S

Tl

JLa

b

c

d

a

ARRAA

1

spcsr(cfloai

ST

h0

ARTICLE IN PRESSG ModelBMB-4170; No. of Pages 8

Journal of Steroid Biochemistry & Molecular Biology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Journal of Steroid Biochemistry and Molecular Biology

j ourna l h omepage: www.elsev ier .com/ locate / j sbmb

he N-terminal domain of the androgen receptor drives its nuclearocalization in castration-resistant prostate cancer cells

avid A. Dara,d, Khalid Z. Masoodia, Kurtis Eisermanna, Sudhir Isharwala, Junkui Aia,aura E. Pascala, Joel B. Nelsona,c, Zhou Wanga,b,c,∗

Department of Urology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15232, United StatesDepartment of Pharmacology and Chemical Biology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15260, United StatesUniversity of Pittsburgh Cancer Institute, University of Pittsburgh School of Medicine, Pittsburgh, PA 15232, United StatesCentral Laboratory College of Science, King Saud University, Riyadh KSA-11451, Saudi Arabia

r t i c l e i n f o

rticle history:eceived 25 November 2013eceived in revised form 14 February 2014ccepted 14 March 2014vailable online xxx

a b s t r a c t

Androgen-independent nuclear localization is required for androgen receptor (AR) transactivation incastration-resistant prostate cancer (CRPC) and should be a key step leading to castration resistance.However, mechanism(s) leading to androgen-independent AR nuclear localization are poorly under-stood. Since the N-terminal domain (NTD) of AR plays a role in transactivation under androgen-depletedconditions, we investigated the role of the NTD in AR nuclear localization in CRPC. Deletion mutagenesiswas used to identify amino acid sequences in the NTD essential for its androgen-independent nuclearlocalization in C4-2, a widely used CRPC cell line. Deletion mutants of AR tagged with green fluores-cent protein (GFP) at the 5′-end were generated and their signal distribution was investigated in C4-2cells by fluorescent microscopy. Our results showed that the region of a.a. 294–556 was required forandrogen-independent AR nuclear localization whereas a.a. 1–293 mediates Hsp90 regulation of AR

nuclear localization in CRPC cells. Although the region of a.a. 294–556 does not contain a nuclear importsignal, it was able to enhance DHT-induced import of the ligand binding domain (LBD). Also, transac-tivation of the NTD could be uncoupled from its modulation of AR nuclear localization in C4-2 cells.These observations suggest an important role of the NTD in AR intracellular trafficking and androgen-independent AR nuclear localization in CRPC cells.

. Introduction

Androgens play a vital role in the development and homeo-tasis of male sex organs [1] as well as the development androgression of benign prostatic hyperplasia (BPH) and prostateancer (PCa) [2–5]. Androgen-deprivation therapy (ADT) is thetandard for treating metastatic PCa, however, patients invariablyecur with more aggressive castration-resistant prostate cancerCRPC) [4,6,7]. During progression to castration resistance, PCaells utilize a variety of cellular pathways in order to survive andourish in an androgen-depleted environment [6,7]. High levels

Please cite this article in press as: J.A. Dar, et al., The N-terminal docastration-resistant prostate cancer cells, J. Steroid Biochem. Mol. Bio

f androgen receptor (AR) expression and renewed expression ofndrogen-regulated genes indicate that AR transcriptional activitys reactivated in CRPC under castration conditions [8]. However, the

∗ Corresponding author at: Department of Urology Shadyside Medical Center,uite G40, 5200 Centre Avenue Pittsburgh, PA 15232, United States.el.: +1 412 623 3903; fax: +1 412 623 3904.

E-mail address: wangz2@upmc.edu (Z. Wang).

ttp://dx.doi.org/10.1016/j.jsbmb.2014.03.004960-0760/© 2014 Elsevier Ltd. All rights reserved.

© 2014 Elsevier Ltd. All rights reserved.

mechanisms leading to AR activation in CRPC remain incompletelyunderstood.

The human AR is a 110 kDa, 919 amino acid protein composed offour domains: (1) the amino terminal activation domain (NTD), (2)the DNA-binding domain (DBD), (3) the hinge region and (4) thecarboxyl ligand-binding domain (LBD) [9]. The NTD (a.a. 1–556)includes the majority of the AR and is the least conserved allow-ing AR to differentially recruit co-regulators conferring androgenspecific transactivation. Proper activation of the AR requires thefirst 30 amino acids of the NTD for the amino-carboxyl terminal(N/C) interaction [10–15]. AR transactivational activity is primarilymediated through the NTD region containing the activation func-tion 1 (AF1) element, which distinguishes AR from the other steroidreceptors that utilize the AF2 region in the LBD [16]. In addition tothe NTD region, two nuclear localization signals have been reportedin the AR. A bipartite nuclear localization signal (NLS1) is present

main of the androgen receptor drives its nuclear localization inl. (2014), http://dx.doi.org/10.1016/j.jsbmb.2014.03.004

in the DBDH region [17,18] and the LBD (a.a. 666–919) contains asecond nuclear localization signal (NLS2) upon androgen binding.Additionally, the LBD contains a nuclear export signal (NES), whichfunctions in the absence of androgens [19].

ARTICLE IN PRESSG ModelSBMB-4170; No. of Pages 8

2 J.A. Dar et al. / Journal of Steroid Biochemistry & Molecular Biology xxx (2014) xxx–xxx

F s upoG -AR, Gs (-DH1 erime

tiItctrgoaiaat

taammfAefimEcda

iduw

ig. 1. Deletion mutagenesis analysis of AR nuclear import and export in COS1 cellFP-DBDH-LBD, and GFP-LBD. (B) COS1 cells were transiently transfected with GFPignal distribution was visualized by fluorescent microscopy 16 h post transfection00 nM DHT for GFP-LBD, or 12 h after androgen withdrawal (Withdrawal). The exp

A key regulatory step in the action of AR is its translocation tohe nucleus. Intracellular trafficking is an important mechanismn the regulation of transcription factors, including AR [20–23].n order for AR to act as a transcription factor it must gain entryo the nucleus. In the prostate, androgens bind to AR in theytoplasm, causing phosphorylation, dimerization, and subsequentranslocation into the nucleus, thereby binding to the androgen-esponse elements within the DNA, with subsequent activation ofenes involved in cell growth and survival. During the progressionf prostate cancer to castration-resistance, the tightly regulatedndrogen signaling pathway is disrupted such that AR can local-ze to the nucleus and activate its target genes in the absence ofndrogens. Our recent studies suggest that Hsp90 is required forndrogen-independent AR nuclear localization in CRPC, althoughhe mechanism involved remains unclear [24,25].

Many studies have revealed various types of AR gene muta-ions that contribute to diseases including spinobulbar musculartrophy (SBMA) [26], androgen insensitivity syndrome (AIS) [27],nd prostate cancer [28,29]. Several reports have revealed thatutant ARs obtained from both AIS and prostate cancer patientsay exhibit abnormal intracellular localization and lower capacity

or ligand-dependent translocation when compared with wild-typeR. These AR mutants fail to achieve normal nuclear import andxhibit a distinct intracellular aggregation profile [18,30–33]. Thesendings suggest that abnormal intracellular localization of the ARutants could be involved in the pathogenesis of these diseases.

lucidation of the nuclear import mechanism of AR will not onlyontribute to an understanding of androgen influence in AR-relatediseases but also provide potential targets to modulate androgenction.

In order to elucidate the mechanism(s) regulating androgen-

Please cite this article in press as: J.A. Dar, et al., The N-terminal docastration-resistant prostate cancer cells, J. Steroid Biochem. Mol. Bio

ndependent AR nuclear localization in CRPC, the effect of AReletion mutagenesis on intracellular localization was testedsing fluorescent microscopy [18]. In this report, C4-2 cellshich were generated through multiple stages of co-culture of

n androgen manipulation. (A) Structures of GFP fusion protein constructs: GFP-AR,FP-DBDH-LBD and GFP-LBD cultured in androgen-free conditions. The subcellular

T), 4 h after treatment (+DHT) with 1.0 nM DHT for GFP-AR and GFP-DBDH-LBD ornt was reproduced 5 times.

androgen-sensitive LNCaP prostate cancer cells with human bonefibroblast MS cells in vivo in castrated male athymic mice [34],were used as a model for CRPC. This study demonstrated animportant role of the NTD in regulating AR intracellular traffickingand its androgen-independent nuclear localization in CRPC.

2. Methods

2.1. Expression vector construction

The expression vector pEGFP-C1 (Clontech) was used togenerate fusion protein constructs with GFP at the N ter-minus of AR and various AR mutants for convenient visu-alization by fluorescent microscopy. Fusion protein constructsgenerated were: pEGFP-AR, pEGFP-LBDAR(666–918), pEGFP-DBDH-LBDAR(557-919), GFP-AR�(294-556), GFP-AR�(1-293), GFP-NTD, GFP-AR(294-556), GFP-AR(1-293)-LBD, GFP-AR(294-556)-LBD, and GFP-NTD-LBD. Various AR mutants were generated by anchor PCR using thefull-length human AR cDNA as the template, which was kindly pro-vided by Dr. Shutsung Liao (University of Chicago). All expressionvectors were verified by sequencing analysis, and plasmids weredouble CsCl gradient purified for transient transfection.

2.2. Cell culture and transfection

COS-1 and LNCaP cell lines were purchased from American TypeCulture Collection (Manassas, VA) and LAPC4 cells were obtainedfrom Dr. Robert Reiter (UCLA). C4-2 cells were obtained from Dr.Leland Chung (Cedars Sinai Medical Center). All cell lines weremaintained in RPMI 1640 medium supplemented with 10% fetalbovine serum (FBS), 1% glutamine, 100 units/ml penicillin, and

main of the androgen receptor drives its nuclear localization inl. (2014), http://dx.doi.org/10.1016/j.jsbmb.2014.03.004

100 �g/ml streptomycin (Invitrogen) at 37 ◦C in the presence of 5%CO2 in a humidified incubator. For androgen-free conditions, RPMI1640 medium was supplemented with FBS stripped 2 times withcharcoal to remove all androgens.

IN PRESSG ModelS

istry & Molecular Biology xxx (2014) xxx–xxx 3

afcpwN

Atiicsua

cgdcwv3

tt(tw

2

3cN(w(C

3

3

pwTawotcwnnGp(oN

Fig. 2. Subcellular distribution of GFP-AR, GFP-DBDH-LBD and GFP-LBD in C4-2cells. C4-2 cells were transiently transfected with GFP-AR, GFP-DBDH-LBD and GFP-LBD and stained with Hoechst. The subcellular localization (A) and quantification(B) were assessed in androgen-free conditions by fluorescent microscopy 16 h aftertransfection. The results are from five transfections for each experimental group. Atleast 200 cells were counted for each transfection to determine the percentage ofcells displaying cytoplasmic (C), even (E), or nuclear (N) localization. The experiment

ARTICLEBMB-4170; No. of Pages 8

J.A. Dar et al. / Journal of Steroid Biochem

AR deletion mutants were transfected into COS-1, LAPC4, LNCaPnd C4-2 cells using LipofectamineTM according to the manu-acturer’s protocol (Invitrogen). Cells were transfected at >60%onfluence in phenol red-free OptiMEM. Localization of GFP fusionroteins was imaged 16 h after transfection or at indicated timeith fluorescence microscopy using either a Nikon TE 2000U orikon TS100 inverted microscope.

For experiments involving nuclear import and export of GFP-R fusion proteins, transfected cells were allowed to express

he appropriate chimeric protein for 16 h prior to treatment ofndicated amount of DHT for 4 h in complete medium prior to imag-ng. For testing the effect of androgen withdrawal, DHT-treatedells were rinsed twice with medium containing charcoal-strippederum at 0, 30, 60, and 120 min for complete removal of any resid-al DHT present in the media. Cells were imaged at 12 h after initialndrogen withdrawal.

Cytoplasmic localization of GFP fusion proteins in transfectedells was defined when the cytoplasmic GFP fluorescence wasreater than that in the nuclei. Likewise, nuclear localization wasefined when nuclear GFP fluorescence was greater than that in theytoplasm. Even distribution was defined when GFP fluorescenceas evenly distributed between the nucleus and cytoplasm as pre-

iously described [19]. Nuclei were stained with 1 mM/L Hoechst3342 solution (Sigma–Aldrich).

To determine the effect of Hsp90 inhibition, C4-2 cells wereransfected with GFP-tagged AR or AR deletion mutants and thenreated with 300 nM 17-AAG (17-AAG was a generous gift of the NCIRockville, MD)) 16 h after transfection. The subcellular localiza-ion was determined by fluorescent microscopy 4 h after treatmentith17-AAG.

.3. TIF2 knockdown and western blot analysis

C4-2 cells were transfected with control siRNA or siTIF2 (sc-8882, Santa Cruz Biotechnology) for 72 h. Cells were lysed in bufferontaining 10 mM NaPO4, 1% Triton-X-100, 1 mM EDTA, 150 mMaCl, 2 mM PMSF, 1 mM Na3VO4, and protease inhibitor cocktail

Sigma). Lysates were subjected to SDS-PAGE and Western blottingas conducted using antibodies against TIF2 (C-20, sc-6976), PSA

A67-B/E3, sc-7316) and GAPDH (FL-335, sc-25778) all from Santaruz Biotechnology.

. Results

.1. The NTD retards nuclear export of AR in COS-1 cells

Transfected GFP-AR and GFP-LBD were localized in the cyto-lasm of COS-1 cells in the absence of DHT, and GFP-DBDH-LBDas evenly distributed between the cytoplasm and nuclei (Fig. 1).

reatment with 1 nM DHT induced nuclear import of GFP-ARnd GFP-DBDH-LBD within 4 h after DHT induction (Fig. 1),hereas 100 nM DHT was required to induce nuclear import

f GFP-LBD. A redistribution of GFP-LBD from the nucleus tohe cytoplasm was observed 12 h after DHT withdrawal (Fig. 1),onsistent with our previous observation [19]. Similarly, hormoneithdrawal resulted in transition of GFP-DBDH-LBD from theucleus to being evenly distributed between the cytoplasm anducleus in COS-1 cells. In contrast, after hormone withdrawal,FP-AR remained predominantly in the nucleus, with the cyto-

Please cite this article in press as: J.A. Dar, et al., The N-terminal docastration-resistant prostate cancer cells, J. Steroid Biochem. Mol. Bio

lasmic GFP signal being slightly but consistently increasedFig. 1), in agreement with our previous findings [19]. Thesebservations suggest the suppression of AR nuclear export byTD.

was reproduced 5 times.

3.2. The NTD is required for androgen-independent nuclearlocalization of AR in C4-2 cells

Transfected GFP-AR is localized to the cytoplasm in AR-negativecell lines such as COS-1 and PC3 in androgen-free medium (Fig. 1)[35,36]. In contrast, GFP-AR is localized to the nucleus in castrationresistant C4-2 cells under the same conditions [25]. To evaluate theimportance of various domains of AR in its androgen-independentnuclear localization in CRPC, we examined the intracellular local-ization of GFP-AR, GFP-DBDH-LBD and GFP-LBD in C4-2 cells. Asexpected, GFP-AR was localized to the nuclei of C4-2 cells inandrogen-free culture medium (Fig. 2), which is distinctly differ-ent from the cytoplasmic localization in COS-1 cells (Fig. 1) andLNCaP cells [37]. However, GFP-DBDH-LBD and GFP-LBD exhib-ited even localization and cytoplasmic localization, respectively,in C4-2 cells, which were the same as their subcellular localiza-tions in COS-1 cells. The above observation suggests that the NTD

main of the androgen receptor drives its nuclear localization inl. (2014), http://dx.doi.org/10.1016/j.jsbmb.2014.03.004

is responsible for the differential subcellular localization betweenC4-2 and COS-1 cells in androgen-free medium and plays a key rolein androgen-independent nuclear localization of AR in C4-2 cells.

ARTICLE IN PRESSG ModelSBMB-4170; No. of Pages 8

4 J.A. Dar et al. / Journal of Steroid Biochemistry & Molecular Biology xxx (2014) xxx–xxx

Fig. 3. Subcellular distribution of GFP-AR, GFP-AR�(294-556) and GFP-AR�(1-293) in C4-2 cells. (A) Diagram of different GFP-AR deletion constructs, GFP-AR, GFP-AR�(294-556),and GFP-AR�(1-293). (B) Representative fluorescent images of C4-2 cells transiently transfected with GFP-AR, GFP-AR�(294-556) or GFP-AR�(1-293) in androgen-free conditions16 h. after transfection. The results were derived from five transfections for each GFP-fusion protein construct. (C) Quantitative analysis of results in (B). At least 200 cellsw ing cy5

3a

NltupHiwrsc

3e

aAcaaA(Nn2ic

3l

nC

ere counted for each transfection to determine the percentage of the cells display times.

.3. Amino acids (a.a.) 294–556 of the NTD are necessary forndrogen-independent AR nuclear localization in C4-2 cells

In order to identify the amino acid sequences within theTD of AR that are essential for androgen-independent nuclear

ocalization in C4-2 cells, we examined the intracellular localiza-ion of the deletion mutants GFP-AR�(294-556) and GFP-AR�(1-293)

sing fluorescence microscopy (Fig. 3A). GFP-AR�(294-556) dis-layed an even distribution between the nucleus and cytoplasm.owever, GFP-AR�(1-293) exhibited predominant nuclear local-

zation in transfected C4-2 cells, which is comparable to whatas observed in the full-length GFP-AR (Fig. 3B and C). These

esults suggest that a.a. 294–556 of the NTD region are neces-ary for androgen-independent nuclear localization of AR in C4-2ells.

.4. The NTD and a.a. 294–556 region alone have no significantffect on tagged-GFP subcellular localization in C4-2 cells

In order to determine whether a.a. 294–556 alone wouldffect tagged-GFP subcellular localization, GFP-NTD and GFP-R(294-556) (Fig. 4A) were transfected into LAPC4, C4-2 and LNCaPells and subcellular localization was assessed. Both GFP-NTDnd GFP-AR(294-556) were localized evenly between the nucleusnd cytoplasm in LAPC4, C4-2, and LNCaP cells, although GFP-R(294-556) had a slight propensity towards being more nuclear

Fig. 4B). Similarly, in the androgen-negative PC3 cells, both GFP-TD and GFP-AR(294-556) exhibited even localization between theucleus and the cytoplasm [38]. These results suggest that a.a.94–556 together with DBDH-LBD are essential for androgen-

ndependent nuclear localization of AR in LAPC4, C4-2 and LNCaPells.

.5. a.a. 294–556 of the NTD enhanced DHT-induced nuclearocalization of GFP-LBD in C4-2 cells

Please cite this article in press as: J.A. Dar, et al., The N-terminal docastration-resistant prostate cancer cells, J. Steroid Biochem. Mol. Bio

Although a.a. 294–556 of the AR does not appear to contain auclear localization signal, it may enhance AR nuclear import in4-2 cells. Since GFP-AR�(1-293) is already in the nuclei of C4-2

toplasmic (C), even (E), or nuclear (N) localization. The experiment was reproduced

cells in androgen-free media, it is not feasible to test the effect ofa.a. 294–556 on nuclear import of this construct, which contains anuclear localization signal (NLS1) in the DBDH region [18,39]. Thus,we transfected C4-2 cells with different GFP deletion constructs:GFP-AR(1-293)-LBD, GFP-AR(294-556)-LBD, and GFP-LBD (Fig. 5A).Intracellular localization of various transfected GFP-fusion pro-teins was examined (Fig. 5B). In the absence of DHT, both GFP-LBDand GFP-AR(1-293)-LBD localized to the cytoplasm of the trans-fected cells, while GFP-AR(294-556)-LBD displayed even localizationbetween the nucleus and cytoplasm (Fig. 5B). Upon treatment withDHT, specifically at higher doses (10–100 nM), GFP-AR(294-556)-LBDwas predominantly localized in the nucleus, while at lower doses(0.1–1.0 nM) GFP-AR(294-556)-LBD still displayed GFP fluorescentcells that were evenly distributed between the nucleus and cyto-plasm of the cell (Fig. 5C). In contrast, GFP-AR(1-293)-LBD localizedevenly in the cell at high doses of DHT (10–100 nM) (Fig. 5C). GFP-LBD also displayed even localization at 10 nM DHT and it took veryhigh levels of DHT (100 nM) for GFP-LBD to begin to enter thenucleus (Fig. 5C). The subcellular localization of GFP-NTD-LBD inresponse to DHT was similar to GFP-AR(1-293)-LBD, suggesting thata.a. 1–293 could counteract the effect of a.a. 294–556 in the con-struct. As a control and as expected, most of the transfected C4-2cells displayed even or nuclear localization of GFP-AR in the absenceof androgens, predominant nuclear localization in the presence of0.1 nM DHT and complete nuclear localization in the presence of1–100 nM DHT (Fig. 5C). These results suggest that a.a. 294–556enhances the androgen stimulated nuclear import of AR in C4-2cells.

3.6. The effect of a.a. 294–556 on androgen-independent nuclearlocalization of AR in C4-2 cells is insensitive to the Hsp90 inhibitor17-AAG

We have previously reported the regulation of androgen-independent localization of AR by Hsp90 [8,25]. However, the

main of the androgen receptor drives its nuclear localization inl. (2014), http://dx.doi.org/10.1016/j.jsbmb.2014.03.004

mechanisms of HSP90 regulation of AR androgen-independentnuclear localization remain unclear. Since the NTD plays an impor-tant role in androgen-independent AR nuclear localization, wetested whether the HSP90 inhibitor 17-AAG can block the effect of

ARTICLE IN PRESSG ModelSBMB-4170; No. of Pages 8

J.A. Dar et al. / Journal of Steroid Biochemistry & Molecular Biology xxx (2014) xxx–xxx 5

F d LNCR C4, Cb ced 5

aAonliG1bciat2i

3a

dAbibTva2ot(ia

ig. 4. Subcellular localization of GFP-AR(294-556) and GFP-NTD in LAPC4, C4-2 anepresentative images of GFP-NTD and GFP-AR(294-556) in transiently transfected LAPy fluorescence microscopy after 16 h of transfection. The experiment was reprodu

.a. 294–556 on androgen-independent nuclear localization of AR.s expected, 17-AAG treatment induced cytoplasmic localizationf GFP-AR in C4-2 cells (Fig. 6) and had no effect on GFP alone (dataot shown) [8]. Interestingly, 17-AAG did not affect the subcellular

ocalization of GFP-DBDH-LBD, which remained evenly distributedn the presence of 17-AAG (Fig. 6). The nuclear localization ofFP-AR�(1-293), which retains a.a. 294–556, was not disrupted by7-AAG. However, GFP-AR�(294-556) was sensitive to 17-AAG inhi-ition, which exhibited even distribution between nucleus andytoplasm in the absence of 17-AAG and cytoplasmic localizationn the presence of 17-AAG. This suggests that 17-AAG inhibitsndrogen-independent nuclear localization of AR and is mediatedhrough a.a. 1–293 of NTD but not a.a. 294–556, although a.a.94–556 are necessary for androgen-independent nuclear local-

zation of AR.

.7. TIF-2 is required for AR activation but notndrogen-independent nuclear localization in C4-2 cells

The region of a.a. 294–556 contains the TAU-5 transactivationomain, which is essential for androgen-independent activation ofR [40,41]. Transcription Intermediary Factor 2 (TIF-2), a mem-er of the steroid receptor co-activator family, plays a key role

n androgen-independent activation of AR in CRPC cells and haseen shown to bind to the TAU-5 motif [39]. Knockdown ofIF-2 has been reported to inhibit androgen-independent AR acti-ation in C4-2 cells [40]. Since TIF-2 binds to TAU-5 within.a. 294–556, we decided to test whether knockdown of TIF-

could also affect androgen-independent nuclear localizationf AR in C4-2 cells. As expected, TIF-2 knockdown inhibited

Please cite this article in press as: J.A. Dar, et al., The N-terminal docastration-resistant prostate cancer cells, J. Steroid Biochem. Mol. Bio

he expression of PSA, a gene regulated by AR transcriptionallyFig. 7A). However, TIF-2 knockdown did not inhibit androgen-ndependent nuclear localization of GFP-AR in C4-2 cells (Fig. 7Bnd C).

aP cells. (A) Diagram of GFP fusion constructs GFP-NTD and GFP-AR(294-556). (B)4-2, and LNCaP cells. The subcellular localization was assessed in complete mediumtimes.

4. Discussion

Multiple AR domains/motifs and their interactions withinAR and with AR co-factors are likely involved in determiningthe ultimate subcellular localization of AR. Identification of thedomains/motifs important in regulating AR intracellular traffickingand/or subcellular localization provides a foundation for furthercharacterization of their interactions essential for androgen-independent AR nuclear localization in CRPC. We have exploredthe mechanism of androgen-independent nuclear localization ofAR by characterizing the subcellular localization of different ARdeletion mutants in castration-resistant C4-2 cells. Our findingssuggest an important role of the NTD in castration- resistance ofprostate cancer, with a.a. 294–556 in the NTD being required forandrogen-independent AR nuclear localization.

Previous studies have shown the important roles of the DBDHand LBD regions in nucleocytoplasmic trafficking of AR, with NLS1present in the DBDH region and NLS2 and NES in the LBD [17–19].Numerous studies have demonstrated the importance of the NTD inAR transactivation [11,36,42–45]. However, the role of the NTD inAR intracellular trafficking is poorly understood, particularly withregards to the androgen-independent nuclear localization of AR.NTD modulation of AR nuclear localization has previously beenreported to involve tubulin, which can be targeted by microtubule-targeting chemotherapy drugs such as docetaxel [37]. Our findingsof the NTD inhibiting nuclear export of AR in COS-1 cells and itsrequirement in androgen-independent nuclear localization of ARin C4-2 cells further argue for an important role of the NTD inmodulating AR subcellular localization.

Regulation of AR intracellular localization and/or trafficking by

main of the androgen receptor drives its nuclear localization inl. (2014), http://dx.doi.org/10.1016/j.jsbmb.2014.03.004

the NTD appears to be complex. The presence of the NTD led to thecytoplasmic localization of full-length AR in COS-1 cells, whereasGFP-DBDH-LBD, which lacks NTD, was evenly distributed in COS-1 cells. This finding is consistent with the presence of a region,

ARTICLE IN PRESSG ModelSBMB-4170; No. of Pages 8

6 J.A. Dar et al. / Journal of Steroid Biochemistry & Molecular Biology xxx (2014) xxx–xxx

Fig. 5. Effect of AR(294-556)on LBD sensitivity to DHT in C4-2 cells. (A) Diagram of GFP fusion constructs GFP-AR, GFP-NTD-LBD, GFP-AR(1-293)-LBD, GFP-AR(294-556)-LBD, andGFP-LBD. (B) Representative images of GFP-tagged protein localization in the absence or presence of 10 nM DHT. (C), Quantification analysis of GFP fusion constructs in C4-2c for eae of sit

AiAtPNidpnint

ells followed by treatment of DHT (0–100 nM) (B). At least 200 cells were counted

ven (E), or nuclear (N) localization. The subcellular localization and quantificationimes.

R50-250, within the NTD capable of promoting cytoplasmic local-zation of the AR in PC3 cells [38]. Furthermore, GFP-AR and GFP-R�(294–556) were localized in the cytoplasm in virtually all of the

ransfected cells whereas GFP-AR�(1–293) was evenly distributed inC3 cells [46]. These observations suggest that a.a. 294–556 of theTD are not sufficient to drive complete nuclear localization of AR

n PC3 cells. It is possible that a.a. 294–556 of the NTD can onlyrive AR nuclear localization in AR-positive castration-resistantrostate cancer cells such as C4-2. The presence of the NTD did

Please cite this article in press as: J.A. Dar, et al., The N-terminal docastration-resistant prostate cancer cells, J. Steroid Biochem. Mol. Bio

ot inhibit androgen-induced nuclear import of AR because DHTnduced efficient nuclear import of GFP-AR when compared to theuclear import of GFP-DBDH-LBD in COS-1 cells (Fig. 1). In con-rast, the NTD significantly retarded export of AR from the nuclei of

ch transfection to determine the percentage of the cells displaying cytoplasmic (C),gnal were assessed by fluorescent microscopy. The experiment was reproduced 3

COS-1 cells after androgen withdrawal (Fig. 1). This suggests thatthe effect of the NTD on AR nuclear export is likely more dramaticthan its influence on AR nuclear import. The mechanisms respon-sible for differential effects of the NTD on nuclear import versusexport are unclear and will require further investigation.

Our studies suggest that the NTD is responsible for the differ-ential localization of AR in C4-2 and COS-1 cells and plays a keyrole in determining androgen-independent nuclear localization ofAR in C4-2 cells. The sensitivity of AR to androgens in different cell

main of the androgen receptor drives its nuclear localization inl. (2014), http://dx.doi.org/10.1016/j.jsbmb.2014.03.004

lines can be very different, with CRPC cells being hypersensitizedto low levels of androgens [47]. Since the NTD is involved in dif-ferential localization of AR in COS-1 and C4-2 cells and is requiredfor androgen-independent nuclear localization of AR, it may play

ARTICLE IN PRESSG ModelSBMB-4170; No. of Pages 8

J.A. Dar et al. / Journal of Steroid Biochemistry & Molecular Biology xxx (2014) xxx–xxx 7

Fig. 6. Localization of GFP-tagged AR and AR deletion mutants in C4-2 cells inthe presence of Hsp90 inhibitor 17-AAG. C4-2 cells transfected with GFP-AR, GFP-AR�(294-556), GFP-AR�(1-293) and GFP-DBDH-LBD were treated with DMSO (control)ob

aClh

icatticui

tTdlHacA1Hv1ti

oCi1Cno

Fig. 7. Effect of TIF2 knockdown on AR subcellular localization and PSA expressionin C4-2 cells. (A) Western blot analysis of C4-2 cells transfected with control siRNAor siTIF2 for 72 h. The Western blots were probed with anti-TIF2, anti-PSA, or anti-GAPDH (loading control) antibodies. (B) Representative images of GFP-AR in C4-2cells co-transfected with GFP-AR and either control siRNA or siTIF2 for 72 h. Local-ization of GFP-AR was determined using fluorescence microscopy. (C) Percentage

r HSP90 inhibitor 17-AAG in androgen-free conditions. Localization was assessedy fluorescent microscopy 4 h after the treatment.

n important role in regulating androgen hypersensitivity of AR in4-2 cells. Understanding the mechanisms by which the NTD regu-

ates AR subcellular localization may provide new insights into ARypersensitization that can lead to castration-resistance.

Androgen-independent activation of AR requires its androgen-ndependent nuclear localization in CRPC cells. However, it is notlear whether AR nuclear localization is sufficient for AR trans-ctivation in CRPC cells. The region of a.a. 294–556 containshe TAU-5 transactivation domain essential for AR transactiva-ion and recognized by TIF2 [48]. In our study, TIF-2 knockdownnhibited AR transactivation but not nuclear localization in C4-2ells, suggesting that transactivation activity of the NTD can bencoupled from its function in modulating AR subcellular local-

zation.The N-terminal a.a. 1–293 region also appears to play an impor-

ant role in modulating AR subcellular localization in C4-2 cells.he deletion mutant GFP-AR�(294-556), which retains a.a. 1–293,isplayed even distribution in transfected C4-2 cells and became

ocalized to the cytoplasm in the presence of 300 nM 17-AAG, anSP90 inhibitor. This suggests that the region of a.a. 1–293 medi-tes the 17-AAG inhibition of nuclear localization of AR in C4-2ells, which is further substantiated by the observation that GFP-R�(1-293) is insensitive to 17-AAG inhibition. The region of a.a.–293 contains a motif essential for N/C interaction within AR.sp90 inhibition by 17-AAG may influence the N/C interactionia the LBD, because Hsp90 can bind to the LBD [8]. Alternatively,7-AAG may affect some signaling molecules that could modulatehe function of a.a. 1–293 and subsequently inhibit androgen-ndependent nuclear localization of AR.

In summary, this study presents evidence for an important rolef the NTD in AR androgen-independent nuclear localization inRPC cells. The region of a.a. 294–556 is required for androgen-

ndependent AR nuclear localization whereas the region of a.a.

Please cite this article in press as: J.A. Dar, et al., The N-terminal docastration-resistant prostate cancer cells, J. Steroid Biochem. Mol. Bio

–293 mediates Hsp90 regulation of AR nuclear localization inRPC cells. Although the region of a.a. 294–556 does not contain auclear import signal, it was able to enhance DHT-induced importf the LBD. This finding is consistent with the findings of Jenster

of cells displaying nuclear or even localization was determined for GFP-AR in cellsco-transfected with either control siRNA or siTIF2. The experiment was reproducedtwo times.

et al. [11], and Zhou et al. [49], which showed that the NTD is presentin both the nucleus and cytoplasm. However, according to a recentreport by Kaku et al., the NTD of AR was localized to the nucleus,suggesting the presence of an NLS in the NTD [18]. Future studieswill be required to determine the reasons for this discrepancy inobservations. Although we found that the NTD could modulate ARlocalization when present in the full-length AR, the region of a.a.294–556 did not confer nuclear localization alone. Our study alsoshowed that transactivation of the NTD can be uncoupled from itsmodulation of AR subcellular localization. Future identification andcharacterization of cellular factors involved in NTD modulation

main of the androgen receptor drives its nuclear localization inl. (2014), http://dx.doi.org/10.1016/j.jsbmb.2014.03.004

of AR subcellular localization will be essential for elucidating themechanisms of androgen-independent nuclear localization andmay lead to new therapeutic targets for the treatment of CRPC.

ING ModelS

8 istry

A

oKiCP

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

ARTICLEBMB-4170; No. of Pages 8

J.A. Dar et al. / Journal of Steroid Biochem

cknowledgements

This investigation was supported in part by National Institutesf Health Grants (R01 CA108675, 1 P50 CA90386, 5 R37 DK51193)..E. is a Post-doctoral Scholar supported by T32 DK007774. K.Z.M.

s a Mellam Scholar and L.E.P. is a Tippins Scholar. Z.W. holds UPMChair in Urological Research and J.B.N. is the Frederic N. Schwentkerrofessor.

eferences

[1] M. Welsh, et al., Identification in rats of a programming window for repro-ductive tract masculinization, disruption of which leads to hypospadias andcryptorchidism, J. Clin. Invest. 118 (April, 2008) 1479.

[2] J.T. Isaacs, D.S. Coffey, Etiology and disease process of benign prostatic hyper-plasia, Prostate (Suppl. 2) (1989) 33.

[3] K.J. O’Malley, et al., The expression of androgen-responsive genes is up-regulated in the epithelia of benign prostatic hyperplasia, Prostate 69(December 1, 2009) 1716.

[4] J.M. Kozlowski, W.J. Ellis, J.T. Grayhack, Advanced prostatic carcinoma. Earlyversus late endocrine therapy, Urol. Clin. North Am. 18 (February, 1991) 15.

[5] A. Jemal, R. Siegel, J. Xu, E. Ward, Cancer statistics, 2010, CA Cancer J. Clin. 60(September–October 2010) 277.

[6] C.D. Chen, et al., Molecular determinants of resistance to antiandrogen therapy,Nat. Med. 10 (January 2004) 33.

[7] O.L. Zegarra-Moro, L.J. Schmidt, H. Huang, D.J. Tindall, Disruption of androgenreceptor function inhibits proliferation of androgen-refractory prostate cancercells, Cancer Res. 62 (February 15, 2002) 1008.

[8] A.J. Saporita, J. Ai, Z. Wang, The Hsp90 inhibitor, 17-AAG, prevents the ligand-independent nuclear localization of androgen receptor in refractory prostatecancer cells, Prostate 67 (April 1, 2007) 509.

[9] E.L. Yong, F. Ghadessy, Q. Wang, A. Mifsud, S.C. Ng, Androgen receptor trans-activation domain and control of spermatogenesis, Rev. Reprod. 3 (September1998) 141.

10] J.A. Simental, M. Sar, M.V. Lane, F.S. French, E.M. Wilson, Transcriptional acti-vation and nuclear targeting signals of the human androgen receptor, J. Biol.Chem. 266 (January 5, 1991) 510.

11] G. Jenster, et al., Domains of the human androgen receptor involved insteroid binding, transcriptional activation, and subcellular localization, Mol.Endocrinol. 5 (October 1991) 1396.

12] G. Jenster, et al., Changes in the abundance of androgen receptor isotypes:effects of ligand treatment, glutamine-stretch variation, and mutation of puta-tive phosphorylation sites, Biochemistry 33 (November 29, 1994) 14064.

13] G. Jenster, H.A. van der Korput, J. Trapman, A.O. Brinkmann, Identification of twotranscription activation units in the N-terminal domain of the human androgenreceptor, J. Biol. Chem. 270 (March 31, 1995) 7341.

14] N.L. Chamberlain, D.C. Whitacre, R.L. Miesfeld, Delineation of two distinct type1 activation functions in the androgen receptor amino-terminal domain, J. Biol.Chem. 271 (October 25, 1996) 26772.

15] T. Gao, M. Marcelli, M.J. McPhaul, Transcriptional activation and transientexpression of the human androgen receptor, J. Steroid Biochem. Mol. Biol. 59(September 1996) 9.

16] J. Reid, S.M. Kelly, K. Watt, N.C. Price, I.J. McEwan, Conformational analysis of theandrogen receptor amino-terminal domain involved in transactivation. Influ-ence of structure-stabilizing solutes and protein-protein interactions, J. Biol.Chem. 277 (May 31, 2002) 20079.

17] T. Ylikomi, M.T. Bocquel, M. Berry, H. Gronemeyer, P. Chambon, Cooperation ofproto-signals for nuclear accumulation of estrogen and progesterone receptors,EMBO J. 11 (October 1992) 3681.

18] N. Kaku, K. Matsuda, A. Tsujimura, M. Kawata, Characterization of nuclearimport of the domain-specific androgen receptor in association with theimportin alpha/beta and Ran-guanosine 5′-triphosphate systems, Endocrinol-ogy 149 (August 2008) 3960.

19] A.J. Saporita, et al., Identification and characterization of a ligand-regulatednuclear export signal in androgen receptor, J. Biol. Chem. 278 (October 24, 2003)41998.

20] Z.X. Zhou, C.I. Wong, M. Sar, E.M. Wilson, The androgen receptor: an overview,Recent Prog. Hormone Res. 49 (1994) 249.

21] E.P. Gelmann, Molecular biology of the androgen receptor, J. Clin. Oncol. Off. J.Am. Soc. Clin. Oncol. 20 (July 1, 2002) 3001.

22] C.S. Chang, J. Kokontis, S.T. Liao, Molecular cloning of human and rat com-plementary DNA encoding androgen receptors, Science 240 (April 15, 1988)324.

Please cite this article in press as: J.A. Dar, et al., The N-terminal docastration-resistant prostate cancer cells, J. Steroid Biochem. Mol. Bio

23] D.B. Lubahn, et al., Cloning of human androgen receptor complementary DNAand localization to the X chromosome, Science 240 (April 15, 1988) 327.

24] K.J. O’Malley, et al., Hsp90 inhibitor 17-AAG inhibits progression of LuCaP35xenograft prostate tumors to castration resistance, Prostate 72 (July 1, 2012)1117.

[

PRESS& Molecular Biology xxx (2014) xxx–xxx

25] J. Ai, et al., HDAC6 regulates androgen receptor hypersensitivity and nuclearlocalization via modulating Hsp90 acetylation in castration-resistant prostatecancer, Mol. Endocrinol. 23 (December 2009) 1963.

26] D.A. Monks, et al., Androgen receptor and Kennedy disease/spinal bulbar mus-cular atrophy, Hormones Behav. 53 (May 2008) 729.

27] I. Mazen, S. Lumbroso, S. Abdel Ghaffar, N. Salah, C. Sultan, Mutation of theandrogen receptor (R840S) in an Egyptian patient with partial androgen insen-sitivity syndrome: review of the literature on the clinical expression of differentR840 substitutions, J. Endocrinol. Invest. 27 (January 2004) 57.

28] C.A. Heinlein, C. Chang, Androgen receptor in prostate cancer, Endocrine Rev.25 (April 2004) 276.

29] W.D. Tilley, G. Buchanan, T.E. Hickey, J.M. Bentel, Mutations in the andro-gen receptor gene are associated with progression of human prostate cancerto androgen independence, Clin. Cancer Res. Off. J. Am. Assoc. Cancer Res. 2(February 1996) 277.

30] H. Kawate, et al., Impaired nuclear translocation, nuclear matrix target-ing, and intranuclear mobility of mutant androgen receptors carrying aminoacid substitutions in the deoxyribonucleic acid-binding domain derived fromandrogen insensitivity syndrome patients, J. Clin. Endocrinol. Metab. 90(November 2005) 6162.

31] M.L. Cutress, H.C. Whitaker, I.G. Mills, M. Stewart, D.E. Neal, Structural basis forthe nuclear import of the human androgen receptor, J. Cell Sci. 121 (April 1,2008) 957.

32] S. Zoppi, et al., Amino acid substitutions in the DNA-binding domain of thehuman androgen receptor are a frequent cause of receptor-binding positiveandrogen resistance, Mol. Endocrinol. 6 (March 1992) 409.

33] L.V. Nazareth, et al., A C619Y mutation in the human androgen receptor causesinactivation and mislocalization of the receptor with concomitant sequestra-tion of SRC-1 (steroid receptor coactivator 1), Mol. Endocrinol. 13 (December1999) 2065.

34] H.C. Wu, et al., Derivation of androgen-independent human LNCaP prostaticcancer cell sublines: role of bone stromal cells, Int. J. Cancer. (Journal interna-tional du cancer) 57 (May 1, 1994) 406.

35] R.K. Tyagi, et al., Dynamics of intracellular movement and nucleocytoplas-mic recycling of the ligand-activated androgen receptor in living cells, Mol.Endocrinol. 14 (August 2000) 1162.

36] A.K. Roy, et al., Androgen receptor: structural domains and functional dynam-ics after ligand-receptor interaction, Ann. New York Acad. Sci. 949 (December2001) 44.

37] M.L. Zhu, et al., Tubulin-targeting chemotherapy impairs androgen recep-tor activity in prostate cancer, Cancer Res. 70 (October 15, 2010)7992.

38] J.A. Dar, et al., N-terminal domain of the androgen receptor contains a regionthat can promote cytoplasmic localization, J. Steroid Biochem. Mol. Biol.(October 4, 2013).

39] R. Chmelar, G. Buchanan, E.F. Need, W. Tilley, N.M. Greenberg, Androgen recep-tor coregulators and their involvement in the development and progression ofprostate cancer, Int. J. Cancer 120 (2007) 719.

40] L. Callewaert, N. Van Tilborgh, F. Claessens, Interplay between two hormone-independent activation domains in the androgen receptor, Cancer Res. 66(January 1, 2006) 543.

41] S.M. Dehm, K.M. Regan, L.J. Schmidt, D.J. Tindall, Selective role of an NH2-terminal WxxLF motif for aberrant androgen receptor activation in androgendepletion independent prostate cancer cells, Cancer Res. 67 (October 15, 2007)10067.

42] G. Han, et al., Hormone status selects for spontaneous somatic androgenreceptor variants that demonstrate specific ligand and cofactor dependentactivities in autochthonous prostate cancer, J. Biol. Chem. 276 (April 6, 2001)11204.

43] G. Han, et al., Mutation of the androgen receptor causes oncogenic transfor-mation of the prostate, Proc. Natl. Acad. Sci. U.S.A. 102 (January 25, 2005)1151.

44] R.J. Andersen, et al., Regression of castrate-recurrent prostate cancer by a small-molecule inhibitor of the amino-terminus domain of the androgen receptor,Cancer Cell 17 (June 15, 2010) 535.

45] S.E. Rundlett, X.P. Wu, R.L. Miesfeld, Functional characterizations of theandrogen receptor confirm that the molecular basis of androgen action is trans-criptional regulation, Mol. Endocrinol. 4 (May, 1990) 708.

46] J.A. Dar, et al., N-terminal domain of the androgen receptor contains a regionthat can promote cytoplasmic localization, J. Steroid Biochem. Mol. Biol. 139(January, 2014) 16.

47] M. Gonit, et al., Hormone depletion-insensitivity of prostate cancer cells issupported by the AR without binding to classical response elements, Mol.Endocrinol. 25 (April, 2011) 621.

48] E. Metzger, J.M. Muller, S. Ferrari, R. Buettner, R. Schule, A novel inducible trans-activation domain in the androgen receptor: implications for PRK in prostate

main of the androgen receptor drives its nuclear localization inl. (2014), http://dx.doi.org/10.1016/j.jsbmb.2014.03.004

cancer, EMBO J. 22 (January 15, 2003) 270.49] Z.X. Zhou, M. Sar, J.A. Simental, M.V. Lane, E.M. Wilson, A ligand-dependent

bipartite nuclear targeting signal in the human androgen receptor: Require-ment for the DNA-binding domain and modulation by NH2-terminal andcarboxyl-terminal sequences, J. Biol. Chem. 269 (May 6, 1994) 13115.