| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
The Population Council and The Rockefeller University, New York, New York 10021
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
(RXR), either in
solution or bound to a DR+4 TRE, recruited SRC-1 in a
T3-dependent manner. The AF-2 of TR was clearly
involved in this process because a TR-RXR heterodimer containing a
mutant TRß (1–450) with impaired AF-2 failed to bind to SRC-1.
Surprisingly, the RXR-specific ligand 9-cis-retinoic acid
induced binding of SRC-1 to the RXR component of the TRE-bound
heterodimer. This novel finding suggests that RXR, as a heterodimeric
partner of TR, has the potential to play an active role in
transcriptional regulation. Our results raise the interesting
possibility that a RXR-specific ligand may modulate
T3-mediated signaling by inducing additional
interactions between TRE-bound TR-RXR heterodimer and the coactivator. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
and ß, which mediate
the physiological actions of thyroid hormones, belong to the nuclear
hormone receptor superfamily of ligand-inducible transcription factors
that also includes receptors for steroid hormones, retinoids, and
vitamin D3 (1, 2, 3, 4). TR interacts with distinct DNA
sequences, termed thyroid hormone response elements (TREs), in the
target gene promoter and functions either as a transcriptional
activator or a repressor depending on the hormonal status of the
receptor (5). In the absence of hormone, TR functions as a silencer of
basal level transcription from the target promoter (6, 7, 8, 9, 10, 11). Ligand
binding to TR releases transcriptional silencing and leads to the
activation of target gene expression (6, 7, 8, 9). The precise mechanisms by
which TR exerts its transcriptional effects on a TRE-linked promoter
are unclear. Recent studies in several laboratories indicate that nuclear receptors repress or enhance transcription by interacting with multiple cellular coregulatory factors, which function as signaling intermediates between the receptors and the RNA polymerase II transcription machinery (for a review see 12 . We and others have demonstrated that unliganded TR associates with a negative coregulatory factor, termed corepressor (13, 14, 15, 16, 17). The receptor-corepressor complex binds to the TRE and actively represses target gene transcription by impairing the activity of the basal transcription machinery (13, 14, 15, 16, 17). The binding of T3 to TR results in the dissociation of the corepressor from the receptor, leading to the reversal of transcriptional repression (13, 14, 15, 16, 17). The hormone-occupied, DNA-bound receptor is then believed to act in unison with a positive coregulatory factor, termed coactivator, to mediate gene activation (9, 14). The coactivators are envisioned to act as bridging molecules between the activation domain(s) of the receptor and the basal transcription machinery.
Studies by Chambon and co-workers (18, 19) utilizing transcriptional
interference or ‘squelching’ between different steroid receptors
first suggested that cofactors in addition to the activated hormone
receptor and the basal transcription apparatus are necessary for gene
activation. We observed recently that ligand-occupied TR, TR or
TRß, inhibits transactivation of a progesterone-responsive reporter
gene by progesterone receptors in CV1 or human breast carcinoma T47D
cells (20). The transcriptional interference occurred in the absence of
a DNA response element for the ‘interfering receptor’ (TR) in the
target promoter. It was also totally dependent on the presence of the
hormonal ligand T3. Conceptually, the target protein(s) of
an inhibitory cross-talk between two different nuclear receptors is one
or more basal transcription factors or coactivator(s). We demonstrated
that T3-occupied TR ligand-binding domain (LBD) could bind
to and functionally deplete a soluble cofactor(s) that is critical for
transactivation of a progesterone-responsive gene in T47D nuclear
extracts, while the basal level of trancription from a minimal TATA
promoter or activated transcription from a control adenovirus
major-late promoter remained unaffected (20). These results provided
strong biochemical evidence in favor of the existence of a common,
limiting coactivator molecule that mediates the interaction of
ligand-bound steroid or TR with the RNA polymerase II transcription
machinery to achieve the activated level of the target gene
expression.
Consistent with this scenario, a number of putative mediator proteins have been isolated during the past 2 yr by a number of laboratories using yeast two-hybrid assay or far-Western cloning (12, 21, 22, 23, 24, 25, 26, 27, 28, 29). Screening of cDNA libraries employing baits containing the LBDs of nuclear receptors led to the isolation of multiple candidate coactivators such as steroid receptor coactivator-1 (SRC-1) (23), SUG-1 (24), transcriptional intermediary factor-1 (TIF-1) (25), receptor-interacting protein-140 (RIP-140) (26), TIF-2 (27), glucocorticoid receptor-interacting protein 1 (GRIP1) (28), p160 (21, 29), and CREB-binding protein (CBP) (29). It is now clear that SRC-1, p160, TIF-2, and GRIP1 are members of a family of structurally related coregulatory factors, which significantly enhance transactivation by several nuclear receptors in the presence of the cognate ligand (23, 27, 28, 29). It has also been reported that each of these proteins interacts directly with a nuclear receptor in a ligand-dependent manner (23, 24, 25, 26, 27, 28, 29). However, the molecular interactions between a coactivator such as SRC-1 and the transactivation domain(s) of a nuclear receptor have not been explored in detail.
In this study, we analyzed the functional interactions between SRC-1
and a TR-retinoid X receptor (RXR) heterodimer to investigate the
mechanisms of action of the coactivator in TR-mediated transactivation.
Using transient transfection experiments, we demonstrated that SRC-1
functions as a transcriptional coactivator for hormone-occupied TRß.
A conserved COOH-terminal amphipathic -helix in the LBD of TRß is
essential for ligand-induced interaction with the coactivator. We
provide strong evidence that this conserved -helix is necessary for
interactions with SRC-1. We also observed that both receptor partners
within a DNA-bound TR-RXR heterodimer display the ability to recruit
SRC-1. Interaction of the coactivator with each partner of the
heterodimer depends on the cognate hormonal ligand. These results
suggest a plausible mechanism by which a combination of TR- and
RXR-specific hormonal signals in a target cell may influence the
transcriptional activity of a TRE-linked promoter by simultaneously
activating both partners within a TR-RXR heterodimer.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, T3 induced a 4- to
5-fold TRß-mediated transactivation of a thyroid hormone response
element (TRE)-linked reporter gene in the absence of any cotransfected
SRC-1 (compare lanes 1 and 2). Cotransfection of an expression vector
containing SRC-1 did not affect transcriptional activity of unliganded
TR but led to a dramatic enhancement in transactivation by liganded TR
(lane 4). The SRC-1-induced enhancement of transactivation was
typically about 4-fold over that mediated by TR alone effecting a net
20-fold T3-dependent transactivation of the TRE-linked
promoter (compare lanes 1 and 4). Consistent with the earlier report by
Onate et al., SRC-1 therefore functions as a coactivator of
TRß-mediated transactivation of a TRE-linked promoter in the presence
of T3.
|
-helix that
contains amino acid residues that are highly conserved among the
nuclear hormone receptors and has been shown to be critical for the
ligand-dependent transactivation by these receptors (9, 30, 31).
Baniahmad et al. (9) previously showed that deletion of the
last six carboxy-terminal amino acids (451–456) in TRß resulted in a
mutant receptor, TRß
C6, which retained the ability to
bind thyroid hormone but was impaired in AF-2 function.
To determine whether the AF-2 core domain of TR is involved in
coactivator function, we employed transient transfection to examine the
effects of SRC-1 on transcriptional activity of TRßC6.
As reported previously, we found that this TR mutant failed to
transactivate a TRE-linked promoter in a T3-dependent
manner (Fig. 1
, lanes 5 and 6). Cotransfection of a vector expressing
SRC-1, which markedly enhanced transactivation by full-length TRß,
did not exhibit any effect on transcriptional activity of the AF-2
mutant either in the presence or in the absence of T3
(lanes 7 and 8). These results clearly implied that conserved amino
acids 451–456 within the C-terminal AF-2 core domain of TRß play a
critical role in the coactivator function of SRC-1.
AF-2 Core Domain of TRß Is Essential for Ligand-Dependent
Interaction with SRC-1
Since SRC-1 functions as a transcriptional coactivator for TR, it
is likely to undergo protein-protein interactions with the ligand-bound
receptor. We therefore examined the interaction between SRC-1 and TRß
in vitro. A recombinant TRß fused to
glutathione-S-transferase (GST) was used in this experiment.
As shown in Fig. 2A, very little binding of
35S-labeled SRC-1 to GST-TRß immobilized on an affinity
matrix was detected in the absence of T3 (lane 2). In the
presence of T3, however, we observed a remarkable
enhancement in the binding of SRC-1 to GST-TRß (lane 3). These
results demonstrated that TRß undergoes direct interaction with SRC-1
in a strictly ligand-dependent manner.
|
, top panel). We observed that the LBD fragment of TRß
containing the amino acids (145–456) exhibited remarkable
ligand-dependent binding to SRC-1 (middle and bottom panels,
lanes 4 and 5). By our estimate, about 20-fold enhancement in SRC-1
binding to GST-TR (145–456) was observed in the presence of
T3 (bottom panel, compare lane 4 with lane 5).
In contrast, a fragment of TRß containing the N-terminal amino acids
(1–260) showed only modest (about 2-fold) ligand-independent binding
to SRC-1 (middle and bottom panels, compare lane 8 or 9 with
lane 4). Interestingly, a receptor mutant GST-TR (82–450) that lacked
the N terminus, as well as the last six COOH-terminal amino acids, but
retained hormone binding, failed to exhibit any interaction with SRC-1
either in the presence or in the absence of T3 (lanes 6 and
7). Thus, the deletion of critical amino acids (451–456) within the
AF-2 core domain of TRß, which results in the loss of ligand-induced
transcriptional activity of this receptor, also disrupts the
interaction between the receptor and SRC-1. Taken together, these
results prompt us to propose that the conserved AF-2 core domain of TR
critically influences the ligand-dependent interaction between SRC-1
and the receptor.
A 20-Amino Acid Synthetic Peptide Containing the AF-2 Core Domain
of TRß Blocks the Interaction between the Receptor and SRC-1
Danielian et al. (30) and others (9, 31) have
identified a 17-amino acid AF-2 core region containing an amphipathic
-helix whose main features are well conserved between all known
nuclear receptors that transactivate in a hormone-dependent manner.
Point mutations within the AF-2 core domain decrease or abolish
ligand-dependent activation, even though the ligand binding, DNA
binding, and dimerization properties remain unaffected (9, 31).
Furthermore, this domain displays an autonomous activation function
when fused to a heterologous DNA binding domain (9, 31). Recent crystal
structure analyses of the nuclear receptors TR, retinoid acid receptor
(RAR), and RXR suggest that induction of AF-2 activity upon ligand
binding corresponds to a major conformational change involving the
repositioning of the -helix containing the AF-2 core domain
(32, 33, 34). These observations raised the possibility that this domain is
involved in creating the surface that interacts directly with a
transcriptional coactivator such as SRC-1. We reasoned that if this
postulation is correct, then a peptide containing the AF-2 core domain
might be able to disrupt the binding of the full-length receptor to
SRC-1 by occluding the coactivator-binding site.
To examine this possibility, we synthesized a peptide containing the
last 20 C-terminal amino acids (437–456) containing the AF-2 core
domain of TRß. A control peptide containing the same region but
harboring mutations in several key amino acids, which are known to be
critical for AF-2 activity (9, 31), was also generated (Fig. 3, top panel). We then incubated ligand-bound
TRß with 35S-labeled SRC-1 in the presence or in the
absence of excess AF-2 core or control peptide. As shown in Fig. 3
(middle panel), incubation with an excess of the control
peptide did not significantly affect the ligand-induced binding of
SRC-1 to TRß (compare lane 2 with lane 5 or 6). Incubation with a
similar excess of the AF-2 core peptide, on the other hand, drastically
inhibited the binding of SRC-1 to hormone-occupied TRß (compare lane
2 with lanes 3 or 4). We observed that approximately 50% of
[35S]SRC-1 binding to 2 pmol of TRß was suppressed in
the presence of 10 nmol of AF-2 core peptide (Fig. 3, bottom
panel). One can consider two different mechanisms by which the
AF-2 core domain may inhibit coactivator binding. It is conceivable
that the AF-2 core helix itself might represent a discrete binding site
for the coactivator. In this scenario, the AF-2 core peptide competes
with the TR LBD for the same binding site on the SRC-1. Alternatively,
the AF-2 -helix may align with another region(s) of the receptor to
generate the proper binding surface for the coactivator. In this latter
scenario, the peptide may inhibit SRC-1 binding by disrupting the
intramolecular interaction that creates the coactivator binding site.
Either way, interference of SRC-1 binding to the full-length receptor
by the AF-2 core peptide strongly suggests that the C-terminal
amphipathic -helix of TRß is critically involved in creating the
surface in the activated receptor that interacts with the
coactivator.
|
, we used
bacterially expressed GST-tagged RXR and hexahistidine-tagged TRß.
The GST-RXR fusion protein was first immobilized on glutathione
affinity resin. The TR-RXR heterodimers were generated in the absence
of DNA by incubating the immobilized RXR with excess TRß. The
formation of heterodimers in solution was ascertained by isolating the
immobilized complexes, followed by analysis of the protein components
by SDS-PAGE and by protein staining. Equivalent intensities of TR- and
RXR-specific bands representing equivalent molar amounts of the
receptors were visualized in the isolated complexes confirming
efficient heterodimer formation under these conditions (Fig. 4A). The
heterodimerization was not significantly affected by the presence of
either a combination of T3 (1 µM) and
9-cis-retinoic acid (RA) (1 µM) (compare lanes
2 and 3 with lanes 6 and 7) or excess direct repeat (DR)+4 TRE (compare
lane 2 with lane 3, lane 6 with lane 7). The resin complexed to TR-RXR
heterodimers was isolated by brief centrifugation and washed
repeatedly. The resin-bound receptors were then treated with either
T3 or 9-cis-RA or solvent follwed by incubation
with 35S-labeled SRC-1.
|
, resin-bound GST-RXR or GST-RXR-TR did not retain
any significant SRC-1 signal in the absence of any hormone (lanes 1, 5,
and 6). Addition of the RXR-specific ligand, 9-cis-RA, led
to a marked enhancement in the binding of SRC-1 to GST-RXR alone
(lower panel, compare lanes 1 and 2). Interestingly,
heterodimerization with TR did not affect 9-cis-RA-dependent
binding of SRC-1 to RXR (compare lane 2 with lanes 3 or 4). Our results
also clearly showed that RXR in a preformed heterodimer could bind to
its ligand and recruit SRC-1 in a ligand-dependent manner (lanes 3 or
4). As expected, addition of T3 to the heterodimer also
markedly induced interaction of SRC-1 with the TR component (lanes 7
and 8). Taken together, these results demonstrated that in a TR-RXR
heterodimer both receptor partners are capable of interacting with a
transcriptional coactivator in a strictly ligand-dependent fashion.
Ligand-Dependent Recruitment of SRC-1 by a DNA-Bound TR-RXR
Heterodimer
It is conceivable that a DNA response element may modulate
interaction of a TR-RXR heterodimer with SRC-1. A recent report
indicated that the polarity or relative configuration of half-sites
within a retinoid response element dictates the interaction of a
RAR-RXR heterodimer with a nuclear receptor corepressor (14). A TR-RXR
heterodimer binds to a DNA response element consisting of two direct
repeat half-sites (DR+4) of consensus sequence AGGTCA in an asymmetric
manner: RXR occupies the upstream half-site and TR occupies the
downstream half-site (36, 37). We therefore examined whether the
binding of a TRß-RXR heterodimer to the DR+4 TRE influences its
interaction with SRC-1. For this purpose, we first assembled TR-RXR
heterodimers on biotinylated DR+4 TRE oligodeoxynucleotides. The
DNA-bound TR-RXR complexes were then isolated by binding to an
avidin-linked resin, and the resin-bound heterodimers were incubated
with 35S-labeled SRC-1 either in the presence or in the
absence of T3 or 9-cis-RA. After stringent
washings, the resin-bound proteins were analyzed by SDS-PAGE for the
presence of 35S-labeled SRC-1. The results of these
experiments are shown in Fig. 5, A and B.
|
, lane 2). This
nonspecific retention of SRC-1 was not significantly affected when the
resin was incubated with biotinylated TRE alone (lane 3) or with TR and
RXR in the absence of TRE (lanes 8 and 9). The results in Fig. 4A
suggest that a heterodimer is generated on the DR+4 TRE when one
receptor partner is incubated with DNA in the presence of an excess of
the other partner. The binding of SRC-1 increased only slightly when
resin-bound TRE-TR-RXR complexes were used in the absence of any
hormonal ligand (lane 4). A pronounced enhancement in the binding of
SRC-1 was observed only when resin-bound TRE-TR-RXR complexes were
incubated with the 35S-labeled protein in the presence of
either T3 or 9-cis-RA (upper and
lower panels, compare lane 4 with lane 5 or 6). These
results were in excellent agreement with those described in Fig. 4B
using TR-RXR heterodimers formed in solution. Coaddition of both
ligands did not produce any significant change in SRC-1 binding
compared with each ligand added singly (compare lane 5 or 6 with lane
7). The limiting amount of SRC-1 in the binding reaction did not allow
us to determine whether the two ligand-occupied receptors in a
heterodimer bound to the coactivator in an additive or a synergistic
fashion.
We also examined the binding of 35S-labeled-SRC-1 to the
DR+4-bound resin when TR or RXR was used singly (Fig. 5B). Previous
reports indicated that TR alone could bind to DR+4 either as a monomer
or homodimer (5, 38, 39). It has been reported that in the presence of
T3, TR homodimers are destabilized whereas the binding of
TR monomers to DR+4 TRE is favored (38, 39). Our results indicated that
TR alone could bind to the TRE under the conditions of our experiment
and recruit SRC-1 in a T3-dependent manner (compare lanes 1
and 2). Upon addition of excess RXR to TR, heterodimers formed on DR+4
TRE, and these complexes also bound to SRC-1 in a
T3-dependent manner (lanes 7 and 8). Although the binding
of TR monomer and homodimer is weaker than the heterodimer, the amount
of SRC-1 recruited by TR-RXR was found to be only slightly greater than
that recruited by TR alone (compare lanes 2 and 8). This observation
was not surprising, however, because the molar concentration of
DNA-bound receptor complexes in the binding reaction far exceeded that
of SRC-1.
We also observed that RXR by itself failed to bind to DR+4 TRE and, consequently, no significant recruitment of SRC-1 was detected on this DNA element when RXR alone was used (lanes 4 and 5). Interestingly, a marked 9-cis-RA-induced SRC-1 binding to the DNA-bound complex was observed only when a combination of RXR and TR was incubated with DR+4 (lanes 9 and 10). The simplest explanation for this observation is that a TR-RXR heterodimer is the predominant complex on the DR+4 TRE when both receptors are incubated together with this DNA. It is also evident from these studies that each receptor partner within the heterodimer positioned at the TRE binds to its cognate ligand and is able to recruit the coactivator in an entirely ligand-dependent manner.
We next investigated whether T3-induced recruitment of
SRC-1 by a TR-RXR heterodimer was dependent on an intact AF-2 of TR.
For this purpose, we used the mutant TRßC6 with impaired AF-2
activity. As shown in Fig. 5B, whereas a DNA-bound heterodimer
containing a full-length TR displayed strong T3-dependent
binding to SRC-1 (compare lanes 1 and 2), a heterodimer of
TRßC6 and RXR failed to exhibit any interaction with
the coactivator either in the presence or in the absence of
T3 (lanes 3 and 4). These results are in agreement with our
finding that the ligand-dependent AF-2 of TR plays an essential role in
coactivator function. Interestingly, addition of 9-cis-RA
elicited binding of SRC-1 to the TRßC6-RXR heterodimer
(lane 5). The 9-cis-RA-induced binding of SRC-1 to RXR is
presumably mediated via the AF-2 of this receptor. Although the
deletion of TRß AF-2 did not appear to significantly influence the
9-cis-RA-induced binding of RXR to the coactivator, we
cannot rule out allosteric control of the activity of one heterodimeric
partner by the other as recently proposed by Schulman et al.
(40), especially because the coactivator is limiting in our assays. We
nonetheless conclude that, in the presence of the cognate ligand, each
receptor component of TR-RXR heterodimer requires the AF-2 core domain
for ligand-dependent interactions with SRC-1. This scenario is entirely
consistent with the proposed role of SRC-1 as a coactivator in nuclear
receptor-mediated transactivation.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Our studies reveal that an AF-2 core domain that is conserved among
nuclear receptors plays a central role in SRC-1-mediated
transactivation. In this study, we demonstrate that a 20-amino acid
peptide containing the AF-2 core domain efficiently inhibits the
binding of TR LBD to SRC-1. More importantly, mutations of key acidic
(Glu 452, Glu 455, Asp 456) and hydrophobic (Phe 454) amino acids,
which impair the ligand-induced transactivation by TR (9, 31), also
abolish the ability of the peptide to interfere with SRC-1 binding to
TR. These results provide compelling evidence that the conserved AF-2
core helix of the receptor is essential for interactions with a nuclear
receptor coactivator. Recent crystal structural analysis of the LBDs of
TR, RXR, and RAR suggests that the AF-2 core domain, which forms an
amphipathic -helix, also known as helix 12, undergoes a striking
conformational transition upon hormone binding (32, 33, 34). In the
unoccupied receptor, the helix 12 projects into the solvent (32). In
the hormone-occupied receptor, the helix folds back toward the receptor
to form a part of the ligand-binding cavity (33, 34). The helix packs
loosely with the hydrophobic residues facing inward toward the
ligand-binding pocket and the charged residues extending into the
solvent (33, 34). It is conceivable that in the hormone-bound receptor,
the helix presents itself as a binding site for the coactivator. As the
hydrophobic residues interact with the ligand-binding core to stabilize
the domain, certain charged residues might be available for direct
interaction with the coactivator. Alternatively, the ligand-induced
repositioning of helix 12 may trigger a reorganization in other parts
of the receptor, creating an interaction surface for the coactivator.
In this scenario, the AF-2 core domain may play a critical role by
participating in intramolecular interactions with another region(s) of
the LBD to help create the proper coactivator-binding site.
Collingwood et al. (41) have recently reported that a
naturally occurring TRß mutant, in which leucine at position 454 in
the AF-2 core domain is changed to valine, still retains its ability to
respond to SRC-1 in transient transfection experiments. The apparent
discrepancy between our results and those of Collingwood et
al. could arise from the use of different TRß mutants in these
studies. In the case of Collingwood et al. (41), the point
mutant L454V retained moderate interactions with SRC-1. Consequently,
in transient transfection assays, overexpressed SRC-1 could still
potentiate transactivation by this mutant TR, presumably by binding to
the altered AF-2 core domain. In contrast, we employed a truncation
mutant, TRßC6, which lacked several conserved amino acids within
the AF-2 helix. This deletion rendered TRßC6 completely
ineffective in SRC-1 binding and as a result, this mutant receptor
failed to respond to the coactivator (Fig. 1).
Although the failure of the AF-2 mutant, TRßC6, to
respond to SRC-1-mediated activation in our transfection experiments
may appear to be simply due to a disruption in coactivator binding,
more complex scenarios can be considered. Recent studies suggest that
AF-2 core domain may influence the interaction of the receptor with
corepressor molecules (40). Unliganded TR or RAR associates with a
corepressor, which is released from the receptor upon hormone binding
(13, 15, 16). Interestingly, addition of hormone did not affect the
transcriptional activity of the AF-2 mutant (Fig. 1, lane 6). This is
presumably due to a failure to release the corepressor. Previous
studies have shown that TRßC6 binds ligand and
undergoes certain conformational changes in response to ligand binding
(42) but remains a constitutive transcriptional repressor (9).
Therefore, hormone binding is not sufficient to dissociate the
corepressor from the AF-2 mutant. Schulman et al. (40) have
recently proposed that the AF-2 core domain itself participates in the
transition from repressive to active state by an as yet unknown
mechanism. One can speculate that the AF-2 helix helps to maintain a
receptor conformation that allows release of corepressor upon hormone
binding. The loss of this structure in the TRßC6
mutant prevents corepressor release. Based on the crystal structure of
TR, however, deletion of six amino acids from the AF-2 core helix is
not expected to disrupt the compact bound state of the receptor or its
hormone-binding cavity, although subtle alterations in the receptor
conformation cannot be ruled out (34). Alternatively, one can imagine
that the binding of the coactivator to the AF-2 helix of a ligand-bound
receptor serves as the switch that triggers a conformation change,
which in turn facilitates corepressor release. Recent observations by
Glass and co-workers (14) however, imply that interactions with the
corepressor are dominant over the recruitment of coactivators. These
studies also suggest that the DNA response elements can allosterically
control the receptor-corepressor association. Further analyses of the
complex interplay between the DNA-bound nuclear receptor and the
corepressor or coactivator are therefore needed to understand how these
coregulatory molecules modulate the gene-regulatory activity of the
receptor.
A surprising finding of the present studies is the observation that the ligand 9-cis-RA induced binding of SRC-1 to RXR in a TR-RXR heterodimer. Previous studies reported that RXR, once heterodimerized, is incapable of binding its ligand (43, 44). It was proposed that RXR in the heterodimer enhances DNA binding by the partner receptor but is a silent partner in transactivation (3, 43, 44). We, however, observed that heterodimerization of RXR with TR did not prevent its binding to 9-cis-RA or its ligand-dependent interaction with SRC-1. RXR, therefore, not only binds to its ligand in the context of the heterodimer, but also maintains its ability to interact with a transcriptional coactivator. Our results, therefore, suggest a potentially active role of RXR in TR-mediated transcription. The reason for the apparent discrepancy between our results and those published previously is not clear. Our results are in good agreement, however, with the work reported previously by Kersten et al. (45) and more recently by Minucci et al. (46). Whereas Kersten et al. (45) using a fluorescence-based method demonstrated that RXR in a RAR-RXR heterodimer can bind to its ligand (45), Minucci et al. employed a differential proteolytic analysis to conclude that RXR in such a heterodimer not only binds its ligand but also functions as a transcriptionally active partner by enhancing retinoid-dependent gene expression (46). However, the role of RXR in transactivation by the TR-RXR heterodimer is yet to be settled unequivocally. In light of the conflicting results from different laboratories, it appears that RXR within a heterodimer may play either a silent or an active role depending on cell, promoter, and ligand contexts.
An interesting issue raised by the present study is the differential transactivation activities displayed by TR and RXR within the heterodimer in response to the cognate ligand. Although both partner receptors can individually bind to the coactivator in a ligand-dependent manner, T3 but not an RXR-specific ligand, when added alone, activated transcription from a TRE-linked promoter in transfection assays (43, 46). Minucci et al. (46) reported that although a RXR-specific ligand, when added singly, failed to activate a retinoid-responsive promoter, it synergistically stimulated transcription when added together with a RAR ligand. Taken together, these studies indicated that a full display of the transcriptional activity of RXR in a TR-RXR or RAR-RXR heterodimer depends on the ligand of its heterodimeric partner. It appears that the transactivation potential of RXR in the heterodimer is somehow masked in the absence of either T3 or the RAR ligand. It is possible that a corepressor, which is complexed with the unliganded TR or RAR in the heterodimer, blocks access of the coactivator to RXR. In this scenario, addition of either T3 or the RAR ligand, which triggers the release of the corepressor from the cognate heterodimeric partner, is obligatory for RXR-dependent transactivation.
In summary, the results presented here demonstrate that a conserved AF-2 core domain of TR plays an essential role in interactions with a transcriptional coactivator, SRC-1, to regulate the activity of a TRE-linked promoter. Our studies further reveal that RXR, the heterodimeric partner of TR, also possesses the potential to modulate T3-mediated signaling by interacting with the coactivator in a 9-cis-RA-dependent manner. These results provide interesting insights into the complex molecular interactions among multiple nuclear receptors and transcriptional coregulatory factors that determine the response of hormone-regulated promoters to diverse ligand signals.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Plasmid Construction
The SRC-1 cDNA used in this study was originally isolated by
Onate et al. (23). The 3432-nucleotide cDNA extends from
positions -62 to + 3361 and contains an open reading frame of 1061
amino acids. The plasmids pCI-TRß and pCI-SRC-1 expressing human
TRß and human SRC-1, respectively, were constructed by inserting
cDNAs encoding these proteins into vector pCI under the control of
human cytomegalovirus immediate-early promoter (Promega).
pCI-TRßC6 was constructed by incorporating the
previously described mutant TRßC6 (9) in the pCI
vector. The construction of the reporter plasmid TRE-tkCAT has been
described previously (20).
Recombinant Nuclear Receptors and Their Mutants
Recombinant nuclear receptors GST-TRß and GST-RXR
containing GST fused to the amino-terminal sequences of TRß and
RXR, respectively, were expressed in Escherichia coli
BL21. Various deletion mutants of TRß were expressed as GST fusion
proteins: GST-TR (145–456), GST-TR (80–450), and GST-TR (1–260). The
receptor fragments were generated by PCR and engineered into the
bacterial expression vector pGX-2T (Pharmacia, Piscataway, NJ). The
construction, expression, and purification of His6-TRß
and His6-RXR were described previously (11). The protein
concentrations of mutant receptors were determined by comparing the
intensity of Coomassie-stained band of each receptor fusion protein
with that of a known BSA standard in SDS-PAGE.
Transient Transfection Experiments
CV-1 cells were maintained in DMEM (GIBCO BRL, Grand Island, NY)
supplemented with 5% FBS (Hyclone Laboratories, Logan, UT).
Semiconfluent cells were transiently transfected using the calcium
phosphate coprecipitation procedure as described previously (20).
Briefly, 5 x 105 cells were plated on 10-cm tissue
culture dishes in phenol red-free DMEM medium containing 5%
charcoal-stripped serum and after 24–48 h were transfected with
plasmid DNAs. Typically, cells received 10 µg chloramphenicol
acetyltransferase (CAT) reporter plasmid, 2 µg pCI-TRß or
pCI-TRßC6, 2.5 µg pCI-SRC-1 or empty pCI vector, and
2 µg of an internal control plasmid pSV-ßgal (Promega), which
contains the gene for ß-galactosidase enzyme. After 12–14 h of
exposure to the calcium phosphate precipitate, the cells were washed
with PBS and incubated in fresh phenol red-free, serum-free medium with
10-8 M T3 or solvent. Cells were
harvested after 18 h for determination of ß-galactosidase and
CAT activities as described previously (20). The amount of cell extract
used per CAT assay was determined after normalization with respect to
the ß-galactosidase activity. Quantification of the CAT activities
was performed by liquid scintillation analysis of the acetylated
[14C]-chloramphenicol product and the remaining
unacetylated substrate. Each transient transfection experiment was
repeated at least three times.
GST Pull-Down Assay
Each GST fusion protein (0.1–1.0 µg) was immobilized on a
glutathione (GSH)-resin (15 µl), and the immobilized protein was
treated with or without T3 (1 µM) or
9-cis-RA (1 µM) at room temperature for 15
min. 35S-labeled SRC-1 was generated from an expression
vector pBK-CMV-SRC-1 by in vitro transcription-coupled
translation in reticulocyte lysates following manufacturer’s
instructions. An aliquot (10 µl) of the translation mix containing
labeled SRC-1 was then incubated with immobilized GST fusion protein in
a binding buffer containing 20 mM Tris-HCl (pH 7.4), 60
mM NaCl, 1 mM dithiothreitol, 15% glycerol,
and 0.1% NP-40. Binding was allowed to proceed for 1 h at 4 C.
The resin was then washed repeatedly with the binding buffer. The
35S-labeled SRC-1 that was retained on the resin was eluted
by boiling with a gel-loading buffer containing SDS, analyzed by
SDS-PAGE, and visualized by fluorography.
In peptide competition experiments, GST-TRß (0.1 µg) was first immobilized on a GSH-resin followed by treatment with T3 (1 µM). The resin-bound hormone-receptor complex was then incubated with 35S-labeled SRC-1 (10 µl translation mix) in the presence of increasing molar excess of either AF-2 core or control peptide. The 35S-labeled SRC-1 that was retained on the resin after stringent washings was then eluted by boiling with a gel-loading buffer containing SDS and analyzed by SDS-PAGE. The AF-2 core and control peptides were synthesized by the Rockefeller University Biopolymer Facility (New York, NY).
Biotinylated DNA Pull-Down Assay
For this assay, the protocol of Kurokawa et al. (14)
was used with certain modifications. The sequence of the sense strand
of the DR+4 TRE was:
5'-GAAGGGGATCCGGGTAAGGTCACAGGAGGTCACGAA-3'. The
sense oligodeoxynucleotide was biotinylated at the 5'-end by using a
biotin-phosphoramidite during synthesis. Equal amounts of sense and
antisense oligonucleotides were annealed to form the double-stranded
DR+4 element. One microgram (
35 pmol) of the double-stranded
oligonucleotide was coupled to 12 µl avidin-linked agarose for 1
h at room temperature. The immobilized biotinylated TRE was then
incubated with purified His6-TRß (20 pmol) or
His6-RXR (80 pmol) or a combination of both for 30 min
at room temperature. The incubation was performed in a binding buffer
containing 20 mM Tris-HCl (pH 7.4), 60 mM NaCl,
1 mM dithiothreitol, 15% glycerol, and 0.1% NP-40. The
resin was collected by brief centrifugation and washed extensively with
the binding buffer. The resin-bound DNA-receptor complexes were then
treated with either T3 (1 µM) or
9-cis-RA (1 µM) or solvent for 20 min at room
temperature. Ten microliters of translation mix containing
35S-labeled SRC-1 were then added to the receptor-DNA
complexes, and incubation was carried out for 30 min at room
temperature. The resin was again collected and washed repeatedly with
the binding buffer. The bound proteins were eluted with a gel-loading
buffer containing SDS and analyzed by SDS-PAGE. 35S-labeled
SRC-1 was visualized by fluorography. The intensity of the full-length
SRC-1 polypeptide was quantitated by densitometry. The optical
densities were corrected by subtracting from the test values the
nonspecific binding of SRC-1 to the resin in the absence of DNA and the
receptors.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
This work was supported by NIH Grant R01DK-50257–02 (to M.K.B).
Received for publication January 31, 1997. Accepted for publication February 28, 1997.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Nature 375:377–382[CrossRef][Medline]
ligand-binding domain
bound to all-trans retinoic acid. Nature 378:681–689[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  R. Rouf, S. Greytak, E. C. Wooten, J. Wu, J. Boltax, M. Picard, E. C. Svensson, W. H. Dillmann, R. D. Patten, and G. S. Huggins Increased FOG-2 in Failing Myocardium Disrupts Thyroid Hormone-Dependent SERCA2 Gene Transcription Circ. Res., August 29, 2008; 103(5): 493 - 501. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Ma, P. Freitag, J. Zhou, B. Brune, S. Frede, and J. Fandrey Thyroid hormone induces erythropoietin gene expression through augmented accumulation of hypoxia-inducible factor-1 Am J Physiol Regulatory Integrative Comp Physiol, September 1, 2004; 287(3): R600 - R607. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. M. Yen Physiological and Molecular Basis of Thyroid Hormone Action Physiol Rev, July 1, 2001; 81(3): 1097 - 1142. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K.-h. Lin and Y.-h. Wu shen-liang chen Impaired Interaction of Mutant Thyroid Hormone Receptors Associated with Human Hepatocellular Carcinoma with Transcriptional Coregulators Endocrinology, February 1, 2001; 142(2): 653 - 662. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. M. Resnick, D. A. Schreihofer, A. Periasamy, and M. A. Shupnik Truncated Estrogen Receptor Product-1 Suppresses Estrogen Receptor Transactivation by Dimerization with Estrogen Receptors alpha and beta J. Biol. Chem., March 15, 2000; 275(10): 7158 - 7166. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. M. Sladek, M. D. Ruse Jr., L. Nepomuceno, S.-M. Huang, and M. R. Stallcup Modulation of Transcriptional Activation and Coactivator Interaction by a Splicing Variation in the F Domain of Nuclear Receptor Hepatocyte Nuclear Factor 4alpha 1 Mol. Cell. Biol., October 1, 1999; 19(10): 6509 - 6522. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Alen, F. Claessens, G. Verhoeven, W. Rombauts, and B. Peeters The Androgen Receptor Amino-Terminal Domain Plays a Key Role in p160 Coactivator-Stimulated Gene Transcription Mol. Cell. Biol., September 1, 1999; 19(9): 6085 - 6097. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Kraichely, J. J. Collins III, R. K. DeLisle, and P. N. MacDonald The Autonomous Transactivation Domain in Helix H3 of the Vitamin D Receptor Is Required For Transactivation and Coactivator Interaction J. Biol. Chem., May 14, 1999; 274(20): 14352 - 14358. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. A. Schreihofer, E. M. Resnick, A. Y. Soh, and M. A. Shupnik Transcriptional Regulation by a Naturally Occurring Truncated Rat Estrogen Receptor (ER), Truncated ER Product-1 (TERP-1) Mol. Endocrinol., February 1, 1999; 13(2): 320 - 329. [Abstract] [Full Text]
|
|
|
|
|
|
S. A. Onate, V. Boonyaratanakornkit, T. E. Spencer, S. Y. Tsai, M.-J. Tsai, D. P. Edwards, and B. W. O'Malley The Steroid Receptor Coactivator-1 Contains Multiple Receptor Interacting and Activation Domains That Cooperatively Enhance the Activation Function 1 (AF1) and AF2 Domains of Steroid Receptors J. Biol. Chem., May 15, 1998; 273(20): 12101 - 12108. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Zhang, M. Jeyakumar, S. Petukhov, and M. K. Bagchi A Nuclear Receptor Corepressor Modulates Transcriptional Activity of Antagonist-Occupied Steroid Hormone Receptor Mol. Endocrinol., April 1, 1998; 12(4): 513 - 524. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
