p53 Transcriptional activity is mediated through the SRC1-interacting domain of CBP/p300.

The tumor suppressor p53 recruits the cellular coactivator CBP/p300 to mediate the transcriptional activation of target genes. In this study, we identify a novel p53-interacting region in CBP/p300, which we call CR2, located near the carboxyl terminus. The 95-amino acid CR2 region (amino acids 2055--2150) is located adjacent to the C/H3 domain and corresponds precisely with the minimal steroid receptor coactivator 1 (SRC1)-interacting domain of CBP (also called IBiD). We show that the region of p53 that participates in the CR2 interaction resides within the first 107 amino acids of the protein. p53 binds strongly to the CR2 domain of both CBP and the highly homologous coactivator p300. Importantly, an in-frame deletion of CR2 within the full-length p300 protein strongly compromises p300-mediated p53 transcriptional activation from a chromatin template in vitro. The identification of the p53-interacting CR2 domain in CBP/p300 prompted us to ask if the human T-cell leukemia virus (HTLV-I) Tax protein, which also interacts with CR2, competes with p53 for binding to this domain. We show that p53 and Tax exhibit mutually exclusive binding to the CR2 region, possibly contributing to the previously reported Tax repression of p53 function. Together, these studies identify and molecularly characterize a new p53 binding site on CBP/p300 that participates in coactivator-mediated p53 transcription function. The identity of the p53.CR2 interaction indicates that at least three distinct sites on CBP/p300 may participate in mediating p53 transactivation.

The tumor suppressor p53 recruits the cellular coactivator CBP/p300 to mediate the transcriptional activation of target genes. In this study, we identify a novel p53-interacting region in CBP/p300, which we call CR2, located near the carboxyl terminus. The 95-amino acid CR2 region (amino acids 2055-2150) is located adjacent to the C/H3 domain and corresponds precisely with the minimal steroid receptor coactivator 1 (SRC1)-interacting domain of CBP (also called IBiD). We show that the region of p53 that participates in the CR2 interaction resides within the first 107 amino acids of the protein. p53 binds strongly to the CR2 domain of both CBP and the highly homologous coactivator p300. Importantly, an in-frame deletion of CR2 within the full-length p300 protein strongly compromises p300-mediated p53 transcriptional activation from a chromatin template in vitro. The identification of the p53-interacting CR2 domain in CBP/p300 prompted us to ask if the human T-cell leukemia virus (HTLV-I) Tax protein, which also interacts with CR2, competes with p53 for binding to this domain. We show that p53 and Tax exhibit mutually exclusive binding to the CR2 region, possibly contributing to the previously reported Tax repression of p53 function. Together, these studies identify and molecularly characterize a new p53 binding site on CBP/p300 that participates in coactivator-mediated p53 transcription function. The identity of the p53⅐CR2 interaction indicates that at least three distinct sites on CBP/p300 may participate in mediating p53 transactivation.
CBP and the highly related protein p300 are very large, highly conserved coactivator proteins that serve to mediate the regulation of gene expression in metazoans. Many transcriptional regulatory pathways converge at CBP and p300 (1)(2)(3)(4). These include pathways that are required for development and differentiation, response to hormonal stimulation, apoptosis, and tumor suppression. A significant number of transcription factors, such as Mdm2, BRCA1, HTLV-I Tax, and SRC1, 1 have been demonstrated to interact with CBP/p300, with several binding at multiple sites on the coactivators (5). The functional significance of these multivalent activator/coactivator interactions is currently unknown.
p53 is a sequence-specific, DNA-binding transcription factor that induces apoptosis or cell cycle arrest in response to genotoxic stress, thus blocking the transmission of DNA mutations to progeny cells (6). Loss of p53 activity has been identified in 60% of the human malignancies examined (7,8), consistent with its critical role in the suppression of malignant transformation. The tumor suppressor functions of p53 are directly linked to its ability to mediate transcriptional activation. To stimulate transcription, p53 binds as a tetramer to specific response elements located in the transcriptional control regions of p53 target genes (6,9). This step initiates the assembly of the complex transcriptional apparatus that initiates RNA synthesis. This critical early step in transcriptional activation is believed to be facilitated by the ability of p53 to simultaneously bind the specific DNA sequences and recruit CBP/p300 to the p53-responsive promoters. CBP/p300 recruitment appears to concomitantly bring RNA polymerase II to the target promoters (10), increasing the rate of preinitiation complex assembly (11). There is also evidence that, following promoter association, CBP/p300 may also recruit or stabilize components of the general transcription machinery, including TFIIB and TBP (12,13). CBP/p300 also facilitates transcriptional activation through nucleosome and transcription factor acetylation. The coactivators have been shown to directly acetylate lysine residues present within the amino-terminal tails of the four core histones (14). Acetylation appears to increase the accessibility of the nucleosomal DNA to transcription factor binding, a critical step in gene activation (15,16). Interestingly, CBP and p300 have also been shown to acetylate p53 at lysine residues 373 and 382 (17). Although acetylated p53 binds short fragments of DNA with a higher affinity than the unacetylated form, this modification does not appear to significantly affect p53 DNA binding activity on chromatin assembled templates (18).
These observations serve to illustrate a prominent role for CBP/p300 in mediating the tumor suppressor functions of p53. However, the molecular details of the physical interaction between the activator and coactivator remain elusive. Several previous studies have indicated that p53 specifically binds to multiple sites on the coactivator, including the KIX domain (19), and an ill-defined carboxyl-terminal region of CBP/p300 (20 -23). The amino-terminal activation domain of p53 has been shown to participate in each of these coactivator interactions (19,24). In studies that attempted to elucidate the precise carboxyl-terminal region of CBP/p300 involved in p53 binding, only the C/H3 domain of CBP (approximate aa 1764 -1850; also called TAZ2 and TRAM) has emerged as a site of p53 interaction (25). However, a recent study using heteronuclear NMR methods to monitor the intermolecular interactions between the activation domain of p53 and C/H3 showed that the binding affinity was weak (K D ϭ 300 M) (24). This result suggests that p53 may make additional contacts within the carboxyl-terminal region of CBP/p300.
In this study, we set out to further characterize the interaction between p53 and the carboxyl-terminal half of CBP/p300. We were interested in determining whether another carboxylterminal site on CBP/p300, alone or in conjunction with C/H3, might account for the observed tight binding of p53 to this region (20 -23). We have identified a new p53-interacting domain on CBP (aa 2055-2150) and p300 (aa 1970 -2193), which we have named CR2. This region corresponds precisely with a domain present on both CBP and p300 that is utilized by steroid receptor coactivator 1 (SRC1) in activated transcription by liganded nuclear hormone receptors (26 -28). Furthermore, this region has been shown to be an important interaction site for numerous transcription factors, including IRF-3 and HTLV-I Tax (29,30). Recently, the solution structure of this domain (IBiD) was solved using heteronuclear NMR and was shown to be composed of three tightly compacted ␣-helices (30). A mutation in CBP that resides in the first of the three ␣-helices in the CR2 region significantly reduces the interaction with p53. We also show that p53 binds to the CR2 domain present in the highly homologous coactivator p300. Importantly, deletion of this region in full-length p300 strongly compromises p53-mediated transcriptional activation in vitro from a template carrying the Mdm2 promoter assembled into chromatin. We identify the first 107 amino acids of p53, which carries the tripartite activation domain, as those involved in the CR2 interaction. Finally, we show that p53 and the HTLV-I Tax protein compete for interaction with CR2 in vitro, possibly contributing to the previously reported Tax repression of p53 transcription function (19,(31)(32)(33).
In the experiments presented in Fig. 3A, full-length p53 and the amino-terminal fragment of p53 (amino acids 1-107) were transcribed and translated using the TNT Quick-Coupled in vitro transcription/ translation system (Promega). Full-length p53 and p53 (aa 1-107) were labeled with [ 35 S]methionine during the in vitro transcription/translation reaction. Because of differences in methionine incorporation (12 methionines in full-length p53 versus four methionines in the aminoterminal p53 fragment), we used three times the amount of the aminoterminal fragment of p53 in vitro transcription/translation product (6 l) in the GST pull-down assay. The in vitro transcription/translation products were incubated with 10 pmol of each GST fusion protein. The amino-terminal fragment of p53 (aa 1-107) was cloned by PCR amplification of the full-length, wild-type p53 cDNA (p53-H-19) (7). The PCR product was inserted into the NdeI/BamHI site of pET15b (Novagen).
Drosophila core histones were purified as previously described (36). The yeast NAP-1 cDNA (37) was cloned into pGEX-2T (Amersham Biosciences, Inc.), and the GST-yNAP-1 fusion protein was expressed in Escherichia coli and purified by glutathione-agarose affinity chromatography and Q-Sepharose. We coexpressed FLAG-tagged ISWI and Acf1 from baculovirus and purified the complex by anti-FLAG affinity batch binding and elution as previously described (38). His 6 -tagged wild-type p300 and p300⌬SRC proteins were expressed from recombinant baculoviruses and purified as previously described (26).
Transcription Template-The p53-responsive Mdm2 P2 G-less plasmid DNA used in the assembly reactions carried the two p53 response elements from the Mdm2 P2 intragenic promoter (39). Briefly, a 567-bp fragment carrying the p53 response elements, TATA sequence, and start site was PCR-amplified and cloned immediately upstream of a 190-bp G-less cassette. The identity of the Mdm2 P2 G-less construct was confirmed by restriction analysis.
In Vitro Transcription Assay-The supercoiled Mdm2 P2 G-less plasmid was assembled into chromatin using GST-yNAP-1, Acf1/ISWI, and Drosophila histones, at a 1.1:1.0 histone:DNA ratio. Following chromatin assembly, preinitiation complexes were formed on the equivalent of 200 ng of the plasmid DNA in the absence or presence of p53 (160 nM), p300 (20 nM), and/or p300⌬SRC (20 nM). All reactions contained 100 M acetyl CoA (United States Biochemical). Nuclear extract (70 g) (42), prepared from CEM cells (a mutant p53 human T lymphocyte cell line) was added immediately following the addition of the activator and/or coactivator. Following a 60-min preincubation reaction at 30°C, RNA synthesis was initiated by the addition of 250 M ATP, GTP, CTP, and 12 M UTP plus 0.8 M [␣-32 P]UTP (3000 Ci/mmol, PerkinElmer Life Sciences). Transcription reactions were processed and analyzed as previously described (43). Molecular weight markers (radiolabeled HpaIIdigested pBR322) were used to estimate the size of the RNA products.
Mammalian Expression Plasmids, Cell Culture, and Transient Cotransfection Assays-Jurkat T-cells (a p53-negative human T lymphocyte cell line) were cultured in Iscove's modified Dulbecco's medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, and penicillin-streptomycin. For transient cotransfection assays, cells were grown to a density of 10 6 cells/ml and transfected with LipofectAMINE (Invitrogen) and a constant amount of DNA for 5 h. The cells were allowed to recover for 19 h before harvest. Cells were lysed, and luciferase activity was measured using the Dual-Luciferase reporter assay system with a Turner Designs model TD 20-e luminometer. Luciferase activity was normalized to a pRL-TK vector (Promega), which encodes the Renilla luciferase from the HSV-TK promoter, as an internal control.

RESULTS
Identification of the p53-interacting CR2 Region of CBP by GST Pull-down Assay-We began this study by testing three large regions of CBP spanning the carboxyl-terminal half of the coactivator (Fig. 1A). Each of these CBP regions were cloned and expressed as GST fusion proteins and tested in GST pull-p53 Interacts at Multiple Sites on CBP/p300 down assays with purified, recombinant, full-length p53. p53 binding to the KIX domain (aa 588 -683) served as a positive control (Fig. 1B, lane 6). We found that p53 bound strongly to only one of the three carboxyl-terminal regions of CBP (Fig. 1B,  lane 4). This region, which we call carboxyl-terminal region 2 (CR2), encompasses CBP amino acids 1894 -2221. Consistent with previous studies (24,25), we found that p53 also bound to C/H3 (TRAM/TAZ2), because our carboxyl-terminal region 1 protein (CR1, aa 1514 -1894) encompasses this domain (Fig.  1B, lane 3; see Fig. 1A). However, p53 binding to this region was significantly less than that observed with either CR2 or KIX in our GST pull-down assay. These data indicate that, at least in our assay, p53 interacts most strongly with the region of CBP, encompassing amino acids 1894 -2221. This observation was confirmed using the yeast two-hybrid assay (data not shown). Interestingly, the CR2 domain (aa 1894 -2221) corresponds closely to the steroid receptor coactivator 1 (SRC1)interacting domain of CBP (aa 1982-2163) (27,28,47). SRC1 is a prominent member of a family of coactivators that utilize CBP/p300 to mediate transcriptional activation of nuclear hormone receptors (48,49).
CBP and p300 are highly homologous proteins, although their precise role in the regulation of gene expression mediated by either protein is unclear. Because CBP and p300 share roughly 50% homology within the CR2 region, we were interested in testing whether p53 also recognizes the CR2-like domain found in p300. To address this question, we cloned the p300 CR2 region (aa 1970 -2193) fused to GST and tested p53 binding in a GST pull-down assay. Fig. 1C shows that p53 binds comparably to the CR2 regions from both CBP and p300 (lanes 3 and 4).
Fine Mapping and Mutational Analysis of the Minimal p53interacting Region of CR2-We were next interested in mapping the minimal region of CR2 competent for p53 interaction. For these studies, we analyzed p53 binding to a series of deletion mutants of CR2 using the GST pull-down assay. Progressive carboxyl-terminal deletions of GST-CR2 revealed that amino acid 2150 represents the carboxyl-terminal border competent for wild type interaction with p53 (GST-CR2 aa1894 -2150 ) ( Fig. 2A, lanes 3-6). Progressive amino-terminal deletions of CR2 revealed that amino acid 2055 represents the aminoterminal border competent for wild type interaction with p53 (GST-CR2 aa2055-2150 ) (Fig. 2A, compare lane 3 with lanes 7-10).
These data show that the minimal region of CR2 competent for interaction with p53 resides within a 95-amino acid fragment, bordered by residues 2055 and 2150 ( Fig. 2A, lane 10). This region precisely overlaps with the minimal CBP sequence (aa 2058 -2130) required for interaction with SRC-1 (28).
To identify critical amino acids within CR2 responsible for interaction with p53, we prepared and characterized a series of double point mutations. The amino acids targeted for mutagenesis were chosen based on conservation between CBP and p300 as well as conservation between the mouse and human CBP. We targeted specific leucine residues within a region that forms amphipathic ␣-helices (and thus, possibly, protein⅐protein contacts). The selected residues were changed to alanines, to minimize effects on secondary and tertiary structure. Four CR2 aa2003-2212 constructs were prepared, each carrying two point mutations as follows: F2101A/I2102A; L2068A/L2071A; L2072A/L2075A; L2140A/L2143A. Fig. 2B shows that only the double point mutant L2068A/L2071A, which resides within the first of the three ␣-helices, had a significant effect on p53 binding (lane 5). These data provide further evidence for the specificity of the p53⅐CR2 interaction. A summary of the p53 interactions with the various CR2 constructs from CBP and p300 is shown in Fig. 2C.
Fine Mapping of the Minimal CR2-interacting Region of p53-Preliminary yeast two-hybrid studies suggested that the site of CBP interaction resides within the first 112 amino acids of p53 (data not shown). Based on this observation, we performed GST pull-down assays using an amino-terminal fragment of p53. In vitro transcribed-translated 35 S-labeled full-length p53 and a 35 S-labeled amino-terminal truncation of p53 (aa 1-107) were tested for their ability to bind the CR2 domain of CBP. Glutathione beads were bound with GST-CR2 aa1894 -2221 or GST-C/H1-KIX aa302-683 and then incubated with the 35 S-labeled in vitro translation products, and the resulting protein⅐protein interactions were detected by Phos-phorImager analysis. Fig. 3A shows that both the full-length and the amino-terminal p53 fragment binds to CR2 (lanes 5 and 8). Although the binding of the amino-terminal truncation fragment to CR2 is clearly specific, the binding appears to be reduced relative to the full-length protein, possibly because the amino-terminal domain in isolation is not structurally identical to the analogous region in the full-length protein. This result is consistent with the observation that amino and carboxyl-ter- FIG. 1. p53 binds strongly to the CR2 domain of CBP. A, schematic representation of the 2441-amino acid cellular coactivator CBP. The regions tested for p53 interaction are indicated. B, p53 binds to the CR2 domain in vitro. Full-length, purified, recombinant p53 (10 pmol) was incubated with GST alone (lane 2) or the indicated GST-carboxyl-terminal region fusion proteins (10 pmol each) (lanes 3-5). As a positive control, we also tested p53 binding to GST-KIX aa588 -683 (lane 6). p53 was detected using an anti-p53 antibody. Onput p53 (5%) is shown (lane 1). Bound p53 and protein molecular weight standards are indicated. C, p53 binds equally well to the CR2 domains derived from CBP and p300. Purified p53 (20 pmol) was incubated with GST alone or the GST-CR2 region from CBP (aa 2003-2212) or p300 (aa 1970 -2193) (20 pmol each). p53 was detected using an anti-p53 antibody. Onput p53 (5%) is shown (lane 1). Bound p53 is indicated.
p53 Interacts at Multiple Sites on CBP/p300 minal interactions in p53 are important for p53 function (50).
To determine whether a previously characterized minimal activation domain of p53 may be involved in the interaction with CR2, we introduced a double point mutation (L22Q/W23S) into this region (19) and tested the ability of the purified mutant protein to bind the minimal CR2 domain (aa 2055-2150). Mutation of these residues has previously been shown to have a dramatic effect on p53 transcription function (51). Fig. 3B shows the results of a GST pull-down assay where we tested the binding of purified wild type and mutant p53 proteins to both CR2 and KIX. Surprisingly, the double point mutation in this minimal p53 activation domain did not have a significant effect on p53 binding to the CR2 domain (Fig. 3B, lanes 7 and 8). As we have previously reported, the double point mutations did significantly reduce p53 binding to the KIX domain (Fig. 3B, lanes 5 and 6) (19). These data suggest that other amino acids in the p53 tripartite activation domain likely participate in CR2 binding.
EMSA Studies on the p53⅐CR2 Interaction-As an alternate method to characterize the p53⅐CR2 interaction, we utilized the electrophoretic mobility shift assay (EMSA). We were inter-ested in determining whether CR2 aa2055-2150 could form a ternary complex with p53 bound to its consensus DNA recognition element. Fig. 4 shows that titration of the purified CR2 domain into p53-containing binding reactions decreased the mobility of the p53⅐DNA complex (lanes 5-7 and 12-14). The change in mobility suggested that CR2 was stably incorporated into the complex. Interestingly, we did not observe a change in the mobility of the p53⅐DNA complex in the presence of increasing amounts of the C/H3-containing CR1 domain (Fig. 4, lanes  2-4). The CR3 domain also had no effect on the migration of the p53⅐DNA complex, consistent with our previous observations (Fig. 4, lanes 8 -10). As a positive control, we titrated the KIX domain of CBP into the p53⅐DNA binding reactions, and compared the ternary complex formation with that observed with CR2. Fig. 4 shows that both CR2 and KIX similarly decreased the mobility of the p53⅐DNA complex (lanes 12-17). The specificity of the DNA binding activity of p53 was confirmed by competition assays using the p53 consensus sequence and antibody supershift assays (data not shown). Finally, CR2, as well as the other CBP domains, did not bind DNA in the absence of  4 -10). As a positive control, p53 binding to full-length GST-CR2 aa1894 -2221 was also tested (lane 3). p53 was detected using an anti-p53 antibody. Onput p53 (5%) is shown (lane 1). Bound p53 and protein molecular weight standards are indicated. B, p53 is defective for an interaction with the CR2 double point mutant L2068A/L2071A. Purified p53 (25 pmol) was assayed for its ability to bind to GST alone or the GST-CR2 aa2003-2212 double point mutants: F2101A/I2102A, L2068A/L2071A, L2072A/L2075, or L2140A/L2143A. (25 pmol). p53 binding to wild-type GST-CR2 aa2003-2212 was tested as a positive control (lane 3). Onput p53 (5%) is shown (lane 1). Bound p53 and protein molecular weight standards are indicated. C, summary of the p53⅐CR2 interactions.
p53 Interacts at Multiple Sites on CBP/p300 p53 (data not shown). The EMSA studies presented here were performed with unacetylated p53, because we have observed no significant differences in the DNA binding activity, or CR2 binding activity, between the CBP/p300-acetylated and unacetylated forms of the protein (data not shown).
Functional Significance of the p53⅐CR2 Interaction in Vitro and in Vivo-To test whether the p53⅐CR2 interaction participated in CBP/p300-mediated p53 transcriptional activation, we examined p53 transcription function in the presence of exogenous wild type p300 or a mutant form of p300 that carries a deletion of the SRC1 domain (26). We selected p300 for these studies, because p53 interacts similarly with the CR2 region of both CBP and p300, and p300 coactivator function has been well characterized in vitro (18,26,41,52,53). To measure coactivator-mediated p53 transcriptional activation, we used a DNA template containing a 567-bp fragment from the Mdm2 intragenic P2 promoter, driving synthesis of a 190-nucleotide guanine-less transcript. This Mdm2 P2 fragment carries two p53 binding sites upstream of the core promoter (54). We chose to analyze transcription in a chromatin context, because several studies have found that analysis of p300 coactivator function in vitro requires nucleosomal templates (18,41,52,53). Chromatin assembly of the p53-responsive G-less template was performed using the recombinant Drosophila assembly proteins Acf1/ISWI, GST-yNAP-1, and purified Drosophila core histones, as previously described (38). These assembly proteins are sufficient for the ATP-dependent formation of evenly spaced nucleosomal arrays (38,55). Fig. 5A shows a DNA topological analysis demonstrating the assembly of native Drosophila core histones onto the p53-responsive G-less template (lanes 3-9). In the presence of the assembly factors, increasing ratios (w/w) of the core histones to the DNA produced a concomitant increase in DNA supercoiling, indicating that nucleosomes were deposited onto the template. The figure shows that a histone/DNA ratio of 1.1:1.0 (w/w) fully assembled the DNA template into chromatin (lane 9); this ratio was used in subsequent in vitro transcription assays. Micrococcal nuclease digestion of the reconstituted chromatin template indicated the presence of evenly spaced nucleosomes on the DNA (data not shown).
We performed in vitro transcription assays on this p53-responsive chromatin template using nuclear extracts from CEM cells (a mutant p53 human T lymphocyte cell line) as a source of basal transcription factors and RNA polymerase. All experiments were performed in the presence of acetyl CoA and in the presence or absence of exogenous p53 and/or p300 or p300⌬SRC. The activator, coactivators, and nuclear extract were added following chromatin assembly. We used unacetylated p53 in this experiment, because a recent study has shown that the unacetylated form of p53 is sufficient for in vitro transcription from a chromatin-assembled template (18). Fig.  5B shows that the addition of purified recombinant p53 alone did not activate transcription from the Mdm2 promoter (lane 3). However, addition of purified recombinant p300 together with p53 produced a significant increase in RNA synthesis from the Mdm2 promoter (18-fold, Fig. 5B, lane 4). Under these same conditions, addition of p300⌬SRC, which carries an inframe deletion of CR2 (aa 2042-2157), activated transcription only 5-fold from these templates; a 3.6-fold reduction in p300 coactivator function (Fig. 5B, lane 5). The absence of p53 reduced both wild type and mutant p300-stimulated transcription, indicating that optimal coactivator function required the presence of p53 (Fig. 5B, lanes 6 and 7). Fig. 5C demonstrates that both the wild type and mutant p300 proteins similarly acetylate p53, confirming that both proteins were equivalently functional with respect to acetyltransferase activity. Furthermore, p300⌬SRC is fully functional for acetylation of free histones as well as nucleosomal core histones (26).
Finally, to determine the functional role of the p53⅐CR2  (lanes 3 and 4), GST-KIX aa588 -683 (lanes 5 and 6), or GST-CR2 aa2055-2150 (lanes 7 and 8) (30 pmol). p53 was detected using an anti-His 6 antibody. Onput wild type and mutant p53 proteins (5%) are shown (lanes 1 and 2). Bound p53 and protein molecular weight standards are indicated. Because CR2 does not have intrinsic activation properties, p53 binding to free CR2 should block the p53 interaction with endogenous (or transfected) CBP/p300 and, therefore, have a dominant negative effect on p53 transcriptional activity. The left panel of Fig. 5D shows that titration of the expression plasmid for CR2 repressed p53 transcriptional activation in a dose-dependent manner (lanes [3][4][5]. We also measured the effect of CR2 on p53 transcriptional activation in the presence of an expression plasmid for CBP. The right panel shows that, in the presence of cotransfected full-length CBP, CR2 again repressed p53-mediated transcription. As expected, the presence of the CBP expression plasmid partially rescued the observed CR2 repression (Fig. 5D, compare lanes 8 and 10). Addition of either the CR2 or the CBP expression plasmids in the absence of p53 had no effect on pG13-luc reporter activity (lane 11 and data not shown). These data support a role for the CR2 domain of CBP/p300 in p53 transcription function in vivo.
HTLV-I Tax and p53 Compete for CR2 Binding in Vitro-Several studies have previously reported that the human T-cell leukemia virus Tax protein represses p53 transcription function (19,(31)(32)(33). Several recent studies suggest that this transcriptional repression may occur as a consequence of direct competition for binding to common regions of CBP/p300, thus compromising p53 promoter recruitment of the coactivator (19, 56 -58). Recently, we reported that the HTLV-I Tax protein binds to the CR2 domain of CBP and p300, and identified CBP aa 2003-2212 as the minimal region competent for interaction with Tax (29). Based on these observations, we hypothesized that the binding of Tax and p53 to CR2 might be mutually exclusive. To directly test this hypothesis, we examined whether increasing concentrations of purified recombinant p53 can displace Tax from CR2 in vitro. Glutathione beads were bound with GST-CR2 aa2003-2212 then incubated with a constant amount of Tax and increasing amounts of p53. The resulting protein⅐protein interactions were detected by Western blot analysis using a solution containing antibodies against both Tax and p53. Fig. 6 shows that increasing amounts of p53 reduced Tax binding to CR2, with a concomitant increase in p53 binding (lanes 3-5). This observation was corroborated in the reciprocal experiment, where increasing concentrations of Tax similarly displaced p53 from CR2 (Fig. 6, lanes 7-9). This showing the Mdm2 P2 G-less transcription template assembled with Drosophila core histones in the presence of dAcf1/ ISWI and GST-yNAP-1. The DNA topoisomers were resolved on an agarose gel, and the DNA was stained with Sybr Gold (Molecular Probes). The supercoiled (S), relaxed (R), and nicked DNA populations, and the histone/DNA ratio, are indicated. B, transcriptional activation on Mdm2 P2 G-less chromatin templates was analyzed in the presence of p53 (160 nM, lanes 3-5), wild-type p300 (20 nM , lanes 4 and 6), and/or p300⌬SRC (20 nM, lanes 5 and 7). Molecular weight size markers, recovery standard, and full-length G-less transcripts are indicated. The fold activation indicated was calculated relative to transcription in the presence of exogenous p53 (lane 3). C, wild-type p300 and p300⌬SRC acetylation of recombinant p53. p53 (235 nM) was acetylated by p300 (lane 2) and p300⌬SRC (lane 3) (30 nM each) in the presence of [ 14 C]acetyl CoA (100 pmol, 57mCi/mmol). p53 acetylation and p300 autoacetylation are indicated. D, the CR2 domain represses p53-activated transcription in vivo. In the left panel, the p53-responsive pG13-Luc reporter plasmid (400 ng) was cotransfected with an expression plasmid for p53 (200 ng, lanes [2][3][4][5], and increasing amounts of an expression plasmid for CMV-CR2 (200, 400, and 800 ng, respectively, lanes [3][4][5]. In the right panel, the p53-responsive pG13-Luc reporter plasmid (400 ng) was cotransfected with an expression plasmid for p53 (200 ng, lanes 7-10). p53 transactivation was assayed in the presence or absence of expression plasmids for fulllength CBP (400 ng, lanes 9 and 10) and CMV-CR2 (800 ng, lanes 8 and 9), as indicated. As a control, cotransfection of an expression plasmid for CMV-CR2 in the absence of p53 is shown (lane 11). The values shown are the mean -fold activation (in triplicate) Ϯ S.D.
p53 Interacts at Multiple Sites on CBP/p300 result is consistent with our observations that both Tax (29) and p53 bind to a similar, overlapping minimal domain of CR2 (aa 2003-2212 and aa 2055-2150, respectively), and that the CR2 double point mutant L2068A/L2071A reduces interaction with both proteins. DISCUSSION In this report, we show that p53 interacts strongly with the carboxyl-terminal region 2 (CR2) of CBP, located between amino acids 2055 and 2150. We also demonstrate that p53 interacts with the corresponding CR2 region of p300, located between amino acids 1970 and 2193. The CR2 region is distinct from the C/H3 domain, the only previously identified region within the carboxyl-terminal half of CBP that has been shown to interact with p53 (24,25). In our assays, p53 interacted more strongly with CR2 than with the region of CBP that encompasses the C/H3 domain (CR1; aa 1514 -1894). We mapped the minimal CR2 region of CBP required for strong interaction with p53 to amino acids 2055-2150. This 95-amino acid minimal CR2 sequence corresponds precisely with the SRC1-interacting domain of CBP, which has been mapped to amino acids 2058 -2130 (28). This domain also corresponds to the CBP region involved in binding to IRF-3 and HTLV-I Tax (29,30). We show that a CR2 double point mutation (L2068A/L2071A), which specifically disrupts the first of the three ␣-helices that resides within this region (30), reduces interaction with p53.
The amino-terminal 107 amino acids of p53 at least partially participate in protein⅐protein interaction with CR2. This is consistent with our observation that CR2 binds well to fulllength p53⅐DNA complexes, suggesting that the DNA binding and tetramerization domains are not involved in CR2 recognition. Previous studies have indicated that the p53 activation domain participates in binding to the KIX domain (19) and C/H3 domain of CBP (20,23). We tested whether a minimal region of the p53 activation domain might interact with CR2 using a double point mutant of p53 (L22Q/W23S). Although we did not observe a significant decrease in the CR2⅐p53 interaction using this mutant, the activation domain of p53 is tripartite, and extends through the first 100 amino acids of the protein. Therefore, other amino acids that reside within this amino-terminal region of p53 likely participate in CR2 complex formation.
Our in vitro transcription studies clearly show that p53 interaction with the p300 CR2 domain is relevant to p53 transcription function. The addition of p300 and p53 strongly stimulated RNA synthesis from the p53-responsive Mdm2 P2 pro-moter assembled into chromatin. However, the p300 deletion mutant p300⌬SRC was significantly reduced in its ability to mediate coactivator function. Our observation that p300⌬SRC retained partial coactivator function in p53-mediated transcription may reflect the ability of p53 to recruit CBP/p300 to the Mdm2 promoter via interaction with other coactivator domains (such as KIX and/or C/H3) (19,20,23,25,59,60). The in vitro transcription result was corroborated using transient transfection assays, confirming a functional role for the CR2 domain in mediating p53 transcription function in vivo.
Previous studies have shown that the HTLV-I Tax protein inhibits many of the tumor suppressor functions of p53 (31)(32)(33)(61)(62)(63). Several recent studies suggest that this may occur through competition for CBP/p300 (19, 56 -58). We have recently shown that, like p53, Tax also recognizes the CR2 region of CBP/p300 (29), raising the possibility that both Tax and p53 bind mutually exclusively to this region. Using a competition binding assay, we directly show that Tax specifically disrupts the p53⅐CR2 interaction, providing further evidence for coactivator competition between these two proteins. It appears that both proteins recognize the same surface structure of CR2, because p53 and Tax are unable to bind the CR2 domain that harbors the double point mutation (L2068A/L2071A) (shown in Fig. 2B, and Ref. 29). Together, these data provide further evidence for a model of Tax repression of p53 transcription function mediated through direct competition, at multiple sites, for CBP/p300. This coactivator competition between Tax and p53 may contribute to the molecular mechanism of HTLV-I-associated malignant transformation.
FIG. 6. Tax and p53 binding to CR2 is mutually exclusive. A constant amount of purified Tax (25 pmol) was incubated with GST alone, or GST-CR2 aa2003-2212 (25 pmol each) in the presence of increasing amounts of purified p53 (25 and 100 pmol, lanes 4 and 5). In the reciprocal experiment, a constant amount of p53 (25 pmol) was incubated with GST alone, or GST-CR2 aa2003-2212 (25 pmol each) in the presence of the increasing amounts of purified Tax (25 and 100 pmol, lanes 8 and 9). Bound proteins were detected by Western blot analysis using both anti-Tax and anti-p53 antibodies. Bound Tax, bound p53, and protein molecular weight standards are indicated.
p53 Interacts at Multiple Sites on CBP/p300