Human Papilloma Virus 16 E6 Oncoprotein Inhibits Retinoic X Receptor-mediated Transactivation by Targeting Human ADA3 Coactivator*

, The expression of human papillomavirus (HPV) E6 oncoprotein is causally linked to high-risk HPV-associ-ated human cancers. We have recently isolated hADA3, the human homologue of yeast transcriptional co-acti-vator yADA3, as a novel E6 target. Human ADA3 binds to the high-risk (cancer-associated) but not the low-risk HPV E6 proteins and to immortalization-competent but not to immortalization-defective HPV16 E6 mutants, suggesting a role for the perturbation of hADA3 function in E6-mediated oncogenesis. We demonstrate here that hADA3 directly binds to the retinoic X receptor (RXR) (cid:1) in vitro and in vivo . Using chromatin immunoprecipitation, we show that hADA3 is part of activator complexes bound to the native RXR response elements within the promoter of the cyclin-dependent kinase inhibitor gene p21. We show that hADA3 enhances the RXR (cid:1) -mediated sequence-specific transactivation of retinoid target genes, cellular retinoic acid-binding

The human papilloma viruses are causally linked to more than 90% of the cases of cervical cancer (1,2). Numerous studies have defined the critical roles of two HPV 1 oncogenes E6 and E7 in oncogenesis. These oncogenes are invariably expressed in HPV-associated carcinomas and cell lines derived from these cancers (3,4). Intact E6 and E7 open reading frames are required for in vitro immortalization of human epithelial cells by the HPV genome, and E6 and/or E7 genes are sufficient to immortalize these cells (5)(6)(7)(8).
A crucial aspect of the oncogenic mechanism of E6 and E7 involves their ability to inactivate two key cell cycle checkpoint proteins, p53 and retinoblastoma protein, respectively (9,10). While E6-induced loss of p53 strongly correlates with E6-induced cellular transformation, recent studies have identified additional cellular targets of E6 that are likely to play important roles in HPV oncogenesis (11,12). Defining the roles of these novel E6 targets is of substantial importance to fully understand the mechanisms of HPV-mediated oncogenesis.
Using the yeast two-hybrid interaction system, we recently identified hADA3 as a novel E6-binding protein (13). hADA3 is the human homologue of yeast transcriptional co-activator yADA3 (alteration/deficiency in activation 3). We have demonstrated that hADA3 binds to the high-risk (cancer-associated) but not to the low-risk HPV E6 proteins, and to immortalization-competent but not to immortalization-defective HPV16 E6 mutants, implying a role for E6-induced hADA3 inactivation in oncogenic transformation.
Genetic studies in yeast have demonstrated that ADA3 functions as a critical component of coactivator complexes that link transcriptional activators, bound to specific promoters, to histone acetylation, and basal transcriptional machinery (14 -16). The core components of this complex include the adapter proteins ADA3 and ADA2 and a histone acetylase GCN5 (general control non-repressed 5) (17). Importantly, hADA3 exists as a component of a yeast ADA-like complex that includes hADA2 and hGCN5, indicating that the functional roles of ADA complex are evolutionarily conserved (18). Studies of mammalian retinoic X receptor (RXR) and growth hormone receptor when expressed in yeast have shown a requirement for the components of yeast ADA complex including the yADA3 gene product (19,20). However, a functional role of hADA3 in nuclear hormone receptor transactivation in mammalian cells has not yet been defined.
Given the important role of the inactivation of p53 transactivation in E6-induced oncogenesis, our earlier studies focused on the novel role of hADA3 as a p53 coactivator (13). Another group has also reported a coactivator function of hADA3 for p53 (21). Our studies also established that the interaction of hADA3 with E6 led to its ubiquitin-mediated degradation (13). In view of the strong yeast genetic data for a potential role of hADA3 as a coactivator for nuclear hormone receptors, our isolation of hADA3 as an E6-binding protein raised the possibility that E6 may inactivate the mammalian nuclear hormone receptor function by targeting hADA3. In this study, we exam-ined if hADA3 associates with and functions as a coactivator for human retinoic X receptors RXR␣ and retinoic acid receptor ␣ (RAR␣)-mediated transactivation, and whether E6 abrogates this function of ADA3. We report here that hADA3 directly binds to RXR␣ but not to RAR␣, allowing its in vitro and in vivo association with RXR homodimers as well as RXR-RAR heterodimers. Using the chromatin immunoprecipitation (ChIP) assay, we demonstrate that hADA3 is a component of the RXR␣-activator complexes bound to native promoter of the cyclin-dependent kinase inhibitor p21. Furthermore, hADA3 enhances the RXR␣-mediated transactivation of target genes, cellular retinoic acid-binding protein II (CRBPII) and p21. Most significantly, we show that HPV16 E6 inhibits RXR␣dependent transactivation. Our results identify RXR␣ as a novel HPV E6 oncoprotein target and implicate the deregulation of retinoid receptor function in HPV oncogenesis.
In Vitro Binding between hADA3 and Retinoid Receptors-The pSG5-RXR and pSG5-RAR expression constructs were used as templates in a rabbit reticulocyte lysate-based (TNT rabbit reticulocyte lysate system; Promega, WI) or an Escherichia coli S-30 extract-based (PROTEIN-script TM -PRO: Ambion, TX) coupled in vitro transcriptiontranslation system in the presence of [ 35 S]methionine to generate 35 S-labeled human RXR and RAR proteins. A GST fusion protein of hADA3 was purified using the glutathione-Sepharose affinity beads from lysates of E. coli transformed with pGEX-4T-1-hADA3. Aliquots of 35 S-labeled proteins were incubated at 4°C with 1 g of GST or GST-hADA3 noncovalently bound to glutathione beads in 300 l of lysis buffer (100 mM Tris, pH 8.0, 100 mM NaCl, 0.5% Nonidet P40) for 2 h, in the absence or presence of 9-cis-retinoic acid (9-cis-RA). The beads were washed with lysis buffer and bound proteins were resolved by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE), and visualized by fluorography.
Co-immunoprecipitation of ADA3 with RXR␣ or RAR␣-5 ϫ 10 5 Saos2 cells were plated in phenol red-free ␣-MEM medium supplemented with 10% charcoal-treated fetal bovine serum. After 48 h, the cells were transfected with the pCR3.1-hADA3 plasmid with or without pSG5 constructs encoding the human RXR␣ or RAR␣, using the Fu-GENE 6 reagent (Roche Molecular Biochemicals). After 24 h, cells were either mock-treated or treated with 100 nM 9-cis-RA. Following a 24-h treatment, cell lysates were prepared in a lysis buffer (100 mM Tris, pH 8.0, 100 mM NaCl, 0.5% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride) and precleared with protein A-Sepharose beads (Amersham Biosciences). 2-mg aliquots of lysate protein were subjected to immunoprecipitation with anti-RXR or anti-RAR␣ antibody (Santa Cruz Biotechnology, Santa Cruz, CA) followed by immunoblotting with an anti-FLAG antibody (M2, Sigma) for detection of associated hADA3. Enhanced chemiluminescence (ECL, Amersham Biosciences) was used for detection.
Northern Blot Analysis of p21 mRNA Expression-Subconfluent 76R-30 cells grown in DFCI-1 medium were treated with different concentration of 9-cis-RA for 24 h. Total cellular RNA was isolated using the Trizol reagent (Invitrogen). The 32 P-labeled p21 and 36B4 cDNA probes were used for Northern blotting using standard procedures (24).
ChIP Assay to Detect the Native p21 Promoter-bound ADA3, RXR␣, and RAR␣-Subconfluent 76R-30 cells grown in DFCI-1 medium were treated with 100 nM 9-cis-RA for 24 h, and native protein-DNA complexes were cross-linked by treatment with 1% formaldehyde for 15 min. The ChIP assay was carried out as reported earlier (25). Briefly, equal aliquots of isolated chromatin were subjected to immunoprecipitation with a rabbit anti-hADA3 antibody (13), or polyclonal antibodies against human RXR or RAR␣. The DNA associated with specific immunoprecipitates or with negative control preimmune serum was isolated and used as a template for the PCR to amplify the p21 promoter sequences containing the retinoid response element. The primers used were: 5Ј-primer, 5Ј-GAGGTCAGCTGCGTTAGAGG-3Ј; 3Ј-primer, 5Ј-GCTCCCATCTACCTCACACC-3Ј. As a specificity control, the glyceraldehyde-3-phosphate dehydrogenase promoter was amplified from the same templates using the following primers: 5Ј-primer, AAAAGCGGG-GAGAAAGTAGG; 3Ј-primer, CTAGCCTCCCGGGTTTCTCT.
Retinoid Receptor-dependent Transactivation of Luciferase Reporters-5 ϫ 10 5 Saos2 or 76R-30 cells were plated per 100-mm dish in phenol red-free ␣-MEM medium supplemented with 10% charcoaltreated fetal bovine serum (Saos2 cells) or in DFCI-1 medium (76R-30 cells) for 48 h and transfected with the expression plasmids, as indicated in the figure legends. Each dish also received 20 ng of SV40 Renilla luciferase reporter (pRL-SV40) to correct for differences in transfection. The total amount of DNA was kept constant by adding the vector DNA. After 24 h, the transfected cells were either mock-treated or treated with 100 nM 9-cis-RA for an additional 24 h. The luciferase activity was measured using a dual-luciferase kit (Promega, WI). Equal aliquots of cell lysates (normalized based on Renilla luciferase activity) were resolved by SDS-PAGE and immunoblotted with anti-RXR, anti-RAR␣, or anti-ADA3 antibodies to assess protein expression.

RESULTS AND DISCUSSION
There is an emerging consensus that cellular transformation by viral oncoproteins disrupts multiple cellular pathways that control cell proliferation, differentiation, migration, and other critical cellular traits. Identification and characterization of cellular targets of human cancer-associated viral oncoproteins is therefore of great interest in cancer research as well as in basic cell biology. We previously identified human ADA3, homologue of the yeast ADA3 transcriptional coactivator, as a novel HPV E6 target (13). Genetic analyses have shown a requirement for yeast ADA3 for transcriptional activation by human RXR and other nuclear hormone receptors when these were expressed in yeast (19). These findings prompted us to examine if hADA3 functions as a coactivator for retinoid receptors in human cells and whether E6 interaction with hADA3 would allow it to perturb the retinoid receptor function.
RXR␣, but Not RAR␣, Directly Binds to hADA3 Protein in Vitro-As a first step, we examined if hADA3 indeed interacts with and coactivates the retinoid receptor-mediated transactivation in mammalian cells. The ligand binding domain (residues 266 -455) of RXR␣, which is known to mediate binding to coactivators (26), was translated in vitro in rabbit reticulocyte lysates or E. coli lysates, and its binding to GST-ADA3 was assessed in the absence or presence of the RXR␣ ligand 9-cis-RA. Whereas RXR␣ failed to bind to GST, as expected, a substantial level of binding to GST-hADA3 was observed. When RXR␣ was translated in rabbit reticulocyte lysates, substantial binding to GST-hADA3 was seen in the absence of retinoic acid; however, this binding was further enhanced by adding the ligand (Fig. 1A, lanes 3 and 4, upper panel). Notably, the E. coli lysate-translated RXR␣ protein bound to GST-ADA3 only in the presence of the ligand (Fig. 1A, upper panel, compare lanes 3 and 4 with lanes 7 and 8). It remains possible that these differences are due to the presence of retinoids or some modification of the RXR␣ receptors in the rabbit reticulocyte lysates. These experiments established that hADA3 can directly interact with the RXR␣ protein in a ligand-dependent manner.
Next, we examined if hADA3 can also interact with the RAR␣ protein as such or in the presence of RXR␣. As RXR␣ and RAR␣ proteins are of a similar size to be able to unambiguously visualize RAR␣ and RXR␣ proteins in the same gel, we used the ligand-binding domain of RXR␣ in these experiments. As shown in Fig. 1B, RAR␣ did not show a detectable level of direct interaction with hADA3 (Fig. 1B, lane 4, upper panel). However, when the ligand-binding domain of RXR␣ was included in the binding reaction, both RAR␣ and RXR␣ could be pulled down with GST-hADA3 (Fig. 1B, lane 5, upper panel). Similar binding was observed when full-length RXR␣ was used (data not shown). Coomassie Blue staining of gels confirmed the presence of respective GST fusion proteins in the binding reactions ( Fig. 1, A and B, lower panels). These results indicate that while RAR␣ does not directly interact with hADA3, it can do so when present as a RXR/RAR heterodimer.
RXR␣ and RAR␣ Associate with hADA3 Protein in Vivo-In view of the in vitro binding, we used co-immunoprecipitation analyses of transfected proteins to assess whether hADA3 interacts with RXR␣ and RAR␣ in Saos2 cells. Anti-RXR immunoprecipitates from transfected cell lysates were subjected to anti-FLAG immunoblotting to detect co-immunoprecipitated FLAG-tagged hADA3. No hADA3 was detected in anti-RXR immunoprecipitates if cells had been transfected only with hADA3 or RXR␣ constructs or when hADA3 plus RXR␣-transfected cells had not been treated with 9-cis-RA ( Fig. 2A, lanes  1-3, right panel). In contrast, hADA3 was readily detectable in anti-RXR immunoprecipitates from hADA3 plus RXR␣-transfected cells treated with 9-cis-RA ( Fig. 2A, right panel, lane 6). Anti-RXR immunoblotting of anti-RXR immunoprecipitates demonstrated the immunoprecipitation of RXR in the appropriate lanes ( Fig. 2A, right lower panel), and anti-FLAG and anti-RXR blotting of whole cell lysates confirmed the expected expression of hADA3 and RXR ( Fig. 2A, left panels).
Given the indirect, RXR␣-mediated binding of RAR␣ to hADA3 in vitro, we also examined if RAR␣ can associate with hADA3 in the presence of RXR␣, the strategy used similar to that used for RXR␣, except that immunoprecipitates were carried out using an anti-RAR␣ antibody. A substantial associa-tion between RAR␣ and hADA3 was observed if RXR was co-expressed, and cells were treated with 9-cis-RA (Fig. 2B,  lane 8, right upper panel). Immunoblotting of anti-RAR␣ immunoprecipitates (right lower panel) showed the RAR␣ protein in the appropriate lanes, and the expected expression of hADA3 or RAR␣ was confirmed by immunoblotting of whole cell lysates (left panels). The above results were confirmed in 293T cells (data not shown). Taken together, our results demonstrate a direct interaction of hADA3 with RXR␣, which allows it to associate with RXR␣ homodimers as well as RXR/RAR heterodimers.
ADA3 Is Present in Activator Complexes Bound to RARE of the Native p21 Promoter-A crucial aspect of coactivator function involves their ability to assemble into complexes with transcriptional activators bound to specific promoters (27). To directly assess if hADA3 is assembled into transcriptional activator complexes bound to a native retinoid response element, we used the ChIP analysis of the retinoid-responsive p21 promoter. As p21 is p53-responsive (28) and hADA3 can function as a p53 coactivator (13,21), we utilized 76R-30 cells, which lack p53 and fail to show a DNA damage-induced increase in the transcription of p21, unlike their normal parental cells (23). Initial experiments established that 9-cis-RA treatment of these cells led to an increase in p21 mRNA when cells were grown in regular DFCI-1 medium, without a need for retinoid deprivation (data not shown). A detectable level of p21 mRNA is seen in 76R-30 cells in the absence of stimulation (Fig. 3A, lane 1); however, 9-cis-RA treatment led to a dosedependent increase in p21 mRNA (lanes 2-4), consistent with the presence of endogenous RXR␣ and RAR␣ mRNA and protein in these cells (data not shown).
For ChIP analysis, 76R-30 cells were either mock-treated or treated with 100 nM 9-cis-RA for 24 h and then fixed with formaldehyde to cross-link the native chromatin-associated protein complexes with the DNA. The chromatin immunoprecipitation were carried out with preimmune (negative control) or anti-ADA3 antibodies (and anti-RAR␣ and RXR antibodies as positive controls), and the co-immunoprecipitated DNA was used as a template for the PCR to amplify the p21 promoter sequences that incorporate RARE (28). As shown in Fig. 3B, a clear PCR amplification product was observed in chromatin immunoprecipitation carried out with anti-ADA3 antibodies but not the preimmune serum (compare lane 7 with lane 9); the intensity of this band was significantly higher in the chromatin immunoprecipitation of 9-cis-RA-treated cells (lane 8), indicating that ADA3 was in association with the native p21 promoter and that this association was enhanced in the presence of the RXR␣ ligand. As expected, PCR products were amplified in chromatin immunoprecipitation carried out with anti-RXR and anti-RAR␣ antibodies; however, no further increase was seen upon ligand treatment of cells (lanes 3-6). These results are consistent with a ligand-independent binding of RXR and RXR/ RAR to the p21 promoter, as observed with other retinoid responsive promoters by gel retardation assay (29). Amplification of promoter sequences of glyceraldehyde-3-phosphate dehydrogenase, which is not a target of retinoids demonstrated the specificity of our results. Overall, these results demonstrate that hADA3 becomes part of the activator complexes bound to a retinoid response element in its native chromatin configuration upon RXR␣ ligand stimulation.
hADA3 Enhances the RXR␣-mediated and RXR␣/RAR␣mediated Transactivation of Reporters Linked to Retinoid Response Elements-Given the ability of hADA3 to interact with RXR␣ and with RXR␣/RAR␣ heterodimers and the previously defined role of ADA3 as a component of the ADA coactivator complex (17), we examined if hADA3 functions as a coactivator of transcription mediated by the RXR␣ homodimer or the RXR␣/RAR␣ heterodimer. To assess the coactivator function of hADA3 for RXR␣, we transiently transfected Saos2 cells with a luciferase reporter linked to the retinoid response element derived from the CRBPII promoter; CRBPII is known to be transactivated by RXR␣ but not by RAR␣ (30). Little CRB-PII-luciferase activity was detected in mock-treated cells (Fig.  4A, lane 1). In contrast, 9-cis-RA treatment of cells transfected with RXR␣ resulted in about a 100-fold induction of the luciferase activity (lane 2). Importantly, co-transfection of hADA3 with RXR␣ led to a substantial, hADA3 dose-dependent increase in ligand-induced luciferase reporter activity (lanes 4, 6, and 8). The reporter activity was not induced if hADA3 was transfected without RXR␣ (data not shown). Western blot of cell lysates showed the expected expression of hADA3 and RXR proteins (Fig. 4B). These results provided evidence that hADA3 can function as a coactivator for RXR␣-mediated transcription in mammalian cells.
Next, we examined the transactivation of a retinoid-responsive promoter that could be transactivated by RXR␣/RAR␣ heterodimers. For this purpose, we employed a luciferase reporter incorporating two copies of the retinoid response element from the p21 promoter (28). When the p21-luciferase reporter was transfected in Saos2 (Fig. 4, C and D) or 76R-30 (Fig. 4, E and F) cells together with RXR␣ and RAR␣, a substantial ligand-inducible increase (about 30-fold in Saos2 cells (Fig. 4C, compare lane 1 with lane 2) and 100-fold in 76R-30 cells (Fig. 4E, compare lane 1 with lane 2) in luciferase activity was observed. Importantly, cotransfection of ADA3 together with RXR␣ and RAR␣ resulted in a dramatic, ADA3 dosedependent, increase in 9-cis-RA-induced luciferase activity in both Saos2 (Fig. 4C, lanes 4, 6, and 8) and 76R-30 cells (Fig. 4E,  lanes 4, 6, and 8). Immunoblotting of cell lysates showed the expected expression of transfected proteins (Fig. 4, D and F). Interestingly, although the levels of ADA3 in 76R-30 cells were much lower than Saos2 cells (probably due to lower transfection efficiency of 76R-30 cells as compared with Saos2), the coactivator function of ADA3 is comparable in both cells (Fig. 4, compare C and D with E and F). Although the reason for this difference is currently unknown, it could reflect differences of cellular origin (epithelial versus fibroblasts). Taken together, these results demonstrated that hADA3 can function as a coactivator for RXR␣ as well as the RXR␣/RAR␣ heterodimer in mammalian cells.
To further confirm the ability of E6 to inhibit the coactivator function of ADA3 for retinoid receptor-mediated transactivation, we utilized the p21 promoter-luciferase reporter; the cells were co-transfected with RXR␣ and RAR␣. As with CRBPII reporter, the p21 luciferase reporter activity was dose-dependently inhibited by the expression of wild-type E6 but not its ⌬9 -13 mutant (Fig. 5C, lanes 6, 8, and 10). Immunoblotting of cell lysates confirmed the expression of hADA3, RXR, and RAR␣ (Fig. 5D). Taken together, our results clearly demonstrate that HPV16 E6 abrogates the coactivator function of hADA3 toward retinoid receptors RXR␣ and RXR␣/RAR␣ in human cells.
The results presented here have significant implications for the potential role of ADA3-containing coactivator complexes in physiological pathways and in oncogenesis. While yeast ADA3 as a component of ADA and other coactivator complexes has been clearly demonstrated to be important, little was known about hADA3 function in mammalian systems except for its ability to function as a p53 coactivator, which others and we have recently uncovered (13,21). Our results that hADA3 functions as a coactivator for retinoid receptors in mammalian cells, together with the ability of yeast ADA3 to function as a coactivator of multiple mammalian nuclear hormone receptors when these were expressed in yeast, strongly suggest that ADA3-containing complexes may participate as coactivators for multiple nuclear hormone receptors. While a recent study failed to detect mouse ADA3 association with estrogen receptor (31), yeast ADA3 has been shown to coactivate estrogen receptor (ER) function, and we have detected a direct interaction between hADA3 and ER. 2 The presence of hGCN5 as well as a hGCN5-related gene product P/CAF in hADA3-containing complexes (18), direct interaction of P/CAF with p300/CBP (32), and the interaction of hADA3 itself with p300 (21) suggest that hADA3 may form multiple, distinct coactivator complexes, as also is the case in yeast (17). Thus, it is likely that additional transcriptional activators will emerge as targets of ADA3 coactivator function.
Our findings that HPV16 E6 abrogates the coactivator function of hADA3 have obvious implications for the potential role of the hADA3-mediated biochemical pathways in oncogenesis. The role of presently known ADA3 targets (p53 and retinoid receptors) in cell growth and differentiation is well established, and these pathways are frequently affected during oncogenesis. Thus, if E6 indeed targets the various hADA3-containing complexes and influences transcriptional pathways in which these complexes play a role, this could represent a significant mechanism for the role of E6 in HPV-mediated oncogenesis.
Our findings support the emerging concept that viral onco- hADA3 is a component of the coactivator complex bound to native retinoid receptor-responsive promoter of the p21 gene. A, induction of p21 mRNA in 76R-30 cells upon 9-cis-RA treatment. 76R-30 cells were grown in DFCI-1 medium, and either mock-treated or treated with indicated concentrations of 9-cis-RA for 24 h. 15 g of total cellular RNA was resolved on a 1% agarose gel, transferred to a nylon membrane, and hybridized with a 32 P-labeled p21 cDNA probe. The p21 mRNA signals were visualized by autoradiography (upper panel). The same blot was reprobed with a 32 P-labeled 36B4 probe as a loading control (lower panel). B, ChIP analysis of hADA3 binding to p21 promoter in 76R-30 cells. 76R-30 cells were either mock-treated (Ϫ) or treated with 100 nM of 9-cis-RA (ϩ) for 24 h, formaldehyde fixed to cross-link the DNA to native chromatin-associated protein complexes, and chromatin lysates prepared as described under "Experimental Procedures." Equal aliquots of chromatin lysates were subjected to immunoprecipitation with antibodies against RXR, RAR␣, and hADA3. Preimmune serum was used as a control. The DNA associated with immunoprecipitates was isolated and used as templates to PCR-amplify the p21 promoter region containing RARE. The PCR products were resolved by agarose gel electrophoresis and stained with ethidium bromide. PCR amplification of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) promoter was used as a specificity control.
FIG. 4. hADA3 enhances the RXR␣-mediated transactivation of reporters linked to RARE. A, Saos2 cells were deprived of retinoids by growth in charcoal-treated phenol-free ␣-MEM medium for 48 h, transfected with 100 ng of CRBPII-RARE-Luc and 40 ng of RXR␣ plasmids together with increasing amounts of pCR3.1-hADA3, using the FuGENE 6 reagent. The total amount of DNA for each transfection was kept constant by adding vector DNA. After 24 h, the cells were either mock-treated (Ϫ) or treated with 100 nM 9-cis-RA (ϩ) for 24 h prior to lysis. 20 ng of SV40 Renilla luciferase reporter (pRL-SV40) was concurrently transfected and used to correct for differences in transfection. Equal amounts of cell lysates were used to measure the luciferase activity. Transactivation results were calculated as fold activation over vector-transfected cell lysates. Data indicates mean Ϯ S.D. of triplicates of a representative experiment. Experiments were repeated at least three times. Ϫ, vector DNA. proteins, such as HPV E6, have attained an ability to perturb the function of multiple transcriptional coactivators apparently by multiple mechanisms. Notably, E6 oncoprotein has been demonstrated to associates with p300, which serves as a coactivator for a number of transactivators, including nuclear hormone receptors (33,34). While the effect of HPV E6 binding to p300 on nuclear receptor-mediated transcription has not been examined, E6 is known to abrogate the coactivator function for p300 toward p53 (33,34). Recently, another E6-binding protein AMF1 (also called G-protein pathway suppressor 2 or GPS2), was shown to be a coactivator for p53 (35,36). These studies are consistent with the idea that E6 has attained an ability to perturb multiple coactivators. How might the E6 perturbation of coactivator function promote oncogenic transformation ? We envision that E6 interaction with hADA3 would disrupt the hADA3-containing HAT complexes, prevent the recruitment of HAT activity to promoter elements, and inhibit the expression of genes that mediate tumor suppressor functions. The role of histone acetylation as a tumor suppressor mechanism is supported by the propensity of CBP Ϯ heterozygous mice to de-velop hematopoietic malignancies (37), and the ability of RAR␣ fusion oncogenes, which recruits histone deacetylases to RAREs (38), to induce promyelocytic leukemia by inhibiting cell differentiation. Indeed, histone deacetylase inhibitors together with retinoid are now being assessed for treatment of such patients (38 -40). Thus, it is reasonable to postulate that ADA3, as a component of HAT-containing coactivator complexes, plays a role in transcription of genes. A large body of evidence supports an important role of retinoids in cell differentiation and cell growth inhibition (41)(42)(43)(44). There is also increasing evidence that retinoids down-regulate the telomerase in a pathway distinct from cell differentiation, implicating the role of retinoids in replicative senescence (45,46). Complementary in vivo studies have demonstrated the ability of retinoids to inhibit tumor formation in carcinogenesis models, typically at the step of tumor promotion (47,48). Importantly, these studies have led to clinical trials of retinoids in chemoprevention of a number of epithelial and non-epithelial cancers (49).
In conclusion, we have shown hADA3 directly interacts with FIG. 5. Inhibition of the hADA3 coactivator function by HPV16 E6. A, Saos2 cells were co-transfected with 100 ng of CRBPII-RARE-Luc and 40 ng of RXR␣, with or without 2.5 g each of pCR3.1-hADA3 and pCR3.1-16 E6 or its mutant, as indicated. After the transfection, the cells were treated with 9-cis-RA, and the luciferase activity was measured in cell lysates, as described in the legend to Fig. 4. B, protein levels of hADA3 (upper panel) and RXR (lower panel) were measured by Western blotting in 9-cis-RA-treated lysates of the experiment shown in A. C, Saos2 cells were co-transfected with 1 g each of p21RARE-Luc, RXR␣, and RAR␣ plasmids, with or without 2.5 g each of pCR3.1-hADA3 and/or pCR3.1-16 E6 or its mutant, as indicated (ϩ). After transfection, the cells were processed as described in the legend to Fig. 4. D, protein level of hADA3 (upper panel), RXR (middle panel), and RAR␣ (lower panel) were analyzed by Western blotting in 9-cis-RA-treated lysates of the experiment shown in C.
B, levels of hADA3 (upper panel) and RXR␣ (lower panel) proteins were analyzed by Western blotting in 9-cis-RA-treated cell lysates of the experiment shown in A. C, Saos2 cells were transfected with 1 g of p21-RARE-Luc and 1 g each of RXR␣ and RAR␣, together with increasing amounts of hADA3, using the FuGENE 6 reagent. After transfection, the cell lysates were analyzed for the luciferase activity, as described in A. D, protein levels of hADA3 (upper panel), RXR␣ (middle panel), and RAR␣ (lower panel) were analyzed by Western blotting in 9-cis-RA-treated lysates of the experiment shown in C. E, 76R-30 cells were transfected with 0.5 g of p21-RARE-Luc and 0.5 g each of RXR␣ and RAR␣, together with increasing amounts of hADA3, using the FuGENE 6 reagent. After transfection, the cell lysates were analyzed for the luciferase activity, as described in A. F, protein levels of hADA3 (upper panel), RXR (middle panel), and RAR␣ (lower panel) were analyzed by Western blotting in 9-cis-RA-treated lysates of the experiment shown in E.
RXR␣ and functions as a coactivator for RXR␣ and RAR␣/ RXR␣-mediated transactivation. In addition, we show that HPV16 E6 by targeting hADA3 for degradation, abrogate the RXR␣-mediated transactivation, implicating disruption of retinoid receptor function in E6 oncogenesis. Given the evolutionary conservation of hADA3, it is likely that viral oncoproteinmediated inactivation of its function plays an important role in oncogenesis.