The Human T-Cell Leukemia Virus Type 1 Tax Protein Confers CBP/p300 Recruitment and Transcriptional Activation Properties to Phosphorylated CREB

Nuclear extract.CEM cells, an HTLV-1-negative human T-cell line, were cultured in Iscove's modified Dulbecco's medium supplemented with fetal bovine serum. Nuclear extracts were prepared as previously described (11).

Expression and purification of recombinant proteins.Bacterially expressed Tax-His₆, Ser¹³³→A CREB (54), and glutathione S-transferase (GST)-KIX (CBP amino acids 588 to 683) (17) proteins were purified to >98% homogeneity as previously described (17). The KIX domain of CBP used in this study is 85% identical to the p300 KIX domain. CREB₃₂₇ was purified to apparent homogeneity and was free of contaminating nucleic acids, as recently described (35). CREB₃₂₇ is a naturally occurring splice variant where serine 119 corresponds to serine 133 in human CREB₃₄₁ (2). To avoid confusion, we used the serine 133 nomenclature throughout this work. Both CREB₃₂₇ and CREB₃₄₁ gave identical results in the CBP/p300 recruitment (data not shown). Full-length His₆-tagged p300 and Flag-tagged CBP were expressed from recombinant baculovirus in Sf9 cells and purified as previously described (27). All proteins were dialyzed against TM buffer (50 mM Tris [pH 7.9], 100 mM KCl, 12.5 mM MgCl₂, 20% [vol/vol] glycerol, 0.025% vol/vol Tween 20, and 1 mM dithiothreitol), aliquoted, and stored at −70°C. CREB was phosphorylated using the catalytic subunit of protein kinase A by incubating 1.6 μM CREB in a reaction mixture containing 3.3 μM ATP, 5 mM MgCl₂, and 55 U of protein kinase A (Sigma) in a 25 mM potassium phosphate buffer, pH 6.6.

Promoter fragments and oligonucleotides.A 643-bp promoter fragment was amplified by PCR from the 4TxRE/G-less plasmid template, which carries four reiterated copies of the HTLV-1 promoter-proximal vCRE cloned upstream of the HTLV-1 core promoter (1). Primers were designed to amplify the sequence between −187 and +456 and generate a fragment carrying a biotin group at the upstream end of the promoter. The native HTLV-1 promoter fragment, carrying sequences upstream of −306, was similarly prepared from the plasmid pHTLV-1/G-less. Primers for PCR amplification of 4TxRE/G-less were as follows: top strand, 5′-Bio/CATCGATAAGCTTCTAG; bottom strand, 5′-CATGATTACGCCAGGC. Primers for pHTLV-1/G-less were as follows: top strand, 5′-Bio/TGCCTGCAGGTCGAC; bottom strand, 5′-GCCTCAGGTAGGGCGGCGGG. The top strand, 5′-biotinylated sequences of the complementary oligonucleotides were as follows (CRE sequences underlined): Viral CRE, 5′-Bio/TTGTCAAGCCGTCCTCAGGCGTTGACGACAACCCCTCACCTCAAA; viral CRE with consensus octanucleotide core, 5′-Bio/TTGTCAAGCCGTCCTCAGGCGTTGACGTCAACCCCTCACCTCAAA; cellular CRE (somatostatin promoter), 5′-Bio/ATCAGGCTTCCTCCTCCTAGCCTGACGTCAGAGAGAGAGGTCGCC. The promoter fragments and the biotinylated double-stranded oligonucleotides were coupled to streptavidin Dynabeads (catalog no. 112.06; Dynal Biotech) at 0.08 pmol of DNA/μl of beads and 0.4 pmol of DNA/μl of beads, respectively. Binding was performed according to the manufacturer's instructions.

Antibodies.The following antibodies were used in Western blots: anti-CREB (sc-186), anti-phospho-Ser¹³³ CREB (sc-7978-R), anti-p300 (sc-584), anti-CBP (sc-583), and anti-GST (sc-138). All were purchased from Santa Cruz Biotechnology. Alexa Fluor IR700 and IR800 goat antimouse and goat antirabbit secondary antibodies were purchased from Molecular Probes. A monoclonal Tax antibody (Hybridoma 168B17-46-92) was obtained from the National Institutes of Health Aids Research and Reference Reagent Program.

Immobilized template assay using purified proteins.Tax, CREB, or pCREB (8 pmol each) was added to 20 μl reactions containing TM buffer with 10% glycerol, 10 μM ZnSO₄, 50 μM ATP, 80 μM acetyl-coenzyme A (CoA), 40 ng/μl of poly(dA-dT)-poly(dA-dT), and Dynabead-bound DNA (1 pmol for PCR fragments and 2 pmol for double-stranded oligonucleotides). Binding reactions were preincubated for 15 min at 30°C. Recombinant p300, CBP, or GST-KIX (2 pmol each) were then added to each sample in a final volume of 40 μl. Samples were mixed at 4°C for 1 h. Beads were magnetically isolated, washed three times with TM buffer, resuspended in 40 μl of Laemmli sample buffer (LSB), and boiled 5 min immediately prior to protein fractionation by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).

Coupled immobilized template assay and in vitro transcription.Following chromatin assembly, preinitiation complexes were formed on 1 pmol of the bead-bound promoter fragment in the absence or presence of Tax (2.5 or 10 pmol) and/or CREB (2.5 pmol) in a 50-μl reaction volume. The transcription factors were incubated with TM buffer, 50 μM ATP, and 50 μM acetyl-CoA for 15 min at 30°C. CEM nuclear extract (100 μg) was added to each binding reaction and incubated for an additional 60 min at 4°C. To analyze bound proteins, 46 μl of the total binding reaction mixture was removed, washed three times with TM buffer, resuspended in LSB, and analyzed by SDS-PAGE and Western blotting. The remaining 4 μl was processed for transcription analysis. The reaction mixture was brought to 30 μl with TM buffer and 50 μM acetyl-CoA, and RNA synthesis was initiated with the addition of 250 μM ATP, 250 μM CTP, 12 μM UTP, and 0.8 μM [α-³²P]UTP (3,000 Ci/mmol). Transcription reactions were processed and analyzed as previously described (32). Molecular weight markers (radiolabeled HpaII-digested pBR322) were used to estimate the sizes of the RNA products. A labeled 622-bp DNA fragment was added to each reaction mixture as a recovery standard.

Chromatin assembly.Recombinant Drosophila histones were assembled into chromatin as previously described (14, 16, 26), except that the assembly was performed on the linear, bead-bound promoter fragments. Assembly was performed at a 4:7 histone:DNA mass ratio. The degree of assembly was confirmed by micrococcal nuclease digestion (26). Micrococcal nuclease digestion of the chromatin-assembled 4TxRE promoter template is shown by Sharma et al. (44).

Protein analysis by SDS-PAGE.Proteins were resolved by SDS-polyacrylamide linear gradient gels and either silver stained or transferred to nitrocellulose and probed with the indicated antibodies. For the chromatin-based binding reactions, two-thirds of each sample was also resolved on an 18% polyacrylamide gel and histone proteins were visualized by Coomassie brilliant blue staining.

Transient-transfection assays.Human Jurkat T cells were cultured in Iscove's modified Dulbecco's medium supplemented with 10% fetal calf serum, 2 mM l-glutamine, and penicillin-streptomycin. Cells were serum starved (0.5% fetal bovine serum) and transfected with a constant amount of DNA using the Lipofectamine reagent (Invitrogen). After 24 h, the cells were harvested and lysed and luciferase activity was measured using the dual-luciferase reporter assay system (Promega) with a Turner Designs model TD 20-e luminometer. Firefly luciferase activity was normalized to Renilla luciferase activity from the herpes simplex virus thymidine kinase promoter (pRL-TK; Promega). The expression plasmid for Tax (pSG-Tax) (42) and the vCRE-Luc reporter plasmid (17) have been previously described. The transient-transfection assay was performed in triplicate and repeated in three independent experiments.

Image processing.The ImageQuant program (Molecular Dynamics) was used to quantify results. Images were processed in Adobe Photoshop, with minor adjustments to brightness/contrast as needed (gamma was kept at 1). No bands were obscured or altered. Images were annotated in PowerPoint. All experiments presented here were shown to be reproducible in at least three independent trials.

Tax requires pCREB for efficient recruitment of full-length p300 and CBP.Our initial goal was to study the coactivator recruitment characteristics of HTLV-1 promoter-bound Tax in complex with unphosphorylated and Ser¹³³-phosphorylated CREB. Nearly all previously published studies characterized the physical interaction of the Tax/CREB/vCRE complex with the KIX domain of CBP/p300. To investigate the recruitment of full-length p300 in vitro, we performed DNA pull-down assays using a promoter fragment carrying four tandem copies of the highly Tax-responsive vCREs (p4TxRE/G-less [1]). A biotin group was incorporated at the upstream end of the DNA, and the fragment was immobilized on magnetic streptavidin-agarose beads. Highly purified, recombinant proteins (Fig. 1A) were incubated with the biotinylated DNA template and washed, and the bound proteins were analyzed by Western blotting. In these studies, we optimized the concentrations of purified Tax and CREB or protein kinase A-phosphorylated CREB to saturate binding at the vCREs. We then added purified p300 at a physiologically relevant concentration (50 nM) (20).

• Open in new tab
• Download powerpoint

FIG. 1.

Recruitment of p300 and CBP to a viral CRE-containing promoter requires pCREB and Tax. (A) Analysis of recombinant, purified proteins by SDS-PAGE and silver staining. Molecular weight markers are indicated. (B) Immobilized 4TxRE promoter DNA (1 pmol) was incubated with full-length p300 (2 pmol) in the presence of CREB, CREB plus Tax, pCREB, and pCREB plus Tax (8 pmol each) as indicated. Bound complexes were washed, and DNA-bound proteins were separated on a 6%-to-12%-gradient SDS-polyacrylamide gel and analyzed by Western blotting. (C) Densitometric analysis of p300 binding from three independent experiments is presented graphically. (D) Analysis of recombinant, full-length CBP recruitment to the immobilized 4TxRE template in the presence of CREB, CREB plus Tax, pCREB, and pCREB plus Tax (8 pmol each), as indicated. The experiment was performed exactly as described for panel B, except purified, recombinant full-length CBP (2 pmol) was used in the binding reaction in place of p300.

We observed dramatically enhanced recruitment of full-length p300 to the Tax/pCREB complex relative to the Tax/CREB complex (Fig. 1B, lanes 1 and 3), indicating that Tax does not bypass the need for CREB phosphorylation in the recruitment of p300. These data are consistent with those of recent studies from our laboratory showing that pCREB stabilizes the complex containing Tax, KIX, and the vCRE much more effectively than the unphosphorylated form of the protein (24). Furthermore, recent studies indicate that Tax is directly responsible for promoting elevated levels of CREB phosphorylation in vivo (24, 48, 49). Together, these data indicate that phosphorylated CREB may play a prominent role in Tax transactivation.

Unexpectedly, the binding of full-length p300 to the HTLV-1 promoter in the presence of pCREB was nearly undetectable (Fig. 1B, lane 2). The addition of Tax produced a significant increase in p300 binding (compare lanes 1 and 3). This is in contrast to studies that used the isolated KIX domain of CBP/p300 to conclude that pCREB strongly recruits the full-length coactivators. DNA pull-down data from three independent experiments was quantified, and a graph showing the relative recruitment of p300 by the various DNA-bound complexes is shown in Fig. 1C. Compared with results for unphosphorylated CREB, pCREB enhanced p300 binding sixfold. The Tax/CREB complex enhanced p300 binding 40-fold relative to results for CREB. The Tax/pCREB complex further enhanced p300 binding 90-fold relative to results for pCREB. These binding assays were performed under a variety of conditions, including the absence and presence of acetyl-CoA and ATP. A recent study found that in the presence of acetyl-CoA, p300 was released from promoter templates following autoacetylation (3). We found that p300 recruitment was unaffected by acetyl-CoA and ATP in our assays (data not shown).

The dramatic results obtained with full-length p300 prompted us to perform the parallel experiment with the full-length paralogue CBP. Both coactivators have been implicated in Tax transactivation and signal-dependent transcription by CREB. The experiment presented in Fig. 1D revealed that Tax and pCREB are both required for CBP recruitment to the Tax-responsive promoter fragment and again that pCREB alone is insufficient for recruitment of the eponymous coactivator. As a control, we tested both CBP and p300 histone acetyltransferase activities and found that the two proteins were similarly active, indicating that they were functionally competent in our assays (data not shown). These data show that recombinant full-length CBP and p300 require both Tax and pCREB to bind the vCRE-containing promoter template in our recruitment assays.

CRE-bound pCREB recruits the isolated KIX domain but not full-length p300.In the experiments presented in Fig. 1, we used a promoter fragment carrying four copies of the Tax-responsive vCRE cloned upstream of a minimal promoter. Because of the significance of the data obtained with this fragment, we were interested in determining whether Tax and pCREB similarly supported the recruitment of p300 to the natural HTLV-1 promoter fragment (pLTR-G-less) (1). Figure 2A shows that p300 recruitment achieved with the natural HTLV-1 promoter was essentially identical to that with the synthetic promoter (see Fig. 1B). As expected, we did not observe Tax binding or p300 recruitment in the absence of CREB (Fig. 2A, lane 3).

• Open in new tab
• Download powerpoint

FIG. 2.

Individual viral and cellular CRE sequences support Tax/pCREB recruitment of p300. (A) The immobilized template assay was performed with the natural HTLV-1 promoter. Binding reactions and analyses were performed as described for Fig. 1B. (B) Partial top-strand sequence of the double-stranded oligonucleotides used in the immobilized template assays shown in panels C and D (see Materials and Methods for the complete sequences). The conserved Tax-binding sequences are indicated with a dashed underline. Nucleotides conforming to the consensus CRE are underlined. (C) The immobilized template assay was performed with the individual 45-bp doubled-stranded CRE sequences shown in panel B. Binding reactions and analyses were performed as described for Fig. 1B. The upper panel is a darker and higher-contrast version of the same p300 blot shown directly below and shows weak detection of p300 in lanes 1 and 2. (D) Immobilized template assay performed in parallel with the experiment shown in panel C, except that GST-KIX was included in the binding reactions in place of p300.

We next analyzed the recruitment of p300 to complexes formed on immobilized double-stranded oligonucleotide templates carrying various individual binding sites. We were interested in comparing pCREB and Tax/pCREB recruitment of p300 to an off-consensus CRE versus a consensus CRE. CREB binds the consensus CRE with an approximately 10-fold greater affinity than that for the off-consensus CREs present in the viral promoter (4). In parallel binding assays, we compared recruitment of p300 by Tax and pCREB to the single, promoter-proximal viral CRE and a consensus cellular CRE, which does not carry the GC-rich sequences necessary for Tax binding (Fig. 2B). Each oligonucleotide was designed with the core CRE sequence starting 23 bp from the upstream biotinylated end. As expected, Tax was not recruited to the cellular CRE (Fig. 2C, lanes 1 and 3). However, we reasoned that since this oligonucleotide contains the optimal CREB binding site, it may allow pCREB alone to recruit p300. Instead, recruitment of p300 by pCREB to the consensus CRE was nearly undetectable (Fig. 2C, lanes 1 to 4). We also designed a binding site carrying a consensus CRE and the GC-rich flanking sequences necessary for Tax recruitment (Fig. 2B). Inclusion of the GC-rich sequences adjacent to the consensus CRE recovered Tax binding and restored p300 recruitment by the Tax/pCREB complex (Fig. 2C, lanes 5 to 8). Together, these data reveal a strict requirement for Tax in complex with pCREB in the recruitment of p300 to the viral promoter.

Our inability to detect pCREB recruitment of the coactivators is significant. Therefore, we wanted to confirm that the full-length coactivators possessed regulatory properties lost in the isolated KIX domain. We next directly analyzed the binding of the various DNA-bound complexes to the KIX domain of CBP. The experiment shown in Fig. 2D was performed in parallel with the experiment shown in Fig. 2C, enabling direct comparison of KIX and p300 recruitment. Importantly, in the absence of Tax, we observed distinct patterns of KIX and p300 recruitment by pCREB (Fig. 2D). Paradoxically, pCREB strongly recruited KIX, despite being defective for p300 recruitment (compare Fig. 2C and D, lanes 1 and 2). As previously observed (17, 28), Tax in complex with unphosphorylated CREB also bound KIX (Fig. 2D, lanes 7 and 11), in contrast to our observations with full-length p300. Together, these data indicate that results of recruitment studies with the KIX domain may not represent the physiological interaction and that strong recruitment of the full-length coactivators requires additional regulators, such as Tax.

Tax/pCREB recruitment of p300 to the viral promoter correlates with strong transcriptional activation.Because of the relevance of our data to Tax and pCREB regulatory function, we tested the recruitment of native p300 to physiologically relevant chromatin templates. For these studies, we assembled the Tax-responsive promoter fragment used in Fig. 1 into chromatin, using purified Drosophila core histones and the recombinant assembly proteins Acf1/ISWI and Drosophila NAP-1 (14, 26, 34). Micrococcal nuclease assays were performed to verify the assembly of nucleosomes on the immobilized promoter fragment. The chromatin template was preincubated with pCREB alone or pCREB in the presence of increasing concentrations of Tax, followed by incubation with T-cell nuclear extract (HTLV-1-negative CEM cells). The immobilized template was washed, and bound proteins were subjected to SDS-PAGE and visualized by Western blotting. To assess the binding of nucleosomes to the washed templates, histones were visualized by Coomassie brilliant blue staining (Fig. 3A, lower panel). Tax in complex with pCREB but not pCREB alone was required for efficient recruitment of endogenous p300 to the promoter template (Fig. 3A, upper panel). We next assayed a portion of each binding reaction shown in Fig. 3A by in vitro transcription. As expected, transcription from the 4TxRE/G-less template was fully repressed in the presence of chromatin (Fig. 3B, lanes 1 and 2) (15, 16). The addition of pCREB had no detectable effect on transcriptional output, despite the strong binding to the template shown in Fig. 3A (lane 2). The absence of a transcription signal with pCREB correlated directly with the weak recruitment of p300 observed in the immobilized template assay. The addition of Tax and pCREB together, however, resulted in potent activation of transcription. We were also interested in testing whether Tax and unphosphorylated CREB could activate vCRE-dependent transcription. However, CREB becomes phosphorylated in the presence of ATP, which is required for the in vitro transcription reaction (data not shown). We therefore analyzed the CREB Ser¹³³-to-Ala mutant (Ser¹³³→A CREB), which cannot be phosphorylated. Tax and Ser¹³³→A CREB were unable to activate transcription, confirming that Tax and pCREB are both necessary for viral transcription (Fig. 3C). We next assayed p300 recruitment together with Tax-pCREB-dependent transcription in the presence of increasing concentrations of GST-KIX. We have previously shown that binding of KIX to the Tax/pCREB/vCRE complex blocks recruitment of full-length p300 (15). GST-KIX (but not GST) simultaneously blocked p300 recruitment and potently inhibited Tax/pCREB-mediated transcription (Fig. 3D). These data directly correlate p300 recruitment with Tax-dependent transactivation.

• Open in new tab
• Download powerpoint

FIG. 3.

Recruitment of endogenous p300 correlates with transcriptional activation from chromatin templates. (A) Immobilized, chromatin-assembled 4TxRE promoter DNA (1 pmol) was incubated in the absence (lane 1) or presence of pCREB (2.5 pmol) (lane 2) and increasing amounts of Tax (2.5 or 10 pmol) (lane 3 and 4), as indicated. Following the preincubation step, CEM nuclear extract was added and samples were incubated for 60 min at 4°C. Immobilized complexes were washed and analyzed by using a 4-to-20% SDS-polyacrylamide gradient gel and Western blotting (upper three panels). A portion of each binding reaction was resolved on an 18% SDS-polyacrylamide gel and stained with Coomassie to detect bound histone proteins (lower panel). Western blot of p300 present in the nuclear extract input (50%) and a Coomassie-stained gel of input histone (100%) are shown. (B) Small aliquots from each binding reaction mixture above (8% of the total) were incubated with acetyl-CoA and nucleoside triphosphates and analyzed in duplicate by in vitro transcription. The positions of molecular weight size markers, recovery standard, and full-length G-less transcripts are indicated. (C) Transcription reactions were performed with the unphosphorylatable mutant Ser¹³³→A CREB (CREB^S133A) and compared with results for pCREB. (All samples were analyzed on a single gel, but lanes 7 and 8 and lanes 9 and 10 were rearranged for this figure to maintain consistency with other experiments.) (D) Protein binding and transcription reactions were performed as described for panels A and B but in the presence of increasing amounts of purified GST-KIX (2.5, 5, and 25 pmol) (lanes 4 to 6), as indicated. GST alone (25 pmol) was used as a negative control (lane 7). Results from the transcription assay are shown in the top panel, and Western blots of bound p300, Tax, GST-KIX, and pCREB are shown in the lower three panels. Input Tax, pCREB, GST-KIX (2.5 pmol purified protein), and p300 (50% of nuclear extract) are shown in lane 8. (E) HTLV-1-negative Jurkat T cells were serum starved (0.5% FBS) and then cotransfected with the vCRE-luc reporter plasmid (100 ng) (17) and a Tax expression plasmid (pSG-Tax, 100 ng) (42). Eight hours prior to harvest, cells were treated with forskolin (FSK) (10 μM) or the dimethylsulfoxide vector and then analyzed by luciferase assay. Transient transfections were performed in triplicate, and each experiment was repeated three times. Relative luciferase activity is shown on the y axis. Corresponding Western blots probed against pCREB and Tax are shown for the samples analyzed in the luciferase assay.

The inability of pCREB to produce a transcriptional response from the vCRE promoter in vitro prompted us to explore the effect of CREB phosphorylation at a comparable promoter in vivo. We performed transient cotransfection assays using a Tax expression plasmid and a luciferase reporter plasmid. The vCRE-luciferase reporter plasmid carried three tandem copies of the vCRE immediately upstream of a core promoter (17). We used this Tax-responsive reporter construct to examine the effect of CREB phosphorylation mediated through the vCREs, without the influence of other promoter elements. Transfection assays were performed in HTLV-1-negative Jurkat T cells in the absence or presence of forskolin, a cAMP agonist that induces Ser¹³³ CREB phosphorylation. Figure 3E shows that forskolin treatment had no effect on vCRE-dependent transcription in the absence of Tax, consistent with the in vitro transcription assays performed with pCREB. Forskolin also had no effect on Tax transactivation, since Tax expression is sufficient to induce CREB phosphorylation in vivo (Fig. 3E, lane 3) (24, 48, 49). Similar results were obtained using the natural HTLV-1 promoter (data now shown). We confirmed that forskolin was active, since treatment with the agonist increased the levels of intracellular CREB phosphorylation (Fig. 3E) (24). These data are consistent with findings of chromatin immunoprecipitation studies that demonstrated the in vivo recruitment of p300 by CREB and Tax (31). These data provide further supporting evidence indicating that CREB phosphorylation alone is insufficient for transcriptional activation.