(Received for publication, November 20, 1996, and in revised form, January 7, 1997)
From the Molecular Endocrinology Group, the
Laboratory of Signal Transduction, NIEHS, National Institutes of
Health, Research Triangle Park, North Carolina 27709, and the
§ Department of Physiology, Dartmouth Medical School,
Lebanon, New Hampshire 03756
Although studies have shown that the mouse glucocorticoid receptor (mGR) contains eight phosphorylation sites (Bodwell, J. E., Ortí, E., Coull, J. M., Pappin, D. J. C., Smith, L. I., and Swift, F. (1991) J. Biol. Chem. 266, 7549-7555), the effect of phosphorylation on receptor function is unclear. We have examined the consequences of single or multiple phosphorylation site mutations on several properties of mGR including receptor expression, ligand-dependent nuclear translocation, hormone-mediated transactivation, ligand-dependent down-regulation of mGR, and receptor protein half-life. Mutations had little effect on receptor expression, subcellular distribution, ligand-dependent nuclear translocation, or on the ability to activate hormone-mediated transcription from a complex (murine mammary tumor virus) promoter. In contrast, the phosphorylation status of the mGR had a profound effect on the ability to transactivate a minimal promoter containing simple glucocorticoid response elements after hormone administration. Similarly, ligand-dependent down-regulation by glucocorticoids of both receptor mRNA and protein was abrogated in mutants containing three or more phosphorylation site alterations. Finally, we show that the phosphorylation status of mGR has a profound effect on the stability of the glucocorticoid receptor protein. Receptors containing seven or eight mutated sites have a markedly extended half-life and do not show the ligand-dependent destabilization seen with wild type receptor. These data show that receptor phosphorylation may play a crucial role in regulating receptor levels and hence control receptor functions.
The glucocorticoid receptor (GR)1 is a member of a family of intracellular ligand-inducible transcription factors termed the steroid/vitamin D/retinoic acid superfamily (2, 3). All of these receptors share certain structural and functional features, such as an amino-terminal transactivation domain, a central Zn2+ finger DNA binding domain, and a carboxyl-terminal ligand binding region. Upon exposure to a specific ligand, the receptor undergoes a transformation process and binds with high affinity to its cognate sequence-specific DNA response element. After DNA binding, the receptor interacts with the basal transcription complex and alters transcription of hormone sensitive genes (4). Many of these proteins, including GRs, are phosphorylated and can become hyperphosphorylated after binding by ligand (1, 5, 6). The role of receptor phosphorylation in receptor function, however, is controversial. Earlier studies suggested that mutation of single or multiple phosphorylated sites in mouse or human GR had little effect on the ability of these mutants to activate transcription (7, 8). However, promoter complexity and context may affect the ability of various phosphorylated forms of the GR to regulate transcription. Therefore, it remains possible that other transcription factors harbored in complex promoters could compensate for potential impaired effects of dephosphorylated GRs.
Another feature shared among some members of the steroid receptor family is that of hormone-mediated down-regulation (9). Our laboratory has shown previously that GR down-regulation occurs primarily at the level of transcription, is restricted to ligands of GRs, and is reversible on hormone withdrawal (10, 11). The genetic elements responsible for hormone-mediated autoregulation reside within the exons of the GR cDNA (12). In addition, GR protein exhibits ligand-dependent destabilization, but little is known about GR turnover or what stimuli alter receptor half-life. However, recent evidence suggests that phosphorylation plays a role in the turnover of other proteins (13-15).
Here we report that decreased phosphorylation in the mouse GR (mGR) decreases transactivation of a hormone-responsive simple promoter. In addition, we show that hormone-dependent autoregulation of the mGR mRNA and protein is abolished in mutants bearing three or more substitutions. Strikingly, receptor half-life is greatly increased with decreased phosphorylation, suggesting that phosphorylation is involved in receptor turnover.
Unless otherwise specified, all reagents were
purchased from Sigma. Dexamethasone
(9-fluoro-16
-methyl-11
,17
,21-trihydroxypregn-1,4-diene-3,20-dione) was purchased from Steraloids (Wilton, NH). The anti-peptide
polyclonal-antibody 1857, characterized previously (16), was used for
both immunohistochemistry and Western analysis of mGRs.
[
-32P]dCTP (3,000 Ci/mmol) was purchased from ICN
Radiochemicals (Irvine, CA). [14C]Chloramphenicol (40-60
Ci/mmol) was purchased from DuPont NEN. Biotrans nylon membranes were
purchased from ICN. Protran nitrocellulose BA85 was purchased from
Schleicher & Schuell. 20 × 20-cm TLC Silica Gel 60 sheets were
purchased from EM Separation Technology (Gibstown, NJ).
Wild type and mutant mGRs were all expressed in the same pSV2wRec vector whose expression is driven by an SV40 promoter. Generation of mutant mGRs in which alanine was substituted for serine or threonine has been described previously (17) (see Table I). The MMTV-CAT reporter plasmid (pGMCS), kindly provided by Dr. Don DeFranco, and the minimal reporter plasmid GRE2-TATA-CAT, have been described previously (18, 19).
|
The method for assessing the subcellular distribution of both wild type mGR and mGR phosphorylation mutants transiently transfected into COS-1 cells has been described previously (20). Briefly, COS-1 cells (African green monkey kidney, ATCC) were grown in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 9 mg/ml glucose, 100 IU/ml penicillin, 100 µg/ml streptomycin and supplemented with 2 mM glutamine and 10% of a 1:1 mixture of fetal calf/calf serum (Irvine Scientific, Santa Ana, CA). Cultures were maintained at 37 °C in a humidified atmosphere of 5% CO2. The cells were passed every 3-4 days and were maintained in culture for no longer than 15 passages. Cells were transfected by the DEAE-dextran method of Sompayrac and Danna (21) as modified by Gorman (22). Cells were incubated with the appropriate DNA/DEAE dextran mixture for 3 h and placed in Dulbecco's modified Eagle's medium supplemented with steroid-stripped fetal calf/calf serum and incubated further at 37 °C for 24 h. Transfected cells were placed in two-chamber glass slides, incubated for an additional 24 h, and then treated with 100 nM dexamethasone or vehicle for 1-2 h. Cells were fixed and processed for immunohistochemical staining as described previously (23).
Chloramphenicol Acetyltransferase (CAT) AssaysTransfections and CAT assays were performed according to the calcium phosphate method described by Gorman (22). COS-1 cells were cotransfected with either wild type or various phosphorylation mutants and either the MMTV-CAT reporter pGMCS or the GRE2-TATA-CAT reporter plasmids. After transfection, cells were either treated with hormone (100 nM dexamethasone) or vehicle and incubated for 24 h. Cell extracts were then analyzed for CAT activity. For our assays, the concentration of acetyl-CoA was increased to 1 mM, and the reaction times and concentration of radiolabeled chloramphenicol required were determined in prior studies to ensure that the percent conversion of radiolabeled chloramphenicol to the acetylated forms was in the linear range. After autoradiography, the acetylated and nonacetylated spots on the TLC plates were excised and counted to determine absolute levels of conversion. Relative levels of activity for the phosphorylation mutants were calculated, and results are reported as fold increase over control levels after hormone treatment.
RNA Isolation and Northern AnalysisLevels of GR mRNA
were evaluated by Northern analysis as described (12). Briefly, COS-1
cells were transfected with either wild type mGR or phosphorylation
mutants as described above. The cells were treated with 100 nM dexamethasone or vehicle and incubated for 24 h. After hormone treatment, the cells were processed according to a
modification of Chirgwin et al. (24) to obtain total RNA. Then, 1-4 µg of poly(A)+ selected RNA was denatured
using glyoxal and dimethyl sulfoxide and separated on a 1% agarose
gel. The RNA was transferred to a nylon membrane, stained with
methylene blue (0.04% methanol, 0.5 M sodium acetate, pH
5.2) to ensure uniformity of transfer, and hybridized with the randomly
primed 32P-labeled mGR cDNA (3-5 × 106 cpm/ml hybridization fluid, 50% formamide, 5 × SSC, 0.05 M sodium phosphate, pH 7.0, 0.25% SDS, 175 µg/ml salmon sperm DNA, and 5 × Denhardt's). Hybridization was
performed for 18-24 h at 52 °C. Subsequently, the membrane was
washed twice for 5 min at 37 °C in 2 × SSC, 0.1% SDS followed
by two washes for 10 min at 65 °C in 0.1 × SSC, 0.1% SDS. The
membrane was exposed to radiographic film at 70 °C and then
developed. The RNA levels were obtained using the video image analysis
software NIH Image 1.56 for the Power PC (Bethesda, MD) which had been
calibrated previously to give results in a linear manner. Relative mGR
RNAs were then assessed by comparison with non-hormone-responsive
-actin mRNA.
Cellular extracts from non-transfected or transiently transfected COS-1 cells that were treated with hormone (100 nM dexamethasone) or vehicle were prepared. Protein levels were quantitated by the Bradford method using a commercial product (Bio-Rad). The extracts were then solubilized in Fairbanks buffer (2% SDS, 20 mM Tris-Cl, pH 7.5, 2 mM EDTA, 10% sucrose, and 20 µg/ml pyronin Y) for 5 min at 100 °C. Typically, 125 µg of protein was separated on a 7.5% polyacrylamide gel as described previously (25, 26). The separated proteins were transferred to a nitrocellulose membrane (27), which was stained with Ponceau S (0.5% in 1% acetic acid) to ensure uniformity of loading and transfer. After destaining with water, the membrane was incubated for 1 h at room temperature in blocking buffer (10% nonfat dry milk, 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.05% Tween 20). Following one 20-min wash and two 5-min washes in 1% milk buffer (1% nonfat dry milk, 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.05% Tween 20) the membrane was incubated for 1 h at room temperature in 1% milk buffer containing the anti-GR epitope-purified polyclonal antibody 1857 (16) at a dilution of 1:2,000. After one 20-min wash and two 5-min washes in 1% milk buffer (1% nonfat dry milk, 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.05% Tween 20) the membrane was incubated further for 1 h at room temperature in 1% milk buffer containing peroxidase-labeled goat anti-rabbit antibody (ECL, Amersham) at a dilution of 1:15,000. The membrane was washed as described above and reacted for 1 min with chemiluminescent reagents commercially supplied (ECL). Fluorograms were analyzed using the video image analysis software NIH Image 1.56 for the Power PC. All hormone-treated samples were compared with vehicle-treated control samples that were arbitrarily set at 100% and expressed as percent of that control with a standard error calculated from at least four (n = 4) separate experiments.
Although the wild type and all of the
phosphorylation mutants were expressed from the same backbone vector,
pSV2wRec (17), we examined the expression obtained in transiently
transfected COS-1 cells to determine if phosphorylation sites
influenced receptor expression. Western blot analysis of total cell
receptor protein using an epitope-purified polyclonal anti-GR antibody
revealed that all of the mutant receptors were expressed in the cells
(Fig. 1). It should be noted that no receptor could be
detected when backbone vector alone was transfected into the COS-1
cells (not shown). Furthermore, no novel degradation products were
noted for any of the phosphorylation mutants. The faster migrating
90-kDa band seen for wild type and all mutants probably represents a ubiquitous GR degradation product. The most substituted forms of the
mGRs, A7 and A8, migrated only slightly faster than the wild type
receptor. This observation is in agreement with results obtained by
Almöf et al. (8), who were able to visualize only a
modest size difference in a human GR mutant when five phosphorylated sites had been abolished in an expression system producing a truncated GR.
Phosphorylation Mutants Translocate to the Nucleus upon Exposure to Hormone
We next analyzed subcellular localization of receptor in
the absence and presence of hormone. Immunohistochemistry of COS-1 cells transiently transfected with either wild type or various phosphorylation mutants of the mGR further demonstrated that all of the
receptors were expressed in the COS-1 cell line. By
immunohistochemistry, relatively similar levels of expression for the
wild type and all of the phosphorylation mutants were observed. The
subcellular distribution of the mGR and phosphorylation mutants in the
transient transfection system revealed that in the absence of hormone,
all receptors have a cytoplasmic location (Fig. 2,
CON panels). However, we see greater variability in the
initial subcellular distribution of the mouse GR compared with human GR
(28). Evaluation of the effect of hormone on nuclear translocation of
wild type and phosphorylation mutants revealed that all of the mGRs
translocate to the nucleus (Fig. 2, DEX panels). Thus,
hormone-mediated nuclear translocation occurred regardless of the
phosphorylation state of the receptor.
Transcriptional Activation by mGR Phosphorylation Mutants
The
ability of the various phosphorylation mutants to activate
transcription of hormone-inducible reporter genes was next assessed.
Previously, Mason and Housley (7) reported that phosphorylation mutants
activated transcription of a glucocorticoid-inducible MMTV-CAT reporter
plasmid as effectively as wild type mouse GR. Similarly, Almöf
et al. (8), using glucocorticoid-responsive elements (GREs)
linked to the complex thymidine kinase promoter, observed comparable
activation between wild type and phosphorylation mutations of the human
GR. We also examined the ability of mouse phosphorylation mutants to
activate the complex promoter from the glucocorticoid responsive
MMTV-CAT plasmid pGMCS. The wild type mGR demonstrated a 12-fold
induction over control when 100 nM dexamethasone was added
to the culture medium for 24 h (Fig. 3A). All of the mutants showed an 8-20-fold
induction of the reporter gene; but, in agreement with previous work
(7, 8), none of these inductions was significantly different from the
wild type receptor.
We next considered if promoter context was an important component in the efficacy of various phosphorylated forms of mGR to activate transcription. To answer this question, we used the reporter plasmid GRE2-TATA-CAT (19) which contains two copies of the GRE from the tyrosine aminotransferase gene positioned just upstream of the minimal adenovirus E1b TATA sequence (Fig. 3B). Strikingly, all but one of the phosphorylation mutants we tested exhibited a decrease in its ability to transactivate transcription after hormone administration (Fig. 3B). These transactivations were only 25-50% of that seen for wild type mGR. Interestingly, the decreases observed were not additive as more phosphorylated sites were mutated, and clearly, all phosphorylation mutants activated transcription. Similar results were obtained when we used a lower dose (1 nM) of hormone to activate the transfected receptors (not shown). Therefore, the differences observed were with the relative potency of transactivation and not in their absolute ability to activate gene expression. These results indicate that, depending on the type of promoter, GR phosphorylation can modulate the magnitude of a hormone-inducible response.
Loss of Autoregulation in Phosphorylation Mutants of the Mouse Glucocorticoid Receptor: Northern AnalysisSince we determined
that promoter context is an important feature in the ability of
phosphorylation-deficient mutants to activate transcription, we wished
to determine if the phosphorylation state of mGR could affect a gene
that is negatively regulated in response to glucocorticoids.
Accordingly, we examined the effect of phosphorylation of mGR on the
down-regulation of its own gene. Previously we determined that the GR
is down-regulated in response to cognate ligands and that this
down-regulation occurs at the level of transcription (11). The genetic
elements responsible for the down-regulation are contained within the
GR coding region (cDNA) (12). Consequently, one can measure the
effect of phosphorylation on GR gene expression by transiently
transfecting the expression vectors of different phosphorylation-deficient mGRs into COS-1 cells and determining the
relative levels of mGR mRNA and protein following administration of
hormone. When wild type mGR was transfected into COS-1 cells the mGR
mRNA typically showed a decrease to less than 50% of control levels after treatment with 100 nM dexamethasone
replicating previous results seen with human GR (12) (Fig.
4A). For those receptors with a single (212A,
220A, 234A) or with two (212/234A or 220/234A) mutated phosphorylation
sites the effect of hormone on receptor mRNA levels was marginally
(55-67% of control levels) different from wild type mGR (45-50% of
control levels) after 100 nM dexamethasone treatment (Fig.
4A). Nevertheless, in all these cases, the differences in
single and double substitution mutant GR mRNA levels after hormone
treatment were not as great as seen for wild type mGR mRNA.
Receptor constructs A3, A4, A5, A5+412, A7, and A8 (Table I) all showed
a substantial reduction in hormone-mediated down-regulation (75-98%
of control levels) of receptor mRNA again compared with wild type
mGR (45-50% of control levels) (Fig. 4A). Therefore, mutation of three or more phosphorylation sites resulted in loss of
hormone-inducible down-regulation of the mouse GR gene. These data
argue that the phosphorylation state of the mGR receptor is extremely
important in attenuating rates of receptor transcription.
Loss of Autoregulation in Phosphorylation Mutants of the mGR Protein
Receptor protein levels were next examined to determine if they were affected by mutation of mGR phosphorylated sites. In agreement with the mGR mRNA data (Fig. 4A), the wild type protein levels decreased to 45% of control levels after treatment with dexamethasone. Neither single or double mutations of phosphorylated sites had significant effects on down-regulation of the mGR protein (Fig. 4B). However, multiple (A3-A8) mutations caused a dramatic loss of hormone-mediated receptor down-regulation, with hormone-treated protein levels 90-160% of untreated receptor levels (Fig. 4B). These results are consistent with those for mGR mRNA where the multiple mutants were not affected by hormone treatment (see A3-A8 in Fig. 4A). Interestingly, the dexamethasone-treated A5, A7, and A8 mutants showed levels of mGR protein even higher than control, suggesting that the dephosphorylated form of the receptor might be more stable (resistant to proteolytic degradation) than the phosphorylated form.
Phosphorylation Status Affects Receptor Half-lifeTo
evaluate the effect that phosphorylation may have on receptor protein
half-life, wild type, A7, and A8 phosphorylation mutants were
transiently transfected into COS-1 cells. The cells were then treated
with 1 µM cycloheximide for 1 h and then left untreated (control), or hormone was added. Relative receptor levels were measured by Western analysis at various times after cycloheximide treatment. Fig. 5A shows that control wild
type mGR has a half-life of about 18 h and that hormone treatment
decreases the half-life to 8-9 h. These results indicate that there is
hormone-dependent destabilization of the mGR protein. In
marked contrast, the A7 mutant (Fig. 5B) has a half-life of
23 h, which is only slightly decreased (to about 18 h) by
hormone treatment. The A8 mutant, in which all identified mouse
GR-phosphorylated sites have been abrogated, has a half-life of 29 h (compared with 18 for control wild type), which if anything, is
increased (to 32 h) by hormone treatment. In this least
phosphorylated form of the mouse GR, the increase in half-life
following hormone treatment could represent a novel receptor
conformation that is resistant to degradation or simply be missing the
signal for proteolytic degradation. These data are the first to show
that the phosphorylation state of the mGR has a profound effect on
protein half-life and hence on receptor function.
The role that phosphorylation contributes to the GR function has been the subject of several recent studies and is controversial (1, 7, 8). Initially, we examined if the phosphorylation status of the GR had any effect on hormone-induced nuclear translocation. None of the mutants we examined showed any diminution of nuclear translocation. In fact, the most mutated form of the mGR (A8) we tested in which all eight phosphorylation sites were abolished, showed translocation properties similar to those of the wild type receptor. It should be noted that the mGR shows more variability in the initial subcellular distribution than its human counterpart (20, 28). However, once the receptors were exposed to hormone, the translocation to the nucleus was complete.
All of the phosphorylation mutants we studied showed a hormone-inducible transcriptional response from the MMTV-CAT reporter system we tested. To date, the other reported studies have used either an MMTV-CAT reporter (7) or a vector with two GREs linked to a complex thymidine kinase promoter (8) and have shown that phosphorylation mutants were capable of eliciting a transcriptional response from a hormone-inducible reporter gene. However, we wished to analyze the transactivation potential of the phosphorylation mutants using a simpler system where additional transcription factors could not compensate for a loss of necessary phosphorylation sites. Therefore, we also tested a hormone-inducible reporter plasmid containing a minimal promoter GRE2-TATA-CAT (19). With the exception of the single substitution 234A the transactivation potential of the phosphorylation mutants was decreased to 25-35% of wild type. Our results suggest that promoter complexity has a significant bearing on transactivation by dephosphorylated receptors. These data imply that when the receptor interacts with the basal transcription complex in the absence of other ancillary factors, the phosphorylation state of the receptor is important in determining the magnitude of the hormone-induced response. A recent report by Kato et al. (29) showing that the human estrogen receptor must be phosphorylated at Ser-118 to achieve full response of the estrogen receptor activation function suggests that a similar status may exist for mGR. Similarly, Weigel and colleagues (30) have shown that phosphorylation mutants of the progesterone receptor give an attenuated response with a hormone-responsive reporter.
We evaluated the effect of phosphorylation status on GR-mediated repression of gene expression by examining hormone-mediated autologous down-regulation of the GR gene. Strikingly, our studies show that hormone-mediated autoregulation of the GR gene is abolished in mutants missing three or more phosphorylation sites. Thus, in this model, repression of gene expression by GR may be dependent on phosphorylation. It is unknown at this time if the effects on GR down-regulation occur by direct interaction of GR with the RNA polymerase II complex (31, 32) or with specific steroid receptor coactivators (33) or directly with the GR gene (11, 12). We speculate that phosphorylation may direct the receptor to act as an activator or repressor of gene transcription. Phosphorylation may represent an adaptive mechanism whereby GR can differentially regulate its own gene expression based on the availability of active kinases and phosphatases in different cell types and tissues.
The observation that the most dephosphorylated forms of the mGR, A7 and
A8, showed higher protein levels when treated with hormone in the
absence of increased mRNA levels suggests that this receptor ligand
complex is more stable than the phosphorylated wild type mGR.
Therefore, we determined receptor half-lives of wild type and both A7
and A8 mutants in the presence and absence of dexamethasone. The wild
type mGR when transfected into COS-1 cells showed a half-life similar
to what Dong et al. (36) observed for rat HTC cells. In a
study using dense amino acids (37) in rat GH cells, GRs in untreated
cells had a half-life of 19 h, whereas GR in cells treated with
triamacinolone acetonide had a half-life of 9.5 h. In our studies,
the wild type control cells showed a half-life of approximately 18 h whereas the dexamethasone-treated (100 nM) cells had a
half-life of 8-9 h (Fig. 5A). These results clearly imply
that there is a hormone-mediated destabilization of the GR. With the A7
and A8 mutants, without hormone, the half-lives were increased to 25 h
(Fig. 5B) and 29 h (Fig. 5C), respectively, and
hormone treatment had only a slight or no effect. These results are
especially interesting in light of the fact that other laboratories have found that phosphorylation targets certain proteins for
proteolysis. For example, Lanker et al. (34) showed recently
that rapid degradation of the G1 cyclin Cln2 occurred after
phosphorylation by a cyclin-dependent protein kinase, and Chen et
al. (35) demonstrated that the nuclear factor B suppressor
I
B
needs to be phosphorylated for kinase-dependent ubiquitination and subsequent protein degradation to occur. Therefore, phosphorylation of the GR is likely to be necessary for receptor turnover. It should be noted that it has been shown that upon hormone
binding the GR becomes hyperphosphorylated (6), supporting the notion
that phosphorylation could target the receptor for hormone-mediated
degradation. Furthermore, these results immediately imply that the
phosphorylation mutants may have a poorer transactivation potential
than observed, as a potentially greater concentration of receptors
would be present in hormone-treated cells transfected with
phosphorylation mutants. Thus, the small effects seen on transactivation of complex promoters as well as the larger effects seen
on simple promoters are likely to be underestimates. These results
demonstrate for the first time that the phosphorylation status of the
mGR receptor has profound effects on different receptor functions.