Human Immunodeficiency Virus Type-1 Tat/Co-activator Acetyltransferase Interactions Inhibit p53Lys-320Acetylation and p53-responsive Transcription*

Patients with AIDS are at increased risk for developing various neoplasms, including Hodgkin's and non-Hodgkin's lymphomas, Kaposi's sarcomas, and anal-rectal carcinomas, suggestive that human immunodeficiency virus type-1 infection might promote establishment of AIDS-related cancers. Tat, the viral trans-activator, can be endocytosed by uninfected cells and has been shown to inhibit p53 functions, providing a candidate mechanism through which the human immunodeficiency virus type-1 might contribute to malignant transformation. Because Tat has been shown to interact with histone acetyltransferase domains of p300/cAMP-responsive element-binding protein (CREB)-binding protein and p300/CREB-binding protein-associated factor, we have investigated whether Tat might alter p53 acetylation and tumor suppressor-responsive transcription. Here, we demonstrate that both Tat and p53 co-localize with p300/CREB-binding protein-associated factor and p300 in nuclei of IMR-32 human neuroblastoma cells and in PC-12 pheochromocytoma cells. Further, p53 trans-activation of the 14-3-3ς promoter was markedly repressed by Tat-histone acetyltransferase interactions, and p53 acetylation by p300/CREB-binding protein-associated factor on residue Lys320 was diminished as a result of Tat-histone acetyltransferase binding in vivo and in vitro. Tat also inhibited p53 acetylation by p300 in a dosage-dependent manner in vitro. Finally, HIV-1-infected Molt-4 cells displayed reduced p53 acetylation on lysines 320 and 373 in response to UV irradiation. Our results allude to a mechanism whereby the human immunodeficiency virus type-1trans-activator might impair tumor suppressor functions in immune/neuronal-derived cells, thus favoring the establishment of neoplasia during AIDS.

Although the most frequent malignancies observed in AIDS patients are non-Hodgkin's lymphomas, central nervous system non-Hodgkin's lymphomas, and Kaposi's sarcomas, compartmentalization of human immunodeficiency virus, type-1 (HIV-1) 1 in the central nervous system might be associated with the recent increase of rare intracranial tumors, such as glioblastomas, anaplastic astrocytomas, and subependymomas (1)(2)(3). The viral trans-activator, Tat, can be endocytosed by surrounding uninfected cells and has been demonstrated to inhibit the G 1 /S arrest-inducing functions of p53, providing a candidate mechanism through which HIV-1 might contribute to malignant transformation in the central nervous system (4 -7). Tat is a 82-101-amino acid peptide that contains an arginine-rich motif required for binding a uracil-containing bulge in the Tat-associated region (TAR) of HIV-1 transcripts (8). Interactions between Tat/TAR-RNA stabilize viral mRNAs; thus, Tat principally acts as an elongation factor to enhance long terminal repeat trans-activation. The arginine-rich motif of Tat interacts with the catalytic histone acetyltransferase (HAT) domains of transcriptional co-activators, p300/CREBbinding protein (CBP) and p300/CBP-associated factor (P/ CAF):hGCN5 (9 -15), and Tat is acetylated by p300 on lysine residues Lys 50 /Lys 51 (Lys 51 is only weakly acetylated) and by P/CAF on Lys 28 (12, 16 -18). Acetylation of Tat on Lys 50 diminishes its binding affinity for TAR-RNA, and acetylation on Lys 28 enhances Tat binding to the Tat-associated kinase complex containing cdk9 and human cyclin T1 (16,17). Importantly, Mujtaba et al. (19) demonstrated that acetylated Tat peptide interacts with the bromodomain of P/CAF, and this interaction could play an important role in dissociation of Tat/ TAR-RNA complexes (20 -22). Tat/co-activator interactions are essential for HIV-1 long terminal repeat trans-activation (9,(12)(13)(14)(15)(16)(17)(18). Because the HAT domains of p300/CBP and P/CAF also target p53 for acetylation on residues Lys 373 /Lys 382 and Lys 320 , respectively, we hypothesized that Tat-HAT interactions might competitively interfere with p53 acetylation and, consequently, tumor suppressor-responsive transcription functions (23)(24)(25).

MATERIALS AND METHODS
Cell Culture, Immunofluorescence Laser Confocal Microscopy, and FACS-IMR-32 human neuroblastoma cells (ATCC number CCL-127) * This work was supported by the NCI, National Institutes of Health, and in part by the Department of Biological Sciences of Southern Methodist University. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed: were cultured in ATCC 2003, Eagle's minimum essential medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin-sulfate (Invitrogen). PC-12 rodent pheochromocytoma cells (ATCC number CRL-1721) were cultured in ATCC F-12K medium supplemented with 5% fetal bovine serum, 15% horse serum, 100 units/ml penicillin, and 100 g/ml streptomycin-sulfate. All of the cells were grown either in tissue culture dishes or eight-chamber slides (Nalge Nunc International) coated with mouse type IV collagen (Invitrogen) and were incubated under 10% CO 2 at 37°C. The transfections were performed using a calcium-phosphate system as recommended by the manufacturer (Invitrogen). Calu-6 carcinoma cells (ATCC number HTB-56) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin-sulfate and were transfected using LipofectAMINE reagent (Invitrogen) as recommended in the manufacturer's protocol. Molt-4 CD4 ϩ lymphoblastic leukemia cells were grown at 37°C under 10% CO 2 in RPMI medium supplemented with 20% human donor serum (Sigma), 100 units/ml penicillin, and 100 g/ml streptomycin-sulfate. The protein extracts were prepared by rapid freeze-thawing followed by centrifugation at 14,000 rpm at 4°C and quantified using the Bradford microassay and spectrophotometric analyses at 595 nm; 20 l from each sample was measured using a luciferase assay kit (Promega Corp.) and a Lumat model LB 9501 luminometer (Berthold, Inc.). All of the experiments were carried out as dose responses in two different cell-lines (IMR-32 and PC-12) or in duplicates for single-point analyses (error bars representative of standard deviations are shown). Immunofluorescence laser confocal microscopy was performed on IMR-32, PC-12, or HIV-1-infected Molt-4 cells using a Leica TCS spectrophotometric confocal microscope equipped with krypton and argon lasers, controlled by a Windows NT-based work station. Relative fluorescence intensities were quantified using TCS linear quantification software. Briefly, the cells were fixed in 0.2% glutaraldehyde and 1% formaldehyde in PBS, and nonspecific antigens were blocked by incubation with 3% (w/v) bovine serum albumin, 0.5% (v/v) Tween 20 in PBS. HIV-1 Tat was detected using the rabbit primary antibody, C-2145, and p53 was detected using either a monoclonal anti-p53 antibody (BP53-12; Upstate Biotechnology, Inc.) or an anti-p53 rabbit polyclonal antibody (Santa Cruz Biotechnologies, Inc.). p300 was detected using a monoclonal anti-p300 CT antibody (RW-128; Upstate Biotechnology, Inc.) or a rabbit polyclonal antibody (N-15; Santa Cruz Biotechnology, Inc.); P/CAF was detected using a goat polyclonal antibody (C-16; Santa Cruz Biotechnology, Inc.). Fluorescent secondary antibodies were used in appropriate combinations: rhodamine red-conjugated donkey anti-mouse IgG, fluorescein isothiocyanate-or Cy-3conjugated donkey anti-goat IgG, fluorescein isothiocyanate-conjugated donkey anti-rabbit IgG, or rhodamine red-conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch Laboratories, Inc.). For apoptosis and cell cycle analyses, transfected IMR-32 cells were treated with an arrest-inducing concentration of adriamycin (100 M) and incubated for 24 h at 37°C. The cells were harvested, washed twice with PBS, and either resuspended in 1ϫ annexin-buffer and stained for annexin-V surface expression for 10 min (Pharmingen, Inc.) or fixed in 3.7% formaldehyde/PBS for 10 min, washed, and permeabilized in 0.1% IGEPAL CA-630 (Sigma)/PBS containing 5 g/ml RNase and incubated on ice for 15 min. The samples were then stained using acridine orange (Molecular Probes, Inc.) at 4°C for 30 min. FACS analyses were performed using a Becton Dickinson, FACSCalibur flow cytometer. The cell cycle analyses were gated to exclude aggregates and fragments.
Protein Purification and in Vitro Acetylation-GST-Tat, GST-Tat K28A/K50A , and GST-p53 fusion proteins were expressed in Escherichia coli, strain DH5␣, by induction with 100 M isopropyl-␤-D-thiogalactopyranoside (Invitrogen) in LB broth containing 100 g/ml ampicillin at 37°C overnight. The cells were pelleted at 5000 rpm, washed twice with PBS, centrifuged, and resuspended in 5 ml of cold PBS containing the protease inhibitors antipain-dihydrochloride, bestatin, chymostatin, leupeptin, and pepstatin (Roche Molecular Biochemicals) at 50 ng/ml each. Bacteria were lysed by sonication over an ice bath using a Bronson sonic dismembrator equipped with a microtip and operated at 70% duty cycle. The lysates were clarified by centrifugation at 14,000 rpm in a Sorvall SS-34 rotor (DuPont/Sorvall) for 30 min at 4°C, and the supernatants were incubated for 2 h at 4°C with 500 l of a 50% mixture of equilibrated glutathione-Sepharose 4B (Amersham Biosciences). Following incubation, the matrices were pelleted by centrifugation at 2500 rpm at 4°C and then washed twice with PBS (with protease inhibitors). Bound GST fusion proteins were eluted in four 200-l fractions of 10 mM reduced glutathione (Sigma) in PBS (with protease inhibitors). The eluted fractions were analyzed by electrophoresis through a 12.5% SDS-polyacrylamide gel; the proteins were visualized by Coomassie staining. Purified proteins were dialyzed overnight at 4°C against 25 mM Hepes, pH 7.9, 5 mM KCl, 0.5 mM MgCl 2 , 0.5 mM EDTA, 0.25 mM dithiothreitol, and 10% glycerol; the fractions were stored in 20-l aliquots at Ϫ80°C. For in vitro acetylation, 100 ng of purified GST-p53 was mixed either with 5 units of purified p300 or ). The samples were incubated at 37°C for 30 min, then the reactions were quenched by the addition of 6 l of 5ϫ SDS-PAGE loading buffer, and the acetylated products were resolved through 4 -20% Trisglycine gels (Invitrogen, Inc.) and visualized by autoradiography/fluorography using ENHANCE reagent (PerkinElmer Life Sciences). Kodak XAR scientific imaging film was exposed for 72 h at Ϫ80°C.
Effects of HIV-1 Infection upon UV-responsive p53 Acetylation-To assess effects of HIV-1 infection upon p53 acetylation in response to UV irradiation in vivo, we infected 3 ϫ 10 6 Molt-4 CD4 ϩ lymphoblastic leukemia cells with 150 l of an infectious stock of HIV-1, HXB2 IIIB (HXB2 IIIB virus stock and Molt-4 cells were generously provided by Dr. G. Roderiquez, Food and Drug Administration, Center for Biologics Evaluation and Research, National Institutes of Health, Bethesda, MD), titered at ϳ2 ϫ 10 5 pg/ml of p24 gag protein determined by a standard anti-p24 gag enzyme-linked immunosorbent assay method. Aliquots were taken from HIV-1-infected samples on consecutive days (4 -6 days post-infection), and half of the volume of infected cells were UV-irradiated for 1.5 min (Fisher UV-Crosslinker at 120 mJ/cm 2 energy level), before culturing for an additional 3 h at 37°C under 10% CO 2 . Induction of p53 protein by UV treatment was confirmed by immunofluorescence laser confocal microscopy and immunoblotting using uninfected or HIV-1-infected cells and a p53-specific monoclonal antibody (Upstate Biotechnology, Inc). Tat expression was measured in lysates prepared from infected cells by immunoblotting using a rabbit polyclonal anti-Tat antibody or by direct visualization by immunofluorescence laser confocal microscopy. Co-localization of Tat and p53 in UVirradiated cells was determined and quantified by immunofluorescence laser confocal microscopy. The status of p53 acetylation on lysine residues Lys 320 and Lys 373 was detected by immunoprecipitating total intracellular p53 from untreated or UV-irradiated, HIV-infected Molt-4 cells. The protein G-agarose immune complexes were washed, and the bound products were resolved by 12.5% SDS-PAGE and immunoblotting using rabbit polyclonal antibodies that specifically recognize Lys 320 -acetylated p53 or Lys 373 -acetylated p53 (Upstate Biotechnology, Inc.).

Inhibition of p53-dependent Transcriptional Activation by
HIV-1 Tat-We first assayed whether Tat-HAT interactions might interfere with neuronal p53-responsive transcription in human IMR-32 neuroblastoma cells and rodent PC-12 pheochromocytoma cells. Because p53 regulates G 2 /M cellular arrest by driving 14-3-3 promoter trans-activation, we investigated the effects of HIV-1 Tat expression upon a 14-3-3 promoter luciferase reporter construct containing three p53responsive elements (other regulatory elements are deleted in this plasmid. Refs. 26 and 27). As shown in Fig. 1, HIV-1 Tat effectively inhibited p53 trans-activation from the 14-3-3 promoter in IMR-32 and PC-12 cells, respectively, whereas neither the Tat K28A/K50A mutant nor green fluorescent protein significantly influenced p53 transcription. Consistently, Tat expression did not seem to alter p53 protein levels, suggesting that Tat causes functional impairment of p53 (Fig. 1A).
Ectopic P/CAF and p300 Counter Inhibitory Effects of HIV-1 Tat upon 14-3-3 Promoter trans-Activation-To evaluate the relative contributions of Tat-p300 or Tat-P/CAF interactions toward the inhibition of p53 transcription, we analyzed a panel A, IMR-32 cells were co-transfected with expression constructs for Tat acetylationdefective mutants (3 g), RSV-Tat K28A and RSV-Tat K50A , and 1 g of CMV-p53 (or CMV-p53 R175H ) with 1 g of 14-3-3 luciferase. The protein levels are shown for p53, p53 R175H , Tat, Tat K28A , Tat K50A , and actin for extracts (30 g of total proteins) used in A. B and C, PC-12 cells (B) and IMR-32 cells (C) were co-transfected with 14-3-3 luciferase and CMV-p53 in the presence of increasing amounts of CMV-E1A 12S or CMV-E1A 12S⌬N (0.5, 1.5, or 3 g). D, IMR-32 cells were cotransfected with 14-3-3 luciferase, CMV-p53, and 3 g of RSV-Tat, either in the presence of increasing amounts of CMV-P/CAF or CMV-p300 (1.5, 2.5, or 5 g). WT, wild type. of previously characterized Tat mutants; the Tat K28A mutant is defective for acetylation by P/CAF, and Tat K50A is impaired for acetylation by p300/CBP (17). The results in Fig. 2A indicate that Tat K28A and Tat K50A mutants are partially impaired in their abilities to repress p53-responsive 14-3-3 transcription as compared with wild-type Tat. The differences in biological activities were not due to fluctuations in protein expression because these mutants were detected at levels comparable with wild-type Tat ( Fig. 2A, lower panel). The trans-dominant-negative p53 mutant, R175H, which is defective for p53-responsive element DNA binding, was used as a control to verify the p53 dependence of 14-3-3 promoter luciferase reporter gene activity ( Fig. 2A and Ref. 26). Because the adenoviral E1A 12S factor also interacts with the CH3/HAT domains of both p300/ CBP and P/CAF, inhibiting their acetyltransferase activities, we compared the effects of E1A 12S upon p53-responsive, 14-3-3 transcription in PC-12 and IMR-32 cells (28 -30). In both cell-types, E1A 12S inhibited p53 trans-activation in a dosage-dependent manner that resembled HIV-1 Tat (Fig. 2, B and C). An amino-terminal deletion mutant, E1A 12S⌬N, defective for interactions with the HAT domains of these co-activators did not affect p53 transcription, which is consistent with the notion that p53 trans-activation functions require co-activator acetyltransferase activities (Fig. 2, B and C, and Refs. 23-25, 30, and 31). Indeed, co-expression of increasing amounts of either P/CAF or p300 significantly countered the repressive effects of Tat upon 14-3-3 transcription, although ectopic p300 restored p53 trans-activation somewhat better than P/CAF (Fig. 2D).
Nuclear and immunostained for laser confocal microscopy using a rabbit polyclonal antibody against p53 (Santa Cruz Biotechnologies, Inc.) and fluorescein isothiocyanate-conjugated anti-rabbit secondary antibody. HIV-1 Tat and Tat K28A/K50A proteins were visualized using an anti-HIV-1 IIIB Tat, rabbit polyclonal antibody (C-2145), and fluorescein isothiocyanate-conjugated anti-rabbit secondary antibody. Endogenous p300 was detected using a monoclonal antibody against the carboxyl terminus of p300 (Upstate Biotechnology, Inc.) and rhodamine redconjugated anti-mouse secondary antibody. Endogenous P/CAF was detected using a goat polyclonal antibody against P/CAF (Santa Cruz Biotechnology, Inc.) and Cy-3-conjugated anti-goat secondary antibody. All of the fluorescence-conjugated secondary antibodies were obtained from Jackson ImmunoResearch Laboratories, Inc. The nuclei were stained with DAPI (Molecular Probes, Inc.) and are shown for reference. The merged images are at the right of each panel. Immunofluorescence laser confocal microscopy was performed using a 63ϫ objective in combination with 13ϫ digital zoom.
in nuclei of transfected IMR-32 neuroblastoma cells, as determined by immunofluorescence laser confocal microscopy, and both proteins displayed varied levels of co-localization with the transcriptional co-activator/acetyltransferase p300 (Fig. 3, A  and B). The subcellular distribution of Tat was observed to be both nucleoplasmic and nucleolar by immunostaining (Fig. 3B). Importantly, the Tat-derived mutant, Tat K28A/K50A , was present in the nuclei of IMR-32 cells (Fig. 3C). We examined the localization of Tat K28A/K50A because the K50A mutation is targeted within the nuclear localization sequence of Tat, and it was therefore necessary to establish whether the mutant protein is properly expressed in the nuclei. In addition, we observed that nuclear P/CAF levels appeared to be somewhat limiting in IMR-32 neuroblastoma cells, as revealed by comparatively weak, P/CAF-specific nuclear immunostaining (Fig.  3, A-C, lower panels).
HIV-1 Tat and Tat Synthetic Peptides Inhibit p53 Acetylation by P/CAF and p300 -Direct interactions of co-activator HATs with Tat, p53, and E1A 12S provide a plausible basis for competitive inhibition of factor acetylation ( Fig. 4A and Refs. 17, 23, 25, and 30). In IMR-32 cells co-transfected with p53 and either Tat or various Tat mutants, the wild-type Tat caused markedly diminished p53 Lys-320 acetylation by P/CAF in vivo, whereas no significant difference was observed for p300-dependent, p53 Lys-373 acetylation (Fig. 4B). The Tat-derived mutants did not affect p53 acetylation, despite having partially repressive effects upon p53-responsive transcription. These discrepancies may derive from weakened co-activator-binding affinities for the mutants or could suggest that alternate signaling interactions contribute to Tat-repression of p53 transcription. Overexpression of p53 did not result in increased levels of p53-acetylated forms (compare lanes 1 and 2 of Fig.  4B), supporting reports that nuclear co-activator concentrations are limiting (32,33). That no apparent difference was observed for p300-dependent, p53 Lys-373 acetylation suggests that the restoring effect of ectopic p300 upon 14-3-3 transcription may have resulted from enhanced factor recruitment (e.g. P/CAF) to p53-responsive elements (Fig. 2E). The results from electrophoretic mobility shift/DNA binding assays using a radiolabeled, p53-responsive oligonucleotide probe revealed that decreased p53-specific DNA binding in nuclear extracts prepared from IMR-32 cells, expressing p53 in the presence of Tat or various Tat mutants, coincided with diminished p53 Lys-320 acetylation (Fig. 4C, lower panel). The DNA binding activities of IMR-32 nuclear extracts were normalized to yield similar binding to a radiolabeled, consensus cyclic AMP-responsive element probe (Fig. 4C, top panel); the same amounts of nuclear extracts were then used to quantify p53-responsive element binding.
We next tested whether GST-Tat or GST-Tat K28A/K50A proteins or various Tat-derived synthetic peptides (Biosynthesis, Inc.) might influence the acetylation of GST-p53 in vitro by recombinant P/CAF or p300 (Upstate Biotechnology, Inc.). Amino acid sequences of full-length Tat and peptide derivatives (wild type and mutants) used in these analyses are provided in Fig. 4D. Relative input levels of GST-Tat, GST-Tat K28A/K50A , and GST-p53 are shown in Fig. 4E. GST-Tat inhibited p53 acetylation by P/CAF and p300 in a dosage-dependent manner; in contrast, GST-Tat K28A/K50A had no significant effect upon acetyltransferase activities (Fig. 4, E and F). Synthetic peptides spanning residues 23-43 (Tat  ) and 41-61 (Tat 41-61 ) of Tat inhibited p53 acetylation by P/CAF and p300, respectively, whereas mutant derivatives of these peptides (Tat K/A23-43 and Tat KK/AA41-61 ) did not (Fig. 4, E and F). Interestingly, the Tat 41-61 peptide did not alter p53 acetylation by P/CAF, whereas p53 acetylation by p300 was slightly inhibited by Tat 23-43 . These results may reflect differences in sub-strate affinities between p300 and P/CAF. The wild-type Tat peptides (residues 23-43 or 41-61) were weakly acetylated in our in vitro assays (data not shown); we did not observe GST-Tat acetylation under the conditions promoting p53 acetylation. In addition, Tat-HAT binding inhibited acetylation of histones H3/H4 in vitro (data not shown). We infer, based upon these results, that the release (or "turnover") kinetics of Tat acetylation by p300 and P/CAF are comparatively slow versus p53/histone acetylation consistent with the acetyltransferase inhibitory functions of Tat.
Because others have previously demonstrated that HIV-1 Tat directly interacts with p53 (6), we attempted to determine whether observed differences in co-activator-mediated p53 acetylation between wild-type Tat and the Tat K28A/K50A mutant might be attributed to differences in p53 binding. The results from co-immunoprecipitation experiments in Fig. 5A indicate that both purified recombinant GST-HIV-1 Tat and GST-Tat K28A/K50A interact with GST-p53 in vitro with similar apparent affinities. In addition, we have also tested whether Tat-P/CAF acetyltransferase domain interactions might interfere with binding of p53 to the catalytic acetyltransferase site of the co-activator and thereby competitively inhibit p53 acetyla-  (125, 250, and 500 ng), and the reactions were immunoprecipitated overnight at 4°C using an anti-p53 monoclonal antibody (Upstate Biotechnology, Inc.) and protein G-agarose. Affinity matrices were washed, the samples were electrophoresed through a 4 -20% SDS-PAGE precast gel (Invitrogen, Inc.), and bound HIV-1 Tat was detected using purified anti-HIV-1 Tat polyclonal antibody. B, p53-null, Calu-6 carcinoma cells were transfected with 3 g of CMV-p53 or CMV-vector control, the extracts were prepared in RIPA buffer, and equivalent amounts of total cellular proteins were used in in vitro binding assays. Increasing amounts of synthetic HIV-1 Tat 23-43 or Tat K/A23-43 peptides (250 ng, 500 ng, and 1 g) were mixed with Calu-6 extracts expressing p53, and 15 units of purified recombinant GST-P/ CAF 352-382 were added to each sample. 30 l of 50% glutathione-Sepharose 4B in PBS was added to each reaction, and GST-pull-down assays were carried out overnight at 4°C. Affinity matrices were washed twice with PBS, the samples were resolved by 12% SDS-PAGE, and bound p53 was detected using an anti-p53 monoclonal antibody. Input levels of synthetic Tat 23-43 and Tat K/A23-43 peptides were detected by electrophoresis through a 4 -20% SDS-PAGE precast gel and Coomassie staining. tion. To address this question, we expressed p53 in the background of p53-null, Calu-6 carcinoma cells and used equivalent amounts of total cellular proteins from prepared extracts in GST pull-down experiments. Increasing amounts of synthetic wild-type Tat  or Tat K/A23-43 mutant peptides were added to binding reactions in the presence of purified recombinant GST-P/CAF 352-382 , comprising the minimal acetyltransferase domain of P/CAF. As shown in Fig. 5B, the p53-P/CAF 352-382 interaction was significantly diminished in a dosage-dependent manner in the presence of wild-type Tat 23-43 peptide but was unaffected by the Tat K/A23-43 mutant peptide. Indeed, because Tat 23-43 peptides lack amino acid residues that are essential for p53 binding by Tat, we infer that HIV-1 Tat competes against p53 for binding to the minimal acetyltransferase domain of P/CAF in our in vitro assays.

/M Cellular Arrest Induced by Adriamycin in IMR-32
Cells-Because 14-3-3 regulates p53-induced G 2 /M arrest under conditions of genotoxic stress, we addressed whether inhibition of p53 functions as a consequence of Tat-HAT binding might weaken the G 2 /M checkpoint in the presence of adriamycin (100 M; Refs. 26 and 27). IMR-32 cells were transfected with Tat or various Tat mutants; after 48 h, the cells were treated with adriamycin and incubated for an additional 24 h prior to FACS analyses to evaluate apoptosis and cellular arrest using annexin-V (Pharmingen Corp.) and acridine orange (Molecular Probes, Inc.) staining methods. As shown in Fig. 6A, neither adriamycin treatment nor Tat alone caused apoptosis in IMR-32 cells. However, cells expressing Tat in the presence of adriamycin exhibited considerable programmed cell death (62.8%) as determined by annexin-V staining (Fig. 6, A and B). Of the Tat mutants, only Tat K50A was associated with significant apoptosis (43.3%), suggesting that P/CAF-binding by Tat might be an important contributing factor for the bypass of G 2 /M arrest. The results from this "mitotic trap" assay suggest that Tat-expressing cells not arrested in G 2 /M enter mitosis (M phase) and are killed through the action of adriamycin. Acridine orange staining/cell cycle analyses revealed that IMR-32 cells were arrested in G 2 /M (G 1 , 27.5%; G 2 /M, 47.0%) as a result of adriamycin treatment (Fig. 6C). Expression of various Tat mutants had no effect upon cellular arrest in G 2 /M (G 1 , ϳ24%; G 2 /M, 50% for each). The wild-type Tat significantly prevented cellular arrest in the presence of adriamycin (G 1 , 43.7%; G 2 /M, 34.4%) coincident with increased apoptosis (Fig. 6C; see also Fig. 6, A and B). A minor population of IMR-32 neuroblastoma cells that contained greater than G 2 /M DNA content were reproducibly detected in each sample, and these cells presumably possess inherent growth-regulatory defects resulting in deregulation of cytokinesis and S phase entry (Fig. 6). This aberrant cellular population did not appear to be affected by expression of HIV-1 Tat or Tat-derived mutants.
HIV-1 Infection Interferes with p53 Acetylation in Response to UV Irradiation-Finally, we examined the influence of HIV-1 infection upon acetylation of the p53 tumor suppressor in response to genotoxic stress caused by UV irradiation (24,25,34). CD4 ϩ Molt-4 lymphoblastic leukemia cells were infected with HIV-1, HXB2 IIIB , over a 6-day period, and consecutive aliquots were removed and exposed to UV irradiation. The cells were cultured for an additional 3 h before extracts were prepared for immunoprecipitation and immunoblot analyses. HIV-1 Tat protein expression was significantly detectable by the fourth day of infection and continued to increase through the sixth day (Fig. 7A). Ultraviolet irradiation is known to induce increased intracellular p53 levels (35,36), and indeed, p53 expression was detectable in uninfected UV-treated cells (Fig. 7B, upper left panel). Uninfected control cells or HIV-1-

FIG. 7. HIV-1 infection inhibits p53 acetylation in response to UV irradiation in Molt-4 cells.
A, Molt-4 cells were infected with HIV-1, HXB2 IIIB , and Tat expression was monitored on consecutive days by immunoblotting using an anti-Tat rabbit polyclonal antibody as described. B, induction of p53 protein expression by UV irradiation was verified by immunofluorescence laser confocal microscopy in uninfected or HIV-1infected cells in the presence or absence of UV treatment. C, nuclear co-localization of Tat and p53 in UV-irradiated, HIV-1-infected Molt-4 cells was detected by immunofluorescence laser confocal microscopy as described. Relative fluorescence intensities and signal overlaps were quantified and are shown below the micrographs. D, actin and p53 protein levels were determined by immunoblotting whole cell extracts prepared from uninfected or HIV-1-infected Molt-4 cells in the presence or absence of UV irradiation (top panels). Total intracellular p53 was immunoprecipitated (IP), and immunoblots were performed using antibodies that specifically recognize Lys 373 -acetylated p53 or Lys 320 -acetylated p53 (bottom panels).
infected Molt-4 cells that were not exposed to UV irradiation did not contain high p53 protein levels as determined by immunofluorescence laser confocal microscopy (Fig. 7B, upper middle and right panels) and immunoblotting using an anti-p53 monoclonal antibody (Fig. 7D). Nuclear Tat and p53 colocalized in HIV-1-infected cells, consistent with reports by others that Tat directly interacts with the p53 tumor suppressor (6); relative fluorescence intensities and signal overlaps for both proteins were measured by linear quantification and are shown below the micrographs in Fig. 7C. To assess the status of p53 acetylation in response to UV irradiation in HIV-1-infected cells, we prepared whole cell extracts and examined the levels of intracellular p53 (Fig. 7D). Actin is shown as an input control. Immunoprecipitations were performed, and the acetylated forms of p53 were identified by immunoblotting using antibodies that discriminate between Lys 320 -acetylated p53 or Lys 373 -acetylated p53 (Fig. 7D). As demonstrated in Fig. 7D, there was slight diminishment of p53 Lys-320 acetylation in HIV-1-infected cells on the fifth day of infection in the absence of UV treatment. More apparent, however, was the decrease in both p53 Lys-373 and p53 Lys-320 acetylation induced by UV irradiation on the fifth and sixth days of HIV-1 infection. Interestingly, the observed reduction in UV-responsive p53 acetylation did not correlate with intracellular levels of Tat, suggesting that other virally encoded proteins or host factors might influence p53 acetylation and responses to genotoxic stress (Fig. 7A). These results indicate that HIV-1 infection significantly prevents acetylation of p53 by co-activators/acetyltransferases in response to UV irradiation, and our in vitro evidence suggests that Tat could play an important role in mediating this inhibitory effect.

DISCUSSION
The tumor suppressor p53 induces the arrest of cell cycle progression in G 1 /S and G 2 /M under conditions of genotoxic stress (26,27,37,38). Because AIDS-affected individuals demonstrate increased risks for developing various neoplasms, inhibition of p53 Lys-320 acetylation/transcription by Tat-HAT binding could promote the acquisition of deleterious mutations by subverting p53-regulated checkpoint defenses. p53 transactivation functions and G 2 /M cell cycle control are regulated by post-translational modifications, e.g. acetylation and phosphorylation, and destabilization of p53 by Mdm2 is dependent upon p400/TRRAP-associated chromatin-remodeling complexes (39 -46). Mdm2 and adenoviral E1A 12S and E1B proteins have been shown to inhibit p53 acetylation by P/CAF, and therefore, the P/CAF co-activator/acetyltransferase may be a key modulator of p53 tumor suppressor-associated activities (47)(48)(49).
Recently, others have demonstrated that Tat inhibits acetyltransferase activities of the co-activators Tip60 and TAFII 250 , thereby causing repression of cellular transcription (50,51). Extracellular Tat produced by surrounding, infected cells might also enter and target p53 in nuclei of adjacent cells to create a local "by-stander" effect that might allow Tat to cooperate with oncogenic factors encoded by ␥-herpesviruses, such as Epstein Barr Virus and Kaposi's Sarcoma herpes-like virus, during co-infections frequently observed in AIDS patients (4,5,52,53). The mechanism by which extracellular Tat enters surrounding cells is largely unknown but requires cell surface heparan sulfate proteoglycans (54). In this study, we have shown that Tat-HAT interactions inhibit p53 Lys-320 acetylation as well as p53 DNA binding and transcription in neuronally derived cells. Both Tat and adenoviral E1A 12S repressed p53 trans-activation to similar degrees through co-activator/HAT binding. The unusually high percentages of cells affected by Tat in our cell cycle and apoptosis analyses are consistent with observations by others that Tat produced in transfected cells enters nuclei of surrounding untransfected cells, inhibiting p53 transcription functions. Our subsequent immunofluorescence microscopy experiments using CMV-HIV-1 Tat-transfected IMR-32 neuroblastoma cells also support this finding (data not shown). Col et al. (10) have reported that inhibition of CBPassociated acetyltransferase activity by HIV-1 Tat prevents histone acetylation but not p53 or MyoD acetylation; however, that study dealt exclusively with basal level p53 acetylation, whereas our present work summarizes effects of Tat-co-activator interactions upon basal level as well as UV-induced p53 acetylation (24,25,34). Our findings indicate that p53 acetylation on lysines 320 and 373 induced by UV irradiation is significantly diminished in HIV-1-infected Molt-4 cells in vivo (24,25,34). Inhibition of p53 acetylation did not strictly correlate with levels of intracellular Tat, and it is likely that other virally encoded factors, such as Nef, that also interact with p53 may influence the p53 response under conditions of genotoxic stress (24,25,34,55). Alternatively, host cellular factors induced by HIV-1 or the viral trans-activator, Tat, could antagonize Tat/co-activator interactions during certain stages of the infection process. Interestingly, the majority of viral factors that interact with p300/CBP are derived from oncogenic viruses. Indeed, Tat has been demonstrated to transform primary B-cells and induce lymphomas in transgenic mice, suggesting that Tat might contribute to certain hematological malignancies (56). Seo et al. (57) have reported that p300 and P/CAF acetyltransferases are inhibited by a cellular complex, INHAT (inhibitor of acetyltransferases), associated with the Set/TAF-I␤ oncoprotein involved in myeloid leukemia, indicative that inhibition of co-activator acetyltransferases might be generally linked to neoplastic transformation. Our results allude to a mechanism whereby HIV-1 Tat might impair tumor suppressor functions in immune/neuro-immune cells of the central nervous system, thus supporting the establishment of AIDS-related cancers.