Nuclear Export Signal of Androgen Receptor (NES^AR) Regulation of Androgen Receptor Level in Human Prostate Cell Lines via Ubiquitination and Proteasome-Dependent Degradation

Abstract

Androgen receptor (AR) plays a key role in prostate development and carcinogenesis. Increased expression and/or stability of AR is associated with sensitization of prostate cancer cells to low levels of androgens, leading to castration resistance. Hence, understanding the mechanisms regulating AR protein stability is clinically relevant and may lead to new approaches to prevent and/or treat prostate cancer. Using fluorescence microscopy, Western blot, and pulse chase assay, we showed that nuclear export signal (NES)^AR, a nuclear export signal in the ligand binding domain (LBD) of AR, can significantly enhance the degradation of fusion protein constructs in PC3 prostate cancer cells. The half-life of GFP-NES^AR was less than 3 h, which was 10 times shorter than that of green fluorescent protein (GFP) control. Further analysis showed that NES^AR can signal for polyubiquitination and that degradation of NES^AR-containing fusion proteins can be blocked by proteasome inhibitor MG132. Ubiquitination of GFP-AR or GFP-LBD was suppressed in the presence of dihydrotestosterone, which is known to suppress NES^AR while inducing nuclear localization signal 2 in AR or LBD, suggesting that the export activity of NES^AR is required for NES^AR-mediated polyubiquitination. Treatment with MG132 also induced aggresome formation of NES^AR-containing fusion proteins in perinuclear regions of the transfected PC3 cells, indicating a role for NES^AR in inducing unfolded protein responses. The above observations suggest that NES^AR plays a key role in AR ubiquitination and proteasome-dependent degradation in prostate cancer cells.

Despite recent progress, prostate cancer continues to be a major cause of cancer-related mortality and morbidity in aging males (1, 2). Prostate cancer is initially androgen dependent, and androgen deprivation therapy remains the standard treatment for patients with metastatic prostate cancer. Unfortunately, patients treated with androgen deprivation therapy often relapse with castration-resistant prostate cancer (CRPC) (3, 4). Newly developed therapeutics for CRPC can only prolong the survival of patients for several months (5–9). New approaches are urgently needed for prevention and/or treatment of prostate cancer. Because androgen signaling is involved in all stages of prostate tumorigenesis (10, 11), including initiation, progression, and therapy resistance, understanding the mechanisms of androgen action will have significant implications in prostate cancer prevention and treatment.

Androgen signaling is mediated through the androgen receptor (AR), an androgen-regulated transcription factor that can activate or repress downstream genes in response to androgen binding. AR is a member of the nuclear receptor superfamily, consisting of N-terminal domain (NTD), DNA-binding domain (DBD), hinge region, and ligand-binding domain (LBD) (12, 13). In the absence of androgens, AR is localized to the cytoplasm in androgen-sensitive cells. Upon addition of androgens, AR translocates to the nucleus to transactivate downstream genes (14–16). According to the literature, AR contains two nuclear localization (NL) signals, NL1 in the DBD and hinge region of AR and NL2 in the androgen-bound LBD, and one nuclear export signal NES^AR within the androgen-free LBD (17–20). In the absence of androgens, NES^AR is dominant over NL1, causing cytoplasmic localization of AR (19). In the presence of androgens, ligand binding to LBD abrogates NES^AR activity while inducing NL2. The presence of NL1 and NL2 would cause efficient NL of androgen-bound AR. Nucleocytoplasmic trafficking of AR represents an important step in the regulation of AR activity because AR must be localized to the nucleus to function as a transcription factor. Abnormal constitutive NL of AR is associated with prostate cancer progression to castration resistance (21, 22).

Another mechanism regulating AR activity is at the level of AR protein stability. AR exhibits an increased half-life and is sensitized to low levels of androgens in CRPC cells (21). In fact, overexpression of AR is sufficient to drive androgen-sensitive cells to become castration resistant (11, 23). A moderate AR overexpression can alter the binding dynamics of the receptor to chromatin and chromatin structures in prostate cancer cells (24). Thus, tight regulation of AR protein levels is important for androgen action in prostate cancer cells. Understanding the mechanisms regulating AR stability/degradation is fundamentally important and clinically relevant.

Systematic protein degradation by the ubiquitin-proteasome system plays an important role in the degradation of intracellular proteins and regulates a variety of biological functions, including cell proliferation, differentiation, and stress response (25, 26). Previous studies have reported the involvement of polyubiquitinylation and proteasome in AR degradation and/or function (27, 28). Inhibition of the ubiquitin-proteasome degradation pathway by MG132 has been reported to increase endogenous AR protein levels in HepG2 and LNCaP cells and result in an increase in the polyubiquitylated AR, which is consistent with AR being a target of proteasomal degradation (29). The putative proline-, glutamate-, serine-, and threonine-rich sequence, a signature motif for ubiquitin-proteasome degradation, has been identified in the hinge region of AR (29). To be degraded by the ubiquitin-proteasome system, AR should be recognized by E3 ligases. A recent study demonstrates that Mdm2 is associated with AR and can function as an E3 ligase for the ubiquitin-proteasome degradation of AR (30). Detailed biochemical binding assays using a glutathione-S-transferase pull-down assay demonstrate that Mdm2 interacts with the AR NH₂-terminal domain plus DBD (30). Inhibition of the proteasome suppresses AR transactivation, nuclear translocation, and interaction with coregulators, whereas proteasome subunits enhance AR transactivation in a dose-dependent manner (31). Due to the critical role of the ubiquitin-proteasome system in AR regulation, targeting AR degradation represents a promising approach (32, 33).

Despite the importance and clinical relevance of AR and progress made in the studies of AR degradation, the mechanisms regulating AR stability/degradation remain incompletely understood. Here we investigated the role of NES^AR on protein half-life, polyubiquitination, and unfolded protein response. The observations in this manuscript provide evidence for NES^AR regulation of AR stability/degradation. Our studies also suggest a role for NES^AR in androgen-regulation of AR protein stability and AR protein level.

Materials and Methods

Expression vector construction

The GFP (green fluorescent protein)-AR, GFP-NES^AR [amino acids (aa) 743–817], GFP-N-terminal regions of AR (GFP-NAR, aa 1–665), GFP-NAR-NES^AR, GFP-LBD (aa 666–918), GFP-c2 (aa 666–715), GFP-NES^ER estrogen receptor (ER), and GFP-NES^MR mineralocorticoid receptor (MR) constructs were generated from human AR, ER, and MR, as previously described (19). His-tagged Ubiquitin expression vector was kindly provided by Dr. James R. Gnarra (University of Pittsburgh).

Cell culture and transfection

The human prostate cancer cell lines PC3, LNCaP, and C4–2 were obtained from American Type Culture Collection (Manassas, VA) and maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum, 1% glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen, Carlsbad, CA) at 37 C in the presence of 5% CO₂ in a humidified incubator.

Various constructs were transfected into PC3, LNCaP, or C4–2 cells using PolyJet In Vitro DNA Transfection Reagent (SignaGen Laboratories, Rockville, MD) according to the manufacturer's protocol. Cells were plated 18–24 h before transfection, and the monolayer cell density reached the optimal 70 to approximately 80% confluence at the time of transfection. Cells were transfected in the presence of serum and collected for expression analysis 24 h after transfection. In some experiments, transfected cells were treated with protein synthesis inhibitor cycloheximide (CHX) (Sigma-Aldrich, St. Louis, MO) at 50 μg/ml, proteasome inhibitor MG132 (Sigma-Aldrich) at 5 μm or indicated concentration, and/or dihydrotestosterone (DHT) (Sigma-Aldrich) at indicated concentrations (10 nm or 100 nm) for various times as described in the figure legends.

Western blot analysis and immunoprecipitation

PC3 cells were lysed in modified radioimmune precipitation assay buffer [50 mm Tris (pH 7.4), 1% Nonidet P-40, 0.25% Na-deoxycholate, 150 mm NaCl, 1 mm EDTA (pH 8.0), 1 mm NaF, 2 mm phenylmethylsulfonylfluoride, 1 mm Na₃VO₄, and protease inhibitor cocktail (Sigma-Aldrich)]. Protein concentration was determined by BCA assay (Pierce Chemical Co., Rockford, IL). For immunoprecipitation, cell lysates were immunoprecipitated with agarose-conjugated anti-GFP (D153–8, MBL International, Woburn, MA) at 4 C overnight. Immunoprecipitates were washed three times in radioimmune precipitation assay buffer before being resuspended in loading buffer and subjected to SDS-PAGE. Western blot was conducted using antibodies specific for GFP (TP401; Acris Antibodies, San Diego, CA), Ubiquitin (sc-8017, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and GAPDH (glyceraldehyde 3-phosphate dehydrogenase) (sc-25778, Santa Cruz Biotechnology), followed with horseradish peroxidase-labeled secondary antibody (sc-2004, Santa Cruz Biotechnology). Signals were visualized using chemiluminescence (ECL Western Blotting Detection Reagents; GE Healthcare, Piscataway, NJ) and were exposed to x-ray film (Fuji film, Stanford, CT). GAPDH served as a loading control. All the experiments were repeated at least three times.

Pulse chase assay

PC3 cells were plated in 6-cm dishes and transiently transfected with expression vectors of GFP or GFP-NES^AR. Transfected PC3 cells were washed with PBS and incubated in 2 ml cysteine- and methionine-free RPMI1640 (R7513, Sigma-Aldrich) complete medium for 30 min, then pulse labeled with [³⁵S]methionine/cysteine for 2 h, at which point labeling medium was washed away and replaced with chase medium containing methionine and cysteine for the chase intervals indicated in Fig. 2C.³⁵S-labeled proteins were immunoprecipitated from cell lysates using agarose-conjugated anti-GFP (D153–8, MBL), analyzed by SDS-PAGE and autoradiograph. The experiments were repeated twice.

Fig. 2.

NES^AR-containing fusion proteins exhibited short half-life. A, Effect of protein synthesis inhibitor (CHX) on the protein levels of GFP, GFP-tagged NES^AR, GFP-tagged NAR, and GFP-tagged NAR-NES^AR at different time in PC3 cells. B, The protein level of the expression vector GFP-c2 containing aa 666–717 of AR was examined by Western blot analysis after treatment of CHX at different times in PC3 cells. C, Pulse chase analysis of GFP and GFP-NES^AR in PC3 cells. IgG immunoprecipitation from a pulse-labeled lysate was included as a control. UT represents untransfected PC3 cell lysates. D and E, Effect of protein synthesis inhibitor (CHX) on the protein levels of GFP, GFP-tagged NES^AR, GFP-tagged NAR, and GFP-tagged NAR-NES^AR at different times in LNCaP cells (D) and C4–2 cells (E). Results of pulse chase analysis are representative of two independent experiments, and other results are representative of at least three independent experiments. IP, Immunoprecipitation.

Immunofluorescence

PC3 cells in six-well plates or on glass coverslips were transiently transfected overnight for visualization. MG132 was used at 5 μm for 20 h after transfection. For immunofluorescence staining, cells on glass coverslips were washed with cold PBS, fixed with 2% paraformaldehyde for 15 min, permeabilized with 0.5% Triton X-100 for 15 min, and then incubated in a humidifying chamber with blocking solution (5% BSA in PBS) for 45 min. Next, coverslips were incubated with the anti-histone deacetylase (HDAC)6 antibody (sc-11420, Santa Cruz Biotechnology) in 0.5% BSA at 4 C overnight. After washes, cells were incubated with Cy3-conjugated antirabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) diluted at 1:500 in 0.5% BSA. Coverslips were washed, stained with Hoechst and mounted onto slides using Dako glycergel mounting medium (DAKO Corp., Carpinteria, CA). Photos were taken by fluorescence microscopy using a Nikon TE 2000U inverted microscope (Nikon, Melville, NY). The experiments were repeated at least three times.

Results

NES^AR-containing fusion proteins exhibit markedly decreased stability

We previously reported a novel putative nuclear export signal NES^AR, localized between aa 743 and aa 817 in the LBD region of AR (19). The diagram of expression vectors containing GFP-tagged NES^AR, NH₂-terminal region of AR lacking LBD (NAR) and both NAR plus NES^AR (NAR-NES^AR) are shown in Fig. 1A. Interestingly in transfected PC3 cells, the fluorescence intensities of NES^AR containing constructs, GFP-NES^AR and GFP-NAR-NES^AR, were much weaker than those without NES^AR, GFP and GFP-NAR (Fig. 1B). Western blot analysis confirmed the different expression levels of these constructs (Fig. 1C). The above findings suggested a potential role for NES^AR in protein degradation in addition to its function as a NES. To determine the effect of NES^AR on protein stability, the protein levels of GFP, GFP-tagged NES^AR, GFP-tagged NAR, and GFP-tagged NAR-NES^AR were investigated in the presence of CHX, which inhibits de novo protein synthesis. Figure 2A showed that fusion proteins containing NES^AR, GFP-tagged NES^AR, and GFP-tagged NAR-NES^AR were not stable and degraded in less than 6 h in the presence of CHX. In contrast, GFP and GFP-tagged NAR remained abundant 24 h after the addition of CHX. As a control, we also tested the stability of GFP-tagged segment outside of NES^AR but within LBD domain of AR (aa 666–715) named as GFP-c2. As shown in Fig. 2B, GFP-NES^AR protein level was significantly reduced within 1 h after CHX treatment; however, the protein level of GFP-c2 remained high 12 h after CHX treatment. To further determine the effect of NES^AR on protein stability, we performed a pulse-chase labeling assay and showed that the half-life of GFP-NES^AR was less than 3 h in transfected cells whereas the GFP control exhibited a half-life longer than 32 h (Fig. 2C). Furthermore, we check the stability of GFP-tagged NES^AR in androgen-sensitive LNCaP cells and castration-resistant C4–2 cells. GFP-NES^AR degraded by more than half within 6 h in both LNCaP and C4–2 cells as indicated by Western blot analysis (Fig. 2, D and E). These observations argued for an important role of NES^AR in regulating protein stability and suggested NES^AR as a novel degron in prostate cancer cells.

Fig. 1.

NES^AR fusion proteins were expressed at low levels in PC3 cells. A, Diagrams of enhanced GFP (EGFP)-tagged fusion proteins. B, Fluorescent intensity of GFP, GFP-tagged NES^AR, GFP-tagged NAR, and GFP-tagged NAR-NES^AR fusion proteins 24 h after transfection in PC3 cells in complete RPMI 1640 medium. Exposure time for all images was 1 sec. C, Western blot analysis of protein levels of transiently transfected GFP, GFP-tagged NES^AR, GFP-tagged NAR, and GFP-tagged NAR-NES^AR fusion proteins in PC3 cells in complete RPMI 1640 medium. Protein level of GAPDH was detected as loading control. Results are representative of at least three independent experiments.

Proteasome inhibitor MG132 blocks degradation of NES^AR-containing fusion proteins

Because proteasome plays a key role in protein degradation, we investigated whether NES^AR-mediated protein degradation was proteasome dependent. As shown in Fig. 3A, the expression levels of various fusion proteins in transfected PC3 cells were examined in the absence or presence of proteasome inhibitor MG132 at different doses. As expected, the expression levels of NES^AR containing fusion proteins, GFP-tagged NES^AR and GFP-tagged NAR-NES^AR, were significantly increased by MG132 treatment in a dose-dependent manner; in contrast, MG132 did not show any significant effect on the protein levels of GFP and GFP-tagged NAR (Fig. 3A). Next, to rule out the possibility that increased protein levels after treatment of MG132 were due to de novo synthesis, we tested whether MG132 could block the degradation of NES^AR fusion proteins when new protein synthesis was blocked. As shown in Fig. 3B, MG132 indeed prevented the degradation of GFP-tagged NES^AR and GFP-tagged NAR-NES^AR in transfected cells in the presence of CHX. In contrast, MG132 had no significant effect on GFP or GFP-tagged NAR protein stability in parallel experiments. These findings indicated that the proteolytic degradation of NES^AR was mediated through the proteasome-dependent protein degradation pathway.

Fig. 3.

Proteasome inhibitor MG132 inhibited degradation of NES^AR-containing fusion proteins. A and B, Effect of MG132 on the expression levels of GFP, GFP-NES^AR, GFP-NAR, and GFP-NAR-NES^AR was shown by Western blot analysis in the absence (A) or presence (B) of CHX at indicated times of treatment. Results are representative of at least three independent experiments.

GFP-NES^AR protein exhibits elevated ubiquitination

Many proteins are targeted for degradation by proteasome via ubiquitination. To be recognized by the proteasome, a target protein needs to be linked to a chain of the small protein ubiquitin. Because proteasome inhibition blocked degradation of NES^AR-containing proteins (Fig. 3), ubiquitination would likely be detectable in NES^AR-containing fusion proteins after the treatment of MG132. To determine whether NES^AR was the target of ubiquitination, PC3 cells were cotransfected with His-tagged ubiquitin (His-Ub) and GFP or GFP-NES^AR and followed by treatment with MG132 or vehicle. Immunoprecipitation with anti-GFP antibody-conjugated beads followed by Western blot analysis using antiubiquitin antibody revealed that the ubiquitination of GFP-NES^AR fusion protein was significantly increased compared with GFP control (Fig. 4). This finding suggested that NES^AR could be a major target in AR for ubiquitination and mediate proteasome-dependent degradation.

Fig. 4.

NES^AR was the target of ubiquitination. Ubiquitination level (upper panel) and immunoprecipitated protein levels (lower panel) of GFP and GFP-tagged NES^AR pulldown by agarose conjugated with anti-GFP in the lysates of PC3 cells were shown by immunoprecipitation and Western blots. HC, Heavy chain; LC, light chain. Results are representative of at least three independent experiments. IB, Immunoblotting; IP, immunoprecipitation.

DHT inhibits polyubiquitination of GFP-LBD and GFP-AR

Previous studies suggested that DHT binding to LBD of AR inactivates NES^AR while inducing NL2 in the LBD region (19). We further tested whether inactivation of NES^AR by hormone binding could suppress NES^AR-mediated ubiquitination of LBD and AR. As shown in Fig. 5A, ubiquitination of GFP-tagged AR and LBD were able to be detected in the presence of MG132, and treatment of DHT at 100 nm dramatically inhibited polyubiquitination of GFP-tagged AR and LBD in the presence of MG132. In contrast, treatment of DHT at 10 nm significantly inhibited the ubiquitination of GFP-tagged AR but not GFP-tagged LBD (Fig. 5B). This was expected because LBD domain alone is less sensitive to DHT than the full-length AR, and treatment with DHT at 10 nm is sufficient to inhibit NES^AR in full-length AR but not in LBD alone. Also, GFP-tagged NAR was not or very weakly polyubiquitinated even in the presence of MG132 (Fig. 5B) and did not respond to androgen treatment, which was consistent with its stability as shown in Fig. 3. This result was in agreement with our expectation that NES^AR in DHT-bound LBD or AR was not functional and failed to mediate efficient ubiquitination.

Fig. 5.

DHT inhibits polyubiquitination of GFP-tagged LBD and GFP-tagged AR. A and B, Ubiquitination levels (upper panel) and immunoprecipitated protein levels (lower panel) of GFP-tagged AR, GFP-tagged LBD, and GFP-tagged NAR pulldown by anti-GFP-conjugated agarose in the lysates of PC3 cells with various treatment were shown by immunoprecipitation and Western blots. DHT was added at 100 nm (A) or 10 nm (B), and MG132 was added at 5 μm. Results are representative of at least three independent experiments. HC, Heavy chain; IB, immunoblotting; IP, immunoprecipitation; LC, light chain.

Proteasome inhibitor MG132 induces aggresome formation of NES^AR-containing fusion proteins in transfected cells

Abnormal polypeptides that escape proteasome-dependent degradation and aggregate in cytosol can be transported via microtubules to aggresomes, the organelles that serve as a storage compartment for protein aggregates and can be actively involved in their refolding and degradation (34). Aggresome represents a protective cellular response to a buildup of aggregating abnormal polypeptides under the conditions when chaperones and ubiquitin-proteasome system machineries fail to handle abnormal species, allowing for sequestration of potentially toxic misfolded proteins and promoting their clearance by autophagy (35–37). Indeed, aggresomes are usually seen in mammalian cells after inhibition of the proteasome (38–40). Recently, a number of factors were implicated in aggresome formation. HDAC6, a microtubule-associated deacetylase, was shown to interact with aggregates of ubiquitinated proteins via its ubiquitin binding BUZ domain. HDAC6 facilitates the association of ubiquitinated proteins with the dynein motor protein that drives this cargo to the aggresome (41). HDAC6 is a component and is required for aggresome formation and also plays an essential role in aggresomal protein degradation (41).

We examined the effects of proteasome inhibitor MG132 on aggresome formation of NES^AR-containing GFP fusion proteins along with controls. The presence of perinuclear inclusions was considered indicative of aggresome formation (42), and HDAC6 served as a marker (43). Aggresome formation of all NES^AR-containing proteins (GFP-tagged AR, GFP-tagged NES^AR, and GFP-tagged NAR-NES^AR) was induced by MG132 in the perinuclear regions of PC3 cells and were colocalized with aggresome marker HDAC6 (Fig. 6). In contrast, MG132 did not induce protein aggregation of GFP or GFP-tagged NAR, both of which were independent of NES^AR (Fig. 6). In conclusion, NES^AR-containing proteins were subject to degradation, which is dependent on ubiquitin-proteasome system in the normal condition; however, inhibition of proteasome induced the aggregation of these undegradated proteins in aggresomes.

Fig. 6.

Proteasome inhibitor MG132 induced NES^AR-containing fusion proteins to form putative aggresome in perinuclear regions in transfected PC3 cells. Cells were treated with MG132 at 5 μm for 20 h. GFP indicated a variety of GFP fusion proteins as green signal. HDAC6, as aggresome marker, was labeled with Cy3 as red signal. Nuclei were stained with Hoechst as blue signal. Solid arrows, Colocalization of GFP signal and HDAC6 staining as aggresomes; open arrows, aggresomes indicated by HDAC6 staining but not GFP signal. Results are representative of at least three independent experiments. DMSO, Dimethylsulfoxide.

NES of ER (NES^ER) and MR (NES^MR) are unstable and ubiquitylated

We have reported that NES is functionally conserved in the steroid receptor superfamily, and both GFP-tagged NES^ER and NES^MR exhibited cytoplasmic localization (19). In this study, we investigated the degradation of GFP-tagged NES^ER and GFP-tagged NES^MR fusion proteins in the presence of CHX. GFP-tagged NES^ER and GFP-tagged NES^MR were not stable and degraded in less than 6 h (Fig. 7A). Immunoprecipitation showed that as NES^AR, both NES^ER and NES^MR were targets of ubiquitination (Fig. 7B). These results suggest that this conserved region in the nuclear hormone receptor family mediates ubiquitin-proteasome system-dependent degradation.

Fig. 7.

GFP-tagged NES^ER and GFP-tagged NES^MR were unstable and ubiquitylated. A, Effect of CHX on the expression levels of GFP-tagged NES^ER and GFP-tagged NES^MR fusion proteins in PC3 cells was shown by Western blot analysis. B, NES^ER and NES^MR were target signals for ubiquitination. Ubiquitinated protein levels (upper panel) and immunoprecipitated protein levels (lower panel) of GFP-tagged NES^ER and GFP-tagged NES^MR pulldown by anti-GFP-conjugated agarose in the lysates of PC3 cells were shown by immunoprecipitation and Western blots. HC, Heavy chain; LC, light chain. Results are representative of at least three independent experiments. IB, Immunoblotting; IP, immunoprecipitation.

Discussion

Understanding the mechanisms regulating AR protein levels in prostate cancer cells is fundamentally important and clinically relevant. Elevated AR protein level is associated with prostate cancer progression to castration resistance (11, 44). In addition to transcriptional regulation, AR expression can also be regulated at the level of protein degradation/stability (45). Previous studies demonstrated that AR protein level and activity can be regulated by the ubiquitin-proteasome degradation pathway. There is a putative proline-, glutamate-, serine-, and threonine-rich sequence, which signals for ubiquitin-proteasome degradation, in the hinge region of AR (29). In the present study, we have generated evidence for NES^AR as another signal for ubiquitin-proteasome degradation and provide new insights into androgen regulation of AR protein stability.

NES^AR is different from any known motifs or signals that regulate ubiquitination and/or protein stability. Protein sequence analysis did not identify any recognizable ubiquitination signal within NES^AR. Thus, NES^AR is a novel signal capable of regulating protein stability. NES^AR is potent in the induction of protein degradation. Fusion protein constructs containing NES^AR were expressed at much lower abundance relative to the control constructs lacking NES^AR. The half-life of GFP or GFP-NAR, which lacks NES^AR, was more than 10-fold longer than GFP-NES^AR or GFP-NAR-NES^AR, which contains NES^AR (Fig. 2A). Thus, NES^AR is likely to play a major role in regulating AR stability.

NES^AR appears to mediate androgen regulation of AR proteins stability. Androgens have been reported to stabilize AR protein in various cells (46). The half-life of AR in LNCaP cells is approximately 3 h in the absence of androgen and is increased to more than 10 h in the presence of 10 nm DHT (21). NES^AR within LBD or AR is inactivated when LBD or AR is bound to androgen (19). Thus, DHT inhibition of polyubiquitination of LBD or AR (Fig. 5) suggests that active NES^AR is required for its effective polyubiquitination and subsequent degradation. Based on the above observations, an important mechanism of androgen stabilization of AR involves the inactivation of NES^AR within androgen-bound AR and subsequent inhibition of NES^AR-dependent ubiquitination, leading to stabilization of androgen-bound AR.

Our studies suggest that polyubiquitination represents an important step in mediating NES^AR regulation of protein stability. However, enzymes involved in the multistep ubiquitination process of NES^AR remain unclear because multiple enzymes exist for each step involved in ubiquitination. Also, the amino acid residue(s) within NES^AR capable of mediating ubiquitination have not been defined. Future studies will be needed to address these important questions. Another interesting question is whether NES^AR export function is associated with its role in mediating polyubiquitination and regulation of AR protein stability. Inhibition of NES^AR export activity by androgen binding in LBD or full-length AR inhibited polyubiquitination and increased protein stability (Fig. 5), suggesting a tight association of NES^AR export function with its activity as a degron. Defining the cellular factors mediating NES^AR export and/or regulating NES^AR-mediated protein degradation will be required to define the potential association between export and degradation directed by NES^AR.

Aggresome formation by NES^AR-containing fusion proteins in the presence of proteasome inhibitor MG132 indicated the capability of NES^AR to trigger unfolded protein responses when proteasome-mediated degradation was inhibited (Fig. 6). Thus, NES^AR is likely to play an important role in the removal of excessive unliganded AR proteins. NES^AR in full-length AR or LBD is active only in the absence of androgens (19). As expected, proteasome inhibitor MG132 induced aggresome formation of GFP-AR in the absence of androgens (Fig. 6). This is consistent with the expectation that androgen-bound AR or LBD is in a folded conformation and does not trigger unfolded protein response.

The protection of full-length AR, but not the LBD, from ubiquitination by 10 nm DHT (Fig. 5B) suggests that the NTD could modulate NES^AR ubiquitination. According to the literature, the interaction between the NTD and LBD of AR enhances the retention of receptor-bound androgen (47). In the absence of NTD, a high rate of DHT dissociation from the LBD is likely to activate the NES^AR, subsequently facilitating NES^AR ubiquitination. However, we cannot rule out the possibility that the androgen-induced interaction between the NTD and LBD can protect the NES^AR from ubiquitination via a different mechanism.

The NESs of other steroid receptors are also likely to play important roles in regulating ubiquitination and degradation. Both GFP-NES^ER and GFP-NES^MR can signal for polyubiquitination and were unstable with a half-life less than 6 h (Fig. 7). This finding further argues that functions of NES are conserved in the steroid receptor superfamily.

Recent studies suggest that CRPC cells can synthesize androgens, which could inhibit NES^AR-mediated AR degradation. One approach to inhibit androgen synthesis in prostate cancer cells is to treat patients with abiraterone, which blocks key steps involved in androgen synthesis (5, 6). Abiraterone is effective in suppressing testosterone levels in patients with CRPC, and therefore it could facilitate AR degradation. It is possible that NES^AR-mediated AR degradation is also regulated by other mechanisms, in addition to DHT binding to LBD. For example, the interaction between NTD and LBD of AR could affect NES^AR degradation. Also, cellular factors involved in NES^AR degradation may be altered in prostate cancer progression, influencing AR stability in prostate cancer cells. Identification and characterization of these cellular factors may lead to new approaches targeting AR degradation pathway in CRPC.

The observations reported here provide new mechanistic insights into androgen regulation of AR protein level and revealed additional functions of LBD. Previous studies have documented multiple functions of LBD, including transactivation, nuclear import, and nuclear export, and regulation of these important functions by androgens. The present study argues that LBD also contains an androgen-regulated degron at the NES^AR region. Induction of NES^AR-mediated AR degradation could effectively block AR action in both androgen-sensitive and castration-resistant prostate cancer cells. Because AR is the key determinant of prostate cancer progression to castration resistance, targeting NES^AR-mediated AR degradation may represent one of the most effective ways to treat patients with prostate cancer, particularly castration-resistant prostate cancer.