583 The retinoblastoma protein-associated transcription repressor RBaK interacts with the androgen receptor and enhances its transcriptional activity

K Hofman, J V Swinnen, F Claessens1, G Verhoeven and W Heyns Laboratory for Experimental Medicine and Endocrinology, Catholic University of Leuven, Gasthuisberg, Herestraat 49, B-3000 Leuven, Belgium 1Division of Biochemistry, Catholic University of Leuven, Gasthuisberg, Herestraat 49, B-3000 Leuven, Belgium

(Requests for offprints should be addressed to W Heyns; Email: [email protected])

Abstract

In search of potential androgen receptor coregulators we performed a yeast two-hybrid screening using the androgen receptor ligand-binding domain as bait and a human prostate cDNA library as prey and found that the carboxy-terminal domain of retinoblastoma-associated Krüppel protein (RbaK), a member of the Krüppel zinc finger protein family, interacts in a ligand-dependent way with the ligand-binding domain of the androgen receptor. RBaK was recently identified as a transcriptional regulator that interacts with the retinoblastoma protein and thereby influences E2F regulated transcription. The interaction of RBaK with the androgen receptor was further documented using mammalian two-hybrid experiments, in vitro binding studies and coimmunoprecipitation. Finally, we demonstrated that both RBaK and the retinoblastoma protein coactivate androgen receptor-mediated transcription in cotransfection experiments. In conclusion, our data show that RBaK interacts with the androgen receptor and increases its transcriptional activity. Moreover, the double interaction of RBaK with the retinoblastoma protein and with the androgen receptor provides a novel link between the androgen receptor and the regulation of the cell cycle. Journal of Molecular Endocrinology (2003) 31, 583–596

Introduction 1996). Moreover, the AR may be activated indepen- dently from androgen action, e.g. by protein kinases, The androgen receptor (AR) is a member of the which in turn may be regulated by growth factors large family of nuclear receptors (NRs) that activate (Nazareth & Weigel 1996, Darne et al. 1998, Wen (or repress) gene expression upon ligand binding et al. 2000). (Evans 1988, Parker 1993, Tsai & O’Malley 1994, During the last few years there has been growing Mangelsdorf & Evans 1995, Mangelsdorf et al. evidence that the activity of steroid hormone 1995) and plays an essential role in the differenti- receptors is modulated further through interaction ation, development and normal function of male with intracellular coregulator proteins, which sex organs. After androgen binding the activated increase or repress the hormone response by receptor interacts with androgen-responsive ele- various mechanisms such as histone acetylation ments present in the regulatory regions of andro- (Pazin & Kadonaga 1997) or chromatin remodeling gen-responsive genes and thereby changes their (Pollard & Peterson 1998). Most of these coregu- transcriptional rate. In addition to such direct lators interact with several steroid hormone genomic effects, androgens also exert more indirect receptors, but receptor specificity has been claimed effects through the regulation of other intracellular for some of them. A common approach for the pathways (Swinnen et al. 1997) or through changes detection of novel coregulators consists of an in the secretion or activity of several growth factors initial screening of a suitable cDNA library for or growth factor receptors (Schuurmans et al. 1991, proteins interacting with an NR by means of the Yan et al. 1992, Zuck et al. 1992, Hall et al. 1993, yeast two-hybrid system, followed by confirmation Horton et al. 1993, Fasciana et al. 1996, Kim et al. of the interaction by various techniques and

Journal of Molecular Endocrinology (2003) 31, 583–596 Online version via http://www.endocrinology.org 0952–5041/03/031–583 © 2003 Society for Endocrinology Printed in Great Britain

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access 584 K HOFMAN and others · Androgen receptor and RBaK interaction

demonstration of coregulatory properties by means the DNA-binding domain (DBD) of Gal4 (Gal4 of cotransfection. DBD)) and pACT2 (expressing the activation In this report we identified one of the proteins domain (AD) of Gal4 (Gal4AD)) were also obtained interacting with the ligand-binding domain (LBD) from Clontech. The yeast vector pGBT9 (Gal4 of the AR in the yeast two-hybrid system as RBaK. DBD) was a gift of Dr Michael R Stallcup RBaK (for retinoblastoma-associated Krüppel pro- (University of Southern California, Los Angeles, tein) was recently described (Skapek et al. 2000) CA, USA). Dr Stallcup also provided pGBT9 based on its interaction with retinoblastoma (Rb), a constructs containing the LBD of the following protein that not only plays a crucial role in the receptors in frame with the Gal4 DBD: human regulation of cell proliferation (Weintraub et al. AR (AR644–919), mouse glucocorticoid receptor 1995, La Thangue 1996, Dyson 1998, Harbour & (GR513–783), human estrogen receptor (ER
274–595), Dean 2000) but also has coactivator properties for human retinoic acid receptor (RAR
186–462), human the AR (Lu & Danielsen 1998, Yeh et al. 1998) thyroid hormone receptor (TR
121–410), human vita- and for other NRs (Singh et al. 1995). RBaK is min D receptor (VDR95–427) and human retinoid a member of a large family of transcriptional X receptor (RXR
197–462). Dr Hinrich Gronemeyer regulators that contain multiple Krüppel type zinc (IGBMC, Strasbourg, France) provided the expres- fingers linked to a KRAB (Krüppel-associated box) sion vectors for TIF2 (pSG5-TIF2 and pSG5- repressor motif at its amino-terminal side. As TIF2·5). pSG5-VDR was a gift of Dr Carsten shown by Skapek et al. (2000), RBaK interacts with Carlberg (Heinrich-Heine Universitaet, Dusseldorf, Rb through the linker sequence between the Germany). Dr Tony Kouzarides (University of KRAB domain and the zinc finger region and Cambridge, UK) kindly provided pcDNA3-Rb. affects the important control function of Rb in the Finally, pMOR (expressing the mouse ER
) cell cycle, through regulation of the activity of and p(Gal4)5-TK-luciferase were provided by E2F transcription factors (Weintraub et al. 1995, Dr Malcolm G Parker (Imperial College, Faculty La Thangue 1996, Dyson 1998, Harbour & Dean of Medicine, London, UK). The modified 2000). pSNATCHII expression vector (expressing the In this communication we show that the VP16 AD) was described in Alen et al. (1999a). carboxy-terminal part of RBaK mediates the interaction of RBaK with the AR. This interaction DNA manipulations was confirmed by binding studies and by coimmunoprecipitation experiments. Finally, we Plasmids were constructed and isolated according demonstrate that RBaK coactivates AR-mediated to standard methods. The Gal4 DBD-AR(LBD) transcription in COS-7 cells and in CV-1 cells, fusion construct was generated by amplification of even when the transcriptional activity of the AR in the hinge region and LBD (amino acids (aa) CV-1 cells is already increased by cotransfection 624–919) of the human AR with the forward with Rb. For that reason RBaK may be considered primer 5
-gcgacgCATATGACTCTGGGAGCCC as a novel AR coactivator. Moreover, it is GGAAGCTGAAGAAACTT-3
and the reverse conceivable that the double interaction of RBaK primer 5
-atctatGGATCCTTCACTGGGTGTG with the AR and with Rb plays a role in the GAAATATAGGGGCTTGAC-3
. The under- regulation of cell proliferation by androgens, a lined sequences are NdeI and BamHI restriction process wherein the control of E2F activity by Rb is sites respectively. The amplified NdeI-BamHI an important step. digested product was then cloned in frame in the homologous sites of the pAS2-1 vector (Clontech) yielding the plasmid pAS2-AR(LBD). The TIF2·5 Materials and methods (aa 624–869) fragment (Voegel et al. 1998) was amplified using 5
-atatatGGATCCGAGAGAGCT Yeast strain and plasmids GACGGGCAGAGCAG-3
as forward primer, The yeast strain Y190 (Saccharomyces cerevisiae) was 5
-atatatGGATCCCTAGCTCTGTGAAATTCG obtained from Clontech Laboratories as a compo- CAGTCG-3
as reverse primer and the pSG5- nent of the MATCHMAKER two-hybrid system. TIF2·5 plasmid as template. The amplified and The yeast shuttle vectors pAS2-1 (which expresses BamHI digested product was then ligated into the

Journal of Molecular Endocrinology (2003) 31, 583–596 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access Androgen receptor and RBaK interaction · K HOFMAN and others 585

BamHI site of pACT2 (Clontech). The complete protocol. Yeast was transformed via the lithium open reading frame (ORF) of RBaK was obtained acetate procedure as in the protocol of Gietz et al. as follows: RNA was prepared from normal human (1995). For the cycloheximide resistance selection, prostate tissue as described by Swinnen et al. (1994). yeast was plated out on medium containing cDNA was synthesized using the RT-PCR kit from 10 µg/ml cycloheximide. Transformants were Life Technologies according to the manufacturer’s allowed to grow at 30 C until the colonies were protocol. The complete ORF of RBaK was sufficiently large to assay for -galactosidase amplified using 5
-aaatGGATCCATGAACACA activity. In short, transformed cells were filter TTGCAGGGGCC-3
as forward primer and plated on a sterile Whatman Number 5 filter. Cells 5
-aaatGGATCCTCAGAGATTTTCCACATCA were then permeabilized by a cycle of freezing in AG-3
as reverse primer. The amplified product liquid nitrogen and thawing at room temperature. was subcloned in pGEM-T (Promega) and Each filter was then soaked into an appropriate sequenced. After insert excision from pGEM-T amount of Z-buffer containing 5-bromo-4-chloro- with BamHI, the RBaK ORF was cloned in the 3-indolyl---galactopyranoside and incubated at sense (S) and anti-sense (AS) direction into a 30 C until blue colonies appeared (between 0·5 pSG5-HA vector which expresses the viral and 8 h). hemagglutinin (HA) epitope, yielding pSG5-HA- For quantitative studies of NR transactivation RBaK(S) and pSG5-HA-RBaK(AS) respectively. activity, a liquid -galactosidase assay was used. The pSG5-HA vector itself was prepared by His+/LacZ+ yeast transformants were grown to cloning a double stranded oligonucleotide (5
- stationary phase in the single dropout (SD) medium CCACCATG*TACCCATACGATGTTCCAGAT SD/Trp/His/Leu containing 40 mM TACGCT-3
with the Kozak sequence in italics 3-amino-1,2,4-triazole (3-AT) and the appropriate and ATG* as start codon) encoding the HA epitope concentration of ligand. The saturated culture was with sticky EcoRI ends in the homologous EcoRI diluted in YPD medium to an A600 of 0·2 and the  site of the pSG5 vector. The pRSET-A-RBaK yeast was grown to midlog phase (A600 0·6; carboxy-terminal domain (CTD: RBaK417–714) and approximately 3 h). Then 1·5 ml of the culture the pSNATCHII-RBaKCTD sense and anti-sense were centrifuged, the pellet was washed once constructs were generated from the pACT2- with 1·5 ml Z-buffer and resuspended in 300 µl RBaKCTD plasmid picked up in the yeast Z-buffer. One hundred and fifty microliters of cell two-hybrid screening by insert excision with BglII. suspension were then permeabilized by two cycles The insert, which also contains the HA epitope of freezing in liquid nitrogen and thawing at 37 C. sequence from the pACT2 vector 5
in frame with Finally, 50 µl were used for chemiluminescent the RBaKCTD sequence, was then cloned in the detection of -galactosidase using Clontech’s corresponding BamHI sites of the pRSET-A luminescent -galactosidase genetic reporter system (Invitrogen) and of the pSNATCHII vector. To II. To test the ligand-dependency of the inter- generate the pSG5-Rb construct, the Rb ORF was action, His+/LacZ+ yeast transformants were amplified from the pcDNA3-Rb construct using grown in SD/Trp/Leu medium supple- 5
-aaatggatCCACCATGCCGCCCAAAACCCCC mented with or without ligand. Subsequently, the CG-3
as forward primer and 5
-aaatggatcc liquid -galactosidase assay was performed. The TCATTTCTCTTCCTTGTTTGAGG-3
as re- following ligands were used: 10 nM methyl- verse primer. The amplified, BamHI digested trienolone (R1881, a synthetic androgen) for the product was cloned in the corresponding BamHI AR, 0·1 µM estradiol for the ER, 50 µM dexam- site of pSG5. The pSG5-3 FLAG-AR expression ethasone for the GR, 1 µM 1,25-dihydroxyvitamin construct, the AR deletion constructs and the AR D3 for the VDR, 1 µM all-trans retinoic acid for the point mutations were constructed as described RAR, 10 µM 9-cis retinoic acid for the RXR and previously (Alen et al. 1999b). 10 µM triiodothyronine for the TR. All hormone stock solutions were prepared in ethanol, except triiodothyronine, which was dissolved in water. Yeast manipulations The final concentration of ethanol never exceeded Standard microbiological techniques and media 0·1%. Where appropriate the same concentration were used according to the manufacturer’s of ethanol was added to the control incubations. www.endocrinology.org Journal of Molecular Endocrinology (2003) 31, 583–596

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access 586 K HOFMAN and others · Androgen receptor and RBaK interaction

Two-hybrid screening the purified (His)6-fusion protein had been bound ff pAS2-ARLBD was transformed into Y190 using or with 50 µl TALON alone. NETN-bu er (0·02% the lithium acetate method. The transformed NP40, 20 mM Tris–HCl pH 8·0, 100 mM NaCl  and 1 mM EDTA) was added to a final volume of colonies were grown overnight in SD/ Trp to  ensure that they contained the bait plasmid. 250 µl. After incubation for 3 h at 4 C on a slowly Thereafter, the culture was transformed with the rotating wheel, the matrix was washed four times with 1 ml NETN buffer. Bound proteins were human prostate cDNA MATCHMAKER library ff (obtained from Clontech) and double transformants eluted by adding 50 µl SDS loading bu er and were plated out on SD/Trp/Leu/His boiling. Finally, the eluted proteins were separated plates containing 40 mM 3-AT and 10 nM R1881. by SDS-PAGE and visualized by autoradiography. Then the plates were incubated at 30 C until colonies appeared. His+ clones were picked up, restreaked at least once on triple minus plates and Transfections of mammalian cells then assayed for the LacZ phenotype. CV-1 cells or COS-7 cells (seeded at a density of Library plasmids from His+, LacZ+ clones 15 000 cells/well of a 48-well plate) were cultured in were rescued via cycloheximide selection and Dulbecco’s modified Eagle’s medium (Life Tech- retransformed alone, with pAS2-ARLBD or with a nologies) containing 10% fetal calf serum. Cotrans- Gal4 DBD-laminin control expression plasmid. fection experiments were performed using the Library plasmids that activated LacZ only in Fugene-6 method (Roche) as previously described the presence of the bait plasmid were then (Hofman et al. 2000) with (per well) 100 ng reporter sequenced using autoread-sequencing kits and an construct (C3 ARE-luciferase), 0·17 ng pSG5-AR ALF automated laser fluorescence sequencer and various amounts of pSG5 expression constructs (Amersham-Pharmacia). as indicated in the figures. For the mammalian two-hybrid experiments, cells were transfected with (per well) 100 ng p(Gal4)5-TK-luciferase reporter construct, 0·67 ng pAB-Gal4 DBD-ARLBD con- In vitro binding experiments struct and 0·67 ng pSNATCHII, pSNATCHII- RBaK(S) or pSNATCHII-RBaK(AS) construct To study the interactions of various 35S-labeled respectively. NRs with the fragment of RBaK isolated dur- ing the screening (RBaKCTD: RBaK417–714), RBaKCTD was prepared as a (His) -fusion protein 6 Preparation of nuclear extracts and immobilized on Talon matrix (Clontech). Therefore the pRSET-A-RBaKCTD plasmid was For the preparation of nuclear extracts, COS-7 transformed in BL21(DE3)pLysS Escherichia coli cells cells were seeded out at a density of 3106 cells (Stratagene). (His)6-fusion proteins were expressed in a 14 cm plate. After 24 h of culture the cells and purified under denaturing conditions on the were transfected with (per plate) 4 µg pSG5-3 TALON-matrix according to the manufacturer’s FLAG-AR and 4 µg pSG5-HA-RBaK. In the protocol. For the expression of the AR, GR, ER, control conditions 4 µg pSG5-AR and 4 µg VDR, Rb or luciferase labeled with [35S]methio- pSG5-HA-RBaK or 4 µg pSG5-HA-RBaK and nine (.1000 Ci/mmol, Amersham) the corre- 4 µg pGEM7zf+ were transfected. pGEM7zf+ was sponding pSG5 expression constructs were used as used to keep the amount of transfected DNA templates in the TNT-coupled T7 transcription constant. Sixteen hours after transfection cells were translation reticulocyte lysate system (Promega). treated for 24 h with 10 nM R1881, a strong AR The synthesis of these receptors was performed in agonist. Then, the cells were washed twice with both the presence (10 nM) or absence of their 3 ml ice-cold PBS, collected in 1·5 ml ice-cold PBS, respective ligand (R1881 for the AR, estradiol for transferred to a prechilled microcentrifuge tube the ER, dexamethasone for the GR and 1,25- and centrifuged down briefly. After removal of the dihydroxyvitamin D3 for the VDR). supernatant the pellet was resuspended in 400 µl To study the interaction, 22·5 µl 35S-labeled freshly made cold hypotonic buffer. This buffer lysate were incubated with 50 µl TALON to which consisted of 4 volumes 10 mM Hepes–KOH buffer

Journal of Molecular Endocrinology (2003) 31, 583–596 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access Androgen receptor and RBaK interaction · K HOFMAN and others 587

(pH 7·9), 10 mM KCl and 1·5 mM MgCl2 and 1 Results volume of a solution containing one Minitablet of Complete Protease Inhibitors (Roche) dissolved in Characterization of RBaK as an AR-interacting 1·5 ml water. The resuspended pellet was incu- protein by yeast two-hybrid screening bated on ice for 10 min, vortexed for 30 s and To screen for proteins interacting with the LBD of centrifuged for 10 s. The supernatant was removed the AR (ARLBD), the yeast strain Y190, previously and the nuclear pellet was incubated in 100 µl transformed with pAS2-ARLBD (expressing the ice-cold high-salt buffer. This buffer consisted of Gal4 DBD-ARLBD fusion protein), was trans- 3 volumes 20 mM Hepes–KOH buffer pH 7·9, formed with a human prostate-derived cDNA 420 mM KCl, 25% glycerol, 1·5 mM MgCl2 and library expressed as fusion proteins with the 1 volume of the protease inhibitor solution. This Gal4AD. Approximately 4106 yeast transform- mixture was further incubated on ice for 20 min ants were screened in the presence of 10 nM and then vortexed for 10 s. Debris were centrifuged R1881, a strong AR agonist. One of the clones down for 2 min at 4 Cat14000g and the showing strong ligand-dependent interaction with supernatant was transferred to a prechilled the ARLBD encodes a 297 amino acid long microcentrifuge tube. Finally, the protein concen- peptide. The complete ORF of this clone was tration of the nuclear extract was measured using sequenced and the corresponding peptide was bicinchoninic acid after trichloroacetic acid precipi- found to be 100% identical to the carboxy-terminal tation (Brown et al. 1989). Nuclear extracts were part of RBaK, a recently identified protein that stored at 70 C. interacts with Rb (Skapek et al. 2000).

Immunoprecipitation RBaK–NR LBD interaction in yeast For the immunoprecipitation of FLAG-tagged AR, In yeast, the fragment of RBaK isolated in the yeast the EZview Red ANTI-FLAG M2 affinity gel two-hybrid screening (Gal4AD-RBaK AA417–741) (Sigma) was used. Five microliters of the 50% bead interacted with the Gal4 DBD-ARLBD in a slurry were used for 50 µg nuclear extract protein ligand-dependent way, suggesting that R1881 (see above). This bead slurry was washed twice with induced a conformational change of the AR 100 µl TBS (50 mM Tris–HCl, pH 7·4 and required for protein–protein interaction. Indeed, 150 mM NaCl) and centrifuged down for 10 s at growth of yeast transformed with pAS2-ARLBD 4 C. After addition of 50 µg nuclear extract and pACT2-RBaKCTD in SD/Trp/Leu/ protein in 100 µl lysis buffer (50 mM Tris–HCl pH His medium was only observed in the presence 7·4, 150 mM NaCl, 1 mM EDTA and 1% Triton of the ligand. Moreover, when the positive clone X-100) the mixture was incubated for 2 h at 4 C was grown in SD/Trp/Leu medium there on a slowly rotating wheel. Then, the mixture was was little or no -galactosidase activity (measured centrifuged down for 10 s at 4 C and the in the liquid assay) in the absence of ligand, but the supernatant was removed. Thereafter, the beads activity increased markedly as a function of the were washed twice with 500 µl TBS by incubation concentration of R1881 (Fig. 1A) with a maximal on a rotating wheel for 15 min at 4 C. Finally, the 60-fold stimulation at 10 nM. Near maximal beads were boiled for 5 min in 20 µl 1sample stimulation was reached at 3 nM. For yeast buffer and the eluted proteins were size fraction- transformed with pAS2-ARLBD alone no ated onto an SDS-PAGE gel and visualized by -galactosidase activity was observed in the immunoblotting. For the detection of HA-tagged presence or in the absence of R1881. RBaK we used a monoclonal antibody directed In order to evaluate receptor specificity, the against the viral HA epitope and directly labeled Gal4AD-RBaKCTD construct was cotransformed with horseradish peroxidase (HRP) from Roche. A in yeast with Gal4 DBD-LBD constructs of the AR, rabbit polyclonal antibody raised against an GR, ER, VDR, RXR, RAR and TR. As positive amino-terminal peptide (aa 1–20) was used for the controls we looked at the interaction between the detection of the AR (Aumüller et al. 1998) in Gal4 DBD-LBD of these NRs and Gal4AD- combination with goat-anti-rabbit immunoglobins/ TIF2·5. Whereas the Gal4AD-TIF2·5 construct HRP from Dako as secondary antibody. produced no galactosidase activity in itself, www.endocrinology.org Journal of Molecular Endocrinology (2003) 31, 583–596

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access 588 K HOFMAN and others · Androgen receptor and RBaK interaction

A 120 construct in the presence of ligand, but these

100 interactions were very weak in the liquid units -galactosidase test (Fig. 1B). 80

60 In vitro interaction of RBaK with the AR -galactosidase β 40 To confirm the RBaK-AR interaction in vitro, 35 20 S-labeled full-length AR was added to (His)6- Relative RBaKCTD coupled to TALON matrix or to the 0 -10 -10 -9 -9 -8 -8 -7 TALON matrix alone. As shown in Fig. 2, a clear 1.103.101.10 3.10 1.10 3.10 1.10 M R1881 interaction of the AR with the RBaKCTD was observed in vitro. This in vitro interaction was ligand-independent – in contrast to the interaction B 12 TIF2.5 in yeast – and not antagonized by the anti- RBaK 10 androgen hydroxyflutamide (data not shown). units Using similar procedures we also tested the in vitro 8 interaction between RBaKCTD and the GR, ER, 6 VDR or Rb. The interaction of RBaKCTD with -galactosidase

β luciferase was used as a negative control. Under 4 these conditions the strongest interaction was 2

Relative observed with the AR but there also was clear

0 interaction with the GR and to a lesser extent with AR GR ER VDR TR RAR RXR the ER (Fig. 2). Only weak interaction was seen Figure 1 Interaction of RBaKCTD with various NRLBDs with the VDR or Rb. The lack of interaction with in yeast. (A) Dose-dependent interaction between Rb is not unexpected, since it is not the RBaKCTD RBaKCTD and the Gal4 DBD-AR(LBD). The but the linker region between the amino-terminal -galactosidase activity was measured in the presence of the indicated concentration of R1881. Yeast were KRAB domain and the zinc finger domain of grown in double selective SD/−Trp/−Leu medium. The RBaK that is required for the interaction with Rb -galactosidase activity in the presence of 10 nM (Skapek et al. 2000). In our hands this region R1881 was arbitrarily set to 100 and other values of RBaK, however, did not interact with the were calculated as relative values. (B) The interaction AR either in vitro or in yeast (data not shown). of RBaKCTD and TIF2·5 with the LBD of several NRs in yeast in double selective SD/−Trp/−Leu To further characterize the region of the AR medium in the presence of the appropriate ligand was required for the interaction with RBaK, AR evaluated by means of the liquid -galactosidase assay deletion mapping experiments were performed (see Materials and methods). The -galactosidase using carboxy-terminally truncated receptors, since activity obtained for the interaction between the RBaK was isolated in a screening with the Gal4 DBD-AR(LBD) and Gal4AD-TIF2·5 was arbitrarily set to 1 and the other values were calculated as ARLBD. Based on the alignment of Würtz et al. relative values. The values shown are means±S.D. (1996), we generated the following deletion -galactosidase activities of three independent mutants: AR-1 (AR1–723) ending after H3 (aa experiments performed in triplicate. 698–720), AR-2 (AR1–697), AR-3 (AR1–684) ending after H1 (aa 677–684), AR-4 (AR1–671) ending after the hinge region (aa 627–670) and significant ligand-dependent interactions with AR-5 (AR1–627) ending after the DBD (aa TIF2·5 were observed for all NRs (Fig. 1B). For the 539–627). As shown in Fig. 3, the binding was well RBaKCTD, however, only the LBD of the AR conserved for the AR-1 mutant, which no longer displayed a strong ligand-dependent interaction binds androgens, but most of the binding was lost (Fig. 1B), which was even stronger than that when H3 was deleted (AR-2). This observation observed for TIF2·5. It should be noted that LacZ+ suggests that the H3-containing domain strongly yeast colonies did arise when the Gal4 DBD- contributes to the interaction between the AR-LBD GR(LBD) or the Gal4 DBD-ER(LBD) construct and RBaK-CTD. Remarkably, point mutations of were cotransformed with the Gal4AD-RBaKCTD residues located in H3 and previously reported to

Journal of Molecular Endocrinology (2003) 31, 583–596 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access Androgen receptor and RBaK interaction · K HOFMAN and others 589

A B

Input Talon RBaK-CTD Talon Input RBaK-CTD Input Talon RBaK-CTD 35 30 25 AR GR VDR 20 15 10 5

% of binding to RBaK-CTD 0 Rb luc ER AR GRER VDR Rb luc Figure 2 In vitro interaction between RBaK and NRs. (A) Full-length 35S-NRs, 35S-Rb and 35S-luciferase were synthesized using the TNT-coupled reticulocyte lysate method and incubated with TALON matrix alone or with

(His)6-RBaKCTD bound to TALON in the presence of 10 nM R1881 (AR), 10 nM dexamethasone (GR), 0·1 µM 1,25-dihydroxyvitamin D3 (VDR), 10 nM estradiol (ER) or ethanol (Rb and luc). The input lanes represent 10% of the amount of 35S-protein applied for each condition. (B) Three independent experiments as shown in (A) were performed and the protein signals were densitometrically quantified. Mean values±S.D. are shown in the graph.

A B

Talon Input RBaK-CTD Talon Input Input Talon RBaK-CTD RBaK-CTD 35 30 25 AR-∆1 AR-∆2 AR-∆3 20 15 10 5

% of binding to RBaK-CTD 0 AR-∆4 AR-∆5 luc

∆1

∆3

∆5

∆2

∆4

luc

WT

AR-

AR-

AR-

AR- C AR-

538 626 WT 1 919

538 626 ∆1 1 723 538 626 ∆2 1 697

538 626 ∆3 1 684

538 626 ∆4 1 671

538 626 ∆5 1 627 Figure 3 In vitro interaction between RBaKCTD and AR deletion mutants. (A) In vitro interaction studies were performed as in Fig. 2 using 35S-AR and 35S-AR deletion mutants as described in the text. (B) Three independent experiments as shown in (A) were performed and the protein signals were densitometrically quantified. Mean values±S.D. are shown in the graph. (C) Schematic representation of the AR deletion mutants. www.endocrinology.org Journal of Molecular Endocrinology (2003) 31, 583–596

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access 590 K HOFMAN and others · Androgen receptor and RBaK interaction

be essential for the interaction of the AR with other these expression constructs, the addition of ligand coactivators (L712R, V715/716A, W718A, A719K increased the Gal4 driven luciferase response for and K720A) did not markedly influence the the RBaK(S) construct (Fig. 5), but not for the binding of RBaK to the AR. Deletions of H1 did other constructs. Consequently, the AR-LBD not reduce the binding further and even after interacts with RBaK in a ligand-dependent way in deletion of the hinge region (AR-5) there still was a mammalian as well as in a yeast two-hybrid stronger binding of the deletion mutant to the system, in contrast to the results of the in vitro RBaK-CTD than observed for the luciferase binding and coimmunoprecipitation experiments, control. This observation suggests that the amino- wherein the interaction of the AR with RBaK was terminal domain, NTD (or DBD) of AR contrib- ligand-independent. utes to the interaction of the whole AR with RBaK, although one cannot fully exclude strong non- Influence of RBaK on AR-mediated specific binding. It should be noted, on the other transcription hand, that the LBD in itself interacts in a ligand-dependent way with RBaK in yeast and To evaluate the potential of RBaK as an AR mammalian (see below) two-hybrid experiments. coregulator, cotransfection experiments were per- formed in CV-1 cells and COS-7 cells. In these experiments the effect of a sense RBaK expression Coimmunoprecipitation of RBaK with the AR construct (RBaK(S)) was compared with an To demonstrate that RBaK interacts in vivo with antisense RBaK expression construct (RBaK(AS)) the AR, coimmunoprecipitation experiments were or to the pSG5 1 control. This construct contains performed using FLAG-tagged AR and an an unrelated cDNA fragment (NuRIP-6) without antibody against the FLAG sequence for immuno- coregulator properties but which has approxi- precipitation. Nuclear extracts were prepared from mately the same size as the RBaK insert and COS-7 cells cotransfected with pSG5-3 FLAG-AR therefore produces nearly the same suppression of together with pSG5-HA-RBaK(S) and grown in the concentration of the AR (Hofman et al. 2000). the presence or absence of 10 nM R1881. In the As shown in Fig. 6, very similar concentrations of control conditions, COS-7 cells were transfected the AR in the presence of ligand were obtained for both with pSG5-AR and pSG5-HA-RBaK(S) or the three constructs. In COS-7 cells, however, the with pSG5-HA-RBaK(S) alone. As shown in expression of the AR was stronger than in CV-1 Fig. 4A, RBaK was nearly undetectable in the cells and therefore required a shorter exposure immunoprecipitate when coimmunoprecipitation time. In both cell lines the cotransfection with experiments were performed with non-flagged AR RBaK(S) markedly increased the ligand-induced or in the absence of AR. However, there was activity of the C3 ARE-Luc-reporter construct, as marked coimmunoprecipitation of HA-RBaK with compared with cotransfection with RBaK(AS) or the AR when the cells were cotransfected with with the inactive 1 control construct. For that flagged AR and HA-RBaK, indicating that the reason, the cotransfection data support a role of RBaK protein was associated with the AR. The RBaK as an AR coactivator. degree of coimmunoprecipitation of RBaK, how- ever, was not influenced by the presence of Influence of Rb on RBaK-mediated AR androgen (Fig. 4B). coactivation Since the amino-terminal half of RBaK interacts Mammalian two-hybrid interaction with Rb (Skapek et al. 2000), which in turn may The interaction between the AR and RBaK was interact with the NTD of the AR and coactivate its also analyzed by means of a mammalian two- function (Singh et al. 1995, Lu & Danielsen 1998, hybrid experiment, using a Gal4 DBD-ARLBD Yeh et al. 1998), there could be a link between these fusion construct and VP16AD-RBaK fusion con- interactions with formation of a ternary complex. structs wherein RBaK was cloned in the sense or in However, the similar degree of AR coactivation in the anti-sense orientation or VP16AD alone as a COS-7 cells and in CV-1 cells suggests that the control. After cotransfection of COS-7 cells with coactivation by RBaK is not mediated by Rb.

Journal of Molecular Endocrinology (2003) 31, 583–596 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access Androgen receptor and RBaK interaction · K HOFMAN and others 591

A d

Input Transfected Precipitate Immunodetection

FLAG-AR AR + HA-RBaK HA-RBaK

AR AR + HA-RBaK HA-RBaK

AR

HA-RBaK HA-RBaK

B

Input Input Precipitated Precipitated

HA-RBaK

- R1881 + R1881 Figure 4 Coimmunoprecipitation of RBaK with the AR. (A) Nuclear extracts were prepared from COS-7 cells, grown in the presence of 10 nM R1881 and cotransfected with expression vectors for the proteins as indicated in the figure. Nuclear extracts were incubated with an ANTI-FLAG M2 affinity gel. Bound proteins were eluted by boiling the gel in 1×sample buffer. Proteins were separated by SDS-PAGE and electroblotted onto a nitrocellulose membrane, which was probed with a polyclonal antiserum against the AR or with a monoclonal antiserum against the HA-epitope (see Materials and methods). Immunoreactive proteins were visualized by chemiluminescence. Input lanes are equivalent to 50% of the amount applied for the AR and 4% of the amount applied for RBaK. (B) Influence of ligand on the coimmunoprecipitation of RBaK. COS-7 cells were cotransfected with pSG5-3FLAG-AR and with pSG5-HA-RBaK and incubated without ligand or with 10 nM R1881. A monoclonal antiserum against the HA epitope was used for immunodetection.

www.endocrinology.org Journal of Molecular Endocrinology (2003) 31, 583–596

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access 592 K HOFMAN and others · Androgen receptor and RBaK interaction

12 Series2- R1881 Discussion

ty Series1

i + R1881

v 10 i From the data presented we may conclude that the 8 LBD of the AR interacts with RBaK, a protein

erase act 6 recently identified based on its interaction with the if

uc Rb protein (Skapek et al. 2000), a key regulator of l 4

ve the cell cycle. Whereas the interaction of RBaK i at l 2 with Rb requires the amino-terminal region of Re RBaK, the interaction of RBaK with the AR 0 occurs through the CTD of RBaK. This inter- VP16 VP16 RBaKCTD(S) RBaKCTD(AS) VP16 action of RBaK with the AR was confirmed by in vivo two-hybrid interaction experiments in yeast Gal4DBD-ARLBD and in mammalian cells, by in vitro binding studies Figure 5 Mammalian two-hybrid interaction between and by coimmunoprecipitation from nuclear Gal4 DBD-ARLBD and VP16AD-RBaKCTD. COS-7 6 extracts. Finally, cotransfection of RBaK with the cells (2×10 cells/well of a 48-well plate) were AR resulted in a marked increase of AR activity, as transiently transfected with 100 ng/well p(Gal4)5- luciferase reporter construct, 0·67 ng expression measured by means of androgen-regulated luci- construct for the Gal4 DBD-ARLBD, and with 0·67 ng ferase activity, showing that RBaK functions as a empty pSNATCHII expression vector (expressing the coactivator under these conditions. In fact, such VP16AD), 0·67 ng expression vector for the coactivation was observed not only for the AR, but VP16AD-RBaKCTD sense or 0·67 ng of a similar anti-sense construct as indicated. Cells were incubated also for the GR (results not shown), which showed for 24 h with 10 nM R1881 (black bars) or with absolute weaker interaction with RBaK in the yeast ethanol as vehicle (white bars). Thereafter, luciferase two-hybrid system and in vitro. For that reason activities were measured. The luciferase activities in the RBaK should not be considered as an AR-specific absence of ligand were arbitrarily set to 1 and relative coactivator. Moreover, it should be kept in mind luciferase values (means±S.D.) were calculated for the conditions supplemented with ligand. that our data show that RBaK interacts with the AR and coactivates its function in cotransfected cells that overexpress both proteins. This does not Indeed, in COS-7 cells, but not in CV-1 cells, the necessarily imply that such interaction and function of Rb is strongly suppressed by expression coactivation occur for the naturally occurring gene of the large T antigen (Ludlow et al. 1989) and this products. applies also to the coactivation of the AR by Rb Whereas the interaction of the ARLBD with the (Lu & Danielsen 1998). In line with their results, we RBaKCTD was clearly ligand-dependent in the observed that cotransfection with increasing mammalian and yeast two-hybrid experiments, this amounts of a Rb expression construct markedly was not the case when this interaction was studied increased AR transactivation in CV-1 cells in spite in vitro or by means of coimmunoprecipitation. The of decreasing AR concentrations (Fig. 7). In COS-7 reasons for this discrepancy remain unclear, but cells, on the other hand, the decrease of AR it is conceivable that the presence of the NTD concentrations was accompanied by a parallel in the latter experiments may have contributed decrease of AR activity. Consequently, Rb to the AR binding in the absence of ligand. It is coactivates the AR in CV-1 cells but hardly at all or also conceivable that the two-hybrid interactions not in COS-7 cells. Moreover, although cotransfec- require a tighter or sterically specific interaction, tion of CV-1 cells with Rb increased the activity of which is only realized in the presence of ligand. In the AR, it did not affect the degree of AR any case such discrepancy is not uncommon for coactivation by RBaK (data not shown). Indeed, AR coregulators (Blanco et al. 1998, Brady et al. the degree of coactivation by RBaK(S) (as 1999, Ting et al. 2002), which very often interact in compared with the RBaK(AS) construct) was 3·1 a ligand-independent way with the AR in vitro. after cotransfection with Rb and 3·0 without the The observation that cotransfection with RBaK latter. For that reason it seems unlikely that the results in coactivation rather than corepression is coactivation of the AR by RBaK and by Rb are somewhat surprising. Indeed, structurally RBaK directly linked. belongs to the subgroup of Krüppel-like zinc finger

Journal of Molecular Endocrinology (2003) 31, 583–596 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access Androgen receptor and RBaK interaction · K HOFMAN and others 593

Figure 6 Influence of RBaK on AR transactivation in CV-1 cells. (A) CV-1 cells or COS-7 cells (1·5×104 cells/well of a 48-well plate) were transiently transfected with 100 ng/well of reporter construct (C3 ARE-luciferase), with 0·17 ng pSG5-AR and with 0·67 ng pSG5 expression construct containing either RBaK(S), RBaK(AS) or an unrelated cDNA insert of similar size (ϕ1) as indicated. Cells were incubated for 24 h with 10 nM R1881 (filled bars) or with ethanol as vehicle (white bars) and luciferase activities were measured. The luciferase activities obtained in the presence of ligand for the pSG5-ϕ1 construct were arbitrarily set to 1 and relative luciferase values were calculated for the other conditions. The results shown are the means±S.D. of three independent experiments each performed in triplicate. (B) AR levels were estimated by means of Western blotting on pooled 3×10 µl aliquots of the same cell extracts. proteins, which contain an amino-terminal KRAB teins (Friedman et al. 1996, Moosmann et al. 1996). domain, separated from the zinc finger domain by Moreover, the KRAB domains can be linked with a linker region (Bellefroid et al. 1991). Several NR functioning via these KAP1/TIF1 proteins reports link KRAB transcriptional repression with which interact with NRs and occur in larger NR functioning (Losson 1997, de Haan et al. 2000), complexes mediating NR activity (Underhill et al. but it has never been reported that a KRAB- 2000). containing protein enhances the activity of an NR. As already mentioned RBaK not only interacts In general, the KRAB motif of these proteins with the AR, but also with the Rb protein, for results in transcriptional repressor activity through which AR coactivation properties have already interactions with components of the basal transcrip- been described in several cell lines (Lu & Danielsen tion factor complex (Margolin et al. 1994, Witzgall 1998, Yeh et al. 1998). In the present report we et al. 1994, Lange et al. 1995) or with other could confirm such coactivation of the AR by Rb in coregulatory proteins, such as KAP1/TIF1 pro- CV-1 cells. Moreover, although cotransfection of www.endocrinology.org Journal of Molecular Endocrinology (2003) 31, 583–596

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access 594 K HOFMAN and others · Androgen receptor and RBaK interaction

Figure 7 Influence of Rb on AR transactivation in CV-1 cells or COS-7 cells. (A) CV-1 cells or COS-7 cells (1·5×104 cells/well of a 48-well plate) were transiently transfected with 100 ng/well of reporter construct (C3 ARE-luciferase), with 0·17 ng pSG5-AR and with increasing amounts (0, 0·33, 0·66 and 1·33 ng) of pSG5-Rb expression construct or an expression construct containing an unrelated cDNA fragment (2) of approximately the same size as the Rb insert without coregulator properties. Cells were incubated for 24 h with or without 10 nM R1881 and luciferase activities were measured. The luciferase activity in the absence of cotransfected construct was arbitrarily set to 1 and relative luciferase values were calculated for other conditions. The results shown are the means±S.D. of three independent experiments each performed in triplicate. (B) AR levels were estimated on pooled 3×10 µl aliquots of the same cell extracts.

RBaK resulted in a further increase of AR activity, The interaction of RBaK with the AR on the one the degree of coactivation by RBaK was similar hand and with the Rb protein on the other hand with or without cotransfection with Rb. Finally, may contribute to the role of androgens in the RBaK produced a similar degree of coactivation in regulation of the cell cycle, wherein the Rb protein COS-7 cells, wherein Rb is largely inactivated as in functions as a crucial control mechanism. The CV-1 cells. For that reason it seems less likely that growth suppressive effects of Rb are best character- simultaneous interactions of the AR and RBaK ized by its role in suppression of transcription with Rb are required for the coactivation of the AR mediated by E2F transcription activators (Dynlacht by RBaK in normal or tumoral cells. et al. 1994, Krek et al. 1994, 1995). In the G1 phase

Journal of Molecular Endocrinology (2003) 31, 583–596 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access Androgen receptor and RBaK interaction · K HOFMAN and others 595 various E2Fs are bound by Rb and hereby unable Aumüller G, Holterhus PM, Konrad L, von Rahden B, Hiort O, Esquenet M & Verhoeven G 1998 Immunohistochemistry and in to promote cell proliferation. Upon phosphoryl- situ hybridization of the androgen receptor in the developing ation of Rb, E2F is released and initiates the human prostate. Anatomy and Embryology 197 199–208. transcription of genes containing E2F sites in their Bellefroid EJ, Poncelet DA, Lecocq PJ, RevelantO&Martial JA 1991 The evolutionarily conserved Kruppel-associated box promoter. Many of these genes are involved in cell domain defines a subfamily of eukaryotic multifingered proteins. cycle progression and DNA replication (Flemington PNAS 88 3608–3612. et al. 1993, Helin et al. 1993, Weintraub et al. 1995, Blanco JC, Minucci S, Lu J, Yang XJ, Walker KK, Chen H, Evans RM, NakataniY&Ozato K 1998 The histone acetylase PCAF La Thangue 1996, Dyson 1998, Harbour & Dean is a nuclear receptor coactivator. Genes and Development 12 2000). As we could previously show in LNCaP 1638–1651. cells (Hofman et al. 2001) the regulation of cell Brady ME, Ozanne DM, Gaughan L, Waite I, Cook S, Neal DE & Robson CN 1999 Tip60 is a nuclear hormone receptor proliferation by androgens is accompanied by coactivator. Journal of Biological Chemistry 274 17599–17604. major changes in E2F activity. Most probably these Brown RE, Jarvis KL & Hyland KJ 1989 Protein measurement changes are largely due to androgen-dependent using bicinchoninic acid: elimination of interfering substances. Analytical Biochemistry 1989 180 136–139. changes in the phosphorylation state of Rb, Darne C, Veyssiere G & Jean C 1998 Phorbol ester causes ligand- mediated in turn by the activity of cyclin- independent activation of the androgen receptor. European Journal dependent kinases and their inhibitors. However, of Biochemistry 256 541–549. since RBaK cotransfection has been shown to Dynlacht BD, Flores O, Lees JA & Harlow E 1994 Differential regulation of E2F transactivation by cyclin/cdk2 complexes. affect the E2F suppressor activity of Rb (Skapek Genes and Development 8 1772–1786. et al. 2000) and since RBaK mRNA is present Dyson N 1998 The regulation of E2F by pRB-family proteins. in many cells and tissues including LNCaP cells (K Genes and Development 12 2245–2262. Evans RM 1988 The steroid and thyroid hormone receptor Hofman & W Heyns, unpublished observations), superfamily. Science 240 889–895. the interaction of RBaK with the AR may provide Fasciana C, van der Made AC, Faber PW & Trapman J 1996 an additional level of control. Androgen regulation of the rat keratinocyte growth factor (KGF/FGF7) promoter. Biochemical and Biophysical Research Communications 220 858–863. Flemington EK, Speck SH & Kaelin WG Jr 1993 E2F-1-mediated Acknowledgements transactivation is inhibited by complex formation with the retinoblastoma susceptibility gene product. PNAS 90 The excellent technical assistance of Hilde 6914–6918. Friedman JR, Fredericks WJ, Jensen DE, Speicher DW, Huang XP, Geeraerts is kindly acknowledged. We thank Dr Neilson EG & Rauscher FJ III 1996 KAP-1, a novel corepressor Hinrich Gronemeyer, Dr Michael R Stallcup, for the highly conserved KRAB repression domain. Genes and Dr Malcolm G Parker, Dr Carsten Carlberg, and Development 10 2067–2078. Gietz RD, Schiestl RH, Willems AR & Woods RA 1995 Studies on Dr Tony Kouzarides for their generous gifts of the transformation of intact yeast cells by the LiAc/SS-DNA/PEG plasmids. This work was supported by a grant procedure. Yeast 11 355–360. ‘Geconcerteerde Onderzoeksactie van de Vlaamse de Haan G, Chusacultanachai S, Mao C, Katzenellenbogen BS & Gemeenschap’, by a research grant from the Fund Shapiro DJ 2000 Estrogen receptor-KRAB chimeras are potent ligand-dependent repressors of estrogen-regulated gene expression. for Scientific Research, Flanders, Belgium (FWO), Journal of Biological Chemistry 275 13493–13501. by a cancer research grant from FB insurances Hall JA, Harris MA, IntresR&Harris SE 1993 Acidic fibroblast and VIVA and by a grant ‘Interuniversity Poles growth factor gene 5
non-coding exon and flanking region from hamster DDT1 cells: identification of the promoter region and of Attraction Program, Belgian State, Prime transcriptional regulation by testosterone and aFGF protein. Minister’s Office, Federal Office for Scientific, Journal of Cellular Biochemistry 51 116–127. Technical and Cultural Affairs’. Harbour JW & Dean DC 2000 Rb function in cell-cycle regulation and apoptosis. Nature Cell Biology 2 E65–E67. Helin K, HarlowE&Fattaey A 1993 Inhibition of E2F-1 transactivation by direct binding of the retinoblastoma protein. References Molecular and Cellular Biology 13 6501–6508. Hofman K, Swinnen JV, Claessens F, VerhoevenG&Heyns W Alen P, Claessens F, Verhoeven G, Rombauts W & Peeters B 1999a 2000 Apparent coactivation due to interference of expression The androgen receptor amino-terminal domain plays a key role in constructs with nuclear receptor expression. Molecular and Cellular p160 coactivator-stimulated gene transcription. Molecular and Endocrinology 168 21–29. Cellular Biology 19 6085–6097. Hofman K, Swinnen JV, VerhoevenG&HeynsW2001 E2F Alen P, Claessens F, Schoenmakers E, Swinnen JV, Verhoeven G, activity is biphasically regulated by androgens in LNCaP cells. Rombauts W & Peeters B 1999b Interaction of the putative Biochemical and Biophysical Research Communications 283 97–101. androgen receptor-specific coactivator ARA70/ELE1 alpha with Horton R, PasupulettiV&Antonipillai I 1993 Androgen induction multiple steroid receptors and identification of an internally of steroid 5 alpha-reductase may be mediated via insulin-like deleted ELE1 beta isoform. Molecular Endocrinology 13 117–128. growth factor-I. Endocrinology 133 447–451. www.endocrinology.org Journal of Molecular Endocrinology (2003) 31, 583–596

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access 596 K HOFMAN and others · Androgen receptor and RBaK interaction

Kim IY, Zelner DJ, Sensibar JA, Ahn HJ, Park L, Kim JH & Lee C Kruppel-associated box family transcriptional repressor that 1996 Modulation of sensitivity to transforming growth factor-beta interacts with the retinoblastoma gene product, RB. Journal of 1 (TGF-beta 1) and the level of type II TGF-beta receptor in Biological Chemistry 275 7212–7223. LNCaP cells by dihydrotestosterone. Experimental Cell Research 222 Swinnen JV, Esquenet M, Heyns W, Rombauts W & Verhoeven G 103–110. 1994 Androgen regulation of the messenger RNA encoding Krek W, Ewen ME, Shirodkar S, Arany Z, Kaelin WG Jr & diazepam-binding inhibitor/acyl-CoA-binding protein in the Livingston DM 1994 Negative regulation of the growth-promoting human prostatic adenocarcinoma cell line LNCaP. Molecular and transcription factor E2F-1 by a stably bound cyclin A-dependent Cellular Endocrinology 104 153–162. protein kinase. Cell 78 161–172. Swinnen JV, Ulrix W, Heyns W & Verhoeven G 1997 Coordinate KrekW,XuG&Livingston DM 1995 Cyclin A-kinase regulation of regulation of lipogenic gene expression by androgens: evidence for E2F-1 DNA binding function underlies suppression of an S phase a cascade mechanism involving sterol regulatory element binding checkpoint. Cell 83 1149–1158. proteins. PNAS 94 12975–12980. Lange R, Christoph A, Thiesen HJ, Vopper G, Johnson KR, TingHJ,YehS,NishimuraK&ChangC2002Supervillin Lemaire L, Plomann M, Cremer H, BarthelsD&Heinlein UA associates with androgen receptor and modulates its 1995 Developmentally regulated mouse gene NK10 encodes a transcriptional activity. PNAS 99 661–666. zinc finger repressor protein with differential DNA-binding Tsai MJ & O’Malley BW 1994 Molecular mechanisms of action of domains. DNA and Cell Biology 971–981. 14 steroid/thyroid receptor superfamily members. Annual Review of La Thangue NB 1996 E2F and the molecular mechanisms of early Biochemistry 63 451–486. cell-cycle control. Biochemical Society Transactions 24 54–59. Losson R 1997 KRAB zinc finger proteins and nuclear receptors: a Underhill C, Qutob MS, Yee SP & Torchia J 2000 A novel nuclear receptor corepressor complex, N-CoR, contains components of possible cross-talk. Biological Chemistry 378 579–581. LuJ&Danielsen M 1998 Differential regulation of androgen and the mammalian SWI/SNF complex and the corepressor KAP-1. Journal of Biological Chemistry glucocorticoid receptors by retinoblastoma protein. Journal of 275 40463–40470. Biological Chemistry 273 31528–31533. Voegel JJ, Heine MJ, Tini M, Vivat V, ChambonP&Gronemeyer Ludlow JW, DeCaprio JA, Huang CM, Lee WH, Paucha E & H 1998 The coactivator TIF2 contains three nuclear Livingston DM 1989 SV40 large T antigen binds preferentially to receptor-binding motifs and mediates transactivation through CBP an underphosphorylated member of the retinoblastoma binding-dependent and -independent pathways. EMBO Journal 17 susceptibility gene product family. Cell 56 57–65. 507–519. Mangelsdorf DJ & Evans RM 1995 The RXR heterodimers and Weintraub SJ, Chow KN, Luo RX, Zhang SH, HeS&DeanDC orphan receptors. Cell 83 841–850. 1995 Mechanism of active transcriptional repression by the Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G, retinoblastoma protein. Nature 375 812–815. Umesono K, Blumberg B, Kastner P, Mark M & Chambon P Wen Y, Hu MC, Makino K, Spohn B, Bartholomeusz G, Yan DH 1995 The nuclear receptor superfamily: the second decade. & Hung MC 2000 HER-2/neu promotes androgen-independent Cell 83 835–839. survival and growth of prostate cancer cells through the Akt Margolin JF, Friedman JR, Meyer WK, Vissing H, Thiesen HJ & pathway. Cancer Research 60 6841–6845. Rauscher FJ III 1994 Kruppel-associated boxes are potent Witzgall R, O’Leary E, Leaf A, OnaldiD&Bonventre JV 1994 The transcriptional repression domains. PNAS 91 4509–4513. Kruppel-associated box-A (KRAB-A) domain of zinc finger Moosmann P, Georgiev O, Le Douarin B, Bourquin JP & proteins mediates transcriptional repression. PNAS 91 4514–4518. Schaffner W 1996 Transcriptional repression by RING finger Würtz JM, Bourguet W, Renaud JP, Vivat V, Chambon P, Moras D protein TIF1 beta that interacts with the KRAB repressor domain & Gronemeyer H 1996 A canonical structure for the ligand-binding of KOX1. Nucleic Acids Research 24 4859–4867. domain of nuclear receptors. Nature Structural Biology 3 87–94. Nazareth LV & Weigel NL 1996 Activation of the human androgen Yan G, Fukabori Y, Nikolaropoulos S, WangF&McKeehan WL receptor through a protein kinase A signaling pathway. Journal of 1992 Heparin-binding keratinocyte growth factor is a candidate Biological Chemistry 271 19900–19907. stromal-to-epithelial-cell andromedin. Molecular Endocrinology 6 Parker MG 1993 Steroid and related receptors. Current Opinion in Cell 2123–2128. Biology 5 499–504. Yeh S, Miyamoto H, Nishimura K, Kang H, Ludlow J, Hsiao P, Pazin MJ & Kadonaga JT 1997 What’s up and down with histone WangC,SuC&Chang C 1998 Retinoblastoma, a tumor deacetylation and transcription? Cell 89 325–328. suppressor, is a coactivator for the androgen receptor in human Pollard KJ & Peterson CL 1998 Chromatin remodeling: a marriage prostate cancer DU145 cells. Biochemical and Biophysical Research between two families? Bioessays 20 771–780. Communications 248 361–367. Schuurmans AL, Bolt J, Veldscholte J & Mulder E 1991 Regulation Zuck B, Goepfert C, Nedlin-Chittka A, Sohrt K, Voigt KD & of growth of LNCaP human prostate tumor cells by growth Knabbe C 1992 Regulation of fibroblast growth factor-like Journal of Steroid Biochemistry and factors and steroid hormones. protein(s) in the androgen-responsive human prostate carcinoma Molecular Biology 40 193–197. cell line LNCaP. Journal of Steroid Biochemistry and Molecular Biology SinghP,CoeJ&HongW1995 A role for retinoblastoma protein in 41 659–663. potentiating transcriptional activation by the glucocorticoid receptor. Nature 374 562–565. Skapek SX, Jansen D, Wei TF, McDermott T, Huang W, Olson EN Received 26 May 2003 & Lee EY 2000 Cloning and characterization of a novel Accepted 20 August 2003

Journal of Molecular Endocrinology (2003) 31, 583–596 www.endocrinology.org

Downloaded from Bioscientifica.com at 10/01/2021 09:06:30AM via free access