DRIP150 coactivation of estrogen receptor alpha in ZR-75 breast cancer cells is independent of LXXLL motifs.

Vitamin D receptor-interacting protein 150 (DRIP150) has been identified as part of mediator-like complexes that enhance transcriptional activation of the estrogen receptor (ER) and other nuclear receptors (NRs). DRIP150 coactivates ligand-dependent ERalpha-mediated transactivation in ZR-75 and MDA-MB-231 breast cancer cells transfected with a (luciferase) reporter construct (pERE3) regulated by three tandem estrogen-responsive elements. Coactivation of ERalpha by DRIP150 in ZR-75 cells was activation function 2-dependent and required an intact helix 12 that typically interacts with LXXLL motifs (NR box) in p160 steroid receptor coactivators. DRIP150 contains C- and N-terminal NR boxes (amino acids 1182-1186 and 69-73, respectively), and deletion analysis of DRIP150 showed that regions containing these sequences were not necessary for coactivation of ERalpha. Analysis of multiple DRIP150 deletion mutants identified a 23-amino-acid sequence (789-811) required for coactivation activity. Analysis of the protein crystal structure data base identified two regions at amino acids 789-794 and 795-804, which resembled alpha-helical motifs in Lanuginosa lipase/histamine N-methyltransferase and hepatocyte nuclear factor 1, respectively. By using a squelching assay and specific amino acid point mutations within each alpha-helix, the NIFSEVRVYN (795-804) region was identified as the critical sequence required for the activity of DRIP150. These results demonstrate that coactivation of ERalpha by DRIP150 in ZR-75 cells is NR box-independent and requires a novel sequence with putative alpha-helical structure.


Vitamin D receptor-interacting protein 150 (DRIP150) has been identified as part of mediator-like complexes that enhance transcriptional activation of the estrogen receptor (ER) and other nuclear receptors (NRs). DRIP150 coactivates ligand-dependent ER␣-mediated transactivation in ZR-75 and MDA-MB-231 breast cancer cells transfected with a (luciferase) reporter construct (pERE 3 ) regulated by three tandem estrogen-responsive elements. Coactivation of ER␣ by DRIP150 in ZR-75 cells was activation function 2-dependent and required an intact helix 12 that typically interacts with LXXLL motifs (NR box) in p160 steroid receptor coactivators. DRIP150 contains C-and N-terminal NR boxes (amino acids 1182-1186 and 69 -73, respectively), and deletion
analysis of DRIP150 showed that regions containing these sequences were not necessary for coactivation of ER␣. Analysis of multiple DRIP150 deletion mutants identified a 23-amino-acid sequence (789 -811) required for coactivation activity. Analysis of the protein crystal structure data base identified two regions at amino acids 789 -794 and 795-804, which resembled ␣-helical motifs in Lanuginosa lipase/histamine N-methyltransferase and hepatocyte nuclear factor 1, respectively. By using a squelching assay and specific amino acid point mutations within each ␣-helix, the NIFSEVRVYN (795-804) region was identified as the critical sequence required for the activity of DRIP150. These results demonstrate that coactivation of ER␣ by DRIP150 in ZR-75 cells is NR box-independent and requires a novel sequence with putative ␣-helical structure.
The estrogen receptor (ER) 1 is a ligand-activated transcription factor and a member of the nuclear hormone receptor superfamily (1)(2)(3)(4)(5). These proteins exhibit conserved structural domains (A-F) including a central DNA binding domain (DBD) (C), and a ligand binding domain that overlaps activation function 2 (AF2) in the C-terminal (E/F) region. Most nuclear receptors also contain a less conserved N-terminal AF1 (A/B) and a flexible hinge domain (D). The ligand-bound ER␣ (or ER␤) forms homo-or heterodimers that interact with specific response elements or other DNA-bound nuclear proteins in target gene promoters. A palindromic estrogen-responsive element (ERE) was first identified in the frog vitellogenin A2 gene promoter (6), and other functional consensus and nonconsensus EREs have been characterized in promoters of several E 2 -responsive genes (7). ER-mediated transactivation from ERE promoters is dependent on recruitment of an array of nuclear proteins including coactivators that facilitate interactions of ligand-bound ER with the basal transcription machinery (8 -14).
Steroid receptor coactivators (SRCs) include SRC1 (NCoA-1), SRC-2 (TIF2/GRIP-1/NCoA-2), SRC-3 (pCIP/ACTR/A1B1), and several related proteins represent one class of coactivators that enhance ER␣-mediated transactivation. Interactions of SRCs with ER␣ and other nuclear receptors is ligand-dependent, and SRCs also exhibit weak histone acetyltransferase and facilitate recruitment of other coregulatory proteins including CBP/300. SRC interactions with ER␣ and other nuclear receptors are dependent on one or more LXXLL motifs that interact with helix 12 in the AF2 region (15)(16)(17)(18)(19). Several other classes of nuclear receptor coactivators have now been identified, and their interactions with ER␣ and other receptors can be LXXLLdependent or -independent. For example, caveolin-1 interacts with ER␣ and coactivates ER␣-mediated transactivation, and the scaffolding domain of caveolin-1 is required for this response (20,21).
Another important class of coactivators resemble the mammalian Mediator complex, and these have been termed vitamin D-interacting proteins (DRIPs) or thyroid hormone-associated proteins (TRAPs) (22)(23)(24)(25). Many of the DRIPs/TRAPs have also been identified in other large coregulatory complexes such as NAT, ARC, CRSP, and SMCC (26 -29). DRIP205 and TRAP220 are identical to peroxisome proliferator-activated receptorbinding protein (30), and several studies show that interaction of ER␣ and nuclear receptors with the DRIP/TRAP complex occurs via DRIP205/TRAP220, which anchors the complex to the receptor.
DRIP205 contains two LXXLL motifs that are required for ligand-dependent interactions with nuclear receptors including ER␣ and ER␤ (31)(32)(33)(34)(35)(36). There is also evidence suggesting that DRIP205, other mediator complex proteins, and p160 coactivators/p300 coordinately interact to enhance ER-dependent transactivation (37,38). For example, a recent study analyzed ER␣coactivator interactions on the pS2 gene promoter by chromatin immunoprecipitation assays and showed a cyclic association and dissociation of DRIP205 and p160 SRCs but in opposite phases (38). This is consistent with different temporal patterns of coactivator-receptor interactions that may cooperatively enhance nuclear hormone receptor-dependent transactivation. It has also been reported that DRIP150 interacts with ER and other nuclear receptors in vitro (37)(38)(39). Moreover, analysis of coactivator assembly on the pS2 gene promoter by chromatin immunoprecipitation shows that at some time points, both ER␣ and DRIP150 are cross-linked to the promoter in the absence of DRIP205 (38). This study investigates coactivation of ER␣ by DRIP150 in ZR-75 breast cancer cells transfected with a construct containing three tandem EREs (pERE 3 ). DRIP150 coactivates ER␣ in ZR-75 cells, and coactivation is AF2-dependent. However, studies with mutant DRIP150 constructs show that coactivation of ER␣ is independent of the two LXXLL in NR box motifs present in the C-and N-terminal regions of DRIP150. Coactivation of ER␣ required a unique 23-amino acid sequence (amino acids 789 -811) in DRIP150, and mutational analysis has further identified a putative helical region at amino acids 795-804 (NIFSEVRVNY) that is required for hormone-induced transactivation.

MATERIALS AND METHODS
Cell Lines, Chemicals, and Biochemicals-The ZR-75 human breast cancer cell line was obtained from the American Type Culture Collection (ATCC, Manassas, VA), and cells were cultured in RPMI 1640 (Sigma) supplemented with 10% fetal bovine serum (FBS) (Summit Biotechnology, Fort Collins, CO). Medium was further supplemented with sodium bicarbonate, glucose, Hepes, sodium pyruvate, and antibiotic/antimycotic solution (Sigma). MDA-MB-231 cells were obtained from the ATCC and maintained in Dulbecco's modified Eagle's medium/ F-12 supplemented with charcoal-stripped FBS. Cells were maintained at 37°C with a humidified CO 2 /air (5:95) mixture. Phenol-free Dulbecco's modified Eagle's medium/F-12 media, phosphate-buffered saline, and E 2 were also obtained from Sigma. [␥-32 P]ATP (3000 Ci/mmol) was purchased form PerkinElmer Life Sciences, and poly[d(I-C)] was from Roche Applied Science. Restriction enzymes, 5ϫ luciferase lysis buffer, luciferin, and TNT7 in vitro translation kit were purchased from Promega (Madison, WI). Reagents for ␤-galactosidase analysis were obtained from Tropix (Bedford, MA), and anti-Xpress antibody was from Invitrogen. ER antibodies for gel mobility shift and coimmunoprecipitation assays and protein G plus-Agarose were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Recombinant human ER␣ protein was obtained from Panvera (Madison, WI), and all other chemicals and biochemicals were obtained from commercial sources at the highest quality available.
Oligonucleotides and Plasmids-The consensus estrogen-response element (ERE) probe used in gel mobility shift assays was synthesized by the Gene Technologies Laboratory (College Station, TX), and the sequence was 5Ј-GTC CAA AGT CAG GTC ACA GTG ACC TGA TCA AAG TT-3Ј. ER␣ expression plasmid was kindly provided by Dr. Ming-Jer Tsai (Baylor College of Medicine, Houston, TX). Expression plasmids for ER␣ mutants with deletion of amino acids 1-178 (HE19) and TAF1 containing D538N, E542Q, and D545N mutations were kindly provided by Dr. Pierre Chambon (Institut de Genetique et de Biologie Moleculaire et Cellulaire, Illkirch, France) and Dr. Donald McDonnell (Duke University, Durham, NC), respectively. cDNA encoding DRIP150 was kindly provided by Dr. Leonard P. Freedman (Merck). The expression plasmid for the GRIP-1 NR box polypeptide GAL4 fusion protein was also provided by Dr. Donald McDonnell (Duke University). The expression plasmid for the AF1 polypeptide was generated in this laboratory by cloning amino acids 1-180 of ER␣ into NheI/EcoRV site of pcDNA 3.0 . pcDNA 3.1 -His-LacZ was purchased from Invitrogen. The pERE 3 reporter containing three consensus ERE sites linked to a luciferase gene was created by cloning an oligonucleotide with three ERE elements into BamHI-HindIII cut pXP-2 plasmid (40). ER␣-GAL4 fusion protein was constructed as follows. First, the GAL4DBD fusion expression vector pM (Clontech) was digested with BamHI and HindIII, and the oligonucleotide sequence GAT CCG TGT CTG CAG ACG TCG ACA was inserted into this digested vector. This oligonucleotide was added to create more space between restriction enzymes BamHI and SalI in the polylinker of vector pM, providing a more efficient digestion of these two enzymes when cut simultaneously. This new vector, pM (ϩ10), was then used for construction of the pM-ER plasmid. Primers used for preparing GAL4DBD fusion protein with ER were upper primer CTG TGG ATC CGT ATG ACC ATG ACC CTC CAC ACC AAA and lower primer TCA TGG TCG ACT CAG ACT GTG GCA GGG AAA CC. After PCR amplification, the ER-cDNA fragment was digested with BamHI and SalI and cloned into pM(ϩ10) digested with BamHI/SalI to give pM-ER. The 17m5-GAL4-Luc plasmid containing five copies of the yeast GAL4 recognition motif linked to a luciferase reporter gene was provided by Dr. Patrick Balaguer (INSERM 458, Montpellier, France) and Tim Zacharewski (Michigan State University, East Lansing, MI).
Cloning of DRIP150 Mutants-The ⌬1145-1454 m1(mutant1) of DRIP150 was generated by the KpnI/XhoI digestion of plasmid pcDNA 3.0 -DRIP150. After cutting with KpnI/XhoI, the fragments were run on 1% agarose gel, and 3.5-kb fraction was eluted and ligated with KpnI/XhoI cut pcDNA 3.0 vector. Except for m1, all other clones expressing DRIP150 mutants were generated by PCR amplification, and primers used for preparing DRIP150 mutants and GAL4DBD fusion proteins are summarized in Table I and Table II. Xpress-tagged m2 and m11 DRIP150 mutants were generated by putting the appropriate fragments into Xpress-tagged HisA-pcDNA 3.1 vector. pMDRIP150 was generated by inserting DRIP150 into pM vector. pM23 and related point mutants were generated by inserting DRIP150 23 aa (789 -811 region) and the 23-aa region with the mutated DRIP150 aa 792 (Ala 3 Pro), 801 (Arg 3 Pro) or double mutant into the pM vector.
After PCR amplification, cDNA fragments of m2 and m3 were digested with KpnI/EcoRI and cloned into pcDNA 3.0 digested with KpnI/ EcoRI. cDNA fragments of m4, m5, m6, m7, m8, m9, m10, m11, and m12 were digested with KpnI/XbaI and cloned into pcDNA 3.0 digested with KpnI/XbaI. The cDNA fragment of Xpress-tagged m2 was digested with EcoRI/XbaI and cloned into HisA-pcDNA 3.1 digested with EcoRI/ XbaI. The cDNA fragment of Xpress-tagged m11 was digested with BamHI/XbaI and cloned into HisApcDNA 3.1 digested with BamHI/XbaI. The cDNA fragments of pMDRIP150, pM23, pM23A792P, pM23R801P, and pM23A792P/R801P were also digested with BamHI/XbaI and cloned into pM digested with BamHI/XbaI. Transient Transfection Assays-Cells were seeded onto 12-well plates in phenol-free Dulbecco's modified Eagle's medium/F-12 supplemented with 2.5% charcoal-stripped FBS. After 18 h, cells were transfected by the calcium phosphate method with 1 g of pERE 3 reporter plasmid, 0.25 g of a cytomegalovirus ␤-galactosidase expression plasmid, the appropriate ER␣ expression plasmid, and the appropriate DRIP150 expression plasmid. After 6 -8 h, cells were shocked with 25% glycerol in phosphate-buffered saline (PBS) for 75 s, rinsed once with PBS, and treated with either Me 2 SO or 10 nM E 2 in Dulbecco's modified Eagle's medium/F-12 plus 2.5% charcoal-stripped FBS for 36 h. Cells were harvested by scraping the plates in 100 l of 1ϫ lysis buffer (Promega). Thirty five l of the cell lysate was used for performing luciferase assays on a Lumicount Luminometer (Packard Instrument Co.). Thirty five l of the cell lysate was used for determining ␤-galactosidase activity on a luminometer. Normalized luciferase values were calculated by dividing the luciferase by the ␤-galactosidase activities for a given sample. Results were expressed as means Ϯ S.E. for at least three separate experiments for each treatment group and were compared with the Me 2 SO control group (arbitrarily set at 1) for each set of experiments.
Gel Electrophoretic Mobility Shift Assays-Five picomoles of synthesized ERE was labeled at the 5Ј end using T4-polynucleotide kinase and [␥-32 P]ATP. Plasmids containing the DRIP150, m1, m2, and m3 cDNAs were used to in vitro transcribe and translate the corresponding protein in a rabbit reticulocyte lysate system (Promega). Three l of recombinant human ER␣ (500 fmol) was mixed with 3 l of bovine serum albumin (500 ng/l), 2 l of poly(dI-dC) (1 g/l), 5 l of 5ϫ binding buffer (20 mM Hepes, 5% glycerol, 100 mM KCl, 5 mM MgCl 2 , 0.5 mM dithiothreitol, 1 mM EDTA), and 1 l of E 2 (3.5 ϫ 10 Ϫ7 M) to give a final concentration of 2.5 ϫ 10 Ϫ8 M E 2 , and incubated on ice for 15 min. In vitro translated DRIP150, m1, m2 or m3 were then added to the above mixture and incubated on ice for 5 min. To balance the volume, in vitro translated pcDNA 3.0 was also added. For supershift experiments, 2 l of normal IgG or ER antibody was added to the mixture after 5 min and then incubated on ice for an additional 5-10 min, and 5 l of 32 P-labeled ERE probe (120,000 cpm) was added to the reaction mixture, giving a final volume of 25 l. The mixture was incubated at 20°C for 15 min. Samples were then loaded onto 5% polyacrylamide gel and run at 110 V in 0.09 M Tris, 0.09 M borate, 2 mM EDTA, pH 8.3, for 2.5 h. The gel was dried and exposed to a phosphoscreen for 12 h, and protein-DNA binding was visualized by autoradiography using a Storm PhosphorImager (Amersham Biosciences).
Coimmunoprecipitation Assays-Two hundred l of reticulocyte lysate was mixed with 40 l of protein G plus-agarose and 17 l of ER antibody and shaken for 1 h at 4°C to preclear ER expressed in the reticulocyte. After 1 h, the mixture was centrifuged at 1,500 ϫ g for 5 min, and the supernatant was used to in vitro translate DRIP150, ER, and pcDNA 3.0 . DRIP150, ER␣, and pcDNA 3.0 were in vitro translated using [ 35 S]methionine in a rabbit reticulocyte lysate system (Promega, Madison, WI), and 10 l of the in vitro translated 35 S-DRIP150 was mixed with 2 l of in vitro translated 35 S-labeled ER and incubated with 13.2 ϫ 10 Ϫ7 M E 2 to give a final concentration of 100 nM E 2 on ice for 15 min. Ten l of ER␣ antibody was then mixed with protein G plusagarose (1:2 ratio) and added to the above mixture and incubated for 3 h at 4°C with shaking every 30 min. For samples not containing 35 S-DRIP150, only 35 S-labeled ER (2 l) was mixed with ER antibodyprotein G plus-agarose mixture and incubated for 3 h at 4°C as described above. PBS (1 ml) was then added to each sample, shaken for 30 s, and then centrifuged at 1,500 ϫ g for 5 min. After centrifugation, the supernatant was discarded, and the pelleted fraction (100 l) was mixed with 20 l of 1ϫ sample buffer (50 mM Tris-HCl, pH 6.8, 2% SDS, 0.1% bromphenol blue, 10% glycerol, and 100 mM dithiothreitol) containing ␤-mercaptoethanol. The sample was then boiled for 5 min, loaded onto SDS-polyacrylamide gel, and run at 150 V for 4 h. The gel was dried and exposed to a phosphoscreen for 3 days, and proteins were visualized by autoradiography using a Storm PhosphorImager (Amersham Biosciences).
Western Immunoblot Assays-COS-7 cells (from ATCC) were seeded in 6-well plates at a concentration of 200,000 cells/well in phenol-free Dulbecco's modified Eagle's medium/F-12 with 2.5% charcoal-stripped FBS. After 24 h, the media were removed, and serum and antibioticfree, phenol-free Dulbecco's modified Eagle's medium/F-12 was added to the wells. X-press-tagged DRIP150 m2 and DRIP150 m11 were trans-fected using the Lipofectamine transfection method (Invitrogen). After 6 h, the media were removed, and phenol-free Dulbecco's modified Eagle's medium/F-12 with 2.5% charcoal-stripped FBS was added, and cells were incubated for 36 h. Cells were then harvested in lysis buffer (50 mM Hepes, pH 7.5, 150 mM NaCl, 10% (v/v) glycerol, 1% Triton X-100, 1.5 mM MgCl 2 , 1 mM EGTA, 10 g/ml aprotinin, 50 mM phenylmethylsulfonyl fluoride, 50 mM sodium orthovanadate), placed on a rocker at 4°C to extract soluble protein, and centrifuged at 14,000 ϫ g for 10 min at 4°C. Protein was quantitated, and an equal amount of protein (150 g) was diluted with loading buffer, boiled, and loaded on 10% SDS-polyacrylamide gel. Samples were electrophoresed at 150 -180 V for 3-4 h. For samples containing in vitro translated Xpresstagged m2 and m11, Xpress-tagged DRIP150 m2 and DRIP150 m11 were translated in vitro in a rabbit reticulocyte lysate system (Promega), diluted with loading buffer, boiled, loaded on 10% SDS-polyacrylamide gel, and electrophoresed at 150 -180 V for 3-4 h. The separated proteins were transferred (in a buffer containing 48 mM Tris-HCl, 29 mM glycine, and 0.025% SDS) to polyvinylidene difluoride membrane (Bio-Rad). Specific proteins were detected by incubation with mouse monoclonal anti-Xpress antibody (1:5000 dilution) for 4 h, rinsed with distilled water (three times), followed by blotting with horseradish peroxidase-conjugated anti-mouse secondary antibody (1: 5000 dilution) for 1.5 h. The membrane was then washed with PBS/ Tween 20 (0.05%), and blots were exposed to chemiluminescent substrate (ECL) (PerkinElmer Life Sciences) and placed on Kodak X-Omat AR autoradiography film. The detected bands were scanned using a Sharp JX-330 scanner (Sharp Electronics Corp., Mahwah, NJ).
Statistical Analysis-Statistical differences between different treat-  GAA TTC CTA TGG TAG AGA ACG TGC AAA TTC  ⌬325-1454 m3 Upper primer, same as m2 pcDNA 3   Upper primer, same as A792P mpM23 pM Lower primer, same as R801P mpM23 ment groups were determined using Student's t test or analysis of variance (Fisher's protected Least Significant Difference), and the levels of significance were noted (p Ͻ 0.05). The results were expressed as mean Ϯ S.E. for at least three replicate determinations for each experiment.

RESULTS
DRIP150 Coactivation of ER␣-DRIP150 is a member of the mediator complex of proteins, and this study investigates co-activation of ER␣ by DRIP150 in ZR-75 cells transfected with pERE 3 . E 2 -dependent transactivation in this cell line is minimal in cells transfected with pERE 3 alone; however, E 2 responsiveness is observed after cotransfection with minimal amounts of ER␣ expression plasmid. This is due, in part, to overexpression of pERE 3 in the transfected cells, which results in limiting levels of ER␣. The three tandem consensus EREs are inserted upstream from a minimal TATA-luciferase, which , and DRIP150 expression plasmid (2.5-7.5 ng) treated with Me 2 SO or 10 nM E 2 , and luciferase activity was determined as described under "Materials and Methods." Significant (p Ͻ0.05) coactivation of E 2 -induced activity is indicated by an asterisk, and results are expressed as means Ϯ S.E. for at least three separate determinations for each treatment group. Coactivation was also observed using 10, 25, or 150 ng of ER␣ expression plasmid (data not shown). Significant coactivation by DRIP150 (or mutants) in this study represents an increase in the fold induction compared with that observed for E 2 alone. Hormone responsiveness was not observed in the absence of cotransfected ER␣. B, coactivation of pM-ER␣ by DRIP150. ZR-75 cells were transfected with pM (empty vector), pM-ER␣ (50 ng), or DRIP150 (200 or 400 ng), treated with E 2 or Me 2 SO, and luciferase activity determined as described under "Materials and Methods." Significant (p Ͻ 0.05) induction by E 2 (*) and coactivation by DRIP150 (**) is indicated. C, coactivation of ER␣ in MDA-MB-231 cells. Cells were treated at described in A, and significant (p Ͻ 0.05) coactivation by DRIP150 is indicated (**). Coactivation of HE19 (D) and TAF1 (E) by DRIP150. ZR-75 cells were treated with 10 nM E 2 or Me 2 SO, transfected with the indicated amounts of plasmids, and luciferase activity determined as described under "Materials and Methods." Significant (p Ͻ 0.05) coactivation is indicated (**). SNURF coactivation of ER␣ (D) and DRIP150 coactivation of ER␣ (E) serve as positive controls for these experiments.
has lower intrinsic E 2 responsiveness compared with constructs containing the human thymidine kinase or frog vitellogenin A2 gene promoters (6). E 2 induced luciferase activity in ZR-75 cells transfected with pERE 3 , and enhancement of this response was variable and dependent on the amount of cotransfected ER␣. Results in Fig. 1A show that E 2 induced a Ͼ10-fold increase in luciferase activity in ZR-75 cells transfected with pERE 3 , and this response was enhanced Ͼ3-fold after cotransfection with DRIP150. Maximal coactivation of ER␣ was observed using 2.5-5.0 ng of DRIP150 expression plasmid, and the coactivation response was decreased or squelched using higher amounts of DRIP150 plasmid. ZR-75 cells were also transfected with GAL4-luc and pM-ER␣, which contained fulllength ER␣ fused to the DBD of GAL4, and E 2 induced transactivation in this mammalian one-hybrid assay. This response was also enhanced by cotransfection with DRIP150 expression plasmid (Fig. 1B). Coactivation required higher amounts of transfected DRIP150 in this assay, and DRIP150 did not affect hormone-induced transactivation in cells transfected with pM alone, which contained the GAL4-DBD but not the ER␣ insert (data not shown). By using an approach analogous to that described in Fig. 1A, it was also shown that DRIP150 coactivates ER␣-mediated transactivation in ER-negative MDA-MB-231 breast cancer cells transfected with pERE 3 , and the fold enhancement was Ͼ5 (Fig. 1C). Thus, comparable coactivation results were obtained in both ER-positive (ZR-75) and ERnegative (MDA-MB-231) breast cancer cell lines.
ER␣ contains two major activation domains, and we therefore investigated the coactivation activity of DRIP150 in cells transfected with HE19 (deletion of AF1) and ER␣-TAF1, which contains three amino acid mutations in helix 12 (D538N, E542Q, and D545N), which inactivates AF2 (41). The results in Fig. 1D demonstrate that in ZR-75 cells transfected with pERE 3 and HE19, treatment with E 2 increased (Ͼ50%) luciferase activity, and this response was further enhanced by cotransfection with DRIP150. As a positive control for this experiment, we also observed coactivation of HE19 by the RING finger protein SNURF as reported previously (40). A higher level of coactivation by SNURF was observed, and this may be due, in part, to the cooperative coactivation of HE19 by SNURF and TATA-binding protein (40). E 2 induced luciferase activity in ZR-75 cells transfected with ER␣-TAF1; however, this response was not enhanced by DRIP150 (Fig. 1E), whereas in the same experiment, DRIP150 coactivated wild-type ER␣ (positive control).
These results suggest that DRIP150 primarily coactivates ER␣ through direct or indirect interactions with the AF2 domain, and this was further investigated by competition (squelching) experiments with NR box and AF1 proteins (AF1p). The results in Fig. 2A demonstrate that increasing amounts of the 2XF6 peptide (42), which contains two NR boxes fused to the yeast GAL4-DBD, significantly decreased DRIP150 coactivation of ER␣ in ZR-75 cells. In contrast, transfection with the AF1 protein that contains amino acids 1-182 from ER␣ did not significantly decrease (or squelch) DRIP150 coactivation of ER␣ (Fig. 2B), whereas this protein inhibited AF1-dependent activation of GC-rich promoter constructs by ER␣/Sp1 (43). These results suggest that coactivation of ER␣ by DRIP150 is primarily AF2-dependent.
Interactions of ER␣ and DRIP150 -Kang and coworkers (44) reported that in nuclear extracts containing DRIP complex proteins, both wild-type and the ligand binding domain of ER␣ interacted with DRIP150 in pulldown assays; however, DRIP150-ER␣ interactions were not observed unless DRIP205 was also expressed. Results illustrated in Fig. 3A used in vitro translated and radiolabeled 35 S-ER and 35 S-DRIP150 in coimmunoprecipitation experiments. ER␣ antibodies coimmunoprecipitate ER␣ alone and in combination with DRIP150, indicating that both proteins directly interact. We also investigated direct interactions of DRIP150 and ER␣ in a mammalian twohybrid assay in MDA-MB-231 cells transfected with pM-DRIP150 plus VP-ER␣ expression plasmids (Fig. 3B). E 2 had no effect in cells transfected with pM/VP (empty vectors), pM/ VP-ER␣, or pM-DRIP150/VP-ER␣; however, a significant induction of luciferase activity by E 2 was observed after cotransfection with pM-DRIP150 and VP-ER␣. This experiment was also carried out as a mammalian one-hybrid assay in MDA-MB-231 cells transfected with pM-DRIP150 and ER␣, and the results showed that ER␣-enhanced transaction was ligand-dependent (Fig. 3C). These results paralleled the complementary coactivation of pM-ER␣ by DRIP150 (Fig. 1B) and confirmed interactions between ER␣ and DRIP150. Interactions of the in vitro translated proteins (unlabeled) were also investigated in gel mobility shift assays using 32 P-ERE (Fig. 3D). Incubation of ER␣ and 32 P-ERE gave a retarded band (Fig. 3D, 1st lane), and coincubation with increasing amounts of DRIP150 increased intensity of the retarded band (2nd to 4th lanes), but a super-
shifted ternary complex was not observed. DRIP150 alone did not form a complex with 32 P-ERE (Fig. 3D, 5th lane), and the DRIP150-enhanced complex was supershifted by ER␣ antibodies (7th lane) but not by nonspecific IgG (6th lane). Thus, DRIP150 enhanced ER⅐ERE complex formation, and similar observations have been reported previously for other transcription factors (including ER␣) in gel mobility shift assays where two interacting proteins did not form a ternary complex with DNA; however, protein-DNA binding of one protein was enhanced by the other protein (45)(46)(47)(48).

Coactivation of ER␣ by Mutant DRIP150 Constructs-Wild-
type DRIP150 contains 1454 amino acids (aa) with two putative NR boxes at amino acids 1182-1186 and 69 -73. Fig. 4A summarizes the effects of wild-type DRIP150 and mutants containing deletions of aa 1145-1454 (DRIP150m1), 789 -1454 (DRIP150 m2), and 325-1454 (DRIP150m3) on coactivation of ER␣ in ZR-75 cells transfected with pERE 3 . The results show that DRIP150m1 was the only one of these deletion mutants that coactivated ER␣, suggesting that the C-terminal NR box was not required for coactivation, and the N-terminal NR box   FIG. 3. DRIP150-ER␣ interactions. A, coimmunoprecipitation. In vitro expression of 35 S-labeled DRIP150 or ER␣ were immunoprecipitated by ER␣ antibodies as described under "Materials and Methods." [ 32 P]pcDNA3 (empty vector) served as a control. B, mammalian two-hybrid assay. MDA-MB-231 cells were transfected with various constructs and treated with Me 2 SO or E 2 , and luciferase activity was determined as described under "Materials and Methods." Results are expressed as means Ϯ S.E. for three separate determinations for each treatment group, and significant (p Ͻ 0.05) induction by E 2 is indicated (*). C, mammalian one-hybrid assay. MDA-MB-231 cells were transfected with various constructs, treated with Me 2 SO or E 2 , and analyzed essentially as described in B. Significant (p Ͻ 0.05) estrogen-dependent coactivation of pM-DRIP150 by ER␣ is indicated (*). D, gel mobility shift assays. 32 P-ERE was incubated with ER␣, in vitro expressed DRIP150 in the presence or absence of IgG (nonspecific) or ER␣ antibodies, and examined in a gel mobility shift assay as described under "Materials and Methods." In a separate experiment, excess unlabeled ERE also decreased intensity of the specifically bound retarded band (Bound DNA).
was not sufficient for coactivation. The results also indicate that the 789 -1144 aa are required for coactivation of ER␣ on an ERE promoter. The results in Fig. 4A were obtained using 2.5 ng of wild-type/mutant DRIP150 expression plasmid; how-ever, similar results were observed for DRIP150 mutants over a range of plasmid concentrations in separate experiments (data not shown). Coactivation of ER␣ was further investigated with a series of DRIP150 mutants with deletions

FIG. 4. Coactivation of ER␣ by wild-type and variant DRIP150 and their interactions in gel mobility shift assays.
Coactivation of ER␣ by wild-type DRIP150 and mutants 1-3 (A), 4 -7 (B), and 9 -12 (C). ZR-75 cells were transfected with pERE 3 (50 ng), and wild-type variant DRIP150 expression plasmids, treated with Me 2 SO or 10 nM E 2 , and luciferase activity were determined as described under "Materials and Methods." Results are expressed as means Ϯ S.E. for three separate experiments for each treatment group. All experiments were carried out over a range of DRIP150 (wild-type/variant), and the maximal enhanced coactivation (fold) is reported. Results in C were carried out in several separate experiments and were combined to show the enhanced coactivation (fold), whereas results in A and B were carried out at the same time. Significant (p Ͻ 0.05) coactivation is indicated by an asterisk. D, gel mobility shift assay. 32 P-ERE and ER␣ were incubated with equal amounts of in vitro expressed wild-type or variant DRIP150 and analyzed in a gel mobility shift assay as described under "Materials and Methods." The specifically bound retarded band (Bound DNA) is indicated. E, expression of DRIP150 mutants. DRIP150m11 and DRIP150m2 (expressed tagged) were in vitro translated (lanes 1 and 2) or transfected into COS-7 cells (lanes 3 and 4), and aliquots of translated protein or whole cell lysates were analyzed by SDS-PAGE and immunoblot analysis as described under "Materials and Methods." of 977-1454 (DRIP150m4), 886 -1454 (DRIP150m5), 870 -1454 (DRIP150m6), and 865-1454 aa (DRIP150m7). The results summarized in Fig. 4B demonstrate that all of the mutants coactivated ER␣ in ZR-75 cells transfected with pERE 3 , and these were also observed with different amounts of expression plasmid. A VXXLL motif was present in DRIP150m6 but not DRIP150m7; however, the activity of both mutants as coactivators of ER␣ suggests that the VXXLL motif was not required for coactivation by the DRIP150 mutant constructs. Results summarized in Fig. 4, A and B, demonstrate that the C-terminal NR box of DRIP150 is not required for coactivation and that amino acids 789 -864 are necessary for coactivation.
An additional series of DRIP150 mutants containing deletions of 850 -1454 (DRIP150m9), 827-1454 (DRIP150m10), 812-1454 (DRIP150m11), and 1-77/865-1454 aa (DRIP150m12) were also investigated as coactivators of ER␣. The activities of these constructs were determined in separate experiments where there was some variability in the fold induction by E 2 and the amount of mutant DRIP150 expression plasmid required to give maximal coactivation. Therefore, data obtained for these constructs are reported as fold enhancement of coactivation compared with cells treated with E 2 alone (no coactivation). The results showed that all the DRIP150 deletion constructs coactivated ER␣. The deletion of the N-terminal NR box (DRIP150m12) did not result in loss of coactivation, showing that this motif was not necessary for DRIP150 coactivation of ER␣. Thus, results of deletion analysis of DRIP150 indicate that the 23 amino acids between aa 789 and 811 were required for coactivation of ER␣ in ZR-75 cells.
We also investigated interactions of ER␣ and 32 P-ERE in the presence or absence of in vitro expressed DRIP150 expression plasmids (Fig. 4C). The ER␣-ERE retarded band (first lane) intensity was enhanced after coincubation with wild-type DRIP150 (second lane) and deletion mutants m1, m2, and m3 (third to fifth lanes). Wild-type DRIP150 alone did not bind 32 P-ERE (fifth lane), and ER␣ antibodies (seventh lane) but not IgG (sixth lane) supershifted the retarded band as indicated in Fig. 3B. The enhanced ER␣-ERE retarded band intensity was observed after coincubation not only with wild-type DRIP150 and mutant m1, which coactivate ER␣, but also with mutants m2 and m3 that are inactive as coactivators. These results suggest that this response may reflect interactions of DRIP150 mutants with ER␣ in vitro, but these interactions did not predict their activities as coactivators of ER␣. DRIP150m2 and DRIP150m11 have similar molecular weights as illustrated in Fig. 4E in which in vitro expressed DRIP150 mutants were analyzed by SDS-PAGE and Western blot analysis (lanes 1 and  2). These proteins were also observed in whole cell lysates after transfection (Fig. 4E, lanes 3 and 4) in COS-7 cells.
Coactivation/Squelching by DRIP150 Coactivation Peptide-DRIP150m8 plasmid expresses amino acids 755-885, which encompass the region of DRIP150 required for coactivation of ER␣. Results in Fig. 5A show that DRIP150m8 coactivates ER␣ at low concentrations and squelches activity at higher concentrations. Moreover, the results summarized in Fig. 5B show that DRIP150m8 inhibits coactivation of ER␣ by wild-type DRIP150. DRIP150 and other deletion mutants also coactivate HE19, and the results in Fig. 5C show that DRIP150m11 coactivates ER␣, and cotransfection with DRIP150m8 inhibits or squelches the coactivation response. pM23 contains the minimal sequence of DRIP150 (aa 789 -811, DIPAHLNIF-SEVRVYNYRKLILC) necessary for coactivation of ER␣, and this peptide is fused to the DBD of the yeast GAL4 protein (Fig.  6A). Transfection of ZR-75 cells with pERE 3 and different amounts of pM23 expression plasmid showed that this chimeric protein coactivates ER␣ and then squelches this response with increasing amounts of transfected plasmid (Fig. 6B). This parallels a similar coactivation/squelching response observed for DRIP150m8 (Fig. 5A). Moreover, pM23 also inhibits DRIP150 coactivation of ER␣ (Fig. 6C), demonstrating that pM23 and DRIP150m8 exhibit comparable coactivation of ER␣ at low concentrations but also squelch transactivation (at higher concentrations) and inhibit DRIP150 coactivation of ER␣. We also examined the protein crystal structure data base for similarities between the DRIP150 amino acid sequence 789 -811 with other proteins. The first six residues DIPAHL fold into an ␣-helix when they occur in Lanuginosa lipase (49) and histamine N-methyltransferase (50). There was also homology between residues 7 and 16 (NIFSEVRVYN) of the DRIP150 23amino acid sequence and an ␣-helical region in hepatocyte nuclear factor 1 (HNF1; NLVTEVRVYN) (51). Results in Fig. 7, A and B, summarize squelching experiments with pM23R801P and pM23A792P, which express the GAL4 -23-amino acid fusion protein with mutations at amino acids 801 (Arg 3 Pro) and 792 (Ala 3 Pro). Proline residues were inserted to disrupt ␣-helical structure. The results show that pM23R801P did not squelch DRIP150 coactivation of ER␣ (Fig. 7A), whereas pM23A792P exhibited wild-type (pM23) activity and squelched DRIP150 coactivation of ER␣ (Fig. 7B). Squelching of DRIP150 coactivation of ER␣ was also not observed using the double mutant pM23A792P/R801P (Fig. 7C). These data suggest that the sequence at amino acids 795-804 in pMDRIP150, which resembles an ␣-helical motif in HNF-1, is an important structural feature of DRIP150 required for coactivation of ER␣. These data suggest that in addition to LXXLL motifs, other helical sequences in coactivators can play a role in coactivation of ER␣ and possible other nuclear receptors in breast cancer cells.

DISCUSSION
Several nuclear coregulatory complexes that associate with transcription factors and potentiate RNA polymerase II transcription have been identified, and many of their individual subunits are identical (23)(24)(25)(26)(27)(28)(29). The functions of the DRIP, TRAP, NAT, ARC, and CRISP coregulatory complexes are similar to that described for Mediator complexes initially purified from yeast. Interactions of these coregulatory complexes with NRs, including ER␣ and ER␤, have been investigated, and there is evidence in some cell lines that DRIP205 anchors the protein complex to NRs (23,25,31,44). Several reports have investigated DRIP205 coactivation of NRs including both ER␣ and ER␤. DRIP205-dependent coactivation of ER depends on both cell context and ER subtype, and the NR boxes of DRIP205 are required for coactivation (33)(34)(35)(36)(37)(38). Both DRIP205 and DRIP150 also directly interact with ER␣/ER␤, and other stud- ies confirm that DRIP150 interacts with the glucocorticoid and androgen receptors (44,52,53).
Previous studies (52) indicate that DRIP150 coactivated glucocorticoid receptor-mediated transactivation, and this response was AF1-dependent; however, coactivation of ER␣ by DRIP150 has not been extensively investigated. DRIP150 coactivates ER␣ in MDA-MB-231 and ZR-75 cells transfected with pERE 3 , and comparable enhancement of transactivation was observed in a mammalian one-hybrid assay in cells transfected with pM-ER␣/GAL4-luc (Fig. 1B). E 2 -dependent coactivation of DRIP150 by ER␣ has also been observed (Fig. 3C). DRIP150 also coactivated HE19 but not ER␣-TAF1 in ZR-75 cells, suggesting that the AF1 domain of ER␣ was not necessary for coactivation and that an intact helix 12 was required.
These results are in contrast with the reported AF1-dependent coactivation of glucocorticoid receptor by DRIP150 (53) but are comparable with previous studies (15)(16)(17)(18)(19) in several different cell lines showing that helix 12 is a critical surface of ER␣ that interacts with NR boxes of p160 coactivators. The importance of the AF2 region of ER␣ for coactivation by DRIP150 was supported by the inhibitory effects or squelching of enhanced transactivation in ZR-75 cells transfected with an NR box expression plasmid ( Fig. 2A). This construct contains two copies of the GRIP1 NR box and inhibits ER␣-mediated transactivation (42). In contrast, overexpression of an AF1 peptide (aa 1-182 of ER␣) ( Fig. 2A) did not affect coactivation of ER␣ by DRIP150, confirming the important role of AF2 of ER␣.
DRIP150 contains two LXXLL NR box motifs in the N- (69 -73) and C-terminal (1182-1186) regions. Their role in coactivation of ER␣ by DRIP150 has not been determined; however, some previous studies show that NR boxes are critical regions for the coactivation of NRs by DRIP205 (32)(33)(34). Research in this laboratory has demonstrated recently (54) that coactivation of ER␣ by DRIP205 in ZR-75 cells is complex, and coactivation does not require the NR boxes. Deletion analysis of DRIP150 (Figs. 4 and 5) shows that coactivation of ER␣ by DRIP150 deletion variants was NR box-independent, and a 23-amino acid sequence (aa 789 -811) was identified as an essential region for DRIP150 coactivation of ER␣ (Fig. 4). Wildtype DRIP150 coimmunoprecipitates ER␣ (Fig. 3A) as reported previously (44), and ligand-dependent DRIP150-ER␣ interactions were observed in a mammalian two-hybrid assay (Fig.  3B). However, in gel mobility shift assays DRIP150 does not form a DRIP150⅐ER␣⅐ERE ternary complex but enhances the ER␣-ERE retarded band intensity (Fig. 3D). The failure to observe a supershifted ternary complex is not unprecedented because previous studies report that ER␣ enhances Sp1/Sp3 DNA binding (48,55), cyclin D1 enhances ER␣ DNA binding, and human T-cell lymphotropic virus type I transcriptional activator (Tax) enhances CREB DNA binding and binding of other transcription factors in gel mobility shift assays (45)(46)(47). Most interestingly, the results also show that DRIP150 and DRIP150m1, which coactivate ER␣, also enhance the ER␣-ERE retarded band. However, mutants that are inactive as coactivators (DRIP150m2 and DRIP150m3) exhibit comparable activity in the gel shift assay (Fig. 4D). This suggests that enhancement of ER␣-ERE binding by DRIP150 variants is not predictive for coactivation of ER␣-mediated transactivation, which requires the 23-amino acid 789 -811 sequence.
We have further investigated the role of the DRIP150 "coactivation sequence" in hormone-induced transactivation using DRIP150m8 that contains amino acids 755-885 and pM23, which contains DRIP150 amino acids 789 -811 fused to the yeast GAL4-DBD. Transfection of either protein gave a biphasic response typical of many coactivators in which low concentrations resulted in coactivation of ER␣ and higher amounts of transfected plasmids subsequently decreased or squelched transactivation (Fig. 5). pM23 and/or DRIP150m8 also inhibit wild-type and mutant DRIP150 coactivation of ER␣ or HE19 (Fig. 5), and these responses are similar to those observed for NR box peptides containing LXXLL sequences (42). Our results confirm that DRIP150 interacts with ER␣ as reported previously (44) and coactivates ER␣ in ZR-75 (and MDA-MB-231) cells transfected with pERE 3 . The coactivator activity of DRIP150 alone in ZR-75 cells contrasts with previous reports showing that ligand-dependent recruitment of mediator complex proteins to ER␣ and other nuclear receptors requires DRIP205 as an anchor component for complex-receptor interactions (23,25,31,44). However, other reports show that DRIP150 alone interacts with nuclear receptors (52,53), and this has now been observed for ER␣ (Fig. 3A). A recent study (56) also reported isolation of a transcriptionally active coactivator CRSP-mediator complex that contained CRSP150/ DRIP150 but not DRIP205/Med220 (or Med70), suggesting that DRIP205 is not always required for a functional mediator coactivator complex. This is also supported, in part, by chromatin immunoprecipitation studies on the time-dependent recruitment of coactivators, such as SRCs and DRIPs, to the ERE of the pS2 gene promoter in MCF-7 cells (38). The results showed that at some time points, DRIP150 was associated with the pS2 promoter in the absence of DRIP205, suggesting a DRIP205-independent role for DRIP150 as a coactivator of ER␣, and this is consistent with the results of this study.
DRIP150 coactivation of ER␣ is independent of the two NR boxes and requires a 23-amino acid sequence DIPAHLNIF-SEVRVYNYRKLILC at 789 -811 (Figs. 4 and 5). By using the protein crystal structure data base, there was not a good match between the 23-amino acid DRIP150 sequence and other known crystalline proteins; however, the first six residues DIPAHL fold into an ␣-helix when they occur in Lanuginosa lipase and histamine N-methyltransferase (49,50). Amino acids 795-804 in DRIP150 are homologous to amino acids 69 -78 in hepatocyte nuclear factor 1, which also fold into an ␣-helix (51). pM23 efficiently squelches DRIP150 coactivation of ER␣ (Fig. 6C), and we used this assay to identify the function of the two helical components within the 789 -811-amino acid region of DRIP150. Results in Fig. 7 show that pM23A792P exhibited wild-type (pM23) squelching activity, whereas pM23R801P and pM23A792P/R801P (double mutant) did not squelch coactivation of ER␣ by DRIP150. These data suggest that the ␣-helical structure within the NIFSEVRVYN (amino acids 795-804) sequence is required for the activity of DRIP150 as a coactivator of ER␣.
In summary, results of this study uniquely identify a novel sequence in DRIP150 required for coactivation of ER␣ and demonstrate that LXXLL boxes in DRIP150 are not required for enhancement of ER␣-dependent transactivation. However, amino acids 795-804 within the critical 789 -811 region of DRIP150 required for coactivation of ER␣ are homologous with an ␣-helical region in hepatocyte nuclear factor 1 and therefore resemble the ␣-helical structure associated with NR boxes. Current studies are focused on the function of DRIP150 and the 789 -811-amino acid sequence in coactivation of other nuclear receptors including ER␣/Sp1-and ER␣/AP1-mediated transactivation and crystallization of the novel ␣-helical structures in DRIP150 (amino acids 789 -811).