| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
INSERM Unité 344 (H.F.-Y., R.F., K.P., E.M.)
Endocrinologie Moléculaire Faculté de Médecine
Necker Paris cedex 15, France 75730
Laboratoire de
Physiologie Générale et Comparée (FY.H., D.F.,
B.D., L.A.) Muséum National d’Histoire
Naturelle UMR 8572 CNRS Paris cedex 05, France 75231
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
(RXR),
inhibited the PRL-induced Stat5a- and
Stat5b-dependent reporter gene expression by up to 60%. This
T3-inhibitory effect studied on Stat5 activity
was partly reversed by overexpression of a TRß1 dominant negative
variant mutated within its nuclear localization signal (TR2A). We next
showed that TRß1 and TR2A in the presence of RXR increased and
decreased, respectively, Stat5 localization into the nucleus regardless
of hormonal stimulation. Thus, our data suggest that TRß1 can be
associated with Stat5 in the cytoplasm and may be involved in Stat5
nuclear translocation. In PRL-treated cells overexpressing
TRß1/RXR, both Stat5 and TRß1 were coimmunoprecipitated,
indicating physical association of the two transcription factors. In
these cells, addition of T3 with ovine (o)PRL
decreased the amounts of total and tyrosine-phosphorylated Stat5 in the
cytoplasm compared with oPRL-treated cells. In the nucleus, no clear
difference was observed on Stat5 DNA-binding after treatment with PRL
and T3 vs. PRL alone in TRß1/RXR
transfected cells. However, antibodies directed against TRß1 lowered
Stat5-DNA binding and addition of the deacetylase inhibitor
trichostatin A (TSA) relieved T3 inhibition on Stat5
transcriptional activity. Thus, we postulated that the negative
cross-talk between TR and Stat5 on target genes could involve histone
deacetylase recruitment by liganded TRß1. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
THs are implicated in numerous processes such as growth, maturation,
development, and metabolism of most tissues in mammals (14, 15, 16).
Addition of T3 leads to the activation of their
nuclear receptors and modulation of target gene expression. Two main
loci encoding thyroid hormone receptor 
(TR) and ß (TRß)
isoforms exist, each isoform differing somewhat in structure (17).
Based on hormone binding assays, it has been proposed that TRs are
nuclear regardless of their ligand status (18). However, the precise
distribution and the functional role of TR isoforms are not yet
entirely understood. Indeed, different functional roles of TRß and
TR were first shown on a rat TRH transcript in vitro
(19), then in vivo by somatic gene transfer (20). Such
findings were confirmed with studies in knockout mice, in which TRß
was shown to be dispensable for normal developmental expression of some
T3-dependent target gene (21, 22, 23). Like many
receptors belonging to the superfamily of steroid/thyroid nuclear
receptors, thyroid hormone receptors (TRs) bind to specific TH-DNA
responsive elements (TRE) upstream of target gene in heterodimeric
complex with the 9-cis retinoid acid receptor (RXR)
(24). Recently, a number of studies have shown a cross-talk between
Stat5a and b with nuclear steroid receptors (25), such as
glucocorticoid receptor (GR) (26, 27), estrogen receptor (ER) (27),
progesterone receptor (PR) (28), and peroxisome proliferator-activated
receptor (PPAR) (29). Unliganded PR or GR are essentially
cytoplasmic. Thus, the observation that progestin or dexamethasone
treatment induces Stat5 translocation into the nucleus, possibly
mediated by the physical association of these liganded nuclear
receptors with inactivated Stat5 (27, 28, 30), is most important.
Nuclear entry of steroid nuclear receptors appears to be mediated by
the interaction of nuclear localization signals (NLSs) within the
proteins and specific NLS-binding proteins in nuclear pores (31, 32, 33).
In contrast, no residue comprising a Stat NLS has been yet
demonstrated, but dimerization of Stat5 molecules upon PRL stimulation
has been postulated to unmask a NLS (34). Thus, the mechanisms
underlying nuclear entry of numerous proteins, such as Stats monomers
or homo- and heterodimers that are present in the cytoplasm, remain to
be clarified.
In the present study, we investigated a possible inhibitory involvement
of T3 in the oPRL-induced activation of a target
gene in the HEK293 cell line. We carried out transfections so as to
express elements of both the PRLR/Stat5a signaling pathway and the
TRß1/RXR pathway. We show that in TRß1/RXR-transfected cells
T3 markedly repressed oPRL-dependent Stat5a or
Stat5b-induced reporter gene expression under the control of a
lactogenic responsive element (LHRE) or the ß-casein promoter. We
also show that overexpression of TRß1/RXR results in an increased
nuclear localization of Stat5a regardless of ligand stimulation.
Conversely, overexpression of a dominant negative TRß1 mutated in its
NLS decreases both Stat5a nuclear localization and
T3-dependent inhibitory effects on PRL-induced
transactivation of the LHRE-thymidine kinase (TK)-luciferase promoter.
We also show the ability of Stat5a to associate with TRß1 in
cytoplasmic extracts of oPRL and oPRL + T3
treated cells. Furthermore, in oPRL-treated cells, addition of
T3 reduces phosphorylated Stat5a in the cytosol.
In electrophoretic mobility shift assay (EMSA) studies, with LHRE
probe, no band was supershifted after addition of antibodies directed
against TRß1. However, TRß1 antibodies lowered Stat5a nuclear DNA
binding when TRß1/RXR was overexpressed. Finally, we show that
addition of the histone deacetylase inhibitor trichostatin A (TSA)
relieved T3 inhibition on Stat5a transcriptional
activity. Thus, our data indicate that TRß1/RXR modifies the
subcellular distribution of Stat5a in unstimulated as well as in
stimulated cells. The inhibitory effect of T3
could result from an heterocomplex formation between Stat5a and
recruitment by liganded TRß1 of a histone deacetylase within the
transcriptional machinery in PRL-stimulated cells.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Inhibits oPRL-Induced Transcriptional
Activity. Without RXR, ovine (o)PRL treatment caused a 10-fold
induction of luciferase expression from the LHRE (Fig. 1A
, lane 2) as compared with controls
(Fig. 1A, lane 1). T3 alone had no effect on
luciferase expression compared with control. However, when
T3 was added with oPRL, it inhibited
LHRE-dependent luciferase expression, by up to 20%, as compared with
oPRL-treated cells.
|
did not modify
the basal or the oPRL-stimulated LHRE-dependent luciferase expression
compared with controls (Fig. 1B
, lane 7). In contrast, addition of
T3 in TRß1/RXR transfected cells resulted in
a higher level of inhibition (<60%) of oPRL-induced
LHRE-dependent luciferase expression (Fig. 1B, lane 8). This
T3-induced inhibition of the PRL transduction
pathway was dose dependent in cells expressing either TRß1 alone or
TRß1/RXR heterodimers. Optimal inhibition occurred at
10-8 M T3
(range from 10-12 M to
10- 6 M, data not shown). As shown
in Fig. 1C, a similar T3 inhibitory
effect on Stat5-mediated LHRE-TK-luciferase activity was obtained
when cells were transfected with GHR following the same experimental
procedure.
We next examined whether this inhibitory effect of
T3 on LHRE-induced reporter gene activity was
specifically dependent on Stat5a or Stat5b (Fig. 2, A and B). For comparison, we studied
the inhibitory effect of TRß1/RXR on the wild-type ß-casein
promoter linked to luciferase reporter gene activity (Fig. 2B). For
this purpose, the same experimental procedure in the presence of
TRß1/RXR with PRLR was applied. As shown in Fig. 2, A and B,
we obtained the same inhibitory effects after stimulation by oPRL +
T3 in cells transiently cotransfected with the
ß-casein promoter and Stat5a or Stat5b relative to those
expressing LHRE (Fig. 2B). However, Stat5a and Stat5b differentially
potentiated LHRE and ß-casein promoter activities. As shown in Fig. 2, comparison of Stat5a/Stat5b-mediated activation shows a more
efficient transactivation potency of Stat5a than Stat5b on both the
ß-casein promoter and the LHRE promoter. Thus, in our studies, Stat5a
seems to be the member of Stat family involved as well as in LHRE or
ß-casein promoter activities. Moreover, the higher sensitivity to
Stat5a-mediated transcriptional activity was more apparent using
LHRE-luciferase promoter than with the ß-casein promoter. Indeed,
similar induction by oPRL was observed in control and cells
overexpressing Stat5b with LHRE (Fig. 2A), whereas oPRL was unable to
potentiate ß-casein promoter activity in control cells in the absence
of added Stat5a or Stat5b (Fig. 2B).
|
). Luciferase
expression estimated after cotransfection of GFP-TRß1 with
RXR, PRLR, and the LHRE-TK-luciferase in the presence of
T3, oPRL, and T3 + oPRL
showed similar fold-induction of luciferase activity, as previously
observed with the wild-type TRß1 (compare Fig. 3A with Fig. 1B).
However, transfection of the NLS-mutated GFP-TR2A sharply attenuated,
from 60% to 20%, the T3-inhibitory effect
on PRL-induced LHRE-TK-luciferase expression obtained with the
wild-type GFP-TRß1 (Fig. 3B).
|
on Stat5a DNA Binding (Fig. 4, A and C) or
with PRLR and Stat5a only (Fig. 4B). Twenty four hours later the cells
were stimulated over a period of 30 min with different combinations of
T3 (10-8 M)
and/or oPRL (1 µg/ml). No Stat5a DNA binding was seen in the
absence of oPRL (with or without T3) (Fig. 4, A
and B, lanes 1–10). When cells were stimulated with oPRL, a specific
association of Stat5a to its DNA binding site was identified (Fig. 4, A
and B, lanes 11–20). Specificity of Stat5a for its consensus sequence
was verified by competition with a 100-fold excess of unlabeled
oligonucleotides (Fig. 4, A and B, lanes 12 and 17). Adding 1 µg of
mStat5a antibodies resulted in the appearance of a supershifted band
containing Stat5a bound to DNA (Fig. 4, A and B, lanes 13 and 18). With
the LHRE probe, no supershifted or shifted material was apparent after
addition of human (h)TRß1 antibodies in extracts of control or
stimulated cells by T3 (Fig. 4, A and B, lanes 4,
5, 9, and 10). This absence of shift seems to be in agreement with a
lack of an apparent canonical consensus TRE on LHRE. In nuclear
extracts from oPRL and oPRL + T3 treated cells,
addition of increasing amounts of hTRß1 antibodies repeatedly
diminished the intensity of the shifted band when TRß1 and RXR
were overexpressed. As illustrated in Fig. 4A, quantification of the
shift showed a reduction of 12% and 30%, respectively, in
oPRL-treated cells (lanes 14 and 15) and in oPRL +
T3-treated cells (lanes 19 and 20). HumanTRß1
antibodies did not affect the shifted band when TRß1 and RXR were
not overexpressed in cells (Fig. 4B, lanes 14, 15, 19, and 20). Such
reduced binding of shifted materials (Fig. 4A) in the presence of
TRß1/RXR is not fully understood but might represent an unstable
low-affinity heterocomplex containing Stat5a, TRß1, and RXR
associated within the LHRE consensus sequence and/or loosening of
specific Stat5a-DNA binding. Similar results were obtained when cells
were transfected with Stat5b instead of Stat5a (data not shown).
Moreover, addition of TRß1/RXR apparently also increased Stat5a
DNA binding as shown in our EMSA when nuclear extracts originated from
T3 + oPRL- or T3-treated
cells, in comparison to cells transfected without TRß1/RXR (Fig. 4A, lanes 11 and 16; compare with Fig. 4B, lanes 11 and 16).
|
). TRß1, RXR,
and Stat5a were overexpressed in 293 cells stimulated during 30
min with T3 and T3 + oPRL.
In both groups specific shifts were observed in nuclear cells
extracts (Fig. 4C, lanes 1–8). In these two treatment conditions the
shifted material was supershifted in presence of hTRß1 antibodies
(Fig. 4C, lanes 4 and 8). A marked inhibition of the shift and
supershift was apparent in nuclear extracts from oPRL +
T3 (Fig. 4C, lanes 5–8) compared with
T3-treated cells (lanes 1–4). When mStat5a
antibodies were used, no supershifted materials could be seen
regardless of the hormonal status of nuclear extracts used (Fig. 4C, lanes 3 and 7). For all combinations of treatment, the specificity of
the shifted bands was validated by addition of 100-fold excess of
unlabeled TRE (Fig. 4C, lanes 2 and 6).
Effect of T3 and oPRL on Stat5a and TRß1
in Whole Cell Extracts
Using Western blotting analysis we next investigated the
interactions between Stat5a and TRß1 and their phosphorylation
status. In all the transfection studies shown in Fig. 5, A and B, cells were cotransfected with
PRLR, Stat5a, RXR, and TRß1. As shown in Fig. 5A, the effects of
oPRL and T3 on Stat5a phosphorylation was
estimated after immunoprecipitation with anti-mStat5a or anti-hTRß1
antibodies and immunoblotted with antiphosphotyrosine antibodies.
Tyrosine phosphorylation of Stat5a was only apparent in cells
stimulated by oPRL (1 µg/ml) (Fig. 5A, Stat5a arrowhead).
A lower level of activated Stat5a tyrosine phosphorylation was
shown in the cytosol of cells incubated simultaneously with both oPRL
and T3 (Fig. 5A, Stat5a arrowhead;
compare lanes 5 and 7). When the same extracts were immunoprecipitated
by anti-hTRß1 antibodies and immunoblotted with antiphosphotyrosine
antibodies, phosphorylated TRß1 was apparent in all extracts,
regardless of treatment (Fig. 5A, TRß1 arrowhead). To
further evaluate whether or not coexpressed TRß1 might complex with
Stat5a, the ability of TRß1 to coimmunoprecipitate with Stat5a was
tested in whole cell extracts after addition of monoclonal hTRß1
antibodies and immunoblotting with polyclonal mStat5a antibodies (Fig. 5B). Comparisons were performed using the same extract
immunoprecipitated and probed with mStat5a antibodies. Western blot
analysis after single immunoprecipitation and immunoblotting with
mStat5a antibodies revealed one specific band of 92 kDa. In whole cell
extracts immunoprecipitated by hTRß1 antibodies and probed with
mStat5a antibodies, Stat5a occurred only in cells stimulated by
oPRL or oPRL + T3 (Fig. 5B, Stat5a
arrowhead). No apparent association of TRß1 with Stat5a
could be identified in unstimulated cells (control) or in cells
incubated with T3 alone. This TRß1-Stat5a
association was also observed in cytoplasmic extracts but was
undetectable in the nucleus (data not shown). Expression of endogenous
Stat5a and Stat5b in untransfected 293 cells was estimated in whole
cell extracts immunoprecipitated and probed with mStat5a or mStat5b
antibodies. As shown in Fig. 5C, we were not able to detect Stat5b,
whereas low levels of Stat5a are present.
|
on Stat5a Nuclear Translocation affects the
subcellular distribution and nuclear translocation of Stat5a. All
transient transfections performed with PRLR and Stat5a were
complemented with TRß1 and RXR (+TR) or control plasmid
pcDNA3 (-TR). Western blots were performed on
cytosolic (Fig. 6A) and nuclear extracts
(Fig. 6B). Extracts were immunoprecipitated and immunoblotted with
anti-mStat5a antibodies. In the presence of TRß1/RXR, decreased
Stat5a levels were apparent in all cytosolic extracts compared with
controls (Fig. 6A). In the nucleus, overexpression of TRß1/RXR
induced an increased nuclear localization of Stat5a (Fig. 6B). In each
group of transfected cells (+TR or -TR), hormonal treatments did not
apparently modify Stat5a expression in cytosolic (Fig. 6A) or nuclear
extracts (Fig. 6B).
|
, A and B). Phosphorylated Stat5a was observed only in cytosolic
and nuclear extracts from cells stimulated with oPRL or with oPRL +
T3 (Fig. 6A, lanes 5–8). In oPRL-treated cells
overexpressing TRß1/RXR, a decreased amount of phosphorylated
Stat5a was apparent in the cytoplasm compared with cells without
TRß1/RXR (Fig. 6A, lane 6 vs. lane 5, pTyr
arrowhead). We found an opposite effect in the nucleus (Fig. 6B, lane 6 vs. lane 5, pTyr arrowhead). This
trend toward Stat5a diminution in the cytoplasm was further lowered in
the presence of T3 + oPRL in both groups of
transfected cells (Fig. 6, A and B, lane 8 vs. 7, pTyr
arrowhead). In contrast, increased amounts of phosphorylated
Stat5a and Stat5a protein could be found in the corresponding nuclear
compartment, when cells were transfected with TRß1/RXR and treated
by oPRL or T3 + oPRL (Fig. 6B, lanes 6 and 8,
pTyr arrowhead) than in cells without TRß1/RXR (Fig. 6B, lanes 5 and 7, pTyr arrowhead) or cells unstimulated by
oPRL (lanes 1–4).
In order to ascertain how the NLS-mutated GFP-TR2A/RXR partly
abrogated the T3-inhibitory effect on PRL-induced
LHRE-TK-luciferase expression, we evaluated its role on the subcellular
distribution of Stat5a (Fig. 6, C and D). Cytoplasmic and nuclear
extracts were immunoprecipitated and immunoblotted with anti-mStat5a
antibodies. No modification in the amount of Stat5a in the cytosol was
evidenced when GFP-TR2A was overexpressed between the different types
of treatment (Fig. 6C). Overexpression of GFP-TR2A/RXR did not
induce nuclear localization of Stat5a in control or
T3-treated cells (Fig. 6C). Thus, overexpression
of GFP-TR2A did not apparently affect the amount of Stat5a present in
cytosolic or nuclear extracts.
The Effects of Histone Acetylation on
T3-Mediated Transcriptional Repression
To assess whether or not histone deacetylase activity is involved
in the T3-mediated repression of Stat5a-dependent
transcription, we tested the ability of TSA, a potent inhibitor of
histone deacetylase, to affect Stat5a-induced promoter activity.
Hormonal treatments were simultaneously performed in 293 cells with or
without TSA. As shown in Fig. 7, in the absence of TSA, T3 addition inhibited
PRL-induced LHRE activity by up to 40%. Addition of TSA (1000
nM) potentiated oPRL-induced LHRE promoter activity as
illustrated by the 10-fold increase in luciferase expression (Fig. 7A).
At the dose of TSA used in our study, TSA had no significant effect on
promoter activity in the absence of oPRL. In contrast, a clear
dose-response effect on oPRL-induced LHRE-TK-luciferase expression was
obtained with increased amounts of TSA (50 to 1000 nM)
(data not shown). As shown in Fig. 7B, T3-dependent repression was almost fully
abrogated in the presence of TSA. Thus, our results suggest that
recruitment of histone deacetylase activity by liganded TRß1 might be
part of its inhibitory effect on LHRE promoter activity.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
heterodimer, T3 induces a clear
inhibition of PRL-induced transcription. This
T3-dependent TRß1 inhibition was apparent with
both Stat5a and Stat5b, using PRLR and GHR on a construct containing
six copies of LHRE. These observations could partially explain the
observation that high levels of TH have been shown to inhibit
PRL-induced mammary gland development and milk production in mammals
(6). In recent years, a number of nuclear steroid receptors have been
identified that are critical for the transcriptional responses
transduced by Stat5. It is well established that phosphorylation of
Stat5 plays a central role in stabilizing the processes of dimerization
and nuclear translocation upon activation of PRL-induced gene
expression (37). However, the mechanisms underlying the formation of an
active Stat dimer and its translocation to the nucleus have not been
fully elucidated. No NLSs have been yet identified on Stat molecules,
raising the question of how they are imported to the nucleus (38).
In our model, after immunoprecipitation and Western blotting with
Stat5a antibodies, we showed that TRß1/RXR, respectively,
decreased Stat5a protein level in the cytosol and increased Stat5a
nuclear protein localization, suggesting their association during the
process of induced nuclear translocation. This process occurred even in
control cells without oPRL and T3 and in
T3-treated cells. However, we were not able to
show in either of these groups the occurrence of stable
association between Stat5a and TRß1 or Stat5 DNA binding on the LHRE.
Thus our data suggest that overexpression of TRß1/RXR induced
protein-protein interaction with Stat5a and increased nuclear
translocation and sequestration of a subpopulation of nonactivated
Stat5 monomer before any treatments in our model, similar to the
model of GR/Stat5 interaction (27). Indeed, similar LHRE- driven
luciferase activity was seen in the presence or absence of
TRß1/RXR in control or oPRL-treated cells (our unpublished
observations). Moreover, expression of RXR in the absence of TRß1
did not induce inhibition of Stat5a transcriptional activity or changes
in nuclear translocation (data not shown). However, RXR potentiated
TRß1 interaction with translocation of inactivated Stat5a into the
nucleus.
The nuclear localization of TRß1 in the absence of ligand remains controversial (36). Recently, an equal distribution of TRß1 isoforms between the nucleus and cytoplasm has been observed in the absence of ligand (36, 39). Alanine substitution of Lys184-Arg185 in the D-domain of the GFP-TR2A mutant, between the DNA and hormone binding domain of TRß1, was shown to inhibit both its nuclear translocation and T3-mediated transactivation (36). Using this mutated construct of TRß1 in our study, a marked attenuation of the T3-inhibitory effect on Stat5a transduction pathways was observed. This loss of T3 inhibition was associated in our Western blotting studies with a lowered amount of Stat5a translocated or sequestrated in the nucleus. These data suggest that nuclear translocation and/or sequestration of latent Stat5a monomers, regardless of the hormonal treatment, might be closely linked to unliganded TRß1 subcellular distribution in our study.
Stat5a phosphorylation plays a critical role in Stat5a dimerization,
nuclear translocation, and DNA binding (3). We show that Stat5a and
TRß1 could be coimmunoprecipitated in whole cell extracts only when
Stat5a phosphorylation was induced by oPRL treatment. Furthermore,
expression of TRß1/RXR in oPRL-treated cells was associated with
increased amount of phosphorylated Stat5a in the nuclear compartment as
evidenced in our Western blotting studies. Thus, oPRL-induced Stat5a
phosphorylation and dimerization in our study likely strengthened the
binding affinity involved in the physical interactions between Stat5a
and TRß1 molecules.
Several functional interactions involving Stat5a or Stat5b and nuclear receptors from the steroid/thyroid receptor family members have been published (25, 26, 27, 28, 29, 30). Among them, reports on GR and PR suggested that those nuclear receptors and Stat might converge and physically interact in the cytoplasm, thereby enhancing Stat5a entry into the nucleus (27, 28). Moreover, overexpression of GR enhanced Stat5 tyrosine phosphorylation and DNA binding (27), an effect possibly related to the predominantly cytoplasmic subcellular localization of GR (40). Interestingly, when overexpressing the GR, Cella et al. (30) found 1.5 times more GR or Stat5 present in the nucleus, respectively, in cells stimulated by oPRL or dexamethasone compared with unstimulated cells, suggesting that inactive partner recruitment occurred before nuclear translocation. Similar results were obtained with the PR (28). Progesterone alone in the T47Dco breast cancer cell line also induced nuclear translocation of Stat5, suggesting that liganded PR, like GR, may be implicated as cotransporter of Stat5 in the nucleus.
Although increased phosphorylated Stat5a was present in the nucleus of
oPRL cells transfected with TRß1/RXR, we were unable to detect,
with EMSA analysis and specific antibodies against TRß1, any
supershifted molecular complex representing a stable association of
Stat5a with TRß1. Nevertheless, addition of increased amounts of
TRß1 antibodies disrupted the shifted DNA binding of Stat5a from oPRL
and T3 + oPRL nuclear extracts when TRß1/RXR
were overexpressed but not in their absence. This latter observation
clearly supports the recruitment of TRß1 by activated Stat5a dimer
within its cognate DNA binding site. However, our data differ somewhat
from those obtained on the cross-talk between ER and Stat5.
Overexpression of ER in the presence of oPRL and estrogen was
associated with a clear inhibition of Stat5-DNA binding (27). In these
studies, the inhibitory effect of ER on Stat5 DNA binding was shown to
arise from a decreased Stat5a phosphorylation. Our data also differ
from the interactions between Stat5a and b with GR, which are among the
best studied compared with other members of this nuclear receptor
superfamily (26, 27, 30, 41, 42). GR does not bind to, but directly
associates with, Stat5 on its DNA response element (LHRE) in the
promoter of the ß-casein gene. Treatment with oPRL and dexamethasone
induced the formation of a GR/Stat5/DNA complex as confirmed in EMSA
where supershifted materials were seen using specific antiserum against
either Stat5 or the GR (26). Thus, the hypothesis we put forward to
explain such discrepancies in the literature is that the complex
containing Stat5a and TRß1 or Stat5 and ER on a Stat5 responsive
element (LHRE) possibly interacts with a lower or a weaker DNA binding
affinity compared with the heterocomplex containing GR and Stat5a. It
is important to mention that among members of the nuclear
steroid/thyroid receptor family, GR and PR differ from thyroid and ERs
in their recognition of DNA response elements (24, 43, 44). Several
half-GREs have been mapped in the ß-casein promoter (45), but no
half-site TRE that binds TR and ER (AGGTCA) has been reported.
Interestingly, PRs and GRs potentiated (25), whereas ER repressed (25),
Stat5 transcriptional activity as TRß1. However, we do not
fully understand how T3 or estrogen negatively
coordinates Stat5 DNA binding and gene transcription through its
responsive element. However, the data presented here are consistent
with the hypothesis that TRß1/RXR and T3 in
the presence of PRL enhance dimerization of Stat5a in a functional
antagonistic heterocomplex with TRß1, thus increasing nuclear
translocation and DNA binding activity of this complex at the LHRE site
present in the promoter of the ß-casein gene. Such increased
DNA-binding activity of Stat5a dimer, as a result of increased
Stat5a/TRß1/RXR nuclear translocation and/or its decreased nuclear
turnover, might represent one of the functional consequences in
TRß1/RXR-mediated inhibition at the promoter level. A similar
model was demonstrated with GR and Stat5, PRL-induced Stat5
dimerization resulting in an increased nuclear localization of
unliganded cytoplasmic GR (27). Furthermore, physical interactions
between TRß1 and Stat5a might be only one component, the
transcriptional antagonism exhibited by this complex likely involving
recruitment of other partners to be operable on PRL-dependent target
genes. The experiments presented here strongly suggest the occurrence
of protein/protein interactions between Stat5a and TRß1 and the
transcriptional machinery in oPRL- and oPRL + T3
-stimulated cells. Consequently, the carboxy-terminal ligand binding
domain of TRß1 was likely not involved in this association, since a
clear inhibition of oPRL transcriptional activity could be obtained
only upon addition of T3.
In our EMSA analysis using an idealized TREpal we repeatedly observed
DNA binding of TRß1 supershifted by hTRß1 antibodies, in nuclear
extracts from T3- and T3 +
oPRL-treated cells. Additionally, a clear inhibition of the shifted and
supershifted TRE probe was seen in oPRL and oPRL +
T3 nuclear extracts compared with those
originating from T3-treated cells. This
observation seems relevant since GH significantly inhibited a
T3-induced TRE-CAT transcriptional activity when
TRß1 was overexpressed (46). Thus, it appears likely that the
physical cross-talk of TRß1 and Stat5a in our studies also affects
thyroid hormone transduction pathway. Such cross-talk observed only
during hormonal treatment might be stabilized by multiprotein complex
components in the transcriptional machinery such as nuclear corepressor
(NcoR) and SMRT (silencing mediator of retinoid and thyroid hormone
receptor) with associated HDACs or sin3 and Rpd3 histone deacetylase
activity or nuclear coactivator (NCoA) with histone acetyltransferase
activity (47). Members of the Stats family have been shown to interact
directly with the related coactivator p300 or with N-myc interactor
(Nmi), a new auxiliary protein downstream of CREB binding protein
(CBP)/p300 (48). CBP/p300 or the steroid receptor coactivator SRC-1
does not appear to participate in the Stat5a-mediated suppression of
the inhibitory effect of Stat5a on glucocorticoid or PPAR responsive
elements (41, 46). However, we found in our present study that TSA,
a potent deacetylase inhibitor, potentiated PRL-induced gene
transcription and abrogated T3 inhibition on
Stat5a transcriptional activity. TSA markedly stimulated PRL
transduction pathway, suggesting that interplay between NCoA and NCoR
both occurred in Stat5a transactivation. TRH and TSH both have a
negative responsive element (nTRE) allowing negative feedback
regulation by TRß1 in the presence of T3 (19, 20, 49). Characterization of nTRE consensus nucleotide sequences
differs in TSH and TRH promoter (49, 50). Depending on the nTRE present
at the promoter level, unliganded or liganded TRß1/RXR exerted its
negative regulation by recruitment of histone deacetylase. Thus, in
our model, liganded TRß1 could allow the recruitment of a complex
that contains histone deacetylase activity in its inhibitory effects on
Stat5a transcription activity as already postulated in the negative
feedback regulation of TSH (49).
Finally, the present data support the concept that those nuclear receptors affecting Stat5 signaling have two key characteristics: they all contain putative NLSs and all interact with the transcriptional machinery either directly or indirectly (38). We postulate that the interactions between Stat5a and TRß1 described in the present study may be part of the mechanisms implicated in the TH-induced inhibition of mammary gland development. Similar antagonistic cross-talk, but on a T3 responsive element, might be also part of the antagonistic effect elicited by PRL during TH-induced metamorphosis in amphibians (13).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
was provided by D. Mangelsdorf (Dallas, TX).
PCI-nGFP-TRß1 and the NLS-mutated construct pCI-nGFP-TR2A were
provided by Dr. S.-Y. Cheng (36) (NIH, Baltimore, MD). CMV
ß-galactosidase cDNA was used as internal control to normalize
LHRE-luciferase and ß-casein-luciferase expression in cell-transient
transfection assays (52), and pcDNA3 vector (Promega Corp., Madison, WI) was used for DNA complementation. Ovine PRL
(NIDDK-oPRL-20) was a gift from the National Hormone and
Pituitary/NIDDK program (Baltimore, MD), T3 was
purchased from Sigma (St. Louis, MO), and TSA was
purchased from Wako Pure Chemical Industries Ltd.
(Richmond, VA). Monoclonal anti-hTRß1 (sc-737 or sc-737X for EMSA),
polyclonal anti-mStat5a (sc-1081 or sc-1081X for EMSA), and polyclonal
anti-mStat5b (sc-835-G or sc-835X for EMSA) used for Western blot were
purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz,
CA). We used also a monoclonal antiphosphotyrosine (4G10; Upstate Biotechnology, Inc., Lake Placid, NY). Protein levels were
determined using Bio-Rad protein assay reagent (Bio-Rad Laboratories, Inc. Hercules, CA) (56).
Cell Culture and Transfections
The human embryonic kidney fibroblast HEK 293 cell line was used
for transient transfection and maintained in DMEMnut
F12 complemented with 10% FCS, 2 mM glutamine, and
antibiotics. Cells were split into six-well plates at a density of
800.000 cells per well (
70% confluency) in DMEM (4.5 g/liter of
glucose, 10% FCS, 2 mM glutamine, antibiotics) before
being transiently transfected using the calcium phosphate technique
(57) with 0.1 µg/well of LHRE-TK-luciferase and ß-galactosidase,
0.05 µg/well of PRLR, TRß1 (GFP-TRß1 or GFP-TR2A), and RXR.
Within 15 h after transfection, the cells were starved in the
absence of FCS in DMEM (4.5 g/liter of glucose) for an additional
7 h of starvation. Cells were then stimulated during 10–12 h for
functional test by oPRL (1 µg/ml), T3
(10-8 M), or TSA (1000
nM) added alone or in combination to culture medium.
Luciferase activity was measured in relative light units (RLU) and
normalized by the estimated ß-galactosidase activity, as previously
described (52).
Cytosolic and Nuclear Cell Extracts Preparation
HEK293 cells were grown in 100-mm culture dish. Transient
transfection was performed as above using 1 µg of cDNA encoding the
PRLR, Stat5a, TRß1, and RXR. After transfection cells were treated
during 30 min with oPRL (1 µg/ml) and with T3
(10-8 M). Cells were subsequently
lysed with 1 ml of lysis buffer (10 mM Tris-HCL, pH 7.5, 5
mM EDTA, 150 mM NaCl, 30 mM sodium
pyrophosphate, 50 mM NaF, 1 mM
Na3VO4, 10% glycerol, 0.5% Triton X-100)
containing a cocktail of protease inhibitors (1 mM
phenylmethylsulfonyl fluoride (PMSF), 1 µg/ml pepstatin, 2 µg/ml
leupeptin, 5 µg/ml aprotinin) for 30 min at 4 C. All extracts were
aliquoted and frozen at -80 C.
Cells were washed with 1x PBS to remove media and serum, scraped, and collected in 1 ml of 1x PBS with 1 mM of Na3VO4. Cells were pelleted and resuspended by repeated pipetting in 4 volumes of hypotonic buffer (10 mM HEPES-KOH, pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM dithiothreitol) supplemented with 1 mM Na3VO4, 20 mM NaF, and 0.2 mM PMSF. The cytoplasmic supernatant was collected and nuclear pellet was resuspended in 1 volume of hypertonic buffer on ice 10–15 min (20 mM HEPES-KOH at pH 7.9, 1.5 mM MgCl2, 25% glycerol, 420 mM NaCl, 0.2 mM EDTA) and complemented with 0.2 mM PMSF, 0.2 mM phosphatase, 20 mM NaF, and 1 mM Na3VO4. Extracts were centrifuged at 10.000 rpm at 4 C for 10 min to remove nuclear debris. All extracts were aliquoted and frozen at -80 C. Each aliquot was thawed only once.
Immunoprecipitation and Western Blotting
For immunoprecipitation, 500 µl of lysates were incubated with
2 µl of anti-mStat5a antibodies (0.8 µg/ml) and anti-hTRß1 (1.2
µg/ml) and mixed with 30 µl of Protein A-Agarose or Protein G-plus
(Santa Cruz Biotechnology, Inc., TEBU) for 2 h
with shaking. The immunoprecipitate was collected by
centrifugation, washed in lysis buffer, and boiled for 5 min in
sample buffer (125 mM Tris, pH 6.8, 5% SDS, 10%
ß-mercaptoethanol, 20% glycerol). Immunoprecipitated proteins
were separated by SDS-PAGE (7.5%) for 2 h at 30 mA. The gel was
transferred onto polyvinylidene difluoride (PVDF) transfer membrane
(Polyscreen, NEN Life Science Products, Boston,
MA), and probed with anti-mStat5a or anti-mStat5b (Upstate Biotechnology, Inc., 1:1000), anti-hTRß1 (Upstate Biotechnology, Inc., 1:1000), antiphosphotyrosine
(Upstate Biotechnology, Inc., 1:10000) conjugated with an
antimouse or an antirabbit IgG-horseradish-peroxidase (1:5000)
and visualized by the ECL detection system (Amersham Pharmacia Biotech, Arlington Heights, IL) for 30 min at room
temperature.
Electrophoretic Gel Mobility Shift Assay (EMSA)
The same experimental procedure was used for cell culture
as described above. The single-strand sequences of LHRE used in our
assay were, respectively: 5'-AGGCTGACTTCTTGGAATTAGCCC-3'(sense),
5'-GGGCTAATTCCAAGAAGTCAGCCT-3'(antisense). These probes were provided
by Genaxis (Saint-Cloud, France). The palindromic TREpal was used for
TRß1 DNA-binding. This sequence contain a TRE underlined as follow:
5'-AGGCTCAGGTCA TGACCTGAGCCC-3' (sense) and
5'-GGGCTCAGGTCATG ACCTGAGCCT-3' (antisense).
The two complementary strands were annealed and end-labeled with
[32P]-dATP by the Polynucleotide Kinase T4
and purified by P-6 Micro Bio spin columns (Bio-Rad Laboratories, Inc.). Eight micrograms of nuclear extracts were used per
reaction, 2 µg of poly(dI)-poly(dC) (Pharmacia Biotech,
Piscataway, NJ) were added to 2 µl of migration buffer
(50 mM Tris-HCl, pH 7.9, 12.5 mM
MgCl2, 1 mM EDTA, 1 mM
dithiothreitol, and 20% glycerol), and total volume was adjusted at 20
µl. The sample were incubated 30 min at 4 C with 40,000 cpm of probe.
When competitor was added to specific concentration as indicated, they
were included with extracts during the incubation of labeled probe. For
all antibodies, 1 µg to 4 µg/reaction were used and incubated 30
min before the reaction binding. Reactions were resolved on 4%
polyacrylamide (38:2 acrylamide/bis-acrylamide, Interchim) gels
containing 0.25x TBE. Gels were pre-run at 50 V overnight and run 1 to
1.30 h at room temperature. They were dried on wathman and
analyzed with Kodak (Eastman Kodak,
Rochester, NY) film after 48 h exposure at -80 C. We used the
ImageQuant logiciel (Molecular Dynamics, Inc., Sunnyvale,
CA) for calculating the percentage of shift intensity.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication August 18, 1999. Revision received May 31, 2000. Accepted for publication June 1, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
and TRß in the control of
thyroid hormone production and post-natal development. EMBO J 18:623–631[CrossRef][Medline]
- and ß-receptor
subtypes in mediation of T3 functions during
postnatal murine intestinal development. Gastroenterology 166:1367–1378[CrossRef]
(PPAR) signaling pathways.
J Biol Chem 274:2672–2681) and glucocorticosteroid receptor (GR).
Exp Cell Res 242:255–264[CrossRef][Medline]
-mediated signaling.
Cell 96:121–130[CrossRef][Medline]
This article has been cited by other articles:
 |
|
 |
|
  M. B. Hapon, A. B Motta, M. Ezquer, M. Bonafede, and G. A Jahn Hypothyroidism prolongs corpus luteum function in the pregnant rat Reproduction, January 1, 2007; 133(1): 197 - 205. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B Anguiano, R Rojas-Huidobro, G Delgado, and C Aceves Has the mammary gland a protective mechanism against overexposure to triiodothyronine during the peripartum period? The prolactin pulse down-regulates mammary type I deiodinase responsiveness to norepinephrine J. Endocrinol., November 1, 2004; 183(2): 267 - 277. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. M. Litterst, S. Kliem, D. Marilley, and E. Pfitzner NCoA-1/SRC-1 Is an Essential Coactivator of STAT5 That Binds to the FDL Motif in the {alpha}-Helical Region of the STAT5 Transactivation Domain J. Biol. Chem., November 14, 2003; 278(46): 45340 - 45351. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Shipley and D. J. Waxman Down-Regulation of STAT5b Transcriptional Activity by Ligand-Activated Peroxisome Proliferator-Activated Receptor (PPAR) {alpha} and PPAR{gamma} Mol. Pharmacol., August 1, 2003; 64(2): 355 - 364. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. E. Timsit and D. S. Riddick Stimulation of Hepatic Signal Transducer and Activator of Transcription 5b by GH Is Not Altered by 3-Methylcholanthrene Endocrinology, September 1, 2002; 143(9): 3284 - 3294. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. R. Simon, J. Fortune, M. Iwahashi, and E. Sutherland Sexual dimorphic expression of ADH in rat liver: importance of the hypothalamic-pituitary-liver axis Am J Physiol Gastrointest Liver Physiol, September 1, 2002; 283(3): G646 - G655. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. F. Bunn, J. A. Neidig, K. E. Freidinger, T. A. Stankiewicz, B. S. Weaver, J. McGrew, and L. A. Allison Nucleocytoplasmic Shuttling of the Thyroid Hormone Receptor {{alpha}} Mol. Endocrinol., April 1, 2001; 15(4): 512 - 533. [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 |
