Activation of Integrated Provirus Requires Histone Acetyltransferase

A unique aspect of the retrovirus life cycle is the obligatory integration of the provirus into host cell chromosomes. Unlike viruses that do not integrate, retroviruses must conserve an ability to activate transcription from a chromatin context. Human immunodeficiency virus (HIV)-1 encodes an unusual and an unusually potent transcriptional transactivator, Tat, which binds to a nascent viral leader RNA, TAR. The action of Tat has been well studied in various reductive model systems; however, the physiological mechanism through which Tat gains access to chromatin-associated proviral long terminal repeats (LTRs) is not understood. We show here that a nuclear histone acetyltransferase activity associates with Tat. Intracellularly, we found that Tat forms a ternary complex with p300 and P/CAF, two histone acetyltransferases (HATs). A murine cell defect in Tat transactivation of the HIV-1 LTR was linked to the reduced abundance of p300 and P/CAF. Thus, overexpression of p300 and P/CAF reconstituted Tat transactivation of the HIV-1 LTR in NIH3T3 cells to a level similar to that observed for human cells. By using transdominant p300 or P/CAF mutants that lack enzymatic activity, we delineated a requirement for the HAT component from the latter but not the former in Tat function. Finally, we observed that Tat-associated HAT is preferentially important for transactivation of integrated, but not unintegrated, HIV-1 LTR.

A unique aspect of the retrovirus life cycle is the obligatory integration of the provirus into host cell chromosomes. Unlike viruses that do not integrate, retroviruses must conserve an ability to activate transcription from a chromatin context. Human immunodeficiency virus (HIV)-1 encodes an unusual and an unusually potent transcriptional transactivator, Tat, which binds to a nascent viral leader RNA, TAR. The action of Tat has been well studied in various reductive model systems; however, the physiological mechanism through which Tat gains access to chromatin-associated proviral long terminal repeats (LTRs) is not understood. We show here that a nuclear histone acetyltransferase activity associates with Tat. Intracellularly, we found that Tat forms a ternary complex with p300 and P/CAF, two histone acetyltransferases (HATs). A murine cell defect in Tat transactivation of the HIV-1 LTR was linked to the reduced abundance of p300 and P/CAF. Thus, overexpression of p300 and P/CAF reconstituted Tat transactivation of the HIV-1 LTR in NIH3T3 cells to a level similar to that observed for human cells. By using transdominant p300 or P/CAF mutants that lack enzymatic activity, we delineated a requirement for the HAT component from the latter but not the former in Tat function. Finally, we observed that Tat-associated HAT is preferentially important for transactivation of integrated, but not unintegrated, HIV-1 LTR.
DNA in eukaryotes is packaged with histones and non-histone proteins into chromatin (reviewed in Refs. 1 and 2). In this setting, active gene expression implies that chromatin-associated template is physically accessed by transcription factors (3,4). One route to the remodeling of chromatin is through acetylation of histone tails which alters nucleosomal structures (reviewed in Ref. 5) facilitating entry of transcription factors to promoters. Consistent with this concept, intrinsic (GCN5, p300/CBP, P/CAF, TAF250, and ACTR; reviewed in Ref. 6) and/or associated (reviewed in Ref. 7) histone acetyltransferase activities have been described for many transcription factors.
Like all retroviruses, an essential step in the life cycle of HIV-1 is the integration of proviral DNA into host cell chromosomes. Physiological mechanisms for proviral transcription must thus account for the activation of chromatin-associated viral genomes (2,(61)(62)(63)(64)(65). For HIV-1, it is abundantly clear that the integrated provirus is organized into nucleosomal forms (64,65). In addition, there is evidence that the activity of the virally encoded Tat protein describes a rate-limiting step for activation of integrated HIV-1 genomes (66). Although Tat has been extensively studied in various reductive paradigms (reviewed in Ref. 43), to date there is no information as to how this activator mediates transcription from nucleosomally packaged long terminal repeats (LTRs). Three possibilities can be contemplated as follows: (i) the activation domain of Tat, like other potent activators (67), could be inherently sufficient for remodeling chromatin; (ii) Tat could associate with SWI/SNF-like chromatin-remodeling protein machines (reviewed in Ref. 68); and/or (iii) Tat could associate with a histone acetyltransferase. Here, we report on an intracellular multiprotein complex that contains Tat, p300/CBP, and P/CAF. In characterizing this complex, we find that the histone acetyltransferase activity of P/CAF is preferentially required for Tat function. Finally, we show that this activity is needed for Tat activation of integrated, but not unintegrated, HIV-1 LTR templates. Tat-associated HAT likely provides a precedent for one function that may be conserved for viruses that integrate into host chromosomes.

EXPERIMENTAL PROCEDURES
Cells-HeLa, HeLa LTR-CAT, and NIH3T3 cells were propagated in Dulbecco's modified Eagle's medium with 10% fetal bovine serum. HeLa cells and NIH3T3 were transfected using calcium phosphate.
Antibodies-Anti-Gal4 monoclonal antibody was purchased from Santa Cruz Biotechnology; anti-p300 monoclonal antibody and anti-EF-1␣ polyclonal antibody were purchased from Upstate Biotechnology Inc.; anti-P/CAF has been described previously (28); and M2 anti-FLAG antibody was purchased from Eastman Kodak Co.
Preparation of Fusion Proteins-E. coli were grown overnight in 50 ml of LB with 100 g/ml ampicillin. A 500-ml LB ϩ ampicillin flask was inoculated with the overnight culture and was grown for 1 more hour at 37°C. Isopropyl-1-thio-b-D-galactopyranoside was added to a final concentration of 0.1 mM to induce fusion protein expression, and the culture was switched to 30°C for an additional 4 h. Cells were collected by centrifugation in a GSA rotor at 5800 ϫ g for 10 min at 4°C. Bacterial pellets were lysed either by sonication or by lysozyme digestion. For lysozyme digestion, the pellet was resuspended in 10 ml of buffer containing 10 mM Tris-HCl, pH 8.0, 1 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride (buffer A). 10 mg of lysozyme was added, and the cells were digested for 1 h on ice. 10 ml of buffer A supplemented with 20 mM of MgCl 2 and 50 units of DNase I (Life Technologies, Inc.) were added, and the mixture was allowed to incubate for 15 min on ice. For sonication, the bacterial pellet was resuspended in 25 ml of phosphate-buffered saline containing 1 mM phenylmethylsulfonyl fluoride and was sonicated (Branson) for 15 pulses (70%) at the maximum microprobe setting. The resulting mixture (either from sonication or lysozyme digestion) was centrifuged in a GSA rotor at 5800 ϫ g for 10 min at 4°C. A second centrifugation in a SS-34 rotor at 23,500 ϫ g for 20 min at 4°C clarified the extract of remaining debris.
Fusion Protein Affinity Chromatography-Truncated variants of HIV Tat-1 (SF2 strain) were expressed as glutathione S-transferase (GST) fusion proteins in E. coli DH5␣ (Life Technologies, Inc.) or BL21 (Pharmacia). Bacterial lysates were prepared as above and were incubated with glutathione-Sepharose overnight. The resin was washed extensively with phosphate-buffered saline and equilibrated with buffer B (20 mM HEPES-KOH, pH 7.9, 20 mM KCl, 1 mM MgCl 2 , 17% glycerol, 2 mM dithiothreitol). HeLa cell (Cell Trends, Middletown, MA) extracts were prepared as described by Dignam et al. (69) with the following modifications. After the first Dounce homogenization, the mixture was centrifuged once at 1500 ϫ g. Prior to dialysis with buffer B, 0.33 g/ml ammonium sulfate was added to precipitate proteins. The pellet was resuspended into 1 packed cell volume of buffer B and was dialyzed against 100 volumes of buffer B with two changes. Cellular extracts were incubated with the various protein-bound resins overnight at 4°C. The resins were packed into columns, and the columns were washed with buffer B containing 0.1 M KCl. Proteins were eluted in a stepwise fashion with buffer B containing 0.25 and 0.5 KCl. The eluates were desalted and concentrated using 10,000 molecular weight cut-off microconcentration tubes (Amicon, Beverly, MA) to a final volume of 100 l in buffer B.
Histone Acetyltransferase Assay-HeLa cell extract was prepared by a modification of the method described by Dignam et al. (69). Columns were eluted by KCl-containing buffers. Samples were concentrated and desalted in Amicon microconcentration tubes. Assays for histone acetyltransferase activity (HAT) were conducted in 30-l reactions at 37 o C for 20 min in buffer containing 50 g/ml histones (Sigma), 50 mM Tris-HCl, pH 8.0, 1 mM dithiothreitol, 0.1 mM EDTA, 150 mM NaCl, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, and 1 Ci/reaction (2.3 Ci/mmol) [ 3 H]acetyl coenzyme A (ICN). Proteins were resolved by 15% SDS-PAGE. Gels were fixed in 30% methanol and 10% acetic acid, washed with water, incubated 2 h with 1 M salicylic acid, dried, and exposed to x-ray film at Ϫ70 o C.
Western Blot Analysis-Column eluates were resolved by SDS-

FIG. 1. Protein-protein interaction between Tat and a histone acetyltransferase.
A, HAT activity associates with Tat. Elutions from GST and GST-Tat 1-101 columns, prepared as described under "Experimental Procedures," were assayed for HAT activity. The reaction was performed without (lanes 1-3) or with either 50 g/ml histones (lanes 4 -6) or 50 g/ml of bovine serum albumin (BSA) (lanes 7-9). [ 3 H]Acetyl coenzyme A (1 Ci/ reaction) was used as donor. Proteins were resolved by 15% SDS-PAGE. Gels were fixed in 30% methanol and 10% acetic acid, washed with water, incubated with 1 M salicylic acid, dried, and exposed to x-ray film at Ϫ70°C. B, HAT activity correlates with HeLa extract. HeLa cell extract was bound and eluted from GST-Tat-(1-101) columns and used in HAT reactions (lanes 1 and 3). Lanes 2 and 4 contain eluates from GST-Tat-(1-101) columns that were incubated with mock HeLa cell extract.
PAGE. Proteins were transferred to polyvinylidene difluoride membrane by semi-dry electroblotting (Millipore, Bedford, MA) for 1 h at 400 mA. Membranes were incubated with the primary antibody overnight at 4 o C, washed, and incubated for 1 h with the appropriate secondary antibody (Tropix) for 1 h. Proteins were visualized by chemiluminescence (Tropix, Bedford) according to the manufacturer's protocol.
Transient Transfection for CAT Assays-HeLa and NIH3T3 cells were transfected using calcium phosphate. Amounts of DNA are as indicated in the figure legends, and total amounts were normalized using carrier plasmids. Chloramphenicol acetyltransferase assays were performed as described (70,71).

RESULTS
An HIV-1 Tat-associated Histone Acetyltransferase Activity-Various studies have indicated that the integrated HIV-1 provirus is organized into nucleosomes (64,65). Histone acetyltransferase is believed to be important for transcriptional activation. However, a direct role of HAT in transcription of chromatin has not been explicitly demonstrated. Thus, we explored whether HAT functions in physiological activation of HIV-1 which is integrated into host cell chromosomes. Accordingly, we first checked for the possibility that a HAT associates with the viral transactivator, Tat. We probed HeLa total cell extract using GST affinity chromatography, comparing eluates from GST and GST-Tat columns for HAT activity. Elutions from GST or GST-Tat resin were tested in reactions without (Fig. 1A, lanes 1-3) or with (Fig. 1A, lanes 4 -6) either histone or bovine serum albumin as substrates (Fig. 1A, lanes 7-9). These incubations revealed specific acetylation of histones by an activity eluted from GST-Tat (Fig. 1A, lane 6). To rule out that this activity might have originated from a contaminating bacterial protein, eluates from a HeLa cell extract-equilibrated GST-Tat 1-101 column (Fig. 1B, lanes 1 and 3) were compared with counterpart eluates from a GST-Tat 1-101 (Fig. 1B, lanes  2 and 4) column incubated with mock extract. Acetylated histones were observed only in the former eluate (Fig. 1B, lane 1), supporting a eukaryotic origin of Tat-associated HAT activity.
Next, the region within Tat that binds HAT was characterized. Fig. 2A shows that HAT associates with both 1-and 2-exon forms of Tat (lanes 2 and 3) suggesting that the first 72 amino acids are wholly sufficient for binding. In control incubations, we observed that Tat-associated HAT showed specific discrimination in that it did not interact with the activation domain of Sp1 (GST-Sp1A; Fig. 2A, lane 4). Binding assays using Tat-deletion mutants revealed that GST- Tat A Tat-associated HAT Is p300 -To identify the Tat-associated HAT, we surveyed column elutions using specific antisera for p300, CBP, or TAF250, three well characterized histone acetyltransferases (Fig. 3). No reactive protein was detected with either anti-CBP or anti-TAF250 (data not shown). On the other hand, p300 reactivity was observed in GST-Tat but not GST eluates (Fig. 3, compare lane 3 to lane 6). The gel migration profile of this band is also consistent with that of p300.
The in vitro binding results prompted us to verify for the existence of an intracellular Tat⅐p300 complex. To accomplish this, we transfected HeLa cells with pCMV-TatGal4, which has the Gal4 DNA-binding domain fused to the C terminus of Tat. 24 h after transfection, cells were immunoprecipitated with either anti-p300 (p300; Fig. 5A, lane 1) , and B, lane 3). The immunoprecipitates were resolved by SDS-PAGE, transferred to membranes, and reciprocally probed in Western analyses with anti-Gal4 (Fig. 5A) or anti-p300 (Fig. 5B). In this manner, we found Tat in p300-specific immunoprecipitates (Fig. 5A, lane 2) and p300 in Tat-specific immunoprecipitates (Fig. 5B, lane 2). In the control anti-EF-1␣ immunoprecipitates, neither Tat nor p300 was found (Fig. 5A,  lane 2, and B, lane 3).
Tat and p300 Show Synergy in the Activation of the HIV-1 LTR-Rodent cells are partially defective for Tat transactivation of the HIV-1 LTR (72,73). A human cell-specific factor(s) has been shown to be required for optimal TAR-dependent transactivation (72,73). In the course of trying to understand Tat/p300 interactions, we wondered whether such interactions might be different between murine (NIH3T3) and human (HeLa) cells. As a first step in addressing this, we checked for the expression of p300 in HeLa and NIH3T3 cells using anti-p300. Surprisingly, in Western blots, we found that NIH3T3 cells had a markedly reduced (at least 10 times lower) ambient amount of p300 when compared with HeLa cells (Fig. 6, compare lanes 2 and 3). Coomassie Blue staining of the membrane verified that this reduction in the NIH3T3 sample was not due trivially to a gel-loading artifact (Fig. 6, lanes 5 and 6).
In view of the low levels of p300, NIH3T3 cells represent an ideal host cell system for assessing the functional implications of the Tat/p300 interaction. We asked in transient transfections how p300 might influence Tat transactivation (Fig. 7). Cotransfection into NIH3T3 cells of pLTR-CAT with Tat produced a 6.6-fold activation (Fig. 7, lane 2) over that of pLTR-CAT alone (Fig. 7, lane 1). Transfection of p300 alone did not noticeably alter the expression of the HIV 1-LTR contained in pLTR-CAT (Fig. 7, lane 3). However, we found that a threeplasmid transfection of Tat, p300, and pLTR-CAT produced 74 times more activity than pLTR-CAT alone (Fig. 7, lane 4). This result indicates a more than additive synergistic interaction between Tat and p300 in activation of the HIV-1 LTR.
Because p300 alone had minimal effects on LTR activation (Fig. 7, lane 3) and because p300 lacks inherent DNA binding activity, its activity at the HIV-1 promoter likely occurs as a consequence of physical recruitment by Tat. To test this hypothesis, we asked whether mutated versions of p300 (e.g. nt, m, or ct; Fig. 4) might interfere with the activity of full-length p300. In direct comparisons, we observed that p300nt (amino acids 1-670) strongly inhibited Tatϩp300 transactivation of the HIV-1 LTR (Fig. 7, lane 9) whereas p300m and p300ct did not (data not shown). That p300nt (but not p300 m nor p300ct) is competent for direct binding to Tat (Fig. 4) is consistent with a physical competition between p300nt and full-length p300 for Tat. P/CAF Is a Coactivator for Tat-P/CAF is a histone acetyltransferase that was identified by its association with p300. Based on results here and elsewhere, it is clear that different domains in p300 interact with P/CAF and Tat. In Fig. 7, we observed that a p300 mutant (p300⌬E1A), defined as being unable to bind P/CAF, failed to support Tat transactivation in NIH3T3 cells (lane 6). This finding raised the possibility that the functional contribution of p300 to Tat activity might be mediated indirectly through P/CAF. In pursuing this hypothesis, we first asked if there is evidence consistent with a possible Tat, p300, and P/CAF ternary complex. Thus we re-examined flow-through fractions , washes, and salt elutions from GST and GST-Tat columns (Fig. 3) searching for P/CAF (Fig. 8A). Intriguingly, P/CAF was found in GST-Tat-eluates but not in GST eluates (Fig. 8A, compare lane 6 to 3). This result is compatible with either Tat and P/CAF interacting directly or Tat and P/CAF interacting indirectly as a consequence of a p300-bridge. To distinguish between these two possibilities, we substituted in the incubations purified recombinant P/CAF (rP/CAF) in place of whole cell lysates. When rP/CAF was equilibrated with either GST or GST-Tat (Fig. 8B), P/CAF was found in eluates from the latter but not from the former (Fig.  8B, compare lane 6 to lane 3). This result supports a direct Tat-P/CAF interaction, suggesting that a Tat⅐p300⅐P/CAF ternary complex is stabilized through multiple contacts.
In Western blots, P/CAF expression in NIH3T3 cells was also found to be exceedingly low. 2 Thus, we asked whether overexpression of P/CAF in NIH3T3 cells could influence Tatϩp300 (Fig. 7) activation of the HIV-1 LTR. Repeating a series of cotransfections, we observed that Tat expression activated the LTR (Fig. 9, lane 2) 7.5-fold over its basal activity (Fig. 9, lane  1); cotransfection of Tat with p300 produced an 80-fold activation (Fig. 9, lane 4). In comparison, P/CAF alone had little effect on Tat activity (Fig. 9, compare lane 1 to lane 6). However, co-expression of Tat, p300, and P/CAF provided an 860-fold activation (Fig. 9, lane 7 and lanes 9 to lane 11). (Note that 860-fold was determined based upon a serial dilution of the assay in Fig. 9, lane 7. The result in lane 7 was in a non-linear range of the CAT assay.) Thus, co-expression of Tat, p300, and P/CAF reconstituted Tat transactivation in NIH3T3 cells to a level comparable to that observed for human cells.
Inhibition of Tat Activity by a P/CAF Mutant Lacking HAT Activity-Both p300 and P/CAF are histone acetyltransferases. To determine whether each moiety independently contributes an "adaptor" activity and/or a histone acetyltransferase activity toward Tat-mediated HIV-1 LTR activation, we checked for functional effects from overexpressing either p300 or P/CAF mutants that lack enzymatic activity. We tested both types of mutants for inhibition of Tat activation of either an integrated or a non-integrated HIV-1 LTR template (Fig. 10).
In side-by-side comparisons, an intriguing pattern emerged (compare Fig. 10, A and B). Neither p300⌬HAT overexpression nor P/CAF⌬HAT overexpression was found to affect Tat activation of a non-integrated LTR-template (Fig. 10B). On the other hand, in an integrated-template setting, P/CAF⌬HAT, but not p300⌬HAT, potently abrogated Tat activation function (Fig. 10A). In view of previous results, two interpretations can be made from this set of results. First, one can conclude that in the Tat⅐p300⅐P/CAF ternary complex, p300 likely contributes a primary adaptor function, whereas P/CAF contributes a histone acetyltransferase function. Second, the histone acetyltransferase function supplied by P/CAF appears to be important for Tat activation of an integrated, but not unintegrated, LTR.  1-3) or GST (lanes 4 -6) were resolved by 8% SDS-PAGE. Proteins were transferred to polyvinylidene difluoride membrane and probed with anti-p300 monoclonal antibody. Visualization was performed using chemiluminescence (Tropix). Arrow points to p300-reactivity (p300). No signal for either CBP or TAF250 was observed when the membrane was probed with either anti-CBP or anti-TAF250 (data not shown).

DISCUSSION
Most eukaryotic DNA exists as chromatin in a nucleosomal format. In intact nuclei, gene transcription is a multi-step process that involves, minimally, the binding of upstream activators, the assembly of general transcription factors into an active complex at the promoter, and the initiation and elongation of transcription by RNA polymerase (reviewed in Ref. 74). How these steps are accomplished on naked DNA templates has been well studied in cell-free systems. How the corresponding steps transpire in the setting of chromatinized DNA is less FIG. 4. Tat interacts with the N terminus of p300. A, schematic representations of functional domains in p300/CBP (top). The p300 mutants used for mapping interaction with Tat are shown (bottom). p300nt contains amino acids 1-670; p300m contains amino acids 671-1194; and p300ct contains amino acids 1135-2414. B, binding of Tat to the N terminus of p300. GST-Tat columns were equilibrated with SF9 extracts expressing flag-tagged full-length p300 (lanes 1-4), p300nt (lanes 5-8), p300m (lanes 9 -12), or p300ct (lanes 13-16). Flow throughs (ft; lanes 1, 5, 9, and 13) were collected, and the columns were extensively washed (w, lanes 2, 6, 10, and 14). Beads were then eluted with stepwise increases of KCl-containing buffer (0.25 M, lanes 3, 7, 11, and 15; 0.5 M, lanes 4, 8, 12, and 16). The presence of p300 and p300 mutants in the elutions were assessed by Western blotting using anti-flag M2 monoclonal antibody. well understood. There is, however, a consensus that nucleosomes structurally hinder access of transcription factors to their cognate sites on DNA. Thus, many transcription factors bind with a 2-log reduced affinity for nucleosomal DNA as compared with naked DNA (reviewed in Ref. 75). Similarly, there is also evidence that polymerase elongation is more effi-cient on naked as compared with nucleosome-organized DNA (76).
Viruses are obligatory parasites of host cells. Unlike some nuclear DNA viruses or cytoplasmic RNA viruses, all retroviruses integrate obligatorily into host chromosomes during the replicative life cycle. Reasonably, then, regulatory mechanisms for productive retroviral infections must conserve functions evolved to accommodate the structural organizations inherent to eukaryotic DNA. Indeed, intracellularly, retrovirus integration studies have shown a strong preference by integrase for nucleosomal rather than nucleosome-free DNA (77,78).
HIV-1 is one of the better studied paradigms for retroviral transcription. Work from various laboratories (64,65,79,80) has shown that chromatin structure is a major contributory factor in the regulation of HIV-1 provirus expression. Hence, it has been suggested that transcription of the same HIV-1 DNA template is controlled by distinct mechanisms depending on its presence in an integrated versus an unintegrated format (38). Additionally, it is known that the HIV-1 LTR in its integrated form is organized into nucleosomal and nucleosome-free structures (65). Disruption of one or more of the LTR-associated nucleosomes by diverse stimuli such as tumor necrosis factor -␣ (81) has been correlated with activation of HIV-1 transcription. Binding of upstream factors such as NF-B and Sp1 to cognate sites has also been linked to re-modeling of positioned nucleosomes in the HIV-1 LTR (79,80). At the same time, a genetic analysis of latent HIV-1 infections has found Tat expression to be the rate-limiting event for activation of proviral transcription (66). Collectively then, these observations suggest that remodeling of chromatin is important for the transcription of proviruses and that in an integrated-HIV-1 context a Tatassociated function might trigger the first step toward derepressing nucleosome-associated templates.
If a Tat-associated activity is rate-limiting for the transcription of the integrated provirus, then it is reasonable that this activity could influence access to chromatin-associated LTR. Currently, many cellular proteins have been described to bind Tat (see Introduction). However none of these Tat-binding factors has been assayed for transcriptional function on chromatin. Furthermore, based on existing knowledge, none would be expected to fulfill a role in chromatin remodeling. Thus, the idea that a Tat-associated factor remodels chromatin leads logically to the consideration of either protein machines such as SWI/SNF/NURF/RSC (reviewed in Ref. 68) or histone acetyltransferases (reviewed in Ref. 11). Because inhibitors of histone deacetylation were recently shown to activate integrated HIV-1 LTRs (81, 82), we were thus directed to explore the existence of a Tat-associated histone acetyltransferase. Our findings show that both p300 and P/CAF exist in an intracellular complex with Tat. In understanding the respective contributions of p300 and P/CAF, we concluded that the HAT activity of P/CAF but not of p300 is preferentially required for Tat activation of integrated LTRs.
What are the functional implications for HATs that associate with Tat? This finding serves to clarify three observations relevant to Tat-mediated transcription. First, it provides a rationale as to how Tat provides a rate-limiting switch for integrated proviruses (66). Plausibly, a small amount of Tat protein, hypothesized to be packaged in virions (83,84), serves as the initial trigger that de-represses the chromatinized LTR promoter. This would explain how Tat participates in initiating access and formation of a RNA polymerase II complex at the integrated promoter and is consistent with the preferential importance of Tat-associated HAT in the activation of integrated but not unintegrated HIV-1 templates (Fig. 10). Second, based on the knowledge that nucleosomes negatively influence  1-3). Visualization was by chemiluminescence (Tropix). After immunoblotting the same membrane was stained with Coomassie Blue (lanes 4-6). Wblot, Western blot. FIG. 7. Tat and p300 show synergy in HIV-1 LTR transactivation. Activities in NIH3T3 cells are shown for various combinations of transfected plasmids. Fold activation is relative to the expression of pLTR-CAT transfected alone which was set as 1. The identity of plasmids used in each transfection is indicated at the bottom of the graph. The amounts of plasmid transfected were pLTR-CAT (5 g), pSVTat (100 ng), pCMVp300 (2 g), pCMVp300⌬E1A (2 g), and pCMVp300nt (2 g). In all cases, pCMV vector DNA was used to normalize total amount of DNA. CAT assay was performed 24 h after transfection, and CAT activities were visualized and quantified by phosphorimaging. polymerase elongation (76), an RNA polymerase-associated HAT activity would also address a processivity defect previously described for the transcription of integrated HIV-1-templates (see Ref. 85 and reviewed in Ref. 42). Recent findings that Tat is associated with RNA polymerase II (56,86,87) suggest that Tat can physically recruit and transfer (85) a HAT activity to an elongating polymerase. If so, this is analogous to the recruitment of HAT to RNA polymerase II complexes by cellular RNA-binding proteins (88, 89) and illustrates diverse means through which viruses and cells conserve important regulatory functions. Finally, the contribution of P/CAF to Tat transactivation is a further explanation for an HIV-1 defect in rodent cells. Thus, in certain cell types with reduced p300 and P/CAF activities, one would expect HIV-1 to have a HATspecified post-integration block to transcription in addition to other characterized obstacles such as that for entry of virus into cells (reviewed in Ref. 90).
FIG. 9. Contribution of P/CAF to Tat transactivation. Fold activation in NIH3T3 cells is expressed relative to the basal expression of LTR-CAT arbitrarily set as 1. The transfected plasmids are as indicated below the graph. The amounts of plasmids transfected were pLTR-CAT (5 g), pSVTat (100 ng), pCMVp300 (2 g), pCMVp300⌬E1A (2 g), and pCMV-P/CAF (2 g). In all cases, pCMV vector was used to normalize for total amounts of DNA. CAT assay was performed 24 h after transfection, and CAT activities were visualized and quantified by phosphorimaging. were pLTR-CAT (5 g), pSVTat (100 ng), pCMVp300 (2 g), pCMVp300⌬HAT (1 g and 5 g), pCMV-P/CAF (2 g), and pCMV-P/CAF⌬HAT (1 g and 5 g). In all cases, pCMV vector was used to normalize for total amounts of DNA. P/CAF histone acetyltransferase activity and presumably a p300 adaptor function in the life cycle of HIV-1. Another salient point that emerges from this study is a direct demonstration (using P/CAF⌬HAT mutant) that histone acetyltransferase activity is needed for activated expression of integrated genes. It is likely that these properties are not restricted to the Tat/ HIV-1 system. Indeed, one could suppose that all integrated DNA and RNA viruses would benefit from a conservation of these functions in one fashion or another. Specifically for HIV-1, a requirement for HAT activity represents another potential window through which to intervene against the virus.