Orphan Receptor TR3 Attenuates the p300-Induced Acetylation of Retinoid X Receptor- (cid:1)

Acetylation modification regulates the functions of histone and nonhistone proteins, including transcriptional activity, protein interaction, and subcellular localization. Although many nuclear receptors have been shown to be modified by acetylation, whether retinoid X receptors (RXRs) are acetylated and how the acetylation is regulated remains un-known. Here, we provide the first evidence of RXR (cid:1) acetylation by p300 on lysine 145. Acetylation of RXR (cid:1) by p300 facilitated its DNA binding and subsequently increased its transcriptional activity. Furthermore, we discovered that TR3, an orphan receptor, exerted a negative regulation on p300-induced RXR (cid:1) acetylation. TR3 significantly reduced the p300-induced RXR (cid:1) acetylation and transcriptional activity, and such inhibition required the interaction of TR3 with RXR (cid:1) . Binding of TR3 to RXR (cid:1) resulted in the sequestration of RXR (cid:1) from p300. 9- cis retinoic acid, a ligand for RXR (cid:1) , enhanced the association of RXR (cid:1) with TR3, rather than acetylation of RXR (cid:1) by p300. Biological function analysis revealed that the mitogenic activity of RXR (cid:1) stimulated by p300 was acetylation dependent and could be repressed by TR3. Upon the treatment of 9- cis retinoic acid, RXR (cid:1)

Acetylation modification regulates the functions of histone and nonhistone proteins, including transcriptional activity, protein interaction, and subcellular localization. Although many nuclear receptors have been shown to be modified by acetylation, whether retinoid X receptors (RXRs) are acetylated and how the acetylation is regulated remains unknown. Here, we provide the first evidence of RXR␣ acetylation by p300 on lysine 145. Acetylation of RXR␣ by p300 facilitated its DNA binding and subsequently increased its transcriptional activity. Furthermore, we discovered that TR3, an orphan receptor, exerted a negative regulation on p300-induced RXR␣ acetylation. TR3 significantly reduced the p300-induced RXR␣ acetylation and transcriptional activity, and such inhibition required the interaction of TR3 with RXR␣. Binding of TR3 to RXR␣ resulted in the sequestration of RXR␣ from p300. 9-cis retinoic acid, a ligand for RXR␣, enhanced the association of RXR␣ with TR3, rather than acetylation of RXR␣ by p300. Biological function analysis revealed that the mitogenic activity of RXR␣ stimulated by p300 was acetylation dependent and could be repressed by TR3. Upon the treatment of 9-cis retinoic acid, RXR␣ was translocated with TR3 from the nucleus to the mitochondria, and apoptosis was induced. Taken together, our data demonstrate the distinct regulatory mechanisms of p300 and TR3 on RXR␣ acetylation and reveal a previously unrecognized role for orphan receptor in the transcriptional control of retinoid receptors. (Molecular Endocrinology 21:

2877-2889, 2007)
N UCLEAR RECEPTORS COMPRISE a large family of ligand-dependent transcription factors that bind to hormone response elements of target genes and regulate their transcriptional activities. Retinoid X receptors (RXRs) are important members of the nuclear receptor superfamily and regulate various physiological processes, such as proliferation, differentiation, and apoptosis in many cell types. RXRs can either homodimerize by themselves or heterodimerize with retinoic acid receptors (RARs) and bind to RXR response element (RXRE) or retinoic acid response element (RARE), thereby positively or negatively controlling the gene transcription and expression. In addition, RXRs also form heterodimers with many other members of nuclear receptors, including vitamin D receptor, peroxisome proliferator-activated receptor, and thyroid hormone receptor (TR) as well as several orphan receptors, such as liver X receptor and TR3 (also termed as Nur77 and NGFI-B) (1,2). RXRs, therefore, play an essential role in integrating multiple nuclear hormone-signaling pathways.
Although RXRs are well known to function restrictedly in nucleus as potent heterodimerization partners for many nuclear receptors, one of the members, RXR␣, has been demonstrated to shuttle between the nucleus and the cytoplasm (3). We have found that RXR␣ serves as a carrier for TR3 mitochondrial targeting in a RXR-ligand-dependent manner (3). Others have also revealed that RXR␣ and its selective ligands are critical regulators for TR3 activity and localization (4). TR3 is a transcriptional factor, and its expression can be rapidly induced by a number of growth factors and mitogens in a variety of cancer cells. In addition, TR3 can translocate from the nucleus to the mitochondria to initiate apoptosis in response to several apoptotic stimuli (3,5,6). Recently, we have demonstrated a unique TR3-p53-MDM2 pathway in which TR3 with p53 cooperatively acts to regulate MDM2 functions in the nucleus (7). Therefore, TR3 may mediate distinct signal pathways, although its regulatory role in these pathways remains to be clarified.
Acetylation represents an important mechanism to regulate the functions of nuclear receptors and their related signal pathways. The candidate acetylation motif (KXKK/RXKK) of many nuclear receptors, such as TR, RAR, peroxisome proliferator-activated receptor, liver X receptor, FXR, vitamin D receptor, glu-cocorticoid receptor, progesterone receptor, hepatocyte nuclear factor, and steroidogenic factor 1, is conserved among different species, including vertebrates, arthropods, and nematodes (8). Other types of motif for acetylation have also been identified, such as GK and SK (9). p300 is a transcriptional coactivator that possesses an intrinsic histone acetyltransferase activity (10). It contributes to the formation of a protein activation complex that bridges various factors to the general transcription machinery. p300 has been shown to acetylate a growing number of nonhistone proteins, notably transcription factors such as p53 (11), E2F1 (12), high-mobility group protein isoform I and Y (13), hepatocyte nuclear factor 4 (14), HIV Tat (15), and nuclear receptors such as androgen receptor (16,17), and estrogen receptor-␣ (18,19).
In the present study, we demonstrated for the first time that RXRs, including RXR␣ and RXR␥, are subjected to p300 acetylation. Such acetylation promotes RXR␣ binding to RXRE, thereby increasing the transcriptional activity of RXR␣. We further revealed that TR3 has a significant inhibitory effect on p300-induced RXR␣ acetylation. TR3 competes with p300 in RXR␣ binding and results in the translocation of TR3/RXR␣ from the nucleus to the mitochondria in response to 9-cis retinoic acid. Although acetylation of RXR␣ by p300 stimulates the growth of HeLa cells, attenuation of p300-induced RXR␣ acetylation by TR3 induces apoptosis through TR3/RXR␣ heterodimerization, translocation, and mitochondrial targeting. Taken together, our results demonstrated that the orphan receptor TR3 plays an important role in p300-induced RXR␣ acetylation by functioning as a negative regulator.

p300 Induces Acetylation of RXR␣ on K145
Retinoid X receptors (RXRs) consist of three different isoforms, RXR␣, RXR␤, and RXR␥. To investigate whether RXRs can be acetylated by p300, green fluorescent protein (GFP)-tagged isoforms of RXRs were transiently transfected into 293T cells. Lysates were prepared from these transfected cells, and GFPtagged proteins were immunoprecipitated and blotted with an acetylation-specific antibody. As shown in Fig.  1A, none of RXR␣, RXR␤, and RXR␥ could be detected by the acetylation-specific antibody when they were transfected into cells alone. However, when p300 was cotransfected, RXR␣ and RXR␥, but not RXR␤, became detectable by the same antibody, indicating that RXR␣ and RXR␥ were acetylated in the presence of p300.
We chose RXR␣ for further study because its acetylation level was higher than RXR␥ (Fig. 1A). The acetylation level of RXR␣ seemed to be enhanced by the increase of transfected p300 (Fig. 1B). However, cotransfection of a p300 mutant (p300DY) deficient in acetylase activity (20) resulted in no acetylation of RXR␣ (Fig. 1B), suggesting that the acetylase activity of p300 is directly responsible for the acetylation of RXR␣. The acetylated RXR␣ could also be clearly detected in HeLa and MCF-7 cells (Fig. 1C, left panel, black arrow), both of which are known to stably express p300 ( Fig. 1C). When small interfering RNA (siRNA)-p300 was introduced into these two cell lines to inhibit the expression of endogenous p300, the acetylation level of RXR␣ greatly decreased (Fig. 1C, right panel, black arrow). Together, these results suggest that p300 is essential and necessary to induce the acetylation of RXR␣.
Sequence analysis of RXR␣ revealed four consensus acetylation motifs located at approximately 144-145 (GK), 212-213 (GK), 362-363 (SK), and 387-388 (SK), respectively ( Fig. 2A). To determine which site is responsible for RXR␣ acetylation, different deletion mutants of RXR␣ were constructed ( Fig. 2A) and cotransfected with p300 into 293T cells for acetylation assays. Wild-type RXR␣ and deletion mutants of RXR␣/D2 and RXR␣/D3, but not RXR␣/D1 mutant, were found to be acetylated (Fig. 2B), indicating that the region around amino acids 133-156 is responsible for RXR␣ acetylation. Because that region contains only a candidate acetylation motif (Lys145), we then constructed a new point mutant of RXR␣ (K145R) in which Lys145 was replaced with Arg. As expected, the K145R mutation significantly abolished the acetylation of RXR␣ induced by p300 in 293T cells (Fig. 2C). These results clearly demonstrate Lys145 as the p300-induced acetylation site on RXR␣.

Acetylation of RXR␣ Increases Its DNA-Binding and Transcriptional Activity
Acetylation of nuclear receptors is often closely related to the regulation of their DNA-binding and transcriptional activity. To determine the correlation between RXR␣ acetylation and its transcriptional activity, we assessed the activity of a luciferase reporter gene linked with RARE in 293T cells. With the presence of endogenous RXR␣, p300 could enhance the reporter activity in a concentration-dependent manner (Fig.  3A), whereas the acetyltransferase-dead mutant of p300 (p300DY) failed to do so (Fig. 3B). When RXR␣ mutant K145R was used for cotransfection, p300 had no effect on the reporter activity (Fig. 3C). Taken together, these results clearly demonstrated a positive regulation of p300-induced acetylation on the transcriptional activity of RXR␣.
We next analyzed the subcellular localization of RXR␣ and p300 by confocal microscopy. When expressed alone in 293T cells, p300 displayed as microspeckles in the nucleus, whereas RXR␣ and its mutant K145R mainly located in the nucleus with diffused pattern (Fig. 3D, first panel). Once coexpressed, p300 and RXR␣ showed a colocalization pattern in about 50% cells by forming obvious speckles in the nucleus (Fig. 3D, second panel). By contrast, such speckles could only be seen in less than 10% of cells that were transfected with either p300DY/RXR␣ (Fig. 3D, third panel) or K145R/p300 (Fig. 3D, last panel). We suspected that the formation of RXR␣ with p300 into speckles might be related to its DNA-binding property. To verify this possibility, EMSA was performed with the use of RXRE as a probe for incubation with nuclear proteins. As shown in Fig. 3E, nuclear proteins extracted from GFP-RXR␣-transfected 293T cells formed a complex with RXRE, and p300 enhanced this complex formation obviously (Fig. 3E, black arrow). The complex band was up-shifted only when the nuclear proteins were preincubated with anti-GFP antibody (aGFP) (Fig. 3E, white arrow), indicating that the complex is composed of homodimer of RXR␣. Chromatin immunoprecipitation (ChIP) assay further demonstrated that p300 could enhance the binding of RXR␣ to the RAR␤ promoter (Fig.  3F). However, the K145R mutant of RXR␣ did not bind to RXRE, even in the presence of p300 (Fig. 3E), and p300 also failed to promote its binding to RAR␤ promoter (Fig.  3F). Together, these data suggested that RXR␣ acetylation by p300 results in its binding to DNA, which is correlated well with the activation of its transcriptional activity.

TR3 Binds to RXR␣ and Attenuates Its Acetylation by p300
The orphan receptor TR3 has been shown to inhibit p300-induced acetylation of p53 (7). Because TR3 heterodimerizes with RXR␣ in vivo (3,21,22), we suspected that TR3 may also be involved in regulation of RXR␣ acetylation. When TR3 was introduced into 293T cells that had been transfected with RXR␣ and p300, we found that the p300-induced RXR␣ acetylation became significantly attenuated (Fig. 4A). A similar result was also observed in HeLa cells that were transfected with TR3 alone (Fig. 4B, black arrow). We further investigated the inhibitory effect of endogenous TR3 on RXR␣ acetylation. When siRNA-TR3 was introduced into HeLa cells to inhibit endogenous TR3 expression, more acetylated RXR␣ could be detected (Fig. 4C, black arrow). Quantitative analysis by densitometry further confirmed the effect of TR3 on regulating RXR␣ acetylation (Fig. 4, B and C A B Fig. 1. p300 Induced the Acetylation of RXR␣ A, Acetylation of RXR isoforms by p300. Different GFP-tagged retinoid receptors, including RXR␣, -␤, and -␥, together with or without p300, were transfected into 293T cells, respectively. After transfection, cells were harvested and cell extracts were prepared. RXRs were immunoprecipitated with anti-GFP antibody followed by Western blotting with acetylation-specific antibody. Tubulin was used to indicate the similar loading of proteins in each lane. B, Acetylation of RXR␣ by p300 in a dosedependent manner. GFP-RXR␣, together with different amounts of HA-p300, was transfected into 293T cells. The level of RXR␣ acetylation was determined as described above. p53 and GFP were used as positive and negative control, respectively. C, Effect of endogenous p300 on RXR␣ acetylation. Cell extracts from HeLa and MCF-7 cells were prepared and immunoprecipitated with anti-RXR␣ antibody, followed by Western blotting with acetylation-specific antibody to show the levels of RXR␣ acetylation (left panel, black arrow). IgG was used as negative control. To diminish the activity of endogenous p300, siRNA-p300 was introduced into HeLa and MCF-7 cells respectively, and the levels of RXR␣ acetylation was determined as above (right panel, black arrow). White arrow indicates a heavy chain. C, bottom panels). Together, these results demonstrated that TR3 has a unique role in antagonizing the p300induced RXR␣ acetylation.
TR3 may inhibit RXR␣ acetylation via interaction. We therefore constructed different deletion mutants of TR3 ( Fig. 4D) to test their interaction with RXR␣. The results showed that TR3/⌬⌵ but not TR3/⌬⌴ could be efficiently precipitated with RXR␣ when coexpressed in 293T cells (Fig. 4D), indicating that the amino acids 299-361 region of TR3 is important for the interaction with RXR␣. Next, we introduced these two deletion mutants of TR3 into 293T cells to analyze their effects on RXR␣ acetylation. The p300-induced acetylation of RXR␣ could be inhibited by TR3/⌬⌵ that interacted with RXR␣ but not by TR3/⌬⌴ that failed to interact with RXR␣ (Fig. 4E). Therefore, it is likely that interaction with RXR␣ is a prerequisite for TR3 to inhibit p300-induced acetylation of RXR␣.

TR3 Competes with p300 for RXR␣ Binding
Because TR3 interacts with RXR␣, it may compete with p300 in the RXR␣ binding to attenuate the p300induced acetylation of RXR␣. Coimmunoprecipitation/ Western blotting confirmed that either TR3 or p300 was able to interact with RXR␣ independently, whereas RXR␣ point mutant K145R almost failed to bind with p300, although it still interacted with TR3 (Fig. 5A). As we expected, increasing the amount of TR3 enhanced the TR3-RXR␣ interaction and decreased the p300-RXR␣ interaction, and vice versa (Fig. 5B), indicating that TR3 and p300 bind to RXR␣ in a competition manner. We further transfected the HA-TR3, HA-p300, and Myc-RXR␣ into 293T cells and detected higher level of TR3 than p300 in RXR␣ immunoprecipitates (Fig. 5C, right panel, lane 3), although HA-TR3 and HA-p300 were expressed at similar levels (Fig. 5C, left panel, lane 3). This result further suggested that RXR␣ has a higher ability to interact with TR3 than p300 when both proteins coexist.
Confocal microscopic analysis revealed that transfection of TR3 inhibited the formation of RXR␣-p300 speckles in 293T cells, with the percentage of cells containing p300/RXR␣ speckles dramatically dropped from 46.8 to 6.1%. Although TR3 colocalized with RXR␣ in a diffused pattern, p300 existed as microspeckles (Fig. 5D). We further found that the TR3/⌬N mutant, but not the TR3/⌬M mutant, could interfere with the formation of RXR␣/p300 speckles (Fig. 5D). The percentage of cells with speckles was much higher in TR3/⌬M-transfected cells (39.3%) than in TR3/⌬N-transfected cells (11.0%). Therefore, the reduced colocalization of RXR␣ with p300 is most likely due to its sequestration by TR3.

9-cis Retinoic Acid Facilitates Translocation of Endogenous RXR␣ with TR3
Because 9-cis retinoic acid is a well known ligand for RXR␣ (23,24), we went on to investigate how it regulates the physical interactions of RXR␣ with p300 and TR3 as well as the acetylation of RXR␣ by p300. Coimmunoprecipitation/Western blotting revealed that 9-cis retinoic acid enhanced the interaction of endogenous RXR␣ with TR3 and simultaneously decreased the interaction of RXR␣ with p300 (Fig. 6A). The increased TR3/RXR␣ association consequently resulted in the translocation of both proteins from the nucleus to the mitochondria, which was seen in more than 65% of HeLa cells (Fig. 6B). On the other hand, the acetylation level of RXR␣ in HeLa cells was reduced to some extent upon 9-cis retinoic acid treatment (Fig.  6C, black arrow). This is most likely due to the attenuation of p300/RXR␣ association (Fig. 6A), because the expression level of endogenous RXR␣ and p300 was not affected by 9-cis retinoic acid (Fig. 6C, top  panel), although the acetylation of RXR␣ was inhibited by 9-cis retinoic acid (Fig. 6C, bottom panel). A similar result was also observed in MCF-7 cells (data not A B C Fig. 2. Identification of the p300 Acetylation Sites on RXR␣ A, Schematic diagrams depict different RXR␣ deletion constructs used in acetylation assays. B, Identification of RXR␣ sequences critical for its acetylation by p300. 293T cells were transfected with p300 and RXR␣ or different RXR␣ truncation mutants as indicated. The levels of RXR␣ acetylation were determined as described in Fig. 1. C, Lysine 145 was responsible for the acetylation of RXR␣. RXR␣ point mutant K145R, together with p300, was transfected into 293T cells, and the levels of RXR␣ acetylation were determined. shown). As a consequence of the reduced acetylation, the transcriptional activity of RXR␣ in HeLa cells remained to be induced by 9-cis retinoic acid (Fig. 6D, black bars) and slightly repressed by the transfection with an increased amount of TR3 (Fig. 6D). Together, these results demonstrate that 9-cis retinoic acid facilitates RXR␣ translocation with TR3 rather than its acetylation by p300 in HeLa cells.

p300 and TR3 Have Distinct Cellular Functions through Regulation of RXR␣ Acetylation
Finally, we investigated the biological significance of p300 positive regulation and TR3 negative regulation on RXR␣ acetylation. p300 has been shown to promote apoptosis as well as cell proliferation, which appears to be highly context dependent (25, 26). We Fig. 3. Acetylation of RXR␣ by p300 Enhanced Its DNA-Binding and Transcriptional Activities A, p300 increased the transcriptional activity of RXR␣. The RARE-linked luciferase reporter gene and ␤-gal gene expression vector, together with increasing doses of p300 expression vector, were transfected into 293T cells. Reporter gene activity was determined and normalized in relation to the cotransfected ␤-gal activity. The bars represent the average Ϯ mean from three independent experiments. B, A p300 mutant that was deficient in acetylase activity (p300DY) was used for the same assay as A. C, Correlation between RXR␣ transcriptional activity and its acetylation. RARE-luciferase reporter gene and ␤-gal gene expression vector, together with different combinations of RXR␣ and its point mutant K145R as indicated, were transfected into 293T cells in the absence or presence of p300. Reporter gene activity was determined and normalized as described above. D, Colocalization of RXR␣ with p300. HA-p300, Myc-RXR␣, and its point mutant K145R were transfected into 293T cells in different combinations as indicated. Cells were immunostained by Myc antibody followed by FITC-conjugated secondary antibody to detect RXR␣ or its point mutant K145R, or by HA antibody followed by Texas Red-conjugated secondary antibody to detect p300. Stained cells were visualized with the confocal microscope. Percentage of cells with speckle RXR␣/p300 is indicated. E, Effect of p300 on RXR␣ and its point mutant K145R binding to DNA. GFP-RXR␣ or GFP-K145R was transfected into 293T cells in the absence or presence of HA-p300 as indicated. Nuclear proteins were prepared, and homodimerization of RXR␣ was analyzed by the method of EMSA and probed with biotin-labeled RXRE oligonucleotides. To determine the formation of RXR␣ homodimer, antibody specific for GFP (aGFP) was preincubated with nuclear proteins for 2 h before assay. NS means nonspecific band. F, The binding of RXR␣ to the RAR␤ promoter. GFP-RXR␣ or its point mutant K145R was transfected into 293T cells in the absence or presence of p300 as indicated. The binding of RXR␣ or K145R to the RAR␤ promoter was analyzed by the method of ChIP. Beads were used as negative control. Input was used to indicate similar loading of DNA in each lane.  Fig. 4. Attenuation of p300-Induced RXR␣ Acetylation by TR3 A, TR3 antagonized the acetylation of RXR␣ induced by p300. Myc-TR3, together with GFP-RXR␣ and HA-p300, were transfected into 293T cells. The levels of RXR␣ acetylation were determined as described as Fig. 1B. B, TR3 inhibited acetylation of endogenous RXR␣. HeLa cells were transfected with Myc-TR3 and the levels of RXR␣ acetylation (black arrow) were determined as described in Fig. 1C. White arrow indicates a heavy chain. C, Effect of siRNA-TR3 on RXR␣ acetylation. HeLa cells were transfected with siRNA-TR3, and the levels of RXR␣ acetylation (black arrow) were determined as described as Fig. 1C. The levels of RXR␣ acetylation regulated by TR3 were quantified by densitometry in B and C. D, Interaction of RXR␣ with TR3 and its deletion mutants. Schematic diagrams of deletion mutants of TR3 are shown at the top. 293T cells were transfected with GFP-RXR␣ and Flag-TR3 or its different deletion mutants as indicated. Cell lysates were immunoprecipitated with anti-GFP antibody. The immunoprecipitates and cell lysates were then analyzed by Western blotting separately using anti-Flag-antibody for TR3 and its deletion mutants and anti-GFP antibody for RXR␣. E, Effect of TR3 deletion mutants on acetylation of RXR␣ by p300. Flag-TR3 and its deletion mutants, together with GFP-RXR␣ and HA-p300, were transfected into 293T cells, and the levels of RXR␣ acetylation were determined as described in Fig. 1B.

Fig. 5. TR3 Competed with p300 Binding to RXR␣
A, Interaction of RXR␣ with TR3 or p300. 293T cells were transfected with HA-TR3, HA-p300, and Myc-RXR␣ or its point mutant K145R as indicated. Cell lysates were immunoprecipitated with anti-HA antibody. The immunoprecipitates and cell lysates were then analyzed by Western blotting separately using anti-HA antibody for TR3, anti-HA antibody for p300, and anti-Myc antibody for RXR␣. B, TR3 and p300 bound to RXR␣ competitively. Increasing amounts of TR3 or p300 together with RXR␣ were transfected into 293T cells as indicated. Cell lysates were immunoprecipitated with anti-Myc antibody to pull down RXR␣. The immunoprecipitates and cell lysates were then analyzed by Western blotting separately using anti-HA antibody for p300 and anti-GFP antibody for TR3. C, The relative binding ability of p300 and TR3 to RXR␣. 293T cells were transfected with Myc-RXR␣, HA-TR3, and HA-p300. Western blot against the common HA tag was used to compare the expression levels of TR3 and p300 (left panel, indicated by arrow). Binding ability of TR3 and p300 to RXR␣ is shown by using anti-HA antibody in RXR␣ immunoprecipitates (right panel, indicated by arrow). D, Effect of TR3 and its deletion mutants on p300-RXR␣ colocalization. Flag-TR3 or its deletion mutants together with Myc-RXR␣ and HA-p300 were transfected into 293T cells as indicated. The transfected cells were immunostained by anti-Flag antibody followed by Alexa fluor 350-conjugated secondary antibody to detect TR3 and its point mutants, by anti-HA antibody followed by Texas Red-conjugated secondary antibody to detect p300, or by anti-Myc antibody followed by FITC-conjugated secondary antibody to detect RXR␣. Stained cells were visualized with the confocal microscope. The same data are plotted to indicate the percentage of cells with speckle RXR␣ (bottom panel).  A, Effect of 9-cis retinoic acid on interactions of RXR␣ with TR3 or p300. HeLa cells were treated with 9-cis retinoic acid (10 Ϫ6 mol/liter) for 12 h. Cell lysates were prepared and immunoprecipitated with anti-TR3 and anti-p300 antibodies, respectively. The immunoprecipitates and cell lysates were then analyzed by Western blotting separately using anti-RXR␣, anti-TR3, and anti-p300 antibodies to detect endogenous TR3, RXR␣, or p300 proteins, respectively. IgG was used as control. B, Cotranslocation of TR3 and RXR␣ from the nucleus to the mitochondria in HeLa cells induced by 9-cis retinoic acid. Cells were treated with or without 9-cis retinoic acid (10 Ϫ6 mol/liter) for 12 h and then immunostained with anti-RXR␣, anti-TR3, and anti-Hsp60 antibodies, followed by their corresponding FITC-, Texas Red-, and Alexa fluor 350-conjugated secondary antibodies to show endogenous RXR␣, TR3, and Hsp60 proteins simultaneously. The fluorescent images were visualized under confocal microscope. C, Effect of 9-cis retinoic acid on RXR␣ acetylation. To examine the effect of 9-cis retinoic acid on endogenous RXR␣ acetylation, HeLa cells were treated with 9-cis retinoic acid (10 Ϫ6 mol/liter) for 12 h before the levels of RXR␣ acetylation were determined. White arrow indicates a heavy chain. The levels of RXR␣ acetylation regulated by 9-cis retinoic acid were quantified by densitometry. D, Effect of 9-cis retinoic acid on RARE activity in HeLa cells. Cells were transfected with Myc-TR3 and then treated with 9-cis retinoic acid (10 Ϫ6 mol/liter) for 6 h. The reporter gene activity was determined as described in Fig. 3A. HeLa cells, as a clear left-shifted curve was seen (third panel, yellow curve). Moreover, like TR3, TR3/⌬⌵ that bound to RXR␣ (Fig. 4D) and blocked p300 acetylation (Fig. 4E) also diminished the p300 effect on stimulating cell growth (third panel, gray curve), whereas TR3/⌬M that did not interact with RXR␣ (Fig. 4D) failed to antagonize p300-induced mitogenic activity (third panel, pink curve). Based on these results and the fact that TR3 heterodimerization with RXR␣ facilitated their shuttling to the mitochondria in response to 9-cis retinoic acid (Fig. 6B), it is possible that TR3 attenuation on RXR␣ acetylation might correlate with apoptosis induction, because mitochondrial targeting of TR3 is critical for triggering apoptosis in cancer cells (3,5,6), just as, for example, when TR3/RXR␣-cotransfected HeLa cells were treated with 9-cis retinoic acid, more apoptotic cells were seen and the apoptotic rate reached to 24.5% (Fig. 7B). Similar results were also observed in TR3⌬⌵/RXR␣ transfected cells (25.7% apoptotic cells) and K145R/ TR3-transfected cells (18.4% apoptotic cells), but not in TR3⌬M/RXR␣-transfected cells (less than 5%). We therefore concluded that p300 positively regulates RXR␣ acetylation to facilitate cell proliferation and TR3 negatively regulates RXR␣ acetylation for apoptosis induction. Fig. 7. Different Functions of p300 and TR3 in Activating Mitogenic Activity and Inducing Apoptosis A, Effects of RXR␣, p300, and TR3 on the mitogenic activity of HeLa cells. Different expression vectors, including RXR␣, p300, TR3, and their related mutants, were transfected into HeLa cells as indicated. After transfection, the cells were maintained in BrdU-containing medium for 2 h. The mitogenic activity was then identified by flow cytometry as described in Materials and Methods. B, TR3 and RXR␣ induced apoptosis in HeLa cells. HeLa cells were transfected with GFP-TR3 and Myc-RXR␣ or their mutants as indicated and then treated with 9-cis retinoic acid (10 Ϫ6 mol/liter) for 48 h. The nuclear morphology stained by 4Ј,6-diamidino-2-phenylindole was visualized under fluorescent microscope, and apoptotic cells were scored by examination of 500 transfected cells.

DISCUSSION
Posttranscriptional modification of nuclear receptors integrates a variety of signaling pathways (8). Although the exact number and variety of nuclear receptors that are acetylated in the cells is still largely unknown, it is clear that far more candidates are modified by this mechanism than initially appreciated. In the current study, we provide the first evidence that RXRs, including RXR␣ and RXR␥, are acetylation targets of p300. Acetylation by p300 enhanced the DNA binding and activated the transcriptional activity of RXR␣, thereby promoting cell proliferation. Furthermore, we found that TR3, an orphan receptor, functions as a negative regulator to attenuate RXR␣ acetylation by competition with p300 for RXR␣ binding. As a result, p300mediated DNA binding of RXR␣ was repressed and RXR␣ turned to form a complex with TR3, which would translocate from the nucleus to the mitochondria to facilitate apoptosis induction in response to 9-cis retinoid acid. These findings allowed us to propose a novel model in which RXR␣ acetylation is positively regulated by p300 and negatively regulated by TR3 (Fig. 8).
RXR␣ processed four candidate acetylation sites, and one of them, the lysine residue 145, could be acetylated by p300. Mutation of Lys145 to Arg greatly abrogated RXR␣ acetylation. As a result of acetylation, the transcriptional activity of RXR␣ was increased in a p300-dependent manner. Because K145 locates in the first zinc finger of the DNA-binding domain of RXR␣ (27), p300-induced RXR␣ acetylation might recruit RXR␣ to its response element. Indeed, in the presence of p300, much more RXR␣ was found to bind to its response element RXRE, whereas RXR␣/ K145R that could not be acetylated by p300 failed to do so. These results demonstrated a good consistency between structural characteristics of RXR␣ and biological function of p300 on regulation of RXR␣ acetylation.
So far, little is known about how the acetylation of nuclear receptors is regulated. We unexpectedly found that the orphan receptor TR3 could act as a negative factor to regulate the p300-induced acetylation of RXR␣. TR3 seems to achieve its inhibitory effect on RXR␣ acetylation by competing with p300 for RXR␣ binding, because the interaction with RXR␣ was a prerequisite for TR3 to inhibit p300-induced RXR␣ acetylation. When RXR␣, TR3, and p300 coexisted in cells, RXR␣ preferred to the interaction with TR3 rather than p300. 9-cis retinoic acid further enhanced the TR3/RXR␣ binding but decreased the p300/RXR␣ interaction. Therefore, TR3 regulation seems to be highly contextual and may be dependent on the relative expression levels of TR3, RXR␣, and p300 in the cell. The cellular level of TR3 and p300 may act as a signal switch leading to diverse pathways: cell growth inhibition or cell proliferation. For example, a high level of TR3 may serve as a potent inhibitor of certain transcription factors by competing with p300 or disturbing p300 histone acetyltransferase activity directly. To address this possibility, more detailed analysis of other transcription factors should be carried out in future.
Previously, we found that RXR-selective retinoids SR11246 and SR11345 can induce the transcriptional activity of RXR␣ (28). However, we showed here that 9-cis retinoic acid, another well-known ligand for RXR␣, promoted RXR␣/TR3 translocation, rather than RXR␣ transcriptional activity. Such an effect of 9-cis retinoic acid appears to be contradictory to its known transcriptional activation of RXR␣. In fact, our previous study found that RXR␣ undergoes nucleocytoplasmic shuttling (3). Like other nuclear receptors, it is not always statically associated with chromatin (29)(30)(31). Thus, it is likely that the fate of cellular RXR␣ may depend on distinct stimuli and cell contexts. For example, p300 facilitates RXR␣ DNA binding through acetylation, whereas 9-cis retinoic acid enhances the interaction of TR3 with RXR␣ and induces TR3/RXR␣ translocation. The difference between current (using 9-cis retinoic acid) and previous (using RXR-selective retinoids SR11246 and SR11345) results can be explained by the possibility that whether transcriptional activity or translocation of RXR␣ will be activated may depend on the natural properties of the compound. In support of this viewpoint, a published report revealed that the nonsteroidal antiinflammatory drug R-etodolac binds RXR␣ and acts as a RXR␣ antagonist to inhibit its transactivation, an event that is asso- Fig. 8. A Proposed Model for TR3 to Negatively Regulate p300-Induced RXR␣ Acetylation ciated with its tumor-selective induction of apoptosis in animals (32). Our current observation that RXR␣ exhibits nucleocytoplasmic shuttling by itself and acts as a carrier for TR3 translocation from the nucleus to the mitochondria in response to 9-cis retinoic acid (3) also provides collateral evidence. Accordingly, RXR␣ nucleocytoplasmic localization appears to be one of the major factors determining 9-cis retinoic acid sensitivity.
Although increasing evidence supports the view that p300 can be under aberrant control in tumor cells (33), the importance of p300 in malignancy remains to be elucidated. RXRs usually form heterodimers with many members of the nuclear receptors (1, 2) and therefore play an essential role in regulation of multiple nuclear hormone-signaling pathways through their unique and potent dimerization capacity. Intriguingly, RXR␣ alone did not impair the BrdU incorporation rate, whereas coexpression of p300 increased the BrdU incorporation rate effectively, indicating that p300 might enhance the mitogenic ability by RXR␣ acetylation in HeLa cells. The fact that p300DY lost its ability to affect cell proliferation even in the presence of RXR␣ further verified that p300-stimulated cell proliferation was correlated with its ability to acetylate RXR␣. More importantly, overexpression of TR3 diminished p300induced cell proliferation by interacting with RXR␣ and cooperated with RXR␣ for apoptosis induction by translocating to the mitochondria in HeLa cells. There is no doubt that gaining further insights into the significance of this modification and translocation will be a benefit in facilitating the design of new approaches toward controlling malignancy of cancers.
In summary, different p300/RXR␣ and TR3/RXR␣ heterodimers in cells may exist in a dynamic equilibrium depending on their cellular environment. Abundant p300 may induce a dimerization interface switch that promotes RXR␣ DNA binding to ensure p300/ RXR␣ colocalization and efficient transcriptional regulation, which finally contributes to cell proliferation. In contrast, TR3 may sequestrate RXR␣ from p300 and preferentially heterodimerize with RXR␣ through their mutual interaction. The TR3/RXR␣ heterodimer then translocates to the mitochondria to induce cell apoptosis. Our study reveals an unexpected role of TR3 in the cross-talk between orphan receptor and retinoid receptors and suggests that competition of TR3 and p300 for the regulation of RXR␣ acetylation might control a dynamic process between cell proliferation and cell apoptosis.

Cell Culture and Transient Transfection
The cell lines cervical cancer HeLa, human breast cancer MCF-7, and human embryonic kidney (HEK) 293T were obtained from American Type Culture Collection (Rockville, MD). All the cell lines were cultured in DMEM supplemented with 10% fetal bovine serum, 100 IU penicillin, and 100 g/ml streptomycin. Transfection was performed using a calcium phosphate precipitation method for 293T cells as described previously (22) and liposomal transfection reagent (Fugene 6; Roche Molecular Biochemicals, Indianapolis, IN) for HeLa and MCF-7 cells according to the manufacturer's instructions.

Coimmunoprecipitation and Western Blotting
Cells were transfected with various plasmids as required and incubated for 36 h before analysis. Cell lysate preparation, immunoprecipitation, and Western blotting were performed as described previously (7). Briefly, cell lysates were incubated with the appropriate antibody for 1 h and subsequently incubated with protein A-Sepharose beads for 1 h. The proteinantibody complexes that were recovered on beads were subjected to Western blot analysis after separation by SDS-PAGE. The immunoreactive products were detected by using enhanced chemiluminescence (Amersham, Arlington Heights, IL).

Luciferase Assay
Cells were transfected with various plasmids, including luciferase-linked reporter gene (Luc-RARE) (34), ␤-galactosidase (␤-gal) expression vector, and the vector expressing RXR␣, TR3, or p300 as required. After 36 h after transfection, luciferase activity was normalized for transfection efficiency using corresponding ␤-gal activity. The ratios of luciferase/␤-gal activity were used as indicators for activity of RARE. All transfections were performed in duplicate, and the data are presented as means Ϯ SD of at least three separate experiments.

EMSA
EMSA was done using a LightShift chemiluminescent EMSA kit (Pierce Chemical Co., Rockford, IL) with biotin-labeled oligonucleotide, corresponding to cellular retinol binding protein type II (CRBPII) RXRE (AGCTTCAGGTCAGAGGTCA-GAGAGC; Invitrogen, Carlsbad, CA). After binding reaction, samples were loaded onto 6% polyacrylamide gel in 0.5ϫ Tris-borate-EDTA buffer and electrophoresed at 100 V at 4 C for 2 h. Biotin-labeled, double-stranded DNA was electrophoretically transferred to positively charged nylon membrane (Hybond-N; Amersham Pharmacia Biotech, Little Chalfont, UK). After cross-linking the transferred DNA to membrane, biotin-labeled DNA was integrated with streptavidin-horseradish peroxidase conjugate. Finally, chemiluminescence was detected using enhanced chemiluminescence. For the supershift experiment, 5.0 g nuclear extract was incubated with 200 ng anti-GFP antibody for 2 h (Santa Cruz) before incubation with the biotin-labeled probe.

ChIP Assay
Cells were cross-linked with 0.75% formaldehyde for 10 min, and then the cross-linking was stopped by adding glycine. Cells were collected in PBS and resuspended in FA lysis buffer [50 mM HEPES-KOH (pH 8.0), 140 mM NaCl, and 1% Triton X-100] with 1ϫ protease inhibitor cocktail (Roche) and then sonicated. After centrifugation, 10% of the total supernatant was saved as total input control. The remaining supernatant was diluted 10-fold in dilution buffer [1.0% Triton X-100, 2 mM EDTA, 150 mM NaCl, and 20 mM Tris-HCl (pH 8.0)] and divided into two parts, incubating with or without 2 g polyclonal anti-Myc antibody (Santa Cruz) overnight at 4 C separately. After immunoprecipitation, 30 l protein A/G-Sepharose was added, and cells were incubated for another 2 h at 4 C. Sepharose beads were washed sequentially in low-salt wash buffer [0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl (pH 8.0), and 150 mM NaCl] three times and in high-salt wash buffer [0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl (pH 8.0), and 500 mM NaCl] one time, and beads were then extracted with 200 l elution buffer (1% SDS, 0.1 M NaHCO 3 ). Two microliters of RNase A (0.5 mg/ml) were added and then heated at 65 C for 5 h to reverse the formaldehyde cross-linking. DNA fragments were purified with a DNA purification kit. For quantitative PCR, the primer sequence (RAR␤ promoter) was as follows: sense 5Ј-CTCCTCCCCTGCTCATTTTA-3Ј and antisense 5Ј-CTG-CCTCTGAACAGCTCACT-3Ј.

BrdU Assay
Cells were incubated with 5-bromo-2Ј-deoxyuridine (5-BrdU, 20 mol/liter; Sigma Chemical Co., St. Louis, MO) for 2 h. After washing with PBS, cells were fixed with 4% paraformaldehyde for 30 min at 4 C and then incubated with saponin (0.1%) for another 10 min. The cells were washed twice with PBS containing 0.1% saponin and resuspended in PBS containing 30 g DNase I. After incubation with anti-BrdU antibody (Santa Cruz) for 1 h, cells were given two PBS washes and then incubated with phycoerythrin-linked antimouse antibody. Finally, cells were washed with PBS, and analyzed by flow cytometry (Beckman Coulter, Inc., Fullerton, CA).