View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Available online at www.sciencedirect.com

Biochimica et Biophysica Acta 1783 (2008) 59–73 www.elsevier.com/locate/bbamcr

Agonist-mediated docking of androgen receptor onto the mitotic chromatin platform discriminates intrinsic mode of action of prostate cancer drugs ⁎ Sanjay Kumar, Nagendra K. Chaturvedi, Subodh Kumar, Rakesh K. Tyagi

Special Centre for Molecular Medicine, Jawaharlal Nehru University, New Delhi 110067, India Received 27 April 2007; received in revised form 2 October 2007; accepted 5 November 2007 Available online 19 November 2007

Abstract

This study documents the analysis of a hitherto unreported dynamic behavior of androgen receptor (AR), a member of the nuclear receptor superfamily. Employing GFP-tagged AR, we observed agonist-mediated docking of AR onto the mitotic chromatin during all the stages of mitosis. When bound to therapeutic drugs with intrinsically absolute or partial agonistic properties, AR concomitantly associated with the mitotic chromatin. Conversely, pure antagonists known to bind and subsequently translocate unliganded AR from cytoplasm to nuclear compartment did not provoke such association. The agonist-mediated docking of AR could not be competed with other transcription factors that constitutively preoccupied the chromosomal docking sites. Amongst the previously reported proteins, AR is first example of a transcription factor whose response on mitotic chromatin platform can be modulated in a ligand-specific manner. However, data from live cell imaging revealed that co- activators of agonist-activated receptor that are recruited into “nuclear foci” of interphase chromatin are dislodged from the mitotic chromatin during cell division. This implies that in absence of critical co-activators, AR transverses mitotic phase in transcriptionally silenced state. Finally, our results indicate that ligand-mediated dynamic relationship of nuclear receptors with mitotic chromatin can be effectively exploited to study, analyze and authenticate therapeutic ligands. © 2007 Elsevier B.V. All rights reserved.

Keywords: Steroid receptor; Mitotic chromatin; Antiandrogen; Prostate cancer

1. Introduction dimerization, ligand binding and activation function 2 (AF-2). For transcription function of AR, interaction between N/C- Androgen Receptor (AR) is a member of the nuclear receptor termini is modulated in ligand-dependent fashion. For example, (NR) superfamily that acts as a ligand-activated transcription agonists induce the N/C interaction while antagonist (like factor. AR plays a major role in the development and main- casodex, hydroxyflutamide etc.) do not induce detectable N/C tenance of male sex characteristics by executing the biological interaction [4–6]. Recently, N/C interaction are reported to play functions of androgens. AR also has an important role in pro- an important role in the control of AR-chromatin binding and state cell proliferation and differentiation; and during androgen- also for the recruitment of SWI/SNF complex that in turn re- responsive prostate cancer, manipulation of AR function by models chromatin to allow AR to bind to AR response element antiandrogenic drugs is the primary mode of treatment of the in interphase chromatin [5]. Also, AR N/C interactions take malignancy [1–3]. As a member of steroid/nuclear receptor place predominantly when AR is mobile, possibly to prevent superfamily, AR is structurally characterized by an amino- unfavorable or untimely cofactor interactions [7]. terminal domain (N-terminal) containing activation function 1 Unliganded AR is a cytoplasmic protein that upon androgen (AF-1); a central DNA binding domain; and, a multifunctional binding translocates to the nucleus. In the nucleus it binds to the carboxyl-terminal domain (C-terminal) involved in receptor regulatory regions of androgen-responsive genes and subse- quently modulates their transcription [6,8,9]. Conversely, most ⁎ Corresponding author. Tel.: +91 11 2674 1544; fax: +91 11 26741781. of the antiandrogenic compounds though bind and translo- E-mail addresses: [email protected], [email protected] cate cytoplasmic AR to nuclear compartment but essentially in- (R.K. Tyagi). hibit or impede its transcriptional function. These antiandrogens

0167-4889/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.bbamcr.2007.11.002 60 S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73 also include the therapeutic drugs like cyproterone acetate, scription by all the three RNA polymerases prevalent in in- flutamide, casodex (bicalutamide), nilutamide etc. that have terphase nuclei of eukaryotic cells is silenced [29,30].Studies gained widespread use in treatment of androgen-dependent with simplified systems revealed that reversible phosphoryla- prostate cancer [10,11]. Since these antiandrogens can efficiently tion of the basal transcription machinery and core histones by inhibit or diminish AR-mediated gene expression, they are cyclin-dependent kinases or its downstream kinases, chromatin therefore useful tools in the treatment of androgen-dependent remodeling complexes and physical compaction of chroma- prostate cancer. Antiandrogens are either complete or partial tin repress transcription during mitosis by dislodging protein inhibitors of AR activity, depending on the nature of the complexes. [29–31]. In contrast to silencing, Sciortino et al. therapeutic molecule. Recent studies demonstrate that AR [32] reported active transcription of cyclin B1 gene during antagonists facilitate AR-DNA binding but inhibit transcrip- mitosis in HeLa cells. Majority of basal transcription factors, tional activation by the recruitment of corepressors [12,13].As RNA polymerases, enhancer-binding factors, histone acetyl compared to androgens, antiandrogens induce a different AR transferases (HATs) and deacetylases (HDACs) are disso- conformation, thereby influencing the recruitment of co- ciated from the mitotic chromatin [29,30,33,34]. However, regulators (co-activators and corepressors) [8,14]. In a classical emerging evidences are indicating that some of the major cel- finding this ligand-selective modulation of AR activity was also lular proteins, including ones having direct or indirect roles in broadened by mutation in AR (T877A and T857A substitution) gene transcription function, remain associated with condensed found in prostate cancer [3,15]. chromatin during mitosis. These include TATA binding pro- Discovery of green fluorescent protein (GFP) and availability teins (TBP), high mobility group protein (HMGB1, HMGB2, of multi-colored fluorescent variants, in conjunction with the HMGA1a), heat shock transcription factor 2 (HSF2), upstream advances in fluorescence microscopy, have helped in revealing binding factor (UBF), RNA polymerase I, topoisomerase II, finer details of protein localization and subcellular dynamics in insulator protein CTCF, co-activator PC4, transcription factors living cells [16]. In principle, all intracellular steroid receptors Fox1 and Runx2 [35–44 and references therein]. During this when bound to their cognate hormones are nuclear and tran- period of initial discoveries, we also reported the constitutive scriptionally active. However, subcellular localization of unli- association of a NR superfamily member to mitotic chromatin ganded receptors varies from being predominantly cytoplasmic throughout mitosis. The finding that a nuclear receptor, Preg- [e.g. AR, glucocorticoid receptor (GR)] to predominantly nane and Xenobiotic Receptor (PXR), associates with con- nuclear [e.g. estrogen receptor (ERα), progesterone receptor]. densed chromatin during all the stages of mitosis was novel and Also, data obtained from GFP-tagged steroid receptors ex- hitherto unanticipated [45]. Nevertheless, with some other pressed and imaged from live cells have consistently shown that proteins associating to mitotic chromatin, attempts have been agonist-bound receptors in nucleus reorganize into speckled made to demonstrate that there is retention of gene expression pattern revealing discrete receptor accumulation into 250–400 patterns by occupation of target gene promoters and that these fluorescent speckles [9,17,18]. These have been commonly promoters are maintained in an open chromatin configuration referred to as “nuclear foci”. The initial speculation that “nuclear even during mitosis [32,41,43,44]. The issue, however, is still foci” represent transcriptionally active form of steroid receptors an important challenge to resolve and needs to be pursued while homogeneously distributed receptor represents transcrip- further to establish the generality of the phenomenon more tionally inactive receptor has drawn substantial experimental concretely. support [18–20]. Subsequent studies in live cells have shown To examine the possibility if some other members of the NR that some of the co-activators or basal transcription factors like superfamily can also associate with the mitotic chromatin, we GRIP1, TIF2, CBP, SRC-1, SRC-3 etc. co-localize with agonist- initially examined two extensively studied members of this activated AR in “nuclear foci” but not with pure antagonist- group. Other than PXR reported previously, our choice was bound steroid receptors [21–25]. Thus, agonist-generated “nu- restricted to other two important members i.e. AR and ERα clear foci” were suggestive of multi-protein complexes required based on the fact that in unliganded state the former is a cyto- for activation of hormone responsive genes [reviewed in plasmic protein while the latter is constitutively localized in the 9,19,26]. Hager's laboratory [6,27] has provided direct evidence nucleus. Clinically, both, AR and ERα are targeted with anti- that the hormone-occupied receptor undergoes rapid exchange hormone drugs for treatments during hormone-responsive stages between interphase chromatin and the nucleoplasmic space. This of prostate and breast cancer respectively[1,11,46]. In the per- interaction of regulatory proteins with target sites in chromatin spective of the dilemma that these cancerous cells transform was observed to be more dynamic process than previously from hormone-responsive to hormone-independent stages, it is believed and lead to the proposal of “hit and run” mechanism of speculated that prevention of transcription function of AR and receptor action. ERα by long-term exposure to anti-hormone drugs may actually During the mitotic stages, interphase chromatin undergoes alter the epigenetic landscape of the targeted cells. This in turn dynamic structural alterations that result in the formation of may impose upon these cells to undertake survival strategies and highly condensed mitotic chromatin with compaction ratio up proliferate via alternative survival routes [46,47]. Since AR and to 1:10,000 [28]. This condensed packaging of eukaryotic ERα share major structural and functional similarities with each genome into chromatin is believedtoplayaroleintheab- other, present work though dealing primarily with AR, could rogation of active gene transcription. In general, it has been also provide important cues for endocrine therapy of ER- assumed that with the onset of mitosis, the active gene tran- responsive malignancies. S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73 61

2. Materials and methods for raising AR antibody were cleared by institutional ethical committee and housed in the central animal house, a core facility for Jawaharlal Nehru Uni- 2.1. Biochemicals versity, New Delhi, India.

Prostate cancer drugs (antiandrogens including spironolactone, cyproterone 2.5. Development of a stably transfected cell line constitutively acetate, nilutamide), 5α-Dihydrotestosterone, nocodazole, Escort III & IV and expressing human wild type AR Cy3-conjugated anti-rabbit IgG raised in sheep, HRP-conjugated anti-rabbit secondary antibodies raised in goat and Hoechst were procured from Sigma HepG2 cells were transfected with wild type AR expression plasmid and a Chemicals Co. (St. Louis, MO, USA). Casadox (bicalutamide) was a gift from vector that contains neomycin resistance gene (pcDNA3.1) in a ratio of 10:1. AstraZeneca (UK). All mammalian cell culture reagents were from Hyclone The cells were allowed to double once under non-selective conditions. Later, (Logan, Utah, USA) and plastic-wares purchased from Corning Costar Corp. cells were supplemented with the complete medium containing 400 μg/ml of (USA). Ni–NTA His–Bind matrix was a product from Qiagen (Valencia, USA). G418 (selective medium) and the medium was replaced every third day. After Other general chemicals used were of analytical grade and were procured from two weeks of selection period, individual colonies were isolated and further different commercial sources. propagated under selective conditions. Individual clones were screened for the stable expression of AR by indirect immunodetection and Western blotting 2.2. Plasmids analysis using AR specific polyclonal antibodies.

GFP-tagged human AR (GFP-AR) and wild type AR (pSG5-AR) were a gift 2.6. Western blot analysis from Prof. O.A. Janne (University of Helsinki, Finland) and has been previously reported [48]. GFP-hERα / GFP-SRC-3 and GFP-TBP were provided by Dr. HepG2, LNCaP, T47D and Hep-AR cells were cultured to a confluency of M. A. Mancini (Baylor College of Medicine, Houston, USA) and Dr. Sui Huang ∼80% in 60 mm culture dishes. Total protein was extracted from these cells using (Northwestern University Medical School, Chicago) respectively. GFP-GRIP1 trizole TRI reagent (Sigma) according to the instructions provided with the reagent. was obtained from Dr. B.M. Paschal (University of Virginia, USA) and is des- The proteins were finally dissolved in SDS-PAGE sample buffer, denatured by cribed previously [23]. Human PXR was subcloned in frame between EcoRI and heating at 95 °C for 5 min. and resolved on a 10% SDS-PAGE. Proteins were BamHI restriction sites of pDsRed-Express-C1 (BD Biosciences, USA) from electro-blotted onto the polyvinyldifluoridine (PVDF) membrane using semi-dry GFP-PXR as previously described [45]. The full-length AR cDNA was isolated transfer system (BioRad). After protein transfer, the blot was blocked with 5% non- from GFP-AR and subcloned between BglII and BamHI restriction sites of fat dry milk in TBS for 2 h at room temperature and then incubated overnight with pDsRed-Express-C1 (BD Biosciences, USA) and is termed as RFP-AR. Full- AR antiserum at dilution of 1:5000 at 4 °C. The membrane was then washed three length TBP cDNAwas isolated from GFP-TBP and sub-cloned between KpnI and times with TBST (TBS with 0.1% Tween-20) and incubated for 1 h with 1:10,000 BamHI restriction sites of pDsRed-Express-C1 and is referred to as RFP-TBP.We dilution of horseradish peroxidase-conjugated anti-rabbit antibody. The bound cloned full-length cDNA for AR into EcoRI restriction site of pET-28a (Novagen) antibody complexes were detected using ECL method. and designated it as pET-AR that is having an N-terminal His-tag sequence. 2.7. Mitotic Chromatin fractionation 2.3. Mammalian cell cultures and transient plasmid transfections COS-1 cells were transfected with GFP-AR and cultured in steroid-stripped ATCC cell lines COS-1, HepG2, T47D and LNCaP were obtained from culture medium. After transfection, the cells were treated either with vehicle alone National Cell repository (NCCS, India). COS-1 and HepG2 were grown in or DHT (10−8 M) along with nocodazole (500 ng/ml) for 20 h. Mitotic chromatin DMEM supplemented with 10% FBS, 100 μg/ml penicillin and 100 μg/ml was isolated from cells after retrieving the nocodazole arrested cells by mitotic streptomycin (complete medium), while LNCaP cells were grown in Nutrient shake-off as described [41]. Cells were collected by centrifugation at 500 ×g for Mixture F-12 HAM, Kaighn's medium supplemented with 10% FBS, antibiotics 5 min at 4 °C, washed with cold PBS and resuspended in 75 mM KCl. After as above and 10−11 M DHT. T47D cells were grown in RPMI-1640 sup- 20 min on ice, cells were pelleted at 500 ×g for 5 min at 4 °C and homogenized in plemented with 10% FBS, antibiotics and 1 mM sodium pyruvate. The cultures disruption buffer (10 mM Tris–HCl, pH 7.4, 120 mM KCl, 20 mM NaCl, 0.1% were maintained in a humidified incubator maintained at 5% CO2 and 95% air Triton X-100, 2 mM CaCl2, 1× protease inhibitor cocktail and 0.1 mM PMSF) by atmosphere at 37 °C. Transient DNA transfections in COS-1 and HepG2 cells drawing through a 26-gauge needle 5–10 times. Cell lysates were pelleted by were performed using Escort III and IV reagent respectively according to the centrifugation at 3300 ×g for 5 min at 4 °C and the supernatant was retained. instruction provided by the manufacturer. Pellets were washed once with disruption buffer and incubated again with disruption buffer containing 0.4 M NaCl. After incubation for 30 min, chromatin- 2.4. Prokaryotic expression and purification of full-length AR for enriched fraction was collected by centrifugation at 40,000 ×g (Beckmann) for raising polyclonal antibodies 15 min at 4 °C. Protein in the samples was estimated by CB-X™ Protein assay kit (GenoTech). Expression and purification of full-length AR in E. coli BL21 cells was performed as described earlier [49] with minor modifications. In brief, pET-AR 2.8. Fluorescence microscopy and live cell imaging transformed E. coli BL21 (DE3) cells were grown up to 0.4 to 0.6 OD600 at 37 °C and the AR expression was induced by 1 mM isopropyl-1-thio-D-galactopyrano- COS-1 cells were seeded into 35 mm plates and allowed to grow to 70–80% side (IPTG) at 30 °C for 3 h. Over expressed protein from the inclusion bodies confluency. The following day, cells were transfected with 500 ng of GFP-AR, fraction was extracted with urea extraction buffer [100 mM sodium phosphate pH GFP-ERα, GFP-PXR, RFP-AR, GFP-GRIP1, GFP-SRC-3, RFP-TBP or RFP- 8, 10 mM Tris–HCl, pH 8, 100 mM NaCl, 8 M Urea, 10 mM imidazol and 1× PXR using Escort III transfection reagent. After 12–15 h of transfection protease inhibitors cocktail] at RT for 60 min. and subsequently loaded onto incubation, cells were supplemented with 5% steroid-stripped serum DMEM 0.5 ml of Ni–NTA His–Bind matrix column. The matrix was washed three times followed by treatment with the respective ligands that included of 10− 8 M5α- with washing buffer (100 mM sodium phosphate pH 8, 10 mM Tris–HCl, pH 8, dihydrotestosterone (DHT), 10− 8 M17β-estradiol and 10− 6 M cyproterone 100 mM NaCl, 8 M Urea, 20 mM imidazol with protease inhibitors) and the acetate, spironolactone, casadox, hydroxyflutamide, nilutamide or 20 μMof matrix-bound protein was eluted with elution buffer [100 mM sodium phosphate rifampicin prepared in DMSO:ethanol (1:1) When required, cells were arrested pH 8, 10 mM Tris–HCl, pH 8, 100 mM NaCl, 8 M Urea, 250 mM imidazol and in mitosis by adding 0.1 μg/ml nocodazole to culture media for 12–16 h. After protease inhibitors]. The eluted protein(s) recovered at different steps or as treatment period, events for nuclear translocation, sub-nuclear compartmenta- fractions was analyzed for purification on SDS-PAGE. Immunization of New lization and mitosis were evaluated under the fluorescence microscope. With Zealand white rabbit with 250 μg of purified AR protein recovered from the DHT, cell treatment between 4–24 h produced same results; however, preparatory SDS-PAGE gel was done as described elsewhere [49]. Rabbits used antiandrogens being relatively slow acting were incubated with cells for 12– 62 S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73

24 h. To facilitate the visualization of the nucleus Hoechst (final concentration −20 °C with 5% and 10% (v/v) acetic acid. Irrespective of the procedures used, 1 μg/ml) was added to the live cells at least 1 h prior to imaging. For the after 20 min. of fixation, cells were rehydrated and permeabilized for 5 min. in statistical analysis of the association of steroid/nuclear receptors with mitotic ice-cold PBS containing 0.1, 0.2 and 0.5% Triton X-100. Cells were then chromatin, the receptors expressing mitotic cells were counted in absence and incubated with blocking buffer (3% BSA in PBS) followed by incubations with in presence of their respective ligands. At any time point 50–100 mitotic cells primary and secondary antibodies. Of all the procedures tested, the approach that were scored. gave the most optimum immunodetection results with the mitotic cells (and also The fluorescent cells were viewed and imaged through a Nikon upright interphase cells) is elaborated in the following protocols. fluorescence microscope (model 80i) equipped with water immersion objectives For indirect immunodetection, Hep-AR cell line and LNCaP cells were and connected to cooled CCD digital camera (model Evolution VF, Media cultured over sterile glass cover slips. After treatment with 10−8 M DHT for Cybernetics, USA). Cell images were captured and analysed with Image ProPlus 24 h, The cells were washed thrice with 10 mM phosphate buffered saline (PBS) version 5.0 software (Media Cybernetics, USA). The images were processed and then fixed with chilled methanol (−20 °C) containing 5% (v/v) acetic acid using standard image processing techniques. for 20 min. on ice. Following fixation, cells were incubated further for 5 min. in ice-cold PBS containing 0.5% Triton X-100. The cells were then blocked in 3% 2.9. Procedure for cell fixation and permeabilization for bovine serum albumin (BSA) in PBS for 1 h. For immunodetection, the cells immunodetection in interphase and mitotic cells were incubated overnight with primary AR polyclonal antiserum at a dilution of 1:300 at 4 °C. After three washes with PBS, the cells were incubated with Cy3- Although a number of cell fixation and permeabilization procedures are conjugated sheep anti-rabbit IgG (1:300 dilution) prepared in PBS with 3% available that are in general use in immunocytology, some of these in several BSA. Hoechst-33342 was included with the secondary antibody preparation for staining and visualization of the nucleus. The cells were rinsed three times in instances have failed to detect the antigenic proteins associated with the condensed mitotic chromatin [35,36,50]. To overcome the circumstantial PBS and the cover slips were mounted on a glass slide with 20% glycerol in PBS limitation when working with condensed chromatin, some of the cell fixation and visualized under the fluorescence microscope. and permeabilization procedures were carried out for systematic and comparative analysis by immuno-probing the mitotic cells. Some of the cell 2.10. Preparation of mitotic chromosome spreads for AR localization fixation procedures or their modification used includes (a) chilled methanol at −20 °C, (b) 4% formaldehyde at RT, (c) 4% paraformaldehyde at RT, (d) chilled To prepare mitotic chromosome spreads, Hep-AR stable cell line and LNCaP ethanol at −20 °C with 5% and 10% (v/v) acetic acid and (e) chilled methanol at cells were incubated in a medium containing10− 8 M DHT and 0.1 μg/ml

Fig. 1. Association of GFP-tagged steroid/nuclear receptors with mitotic chromatin in nocodazole arrested cells. A choice of three nuclear receptors was made based on their subcellular localization in unliganded forms (i.e. AR = cytoplasmic; ERα and PXR = nuclear). For transient transfection, 500 ng of each of the GFP-tagged receptors was expressed separately in COS-1 cells under steroid-free conditions as described under “Materials and methods”. Following receptor expression, the control cells (unliganded) were treated with the vehicle alone or treated (liganded) with their cognate agonists (AR with DHT 10− 8 M, ERα with 17β-estradiol 10−8 M, PXR with rifampicin 20 μM) as indicated in the figure. Ligand treatments were given after arresting the cells in mitosis with 0.1 μg/ml nocodazole for 12–16 h. Panel A shows the differences in chromosomal association of the three unliganded receptors while panel B shows a uniform agonist-mediated association of the three receptors with the mitotic chromatin. In the panels A and B, the left panels show the GFP fluorescence for the receptors, the middle panels show the Hoechst stained DNA of the corresponding cells while the right panels show the merged images of the two fluorescence. Panel C shows a comparison (in percent of the cells), with or without chromosomal association in presence or absence of the corresponding agonists. For the purpose of quantitation, 50–100 cells were scored per sample. The values represent means±s.d. of three separate experiments. Scale bar, 5 μm. S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73 63 nocodazole for 4–5 h. Following nocodazole treatment the cells were mildly as a homodimer or heterodimer with RXR under liganded as dislodged by swirling the media. The mitotic cell enriched media was centrifuged well as unliganded states [53]. Whether dimerization of these and resuspended in hypotonic buffer (0.25% (w/v) sodium citrate, 37.5 mM KCl). After incubation for 20 min. at 37 °C, cells were pelleted at 900 ×g for receptors in unliganded states could attain increased affinity to 10 min., re-suspended in 0.5 ml hypotonic buffer and 5 ml ME buffer (methanol/ chromatin without initiation of transcription is not clear. Also, at glacial acetic acid in ratio of 4:1) was added and mixed gently. Cells were the initial stages of these findings, it is difficult to assess if all repelleted and resuspended in fresh ME buffer. After 1 h incubation at room the cellular transcription factors encompass this behavior or it is temperature, cells were centrifuged as above and spread onto a coverslip and air- restricted to only some of the critical receptors executing un- dried. The cell spreads were then kept in blocking buffer and subjected to indirect immunofluorescence with anti-AR antibody as described above. known but dedicated cellular functions.

3.2. DHT-activated AR binds mitotic chromatin during all the 3. Results stages of cell division

3.1. Association of nuclear receptor superfamily members with Transcription modulation by AR is a function that it imparts the mitotic chromatin: is it a general phenomenon? in response to interaction with its physiological ligands. When in ligand-bound state AR resides in the nuclear compartment Recently, while resolving the controversial issue of sub- like all other members of the NR superfamily. However, in cellular localization of PXR we reported the first case of a unliganded state AR remains anchored to a number of cha- member of the NR superfamily that can associate with the perone proteins (Hsp70, Hsp90, immunophilins etc.) that help mitotic chromatin [45]. Whether liganded or not, PXR was in retaining the receptor primarily in the cytoplasmic compart- found to be a nuclear protein that constitutively associated with ment [26]. For the purpose of studying the behavior of AR the mitotic chromatin during all the stages of cell division. This during mitotic stages of normally proliferating cells, GFP-AR finding prompted us to undertake a systemic approach with was transiently expressed in COS-1 cells. As observed in no- some other members of the superfamily. For this purpose, other codazole arrested mitotic cells, in naturally dividing cells also, than PXR, we selected two well-characterized members of the the unliganded AR did not associate with the mitotic chromatin NR superfamily i.e. AR and ERα. It has been well established while DHT-bound receptor showed a clear chromatin binding that in unliganded state AR protein is localized in the cyto- during all the mitotic stages. Fig. 2 represents a typical profile of plasmic compartment while the ERα and PXR reside pre- unliganded and DHT-bound AR during interphase and in dif- dominantly in the nuclear compartment [9 and references cited ferent stages of mitosis. When cultured in steroid-free condi- therein]. For this purpose GFP-tagged receptors (AR, ERα, tions these mitotic stages clearly demonstrated total absence of PXR) were transiently expressed in COS-1 cells under ligand- AR association with the condensed mitotic chromatin (Fig. 2A). free culture conditions. For arresting the cells in mitosis no- As seen in Fig. 2B, DHT-bound AR was observed to be as- codazole was added to the culture media at the final con- sociated with mitotic chromatin during prophase, metaphase, centration of 0.1 μg/ml. The sets of experimental cells were anaphase, telophase and cytokinesis. In interphase nucleus, treated at optimally known concentrations of their respective agonist-activated AR has been previously reported to redis- ligands while the control cells were treated with vehicle alone tribute into fine structures that have been generally termed as (DMSO:ethanol 1:1). AR expressing cells were treated with “nuclear foci” [18,19,23]. Interestingly these “nuclear foci” 10− 8 M DHT, ERα expressing cells with 10− 8 Mof17β- appeared to co-migrate with the condensing chromatin and this estradiol while PXR expressing cells were incubated with dynamic movement was observable up to prophase and me- 20 μM rifampicin. Following treatments, the cells were imaged taphase of mitosis (Fig. 2B). However, after metaphase these and scored under fluorescence microscope. Typical intracellular “nuclear foci” could not be resolved as separate entity probably profiles of receptor distribution during unliganded (Fig. 1A) and due to extreme condensation of chromatin. Observation made liganded states (Fig. 1B) are shown with GFP fluorescence. above with live cell studies were further confirmed by bio- Corresponding cells showing chromosomal DNA stained with chemical fractionation of mitotic chromatin from nocodazole Hoechst and merged images of the two fluorescence are also arrested COS-1 cells transiently expressing GFP-AR in pre- shown. Fig. 1C shows the percent of mitotic cells revealing the sence or absence of DHT (Fig. 2C). The supernatant and the mitotic chromatin binding for unliganded and liganded recep- chromatin fractions from DHT-treated cells were analysed by tors. From the results it was evident that in liganded state all the western blotting using the AR polyclonal antibody described three receptors were capable of binding to the condensed mitotic herein. Results show that AR band is significantly more pro- chromatin. However, this behavior in mitotic cells differed minent in chromatin of DHT-treated cells as compared to the when these receptors were in unliganded state. In case of GFP- untreated control. In context to the studies conducted so far with AR, the unliganded receptor did not associate with the mitotic other relevant proteins, AR appears to be the first case of a chromatin. On the contrary, unliganded GFP-ERα was observed transcription factor whose association with the mitotic chroma- to be associated only partially (in ∼20% cells) while unliganded tin can be modulated by its natural ligand. A few of the other GFP-PXR in all the mitotic cells was seen to be consistently transcription factors (PXR, FOX1, zinc finger protein CTCF, associated with the mitotic chromatin. Interestingly, unlike AR, Runx2) showing association with the mitotic chromatin have ERα has been reported to exist as a homodimer even in absence been reported during last couple of years [41,43–45]. However, of its ligand [51,52]. PXR on the contrary is reported to exist their association with the mitotic chromatin has been observed 64 S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73

Fig. 2. Agonist-mediated binding of GFP-AR to the mitotic chromatin throughout the mitotic stages. COS-1 cells were transiently transfected with 500 ng of GFP-AR and incubated in steroid-free medium for 24 h. Following receptor expression the cells were treated with DMSO:ethanol (vehicle) for control cells or with 10− 8 Mof 5α-dihydrotestosterone (DHT). After incubation with DHT, GFP-AR that associated with mitotic chromosomes was recorded in live cells by fluorescence microscopy. Hoechst was used as a fluorescent dye to visualize corresponding nuclei/condensed chromosomes. Panel A shows unliganded AR that does not associate with the condensed chromosomes during mitosis. Panel B shows DHT-activated AR that binds to the condensed chromosome during all the stages of mitosis. In the “unliganded GFP-AR” and “GFP-AR+DHT” set of images, the left panels show the GFP-tagged AR, the middle panels show the Hoechst stained DNA of the corresponding cells while the right panel shows the merged images of the two fluorescence. The interphase cells and the mitotic stages of the naturally dividing cells are indicated. Panel C: To confirm the above observation, biochemical fractionation of mitotic chromatin was performed in COS-1 cells transiently expressing GFP- AR in unliganded and DHT-bound conditions as described in “Materials and methods”. The supernatant and chromatin-enriched fractions were analyzed by western blot using the anti-AR polyclonal antibody. Western blot shows the supernatant and corresponding mitotic chromatin fractions (chromatin fraction) in absence and in presence of DHT. As shown in figure GFP-AR is present in both the supernatant fractions in the absence and in the presence of DHT. While GFP-AR band is more prominent in the mitotic chromatin fraction of DHT-treated cells. A faint band is detected in the chromatin fraction of untreated cells. These results further confirm that GFP-AR binding to mitotic chromatin is DHT-dependent. Scale bar, 5 μm. to be constitutive and cannot be modulated by similar molecular immunodetection of untagged TBP in fixed cells failed to show events. existence of similar phenomenon [35]. Unavailability of an- tigenic epitope(s) by the antibody used was also suggested as 3.3. A comprehensive analysis to overcome artifacts a plausible reason for this failure. In this context, the limitations introduced during mitotic cell fixation procedures of chemical fixatives and permeabilization reagents were also highlighted by others when working with HMGB1 and HMGB2 Association of PXR with mitotic chromatin was an un- (high mobility group box) proteins [36]. anticipated finding [45] and similar behavior of AR and ERα For confirming our observations more stringently with un- during mitosis is somewhat intriguing. Though these studies tagged AR, we first purified full-length human AR and raised were conducted in living cells using GFP-tagged chimera, the polyclonal antibodies in rabbit against the whole protein. Ad- concern that GFP tagging to receptor can produce an un- vantages of using antibodies against full-length antigen were physiological receptor response needed to be experimentally perceived to be advantageous since such antibodies will access ascertained. In earlier instances, GFP-tagged TBP was found to multiple epitopes making immunodetection possible even if be associated with condensed mitotic chromatin but indirect some antigenic epitopes remain buried in the condensed S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73 65 chromatin. Other than LNCaP cells expressing a mutant AR, expected position of 110 kDa was clearly visible in both the cell another cell line stably transfected with wild type AR in HepG2 lines expressing AR (Fig. 3B). In LNCaP, a prostate cancer cell cells was prepared to authenticate our results obtained from line, we detected AR in varied conditions. In normally pro- living cells. In this context, Fig. 3A shows results for AR liferating cells DHT-bound AR was marginally at higher level immunodetection in cell line expressing untagged wild type than in unliganded or antiandrogen bound state. Similarly, in receptor with the antibody raised against full-length receptor. In nocadozole arrested cells the AR relative levels also varied with our experiments, we could specifically detect the unliganded ligand status. However, it was clear that AR levels in interphase AR in cytoplasm and DHT-bound receptor in the nucleus. The cells were maintained at higher levels than in cells arrested in antibodies generated were tested further by western blot ana- mitosis (Fig. 3C). These results are in agreement with the lysis using whole cell extracts from low AR expressing T47D findings that agonist-bound receptor is more stable in cellular cells and also in stably transfected HepG2 (Hep-AR) cells. milieu [54] and AR levels during mitosis are maintained at When compared to normal HepG2 cells (control) a band at the reduced levels [55]. Once the experimental tools were opti- mized we employed these in our forthcoming experiments. In search for a suitable protocol for working with mitotic chromatin we carried out a systematic and comparative analysis of some of the prevalent procedures by probing DHT-activated AR in LNCaP and Hep-AR cells transversing mitotic stages. Amongst the various protocols described under “Methods”, the one that gave the most optimum immunodetection results with condensed chromatin was with chilled methanol with 5% acetic acid (v/v) followed by permeabilization with 0.5% Triton X-100. Chilled ethanol with 5% (v/v) acetic acid used by another group in related studies with other transcription factor also gave com- parable results [41]. Representative results from three major fixation procedures evaluated in the present study are shown in Fig. 4. AR immunodetection results obtained from mitotic cells using two cell types expressing the mutant or wild type AR divulged that cells fixed with chilled absolute methanol or paraformaldehyde do not show significant AR bound to mitotic chromatin. Majority of the immunodetected AR protein ap- peared abrogated adjacent to mitotic chromatin with only some cells showing a minor fraction of AR associated with condensed chromatin. However, in our screening, cells fixed with chilled methanol with 5% acetic acid (v/v), revealed a consistent as- sociation of immunodetected AR with the condensed chro- matin. The efficacy of the fixation methods was further tested by fixing the COS-1 cells that expressed GFP-AR associated with mitotic chromatin in presence of DHT. Using different cell fix- Fig. 3. Preparation of polyclonal antibody against full-length AR and a cell line ation methods, only methanol–acetic acid appeared superior in stably expressing wild type AR. Prokaryotic expression of AR in E. coli BL21 and final purification with Ni–NTA His–Bind matrix was performed as retaining the receptor protein on mitotic chromatin (not shown). described under “Materials and methods”. Using AR antibody, stably transfected Finally, with methanol–acetic acid cell fixation procedure, Hep-AR cells were specifically examined for AR in hormone-free and hormone- mitotic stages of LNCaP and Hep-AR extending from prophase bound conditions (A). The unliganded AR was predominantly cytoplasmic to cytokinesis were produced and are represented in Fig. 5.Itis − ( DHT) while the hormone-bound receptor was nuclear (+DHT) indicating the evident that untagged AR bound to DHT associates with mitotic specificity of the antibody. Panel B shows the western blot analysis for AR constitutively expressed by HepG2 cells stably transfected with wild type AR chromatin throughout mitosis as was observed with the GFP- (Hep-AR) while the normal (untransfected) HepG2 cells do not show the tagged receptor in live cell experiments. corresponding AR band. Human breast cancer cell line T47D revealed the presence of endogenously expressed AR albeit at lower level. Molecular weight 3.4. Mitotic chromosomal spreads reveal a higher order markers by symbol M. Panel C shows the western blot analysis for the systematic distribution profile of AR on condensed chromatin expression level of endogenous AR by human prostate cancer cell line LNCaP in interphase and mitosis in unliganded and liganded conditions. LNCaP cells are arrested in mitosis by incubating with 500 ng/ml nocadozole for 20 h and treated In a whole cell undergoing mitosis, the chromosomal details with 10− 8 M DHT or 10−6 M of pure antiandrogen. Western blot analysis shows cannot be resolved as explicitly as is possible with chromosomal that AR expression is more in interphase cell extracts as compared to mitotic cell spreads. In a few related studies chromosomal spreads have extracts seen in (interphase versus mitotic phase lanes). Also, DHT-bound AR provided better details [37,41,42], and therefore, were taken (+H) level is more in LNCaP cells while lower in unliganded (−H) or antiandrogen-bound (+A) irrespective of cell cycle arrest. β-actin is served as a advantage of in the present study. In our experiments, chromo- protein loading control. Molecular weight markers are shown in lane M. Scale somal spreads were prepared from Hep-AR and LNCaP bar, 20 μm. cell lines after treatments with DHT. Following hormone 66 S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73

Fig. 4. Comparative analysis of cell fixation and permeabilization procedures for determining the optimum conditions in immunodetection of AR associated with mitotic chromatin. AR expressing LNCaP and stable Hep-AR cell lines were used for the immunodetection of agonist-activated receptor during mitosis. LNCaP and Hep-AR cells were cultured onto the glass cover slips, treated with 10−8 M DHT and allowed to incubate for 24 h. Subsequently, the cells were fixed with either chilled methanol alone (first row) or 4% paraformaldehyde (second row) or with chilled methanol +5% acetic acid (third row). AR was immuno-stained with the anti-AR primary antibody and a Cy3-conjugated secondary antibody (red), and DNA was stained with Hoechst (blue). The merged images show the colocalization of AR with mitotic chromatin. For each fixation procedure two representative cells are shown. Scale bar, 5 μm.

treatment, cells were processed for chromosomal spreads and rapidly reorganize into punctate pattern (termed “nuclear foci”) immunodetection with anti-AR antibodies as described under upon agonist binding [9,18,26]. From the studies in several “Methods”. Results from mitotic chromosomal spreads prepared laboratories it is evident now that only agonist-bound receptor similarly from the two cell lines are shown in Fig. 6.As having transcriptional competency could reorganize into “nu- expected, AR was clearly associated with all the mitotic clear foci” representing potential sites engaged in gene tran- chromatin. However, along the condensed chromatin the scription. Pure antagonist, though efficiently bind and translocate receptor is visibly distributed unevenly. Spherical regions of the cytoplasmic receptor into nuclear compartment, revealed high receptor density can be easily distinguished from the areas only homogeneous intra-nuclear pattern. Additionally, since of little or no binding (Fig. 6 inset magnified). Similar results on mixed agonist–antagonists have some innate agonist property, mitotic chromatin were also reported with transcription factor they also provided similar punctate pattern. This correlated well CTCF [41]. Interestingly, in three-dimensional image analysis of with transcriptional status of the receptor and was considered interphase nucleus, agonist-mediated “nuclear foci” formation indicative of an important regulatory process by numerous other by AR has been roughly estimated to be between 250 and 400 in reports. Recent studies with GFP-AR using FRAP revealed number [17]. Whether these spherical, receptor-enriched regions significant differences in AR mobility in interphase nucleus/ on chromosomes are representatives of pre-existing “nuclear chromatin depending on the nature of ligand i.e. agonists or foci” that were apparently engulfed during chromatin condensa- antagonists. Unliganded or antagonist-bound AR was observed tion, is a subject of speculation and warrants further to be significantly more mobile as compared to agonist-bound investigation. AR. It is postulated that antagonist-bound AR has extremely short residency on chromatin that is not enough to support 3.5. Association of GFP-AR with mitotic chromatin is transcriptional activation [6,20]. discernible with the agonistic or antagonistic nature of ligand(s) Keeping in view the above background, we selected DHT as the most preferred AR agonist along with some other well- Over the last few years a number of laboratories have shown studied antiandrogens that are used in treatment of androgen- that intranuclear homogenous distribution of steroid receptors dependent prostate cancer. These antiandrogenic drugs have S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73 67

Fig. 5. Association of untagged agonist-activated AR with the mitotic chromatin performed under optimal cell fixation and permeabilization procedure. Indirect immunofluorescence experiments were carried out on LNCaP and Hep-AR stable cells fixed with chilled methanol +5% acetic acid. Following immunodetection, different mitotic stages that included prophase, metaphase, anaphase, telophase and cytokinesis in normally dividing cell were imaged and presented above. AR was stained with the primary anti-AR antibody and a Cy3-conjugated secondary antibody (red), and DNA was stained with Hoechst (blue). The merged images show the colocalization of AR and mitotic chromatin. Scale bar, 5 μm. been previously reported to bind and translocate the unliganded association of AR with mitotic chromatin will also distinguish cytoplasmic AR to the nuclear compartment. Of these, three between agonist and antagonist property of these ligands was drugs i.e. casodex, hydroxyflutamide and nilutamide are pure revisited in live cell experiments using GFP-AR. In this attempt, antagonist whereas cyproterone acetate and spironolactone other than DHT, we tested all the above-mentioned therapeutic are of mixed agonist–antagonist nature. Whether dynamic antiandrogens. As expected, the two mixed agonist–antagonist

Fig. 6. Distribution of immunodetected AR on mitotic chromosome spreads. Mitotic chromosome spreads were prepared with DHT-activated prostate cancer cell line LNCaP and stably transfected Hep-AR cell line. Indirect immunofluorescence was carried out with the anti-AR antibody and a Cy3-conjugated secondary antibody (Red, left panel). DNA was stained with Hoechst (Blue, middle panel). The merged images show the colocalization of AR and DNA (right panel). A portion of the merged image from Hep-AR cell line expressing wild type AR is highlighted and indicated with an arrow. The magnified merged image on extreme right exhibits multiple areas with intense focal distribution of AR on mitotic chromatin. These higher order structures were visible in both the cell lines represented here. Scale bar, 5 μm. 68 .Kmre l iciiae ipyiaAt 73(08 59 (2008) 1783 Acta Biophysica et Biochimica / al. et Kumar S. – 73

Fig. 7. GFP-AR activated by pure or partial agonists reveal formation of “nuclear foci” during interphase and chromosomal docking during mitosis. GFP-AR was expressed in COS-1 cells and challenged with different ligands including drugs used in prostate cancer treatment. In interphase cells, unliganded GFP-AR remains partitioned in the cytoplasmic compartment (A row 1). Upon treatment with pure agonist (DHT) or mixed agonist–antagonist (CA, spironolactone) GFP-AR enters nucleus and forms “nuclear foci” (A, row 2 to 4) and also associates with mitotic chromatin (B, row 2 to 4). However, with pure antagonists (casodex, nilutamide, hydroxyflutamide), the receptor is homogeneously distributed, marked by absence of “nuclear foci” in the nucleoplasm (A, row 5 to 7), and, does not associate with chromatin during mitosis (B, row 5 to 7). The results from GFP-AR shown in panels A and B were confirmed by immunocytology in a cell line stably expressing untagged wild type AR and are shown similarly arraged in panels C and D. The results from GFP-tagged AR were in agreement with immunodetected wild type AR. In panels A to D, the left panels show AR, the middle panels show the nuclear/chromatin staining with Hoechst and the right panels exhibit the merged images of the corresponding fluorescent cells. CA = cyproterone acetate. Scale bar, 5 μm. S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73 69

agonist-dependent and could be correlated with the transcrip- tionally active state of the receptor [9,18]. With respect to AR and ER, studies primarily based on live cell imaging, have already indicated that nuclear receptor co-activators like GRIP1, TIF2, SRC1, SRC3 and etc. co-localize in “nuclear foci” with the agonist-activated but not with pure antagonist-bound re- ceptor [21–25]. Essentially, pure antagonist-bound nuclear receptors have never been reported to form “nuclear foci” but were observed to display a homogeneous nucleoplasmic recep- tor distribution. Therefore, in this context, it was interesting to explore whether co-activators are colocalized with AR on con- densed chromatin in mitotic cells. Results obtained from this query are presented in Fig. 8. COS-1 cells were co-transfected with RFP-AR and the co-activator GFP-GRIP1 and cultured for protein expression in steroid-free media. After expression period, the relevant set of cells were treated with DHT

Fig. 8. Agonist-activated AR associates with mitotic chromatin but does not recruit coactivator GRIP1. COS-1 cells were co-transfected with RFP-AR and GFP-GRIP1 and cultured in steroid-free culture medium for receptor expression. In the interphase nuclei, unliganded RFP-AR is visible as a predominantly cytoplasmic protein while GFP-GRIP1 is seen as nuclear protein (untreated interphase row). During mitosis, both the unliganded RFP-AR as well as GFP- GRIP1 do not associate with the mitotic chromatin (untreated metaphase row). Upon treatment with DHT, RFP-AR enters nucleus and forms nuclear foci and colocalizes with GFP-GRIP1 in interphase cells (DHT-treated interphase row). However, in mitotic cells, agonist-activated AR associates with mitotic chro- matin but fails to recruit GFP-GRIP1 (DHT treated metaphase row). From left, panel 1 shows the RFP fluorescence for AR while panel two shows the GFP fluorescence for GRIP1. Panel 3 illustrates the merged images from the two fluorescence in panel 1 and 2. Panel 4 exhibits the Hoechst stained nuclei/mitotic chromatin of the corresponding cells.

(cyproterone acetate, spironolactone) showed AR distribution into “nuclear foci” while the pure antagonists (casodex, nilutamide, hydroxyflutamide) showed homogeneous AR distribution in the interphase nucleoplasm (Fig. 7A). Interest- ingly, AR showed binding to mitotic chromatin only with the drugs (cyproterone acetate, spironolactone) that exhibited induction of “nuclear foci” (Fig. 7B). On the contrary, association of AR with the mitotic chromatin was abrogated when bound to pure antagonists. The above results obtained from GFP-tagged AR, transiently expressed in COS-1, were authenticated in a cell line (Hep-AR) that stably expresses untagged wild type AR analogous to endogenous levels in LNCaP expressing the mutant AR. The effect of different ligands on AR binding to chromatin, obtained from live cells studies were in agreement with the results observed with cell line expressing wild type AR (Fig. 7C and D). These results conclusively established that of the five therapeutic drugs only Fig. 9. Agonist-mediated binding of AR onto the mitotic chromatin of living those having pure or partial agonistic properties could inflict cells is neither impeded nor prevented by coexpressed TBP or PXR. COS-1 cells were co-transfected with GFP-AR and RFP-TBP (A) or RFP-PXR (B) and chromosomal binding of the receptor. cultured in steroid-free culture medium for receptor expression. In the interphase cells, unliganded GFP-AR is visible as a predominantly cytoplasmic protein 3.6. GRIP1 that colocalizes with AR during interphase is while RFP-TBP and RFP-PXR are seen as nuclear proteins. In mitotic meta- aborted from mitotic chromatin during cell division phase, unliganded GFP-AR did not associate while RFP-TBP and RFP-PXR constitutively associated with the mitotic chromatin. Though the mitotic chromatin are preoccupied with RFP-TBP or RFP-PXR, DHT-activated GFP- Recent studies have convincingly shown that some of the AR efficiently occupied the binding sites on the mitotic chromatin implying that critical co-activators of nuclear receptors are recruited into the the association with chromatin is apparently non-competitive in nature. Scale “nuclear foci” of the interphase cells. Formation of these foci is bar, 5 μm. 70 S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73

(10− 8 M) according to scheme described in the figure. It was the chromosomal docking sites are distinct or specific for these evident that under steroid-free culture condition RFP-AR was transcription factors and cannot be competed out by each other. expressed as a cytoplasmic protein while GFP-GRIP1 resided predominantly in the nuclear compartment of the interphase 4. Discussion cell. Under similar culture conditions in the mitotic phase none of the two proteins associated with the mitotic chromatin. During the last few years, several key components of gene However, in DHT-activated state, both, RFP-AR and GFP- transcriptional machinery including transcription factors that lie GRIP1 were observed to be predominantly residing in the within or outside the NR superfamily have been reported to nucleus. In agreement with the previous reports, the merged undertake a dynamic association with the mitotic chromatin images of the red and green fluorescence showed appearance of [41,43–45]. This dynamic docking of transcription factors onto significant level of yellow fluorescence at the “nuclear foci” the condensed mitotic chromatin should have important phys- which is indicative of co-localization of two proteins. However, iological significance. In this report, we have documented ana- results observed from DHT-activated mitotic chromatin lysis of a hitherto unreported dynamic behavior of AR, an revealed that the two proteins that were reported to co-localize extensively studied member of this superfamily. A significant in “nuclear foci” of interphase cells reveal dissimilar behavior. amount of recent data showing association of proteins to mitotic While the DHT-activated AR associated with the mitotic chromatin in living cells has been generated using expression of chromatin, the co-activator GRIP1 was not recruited along GFP-tagged chimeric proteins. Notably, in these experiments with the AR. This implies that the formation of transcriptionally there have been instances when the observations made in live competent complex is an event specific to interphase nucleus cells could not be reproduced in fixed cells used for indirect and not to mitotic chromatin. Recently, another co-activator immunodetection [35,36]. In agreement with Pallier et al. [36], SRC-3 has been similarly observed to be co-localized with ER we also observed that use of chemical fixatives could be a in interphase cells [25]. In our preliminary study, GFP-SRC-3 determining factor in success or failure of immunodetection also did not exhibit any association with the mitotic chromo- results obtained with mitotic chromatin. These variations can somes (not shown). also be protein specific since in our experience PXR association with mitotic chromatin was obtainable with most of the cell 3.7. Mitotic chromatin pre-occupied by other over-expressed fixation procedures while AR gave optimal results with chilled transcription factors does not prevent subsequent binding of methanol plus 5% acetic acid. These differences may be attri- agonist-activated AR buted, at least in part, to the affinity with which the observed protein binds to the mitotic chromatin. Nonetheless, this is a fine TBP and PXR are among the first few proteins that are example of experimental artifacts that may be introduced by reported to bind to mitotic chromatin and are also involved in some of the cell fixation procedures and also justifies the ge- gene transcription. The finding that DHT-activated AR also neral superiority of live cell studies over the fixed cells. Ad- associates with the mitotic chromatin raised the pertinent ques- ditionally, the limitations of the cell fixation procedures appear tion if the chromatin sites for receptor docking are specific or to have contributed in delaying the observation of this sig- general in nature for each transcription factor. Since docking of nificant phenomenon. Notably, in normal cell culture conditions AR could be modulated by DHT, it was possible to preoccupy mitotic cells are not present in large numbers and the adherent the chromosomal docking sites constitutively by another tran- mammalian cells undergoing mitotic division will generally scription factor and then attempt to load the AR by DHT acti- become round with nuclear structure apparently visible as dis- vation. If the docking sites are general in nature and stay pre- integrated. Such cells, therefore, can be easily confused with the saturated by another transcription factor, DHT-activated AR will apoptotic cells, thereby, escaping attention of the observer be subsequently prevented from chromosomal binding. Fig. 9 during protein localization. shows results from live cell imaging where either RFP-tagged Employing GFP-tagged AR and ERα we observed agonist- TBP (RFP-TBP) or RFP-tagged PXR (RFP-PXR) is co- mediated docking of these receptors onto the mitotic chromatin expressed along with GFP-tagged AR. Fluorescent distribution during cell division. In our extended studies using GFP-AR, we in interphase cells shows that TBP and PXR are nuclear proteins have shown a distinctly differential response to androgenic and while unliganded AR is compartmentalized in the cytoplasm. pure anti-androgenic therapeutic drugs used in targeting AR. Results from mitotic cells, co-expressing RFP-TBP and GFP- When bound to drugs with intrinsically absolute or partial ag- AR in steroid-free culture conditions, showed that TBP remains onistic properties, AR associated with the mitotic chromatin. constitutively associated with the mitotic chromatin while AR Conversely, the pure antagonists that are previously reported to remains unbound. Addition of DHT to culture media induced bind and subsequently translocate unliganded AR from cyto- efficient AR binding to the mitotic chromatin without any ap- plasm to nuclear compartment do not provoke such association parent hindrance (Fig. 9A). We further conducted similar ex- with the mitotic chromatin. Conceivably, with similar drugs, a periment by co-expressing RFP-PXR and GFP-AR since former correlation to AR dynamic response analogous to formation of is also a member of the NR superfamily (Fig. 9B). Similar to the “nuclear foci” in interphase nuclei is also evident on mitotic results obtained with TBP, PXR that pre-occupied the mitotic chromatin. Only those drugs that generated “nuclear foci” were chromatin did not inhibit or impede the DHT-activated AR potent in imparting docking of AR onto mitotic chromatin while docking onto the mitotic chromatin. These results suggested that others that do not induce “nuclear foci” abrogated receptor S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73 71 association with the mitotic chromatin. These observations were the chromatin binding of AR. Such competitive interaction further substantiated by immunocytology of untagged wild type between the non-site-specific chromatin binding factors are AR in a cell line stably expressing the receptor near endogenous well described in case of global non-specific binding factor H1 levels. where its interaction on chromatin is modulated by a network Previous studies of GFP-AR, using FRAP analysis, revealed of other non-site specific competitors such as HMGs which significant differences in dynamics of AR-chromatin association lead to reduction in the association of H1 with chromatin [58]. with unliganded/antagonist-bound AR when compared with However, in our case results obtained with TBP and PXR that agonist bound AR [6]. The rate of dissociation of pure anta- pre-occupied the mitotic chromatin did not influence the gonists (casodex, hydroxyflutamide) bound AR from chromatin ligand-mediated docking of AR onto mitotic chromatin. This was significantly higher than agonist (DHT) and partial agonists study has important significance since specificity and non- (cyperoterone acetate) bound AR [6,20]. These results support specificity of interaction on mitotic chromatin could be a our observation of ligand-dependent differential behaviour of relevant query. In this context, our observations suggest that AR on mitotic chromatin. The agonist/partial agonists bound AR these factors associate on distinct binding sites on to the mitotic associates more stably with interphase chromatin. On the con- chromatin. trary, in the presence of antagonists the resident time of AR on The present study has significant relevance in the perspective chromatin is much less as compared to agonist. Whether similar of the recently emerging concepts of “epigenetic marking” and ligand-dependent phenomena are also occurring on the mitotic “gene book-marking” [59,60]. “Epigenetic regulatory processes” chromatin platform needs to be established concretely. Alter- govern chromatin structure through covalent modification of natively, ligand-dependent N/C-termini interactions of AR may histone substrates via methylation, acetylation, phosphorylation be involved in inflicting differences in mitotic chromatin bind- or ubiquitination processes [61,62]. In simple analogy to these ing. In this context, N/C-terminal interactions have already been epigenetic regulatory processes, “gene book-marking” by ligand- reported to stabilize receptor association with interphase chro- activated transcription factors can be broadly defined as a mech- matin [5,6]. It is believed that only AR agonists lead to N/C anism used in transmitting the pattern of active gene transcription interactions whereas AR antagonists prevent these intramole- from interphase nucleus, via mitosis, to the emerging daughter cular interactions [4]. Future studies are expected to unravel cells, thereby, ensuring inheritance of progenitor's active gene the AR dynamics/mobility on mitotic chromatin and also ex- transcripts to the progeny [45,60,63]. The cyclic process of plain how agonistic–antagonistic nature of different ligands al- transmitting the pre-existing transcription status from the un- ters receptor relationship with mitotic chromatin. condensed chromatin of interphase cell to condensed chromatin A remarkable difference in behavior of agonist-activated AR during mitosis and then final re-emergence of active transcription was with recruitment of co-activators. Recruitment of ligand- machinery during return to de-condensation state may be a activated transcription factors and co-activators in the “nuclear natural process. With respect to ligand-modulated transcription foci” of interphase cells have been documented in significant factors this feed-forward transmission, by receptors themselves, details [21–25]. Contrary to observations in interphase nuclei over the generations, may also encounter active gene pattern we found that AR though associates with the mitotic chromatin modifications depending on appearance and disappearance of in agonist-dependent manner, the co-activator(s) stay aborted. physiological ligands. However, in view of the “hit and run” This is suggestive of a transition in behavior of agonist-bound mechanism of action of steroid/nuclear receptors, it remains to be transcription factor from being transcriptionally active (during investigated if the nature of ligand-activated transcription factors interphase) to transcriptionally inactive or silenced (during mi- with mitotic chromatin is as dynamic as during interphase or is tosis). In this perspective, amongst the previously reported relatively static [6,27]. proteins, AR is first example of a transcription factor whose In some of the human malignancies (like prostate or breast association on mitotic chromatin platform can be modulated in a cancer) the involvement of some of the members of ligand- ligand-specific manner. If found to be a general phenomenon, modulated transcription factors (primarily AR and ERα)is this dynamic property of steroid/nuclear receptors by therapeu- physiologically and clinically important. Aberrant behavior or tic ligands may provide a composite screening protocol to some dysfunction of receptor signaling is known to lead to some of of the emerging technologies that are being forwarded for these human ailments. Testimony towards importance of these analyzing newer drugs and ligands [50,56,57]. Also, it will be receptors as therapeutic targets is evident by the fact that a large interesting to determine if these properties could be exploited to number of commonly prescribed drugs act via nuclear receptors. establish the nature of molecular interventions that confers most However, in prostate and breast cancers treatments with anti- desirable therapeutic benefits. hormone ligands have provided only interim relief, but sub- We have shown that agonist-mediated binding of AR onto sequently, lead to re-emergence of advanced stages of disease the mitotic chromatin of living cells is neither competed nor progression when these drugs become ineffective. It is en- prevented by co-expression of other mitotic chromatin binding visaged that silencing of transcription function of targeted proteins. These results suggested that the chromosomal dock- transcription factors by anti-hormone drugs over long period ing sites are distinct or specific for these transcription factors may alter the natural gene book-marking profile, thereby, in- and cannot be competed out by each other. If these factors flicting upon cells to undertake alternative cell survival and were bound to the non-specific sites, an increase in cellular proliferation strategies. Hence, during anti-hormone therapy, abundance of other non-site-specific competitor would affect active gene transcripts offered to progeny from its progenitor for 72 S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73 inheritance of active gene transcripts will not be preserved [13] P. Zhu, S.H. Baek, E.M. Bourk, K.A. Ohgi, I. Garcia-Bassets, H. Sanjo, S. which may lead to either cell elimination or survival as ad- Akira, P.F. Kotol, C.K. Glass, M.G. Rosenfeld, D.W. Rose, Macrophage/ cancer cell interactions mediate hormone resistance by a nuclear receptor vanced progression of a malignancy. Though speculative, it is derepression pathway, Cell 124 (2006) 615–629. possible that in absence of such molecular intervention, for [14] C.E. Bohl, W. Gao, D.D. Miller, C.E. Bell, J.T. Dalton, Structural basis for example, normal transmission of inheritable active gene tran- antagonism and resistance of bicalutamide in prostate cancer, Proc. Natl. scripts carried during mitosis help a “prostate or breast cell” to Acad. Sci. U. S. A. 102 (2005) 6201–6206. emerge as “prostate or breast cell” rather than a new genre. [15] J.P.Gaddipati, D.G. McLeod, H.B. Heidenberg, I.A. Sesterhenn, M.J. Finger, J.W. Moul, S. Srivastava, Frequent detection of codon 877 mutation in the In conclusion, we highlight an example of agonist-mediated androgen receptor gene in advanced prostate cancers, Cancer Res. 54 (1994) docking of a member of the ligand-activated transcription factors 2861–2864. onto the mitotic chromatin, in both, living and fixed cells. Since [16] A. Muller-Taubenberger, Application of fluorescent protein tags as the association of AR onto the mitotic chromatin platform obeys reporters in live-cell imaging studies, Methods Mol. Biol. 346 (2006) – the intrinsic mode of action of natural and clinical ligands, 229 246. [17] A. Tomura, K. Goto, H. Morinaga, M. Nomura, T. Okabe, T. Yanase, R. therefore, it may offer a novel paradigm for understanding the Takayanagi, H. Nawata, The subnuclear three-dimensional image analysis concerted action of ligand–receptor interactions in perspective of androgen receptor fused to green fluorescence protein, J. Biol. Chem. of normal physiology and clinical background. 276 (2001) 28395–28401. [18] B.E. Black, B.M. Paschal, Intranuclear organization and function of the – Acknowledgments androgen receptor, Trends Endocrinol. Metab. 15 (2004) 411 417. [19] R.K. Tyagi, Y. Lavrovsky, S.C. Ahn, C.S. Song, B. Chatterjee, A.K. Roy, Dynamics of intracellular movement and nucleocytoplasmic recycling of the The research work presented in this paper was financially ligand-activated androgen receptor in living cells, Mol. Endocrinol. 14 (2000) supported by research grants to RKT from the Council of 1162–1174. Scientific and Industrial Research [CSIR grant nos. 37(1249)/ [20] P. Farla, R. Hersmus, J. Trapman, A.B. Houtsmuller, Antiandrogens pre- 06/EMR-II and 37(1146)/03/EMR-II], India and partially by vent stable DNA-binding of the androgen receptor, J. Cell Sci. 118 (2005) 4187–4198. Indian Council of Medical Research (ICMR). Sanjay Kumar, [21] U. Karvonen, O.A. Janne, J.J. Palvimo, Pure antiandrogens disrupt the Nagendra K. Chaturvedi and Subodh Kumar acknowledge recruitment of coactivator GRIP1 to colocalize with androgen receptor in CSIR for the award of Research Fellowships. nuclei, FEBS Lett. 523 (2002) 43–47. [22] M. Saitoh, R. Takayanagi, K. Goto, A. Fukamizu, A. Tomura, T. Yanase, References H. Nawata, The presence of both the amino- and carboxyl-terminal do- mains in the AR is essential for the completion of a transcriptionally active form with coactivators and intranuclear compartmentalization common to [1] S. Hirawat, D.R. Budman, K. Willi, The androgen receptor: structure, the steroid hormone receptors: a three-dimensional imaging study, Mol. mutation and antiandrogens, Cancer Investig. 21 (2003) 400–417. Endocrinol. 16 (2002) 694–706. [2] C.A. Heinlein, C. Chang, Androgen receptor in prostate cancer, Endocr. [23] B.E. Black, M.J. Vitto, D. Gioeli, A. Spencer, N. Afshar, M.R. Conaway, Rev. 25 (2004) 276–308. M.J. Weber, B.M. Paschal, Transient, ligand-dependent arrest of the an- [3] C.L. Bevan, Androgen receptor in prostate cancer: cause or cure? Trends drogen receptor in subnuclear foci alters phosphorylation and coactivator Endocrinol. Metab. 16 (2005) 395–397. interactions, Mol. Endocrinol. 18 (2004) 834–850. [4] H. Dotzlaw, U. Moehren, S. Mink, A.C. Cato, J.A. Iniguez-Lluhi, A. [24] L.N. Song, E.P. Gelmann, Interaction of beta-catenin and TIF2/GRIP1 in Baniahmad, The amino terminus of the human AR is target for core- transcriptional activation by the androgen receptor, J. Biol. Chem. 280 (2005) pressor action and antihormone agonism, Mol. Endocrinol. 16 (2002) 37853–37867. 661–673. [25] L. Amazit, L. Pasini, A.T. Szafran, V. Berno, R.C. Wu, M. Mielke, E.D. [5] J. Li, J. Fu, C. Toumazou, H.G. Yoon, J. Wong, A role of the amino- Jones, M.G. Mancini, C.A. Hinojos, B.W. O'malley, M.A. Mancini, Re- terminal (N) and carboxyl-terminal (C) interaction in binding of androgen gulation of SRC-3 intercompartmental dynamics by estrogen receptor and receptor to chromatin, Mol. Endocrinol 20 (2006) 776–785. phosphorylation, Mol. Cell Biol. 27 (2007) 6913–6932. [6] T.I. Klokk, P. Kurys, C. Elbi, A.K. Nagaich, A. Hendarwanto, T. [26] A. Griekspoor, W. Zwart, J. Neefjes, R. Michalides, Visualizing the action Slagsvold, C.Y. Chang, G.L. Hager, F. Saatcioglu, Ligand-specific dyna- of steroid hormone receptors in living cells, Nucl. Recept. Signal. 5 (2007) mics of the androgen receptor at its response element in living cells, Mol. e003. Cell Biol. 27 (2007) 1823–1843. [27] J.G. McNally, W.G. Muller, D. Walker, R. Wolford, G.L. Hager, The [7] M.E. van Royen, S.M. Cunha, M.C. Brink, K.A. Mattern, A.L. Nigg, H.J. glucocorticoid receptor: rapid exchange with regulatory sites in living Dubbink, P.J. Verschure, J. Trapman, A.B. Houtsmuller, Compartmenta- cells, Science 287 (2000) 1262–1265. lization of androgen receptor protein–protein interactions in living cells, [28] G. Li, G. Sudlow, A.S. Belmont, Interphase cell cycle dynamics of a late- J. Cell Biol. 177 (2007) 63–72. replicating, heterochromatic homogeneously staining region: precise chor- [8] C.A. Heinlein, C. Chang, Androgen receptor (AR) coregulators: an over- eography of condensation/decondensation and nuclear positioning, J. Cell view, Endocr. Rev. 23 (2002) 175–200. Biol. 140 (1998) 975–989. [9] S. Kumar, M. Saradhi, N.K. Chaturvedi, R.K. Tyagi, Intracellular [29] M.A. Martinez-Balbas, A. Dey, S.K. Rabindran, K. Ozato, C. Wu, Dis- localization and nucleocytoplasmic trafficking of steroid receptors: an placement of sequence-specific transcription factors from mitotic chroma- overview, Mol. Cell. Endocrinol. 246 (2006) 147–156. tin, Cell 83 (1995) 29–38. [10] C. Bai, A. Schmidt, L.P. Freedman, Steroid hormone receptors and drug [30] J.M. Gottesfeld, D.J. Forbes, Mitotic repression of the transcriptional discovery: therapeutic opportunities and assay designs, Assay Drug Dev. machinery, Trends Biochem. Sci. 22 (1997) 197–202. Technol. 1 (2003) 843–852. [31] S. Sif, P.T. Stukenberg, M.W. Kirschner, R.E. Kingston, Mitotic inacti- [11] H. Miyamoto, E.M. Messing, C. Chang, Androgen deprivation therapy for vation of a human SWI/SNF chromatin remodeling complex, Genes Dev. prostate cancer: current status and future prospects, Prostate 61 (2004) 12 (1998) 2842–2851. 332–353. [32] S. Sciortino, A. Gurtner, I. Manni, G. Fontemaggi, A. Dey, A. Sacchi, K. [12] Y. Shang, M. Myers, M. Brown, Formation of the androgen receptor tran- Ozato, G. Piaggio, The cyclin B1 gene is actively transcribed during scription complex, Mol. Cell 9 (2002) 601–610. mitosis in HeLa cells, EMBO Rep. 2 (2001) 1018–1023. S. Kumar et al. / Biochimica et Biophysica Acta 1783 (2008) 59–73 73

[33] M.J. Kruhlak, M.J. Hendzel, W. Fischle, N.R. Bertos, S. Hameed, X.J. [47] C.D. Chen, D.S. Welsbie, C. Tran, S.H. Baek, R. Chen, R. Vessella, M.G. Yang, E. Verdin, D.P. Bazett-Jones, Regulation of global acetylation in Rosenfeld, C.L. Sawyers, Molecular determinants of resistance to an- mitosis through loss of histone acetyltransferases and deacetylases from tiandrogen therapy, Nat. Med. 10 (2004) 33–39. chromatin, J. Biol. Chem. 276 (2001) 38307–38319. [48] H. Poukka, U. Karvonen, N. Yoshikawa, H. Tanaka, J.J. Palvimo, O.A. [34] S. Dovat, T. Ronni, D. Russell, R. Ferrini, B.S. Cobb, S.T. Smale, A Janne, The RING finger protein SNURF modulates nuclear trafficking of common mechanism for mitotic inactivation of C2H2 zinc finger DNA- the androgen receptor, J. Cell Sci. 113 (2000) 2991–3001. binding domains, Genes Dev. 16 (2002) 2985–2990. [49] M. Saradhi, B. Krishna, G. Mukhopadhyay, R.K. Tyagi, Purification of full- [35] D. Chen, C.S. Hinkley, R.W. Henry, S. Huang, TBP dynamics in living length human Pregnane and Xenobiotic receptor: polyclonal antibody pre- human cells: constitutive association of TBP with mitotic chromosomes, paration for immunological characterization, Cell Res. 15 (2005) 785–795. Mol. Biol. Cell 13 (2002) 276–284. [50] V. Berno, C.A. Hinojos, L. Amazit, A.T. Szafran, M.A. Mancini, High- [36] C. Pallier, P. Scaffidi, S. Chopineau-Proust, A. Agresti, P. Nordmann, M.E. resolution, high-throughput microscopy analyses of nuclear receptor and Bianchi, V. Marechal, Association of chromatin proteins high mobility coregulator function, Methods Enzymol. 414 (2006) 188–210. group box (HMGB) 1 and HMGB 2 with mitotic chromosomes, Mol. Biol. [51] M. Salomonsson, J. Haggblad, B.W. O'Malley, G.M. Sitbon, The human Cell 14 (2003) 3414–3426. estrogen receptor hormone binding domain dimerizes independently of [37] M. Harrer, H. Luhrs, M. Bustin, U. Scheer, R. Hock, Dynamic interaction ligand activation, J. Steroid Biochem. Mol. Biol. 48 (1994) 447–452. of HMGA1a proteins with chromatin, J. Cell Sci. 117 (2004) 3459–3471. [52] F. Schaufele, X. Carbonell, M. Guerbadot, S. Borngraeber, M.S. Chapman, [38] H. Xing, D.C. Wilkerson, C.N. Mayhew, E.J. Lubert, H.S. Skaggs, M.L. A.A. Ma, J.N. Miner, M.I. Diamond, The structural basis of androgen Goodson, Y. Hong, O.K. Park-Sarge, K.D. Sarge, Mechanism of hsp70i receptor activation: intramolecular and intermolecular amino–carboxy in- gene bookmarking, Science 307 (2005) 421–423. teractions, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 9802–9807. [39] D. Chen, M. Dundr, C. Wang, A. Leung, A. Lamond, T. Misteli, S. Huang, [53] S.M. Noble, V.E. Carnahan, L.B. Moore, T. Luntz, H. Wang, O.R. Ittoop, J.B. Condensed mitotic chromatin is accessible to transcription factors and Stimmel, P.R. Davis-Searles, R.E. Watkins, G.B. Wisely, E. LeCluyse, A. chromatin structural proteins, J. Cell Biol. 168 (2005) 41–54. Tripathy, D.P. McDonnel, M.R. Redinbo, Human PXR forms a tryptophan [40] M. Kawagishi, T. Akashi, A. Kikuchi, Dynamic association of topo- zipper-mediated homodimer, Biochemistry 45 (2006) 8579–8589. isomerase II to the mitotic chromosomes in live cells of Aspergillus [54] T. Furutani, T. Watanabe, K. Tanimoto, T. Hashimoto, H. Koutoku, M. nidulans, Biochem. Biophys. Res. Commun. 334 (2005) 324–332. Kudoh, Y. Shimizu, S. Kato, H. Shikama, Stabilization of androgen receptor [41] L.J. Burke, R. Zhang, M. Bartkuhn, V.K. Tiwari, G. Tavoosidana, S. protein is induced by agonist, not by antagonists, Biochem. Biophys. Res. Kurukuti, C. Weth, J. Leers, N. Galjart, R. Ohlsson, R. Renkawitz, CTCF Commun. 294 (2002) 779–784. binding and higher order chromatin structure of the H19 locus are [55] I.V. Litvinov, D.J. Vander Griend, L. Antony, S. Dalrymple, A.M. De maintained in mitotic chromatin, EMBO J. 24 (2005) 3291–3300. Marzo, C.G. Drake, J.T. Isaacs, Androgen receptor as a licensing factor for [42] C. Das, K. Hizume, K. Batta, B.R. Kumar, S.S. Gadad, S. Ganguly, S. DNA replication in androgen-sensitive prostate cancer cells, Proc. Natl. Lorain, A. Verreault, P.P. Sadhale, K. Takeyasu, T.K. Kundu, Transcrip- Acad. Sci. U. S. A. 103 (2006) 15085–15090. tional coactivator PC4, a chromatin-associated protein, induces chromatin [56] R.K. Tyagi, Dynamics of subcellular compartmentalization of steroid condensation, Mol. Cell. Biol. 26 (2006) 8303–8315. receptors in living cell as a strategic screening method to determine the [43] J. Yan, L. Xu, G. Crawford, Z. Wang, S.M. Burgess, The Forkhead biological impact of suspected endocrine disruptors, Med. Hypotheses transcription factor FoxI1 remains bound to condensed mitotic chromo- 60 (2003) 501–504. somes and stably remodels chromatin structure, Mol. Cell Biol. 26 (2006) [57] E.D. Martinez, G.V. Rayasam, A.B. Dull, D.A. Walker, G.L. Hager, An 155–168. estrogen receptor chimera senses ligands by nuclear translocation, J. Steroid [44] D.W. Young, M.Q. Hassan, X-Q. Yang, M. Galindo, A. Javed, S.K. Zaidi, P. Biochem. Mol. Biol. 97 (2005) 307–321. Furcinitti, D. Lapointe, M. Montecino, J.B. Lian, J.L. Stein, A.J. van [58] F. Catez, T. Ueda, M. Bustin, Determinants of histone H1 mobility and Wijnen, G.S. Stein, Mitotic retention of gene expression patterns by the cell chromatin binding in living cells, Nat. Struct. Mol. Biol. 13 (2006) 305–310. fate-determining transcription factor Runx2, Proc. Natl. Acad. Sci. U. S. A. [59] A.D. Goldberg, C.D. Allis, E. Bernstein, Epigenetics: a landscape takes 104 (2007) 3189–3194. shape, Cell 128 (2007) 635–638. [45] M. Saradhi, A. Sengupta, G. Mukhopadhyay, R.K. Tyagi, Pregnane and [60] K.D. Sarge, O.K. Park-Sarge, Gene bookmarking: keeping the pages open, Xenobiotic receptor (PXR) resides predominantly in the nuclear compart- Trends Biochem. Sci. 30 (2005) 605–610. ment of the interphase cell and associates with the condensed chromo- [61] T. Kouzarides, Chromatin modifications and their function, Cell 128 (2007) somes during mitosis, Biochim. Biophys. Acta 1746 (2005) 85–94. 693–705. [46] R. Clarke, M.C. Liu, K.B. Bouker, Z. Gu, R.Y. Lee, Y. Zhu, T.C. Skaar, B. [62] B. Li, M. Carey, J.L. Workman, The role of chromatin during transcription, Gomez, K. O'Brien, Y. Wang, L.A. Hilakivi-Clarke, Antiestrogen resis- Cell 128 (2007) 707–719. tance in breast cancer and the role of estrogen receptor signaling, Oncogene [63] G.L. Zhou, D.P.Liu, C.C. Liang, Memory mechanisms of active transcription 22 (2003) 7316–7339. during cell division, Bioessays 27 (2005) 1239–1245.