Dynamics of Intracellular Movement and Nucleocytoplasmic Recycling of the Ligand-Activated Androgen Receptor in Living Cells
Rakesh K. Tyagi, Yan Lavrovsky, Soon C. Ahn, Chung S. Song, Bandana Chatterjee, and Arun K. Roy Department of Cellular and Structural Biology (R.K.T., Y.L., S.C.A., C.S.S., B.C., A.K.R.) The University of Texas Health Science Center at San Antonio and Audie L. Murphy Memorial Veterans Affairs Hospital (B.C.) San Antonio, Texas 78284
An expression construct containing the cDNA encoding a modified aequorea green fluorescent protein (GFP) ligated to the 5ⴕ-end of the rat androgen receptor (AR) cDNA (GFP-AR) was used to study the intracellular dynamics of the receptor movement in living cells. In three different cell lines, i.e. PC3, HeLa, and COS1, unliganded GFP-AR was seen mostly in the cytoplasm and rapidly (within 15–60 min) moved to the nuclear compartment after androgen treatment. Upon androgen withdrawal, the labeled AR migrated back to the cytoplasmic compartment and maintained its ability to reenter the nucleus on subsequent exposure to androgen. Under the condition of inhibited protein synthesis by cycloheximide (50 g/ml), at least four rounds of receptor recycling after androgen treatment and withdrawal were recorded. Two nonandrogenic hormones, 17-estradiol and progesterone at higher concentrations (10ⴚ7/10ⴚ6 M), were able to both transactivate the AR-responsive promoter and translocate the GFPAR into the nucleus. Similarly, antiandrogenic ligands, cyproterone acetate and casodex, were also capable of translocating the cytoplasmic AR into the nucleus albeit at a slower rate than the androgen 5␣-dihydrotestosterone (DHT). All AR ligands with transactivation potential, including the mixed agonist/antagonist cyproterone acetate, caused translocation of the GFP-AR into a subnuclear compartment indicated by its punctate intranuclear distribution. However, translocation caused by casodex, a pure antagonist, resulted in a homogeneous nuclear distribution. Subsequent exposure of the casodextreated cell to DHT rapidly (15–30 min) altered the homogeneous to punctate distribution of the already translocated nuclear AR. When transported into the nucleus either by casodex or by DHT, GFP-AR was 0888-8809/00/$3.00/0 Molecular Endocrinology 14(8): 1162–1174 Copyright © 2000 by The Endocrine Society Printed in U.S.A.
resistant to 2 M NaCl extraction, indicating that the homogeneously distributed AR is also associated with the nuclear matrix. Taken together, these results demonstrate that AR requires ligand activation for its nuclear translocation where occupancy by only agonists and partial agonists can direct it to a potentially functional subnuclear location and that one receptor molecule can undertake multiple rounds of hormonal signaling; this indicates that ligand dissociation/ inactivation rather than receptor degradation may play a critical role in terminating hormone action. (Molecular Endocrinology 14: 1162–1174, 2000)
INTRODUCTION Steroid hormone receptors are ligand-activated transcription factors. Initial identification of the estrogen receptor in the early 1960s and studies on its mechanism of action through subcellular fractionation led to the development of a two-step model for steroid hormone action (1, 2). In this model, unliganded receptor was thought to be localized in the cytoplasm as a heteromeric complex with heat shock proteins and, upon ligand binding, to undergo conformational transition, dissociation from the heteromeric partners, homodimerization, and nuclear translocation to initiate target gene regulation. Subsequent studies with different approaches, such as immunostaining, autoradiography with labeled steroids, and cytofusion analysis, have generated conflicting pictures with respect to the intracellular distribution of unliganded steroid hormone receptors (3–15). In the case of androgen receptor, primary localization of the unliganded receptor either in the cytoplasmic (9) or in the nuclear compartment (10) of transfected cells has been reported. The prevalent notion that, with the exception of the glucocorticoid receptor, unliganded steroid receptors are nuclear proteins is largely based on the assumption that observations of the cytoplasmic existence of 1162
Intracellular Movement of AR
other unliganded receptors by cell fractionation may reflect nuclear leakage of these proteins due to their loose association with the chromatin (16, 17). In addition to introducing experimental artifacts, both biochemical and histochemical approaches can provide only a static picture of the otherwise dynamic state of a living cell. Recent advances in imaging techniques allow monitoring the intracellular molecular movements in real time (18). Chimeric gene constructs expressing proteins of interest tagged with the modified and improved version of aequorea green fluorescent protein (GFP) can be used to follow movement of proteins in real time through wide-field water immersion objectives and a high-resolution charge-coupled detection device (19). We have used this approach to explore the intracellular movement and subnuclear compartmentalization of the androgen receptor (AR) after ligand treatment and ligand withdrawal. We report that most of the unliganded AR under the steadystate condition resides in the cytoplasm and, upon hormone exposure, rapidly migrates into the nucleus. Furthermore, androgen withdrawal releases the receptor from its chromatin association and exports it back into the cytoplasmic compartment for recycling when the hormone is reintroduced. Additionally, we show that ligands with both transactivation potential and inhibitory action can cause translocation of the AR from cytoplasm to nucleus albeit to a varying extent. However, only ligands with agonist activity and not the pure antagonist, i.e. casodex, are capable of translocating the AR into a distinct subnuclear compartment. The role of ligands in nuclear translocation of GFP-AR with a truncated AR and a weaker version of GFP has been reported previously (20).
1163
We then examined the localization of the GFP-AR in three different cell lines, i.e. prostate-derived PC3, uterus-derived HeLa, and kidney-derived COS1 cells. In the absence of hormone, GFP-AR was found to be mostly cytoplasmic in all three cell types (Fig. 1, A–C). However, addition of DHT to the culture medium resulted in a time-dependent translocation of the GFP-AR into the nuclear compartment. Nuclear migration of the receptor was rapid and clearly evident within 15 min after hormone treatment, and the receptor became primarily or exclusively nuclear within 60 min. A similar DHT-mediated nuclear translocation was also observed by immunostaining of both the endogenous AR in AR-positive LNCaP cells and COS1 cells transfected with the pCMV-AR expression vector (Fig. 2, A and B). We also observed that, at any particular hormone concentration, there were cell-to-cell variations in the extent of nuclear translocation, i.e. some of the transfected cells were more responsive than others in the androgen-dependent nuclear import of the GFP-AR. Therefore, to obtain a normalized rate of nuclear import, we scored 100 transfected cells within a field for the nucleocytoplasmic fluorescence intensity. Cells with exclusively nuclear and predominantly nuclear fluorescence (N⬎C) within the same field were counted at different time points after hormone exposure. Similar quantitation was also performed with immunostained COS1 cells transfected with the wildtype rat AR expression plasmid. Results presented in Fig. 3 show the kinetics of hormone-dependent nuclear translocation of the GFP-AR and immunodetected wild-type AR in transfected COS1 cells. In both cases, a linear increase in nuclear import after DHT treatment was observed, and the import was complete between 60 to 90 min.
RESULTS Ligand-Dependent Transactivational Activity of the Chimeric GFP-AR and Kinetics of Its Nuclear Translocation A fusion protein containing an amino-terminal extension of the intact rat AR with GFP was used for this study. We first tested the ability of this chimeric protein for androgen-dependent transactivation of the mouse mammary tumor virus-chloramphenicol acetyltransferase (MMTV-CAT) promoter-reporter in AR-negative PC3 cells. In the presence of 10⫺9 M 5␣-dihydrotestosterone (DHT), cells transfected with the chimeric (GFP-AR) expression vector showed about one third transactivation function as compared with the nonchimeric AR (relative CAT activities of 183 ⫾ 19 for GFP-AR vs. 588 ⫾ 11 for AR). The partial loss of transactivational activity of the GFP-AR may be due to the GFP interference with the amino-terminal TAF-1 function of the AR protein. However, transactivation function of both of these proteins was almost totally dependent on the presence of the androgen (relative CAT activity without DHT of less than 3).
Nucleocytoplasmic Recycling of AR after Androgen Treatment and Androgen Withdrawal The dynamic nature of the nucleocytoplasmic movement of the AR and its hormonal control was explored further by imaging of the GFP-AR fluorescence after androgen treatment and withdrawal. For this set of experiments, we used COS1 cells, where the expression plasmid is subjected to intracellular amplification, yielding a higher level of GFR-AR that is necessary for monitoring the decaying fluorescent protein after prolonged cycloheximide inhibition of new protein synthesis. Results presented in Fig. 4A show a reversal of cytoplasmic to nuclear migration of the GFP-AR after hormone withdrawal in the presence of 50 g/ml cycloheximide. At 12 h after hormone withdrawal, no significant difference in the intensity of fluorescence between nuclear and cytoplasmic compartments can be distinguished. Since export of the receptor from the nuclear to the cytoplasmic compartment only became evident between 6 to 12 h after hormone withdrawal, it was of interest to compare the rate of ligand dissociation with nuclear export. Data presented in Fig. 4B
MOL ENDO · 2000 1164
Vol 14 No. 8
Fig. 1. Androgen-Dependent Nuclear Translocation of GFP-AR in Transfected Cells A, PC3; B, HeLa; C, COS1. Images from the same cells were acquired at 0, 15, 30, and 60 min after treatment with 10⫺8 M DHT. Bar, 15 m.
show about 50% ligand dissociation within the same time frame when nuclear export became clearly evident. A discordant relationship between ligand dissociation and nuclear export of GR has been reported earlier (21). We speculate that a dynamic process involving capture of the dissociated ligand from one receptor molecule by another unoccupied receptor provides a threshold effect of nuclear retention after hormone withdrawal. The fact that the concentration of cycloheximide used in these experiments (50 g/ml) caused a total inhibition of immunodetectable GFP-AR synthesis is demonstrated by results presented in Fig. 4C. Even at 10 g/ml, no immunodetectable GFP-AR can be seen within 15 h of culture period while untreated cells produced a high level of this protein. The GFP-AR exported from the nucleus after 12 h of hormone withdrawal could again be transported into the nuclear compartment after reexposure to its hormonal ligand (Fig. 5A). Quantitative analysis of this reutilization process for up to four cycles of import and three cycles of export by sequential hormone treatment and withdrawal (Fig. 5B) showed a gradual decline of the recycling competency of the receptor protein. We interpret this observation to indicate a time-dependent structural damage (degradation) of the receptor protein. However, a decreased level of importins or other
proteins in nucleocytoplasmic transport after inhibition of protein synthesis by cycloheximide may also be a contributing factor. During the first two rounds of recycling, the receptor was almost equally translocation competent. Upon subsequent hormone treatment the same batch of cells at their third and fourth import cycles showed a gradual decline of ligand-dependent nuclear import. Thus, the same receptor may be able to mediate multiple rounds of hormonal signaling, which suggests that inactivation and/or dissociation of hormonal steroids, rather than nuclear degradation of the receptor protein after its initial import, may play a dominant role in the termination mechanism of hormone action. The nuclear import of steroid receptors mediated by importins interacting with the nuclear localization signal (NLS) has been fairly well characterized (9, 22–27). However, the corresponding nuclear export signal (NES) of the steroid hormone receptors has not been identified, and the mechanism of the nuclear export of steroid receptors is not clearly understood (27). Leptomycin B (LMB) is a specific inhibitor of exportins, a class of chaperone proteins that bind to NES and facilitate the export of nuclear proteins into the cytoplasmic compartment (28, 29). Recently the role of exportin in the nucleocytoplasmic recycling of a cytoplasmically sequestered sig-
Intracellular Movement of AR
1165
Fig. 2. Androgen-Dependent Nuclear Translocation of Untagged AR Detected by Immunostaining A, Untransfected LNCaP cells. B, COS1 cells transfected with pCMV-AR expression vector. Cells in separate culture dishes were treated with 10⫺8 M DHT and fixed at different time intervals (0, 15, 30, and 60 min) as indicated on the top of the figure. Bar, 15 m.
ically active and that the nuclear export of the liganddissociated AR may involve an as-yet-uncharacterized exportin-independent transport mechanism. Nuclear Translocation and Differential Subnuclear Compartmentalization of GFP-AR by CrossReactive AR Agonists and Inhibitors of AR Function
Fig. 3. Time course of Nuclear Translocation of GFP-AR and Untagged AR in Transfected COS1 Cells Cells in separate culture dishes were exposed to DHT (10⫺8 M) and fixed at different time intervals for either direct observation (GFP-AR) or immunostaining (untagged AR). Each point represents percent cells with predominantly nuclear fluorescence. Open circles, GFP-AR; closed circles, AR without the GFP tag.
nal-activated transcription factor, i.e. p65 subunit of nuclear factor-B (NFB), has been described (30). Using p65-GFP as our positive control, we examined the effect of LMB on the nuclear export of the GFP-AR after androgen withdrawal. Although LMB failed to prevent the export of GFP-AR from the nuclear to the cytoplasmic compartment, the conditioned medium derived from the GFP-AR transfected cells containing LMB was able to prevent the nuclear export of p65-GFP with consequent accumulation of this protein in the nucleus (Fig. 6). The latter observation shows that LMB added to the culture medium of cells transfected with GFP-AR is still biolog-
The ligand-binding domain of the AR is known to cross-react with 17-estradiol (31). Activation of steroid hormone receptors by growth factors such as epidermal growth factor (EGF) has also been reported (32–36). We have examined relative translocation and transactivation potentials of these and other potential AR modulators in the living cell. Results presented in Fig. 7A show that both 17-estradiol and progesterone, especially at a higher concentration (10⫺7 and 10⫺6 M), were effective in translocating the GFP-AR into the nucleus. However, the gluococorticoid dexamethasone and EGF were totally ineffective. Two antiandrogens, cyproterone acetate (a mixed agonist/ antagonist) (37) and casodex (a pure antagonist) (38), also showed a relatively limited ability to translocate GFP-AR into the nucleus. When these hormones and antihormones were tested for their relative transactivation function for the AR-dependent promoterreporter construct derived from the rat probasin gene (i.e. ARR3-TK-Luc) (39), it showed that in addition to DHT, 17-estradiol, progesterone, and cyproterone acetate were transactivation competent (Fig. 7B). However, the relative abilities of these steroids for nuclear import and transactivation function were different. Next to DHT, estradiol showed the best translocation function, while cyproterone acetate was more competent in transactivation than estradiol. Among
MOL ENDO · 2000 1166
Vol 14 No. 8
Fig. 4. Export of the AR into the Cytoplasmic Compartment after Hormone Withdrawal A, Time course of nuclear-to-cytoplasmic export of the AR after androgen withdrawal. COS1 cells were first exposed to DHT for 4 h to translocate the GFP-AR into the nucleus and then incubated in the hormone-free medium containing cycloheximide (50 g/ml) and imaged at different time intervals (0, 2, 6, and 12 h). B, Intracellular dissociation kinetics of 3H-DHT in the presence of excess unlabeled DHT. C, Western blot showing the effect of 10 and 50 g/ml cycloheximide on GFP-AR synthesis in transfected cells. Cycloheximide was added 2 h after transfection, and cells were harvested 15 h later for Western blotting. Lane 1 contains sample derived from untransfected cells cultured in the presence of 50 g/ml cycloheximide; lanes 2, 3, and 4 contain samples derived from transfected cells in the presence of 0, 10, and 50 g/ml cycloheximide, respectively. The location of the GFP-AR band is marked with the arrowhead. The lower band marked with an asterisk represents a nonspecific cellular protein that cross-reacts with the antibody.
the translocation-competent ligands that we have tested, casodex, a pure antagonist, was found to be the least effective. Again, some cells were more sensitive to casodex-mediated translocation than others and after a prolonged exposure (⬎20 h) to 10⫺6 M casodex, about 70% of the cells showed primarily nuclear fluorescence (data not shown). Additionally, we noted that the pattern of nuclear distribution of the GFP-AR was distinctly different for ligands with or without any transactivation activity. DHT (a pure hormonal agonist), 17-estradiol (a crossreactive hormonal agonist), and cyproterone acetate (a mixed agonist/antagonist) showed a punctate distribution pattern indicative of the association of the translocated receptor within a subnuclear compartment. However, in the case of casodex (a pure antagonist) the translocated receptor was evenly distributed (Fig. 8). Since proteasome-dependent nuclear degradation of the estrogen receptor has been observed (40), we wanted to rule out the possibility of protea-
somic inclusion of GFP-AR as the cause of the punctate distribution. Treatment of GFP-AR transfected cells with the proteasome inhibitor MG132 (3 M) before DHT treatment did not alter the intranuclear punctate distribution of the GFP-AR (data not presented). This observation indicates that the formation of the punctate foci is an intermediate step in the signaling rather than degradation process. That ligand-mediated nuclear translocation and subnuclear compartmentalization represent two distinct steps in the process of hormonal signaling was further indicated by a rapid conversion of the nucleoplasmic GFP-AR of the casodex-treated cell by subsequent treatment with DHT (Fig. 9A). In this case punctate appearance was evident within 15 min, and formation of the punctate foci was complete within 1 h after DHT treatment. Salt extraction of nuclear proteins followed by immunoblot analysis showed that most of the GFP-AR within the nuclei of either casodex-treated or DHT-treated cells is present within the salt (2 M NaCl)-resistant nuclear
Intracellular Movement of AR
1167
Fig. 5. Nucleocytoplasmic Recycling of GFP-AR after Androgen Treatment and Withdrawal A, Intracellular distribution of GFP-AR before hormone treatment (a), 4 h after treatment with 10⫺8 M DHT (b), 12 h after hormone withdrawal in the presence of 50 g/ml cycloheximide (c), and 4 h after hormone treatment (in the presence of cycloheximide) to hormone-withdrawn cells (d). Bar, 15 m. B, Quantitative analysis of cells with exclusively nuclear fluorescence after four cycles of DHT-mediated import and three cycles of export after hormone withdrawal. Each histogram represents average values from four independent experiments ⫾ SD.
matrix fraction. Thus, the punctate foci may represent a distinct step in steroid receptor function after its binding to the nuclear matrix.
DISCUSSION Until recently, with few exceptions (14, 20, 41, 42), technical limitations did not allow real time imaging of nucleocytoplasmic movement of steroid hormone receptors within a single cell. Although static pictures have generated a wealth of information, fixation artifacts and nuclear leakage during the cell fractionation
procedure have often led to controversial conclusions. The general consensus, based on these static approaches, suggests that, in the unliganded state, certain members of the steroid hormone receptor family, such as the glucocorticoid receptor, reside in the cytoplasmic compartment, whereas another class typified by the estrogen receptor, irrespective of the ligand-receptor interaction, is always located in the nucleus (4, 7, 25, 43). However, the subcellular location of the unliganded AR has still remained controversial (9, 10, 16). Recent improvements in the imaging technology with wide-field water immersion objectives can be used to examine the intracellular movement of
MOL ENDO · 2000 1168
Vol 14 No. 8
Fig. 6. LMB-Insensitive Nuclear Export of GFP-AR The upper frames show that LMB is ineffective in preventing the nuclear export of GFP-AR after androgen withdrawal. The lower frames show that nucleocytoplasmic recycling of p65-GFP is almost totally inhibited when the conditioned medium containing LMB was added to cells transfected with the chimeric expression vector producing p65 subunit of NFB tagged with GFP. Frame a, COS1 cells treated with DHT for 4 h without LMB; frame b, after 12 h of hormone withdrawal in the presence of LMB. Both frames a and b show cells transfected with GFP-AR. Frames c and d, p65-GFP transfected cells before and 60 min after incubation with LMB containing conditioned medium retrieved from DHT-withdrawn cells used for frame b. Bar, 15 m.
GFP-labeled proteins in the living cell with a high degree of resolution and in real time. Additionally, the newly modified version of the jellyfish GFP is not only stable at 37 C in mammalian cells but also emit stronger fluorescence signal allowing its accurate localization within various subcellular compartments. This avoids the need for disruption of the cellular dynamics or lowering of the temperature for stabilizing the chimeric protein. The GFP-AR chimera that we have used maintains its ligand-dependent transactivation function at 37 C, albeit to a lesser extent than that of the unmodified AR. Our unpublished results show that introduction of a short histidine tag (Met-Arg-Gly-SerHis-His-His-His) at the N-terminal part of the AR also reduces its transactivation potential to a similar extent. In both cases, functional interference with the Nterminal TAF-1 domain seems to be the likely possibility. In an earlier study a chimeric GFP-AR was also used to examine its nuclear translocation (20). Due to the lower fluorescence intensity of the earlier version of GFP, use of an N-terminally truncated form of AR, whole-cell imaging, and culturing of cells at 30 C to stabilize the GFP-AR, observations reported by these authors are different from those presented in this article. The newly modified GFP (i.e. EGFP) emits ap-
proximately 35-fold more intense fluorescence than the wild-type GFP and the chimeric protein is stable at 37 C. Additionally, optical sectioning allows finer delineation and uniform imaging than the whole-cell images, which blur the distribution pattern of the fluorescent protein due to differences in the thickness of various parts of the cell. Because of a combination of all of these experimental problems, these investigators observed only very low transactivation function of GFP-AR, could not detect complete nuclear translocation after hormone treatment, and failed to observe any subnuclear sequestration of AR (20). In all of the cell lines that we have examined, the unliganded GFP-AR is primarily localized in the cytoplasmic compartment, and upon androgen exposure, rapidly migrates into the nucleus. Translocation of AR can be detected within a few minutes after hormone treatment and is almost complete within 60 min. Translocation kinetics of the GFP-AR are very similar to the pattern observed in immunostained COS1 cells containing unmodified AR. We have also observed that both the unliganded GFP-ER and GFP-peroxisome proliferator-activated receptor (PPAR␣) are primarily nuclear in COS1 cells (data not shown). From all of these results and the existing literature it is reason-
Intracellular Movement of AR
1169
Fig. 7. Relative Effects of Hormones and Antihormones on Nuclear Translocation and Transactivation Function of GFP-AR A, Percent cells with exclusively nuclear fluorescence at 3 h after hormone/antihormone treatment. Each histogram represent average of two independent experiments. Sixty-seven percent of the transfected cells treated with 10⫺8 M DHT showed exclusively nuclear fluorescence, which was used as 100% for comparing the potencies of other agents. B, Transactivation of ARR3-TK-Luc by GFP-AR after treatment with different doses of hormones and antihormones as labeled. E2, 17-estradiol; Prog, progesterone; Dex, dexamethasone; EGF, epidermal growth factor (50, 100, and 200 ng/ml); CA, cyproterone acetate; Cdx, casodex. Each histogram represents average values of six independent experiments ⫾ SD.
able to conclude that members of the steroid hormone receptor family can be classified into two distinct groups, one that requires ligand binding for nuclear translocation [e.g. glucocorticoid receptor (GR) and AR] and another that is nuclear, even without ligand binding [e.g. estrogen receptor (ER), progesterone receptor (PR), and PPAR␣]. Structural features that differentiate these two groups are presently unknown. Cytoplasmic proteins that are larger than 60 kDa are transported into the nuclear compartment through a chaperone-mediated transport process (44). In the case of unliganded GR and AR, the NLS, which interacts with chaperone proteins (importins), is likely to be masked by receptor-associated heat shock proteins and immunophilins. This group of receptors may require ligand-mediated conformational change to unmask the NLS site for its appropriate interaction with importins, whereas in the case of ER and PPAR␣, the NLS may be easily accessible to importins, even in the absence of their respective ligands.
Our observation of the nucleocytoplasmic recycling of the GFP-AR in living cells has an important bearing on the termination mechanism of the signaling process. We have observed multiple cycles of the DHT-dependent import-export process of GFP-AR after inhibition of new protein synthesis in the presence of cycloheximide. These results suggest that termination of nuclear signaling may depend on dissociation of the hormonal ligands due to ligand withdrawal or ligand inactivation. However, the receptor protein, after its dissociation from the ligand and chromatin, may be reutilized for another round of the signaling process when the ligand is again made available. A recent report has suggested an alternative mechanism for the termination of hormonal signal by nuclear degradation of the specific receptor protein through ubiquitination followed by proteasomal degradation (40). However, the quantitative contribution of such a process of receptor degradation in the termination of hormonal signaling has not been established. Our results indicate that, upon ligand withdrawal, most of the GFP-AR is
MOL ENDO · 2000 1170
Vol 14 No. 8
Fig. 8. Subnuclear Localization of GFP-AR after Nuclear Translocation by Hormones and Antihormones The pictures show the pattern of nuclear localization in transfected COS1 cells after translocation induced by DHT (10⫺8 M), cyproterone acetate (CA, 10⫺6 M), casodex (10⫺6 M), and 17-estradiol (E2, 10⫺6 M). Except casodex (a pure antagonist), all other ligands produced a punctate nuclear distribution. Bar, 10 m.
exported back into the cytoplasmic compartment and is still competent to undergo the ligand-dependent translocation process. Multiple rounds of hormonal signaling by recycled receptors and cross-reactive ligands under certain clinical conditions, such as recurrent forms of AR-positive prostate cancer, may contribute to the etiology of the disease. In a large number of these cases there are abnormally high levels of AR, due to either AR gene amplification or possibly to AR up-regulation after prolonged periods of androgen deprivation and antiandrogen therapy (45, 46). These recurrent cancer cells expressing high levels of AR may experience adequate androgenic response caused by multiple rounds of receptor recycling through intermittent exposure of the AR to cross-reactive partial agonists of nongonadal origin. The role of exportin/CRM1 chaperone proteins in the nuclear-to-cytoplasmic export of a number of proteins has been well established (27, 47–49). This mechanism is exquisitely sensitive to inhibition by LMB. LMB sensitivity to nuclear export of PR and GR by the immunostaining method has been examined earlier. Nuclear export of PR is insensitive to LMB (50), and contradictory results have been reported for GR (51, 52). Our results are consistent with the conclusion that the nuclear export of AR is mediated through an LBM-insensitive exportin-independent process, the nature of which has yet to be established. We feel that the results presented in this report concerning the LMB insensitivity of the export process are especially convincing, because 1) our observations were made in the living cell, without any additional intervention; and 2) inhibition of the nuclear export of GFP-
p65 by the LMB-containing conditioned medium derived from the GFP-AR-transfected cells. We have observed a general correlation between the dose-dependent transactivational function of AR by two other natural steroid hormones (i.e. 17-estradiol and progesterone, that are capable of cross-reacting with AR), and their effectiveness for translocating the receptor into the nucleus. Both dexamethasone and EGF were unable either to cause translocation of GFP-AR into the nucleus or to transactivate the ARR3-thymidine kinase (TK)-Luc in the cell transfection assay. A number of studies have suggested the potential role of EGF in enhancing steroid receptor function (32–36). The results presented in this article indicate that such a modulating influence of EGF may not be due to a direct influence on either the nuclear translocation or transactivation function of the unliganded AR and may be dependent on initial androgenic activation (32). It is also of interest to note that both a mixed agonist/ antagonist (cyproterone acetate) and a pure antagonist (casodex) were moderately effective in mediating the nuclear translocation of GFP-AR as compared with DHT. However, the patterns of nuclear distribution of the translocated receptor in these two cases are distinctly different. Similar to other agonists, the cyproterone acetate treatment led to a punctate nuclear distribution of GFPAR, and, in the case of casodex, the fluorescence was homogeneously distributed within the nuclear compartment. However, upon subsequent exposure to DHT, the receptor moved to a distinct subnuclear compartment. A similar difference in the distribution pattern of GR and ER
Intracellular Movement of AR
1171
Fig. 9. Conversion of Homogeneous to Punctate Nuclear Distribution of Casodex-Treated Cells Subsequently Exposed to DHT A, Fluorescence pattern of a transfected COS1 cell initially treated with casodex (10⫺6 M) for 3 h and subsequently treated with 10⫺8 M DHT. A single cell showing exclusively nuclear fluorescence was imaged at 0, 15, 30, and 60 min after DHT treatment. B, Western blot of nuclear extracts of cells treated with casodex and DHT. COS1 cells were cultured for 15 h in the presence of 10⫺6 M casodex (CDX), for 4 h in the presence of 10⫺8 M (DHT), and 19 h in the presence of 10⫺6 M casodex followed by 4 h in the presence of 10⫺8 M DHT (CDX 3 DHT). Monoclonal mouse antibody to GFP was used to detect the GFP-AR. Three lanes for each of the treatment condition represent as follows: F1, supernatant after extraction with 0.25 M ammonium sulfate; F2, supernatant after extraction with 2 M NaCl; F3, pellet containing nuclear matrix.
after agonist and antagonist treatment has also been reported (41, 53, 54). Similar to ER (53), our results show that both homogeneous and punctate forms of GFP-AR are associated with the nuclear matrix. Thus, formation of the punctate foci appears to be a distinct step in the mechanism of steroid hormone action, and this step lies beyond the initial matrix association of the receptor protein. It is of interest to note that only ligands with agonist or partial agonist activity can carry the process up to this step and beyond to initiate transcriptional activation.
region of the modified aequorea GFP (pEGFP-C2, CLONTECH Laboratories, Inc. Palo Alto, CA). Chimeric GFPER␣ was constructed similarly by inserting the human ER␣ cDNA into the BamHI site of the pEGFP-C2 vector. For generating the chimeric GFP construct of the p65 subunit of NFB, the human p65 cDNA fragment was excised from the pCMV-p65 expression vector (provided by Dr. John Cidlowski, NIEHS, Research Triangle Park, NC) and inserted into the BamHI site of the GFP expression plasmid, pEGFPN3. Nucleotide sequences of all DNA constructs were authenticated by manual sequencing. Construction of the expression vectors pCMV-AR and ARR3-TK-Luc has been described previously (39). The promoter-reporter construct pMMTV-CAT was a gift from Dr. Stephen Harris (The University of Texas Health Science Center, San Antonio, TX).
MATERIALS AND METHODS Plasmid Constructs
Cell Culture, Transfection, Cell Fractionation, and Western Blot Analysis
Chimeric GFP-AR was generated by inserting the rat AR cDNA (provided by Dr. Shutsung Liao, University of Chicago, Chicago, IL) into the BamHI site of the 3⬘- end of the coding
PC3, HeLa, LNCaP, and COS1 cells were obtained from American Type Culture Collection (Manassas, VA) and cultured in serum-containing media as recommended by the
MOL ENDO · 2000 1172
supplier. Cells were plated in six-well culture flasks at 150 ⫻ 103 cells per well in the growth medium (MEM supplemented with 5% charcoal-stripped FBS), grown overnight, and cotransfected with either pARR3-TK-Luc or pMMTV-CAT promoter-reporter (1 g) along with either wild-type steroid receptor-expression plasmid (pCMV-rAR) or the chimeric GFP plasmids (pCMV-GFP-rAR, pCMV-GFP-ER), using FuGENE 6 (Roche Molecular Biochemicals, Indianapolis, IN) reagent, 6 l/well, or LipofectAMINE (Life Technologies, Inc., Gaithersburg, MD), 8 l/ml. Media were changed at 24 h after transfection and were replaced with fresh media with or without the hormonal ligands. Cells were incubated for an additional 24 h before harvesting and extraction of proteins. Cell extracts were assayed for protein concentration (Bradford protocol), luciferase activity (assay kit, Promega Corp., Madison, WI), and CAT by enzyme-linked immunoassay (CAT ELISA kit, Roche Molecular Biochemicals). For CAT activity results were expressed as optical density (⫻ 103) per g protein. For experiments involving imaging of GFP fluorescence, after transfection, cells were allowed to express the appropriate chimeric protein for 30 h before any hormonal treatment. For import/export experiments after hormone treatment and withdrawal, hormone-treated cells were rinsed at 0, 30, 60, and 120 min (twice at each time point with 3 ml of the culture medium containing the charcoal-stripped serum) before the next round of the translocation process was initiated. To prevent de novo protein synthesis, the replacement medium was supplemented with cycloheximide (50 g/ml). Cycloheximide-mediated inhibition of GFP-AR synthesis in transfected cells was monitored by Western blot analysis with monoclonal antibody to GFP (primary antibody) and horseradish peroxidase-conjugated antimouse IgG. Peroxidase signal was visualized with ECL plus reagent according to the manufacturer’s recommendation (Amersham Pharmacia Biotech, Arlington, IL). Leptomycin B (a gift from Dr. M. Yoshida, University of Tokyo, Tokyo, Japan) was used at a final concentration of 15 ng/ml. Isolation of cell nuclei and nuclear fractionation were performed according to the modified procedure described by Htun et al. (53). Briefly, freshly harvested cells were suspended in TNM buffer (10 mM Tris, pH 8.0, 300 mM sucrose, 100 mM NaCl, 2 mM MgCl2, 1% dithioglycol, and 1 mM phenylmethylsulfonyl fluoride) and after homogenization cells were treated with 0.5% Triton X-100. Isolated nuclei were washed in TNM buffer and resuspended in DIG buffer (TNM buffer with 50 mM NaCl and 3 mM MgCl2) and digested with 168 /ml DNase I. Digested nuclei were first fractionated with 0.25 M ammonium sulfate (fraction 1), and the pellet was subjected to two sequential extractions with 2 M NaCl (combined supernatants were used as fraction 2), and the remaining pellet containing the nuclear matrix was used as the fraction 3. After freeze-drying, all three fractions were dissolved in 100 l Laemli’s SDS buffer for Western blot analysis according to the procedure described above. Fluorescence Imaging Fluorescence imaging of live cells was performed through a E400 Eclipse epifluorescence microscope and water immersion objectives (Nikon, Melville, NY) connected to a video monitor through a cooled charge-coupled device camera (Princeton Instruments, Princeton, NJ; Roper Scientific Inc., Trenton, NJ). The microscope contains a temperaturecontrolled stage, a stepper motor for optical sectioning, and an automatic filter wheel for appropriate filter selection. Optical sectioning was performed at 1-m steps and images were reconstructed through Metamorph software (Universal Imaging Corp., West Chester, PA). For immunodetection, cells were cultured in two chambered glass slides and fixed with 100% methanol at ⫺20 C for 10 min. Fixed cells were rinsed twice with PBS (10 min each wash) and air dried. Before antibody treatment, slides
Vol 14 No. 8
were incubated in a humid chamber for 30 min, followed by overnight incubation (at 4 C) with polyclonal anti-AR antibody (affinity-purified IgG) produced in the rabbit. The antibody was generated by conjugating the first 20 N-terminal amino acids of AR to keyhole hemocyanin. After removal of the anti-AR antibody by rinsing three times with PBS, cells were treated for 1 h with sheep antirabbit IgG antibody conjugated with the fluorescent dye CY3. This step was followed by rinsing and mounting with coverslips and observation under the fluorescence microscope. Dissociation Rate of AR Binding of 3H-DHT and its rate of dissociation were assayed according to Zhou et al. (55). Briefly, COS1 cells transfected with GFP-AR expression plasmid for 48 h were subsequently placed in a serum-free medium containing 5 nM 3H-DHT in the presence or absence of a 100-fold molar excess of unlabeled DHT and incubated for 2 additional hours. Cells were then washed twice in PBS and incubated further in the serum-free medium containing 50 M unlabeled DHT for various time periods. Cell samples collected at different time intervals were washed twice with PBS, dissolved in Tris-SDS-glycerol buffer, and counted for radioactivity.
Acknowledgments We thank Gilbert Torralva for dedicated technical assistance and Lita Chambers for secretarial help.
Received December 31, 1999. Revision received April 4, 2000. Accepted April 24, 2000. Address requests for reprints to: Arun K. Roy, Ph.D., Department of Cellular & Structural Biology, University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, Texas 78284-7762. E-mail:
[email protected]. B.C. is a career scientist with the Department of Veterans Affairs. S.C.A. was partially supported by a fellowship from the Korean Science and Engineering Foundation. This work was supported by NIH Grants DK-14744 and AG-10486.
REFERENCES 1. Gorski J, Toft D, Shyamala G, Smith D, Notides A 1968 Hormone receptors: studies on the interaction of estrogen with the uterus. Recent Prog Horm Res 24:45–80 2. Jensen EV, Suzuki T, Kawashima T, Stumpf WE, Jungblut PW, DeSombre ER 1968 A two-step mechanism for the interaction of estradiol with rat uterus. Proc Natl Acad Sci USA 59:632–638 3. Smith AA, McLean WS, Hansson V, Nayfeh SN, French FS 1975 Androgen receptor in nuclei of rat testis. Steroids 25:569–586 4. Welshons WV, Lieberman ME, Gorski J 1984 Nuclear localization of unoccupied oestrogen receptors. Nature 307:747–749 5. Perrot-Applanat M, Logeat F, Groyer-Picard MT, Milgrom E 1985 Immunocytochemical study of mammalian progesterone receptor using monoclonal antibodies. Endocrinology 116:1473–1484 6. Ennis BW, Stumpf WE, Gasc JM, Baulieu EE 1986 Nuclear localization of progesterone receptor before and after exposure to progestin at low and high temperatures: autoradiographic and immunohistochemical studies of chick oviduct. Endocrinology 119:2066–2075 7. Liao SS, Kokontis J, Sai T, Hiipakka RA 1989 Androgen receptors: structures, mutations, antibodies and cellular dynamics. J Steroid Biochem 34:41–51
Intracellular Movement of AR
8. McGimsey WC, Cidlowski JA, Stumpf WE, Sar M 1991 Immunocytochemical localization of the glucocorticoid receptor in rat brain, pituitary, liver, and thymus with two new polyclonal antipeptide antibodies. Endocrinology 129:3064–3072 9. Simental JA, Sar M, Lane MV, French FS, Wilson EM 1991 Transcriptional activation and nuclear targeting signals of the human androgen receptor. J Biol Chem 266: 510–518 10. Jenster G, van der Korput HA, van Vroonhoven C, van der Kwast TH, Trapman J, Brinkmann AO 1991 Domains of the human androgen receptor involved in steroid binding, transcriptional activation, and subcellular localization. Mol Endocrinol 5:1396–1404 11. Jenster G, Trapman J, Brinkmann AO 1993 Nuclear import of the human androgen receptor. Biochem J 293: 761–768 12. Liang T, Hoyer S, Yu R, Soltani K, Lorincz AL, Hiipakka RA, Liao S 1993 Immunocytochemical localization of androgen receptors in human skin using monoclonal antibodies against the androgen receptor. J Invest Dermatol 100:663–666 13. Jewell CM, Webster JC, Burnstein KL, Sar M, Bodwell JE, Cidlowski JA 1995 Immunocytochemical analysis of hormone mediated nuclear translocation of wild type and mutant glucocorticoid receptors. J Steroid Biochem Mol Biol 55:135–146 14. Fejes-Toth G, Pearce D, Naray-Fejes-Toth A 1998 Subcellular localization of mineralocorticoid receptors in living cells: effects of receptor agonists and antagonists. Proc Natl Acad Sci USA 95:2973–2978 15. Guiochon-Mantel A, Delabre K, Lescop P, Milgrom E 1996 The Ernst Schering Poster Award. Intracellular traffic of steroid hormone receptors. J Steroid Biochem Mol Biol 56:3–9 16. Kokontis JM, Liao S 1999 Molecular action of androgen in the normal and neoplastic prostate. Vitam Horm 55: 219–307 17. Press MF, Xu SH, Wang JD, Greene GL 1989 Subcellular distribution of estrogen receptor and progesterone receptor with and without specific ligand. Am J Pathol 135:857–864 18. Rizzuto R, Carrington W, Tuft RA 1998 Digital imaging microscopy of living cells. Trends Cell Biol 8:288–292 19. Tsien RY 1998 The green fluorescent protein. Annu Rev Biochem 67:509–544 20. Georget V, Lobaccaro JM, Terouanne B, Mangeat P, Nicolas JC, Sultan C 1997 Trafficking of the androgen receptor in living cells with fused green fluorescent protein-androgen receptor. Mol Cell Endocrinol 129:17–26 21. Yang J, Liu J, DeFranco DB 1997 Subnuclear trafficking of glucocorticoid receptors in vitro: chromatin recycling and nuclear export. J Cell Biol 137:523–538 22. Picard D, Yamamoto KR 1987 Two signals mediate hormone-dependent nuclear localization of the glucocorticoid receptor. EMBO J 6:3333–3340 23. Guiochon-Mantel A, Loosfelt H, Lescop P, Sar S, Atger M, Perrot-Applanat M, Milgrom E 1989 Mechanisms of nuclear localization of the progesterone receptor: evidence for interaction between monomers. Cell 57: 1147–1154 24. Ylikomi T, Bocquel MT, Berry M, Gronemeyer H, Chambon P 1992 Cooperation of proto-signals for nuclear accumulation of estrogen and progesterone receptors. EMBO J 11:3681–3694 25. Cadepond F, Gasc JM, Delahaye F, Jibard N, SchweizerGroyer G, Segard-Maurel I, Evans R, Baulieu EE 1992 Hormonal regulation of the nuclear localization signals of the human glucocorticosteroid receptor. Exp Cell Res 201:99–108
1173
26. Zhou ZX, Sar M, Simental JA, Lane MV, Wilson EM 1994 A ligand-dependent bipartite nuclear targeting signal in the human androgen receptor. Requirement for the DNA-binding domain and modulation by NH2-terminal and carboxylterminal sequences. J Biol Chem 269:13115–13123 27. DeFranco DB 1999 Regulation of steroid receptor subcellular trafficking. Cell Biochem Biophys 30:1–24 28. Wolff B, Sanglier JJ, Wang Y 1997 Leptomycin B is an inhibitor of nuclear export: inhibition of nucleo- cytoplasmic translocation of the human immunodeficiency virus type 1 (HIV-1) Rev protein and Rev-dependent mRNA. Chem Biol 4:139–147 29. Kudo N, Matsumori N, Taoka H, Fujiwara D, Schreiner EP, Wolff B, Yoshida M, Horinouchi S 1999 Leptomycin B inactivates CRM1/exportin 1 by covalent modification at a cysteine residue in the central conserved region. Proc Natl Acad Sci USA 96:9112–9117 30. Harhaj EW, Sun SC 1999 Regulation of RelA subcellular localization by a putative nuclear export signal and p50. Mol Cell Biol 19:7088–7095 31. Yeh S, Miyamoto H, Shima H, Chang C 1998 From estrogen to androgen receptor: a new pathway for sex hormones in prostate. Proc Natl Acad Sci USA 95: 5527–5532 32. Reinikainen P, Palvimo JJ, Janne OA 1996 Effects of mitogens on androgen receptor-mediated transactivation. Endocrinology 137:4351–4357 33. El Tanani MK, Green CD 1997 Two separate mechanisms for ligand-independent activation of the estrogen receptor. Mol Endocrinol 11:928–937 34. Gupta C 1999 Modulation of androgen receptor (AR)mediated transcriptional activity by EGF in the developing mouse reproductive tract primary cells. Mol Cell Endocrinol 152:169–178 35. Ram PT, Kiefer T, Silverman M, Song Y, Brown GM, Hill SM 1998 Estrogen receptor transactivation in MCF-7 breast cancer cells by melatonin and growth factors. Mol Cell Endocrinol 141:53–64 36. Wessely O, Deiner EM, Beug H, von Lindern M 1997 The glucocorticoid receptor is a key regulator of the decision between self-renewal and differentiation in erythroid progenitors. EMBO J 16:267–280 37. Neumann F 1994 The antiandrogen cyproterone acetate: discovery, chemistry, basic pharmacology, clinical use and tool in basic research. Exp Clin Endocrinol 102:1–32 38. Fuhrmann U, Bengtson C, Repenthin G, Schillinger E 1992 Stable transfection of androgen receptor and MMTV-CAT into mammalian cells: inhibition of cat expression by anti-androgens. J Steroid Biochem Mol Biol 42:787–793 39. Chan J, Song CS, Matusik RJ, Chatterjee B, Roy AK 1998 Inhibition of androgen action by dehydroepiandrosterone sulfotransferase transfected in PC-3 prostate cancer cells. Chem Biol Interact 109:267–278 40. Nawaz Z, Lonard DM, Dennis AP, Smith CL, O’Malley BW 1999 Proteasome-dependent degradation of the human estrogen receptor. Proc Natl Acad Sci USA 96: 1858–1862 41. Htun H, Barsony J, Renyi I, Gould DL, Hager GL 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 42. Michigami T, Suga A, Yamazaki M, Shimizu C, Cai G, Okada S, Ozono K 1999 Identification of amino acid sequence in the hinge region of human vitamin D receptor that transfers a cytosolic protein to the nucleus. J Biol Chem 19:33531–33538 43. Guiochon-Mantel A, Lescop P, Christin-Maitre S, Loosfelt H, Perrot-Applanat M, Milgrom E 1991 Nucleocytoplasmic shuttling of the progesterone receptor. EMBO J 10:3851–3859
MOL ENDO · 2000 1174
44. Jans DA, Chan CK, Huebner S 1998 Signals mediating nuclear targeting and their regulation: application in drug delivery. Med Res Rev 18:189–223 45. Koivisto P, Kononen J, Palmberg C, Tammela T, Hyytinen E, Isola J, Trapman J, Cleutjens K, Noordzij A, Visakorpi T, Kallioniemi OP 1997 Androgen receptor gene amplification: a possible molecular mechanism for androgen deprivation therapy failure in prostate cancer. Cancer Res 57:314–319 46. Kokontis JM, Hay N, Liao S 1998 Progression of LNCaP prostate tumor cells during androgen deprivation: hormone-independent growth, repression of proliferation by androgen, and role for p27Kip1 in androgen-induced cell cycle arrest. Mol Endocrinol 12:941–953 47. Ossareh-Nazari B, Bachelerie F, Dargemont C 1997 Evidence for a role of CRM1 in signal-mediated nuclear protein export. Science 278:141–144 48. Fornerod M, Ohno M, Yoshida M, Mattaj IW 1997 CRM1 is an export receptor for leucine-rich nuclear export signals. Cell 90:1051–1060 49. Nakielny S, Dreyfuss G 1999 Transport of proteins and RNAs in and out of the nucleus. Cell 99:677–690 50. Tyagi RK, Amazit L, Lescop P, Milgrom E, GuiochonMantel A 1998 Mechanisms of progesterone receptor
Vol 14 No. 8
51.
52.
53.
54.
55.
export from nuclei: role of nuclear localization signal, nuclear export signal, and ran guanosine triphosphate. Mol Endocrinol 12:1684–1695 Savory JG, Hsu B, Laquian IR, Giffin W, Reich T, Hache RJ, Lefebvre YA 1999 Discrimination between NL1- and NL2-mediated nuclear localization of the glucocorticoid receptor. Mol Cell Biol 19:1025–1037 Liu J, DeFranco DB 2000 Protracted nuclear export of glucocorticoid receptor limits its turnover, does not require the exportin 1/CRM1-directed nuclear export pathway. Mol Endocrinol 14:40–51 Htun H, Holth LT, Walker D, Davie JR, Hager GL 1999 Direct visualization of the human estrogen receptor alpha reveals a role for ligand in the nuclear distribution of the receptor. Mol Biol Cell 10:471–486 Stenoien DL, Mancini MG, Patel K, Allegretto EA, Smith CL, Mancini MA 2000 Subnuclear trafficking of estrogen receptor-␣, steroid receptor coactivator-1. Mol Endocrinol 14:518–534 Zhou ZX, Lane MV, Kemppainen JA, French FS, Wilson EM 1995 Specificity of ligand-dependent androgen receptor stabilization: receptor domain interactions influence ligand dissociation and receptor stability. Mol Endocrinol 9:208–218
