Trafficking of the androgen receptor in living cells with fused green fluorescent protein–androgen receptor

Trafficking of the androgen receptor in living cells with fused green fluorescent protein–androgen receptor

Molecular and Cellular Endocrinology 129 (1997) 17 – 26 Trafficking of the androgen receptor in living cells with fused green fluorescent protein–and...

518KB Sizes 0 Downloads 21 Views

Molecular and Cellular Endocrinology 129 (1997) 17 – 26

Trafficking of the androgen receptor in living cells with fused green fluorescent protein–androgen receptor V. Georget a, J.M. Lobaccaro a, B. Terouanne a, P. Mangeat b, J.-C. Nicolas a, C. Sultan a,c,* a

Institut National de la Sante´ et de la Recherche Me´dicale, INSERM U439, Pathologie Mole´culaire des Re´cepteurs Nucle´aires, 70 rue de Na6acelles, 34090 Montpellier, France b CNRS UMR 5539, Montpellier, France c Unite´ BEDR, Hoˆpital Lapeyronie, Montpellier, France Received 27 December 1996; accepted 5 February 1997

Abstract The trafficking of the androgen receptor (AR) in transfected cells was studied using a green fluorescent protein (GFP)-AR chimera. The reporter molecule GFP enabled the localization of AR in living cells with a good spatial and temporal resolution. After the construction of GFP-AR and verification of the size of the fusion protein produced, we demonstrated that GFP-AR conserves the functional properties of the AR: GFP-AR had the same androgen-binding affinity as AR, and GFP-AR efficiently transactivated an androgen-responsive gene in response to synthetic androgen at 30°C. The fusion protein was first detected throughout the cytoplasm without hormone, fluorescence becoming nuclear rapidly after androgen incubation. This hormone dependence of AR trafficking was confirmed by the use of the mutant GFP-AR-del4, which lacked the androgen-binding function. The mutant was localized in the cytoplasm in the absence of hormone, but the distribution was not modified by androgen incubation. An ACAS 570 scanning laser cytometer was used to quantify fluorescence in a single living cell, first without and then with hormone. Different hormones and antihormones were tested to determine the dynamics of GFP-AR translocation into the nucleus. All the drugs used were able to induce nuclear translocation, and steady state level was rapidly attained within 1 h. The ratio of receptors in cytoplasmic and nuclear compartments was related to both affinity and concentration of ligand. The data from this follow-up study demonstrated for the first time the intracellular dynamics of the hormone-dependent trafficking of AR in a single living cell. © 1997 Elsevier Science Ireland Ltd. Keywords: Androgen receptor; Nuclear translocation; Androgen insensitivity syndrome; Antiandrogens; Intracellular reporter

1. Introduction Steroid receptors may often act as transcriptional factors by binding the hormone-regulatory elements of specific genes to trigger a cascade of transcriptional events. Before exerting their function as transcription Abbre6iations: AR, androgen receptor; GFP, green fluorescent protein; GR, glucocorticoid receptor; TAT-tk-luc, tyrosine amino transferase-thymidine kinase-luciferase; DIC, differential interface contrast; CA, cyproterone acetate; OH-flu, hydroxy-flutamide; DHT, dihydrotestosterone; ACAS, adherent cell analysis and sorting. * Corresponding author. Tel.: + 33 467338696; fax: + 33 467338327; e-mail: [email protected]

factors, these proteins have to pass the nuclear membrane. The intracellular trafficking of steroid receptors has yet to be fully understood [1], and a better definition of these processes could provide information concerning the physiology of the androgen receptor (AR) in target and non-target cells in the presence or absence of ligand, the physiopathology of mutated forms of AR, and the mechanism of action of antihormones. While it is generally agreed that the estrogen and progesterone receptors are predominantly nuclear [2,3], with a continuous shuttle between the nucleus and the cytoplasm [4], the findings concerning the mineralocorticoid receptor, the glucocorticoid receptor (GR), and

0303-7207/97/$17.00 © 1997 Elsevier Science Ireland Ltd. All rights reserved. PII S 0 3 0 3 - 7 2 0 7 ( 9 7 ) 0 4 0 3 4 - 3


V. Georget et al. / Molecular and Cellular Endocrinology 129 (1997) 17–26

the androgen receptor (AR) are more controversial. The mineralocorticoid receptor was first localized using an anti-idiotypic antibody. It was found in both the cytoplasm and the nucleus, independent of the presence of the ligand [5]. Using an antipeptide antibody, the mineralocorticoid receptor was found in the cytoplasm when the ligand was absent, and in the nucleus when present [6]. Different studies report varying localizations of the unliganded GR, depending on the cell line expressing the receptor as well as the fixation and permeabilization conditions of the immunocytochemical procedure [7 – 9]. Contradictory findings have been published for the AR. Unliganded AR is found in the nucleus in target tissues [10,11], whereas in transfected cells overexpressing AR it is located either in the cytoplasm, in both the cytoplasm and the nucleus, or exclusively in the nucleus, according to the cell type [12,13]. In the presence of ligand, AR has been found to be nuclear in transfected cells [12,13] and target tissues. Previous studies of intracellular localization of AR have been performed by immunohistochemistry or immunocytochemistry, using an anti-AR antibody that requires fixing and killing the cells. This can lead to artefacts in the pattern of localization. For our study we developed a new model, using a chimera of AR and a reporter molecule, the green fluorescent protein (GFP) [14], which is a 238 amino acid protein from the jellyfish Aequorea victoria. A recently described GFP mutant (GFPS65T) emits bright green light when excited with blue light [15]. GFP fluorescence can be detected using standard epifluorescence filters for fluorescein [14] without requiring substrate or cofactor. The GFP tag yields two important improvements: first, it eliminates the use of fixation, cell permeabilization, and antibody incubation, these being normally required when using antibodies tagged with chemical fluorophores and second the use of the GFP tag permits kinetic studies of protein localization and trafficking due to the resistance of the GFP chromophore to photobleaching [16]. The properties of GFP thus make this protein a good reporter for studying intracellular protein trafficking in living cells [17]. In the present study, after construction of the fusion protein, we first determined the biochemical and functional characteristics of GFP-AR. We compared molecular weight, androgen-binding capacity and affinity, and transcriptional activation of GFP-AR with those of wild type AR. We next demonstrated that GFP was fluorescent when tagged to the AR, and was nuclear in the presence of synthetic androgen R1881. The model was validated using a mutant GFP-AR-del4 that lacks its carboxy-terminal function. In the second part of this work, we analyzed the dynamics of GFP-AR nuclear/ cytoplasmic localization in response to different androgens and antiandrogens using spatial and temporal resolution. This new model has enabled us to report for

the first time the intracellular dynamics of the hormonedependent trafficking of AR in a single living cell.

2. Materials and methods

2.1. Plasmid construction The pS65T-Cl-GFP plasmid (Clontech, Palo Alto, CA) contains the cDNA of a variant of wild-type GFP in which a mutation described by Heim et al. [15] has been introduced. Expression is driven from a cytomegalovirus promoter (Fig. 1). To digest pS65T-ClGFP with XbaI, the vector was transformed into DMl cells (dam-host) (Gibco-BRL, Cergy-Pontoise, France) to unmethylate the XbaI site. The hAR cDNA was isolated from pCMV5-hAR [18] by XmaI–XbaI digestion and inserted at the cognate sites of pS65T-Cl-GFP. In GFP-AR, GFP was fused at the amino-terminal of an hAR that lacked the first 36 amino-acid residues. The construction of the mutant pGFP-ARdel4 was obtained in the same manner as from the plasmid pCMV5-hAR-del4 previously described [19]. AR-del4 is an AR with a deletion of 13 bp in exon 4, which introduces a premature stop codon at position 783 after a frameshift. As previously described [19], this AR mutant failed to induce any androgen-regulated transcriptional activity, mainly due to the presence of an abnormal ligand-binding domain. Constructions were verified by enzymatic digestion, and were sequenced in ligation fragments to verify the correct reading frame. For transfection, all plasmids were purified with Nucleobond-AX cartridges (Macherey–Nagel, Hoerdt, France).

2.2. Cell culture and biochemical analysis Mammalian cells expressing GFP have exhibited stronger fluorescence when grown at temperatures lower than 37°C [17]. All experiments were thus performed at 30 and 37°C. COS-7 cells and CV1 cells were cultured in Dulbecco’s minimal essential medium (DMEM) (Gibco-BRL, CergyPontoise, France) supplemented with 10% fetal calf serum and penicillin (100 U/ml) and streptomycin (100 mg/ml). Cells were transfected by the calcium phosphate DNA precipitation method. For immunoblot analysis, the cells were solubilized in the lysis buffer (160 mM Tris, pH 6.9, 200 mM dithiothreitol, 4% SDS, 20% glycerol, 0.004% bromophenol blue) in the presence of protease inhibitors (1 mM PMSF, 0.05 mM leupeptin, 0.01 mM pepstatin), boiled 5 min and centrifuged at 13 000 ×g for 10 min. The supernatant was analyzed on SDS-polyacrylamide gel electrophoresis (SDS-PAGE). For each extract of transfected COS-7 cells, immunoblot was performed with both anti-AR antibody (SpO61) diluted 1/2000

V. Georget et al. / Molecular and Cellular Endocrinology 129 (1997) 17–26


Fig. 1. Construction of GFP expression plasmids and structure of GFP fusion proteins. (A) Schematic structure of pS65T-C1-GFP, encoding full-length of a red-shifted variant. This mutant can be used for fusing heterologous proteins to the C-terminus of GFP. The GFP vector contains a Kozak consensus sequence to maximize the efficiency of translation in eukaryotic cells. (B) Schematic structure of the human AR cDNA with its three domains: the activation domain (AD), the DNA-binding domain (DBD) and the ligand-binding domain (LBD). The amino-terminal truncated hAR cDNA (XmaI–XbaI) was fused with the GFP-S65T. (C) Structure of fusion proteins: GFP-AR and the mutant GFP-AR-del4. hAR-del4 is an AR with a deletion of 13 pb at position 2780 that induces a frameshift (hatched area) to a premature stop codon at position 2945. This mutant AR lacked its capacity for binding the hormone.

[20] and anti-GFP antiserum (anti-rGFP, Clontech, Palo Alto, CA) diluted 1/1000 using chemiluminescent detection (Amersham, les Ulis, France) as previously described [19]. For binding assay, COS-7 cells were transfected in 12-well tissue culture dishes with 0.1 mg of either pCMV5-hAR or pGFP-AR expression vectors and 0.1 mg of pCMV-b-galactosidase. Binding of [3H]R1881 (DuPont de Nemours, Les Ulis, France) was measured after cultivating transfected cells in serumfree medium one day before harvest. Cells were incubated with varying concentrations of [3H]R1881 (0.05 nM–3 nM) in the presence or absence of 100 nM unlabeled R1881 for 2 h. The cells were washed with phosphate buffered saline (PBS) and harvested in lysis buffer: 10 mM Tris – H3PO4 (pH 7.8), 2% SDS and 10% glycerol. Aliquots were counted for radioactivity with specific binding representing the difference between counts in the presence and absence of excess unlabeled hormone. The b-galactosidase activity was measured with aliquot of lysate. The results are reported by Scatchard representation. For transcription regulation studies, CV-1 cells were transfected in 6-well tissue culture dishes with 0.2 mg of either pCMV5-hAR or pGFP-AR expression vectors, 2.5 mg of p-tyrosine amino transferase-thymidine kinase-luciferase (TAT-tk-luc), a reporter plasmid construct (firefly structural luciferase gene under control of a rat tyrosine amino-transferase promoter) used as an-

drogen-regulated gene [19] and 0.3 mg of pCMV-bgalactosidase. The cells were then incubated in serum-free medium with hormones diluted from a 1000× concentrated stock solution in ethanol. The incubation with the synthetic androgen, R1881; the steroidal antiandrogen, cyproterone acetate (CA) and the non-steroidal antiandrogen, hydroxy-flutamide (OH-flu) continued for 24 h. For luciferase assay, cells were lysed by 300 ml of lysis buffer: 25 mM Tris– H3PO4 (pH 7.8), 2 mM DTT, 2 mM EDTA, 10% glycerol and l% Triton X-100. After 10 min of incubation, 100 ml of the supernatant was collected and reacted with 100 ml luciferine solution: 270 mM Coenzyme A, 470 mM luciferine, 530 mM ATP, 20 mM Tris–H3PO4, 1.05 mM MgCl2, 2.7 mM MgSO4, 0.1 mM EDTA and 33 mM DTT. Luciferase activity was measured on an LKB luminometer. Each incubation was performed in duplicate.

2.3. Obser6ation of li6ing cells and microscopic analysis For qualitative analysis of fluorescence, COS-7 cells were cultured at 37°C on coverslips and transfected with 2 mg of either pS65T-Cl-GFP, pGFP-AR or pGFP-AR-del4. After 36 h, the transfected cells were cultured at 30°C for at least 12 h. The medium was replaced with phenol red-free DMEM before microscopy because the DMEM exhibited a high back-


V. Georget et al. / Molecular and Cellular Endocrinology 129 (1997) 17–26

ground fluorescence. Living cells on coverslips were placed on a glass slide with the medium containing 10 − 6 M R1881. The transfected cells were examined immediately using a Reichert epifluorescence microscope (Polyvar) attached to a SIT Camera (Lhesa Electronique), which permitted us to observe cells by differential interference contrast (DIC), and in parallel in fluorescence with fluorescine isothiocyanate (FITC) filters. The observation of the cells was performed at 37°C. For quantitative analysis of GFP-AR fluorescence, COS-7 cells were cultured at 37°C and transfected with 2 mg pGFP-AR in 3 cm culture dishes. The transfected cells were cultured at 30°C for at least 12 h before observation. Cells were then analyzed directly in the dish using the ACAS 570 (adherent cell analysis and sorting) (Meridian). This laser scanning cytometer allows the localization and the measurement of the fluorescence, ACAS permits kinetic studies of live cells. Culture dishes were viewed for extended time periods in air maintained at 37°C. Cells were first studied without hormone and then, after hormonal addition to the same culture dish, they were observed and recorded every 15 min. For quantification of nuclear/cytoplasmic ratio, the data obtained with ACAS were collected and quantitated using NIH-image software. For each set of conditions, the intensities of pixels were summed within the individual nuclei and total cellular areas and corrected for background fluorescence. The percentages of nuclear fluorescence were calculated and pooled for each point.

In addition, a minor band at 65 kDa was detected with anti-rGFP, which seems to be unspecific because it was also present in AR cell extract. The weight of GFP-AR was confirmed by immunoblot with SpO61. AR was detected with an apparent molecular mass of 110 kDa as expected and we observed the same specific band for GFP-AR at 130 kDa. Other specific bands at 90 and 75 kDa detected for AR and 110, 90 and 75 kDa for GFP-AR may be degradation products or products of

3. Results

3.1. Expression of AR and GFP fusion proteins To develop an efficient model for studying AR trafficking, we generated a GFP-AR chimera with a 27 kDa GFP variant fused with an amino-terminal truncated AR. The GFP-S65T is a mutant GFP that emits a fluorescence four to six times more intense than wild-type GFP, and it has a single redshifted excitation peak at 490 nm [15]. To investigate whether the construction of pGFP-AR could be translated, and to confirm the correct weight of the protein produced, COS-7 cells were transfected with either pCMV5-hAR, pS65T-Cl-GFP or pGFP-AR, and expressed proteins were analyzed by immunoblotting. Western blots were performed for the same cell extract with two antibodies: anti-rGFP, an antiserum against the recombinant GFP, and SpO61, a polyclonal antibody directed against the N-terminal domain from amino-acids 301 – 320 of the hAR. Fig. 2A shows that GFP-AR was detected by anti-rGFP in one major band at 130 kDa and GFP was detected with an apparent molecular mass of 27 kDa.

Fig. 2. Expression and binding characteristics of GFP-AR compared with AR. (A) Western blot analysis of protein products of pS65T-C1GFP, pCMV5-hAR and pGFP-AR expressed in COS-7 cells at 30°C and analyzed by SDS/11% PAGE. Expressed fusion protein and GFP were detected when 3-fold more extract cells were present. The GFP was visualized by immunostaining with the antiserum anti-rGFP (on the left) and AR was detected with the polyclonal antibody SpO61 (on the right). Molecular masses are shown in kilodaltons. (B) Scatchard plot analysis of androgen ([3H]R1881) binding in COS-7 cells transfected with pCMV5-hAR or pGFP-AR at 30 and 37°C. The constant of dissociation (Kd) was computed from the slope of the line. B, bound (fmol/mg protein); B/F, bound/free.

V. Georget et al. / Molecular and Cellular Endocrinology 129 (1997) 17–26


minor alternative initiation sites. No band was observed for GFP cell lysate. The use of the two antibodies revealed that GFP-AR is very sensitive to degradation because in both cases a 110 kDa band was observed. Western blot was performed with extract of transfected cells cultured at 30 and 37°C. The expression level appeared to be identical (data not shown).

3.2. Androgen-binding characteristics To confirm that the 5% end hAR tagged with GFP did not alter the carboxy-terminal ligand-binding characteristics of AR, the specific high affinity androgen binding of the AR and GFP-AR was measured in the whole cell binding assays with [3H]R1881. This experiment was performed at 30 and 37°C. The dissociation constants (Kd) of GFP-AR were similar to that of AR (0.11 nM and 0.08 nM, respectively, compared with 0.2 and 0.14 nM) (Fig. 2B). No differences were detected between 30 and 37°C, and the androgen-binding capacities of AR and GFP-AR were identical. In all experiments, we observed only a decreased binding capacity of the GFPAR protein compared with AR, while the binding affinity of GFP-AR was conserved.

3.3. Transcription acti6ation The expression vector for either the GFP-AR and the AR was co-transfected with the androgen-regulated reporter gene TAT-tk-luc to investigate the transcriptional activity of GFP-AR compared with AR in presence of androgens or antiandrogens. Exposure of the cells transfected at 30°C with pCMV5-hAR to R1881 at concentrations of 3 ×10 − 12, 10 − 11, 10 − 10, 10 − 9 and 5× 10 − 9 M induced an increase in the luciferase activity, which was dose dependent and reached a maximum of 10-fold induction at 10 − 10 M R1881 (Fig. 3A). CV-l cells transfected with pGFP-AR demonstrated an induction of luciferase activity, which was dose dependent in the same manner as AR and which reached a 5-fold induction at 10 − 10 M R1881 (Fig. 3A). When the same experiment was performed at 37°C, the fold induction obtained with AR was higher (15-fold) and with GFP-AR, lower (2-fold) (data not shown). This sensitivity of GFP to temperature for transcriptional activation assay has already been described for the GR [21]. When cells were maintained at 37°C, GFP-GR did not transactivate the reporter gene, although efficient transactivation by GFP-GR was seen by these authors when the incubation temperature was shifted to 30°C. To compare the effects of antiandrogens on transcription activation with AR and GFP-AR, we studied the antagonistic effect when the cells were incubated at 30°C with R1881, and either CA or OH-flu. The luciferase activity measured in response to 10 − 10 M of R1881, defined as 100%, was reduced significantly by

Fig. 3. Transcriptional activation of an androgen responsive gene (TAT-tk-luc) by AR and GFP-AR. (A) Luciferase activity induced by AR and GFP-AR after incubation of transfected CV-1 cells at 30°C with various concentrations of R1881. The fold increase in luciferase was determined relative to basal activity in the absence of R1881. Each point represents the mean of three independent experiments with duplicate determinations. (B) Transcriptional activity of the AR and GFP-AR induced by R1881 and by CA or OH-flu and transcriptional inhibition by CA or OH-flu vith R1881. The relative luciferase activity is expressed as percentage of the maximal induction observed with 0.1 nM R1881 in the absence of antiandrogens defined as 100%. CA and OH-flu were used at 10 − 5 M. Student’s t-test was used for statistical analysis compared with the induction obtained for R1881 alone. *, P B0.05; **, PB 0.01; ***, PB 0.005.

the addition of high concentrations of CA or OH-flu (Fig. 3B). CA reduced luciferase activity at 10 − 5 M to 80%, whereas 10 − 5 M OH-flu decreased the induction of luciferase activity to 50% (Fig. 3B). An important agonist effect of CA and OH-flu was measured when the cells were incubated with the antihormone alone. These antagonist and agonist effects were detected in the cells with both AR and GFP-AR.


V. Georget et al. / Molecular and Cellular Endocrinology 129 (1997) 17–26

Fig. 4. Androgen dependence of the trafficking of GFP-AR. The fusion protein was detected in live transfected COS-7 cells using an epifluorescence video microscope. A single living cell was observed without hormone (time 0 min) and after incubation with 10 − 6 M R1881 (time 30 min) by epifluorescence and by differential interference contrast (DIC) to determine the localization of the fluorescence. The arrows indicate an untransfected cell.

These experiments demonstrated that fusion proteins respond to androgens and antiandrogens in the same range as that observed for AR.

3.4. Qualitati6e analysis of fusion protein intracellular localization To study the localization of GFP-AR, COS-7 cells were transfected with either pS65T-Cl-GFP, pGFP-AR or GFP-AR-del4 to visualize the fluorescence without hormone, and to study the distribution of this fluorescence after hormonal addition. Living cells on coverslips were placed on a glass slide, and observed by DIC and by fluorescence with the Reichert microscope. Transfected cells were identified by GFP fluorescence. The level of fluorescence was very different from the cellular autofluorescence detected in untransfected cells and is indicated by arrows in Fig. 4. A greater fluorescence was obtained when the cells were cultured at 30°C after transfection. But the analysis of trafficking of GFP-AR with hormones was performed at 37°C to reproduce physiological conditions as much as possible. Cells were initially transfected by pS65T-Cl-GFP to control the localization of GFP alone, and after incubation with 10 − 6 M R1881. GFP appeared uniformly distributed throughout the cytoplasm and nucleus (data not shown). This localization has already been described for GFP wild-type into COS-l cells [21]. GFP is small enough (27 kDa) to pass through the nuclear pores by passive diffusion. The distribution was not affected by incubation with androgen (data not shown), and cells transfected with pGFP-AR without hormone showed a diffuse predominandy cytoplasmic fluores-

cence (Fig. 4). The intensity of the detected fluorescence depends on the cell depth. Thus the higher perinuclear fluorescence exhibited by some cells is due to a thicker cytoplasm rather than a particular staining pattern. When cells were incubated with 10 − 6 M R1881, after 30 min fluorescence became predominantly nuclear. This experiment demonstrated a hormone-dependent trafficking of GFP-AR, as previously described for the AR [22]. Simultaneous observation of the cell by DIC confirmed that the GFP-AR was in the total nuclear area, except in the nucleolus. As control, we used the pGFP-AR-del4 construction encoding for a mutant that lacks the carboxy-terminal function and does not present any transcriptional activity. This mutant was used to locate the fusion protein in COS-7 cells without and with hormone. GFP-AR-del4 was distributed in the cytoplasm without hormone in the same manner as GFP-AR, and the localization was not modified after addition of R1881 10 − 6 M for 1 h (data not shown). This mutant confirmed the hormone-dependent trafficking of AR, and clearly demonstrated the efficiency of our model. Similar experiments were performed with PC3 cells derived from prostate carcinoma transfected by GFP-AR. The same distribution was observed without hormone and the cytoplasmic fluorescence became nuclear after R1881 addition (manuscript in preparation).

3.5. Quantitati6e analysis of GFP-AR trafficking To carry out real-time imaging of protein trafficking in a single living cell, we used an interactive laser cytometer to define the intracellular localization and

V. Georget et al. / Molecular and Cellular Endocrinology 129 (1997) 17–26

Fig. 5. Dynamics of the translocation of GFP-AR induced by R1881. COS-7 cells expressing GFP-AR were cultured in a culture dish at 30°C and the fluorescence was analyzed directly by the ACAS 570 scanning laser cytometer. Living cells were observed and recorded without hormone (a) and with 10 − 6 M R1881 after 15 min (b), 30 min (c) and 60 min (d).

quantify the traffic of GFP-AR after R1881 incubation. Initially, we spotted a few fluorescent cells in the culture dish, each cell being recorded with the position parameters defined. After addition of R1881, each identified fluorescent cell was recorded approximately every 15 min. In this manner, the dynamics of GFP-AR with R1881 10 − 6 M (Fig. 5) was determined. Cytoplasmic fluorescence was observed without hormone as previ-


ously described and, after 15 min of androgen incubation at 37°C, the cytoplasmic fluorescence diminished whereas the nuclear fluorescence increased. After 60 min the fluorescence signal was predominantly nuclear, although a complete translocation was never observed. All the different cells in the dish exhibited the same kinetics. No trafficking was observed at room temperature even after 3 h of incubation with R1881, suggesting thus an active phenomenon. The ability of various androgens and antiandrogens to induce nuclear import of the GFP− AR was also investigated. Image analysis was used to estimate the ratio of GFP-AR nuclear transfer induced by hormones. Cells were incubated with a synthetic androgen, R1881, and two natural androgens, dihydrotestosterone (DHT) and testosterone, and two antiandrogens, steroidal antiandrogen CA and nonsteroidal antiandrogen OH-flu at 10 − 6 M. All the tested compounds were able to direct the GFP-AR into the nucleus with the same speed (Figs. 6 and 7A), and the only difference was the amount of nuclear fluorescence at steady state level (Fig. 7A). R1881 and DHT 10 − 6 M were able to induce import of nearly 70% of the fluorescence into the nucleus, while testosterone 10 − 6 M induced only 56% of nuclear fluorescence. CA and OH-flu induced, respectively, 36% and 28% of total fluorescence into the nucleus at 10 − 6 M, whereas no translocation into the nucleus was observed for 10 − 9 M. We also compared the trafficking of GFP-AR by androgens using two different concentrations: 10 − 9 M and 10 − 6 M. The proportion of nuclear import was hormone-concentration dependent (Fig. 7B). The use of GFP and the interactive laser cytometer thus enabled the analysis of trafficking with androgens and antiandrogens, which would not have been possible with immunotechniques.

Fig. 6. Hormone specificity of the trafficking of GFP-AR. COS-7 cells were incubated with a synthetic androgen, R1881, with two natural androgens, dihydrotestosterone and testosterone, and with two antihormones, cyproterone acetate and OH-flutamide, at 10 − 6 M for 60 min. The fluorescent cells were observed by the laser cytometer.


V. Georget et al. / Molecular and Cellular Endocrinology 129 (1997) 17–26

Fig. 7. Hormone specificity of the kinetics of the GFP-AR nuclear translocation. (A) Transfected cells were incubated with the different hormones and antihormones at 10 − 6 M. The fluorescence in a single living cell was recorded every 15 min. Fluorescence was measured in total cellular and nuclear area, and the percentage of nuclear fluorescence was calculated. (B) The nuclear translocation of GFP-AR was studied with two different concentrations of androgens (10 − 9 and 10 − 6 M). The results are expressed by the percentage of nuclear fluorescence in the saturating phase of trafficking.

4. Discussion In this study, we used GFP-AR to determine the dynamics of hormone-dependent trafficking of the androgen receptor. The GFP fluorescence is higher when cells are cultured at 30°C, Ogawa et al. [21] having described this temperature sensitivity for GFP transfected in vertebrate cells. They demonstrated that for fluorescence detection it is important to shift the incubation temperature after transfection from 37 to 30°C at least 4 h before fluorescence observation. The GFPhGR that they described efficiently transactivated the mouse mammary tumor virus (MMTV) promoter in response to dexamethasone at 30°C but not at 37°C, indicating that temperature is an important parameter

for the function of GFP fusion protein [21]. The underlying cause of this temperature dependence is still unknown. Ogawa et al. suggested that it may be associated with the folding and/or redox state of GFP within the cell. The decreased binding capacity of GFPAR observed by Scatchard analysis suggests a lower expression of GFP-AR compared with AR, possibly due to differences between the pCMV expression vectors ligated either with the AR cDNA or the GFP-AR cDNA. GFP-AR did, however, conserve essential functional characteristics of AR in the cell, such as the binding affinity of R1881 and transcriptional activation. The lower fold induction of GFP-AR with R1881 compared with AR could be due to the lower expression of the fusion protein, or could also be a consequence of the 1–36 amino acid deletion of the amino-terminal cDNA of AR in GFP-AR. Simental et al. [13] described a mutant with amino acids 1–141 deleted, which had a normal transcriptional activation of MMTV-CAT when cotransfected in CV-l cells. Jenster et al. [23], on the other hand, reported that N-terminal deletion of the first 141 amino acid residues resulted in a reduction in CAT activity to : 60% of activity as compared to the wild-type AR into Hela cells. The lower induction of GFP-AR could also be due to a more cumbersome conformation of the fusion protein, less appropriate than AR to transactivate TAT-tk-luc. We demonstrated that GFP can be fused to a androgen receptor without affecting its hormone binding function or its transcriptional capacity. R1881 10 − 6 M induced the trafficking of GFP-AR into the nucleus of COS-7 cells in a rapid process, as previously described for AR [12]. The characteristics of this fusion protein are comparable to wild-type AR. It contains the functional properties of a transcription factor that can be addressed through hormone-dependent transactivation of reporter genes such as TAT-tk-luc. GFP-AR, which is cytoplasmic in the absence of hormone, translocates to the nucleus upon binding to the hormone. Two in vitro analyses have been performed to apply the GFP model to GR: one [24], with GFP fused in frame to the second amino acid of rat GR (GFP-GR) was transfected in cell line 1471.1 derived from the murine adenocarcinoma C127 cell, and the other one [25] with GFP fused to the carboxy-terminal of GR (GR-GFP) was transfected in BHK21 cells (baby hamster kidney cells). In the first analysis, GFP-GR activated the transcription of the reporter gene with dexamethasone. Upon binding to dexamethasone, GFP-GR moved vectorially toward the nucleus. Shortly after GFP-GR entered the nucleus, bright fluorescence foci appeared for dexamethasone-induced cells. In the second analysis, Carey et al. [25] confirmed the hormone dependence of the nuclear translocation of GR-GFP by incubation of cells with 10 − 6 M dexamethasone, but no foci of

V. Georget et al. / Molecular and Cellular Endocrinology 129 (1997) 17–26

GR-GFP in the nucleus were described. In our study, we did not observe any foci in transfected cells with GFP-AR. This difference could be due to the high quantity of fusion protein produced in COS-7 cells compared with cell line 1471.1, which prevented us from observing foci in the high fluorescence background. Another possibility could be the difference in receptor and GFP-receptor construction. The ratio of GFP-AR translocation in the nucleus was hormone and concentration dependent, whereas translocation velocities could not be determined from the sparse time course used. Using GFP-AR, the trafficking was never complete even when high concentrations of androgens and antiandrogens were added. The partial transfer of GFP-AR is most likely related to the steady state level between the concentrations of the nuclear AR and cytoplasmic AR. This may reflect the partial shuttle of AR in the cell, as previously described for the progesterone receptor [4], which was shown to shuttle continuously between the nucleus and the cytoplasm. Guiochon-Mantel et al. [26] showed that the nuclear import requires energy, whereas the nuclear export does not. These results suggest that nucleocytoplasmic shuttling could be a general phenomenon for steroid receptors. The percentages of fluorescence found in the nucleus upon hormone treatment seem to be related to the Kd of AR for these hormones. Indeed, the affinities for the AR of DHT and R1881 are the greatest, followed by testosterone and then CA and OH-flu. It is possible that these affinities are determinant to balance the ratio between the cytoplasm and nucleus. No differences in the pattern of localization were noted between androgens and antiandrogens. Htun et al. [24] reported that the translocation rate of GFP-GR is dependent on hormone concentration, reflecting the dose and time dependence of GR action. The antagonist RU486-activated GFP-GR accumulates almost exclusively in a reticular pattern throughout the nucleus except in the nucleoli, whereas dexamethasone induces GFP-GR in foci in the nucleus. It is this former distribution of GFP-AR that we have described for androgens and antiandrogens. In conclusion, these results demonstrate that GFP is an effective reporter providing good temporal and spatial resolution in living cells. This new model has permitted us, for the first time, to analyze the dynamics of GFP-AR trafficking under androgen incubation in the same living cell. This new model could provide a novel approach for studying the dynamics of AR in target cells, such as prostatic cells or neurons. We believe that this tag would also be useful for investigating the structure-function relations of AR. This could be done by using different natural mutants found in androgen insensitivity syndromes, as has been demonstrated for hAR-del4. Moreover, since the fate of AR after its transcriptional action is not known, this new approach


could afford new insights into this field of investigation. Finally, GFP-AR can be used to screen new antagonists that are unable to induce AR nuclear translocation, which would provide more potent antiandrogens for treatment. Acknowledgements We are grateful to Dr P. Vago, INSERM U254, Montpellier France, for helpful technical discussions. Part of this work was supported by ARC (C.S.) from the Association de Recherche contre le Cancer and FNCLCC705193 (C.S.) from the Fe´de´ration Nationale des Centres de Lutte contre le Cancer and BMH4CT96-0181 from Biomed Program. References [1] Guiochon-Mantel, A., Delabre, K., Lescop, P. and Milgrom, E. (1996) Intracellular traffic of steroid hormone receptors. J. Steroid Biochem. Mol. Biol. 56, 1 – 6. [2] King, W.J. and Greene, G.L. (1984) Monolonal antibodies localize oestrogene receptor in the nuclei of target cells. Nature 307, 745 – 747. [3] Perrot-Applanat, M., Logeat, F., Groyer-Picard, M.T. and Milgrom, E. (1985) Irnmunocytochemical study of mammalian progesterone receptor using monoclonal antibodies. Endocrinology 116, 1473 – 1484. [4] Guiochon-Mantel, A., Lescop, P., Christin-Maitre, S., Loosfelt, H., Perrot-Applanat, M. and Milgrom, E. (1991) Nucleocytoplasmic shuttling of the progesterone receptor. EMBO J. 10, 3851 – 3859. [5] Lombes, M., Farman, N., Oblin, M.E., Baulieu, E.E., Bonvalet, J.P., Erlanger, B.F. and Gasc, J.M. (1990) Immunohistochemical localization of renal mineralocorticoid receptor by using an anti-idiotypic antibody that is an internal image of aldosterone. Proc. Natl. Acad. Sci. USA 87, 1086 – 1088. [6] Robertson, N.M., Schulman, G., Karnick, S., Alnemri, E. and Litwack, G. (1993) Demonstration of nuclear translocation of the mineralocorticoid receptor (MR) using an anti-MR antibody and confocal laer scanning microscopy. Mol. Endocrinol. 7, 1226 – 1239. [7] Martins, V.R., Pratt, W.B., Terracio, L., Hirst, M.A., Ringold, G.M. and Housley, P.R. (1991) Demonstration by confocal microscopy that unliganded overexpressed glucocorticoid receptors are distributed in a non random manner throughout all planes in the nucleus. Mol. Endocrinol. 5, 217 – 225. [8] Akner, G., Wikstrom, A.C. and Gustafsson, J.A. (1995) Subcellular distribution of the glucocorticoid receptor and evidence of its association with microtubules. J. Steroid Biochem. Mol. Biol. 52,1 – 16. [9] Brink, M., Humbel, B.M., De Kloet, E.R. and Van Driel, R. (1992) The unliganded glucocorticoid receptor is localized in the nucleus, not in the cytoplasm. Endocrinology 130, 3575–3581. [10] Sar, M., Lubahn, D.B., French, F.S. and Wilson, E.M. (1990) Immunohistochemical localization of the androgen receptor in rat and human tissues. Endocrinology 127, 3180 –3186. [11] Husmann, D.A., Wilson, C.M., McPhaul, M.J., Tilley, W.D. and Wilson, J.D. (1990) Antipeptides antibodies to two distinct regions of the androgen receptor localize the receptor protein to the nuclei of target cells in the rat and human prostate. Endocrinology 126, 2359 – 2368.

V. Georget et al. / Molecular and Cellular Endocrinology 129 (1997) 17–26


[12] Jenster, G., Trapman, J. and Brinkmann, A.O. (1993) Nuclear import of the human androgen receptor. Biochem. J. 293, 761 – 768. [13] Simental, J.A., Sar, M., Lane, M.V., French, F.S. and Wilson, E.M. (1991) Transcriptional activation and nuclear targeting signals of the human androgen receptor. J. Biol. Chem. 266, 510 – 518. [14] Chalfie, M., Tu, Y., Euskirchen, G., Ward, W.W. and Prasher, D.C. (1994) Green fluorescent protein as a marker for gene expression. Science 263, 802–805. [15] Heim, R., Cubitt, A.B. and Tsien, R.Y. (1995) Improved green fluorescence. Nature 373, 663–664. [16] Flach, J., Bossie, M., Vogel, J., Corbett, A., Jinks, T., Willins, D.A. and Silver, P.A. (1994) A yeast RNA-binding protein shuttles between the nucleus and the cytoplasm. Mol. Cell. Biol. 14,8399 – 8407. [17] Pines, J. (1995) GFP in mammalian cells. Trends Genet. 11, 326 – 327. [18] Brown, T.R., Lubahn, D.B., Wilson, E.M., French, F.S., Migeon, C.J. and Corden, J.L. (1990) Functional characterisation of naturally occurring mutant androgen receptors from subjects with complete androgen insensitivity. Mol. Endocrinol. 4, 1759 – 1772. [19] Lobaccaro, J.M., Lumbroso, J.M., Poujol, N., Georget, V., Brinkmann, A.O., Malpuech, G. and Sultan, C. (1995) Complete androgen insensitivity syndrome due to a new frameshift deletion in exon 4 of the androgen receptor gene: functional analysis of the mutant receptor. Mol. Cell. Endocrinol. 111, 21 – 28. [20] van Laar, J.H., Voorhost-Ogink, M.M., Zegers, N.D., Boersma, W.J.A., Claassen, E., van der Korput, J.A.G.M., Ruizeveld de







Winter, J.A., Van der Kwast, T.H., Mulder, E., Trapman, J. and Brinkmann, A.O. (1989) Characterization of polyclonal antibodies against the N-terminal domain of the human androgen receptor. Mol. Cell. Endocrinol. 67, 29 – 38. Ogawa, H., Inouye, S., Tsuji, F.I., Yasuda, K. and Umesono, K. (1995) Localization, trafficking, and temperature – dependence of the aequorea green fluorescent protein in cultured vertebrate cells. Proc. Natl. Acad. Sci. USA 92, 11899 – 11903. Jenster, G., van der Korput, H.A.G.M., van Vroonhoven, C., van der Kwast, T.H. and Brinkmann, A.O. (1991) Domains of the human androgen receptor involved in steroid binding, transcriptional activation, and subcellular localization. Mol. Endocrinol. 5, 1396 – 1404. Jenster, G., Vanderkorput, H.A.G.M., Trapman, J. and Brinkmann, A.O. (1995) Identification of two transcription activation units in the n-terminal domain of the human androgen receptor. J. Biol. Chem. 270, 7341 – 7346. Htun, H., Barsony, J., Renyi, I., Gould, D.L. and Hager, G.L. (1996) Visualization of glucocorticoid receptor translocation and intranuclear organization in living cells with a green fluorescent protein chimera. Proc. Natl. Acad. Sci. USA 93, 4845–4850. Carey, K.L., Richards, S.A., Lounsbury, K.M. and Macara, I.G. (1996) Evidence using a green fluorescent protein –glucocorticoid receptor chimera that the RAN/TC4 GTPase mediates an essential function independent of nuclear protein import. J. Cell Biol. 133, 985 – 996. Guiochon-Mantel, A., Delabre, K., Lescop, P. and Milgrom, E. (1994) Nuclear localization signals also mediate the outward movement of proteins from the nucleus. Proc. Natl. Acad. Sci. USA 91, 7179 – 7183.

. .