Recruitment of Thyroid Hormone Receptor/Retinoblastoma-interacting Protein 230 by the Aryl Hydrocarbon Receptor Nuclear Translocator Is Required for the Transcriptional Response to Both Dioxin and Hypoxia*

The aryl hydrocarbon receptor nuclear translocator/ hypoxia-inducible factor (ARNT/HIF-1 (cid:1) ) mediates an organism’s response to various environmental cues, including those to chemical carcinogens, such as 2,3,7,8-tetrachlorodibenzo- (cid:2) -dioxin (TCDD or dioxin), via its formation of a functional transcription factor with the ligand activated aryl hydrocarbon receptor (AHR). Sim-ilarly, tissue responses to hypoxia are largely mediated through the HIF-1 heterodimeric transcription factor, comprising hypoxia-inducible factor-1 (cid:3) (HIF-1 (cid:3) ) and ARNT. The latter response is essential for a metabolic switch from oxidative phosphorylation to glycolytic present data that define the role of TRIP230 as a transcriptional coactivator of ARNT-mediated transcriptional responses.

The ability to adapt to environmental changes is essential for an organism's survival. The aryl hydrocarbon receptor nuclear translocator (ARNT) 1 /hypoxia-inducible factor-1␤ (HIF-1␤) is the dimerization partner of a large family of transcriptional factors that act as environmental sensors and control an organism's response to a wide array of environmental stimuli (1). This group of proteins comprises the basic helix-loop-helix-PER-ARNT-SIM domain (bHLH-PAS) family of transcription factors. One such factor, the aryl hydrocarbon receptor (AHR), is a ligand-activated transcription factor that binds environmental contaminants such as dioxin/TCDD. Ligand activation of AHR results in the disruption of the cytoplasmic AHR complex that contains two molecules of HSP90, a co-chaperone, p23, and one molecule of the immunophilin-like protein hepatitis B virus X-associated protein 2 (2)(3)(4)(5)(6). This, in turn, leads to the nuclear translocation of AHR, allowing it to form a functional heterodimeric transcription factor with ARNT, thereby activating target gene transcription (7,8).
Another bHLH-PAS transcription factor, HIF-1␣, forms a heterodimeric complex with ARNT/HIF-1␤. The HIF-1 complex (HIF-1␣/ARNT) activates at least 40 known target genes through binding its cognate DNA response element, the HIF-1␣ response element (HRE) (9), and is necessary for development, the inflammatory response, regulation of angiogenesis, glucose metabolism, and solid tumor growth and metastasis (10 -12). The mechanisms by which HIF-1␣ is regulated under normoxic conditions are reasonably well understood. Normoxic conditions supply proline hydroxylases with O 2 as a substrate for the hydroxylation of HIF-1␣ proline residues 402 and 564 (13,14). These hydroxylated residues are recognized by the Von Hippel-Lindau tumor suppressor E3 ligase complex (15). This in turn targets HIF-1␣ for ubiquitination and ultimately degradation by the 26 S proteasome complex (16 -19). HIF-1␣ is further regulated under normoxic conditions by hydroxylation of an asparaginyl group at amino acid 803 in its carboxyl-terminal activation domain (20). This asparaginyl hydroxylase, termed * This work was supported by NCI, National Institutes of Health, Grant CA28868. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
¶ Supported in part by a fellowship from the University of California Toxic Substances Research and Teaching Program.
ʈ Supported by Research Supplement to Underrepresented Minorities to CA28868.
Beyond the DNA binding of its cognate response element, the mechanisms whereby HIF-1 activates gene transcription are understood in only broad terms. Upon hypoxic stabilization, HIF-1␣ translocates to the nucleus and forms a functional heterodimeric transcription factor with ARNT. This complex is then capable of recruiting CBP through direct interactions with either HIF-1␣ or ARNT (24 -26). Furthermore, ARNT is capable of recruiting the p160/bHLH-PAS co-activators SRC-1 and nuclear co-activator-2/GRIP1/TIF2 (27), which also interact with AHR as well as other transcriptional coactivators (28,29). These proteins are incorporated into multimeric complexes, which interact with and modulate the activity of the core transcriptional machinery as well as modifying local chromatin structure (29).
The intrinsic functions of transcription factors are therefore modulated by the recruitment of ancillary proteins termed co-activators and co-repressors. The wide diversity of transactivators appears to recruit protein complexes consisting of common pools of these co-activator/co-repressor proteins. The recruitment and assembly of these proteins is dictated, at least in part, by the DNA binding transcription factor and the cisacting element it binds, allowing transmission of the genetic information in the transcriptional regulatory regions of target genes to the core transcriptional machinery. This level of specificity is required for the appropriate and controlled regulation of target gene transcription. Heterodimeric ARNT complexes facilitate the activation of target gene transcription by the formation of complexes with several known transcriptional modifiers, including those stated above as well as with Brahma-related gene 1 and components of the mediator complex (31,32). Elucidation of the molecular mechanisms underlying activated transcription by the AHR and HIF heterodimeric complexes is central to our understanding the complex pathologies responsible for a wide spectrum of human diseases including chemical carcinogenesis, solid tumor growth, and atherosclerosis.
The thyroid hormone receptor/retinoblastoma protein-interacting protein 230 (TRIP230) was identified as a potential co-activator of the thyroid hormone receptor (TR) (33). TRIP230 is capable of co-activating TR-mediated gene transcription in a ligand-dependent manner and harbors an LXXLL motif in its putative carboxyl-terminal TR interaction domain. We identified TRIP230 in a yeast two-hybrid screen as a potential interaction partner for ARNT. We reasoned that it might act as a transcriptional co-activator of ARNT-dependent gene activation. In this report, we present strong evidence that ARNT recruits TRIP230 as a transcriptional co-activator for the activation of both AHR-and HIF-1␣-regulated genes. Furthermore, we demonstrate that TRIP230 is essential for transactivation by both the AHR and HIF complexes.

MATERIALS AND METHODS
Plasmid Constructs and Vectors-The ARNT bait, pMP17c-ARNT⌬Q, used for yeast two-hybrid library screening has been described previously (27) and was used to screen a human fetal brain cDNA library cloned into pJG-4.2. For the identification of the ARNT interaction domain within TRIP230, ARNT⌬Q was transferred into the bait plasmid pBTM117c. TRIP230 yeast two-hybrid constructs have been described previously (33) or were generated by PCR and cloned in frame into the yeast expression vector pACT2. The hemagglutinin (HA)-tagged mARNT cDNA expression construct was the kind gift of Jerry Pelletier (34). Full-length mouse and human HIF-1␣ cDNAs were isolated from mouse and human yeast two-hybrid prey libraries and cloned into pcDNA3.1HisC. The VP-16 TRIP230 deletion mutants were constructed by excising the existing TRIP230 deletion mutants from pACT2 and cloned in frame into VP-16.
Cloning of TRIP230 by Yeast Two-hybrid Interaction Cloning-The yeast strain, L40C, has been described previously (35) and was a generous gift of Dr. John Colicelli. Approximately 2 ϫ 10 6 recombinants were transformed into L40C containing pMP-17c-ARNT⌬Q and plated onto medium lacking leucine, tryptophan, and histidine. Transformants that conferred growth were picked and isolated and reintroduced to the bait in L40C to confirm interaction. In order to identify the strongest interacting clones, single colonies were picked and grown individually in synthetic complete medium lacking leucine and tryptophan. To assay for dimerization capability, equal amounts of plasmid-containing L40C were resuspended in 500 l of modified Z-buffer (40 mM NaH 2 PO 4 , 60 mM Na 2 HPO 4 , 10 mM KCl, pH 7.0). Fifty l of 0.1% SDS and three drops of CHCl 3 were added, and samples were vortexed for 1 min. Two hundred l of 4 mg/ml O-nitrophenyl-␤-D-galactopyranoside in 0.1 M NaPO 4 (pH 7.0) was added to each sample, and reactions were incubated at 37°C. Cellular debris was pelleted by centrifugation, and the absorbance of supernatants was measured at 420 nM. All values were derived from two or more transformations; each determination was performed in triplicate, and final values are the result of four separate experiments. For mapping of the ARNT interaction domain within TRIP230, the EGY48 yeast strain was transformed with minimal deletions of TRIP230, cloned into pACT2 and pBTM117c-ARNT⌬Q, grown in the appropriate medium, and subjected to the liquid culture assay, as described above.
In Vivo Co-immunoprecipitation-For interaction studies with TRIP230, 293T cells were transfected with 5 g of either AHR, HIF-1␣, or HA-tagged mARNT or 5 g of untagged mARNT and 5 g of TRIP230 expression plasmid, and samples were prepared as described previously (27). Approximately 24 h after transfection, cells were harvested in 1ϫ LysS buffer (50 mM Tris (pH 7.8), 300 mM NaCl, 0.5% Nonidet P-40, 1 mM EDTA, 10% glycerol, 5 mM ␤-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride, 2 mM benzamidine, and 10 g each of leupeptin, antipain, and aprotinin per ml) for 30 min on ice. Cellular debris was pelleted by ultracentrifugation at 60,000 ϫ g in a Beckman L-65 centrifuge with an SW55 Ti rotor for 30 min at 4°C. Lysates were assayed for protein concentration with Bio-Rad protein assay solution, following the manufacturer's protocol. Affinity-purified antibodies directed against the influenza virus HA tag epitope or control mouse IgG were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Mouse antisera generated against TRIP230 have been described previously (33). For each reaction, 1 mg of cell lysate was incubated with 5 g of affinity-purified anti-HA or the corresponding preimmune IgG overnight at 4°C. Complexes were then mixed with 50 l of protein A-Sepharose beads (50% slurry) on a rotating wheel at 4°C for 1 h. The beads were then washed six times with LysS buffer, and proteins were eluted in SDS sample buffer and boiled for 3 min. Complexes were fractionated by SDS-PAGE, and Western blotting was performed with anti-TRIP230 antisera. Precipitated complexes were detected with horseradish peroxidase-linked goat anti-rabbit IgG and an ECL kit (Pierce).
Mapping of the ARNT Interaction Domain in TRIP230 -For mapping of the ARNT interaction domain within TRIP230, the EGY48 yeast strain was transformed with deletion mutants of TRIP230 cloned into pACT2 together with pBTM117c-ARNT⌬Q, grown in the appropriate medium, and subjected to the liquid culture assay, as described above. For mammalian two-hybrid assays with the GAL4-DBD-ARNT fusion, 250 ng of pG5E4T and 250 ng of pCMV-␤-galactosidase were co-transfected either alone or with 1 g of the GAL4-DBD-ARNT construct and 1 g of TRIP230 deletion mutant expression plasmid or empty plasmid vector into HEK 293T cells (hereafter referred to as 293T cells). Cells were harvested 20 h after transfection and analyzed as described above.
Transient Transfections and Reporter Gene Assays in Mammalian Cells-Luciferase reporter gene assays for analyzing the response to dioxin were performed as described previously (27). In the hypoxia experiments, Hepa-1 cells were maintained in ␣-minimal essential medium supplemented with 10% fetal calf serum and 100 units of streptomycin and penicillin per ml. Typically, 500 ng each of the 4ϫ HREdriven luciferase and pCMV-␤-Gal reporter plasmids was cotransfected with either 2 or 5 g of either TRIP230 cDNA expression plasmid or empty vector into 6-well plates. The final DNA concentration was adjusted to 6 g/well with empty expression vector, and transfection was achieved with 15 l of Superfect transfection reagent (Qiagen, Valencia, CA), according to the manufacturer's protocol. Approximately 16 -20 h after transfection, the plates were transferred to a hypoxia incubator and maintained at 5% CO 2 and 1% O 2 for another 18 h. Cells were then harvested and assayed as described previously (27).
Chromatin Immunoprecipitation Assay-Hepa-1 and MCF-7 cells were treated with either 5 ϫ 10 Ϫ9 M TCDD or vehicle for 30 -90 min or exposed to hypoxia for 6 -8 h. Cross-linking was achieved by adding formaldehyde to a final concentration of 1% in 7.8 M HEPES buffer at room temperature for 10 min. Cells were washed twice with ice-cold phosphate-buffered saline and collected in 1 ml of harvest buffer (10 mM Tris, pH 8.3, 10 mM dithiothreitol, 50 mM NaOH) and incubated at 42°C for 15 min, and then collected by centrifugation at 2,000 ϫ g. Cells were washed sequentially in ice-cold phosphate-buffered saline, Buffer 1 (0.25% Triton X-100, 10 mM EDTA, 0.5 mM EGTA, 10 mM HEPES, pH 6.5) and Buffer 2 (200 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 10 mM HEPES, pH 6.5). Cells were pelleted at 200 ϫ g at 4°C and resuspended in 0.3 ml of cell lysis buffer (50 mM Tris-HCl (pH 8.1), 10 mM EDTA, 1% SDS, Roche complete protease inhibitor mixture). Cell lysates were sonicated to yield DNA fragments ranging in size from 200 to 900 bp. Samples were centrifuged for 10 min at 4°C. Supernatants were diluted 10-fold to a final solution containing 20 mM Tris-HCl (pH 8.1), 1% Triton X-100, 2 mM EDTA, 150 mM NaCl, and complete protease inhibitor mixture. The solutions were precleared with 50 l of 50% slurry protein A or protein G-Sepharose containing 2.0 g of sheared salmon sperm DNA and the appropriate control IgG for 1 h at 4°C. Eluates were then incubated with affinity-purified antibodies specific to various factors overnight at 4°C followed by the addition of 50 l of 50% slurry of protein A or protein G-Sepharose and incubated at 4°C for an additional 2 h. Sepharose beads were pelleted and washed sequentially for 10 Table I. Single Cell Microinjection Assay-Hepa-1 or MCF-7 cells were seeded on poly-L-lysine-treated glass coverslips at subconfluent density, and microinjections were performed as previously described (36) with minor modifications. Briefly, cells were maintained in 10% fetal calf serum. Immediately after injection, cells injected with the xenobiotic compound-responsive element (XRE)-LacZ reporter were treated with either 5 ϫ 10 Ϫ9 M TCDD or 0.1% Me 2 SO and incubated for a further 18 h. Cells injected with the HRE-LacZ reporter were incubated overnight under normoxic conditions, and then cells were incubated for a further 24 h at either 1 or 20% O 2 .
RNA Interference, Reverse Transcription, and Real Time PCR Assays-To negate the expression of TRIP230, a small interfering RNA duplex targeting the sequence 5Ј-AAACTGCAGTCAGCTGCTCAG-3Ј of human (and mouse) TRIP230 as well as a scrambled control oligonucleotide, 5Ј-AATTCTCCGAACGTGTCACGT-3Ј, were synthesized by Qiagen. Hepa-1 cells were transfected with 100 nM small interfering RNA using Lipofectamine 2000 (Invitrogen) transfection reagent according to the manufacturer's protocol. 24 h post-transfection, the cells were treated with 10 Ϫ8 M TCDD for an additional 24 h or either 1% O 2 or 100 M CoCl 2 for an additional 9 h. To assess the effects of the small interfering RNA on TRIP230 protein levels, whole cell extracts were obtained, and 30 g of protein was subjected to SDS-PAGE. The gel was then transferred to a polyvinylidene difluoride membrane, and TRIP230 protein levels were assessed by Western blot analysis as described above.
For analysis of the effect of small interfering TRIP230 on endogenous CYP1A1 and VEGF induction, total RNA was isolated using Trizol reagent (Invitrogen), and reverse transcription was performed using a Taqman reverse transcriptase kit (Applied Biosystems) according to the manufacturer's protocols. 5 g of total RNA was used in a 20-l reaction and amplified by cycling between 25°C for 10 min, 48°C for 30 min, and 95°C for 5 min (Bio-Rad Icycler). To generate a standard curve, 5 l of each diluted cDNA was combined into one tube (10,000ϫ), and 10-fold serial dilutions were performed until a 1ϫ sample was reached. Endogenous CYP1A1, VEGF, and the ribosomal subunit 36B4 mRNA were analyzed by the Taqman real time PCR assay. Taqman primers for 36B4 (5Ј-AGATGCAGCCAGATCCGCAT-3Ј and 5Ј-GTTCTTGCCCAT-CAGCACC-3Ј), for CYP1A1 (5Ј-GCTTGAGTGAGAAGGTCACTCTCT-T-3Ј and 5Ј-CGATCGGCCAATGGTCTCT-3Ј), and VEGF (5Ј-CGCTGG-TAGACGTCCATGAA-3Ј and 5Ј-CTGTACCTCCACCATGCCAAGT-3Ј) and dual labeled Taqman probes containing Fam TM fluorescent dye on the 5Ј end and TAMRA fluorescent dye quencher on the 3Ј end (Integrated DNA Technologies Inc., Coralville, IA) for mouse CYP1A1 (5Ј-T-TTGGGCAAGCGAAAGTGCATCG-3Ј), mouse VEGF (5Ј-TCCCAGGC-TGCACCCACGACA-3Ј) and 36B4 (5Ј-CGCTCCGAGGGAAGGCCG-3Ј) were synthesized by Integrated DNA Technologies, Inc. CYP1A1 and VEGF mRNA levels were normalized relative to the amount of 36B4 mRNA. Taqman assays were performed using an Applied Biosystems Inc. 7700 machine. Real time PCR parameters were 50°C for 2 min and 95°C for 10 min, followed by 40 cycles at 92°C for 15 s and 60°C for 1 min.

RESULTS
Identification of TRIP230 as an ARNT-interacting Protein-In order to identify ARNT-interacting proteins, an exhaustive screen of several yeast two-hybrid cDNA libraries was performed, using as bait a cDNA (ARNT⌬Q) encoding the amino-terminal 687 amino acids of the mARNT protein fused to the LexA DNA-binding domain. The L40C yeast strain, which has been described previously (35), was transformed with the ARNT⌬Q bait plasmid, followed by transformation with one of several cDNA libraries fused to the B42 acid patch transcriptional activation domain. Clones positive for growth on medium lacking histidine were selected, isolated, and sequenced. (Fulllength ARNT could not be used as bait, because this construct generated colonies on the histidine plates at very high frequency, even in the absence of an interacting prey construct, due to the transactivation domain located toward the C terminus of ARNT.) Several of the clones encoded either aminoterminal fragments or, in one case, full-length HIF-1␣, suggesting that at least some of the clones identified encoded bona fide ARNT-interacting proteins. Clones with long open reading frames were reintroduced into yeast containing the bait plasmid and tested in a liquid culture assay for ␤-galactosidase activity. Several novel strongly interacting clones based on ␤-galactosidase activity were identified from the human fetal brain cDNA library screen. One of the clones isolated contained a cDNA with an open reading frame predicting a 395-amino acid polypeptide. This cDNA and corresponding polypeptide had 100% sequence identity with the COOH-terminal 395 amino acids and several hundred bases of the 3Ј-untranslated region of the 230-kDa thyroid hormone receptor/retinoblastoma-interacting protein, TRIP230 (33). Whereas this fragment retained the ability to interact with ARNT⌬Q after purification and reintroduction into yeast containing ARNT⌬Q, it did not interact with a full-length mouse AHR bait construct or the bHLH-PAS domain of HIF-1␣ (data not shown).
Interaction Studies with TRIP230 and ARNT-In order to determine whether a physical interaction between TRIP230 and ARNT can occur in mammalian cells, full-length TRIP230 cDNA (37) was co-expressed with either HA-tagged or wildtype ARNT in HEK 293T cells. Lysates precipitated with anti-HA IgG were subjected to SDS-PAGE, and immunoprecipitates were transferred to nitrocellulose. After Western blot analysis using anti-TRIP230 mouse antisera (33) for detection, a single band was identified at ϳ230 kDa after autoradiography (Fig. 1A). This band was not detected in samples precipitated with control mouse IgG or in samples transfected with untagged mouse ARNT expression plasmid precipitated with anti-HA. We were also interested in determining whether TRIP230 could interact with other bHLH-PAS proteins in mammalian cells; therefore, we also co-transfected 293T cells with the TRIP230 expression plasmid and either HA-tagged full-length AHR or untagged full-length HIF-1␣ cDNAs. For AHR studies, cells were exposed to either vehicle or 5 ϫ 10 Ϫ9 M TCDD for 1 h, or, in the case of HIF-1␣, cells were maintained under normoxic conditions or treated with the 100 mM CoCl 2 for 6 h to mimic hypoxia (38). Lysates were prepared as described above and subjected to SDS-PAGE and Western blot analysis. Neither affinity-purified anti-HA-nor anti-HIF-1␣treated lysates pulled down TRIP230, confirming in mammalian cells our previous observation in yeast that TRIP230 does not interact with either AHR or HIF-1␣ (data not shown).
We next chose to determine the ARNT interaction domain in TRIP230. Toward this end, we employed yeast and mammalian two-hybrid assays using deleted TRIP230 polypeptides. L40C yeast cells transfected with deletion mutant TRIP230 constructs fused to the GAL4 activation domain and assayed for ␤-galactosidase activity revealed that the interaction domain for ARNT overlaps amino acids 1583-1716 and that the carboxyl terminal LXXLL motif putatively required for interaction with the thyroid hormone receptor (33, 39) may not be required for interaction with ARNT ( Fig. 2A). To confirm this observation, plasmids containing TRIP230 cDNA fragments fused to VP16 and full-length mouse ARNT fused to the GAL4 DNA binding domain were co-transfected into HEK 293T cells along with a GAL4 UAS-thymidine kinase promoter luciferase reporter construct. TRIP230 mutants containing amino acids 1340 -1755 and 1583-1978 were equally capable of interacting with ARNT, in that these minimal mutants activated the GAL4 UAS reporter ϳ2-fold over ARNT alone, which was capable of robustly activating this reporter by virtue of the fact that full-length ARNT contains a transactivation domain capable of activating transcription in heterologous reporter systems. These assays again demonstrated that the deleted TRIP230 construct encoding amino acids 1716 -1978 containing the thyroid hormone receptor interaction domain failed to interact with ARNT (Fig. 2B). None of the TRIP230-VP16 mutants was capable of activating transcription when co-transfected with the empty GAL4 DNA binding domain vector (data not shown).
Taken together, these data suggest that the ARNT interaction domain in TRIP230 overlaps the 192-amino acid stretch from 1583 to 1755. Furthermore, the ARNT interaction domain is distinct from the thyroid hormone receptor interaction domain and may not require the carboxyl-terminal LXXLL motif.
TRIP230 Coactivates TCDD-mediated Gene Transcription-TRIP230 has been previously reported to act as a transcriptional coactivator for thyroid hormone receptor-dependent gene transcription (33). Given this known function of TRIP230, we undertook studies to investigate whether TRIP230 is involved in ARNT-dependent transcriptional processes. We first studied the TCDD response. In Hepa-1 cells, the transcriptional response to 5 ϫ 10 Ϫ9 M TCDD of the CYP1A1 promoter-driven luciferase vector, pGL4.2, was ϳ25-30-fold greater than that in unstimulated cells. This transcriptional activity was maximally enhanced ϳ2.5-fold when the cells were cotransfected with 5 g of TRIP230 expression plasmid (Fig. 3A), similar to its maximal coactivation of TR-dependent reporter gene activity (33).

FIG. 2. Mapping of the ARNT interaction domain within TRIP230 with yeast and mammalian two-hybrid assays.
A, the yeast strain L40C was co-transformed with the ARNT⌬Q fusion and TRIP230 minimal mutants fused to the GAL4 activation domain in the prey plasmid pGAD10. LacZ activity was determined using a liquid culture assay. A.A., amino acid. B, HEK 293T cells were co-transfected with 5ϫ GAL4 UAS-LUC and either control plasmid or 1 g of GAL4-DBD-ARNT and 1 g of different TRIP230 minimal mutants fused to the VP16 activation domain.
We employed the chromatin immunoprecipitation assay using specific antibodies against TRIP230 to determine whether this protein is present at activated sites during TCDD-dependent gene induction in mouse Hepa-1 and human MCF-7 cells. Primer sets used to amplify the promoter-proximal XRE element located ϳ500 nucleotides upstream of the transcriptional initiation sites of the mouse and human CYP1A1 genes are listed in Table I. Recruitment of ARNT and TRIP230 to the mouse and human CYP1A1 proximal enhancer region was monitored over a 90-min time span in response to TCDD. Cells harvested immediately after treatment with 5 ϫ 10 Ϫ9 M TCDD showed no detectable ARNT or TRIP230 over this element in either species. Strong ARNT binding to the XRE was seen at both 30 and 90 min on the human site and at 30 min at the mouse site. These data demonstrate that TRIP230 is recruited to the regulatory regions of the dioxin-inducible CYP1A1 gene in a dioxin-dependent fashion. Furthermore, occupation by TRIP230 at the proximal enhancer site is coincident with ARNT occupation, consistent with the notion that TRIP230 has access to these sites through a direct interaction with ARNT. Whereas we were unable to find an antibody directed against AHR that could precipitate CYP1A1 chromatin-containing complexes, it is well established that AHR is present at the XREs within the upstream enhancer and promoter regions during activated transcription (40,41).
TRIP230 Co-activation of Hypoxia-mediated Gene Transcription-To test the effect that TRIP230 might have on hypoxic gene induction, Hepa-1 cells were transfected as described above but with a 4ϫHRE-driven luciferase reporter plasmid. Hypoxia (1% O 2 ) alone caused an approximate 2.4 -3.0-fold increase in reporter gene activity. Increasing amounts of 1.0 -5.0 g of TRIP230 expression plasmid increased this activity in a dose-dependent fashion a further 15-fold (maximally) over what was observed under hypoxic conditions alone or an approximate 40-fold over normoxia (Fig. 4A). In contrast, in MCF-7 cells, TRIP230 only enhanced 4ϫ HRE-driven luciferase reporter activity 2-fold (Fig. 4B). Transient transfections in MCF-7 cells were repeated using 100 mM CoCl 2 in place of hypoxia, and results were nearly identical (data not shown), suggesting that hypoxia-mimicking agents can be substituted for hypoxia in these studies.
In order to ascertain whether HIF-1␣, ARNT, and TRIP230 are present at the regulatory regions of hypoxia-inducible genes during activated transcription, we performed chromatin immunoprecipitation analysis. MCF-7 cells were exposed to 1% O 2 for 6 h or maintained under normoxic conditions. Similar to the situation observed with the CYP1A1 enhancer, after hypoxic stimulation, we were able to record an increase in both ARNT and TRIP230 at the regulatory regions of two HIF-1-inducible genes in MCF-7 cells, namely the VEGF promoter and the EPO enhancer (Fig. 4C). Recruitment only occurred at the VEGF promoter in Hepa-1 cells (data not shown). This latter observation is consistent with previous reports that the EPO gene is silenced in the Hepa-1 cell line (42). Furthermore, we were able to record hypoxic recruitment of HIF-1␣ and the p160 coactivator, SRC-1 at these sites, consistent with their transcriptional activation by hypoxia. Control rabbit IgG was incapable of precipitating chromatin corresponding to the VEGF promoter or EPO enhancer, suggesting that the observed enrichment of these signals by specific antibodies was the result of a direct interaction between each antibody and its corresponding antigen. Similar chromatin immunoprecipitation and luciferase results were obtained using 100 mM CoCl 2 (data not shown).
TRIP230 Is Essential for ARNT-mediated Gene Transcription-Previous reports have demonstrated that CBP/p300 can contribute to the transcriptional activation of hypoxiaresponsive genes (24,26). Furthermore, we have previously demonstrated that SRC-1 is recruited by ARNT for dioxindependent gene activation (27). In order to determine the function of TRlP230 in TCDD-and hypoxia-dependent gene activation, we injected nuclei of Hepa-1 cells with mouse IgG raised against the N-terminal portion of TRIP230 (33) together with lacZ reporter genes. The reporter gene lacZ was placed under the control of the p36 minimal promoter and either four copies of a classic XRE or four copies of the HRE found in the mouse VEGF promoter (43). Injection of anti-TRIP230 IgG greatly reduced the effectiveness of TCDD at activating the XRE-dependent transcription unit (Fig. 5A). Similarly, co-injection of anti-TRIP230 with the HRE-LacZ reporter blocked activation after a 24-h incubation in hypoxic conditions (1% O 2 or 100 M CoCl 2 ) (Fig. 5B). Injection of  control IgG had no effect on either the dioxin or hypoxic response when co-injected with the corresponding reporter. However, in each case, co-injection of the TRIP230 expression plasmid with anti-TRIP230 IgG was sufficient to restore the respective reporter gene activation. Cells were injected with a mixture containing rhodamine-conjugated dextran and, therefore, could be scored by immunofluorescence (Fig. 5C).
These results indicate that the TRIP230 antibody specifically negates TCDD and hypoxia stimulation of an XRE or HRE reporter gene, respectively.
To corroborate these findings, we employed RNA interference to deplete endogenous levels of TRIP230 in Hepa-1 cells and assessed the effect of this treatment on the transcription of both TCDD-and hypoxia-inducible genes. Hepa-1 cells were transfected with a small inhibitory RNA oligonucleotide specific for (both mouse and human) TRIP230 or with a scrambled control oligonucleotide. Lysates were obtained from transfected cells at various times after transfection, and TRIP230 levels were assessed by Western blot analysis (Fig. 5D). TRIP230 levels were significantly reduced after 24 h, and the observed depletion persisted for at least 72 h after transfection. Twentyfour h after transfection, cells were treated with either Me 2 SO or 5 ϫ 10 Ϫ9 M TCDD for 24 h. CYP1A1 RNA levels were assessed by real time PCR and normalized against levels of the mRNA for the ribosomal protein, 36B4. Twenty-four hours after TCDD treatment, CYP1A1 levels were markedly reduced by over 60% compared with the levels observed after transfection with the control scrambled RNA (Fig. 5E). We then assessed the effects of TRIP230 depletion on the hypoxic induction of VEGF. Similar to the situation described above, depletion of TRIP230 significantly compromised VEGF induction (Fig. 5F). Whereas CoCl 2 treatment did not cause the same degree of induction of VEGF RNA as did exposure to 1% O 2 , transfection with small interfering TRIP230 also decreased the response to this agent.
Taken together, these last results provide unambiguous evidence that TRIP230 is essential for both TCDD-and hypoxiadependent transcriptional activation and indicate that the available nuclear pool of TRIP230 molecules is sufficient to coordinate activation of ARNT target genes. were injected with control rabbit IgG or TRIP230 anti-IgG and, where indicated, TRIP230 expression vector. C, the cells were probed with rhodamine-conjugated dextran to identify injected cells and stained for ␤-galactosidase activity. Photomicrographs of rhodamine-stained cells and the corresponding light micrographs after 5-bromo-4-chloro-3-indolyl-␤-galactopyranoside (X-gal) staining are shown. D, lysates extracted from Hepa-1 cells before and at regular intervals after transfection with TRIP230 small interfering RNA (siRNA) were subjected to SDS-PAGE and probed for TRIP230 with mouse TRIP230 antisera. E, Hepa-1 cells were transfected with small interfering TRIP230 (siTRIP230) or scrambled RNA duplex and treated 24 h later with either Me 2 SO or 5 ϫ 10 Ϫ9 M TCDD for 24 h. Total RNA was isolated and subjected to reverse transcription. CYP1A1 mRNA levels normalized against the mRNA levels for the ribosomal protein 36B4 were assessed by real time PCR, using specific primers for these gene products. F, similar procedures were performed as described in E except that cells were left under normoxic conditions or were treated with 100 M CoCl 2 or subjected to 1% O 2 for 9 h. Levels of VEGF mRNA were determined as described above.

DISCUSSION
In order to extend our knowledge of the molecular events underlying ARNT-dependent transcriptional processes, we employed a yeast two-hybrid screening method to identify putative interaction partners of the ARNT protein. Out of this screen, we identified several novel clones, one of which was subsequently described as the 230-kilodalton thyroid hormone receptor/retinoblastoma protein-interacting protein, TRIP230 (33). We present data that define the role of TRIP230 as a transcriptional coactivator of ARNT-mediated transcriptional responses.
Previous studies have demonstrated a requirement of coactivators for AHRC-mediated transactivation, including the p160 family of coactivators, receptor-interacting protein 140, estrogen receptor-associated protein, CBP/p300, the mediator complex, and components of Swi/Snf (26 -28, 31, 32, 44, 45). We can now extend this list to include TRIP230. Transient transfection of a TRIP230 expression plasmid with heterologous reporters responsive to either TCDD or hypoxia indicates that TRIP230 serves a co-activation function for both environmental responses. Furthermore, as assessed by chromatin immunoprecipitation assay, TRIP230 is recruited to sites of activated transcription at the CYP1A1 promoter in vivo in a TCDD-dependent fashion and to the VEGF promoter and the EPO enhancer in a hypoxia-dependent fashion. The functional ablation of TRIP230 by microinjection of antisera specific for TRIP230 abrogated expression of lacZ reporter genes governed by either AHRC or HIF-1. Furthermore, "knockdown" of TRIP230 by RNA interference substantially decreased transcription of endogenous dioxin-and hypoxia-responsive genes under activating conditions. TR appears to be the only member of the NR family of transactivators that is capable of interacting with TRIP230.
Similarly, TRIP230 appears to interact with ARNT but not AHR or HIF-1␣. Two independent studies had previously identified TRIP230 or a portion of it as a thyroid hormone-dependent interactor of TR (33,39), through a domain that harbored an LXXLL amino acid motif. LXXLL motifs are signatures of bona fide nuclear receptor co-activators (36, 46 -50). However, several studies have shown that these motifs are often not required for interaction with bHLH-PAS transcription factors (27,32). Likewise, we have determined that the ARNT interaction domain within TRIP230 is distinct from the TR interaction domain and that ARNT may not utilize the LXXLL motif of TRIP230 as an interaction interface. The specific determinants for ARNT interaction with TRIP230 and for other co-activators such as SRC-1 remain unknown.
The ARNT putative COOH-terminal activation or "Q-rich" domain is largely dispensable for transcriptional activation by the AHRC (41,51). We have previously shown that ARNT recruits SRC-1 through its amino-terminal helix-2 domain (27), and we have demonstrated here that ARNT also does not require its putative COOH-terminal activation domain to recruit TRIP230. It is likely that the true activation domain(s) of ARNT lie in the interface(s) it uses to recruit co-activators such as SRC-1 and TRIP230. Of interest, Kim et al. (52) recently demonstrated that the nuclear receptor co-activator CoCoA on its own did not co-activate estrogen receptor-mediated transcription in a heterologous reporter assay, but in the presence of the p160 co-activator GRIP1 (nuclear co-activator-2/TIF2), it enhanced estrogen receptor-mediated transcription to the same magnitude as other co-activators. Additionally, CoCoA transduced the GRIP1 co-activation signal directly through the GRIP1 bHLH-PAS domain, suggesting that simple co-transfection with a heterologous reporter can, under some circumstances, be insufficient to assess an activator's activation po-tential and, furthermore, that bHLH-PAS domains can serve as activation domains.
In conclusion, we have demonstrated that TRIP230 is indispensable for TCDD and hypoxia-mediated gene activation. ARNT is a common component of a large family of transcription factors. TRIP230 expression is likely to impact many of the corresponding ARNT-mediated transcriptional events.