compelling correlation between acetylation-dependent nucleosome eviction and activator-dependent transcriptional activation and establish H3 K14 as the single functionally relevant p300 target required for Nap1-mediated nucleosome eviction. These previously undescribed findings provide evidence demonstrating a direct role for the histone chaperone Nap1 in the disassembly of the nucleosome octamer following a highly specific acetylation event (for review, see ref. 32). Further, these studies demonstrate a direct functional linkage between site-specific histone acetylation, promoter nucleosome disassembly by Nap1, and gene activation. A model incorporating these sequential steps in the conversion of a silent gene to a transcriptionally active gene is shown in Fig. 5.

Fig. 5.

Model showing the sequential events in acetylation-dependent nucleosome disassembly. p300 is recruited to the promoter via interaction with activators, whereupon it acetylates the histone tails of the promoter-associated nucleosomes. Following H3 K14 acetylation, Nap1, via interaction with p300 (37, 38) mediates disassembly of the nucleosome octamer from the promoter DNA, producing a nucleosome-free region (NFR) that is permissive to the assembly of the general transcription machinery.

Identifying an exclusive role for the H3 tail in acetylation-dependent nucleosome eviction was unexpected, as both the H3 and H4 tails are well-established acetylation targets in nucleosome dynamics and gene activation (5, 6, 32). The H4 tail has been shown to mediate internucleosomal interactions and stabilize higher-order chromatin structures (13, 33–35). Specifically, acetylation of H4 K16 facilitates relaxation of the chromatin fiber, creating a template more permissive to gene activation (13, 14, 33). The natural promoter fragment used in our assays accommodates up to three randomly positioned nucleosomes, and immobilization of this fragment on the streptavidin-coated magnetic bead enforces a physical distance between the templates. As such, the chromatin used in our model system is unlikely to undergo extensive intra- and/or interfiber interactions, and the requirement for H4 tail acetylation is therefore bypassed. Taken together, our findings reveal the functional uncoupling of these two critical, yet mechanistically distinct, histone acetylation pathways. Importantly, the segregation of these two pathways enabled identification of H3 K14 as the precise acetylation target required for disassembly of the histone octamer from the DNA.

Based on chromatin modification studies that span decades, two distinct, yet mutually compatible models have emerged to consider the outcome of histone acetylation on chromatin dynamics (for review, see ref. 32). In the first model, histone tail acetylation induces changes in the physical properties of the nucleosome that reduce inter- or intranucleosomal contacts, producing a more relaxed chromatin fiber. In the second model, acetylated lysines serve as recognition elements to mediate recruitment of chromatin modifying proteins via bromodomains. Our studies described herein are inconsistent with the first model, as the acetylation of a single lysine—at a very specific position within the nucleosome—largely negates a role for charge neutralization and/or reduced nucleosome stability. The specific identification of H3 K14, however, resonates with numerous previous studies that may provide clues to the mechanism of disassembly. Analysis of the acetylation patterns of native histones revealed that H3 K14 acetylation is very specific and ordered and is one of the lysine positions found to be highly acetylated in vivo (7). Consistent with these observations are the ubiquitous findings that H3 lysine 14 is the primary or major acetylation target of multiple HATs. These HATs are expressed throughout eukaryotes and include purified Gcn5, the Spt-Ada-Gcn5 acetyltransferase (SAGA), Ada2, and NuA3 complexes, CBP/p300-associated factor (PCAF), CREB binding protein (CBP), and p300 (8–11). The highly specific targeting of H3 K14 by multiple HATs provides strong evidence for a significant biological outcome associated with acetylation at this position. Of note, the ATP-dependent, bromodomain-containing, yeast chromatin-remodeling complex RSC (chromatin structure remodeling complex) has been shown to specifically bind H3 K14Ac, but not H3 K9Ac (or H3 K14 → Q) (15). In our system, however, the p300 bromodomain does not participate in recruitment of the coactivator to the promoter, because the chromatin template is assembled with unacetylated histones, and p300 is brought to the promoter via KIX domain interaction with DNA-bound Tax/pCREB (29, 36). Furthermore, we show herein that nucleosomes carrying K → R mutations throughout the four core histone tails—except H3 K14—support eviction (see Fig. 4C). Therefore, additional acetyl-lysine interactions are unavailable to the p300 bromodomain during K14 acetylation. We cannot exclude, however, a role for a bromodomain/K14Ac interaction during disassembly of the octamer by Nap1. In support of this model, previous studies have shown an interaction between Nap1 and CBP/p300 in vivo and in vitro (37, 38). Determining the detailed mechanism by which the Nap1/p300 complex facilitates eviction of the H3 K14 acetylated nucleosomes is of significant interest.

In summary, the downstream effects of histone N-terminal tail acetylation have been known for decades, yet defining the precise mechanism(s) by which histone acetylation promotes gene expression has been elusive. We demonstrate here that p300 acetylation of H3 K14 is both necessary and sufficient for Nap1-mediated disassembly of the histone octamer from the promoter DNA, and concomitant activator and Ac-CoA-dependent transcription in vitro. These unique findings link a specific histone tail acetylation event, NFR formation, and transcriptional activation in vitro. As such, they provide a molecular framework to interpret specific histone acetylation events in the context of NFRs and gene activation in vivo. Although our studies establish a direct connection between p300 acetylation of H3 K14 and histone chaperone mediated nucleosome disassembly, it is likely that this mechanism works in concert with additional posttranslational histone modifications and chromatin effector proteins to fine-tune the transcriptional response of a gene appropriate for a specific biological outcome.

Materials and Methods

Protein Expression and Purification.

pCREB (untagged) (39), Tax-His₆ (40), human Nap1 (GST-hNap1L1) (41), and His₆-p300 were purified using the methods described in their accompanying references (see also ref. 27). Xenopus histones carrying the indicated H3 point mutations were prepared by site-directed mutagenesis and confirmed by sequence analysis. Histone proteins were expressed and purified as previously described (42). All the purified proteins are shown in Fig. S1B. The GST-hNap1 expression plasmid was a gift from M. Giacca (International Centre for Genetic Engineering and Biotechnology, Trieste, Italy) and G. Steger (Institute of Virology, University of Cologne, Cologne Germany) (41), and the plasmids encoding the four core histone K → R point mutants (shown in Fig. 1B) were a gift from R. Roeder (Rockefeller University, New York, NY) (43).

Chromatin Assembly, Nucleosome Disassembly, in Vitro Transcription Assays.

A 588-bp fragment carrying the complete HTLV-1 promoter linked to a 290-bp G-less cassette was prepared by PCR, incorporating a biotin group at the upstream end of the fragment and immobilized on streptavidin-coated magnetic Dynabeads (see refs. 27 and 36). The PCR fragment was assembled into chromatin by salt deposition, as previously described (27). Nucleosome disassembly (see schematic, Fig. S1A) and in vitro transcription assays were performed as described previously (27, 36). Although not indicated in each figure, all nucleosome eviction reactions were performed with highly purified Tax, pCREB, p300, hNap1, and acceptor DNA.