| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 33, 30605-30613, August 15, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Department of Cell Biology, University of Cincinnati College of Medicine, Cincinnati, Ohio 45267-0521
Received for publication, May 1, 2003 , and in revised form, May 28, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
In general, AR co-activators serve to stabilize the receptor complex, promote DNA unwinding, and/or recruit the RNA polymerase II complex (19, 20). Type I co-activators, such as SRC-1, TIF2, and pCAF/p300, are recruited by agonist-bound AR. These co-activators either harbor intrinsic HAT activity or recruit HATs to the AR complex. It is believed that the acetylation of lysine residues on core histones loosens chromatin and permits access for transcription factors to bind and activate promoters (19, 21–24). Type II co-activators stabilize the ligand bound conformation of the AR homodimer, which leads to increased activation of androgen responsive genes. For example, the type II co-activator ARA70 stabilizes N- to C-terminal interactions of the androgen receptor, dependent on ligand binding. The increased stability caused by co-activator binding promotes AR activity (19, 25, 26). Lastly, the TRAP/mediator complexes enhance ligand-dependent AR activity, likely through direct recruitment of RNA polymerase II (27). Several other cofactors have been identified, but the mechanisms by which they modulate AR activity have yet to be determined.
For several transcription factors, including selected nuclear receptors, the SWI/SNF ATP-dependent chromatin-remodeling complexes have been implicated in target gene regulation (28–37). The SWI/SNF family encompasses a number of large, multisubunit complexes comprising eight or more proteins. Invariantly, the SWI/SNF complex contains one of two core ATPases, BRG1 or hBRM. These complexes regulate the activity of transcription factors by re-organizing chromatin structure through either facilitating nucleosome condensation (which induces transcriptional repression) or nucleosome dispersion (assisting in transcriptional activation) (38–43). For example, we and others have shown that activity of the retinoblastoma tumor suppressor protein, RB, requires SWI/SNF activity for transcriptional repression of key cell cycle targets, including cyclin A (44–50). Loss of SWI/SNF activity abrogates RB-mediated transcriptional repression and the ability of RB to arrest cellular proliferation. In contrast, RB enhances association of the glucocorticoid receptor (GR) with either BRG1 or hBRM, thus resulting in increased GR activity (51–53). The association between BRG1 or hBRM and GR is known to be direct, and recruitment of SWI/SNF complexes to GR may actually stimulate displacement of GR from chromatin, allowing GR to return to the remodeled chromatin with additional transcription factors (30). Even though SWI/SNF activity clearly is necessary to repress RB targets and activate GR targets, both events can induce a cessation of cellular proliferation (cell cycle arrest and/or apoptosis) (44–46, 50, 52–54). Collectively, these observations contribute to the current hypothesis that the SWI/SNF complex can interact with sequence specific transcription factors to either promote or repress target gene activation, dependent on promoter context and complex content.
In addition to the GR, SWI/SNF activity has been shown to be essential for
estrogen receptor, ER
, activity. ER fails to activate
estrogen-responsive elements in SWI/SNF-defective cells, as demonstrated by
transient reporter assay. Using this same system, ER activity was
restored upon co-expression of BRG1
(33,
34). The hypothesis that
ER requires BRG1 activity was further substantiated by the observations
that BRG1 can be detected at endogenous ER target promoters in the presence of
ligand and that a BRG1-associated factor, BAF57, enhances ER/SRC-1
activation of estrogen-responsive promoters
(35). Presumably, such
induction of ER activity contributes to ER signaling and ER-dependent
proliferation in mammary epithelia and carcinoma cells.
Here, we probed the importance of SWI/SNF activity for AR function. We show that complete ligand-dependent activation of two distinct AR target promoters (PSA and probasin) requires SWI/SNF function. For PSA, SWI/SNF activity was essential, whereas for probasin low level activation was observed in the absence of SWI/SNF, indicating that this target can be at least partially activated independently of chromatin remodeling. Both core promoters were strongly stimulated by the restoration of hBRM. The re-introduction of BRG1 weakly activated the probasin promoter but was unable to stimulate the transcription of the PSA promoter by the AR. We also demonstrate that an ER-responsive promoter can be equally activated by either BRG1 or hBRM. These data suggest a strong preference for hBRM as the core ATPase for AR activity. Lastly, we report that the addition of a distal enhancer region to the proximal promoters relieved SWI/SNF dependence for PSA but did not impact the probasin promoter in remodeling-dependent target promoters. Thus, upstream elements can influence the requirement of SWI/SNF factors for AR action. Collectively, the data contained herein establish a role for hBRM-containing SWI/SNF complexes on AR activity. Moreover, these data demonstrate a role of enhancer regions in controlling the SWI/SNF requirement and reveal a disparity in AR target promoters with regard to the chromatin-remodeling requirement.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-galactosidase construct was the gift of Dr. Jean Wang (University
of California at San Diego, La Jolla, CA). The PSA61-luc reporter was kindly
provided by Dr. Kitty Cleutjens (Erasmus Universiteit, Netherlands) and has
been previously described
(55). pcDNA 3.1 was purchased
from Invitrogen. The pBJ5-BRG1 plasmid was provided by Dr. S. Goff (Columbia
University, New York, NY)
(56). The pCG-hBRM plasmid was
given to us by Dr. M. Yaniv (Institut Pasteur, Paris, France)
(29). The pBabe-dnBRG1 was a
gift from Dr. B. Weissman (University of North Carolina, Chapel Hill, NC).
pBS3X-ERE-luc and pCMV5-hER were kindly provided by Dr. S. Khan
(University of Cincinnati, Cincinnati, OH). Probasin-luc (ARR2-luc) was
constructed from ARR2-PB, a gift of Dr. R. Matusik (Vanderbilt University)
(57). Briefly, ARR2-PB
(57) was digested with
KpnI and BamHI. The probasin fragment was subcloned into the
KpnI and BglII sites in pGL2-Basic. PSA-luc was provided by
Dr. A. Puga (University of Cincinnati) and was generated by digesting
PSA61-luc (55) with
BglII and HindIII. The PSA promoter was inserted into the
BglII and HindIII sites in pGL3-Basic (Promega). ARR2+E-luc
was constructed by inserting of the 1.3-kb EcoRV fragment of the PSA
enhancer into the SmaI site of ARR2-luc. Dihydrotestosterone (DHT)
was purchased from Sigma (St. Louis, MO). -Tubulin, BRG1, and hBRM
antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz,
CA). CD44 antibody was a gift from Dr. L. Sherman (Oregon Health Sciences,
Portland, OR). Cell Culture and Treatment—SW13, C33A, and CV1 cell lines were obtained from ATCC and maintained in a 37 °C, 5% CO2 incubator. SW13, C33A, and CV1 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum (FBS, Atlanta Biologicals, Norcross, GA), 2 mM L-glutamine, and 100 units/ml penicillin/streptomycin (Mediatech, Herndon, VA). For SW13, C33A, and CV1 reporter assays, 10% charcoal dextran-treated FBS (CDT, HyClone Laboratories, Logan, UT) and phenol red-free Dulbecco's modified Eagle's medium were utilized.
Transfection and Reporter Assays—SW13, C33A, and CV1 cells
were transfected in charcoal dextran-treated (CDT) serum, which lacks steroid
hormones, with the indicated plasmid constructs using the BES/calcium
phosphate protocol (58). 0.75
µg of receptor expression plasmid (SG5-AR or ER), 0.5 µg of
reporter construct (ARR2-luc, PSA-luc, PSA61-luc, ARR2+E-luc, or 3X-ERE-luc),
2.65 µg of either pBJ5-BRG1 or pCG-hBRM, 0.1 µg of
CMV--galactosidase, and empty vector (pcDNA3.1) to a total of 4 µg
were transfected for each well of a 6-well dish. Following the transfection,
the cells were washed with sterile phosphate-buffered saline. SW13, C33A, or
CV1 cells were then allowed to recover for a period of 5–6 h and then
supplemented with 0.1 nM DHT (Sigma) or 0.1% ethanol vehicle (EtOH)
for 20–24 h. Following stimulation, cells were harvested and luciferase
activity was quantified using the Promega luciferase assay kit (Promega,
Madison, WI). -Galactosidase activity was used as an internal control
for transfection efficiency and was measured utilizing the manufacturer's
recommended protocol for the Galacto-Star kit (Tropix, Bedford, MA). Basal
activity (in the presence of ethanol vehicle) was set to "1," and
relative luciferase activity is shown. Experiments were performed at least in
triplicate. Averages and standard deviations are shown.
Immunoblotting—1 x 105 C33A cells were seeded into a 6-cm dish. The cells were then transfected with 8 µg of either pcDNA 3.1, pBJ5-BRG1, or pCG-hBRM plasmid for 12 h via BES/calcium phosphate. The cells were then washed with sterile phosphate-buffered saline. After 16 h, the cells were trypsinized and pelleted. The cell pellet was lysed in radioimmune precipitation assay buffer (150 mM NaCl, 1.0% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, and 50 mM Tris, pH 8.0) supplemented with a protease inhibitor mixture containing 10 µg/ml 1,10-phenanthroline, 10 µg/ml aprotinin, 10 µg/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride. Lysis of cells proceeded on ice for 10 min. After brief sonication, lysate was clarified by centrifugation at 14,000 x g, 4 °C for 10 min. The supernatant was quantified by using a DC protein assay (Bio-Rad), and 30 µg of total protein was separated by SDS-PAGE and transferred onto an Immobilon-P membrane (Millipore). The membrane was immunoblotted with the indicated antibodies via standard techniques, and antibodies were detected using Western Lightning chemiluminescence (PerkinElmer Life Sciences, Boston, MA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
) and CMV--galactosidase (as an
internal control for transfection efficiency), and the 3X-ERE-luc reporter
construct. ER activated the ERE reporter 4.9-fold in the presence of 10
nM 17-estradiol (E2) (left panel). In contrast,
addition of 10 nM E2 failed to activate the 3X-ERE-luc reporter in
SW13 cells (right panel). These observations are consistent with
previous reports demonstrating the requirement of SWI/SNF for ER
activity (34).
|
To determine if the AR requires the SWI/SNF complex for transactivation,
similar experiments were performed. Briefly, SWI/SNF-competent (CV1) or
-defective (SW13 and C33A) cells were cultured for 24 h in the absence of
steroid hormone (phenol red-free media supplemented with 10% CDT). Cells were
then transfected with plasmid encoding the androgen receptor (pSG5-AR),
CMV--galactosidase, and the ARR2-luc reporter plasmid at the amounts
indicated. ARR2-luc is a well characterized, AR-specific reporter construct
that contains two copies of the probasin promoter in tandem and directs
prostate-specific expression in transgenic mouse models
(57). Post-transfection, cells
were treated with vehicle (0.1% ethanol) or physiological levels of androgen
(0.1 nM DHT) for 20–24 h. Cells were then harvested and
processed to detect luciferase and -galactosidase activity. As can be
seen in Fig. 1B, the
probasin promoter was induced 138.0-fold by androgen in CV1 cells (left
panel) but only 6.6- and 10.0-fold in SWI/SNF-defective SW13 and C33A
cells, respectively (middle and right panels). This
significant reduction in activity is comparable to that observed with GR in
C33A and SW13 cells (29).
Thus, the AR exhibits significant deficiency in activating the probasin
promoter in the absence of core SWI/SNF activity.
AR-dependent Activation of the Probasin Promoter Is Induced by the
Restoration of BRG1 but Preferentially by hBRM—It has been
previously shown that SWI/SNF activity can be restored in SW13 and C33A cells
through ectopic expression of BRG1 or hBRM. For example, CD44 expression is
lost in BRG1/hBRM-deficient cell lines. The expression of CD44 expression can
be restored by transfecting either BRG1 or hBRM into these cells (SW13 or
C33A) (44,
45,
48). In addition, we and
others have previously shown that BRG1 or hBRM is required for RB-mediated
repression of the cyclin A promoter. Re-introduction of BRG1 or hBRM into a
SWI/SNF-deficient background restores the ability of RB to repress cyclin A
(as observed both through reporter assay and through examination of endogenous
protein) (46,
50). Thus, restoration of
SWI/SNF activity can rescue transcriptional defects from both ectopic reporter
constructs and endogenous promoters. As shown in
Fig. 2A, the function
of the BRG1 and hBRM expression constructs were confirmed by transfecting the
expression plasmids into C33A cells and immunoblotting for CD44 expression
(compare lane 1 to lanes 2 and 3). As can be seen,
both BRG1 and hBRM were effective at restoration of CD44 expression. hBRM was
more efficient in the induction of CD44, which is consistent with previous
reports (48). BRG1 and hBRM
expression levels are shown, and immunoblots against -tubulin were
included as a loading control. To determine if restoration of SWI/SNF ATPase
activity could induce transcription of the probasin promoter, reporter assays
were repeated in the presence of ectopic BRG1 or hBRM. Briefly, SW13 cells
were cultured, as in Fig. 1, in
the absence of steroid hormone and transfected with plasmid encoding AR, BRG1
or hBRM, -galactosidase, and the ARR2-luc reporter at the amounts
indicated. Post-transfection, cells were stimulated with either DHT or ethanol
vehicle. BRG1 or hBRM expression had no effect on basal AR activity (in the
presence of ethanol) (Fig.
2B). In the presence of DHT, BRG1 marginally stimulated
ligand-dependent promoter activation (2.5-fold, as compared with the absence
of BRG1). In comparison, hBRM markedly stimulated AR activity (8.4-fold,
compared with the absence of hBRM or BRG1). Thus, AR exhibited a preference
for hBRM as the SWI/SNF ATPase component for transcriptional activation of the
ARR2-luc reporter. To demonstrate that these effects were attributed to ATPase
activity and not to simple assembly of the SWI/SNF complex, dominant negative
BRG1 (dnBRG1) was utilized. This allele harbors point mutations in the ATPase
domain, rendering the resulting mutant protein capable of assembling the
SWI/SNF complex, but incapable of remodeling chromatin
(56). As can be seen, dnBRG1
lacked any ability to restore AR activity on the probasin promoter
(Fig. 2B).
|
Given the preference of AR for hBRM containing SWI/SNF activity, we
examined the effect of BRG1 versus hBRM on ER activity. It has
been shown that the re-introduction of BRG1 into SW13 cells restores
ER-mediated activation of the 3X-ERE-luc reporter and that dnBRG1 does not
allow for the activation of the same reporter
(34). These data were
confirmed in Fig. 2C,
wherein SW13 cells were transfected with expression plasmid for BRG1 or dnBRG1
and the ERE-luc reporter. Previous reports did not examine the relative
efficacy of hBRM for restoration of ER activity. As depicted in
Fig. 2C, ER
activity was equally restored by BRG1 versus hBRM in SW13 cells (3.3-
and 3.6-fold above background for BRG1 and hBRM, respectively). These
collective data imply that AR has a preference for hBRM as the core ATPase,
whereas BRG1 and hBRM are equally effective in stimulating ER
activity.
Activation of the Proximal PSA Promoter Requires hBRM-dependent
SWI/SNF Activity—Given the unique preference of AR for
hBRM containing SWI/SNF complexes, these results prompted examination of
additional AR-dependent promoter constructs. The prostate-specific antigen
(PSA) promoter is the most well characterized AR target, is tightly regulated
by androgens, and demonstrates prostate-specific expression
(55). The proximal PSA
promoter contains two ARE sites and was cloned upstream of luciferase to
generate the PSA-luc reporter. As can be seen in
Fig. 3A, AR activated
the PSA promoter 
10.1-fold in the presence of ligand in CV1 cells
(left panel). In contrast, no activation of PSA-luc was observed in
SW13 cells (right panel). To analyze the effect of SWI/SNF activity
on activation of this promoter, BRG1 or hBRM were restored by transfection, as
seen in Fig. 2B.
Strikingly, restoration of BRG1 failed to rescue PSA-luc activation, whereas
hBRM strongly induced promoter activity, yielding a 278% increase over
activity in the presence of ligand but absence of hBRM
(Fig. 3B). Thus,
activation of both the proximal probasin promoter and the proximal PSA
promoter requires SWI/SNF activity, with a strong preference for hBRM as the
core ATPase. Interestingly, addition of hBRM in SWI/SNF-competent CV1 cells
had no impact on the proximal PSA promoter
(Fig. 3C), or the
proximal probasin promoter (data not shown), indicating that excess hBRM does
not simply enhance AR activity.
|
The PSA Enhancer Region Facilitates a Bypass of the SWI/SNF
Requirement—The preceding data revealed a requirement for SWI/SNF
activity on the PSA promoter. For this target, a distal regulatory element has
been characterized upstream of the proximal promoter. This enhancer region is
known to harbor a site of AR binding (ARE III, 4 kb upstream) and to
synergize with the proximal PSA promoter for androgen-dependent gene
activation. To analyze the impact of the enhancer region on SWI/SNF activity,
the previously characterized PSA61-luc construct was utilized
(55). As shown in
Fig. 4A, this
construct contains both the proximal promoter and the upstream enhancer region
(–6000 to +12). Similar to Fig.
1, SWI/SNF-defective (SW13 and C33A) and competent (CV1) cells
were transfected in the absence of steroid hormone with the expression plasmid
for AR, -galactosidase, and the PSA reporter construct (PSA61-luc).
Strikingly, the ability of AR to activate the PSA61-luc construct was
comparable in each cell line tested (Fig.
4B), with the highest activation above basal levels
(13.0-fold) being observed in C33A cells (3.8-fold in SW13 and 5.7-fold in
CV1). Thus, the presence of the enhancer region in PSA facilitates a bypass of
the SWI/SNF requirement.
|
To determine if re-introduction or overexpression of hBRM could further augment transcription of the PSA61-luc reporter, SW13 and CV1 cells were transfected as above with expression plasmid for hBRM at the concentrations indicated, and stimulated with either vehicle or DHT post-transfection. As shown, hBRM did not significantly increase the transcriptional levels of the PSA61-luc reporter in either cell background (Fig. 4, C and D). BRG1 was also incapable of potentiating the response to ligand (data not shown). Lastly, re-introduction of BRG1 or hBRM into androgen-dependent LNCaP cells did not significantly increase expression of the endogenous PSA gene (data not shown). Therefore, not only does the enhancer region effectively reverse the requirement for SWI/SNF activity but it renders the proximal PSA promoter refractory to SWI/SNF action.
To examine the effect of the enhancer on the proximal probasin promoter, the ARR2+E-luc construct was generated, which positions the enhancer region of PSA (containing the ARE III) directly upstream of the ARR2 promoter (Fig. 5A) (55, 57). Transfection of this reporter construct into SW13 or C33A cells with AR expression plasmid, as described for Fig. 1, revealed that, unlike the PSA61 reporter, activity of the enhancer addition to the probasin promoter did not alter the response to SWI/SNF activity (Fig. 5, B and C). Ligand-dependent activation of the reporter constructs was similar between ARR2-LUC and ARR2+E-LUC (6.6- versus 6.5-fold, respectively). As with the proximal promoter, activity of ARR2+E was restored upon co-transfection of hBRM (55.0- versus 42.0-fold, respectively). Thus, the impact of the enhancer on SWI/SNF requirement is variant and dependent on promoter context.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
, wherein BRG1 and
hBRM demonstrated equal efficacy in the ability to restore ER
transactivation potential. In addition, we show that the distal PSA regulatory
element strongly influences the requirement of hBRM activity for AR-dependent
transcription on the PSA promoter. The presence of the upstream enhancer
region bypassed the requirement for SWI/SNF and rendered the element
refractory to excess SWI/SNF activity. In contrast, addition of this same
enhancer element to the proximal probasin promoter did not alter the
requirement for the SWI/SNF complex. Together, these data indicate that the
requirement of AR for SWI/SNF activity is dependent on promoter context and is
more receptive to SWI/SNF complexes with hBRM as the core ATPase. SWI/SNF activity has been implicated in the appropriate regulation of specific genes. Here, we show that the PSA promoter shows a distinct requirement for SWI/SNF. Probasin, however, retains some activity in the absence of SWI/SNF (6-fold induction). The disparity in requirement for SWI/SNF activity between the two core promoters indicates that the probasin promoter does not explicitly require chromatin remodeling for responsiveness, suggesting that AR likely binds to the probasin AREs in the absence of these factors. However, the probasin promoter was still receptive to SWI/SNF activity (especially hBRM), suggesting that remodeling may assist in a secondary step of activation (e.g. promoter clearance). In contrast, the PSA promoter may require remodeling activity at an earlier step of transcriptional activation (e.g. AR access). It is interesting to note that basal PSA activity (in the absence of ligand) is significantly reduced upon the re-introduction of BRG1 or hBRM (Fig. 3, B and C), whereas no effect on basal activity was observed with the probasin promoter. Because remodeling proteins can influence several classes of transcription factors (including the loading of both activators and repressors), it is possible that these observations reflect the loading of repressors to the promoter in the absence of hormone.
Strikingly, the presence of the distal PSA enhancer region alleviated the SWI/SNF requirement for the PSA promoter. This region has been well characterized and is known to contain AR binding sites (e.g. ARE III) (55). Recently, it has been shown through chromatin immunoprecipitation experiments that ARE III is a key site of AR interaction. Using quantitative PCR to analyze chromatin immunoprecipitation data, a recent report revealed that AR binds first to ARE III, preceding histone acetylation on this region, whereas AR binding to the promoter region (ARE I) was delayed and occurred after histone acetylation in that region (17). Thus, the distal enhancer region may be the initial site of AR action, capable of influencing events at the proximal promoter. A contrasting earlier report demonstrated that AR binding to ARE III and the ARE I region was simultaneous but that co-repressors were recruited only to the proximal promoter (60). In either scenario, co-activators known to bind the ARE III region may alleviate the requirement of SWI/SNF function in the proximal promoter. The data contained herein suggest that these events do not require SWI/SNF but that ablation of enhancer activity results in a SWI/SNF-dependent proximal promoter. Interestingly, addition of the PSA enhancer region immediately 5' to the proximal probasin promoter did not relieve SWI/SNF dependence, indicating that structural or positional effects of this region (especially spacing between cis-acting elements) could influence a bypass of the SWI/SNF requirement. As of yet, there is scant knowledge of how the enhancer and proximal PSA promoter (a 4-kb separation) interact to regulate AR-dependent transcription. Although the enhancer does contain ARE III, the minimal enhancer region requires a broad 440-bp core known to contain at least three separate regions of activity required for its function (55). Thus, it is evident that multiple cis-acting elements contribute to ARE III function in the enhancer region. The role of the enhancer in controlling the influence of the SWI/SNF complex will be the focus of future study.
No studies have investigated the relative contribution of BRG1
versus hBRM complexes for nuclear receptor activation. Although both
BRG1 and hBRM have been shown to associate with the GR, no comparison of
preference for the ATPases has been performed
(29–31).
Here, we report that AR function on multiple target promoters shows a strong
preference for hBRM activity, whereas ER activity was effectively
restored by either BRG1 or hBRM. Enhancement of AR activity by hBRM on a
nonspecific steroid hormone reporter (murine mammary tumor virus) was also
recently reported to occur in SWI/SNF-competent cells, which is consistent
with our findings (61).
Preference for a core ATPase does have precedent in the literature. For
example, two ankyrin repeat proteins integral to the Notch pathway selectively
associate with BRM and not BRG1. In contrast, several zinc finger proteins
were recently shown to associate with and be activated by BRG1, but not BRM,
through virtue of unique N-terminal sequences present only on BRG1. In that
study, interaction with BRG1 was shown to occur through the zinc finger
protein DNA binding domain, including that of the retinoic acid receptors (RXR
and RAR) (62). It is
interesting to note that both AR and ER contain a highly conserved
central DNA binding domain comprised of two zinc finger-like modules, yet
neither shows a preference for BRG1-containing complexes
(63,
64). Thus, divergent
N-terminal sequence in the nuclear receptors may influence activation by BRG1-
versus hBRM-containing complexes.
The biological outcome of hBRM or BRG1 preference is unclear. Studies with knockout mice have demonstrated that ablation of BRG1 expression is lethal in the developing embryo but is not required for viability in primary embryonic fibroblasts, as determined via Cre-lox technology (65). Thus, BRG1 is specifically required for development. In contrast, BRM knockout mice are viable, indicating that BRM function is dispensable for development (66). In differentiating cells, BRM is expressed at high levels. Interestingly, loss of BRM also results in a small decrease in the size of the testis (an androgen-dependent tissue) (67) as compared with wild-type controls (66). The impact of hBRM loss on the growth of the prostate has yet to be determined. Relative expression of BRG1 and hBRM in prostatic adenocarcinomas has not been addressed, although we have detected both proteins in LNCaP (AR-positive, androgen-dependent human prostatic adenocarcinoma) cells (data not shown). The BRG1 locus is reported to be mutated in AR-negative, androgen-independent DU145 (prostate cancer) cells, resulting in a frameshift prior to the ATPase domain and likely encoding a non-functional protein (68). The impact of this mutation on cellular function is unknown, although DU145 cells are known to support AR activity upon re-introduction of the protein (69), thus supporting our observations that hBRM is the preferred ATPase for AR action. In addition, the BRG1 locus (19p13) is a known region of loss of heterozygosity in prostate cancer, and a re-introduction of 19p13.1–13.2 in the Dunning rat R3327 prostate cancer model suppresses tumor growth (70). This function of BRG1 may be attributed to the requirement of RB for SWI/SNF function, and the effect of BRG1 loss on androgen receptor activity in prostate tumor cells is unknown. Examination of hBRM expression in prostate cancer progression has not been addressed. However, loss of hBRM expression is known to result in induction of BRG1, indicating that compensation between core ATPases takes place (66). Should this be the case, our data would suggest that loss of BRG1 and concomitant induction of hBRM could serve to enhance AR action in the prostate.
In summary, we report a functional interaction between AR and hBRM-containing SWI/SNF complexes and demonstrate that different AR targets show disparity in the requirement for SWI/SNF. Our data also indicate that access to enhancer regions may influence target specificity for SWI/SNF function. These findings provide the first link between AR and hBRM function and provide the impetus to study SWI/SNF impact on multiple AR targets and regulatory elements.
|
FOOTNOTES |
|---|
Supported by the Albert J. Ryan Foundation and the University of Cincinnati
Distinguished Graduate Award.
To whom correspondence should be addressed. Tel.: 513-558-7371; Fax:
513-558-4454; E-mail:
Karen.Knudsen{at}UC.edu.
1 The abbreviations used are: AR, androgen receptor; ER, estrogen receptor;
GR, glucocorticoid receptor; DHT, dihydrotestosterone; PSA, prostate-specific
antigen; FBS, fetal bovine serum; CDT, charcoal-dextran treated; RB,
retinoblastoma tumor suppressor protein; HAT, histone acetyltransferase; BES,
N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid; CMV,
cytomegalovirus; E2, 17-estradiol; dn, dominant negative; ARE, androgen
response element.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K. A. Link, S. Balasubramaniam, A. Sharma, C. E.S. Comstock, S. Godoy-Tundidor, N. Powers, K. H. Cao, A. Haelens, F. Claessens, M. P. Revelo, et al. Targeting the BAF57 SWI/SNF Subunit in Prostate Cancer: A Novel Platform to Control Androgen Receptor Activity Cancer Res., June 15, 2008; 68(12): 4551 - 4558. [Abstract] [Full Text] [PDF]
|
|
|
|
<
