Steroid Receptor Coactivator-1 Splice Variants Differentially Affect Corticosteroid Receptor Signaling
http://www.100md.com
内分泌学杂志 2005年第3期
Division of Medical Pharmacology (O.C.M., S.v.d.L., P.J.S., S.H.H., T.F.D., E.R.d.K.), Leiden/Amsterdam Center for Drug Research and Leiden University Medical Center, Leiden University, 2300 RA Leiden, The Netherlands; Department of Metabolic and Endocrine Diseases (E.K.), University Medical Center Utrecht, 3508 AB Utrecht, The Netherlands; and Division of Nephrology (D.P.), Department of Medicine, University of California, San Francisco, California 94143
Address all correspondence and requests for reprints to: O. C. Meijer, Division of Medical Pharmacology, Leiden/Amsterdam Center for Drug Research, P.O. Box 9503, 2300 RA Leiden, The Netherlands. E-mail: o.meijer@lacdr.leidenuniv.nl.
Abstract
The mechanisms of receptor- and cell-specific effects of the adrenal corticosteroid hormones via mineralo- (MRs) and glucocorticoid receptors (GRs) are still poorly understood. Because the expression levels of two splice variants of the steroid receptor coactivator-1 (SRC-1) 1a and 1e, can differ significantly in certain cell populations, we tested the hypothesis that their relative abundance could determine cell- and receptor-specific effects of corticosteroid receptor-mediated transcription. In transient transfections, we demonstrate three novel types of SRC-1a- and SRC-1e-specific effects for corticosteroid receptors. One is promoter dependence: SRC-1e much more potently coactivated transcription from several multiple response element-containing promoters. Mammalian 1-hydrid studies indicated that this likely does not involve promoter-specific coactivator recruitment. Endogenous phenylethanolamine-N-methyltransferase mRNA induction via GRs was also differentially affected by the splice variants. Another type is receptor specificity: responses mediated by the N-terminal part of the MR, but not the GR, were augmented by SRC-1e at synergizing response elements. SRC fragment SRC988–1240 by the MR but not the GR N-terminal fragment in a 1-hybrid assay. The last type, for GRs, is ligand dependence. Due to effects on partial agonism of RU486-activated GRs, different ratios of SRC-1a and 1e can lead to large differences in the extent of antagonism of RU486 on GR-mediated transcription. Furthermore, we show that SRC-1e but not SRC-1a mRNA expression was regulated in the pituitary by corticosterone. We conclude that the cellular differences in SRC-1a to SRC-1e ratio demonstrated in vivo might be involved in cell-specific responses to corticosteroids in a promoter- and ligand-dependent way.
Introduction
THE NUMEROUS EFFECTS of adrenal corticosteroid hormones in the body are mediated by glucocorticoid (GRs) (1) and mineralocorticoid receptors (MRs) (2). Both receptors are members of the large nuclear receptor family and accordingly act as transcription factors to transactivate or transrepress specific target genes. The MRs and GRs have a high degree of homology in their DNA binding and ligand binding domain but differ considerably in the N-terminal part of the receptor (2). Thus, these receptors recognize the same response elements on the DNA, but they differ in their transactivational (3, 4) and transrepressive (5, 6) properties.
In some tissues the two receptor types mediate in a coordinate fashion signaling by cortisol and corticosterone, the most important glucocorticoids in man and rodents, respectively. In cells expressing both receptor types, MRs and GRs mediate different and at times opposite effects on cellular physiology, underscoring the importance of differential transcriptional properties of these two receptors (7). Besides receptor specificity, there are many instances of cell-specific gene regulation by corticosteroids, unexplained by receptor expression. For instance, in the rat brain, CRH mRNA is down-regulated by corticosterone via GRs in the hypothalamus but up-regulated in the amygdala nucleus, presumably via cell specific transcriptional effects (8). The mechanisms underlying such cell and receptor specificity are as yet largely unknown.
Over the last several years, a large number of coregulatory proteins that influence transcriptional responses of nuclear receptors has been discovered (9). The family of p160 steroid receptor coactivators (SRCs) consists of three genes (10), and each of their products can bind to the activation function (AF)-2 of nuclear receptors through interactions with LxxLL motifs or nuclear receptor (NR) boxes (11). In addition, SRCs may also interact with the less conserved N-terminal domains of steroid receptors (12, 13). The actual stimulation of transcription depends on direct histone acetyl transferase activity and recruitment of cointegrators such as cAMP response element-binding protein (CBP)/p300 (14) or the methyl transferase CARM-1 (15). The interaction with at least one of the SRCs, which are expressed in a cell-specific manner (16, 17), is thought to be necessary for transcriptional stimulation to occur (18, 19). The interaction between the SRC variant and the specific NR is thought to determine the nature of the interaction and magnitude of coactivation.
Five different splice variants of SRC-1 were originally reported (20), of which SRC-1a and 1e have been consistently found (21, 22). These differ only at their carboxy terminus, which is shorter in SRC-1e. The SRC-1a-specific 56 amino acids contain an extra NR box and a potential suppressor domain. These splice variants were shown to differ in their interactions with the (isolated) ligand-binding domains (LBDs) of NRs (21) and their functional interactions with the estrogen receptor- (22). In addition, the expression of SRC-1a and 1e mRNA is cell specific. In crucial corticosteroid-sensitive cell populations in the brain, considerable differences in the expression of the splice variants were observed (23).
In view of the observed differences between SRC-1a and -1e, we tested, using cultured cells as a model, the hypothesis that there are differential interactions between SRC-1 splice variants and corticosteroid receptors, which may contribute to cell- and receptor-specific corticosteroid effects. Hence, we tested the functional and physical interactions of SRC-1a and 1e with MRs and GRs, using transient reporter assays and mammalian 1-hybrid studies. We also tested the hypothesis that the abundance of SRC1a and SRC-1e mRNA is differentially regulated. We find that these splice variants differentially affect transcription in a receptor-, promoter-, and ligand-dependent fashion and that there can be specific physical interactions between SRC-1 and the MR N-terminal fragments. In addition, we show that the relative abundance of SRC-1a and SRC-1e mRNA is subject to regulation in the pituitary in vivo.
Materials and Methods
Transient transfections
CV-1 cells were grown in DMEM supplemented with 5% fetal calf serum (GIBCO, UK), A549 cells in DMEM/HAM F12 mix with 10% fetal calf serum (GIBCO). For transfections, cells were plated in 24-well plates (Greiner Bio-One, Alphen aan den Rijn, The Netherlands) at 3 x 104 cells/well, and charcoal-stripped serum was used. The cells were transfected using SuperFect (Promega, Madison, WI) at a DNA to superfect ratio of 1:2 (CV-1) or 1:4 (A549). For CV-1 cells steroid receptor expression plasmids were used at 100 ng/well. In CV-1 cells 100 ng of SRC-1 expression plasmid and reporters were used per well, based on a number of initial titration studies (showing qualitatively similar but quantitatively increasing effects from 50 to 400 ng plasmid/well), whereas for A549 cells, we used 200 ng of SRC expression vector and reporter plasmid per well. The transcriptionally inert plasmid pSP65 was used to bring the total amount of DNA to 1 μg/well. As control plasmid we used 1 ng of pCMV-R (Promega) coding for Renilla luciferase controlled by cytomegalovirus (CMV) promoter. We found that expression of this promoter is not influenced by activation of either MR or GR in the cell, in contrast to the pTK-R reporter (Promega), which was consistently repressed by activated GR (data not shown). One day after transfection, the cells were treated with corticosterone (CV-1 cells/rat receptors), cortisol (endogenous human receptor in A549 cells), aldosterone and/or antagonists RU486 (GR), and spironolactone (MR). Unless indicated otherwise, all ligands were given at a dose of 10–7 M, which maximally activates both MRs and GRs, except for aldosterone, which was used at the saturating concentration of 5 x 10–9 M. After 24 h the cells were harvested and firefly and Renilla luciferase activity was determined using the Promega dual label reporter assay and a luminometer (LUMAT LB 9507, Berthold, Bad Wildbad, Germany).
All conditions were tested at least in triplicate, and each experiment was repeated three times with essentially similar results. Data are expressed as average ± SEM. All results were analyzed by ANOVA and statistically tested with post hoc Tukey Kramer test. For the ligand-activated receptors, statistical significance (P < 0.05) was attributed based on differences in fold induction by hormone (i.e. all luciferase values from glucocorticoid response element (GRE)-containing reporters were normalized to Renilla luciferase. Induction was calculated for each steroid-treated sample as signal in the presence of hormone divided by the average signal from the same condition in absence of hormone).
Plasmids
Most plasmids we used have been described previously. Human MR and GR expression plasmids (3) were kindly provided by Dr D. Spengler, with permission of Dr. R. Evans (Salk Institute, San Diego, CA). For rat MRs and GRs, and truncations and chimeras thereof, we used the 6RMR and 6RGR-based plasmids (5, 24). These and the pODLO-based reporter plasmids TAT-1-TATA and TAT3-TATA were described earlier (25), as were the reporters driven by rat phenylethanolamine-N-methyltransferase (PNMT) promoter fragment containing multiple GREs and endogenous TAT 5' regulatory region (26). pSG-5 based expression vectors for full-length and mutant SRC-1a and SRC-1e were also described (22). The SRC-1 plasmids used in this study contain a 2x FLAG-tag at the 5' end of the cDNA. The Q plasmid (SRC-1e 1053–1123) was kindly provided by Dr. C. Bevan (Imperial College, London, UK). The pPRE1-tk-Luc and pPRE2-tk-Luc, and mouse mammary tumor virus (MMTV)-luciferase (LUC) plasmids were kindly provided by Dr. G. Jenster (Rotterdam University Medical Center, Rotterdam, The Netherlands). The pPRE2-tk-Luc reporter was constructed by inserting the XhoI-digested (PRE)2TK fragment from pBL2(PRE)2TK-CAT plasmid (27) into the XhoI-digested pGL3basic vector (Promega). The p(PRE)1-tk-Luc reporter was generated by deleting one progesterone response element (PRE) (AGAACAtccTGTACA) from the p(PRE)2-tk-Luc plasmid using the two flanking BsrGI sites. The expression vectors for SRC-1 fragments fused to the VP16-AD were generated by excision of EcoRI fragments from the pSG424 constructs (21) and cloning them into pSG5-VP16 (28).
Western blots
Cells from wells parallel to luciferase experiments were lysed in 0.5 x radioimmunoprecipitation assay and subjected to SDS-PAGE. Proteins were transferred onto Immobilon membranes (Millipore Corp., Bedford, MA) and processed as described (29). Blots were blocked in 10 mM Tris-HCl (pH 8.0), 150 mM NaCl, and 0.05% Tween 20 containing 5% nonfat dried milk powder and subsequently incubated with the M2 primary anti-FLAG antibody (1:4000, Sigma, St. Louis, MO), the M-20 GR antibody (1:500, Santa-Cruz Biotechnology, Santa Cruz, CA), or the monoclonal anti--tubulin antibody (Sigma, 1:1000). After a wash, blots were incubated with peroxidase-conjugated antibodies (1:10,000; Jackson ImmunoResearch Laboratories, West Grove, PA). Blots were washed again, and immunoreactive bands were visualized by enhanced chemiluminescence.
Quantitative PCR quantification of PNMT mRNA
Neuroscreen-1 cells (Cellomics Europe) were grown in RPMI 1640 medium with L-glutamine (Life Technologies, Inc., Grand Island, NY), supplemented with 10% horse serum, and 5% fetal bovine serum. Cells were transfected with 3 μg of SRC-1A or 1E plasmid per 2 million cells or mock transfected, using a Nucleofector and electroporation buffer V (Amaxa Biosystems). Cells were plated at 106 per well in collagen-coated 6-well plates and after 1 d were exposed to dexamethasone (10–6 M) or vehicle for 18 h. Total RNA was isolated using an RNeasy kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. During RNA extraction, samples were treated on-column with RNase-free DNase I (Qiagen). Total RNA required for standard curves was obtained from the rat adrenal and isolated with Trizol reagent (GIBCO BRL). Potential genomic DNA was degraded with DNase I (Invitrogen Corp., Carlsbad, CA). The purity and concentration of isolated total RNA was measured by the ND-1000 spectrophotometer (NanoDrop Technology) where the ratio of absorption at 260 and 280 nm for all samples was at least 1.9. First-strand cDNA was synthesized with SuperScript II reverse transcriptase (Invitrogen) and Oligo(dT)12–18 (Invitrogen) in accordance with the manufacturer’s specifications. Real-time PCR was performed on a LightCycler 2.0 (Roche Diagnostics, Indianapolis, IN) using LightCycler FastStart DNA Masterplus SYBR Green I (Roche Diagnostics). The expression of PNMT was normalized against ?-actin. The sequences of the used primers are: Actb forward, tgaccgagcgtggctaca; Actb reverse, cagcttctctttaatgtcacgca; PNMT forward, gcgagggtgaagcgagtcttgcc; and PNMT reverse, tacccccgggcctcagcagc. Statistical significance of differences in normalized mRNA levels (four wells per condition) was determined by the Kruskal-Wallis test followed by the Mann-Whitney U test.
In situ hybridization
For determination of pituitary mRNA levels of SRC-1a and 1e, male rats (240 g Wistar rats, Charles River, Germany) were adrenalectomized or sham operated (n = 6) under gas anesthesia using the dorsal approach. Six animals per group were sc implanted with a 100-mg pellet containing 100% cholesterol, 20% corticosterone, or 100% corticosterone. All adrenalectomized animals had access to saline. After 1 wk the animals were decapitated, and pituitaries were collected and frozen at –80 C. The animal experiments were performed in accordance with the European Communities Council Directive 86/609/EEC and with approval from the animal care committee of the Faculty of Medicine, Leiden University (UDEC no. 99062). The tissue was cut at 14 μm on a cryostat and thaw mounted on poly-L-lysine-coated slides, each slide containing two sections from four rats, each rat belonging to a different treatment group. In situ hybridization with SRC-1a- and 1e-specific end-labeled oligonucleotide probes was performed as described (23). The signal was quantified from film (X-OMAT AR, Kodak, Rochester, NY) using an automated image analysis system (Paes Nederland, Zoeterwoude). The signals were in the linear range of gray values according to the [14C] RPA 504 microscales (Amserham, Aylesbury, UK). Four sections per animal, i.e. 24 sections per treatment group, were analyzed by ANOVA and post hoc Tukey-Kramer test.
Results
Promoter-specific coactivation
As a basic approach to study possible differential coactivation of rat MRs and GRs by SRC-1a and SRC-1e, we chose to use transient transfection of CV-1 cells. These cells contain no endogenous corticosteroid receptors and therefore allow good definition of the type of steroid hormone receptors that are present. No induction of any reporter was observed in the absence of transfected receptor in these cells (data not shown). At the TAT-1-Luc reporter, which harbors a single GRE in front of the Drosophila alcohol dehydrogenase promoter TATA box, corticosterone-activated MRs (10–7 M) transactivated very modestly in the absence of overexpressed coactivator. Statistical analysis indicated significant difference in hormone induction among the three coactivator conditions. Overexpression of SRC-1a led to a significantly enhanced reporter activity. Transactivation in the presence of hormone was increased in the presence of SRC-1e, but this did not reach significance relative to the no-coactivator condition when calculated as fold induction by hormone (Fig. 1A). Transactivation via GR was similarly different between conditions, i.e. significantly enhanced in the presence of overexpressed SRC-1a but not significantly by SRC-1e (Fig. 1B).
FIG. 1. Promoter-specific coactivation by SRC-1a and -1e. SRC-1a or -1e was overexpressed in CV-1 cells containing MRs (A, C, and E) and GRs (B, D, and F), without hormone (open bars) and in the presence of 10–7 M corticosterone (closed bars) using the TAT-1 (A and B), TAT-3 reporter (C and D), and reporter backbone ODLO (E and F). Reporter activity is expressed as luciferase normalized by CMV promoter-driven Renilla luciferase. Fold induction is indicated in numbers above the bars. Asterisks indicate significant differences in fold induction by corticosterone relative to both other conditions, unless a specific comparison is indicated (P < 0.05). A and B, At the TAT-1 reporter, SRC-1a overexpression leads to modestly stronger coactivation for both receptors. C and D, At the TAT-3 reporter, SRC-1a has no potentiating effect for either receptor, but SRC-1e shows very strong coactivation. In boxes are the fold coactivation by SRC-1a and -1e and the ratio between these two values. G, Expression of FLAG-tagged SRC-1A and SRC-1E on Western blot from transfections in which the differential coactivation was observed, indicating equal expression levels of the splice variants. H, Expression of the steroid receptor expression plasmid as measured by GR immunoreactivity on Western blot is not affected by hormone treatment or coactivator expression in samples run in parallel to luciferase assays. Bottom row shows tubulin expression as control for loading.
On the TAT-3-Luc reporter, containing a tandem repeat of three consensus GREs in front of the alcohol dehydrogenase promoter TATA box, transactivation in absence of overexpressed SRC-1 is much stronger, notably for GRs, which are known to act synergistically at multiple response elements (30, 31). At this reporter coactivation by SRC-1 splice variants showed large differences in absolute transactivation and fold induction relative to the no-hormone condition. SRC-1a overexpression did not lead to any potentiation of the transcriptional response for either the MR or GR. In contrast, SRC-1e strongly potentiated the MR-mediated response; absolute levels of reporter activity were of the same order as those observed for GRs in the absence of overexpressed coactivator (Fig. 1C). Also, GR-mediated action was enhanced by overexpression of SRC-1e (Fig. 1D). MRs and GRs differed qualitatively in their coactivation by the SRC-1 splice variants. Whereas in both cases SRC-1e is the stronger coactivator on the TAT-3-Luc reporter, the SRC-1e to 1a ratio was consistently higher for the MR.
The pODLO backbone of the TAT-1 and TAT-3 plasmids was not influenced by corticosterone activation of either corticosteroid receptor and/or overexpression of SRC-1a and SRC-1e (Fig. 1, E and F). Furthermore, SRC-1a and SRC-1e were expressed at similar levels under our experimental conditions (Fig. 1G), Also, the expression vector of the receptors was not affected by coactivator expression or hormone treatment (shown for GR in Fig. 1H). Human MRs and GRs showed the same promoter dependence in coactivation as the rat receptors in transient transfection studies in CV-1 cells (data not shown).
The results suggested that the number of GREs may be a critical parameter that determines differential coactivation by SRC-1a and -1e. Hence, we tested in CV-1 cells, in parallel to the TAT-1 and TAT-3 reporters, a number of other reporters that differ only in the number of GREs (Table 1). PRE-1-tk-Luc and PRE-2-tk-Luc contain one and two GREs upstream of a minimal tk promoter, respectively. The rat PNMT has a multiple GRE-containing fragment of the rat PNMT gene 5' flanking region upstream of a SV40 basal promoter. The PNMT 3 reporter contains mutations in three GRE half-sites and behaves like a single GRE-containing promoter (26). As shown in Table 1, in all cases the coactivation by SRC-1a was stronger on the single response element-containing reporter, whereas SRC-1e coactivated more strongly at the multiple GREs. Also shown is the ratio of coactivation by SRC-1a and SRC-1e. For the TAT-3 reporter, this ratio is consistently higher for coactivation of the MR, compared with the GR. Coactivation of GR at the often-used GRE-containing MMTV reporter was not different between SRC-1a and SRC-1e (data not shown).
TABLE 1. The fold coactivation of 10–7 M corticosterone-activated MRs and GRs brought about by SRC-1a and -1e on related promoters that contain single or multiple GREs, and the ratio between the SRC-1e and 1a effect
Next, we were interested to see whether these promoter-specific SRC-1a/1e effects would be reproducible in a different cell line, preferably one that is commonly used, endogenously expresses one of the receptor types, and is of human origin (given that we expressed human SRC-1 variants). The human A549 lung carcinoma cell line fulfills these criteria, and we used it to study coactivation of the endogenous human GR (Fig. 2). Transcription synergy at multiple GREs was notably less strong than in CV-1 cells. Coactivation by SRC-1a and -1e of the simple TAT-1 and TAT-3 reporters was comparable to that in CV-1 cells (Fig. 2A): also in this setting SRC-1a overexpression leads to coactivation at the single GRE-driven TAT-1 reporter, whereas at the TAT-3 reporter, it does not coactivate. SRC-1e coactivation did not reach statistical significance on the TAT-1 reporter (calculated as fold induction by hormone), whereas at the TAT-3 reporter, SRC-1e potently coactivated. Induction of transcription was also achieved by treatment of the cells with 1 nM dexamethasone but not with 1 nM of aldosterone (not shown), pointing to GR rather than MR as the mediator of effects of the mixed agonist corticosterone.
FIG. 2. Coactivation of endogenous GRs in A549 cells by SRC-1a and SRC-1e at synthetic and endogenous promoter fragments depends on the number of response elements. A, Coactivation of the TAT-1 and TAT-3 reporters. Open bars, no hormone; closed bars, 10–7 M cortisol. Reporter activity is normalized with pCMV-Renilla. Fold induction by hormone is indicated above the bars. Asterisks indicate statistically significant differences from the no-coactivator condition at P < 0.05. At the TAT-1-luc reporter by SRC-1a is a modestly stronger coactivator; at the TAT-3-Luc reporter, SRC-1e is much stronger. There are no effects of hormone or SRC expression on the pODLO empty vector. B, Fold induction by 10–7 M cortisol of two endogenous regulatory fragments that harbor a single (eTAT) or multiple (PNMT) response elements again shows preferential coactivation of multiple response elements by SRC-1e. Values represent normalized reporter activity in the presence of hormone divided by those in the absence of hormone. C, Induction of PNMT mRNA in neuroscreen cells via GR activation is lower in the presence of overexpressed SRC-1, but SRC-1e coactivates stronger than SRC-1a. Asterisks indicate statistically significant difference from both other groups by Mann-Whitney U test (P < 0.05).
We also tested coactivation of GRE-containing regulatory fragments from two glucocorticoid-regulated genes, which both were activated via endogenous GR in A549 cells (shown in Fig. 2B as fold induction by hormone because of large differences in absolute luciferase levels from the used reporters). Glucocorticoid responsiveness of the eTAT-luc reporter depends on the regulatory 5' steroid-responsive fragment of the tyrosine aminotranferase promoter, containing a single GRE (26). SRC-1a but not SRC-1e overexpression led to significantly increased transactivation of the eTAT-luc reporter. In contrast, the glucocorticoid-responsive multiple GRE-containing fragment from the PNMT 5' regulatory region showed coactivation only in the presence of overexpressed SRC-1e.
We next studied whether differential coactivation by the splice variants would hold in the context of an endogenous gene. As a model we took induction of the PNMT gene via GR by dexamethasone in neuroscreen-1 cells, a subclone of the PC-12 cell line for which the phenomenon was reported (32). Without addition of dexamethasone to the medium, no PNMT mRNA could be measured by quantitative PCR (data not shown). After treatment with 1 μM dexamethasone, there was a clear induction of PNMT mRNA in mock-transfected cells (Fig. 2C). Overexpression of SRC-1A and 1E both led to lower levels of PNMT mRNA, but expression was modestly, but significantly, higher in cells overexpressing SRC-1E than in cells overexpressing SRC-1A, as hypothesized based on the results with the PNMT GRE-containing region in the context of a reporter plasmid.
Thus, the promoter-dependent coactivation in relation to the presence of multiple GREs was observed using overexpressed as well as (for GR) endogenous receptors, in the contexts of simple and more complex regulatory regions, and held for induction of the endogenous PNMT gene.
Input domains of the SRCs
The promoter-dependent coactivation by SRC-1 splice variants may reflect differential recruitment of the coactivators by the receptors. To test this possibility, we studied the interactions between MR and GR with different domains of the SRC-1 protein at the TAT-1 and TAT-3 reporter, coexpressing in CV-1 cells MR or GR with hybrids of SRC-1 fragments and the VP16 activation domain (AD) as shown in Figure 3A. The hypothesis of preferential coactivator recruitment as the basis for promoter-specific effects would be supported by, for example, a higher degree of receptor interaction of the SRC-1a-specific fragment at the TAT-1 reporter, compared with the TAT-3 reporter. At both reporters the expression of the central LxxLL (NR box) containing fragment SRC-1570–780 fused to VP16-AD led to increased reporter activity after activation of either MR and GR by 10–7 M corticosterone, compared with expression of the VP16-AD alone (Fig. 3). Also, the other NR box-containing fragment SRC-1A1241–1441 led to increased reporter activity at both reporters. The SRC-1e C-terminal region (1241–1399) or that of SRC-1a from which the NR box was deleted (1241–1385) showed no sign of interaction with the steroid receptors. These data provide no evidence for promoter-specific recruitment of SRC-1 splice variants as the cause of promoter-specific coactivation because there was no difference in recruitment of the splice variant-specific carboxy-terminal parts of SRC-1. The only other notable difference was a decreased activation for the fragment SRC-1988–1240 at the TAT-3 reporter, which was specific for GR, but did not reach significance in the ANOVA post hoc test (Fig. 3C; also see below).
FIG. 3. Mammalian 1-hybrid experiments suggest that the promoter-dependent coactivation is not due to promoter-dependent interactions between steroid receptors and SRC-1 NR box-containing fragments. A, Schematic representation of SRC-1 protein. The underlined areas were fused to the VP16 activation domain and used to probe interactions with MRs and GRs. Both at the TAT-1 reporter (B) and TAT-3 reporter (C), the LxxLL (or NR box)-containing fragments of SRC-1 led to increased reporter activity, compared with the condition of steroid receptor incubated in the presence of the VP16 AD alone in CV-1 cells. The SRC-1a-specific C-terminal NR box is present in the 1241–1440 fragment but not in the truncated 1241–1385 or SRC-1e-specific fragment 1241–1399. It clearly interacts with MRs and GRs, irrespective of the promoter context. Asterisks indicate significantly increased induction, compared with receptor in the presence of nonfused VP16 AD at P < 0.01.
Because we found that, at least on the TAT-3 reporter, the differences between splice variants were larger for MRs by SRC-1e than for GRs, we studied coactivation of N- or C-terminally truncated receptors with this reporter. We used CV-1 cells to allow distinction between effects on the two activation functions AF-1 (located in the N terminus) and AF-2 (located in the C-terminal LBD) of MRs and GRs. Members of the p160 SRC family principally coactivate the AF-2 of nuclear receptors but may also act at the N-terminally located AF-1 (13). As in the earlier experiments, full-length GRs and MRs were coactivated strongly by SRC-1e but not by SRC-1a on the TAT3 reporter (not shown). In line with the original discovery of SRCs as LBD-interacting proteins, the N-terminally deleted GG and MM were strongly and comparably coactivated by SRC-1e and not by SRC-1a at this reporter (Fig. 4, A and C). The C-terminal truncation MM and GG are constitutively active transcription factors (Fig. 4, B and D). Transactivation by GG was not influenced by overexpression of SRC-1 splice variants (Fig. 4D). For MM we observed a modest but consistent (approximately 2-fold) coactivation by SRC-1e (Fig. 4B).
FIG. 4. The MR, but not the GR N terminus, can be coactivated by SRC-1e on the TAT-3-Luc reporter. SRC-1a and SRC-1e were overexpressed in CV-1 cells with truncations of the corticosteroid receptors containing the ligand-dependent AF-2 (NN; A and C) or ligand-independent AF-1 (NN; B and D). Open bars, no hormone, closed bars, 10–7 M corticosterone. The AF-2 of both receptors shows coactivation similar to full-length receptors. Fold induction by hormone is indicated above the bars. The AF-1 containing MR fragment (C) but not the GR fragment (D) is consistently coactivated by SRC-1e in this setting. Asterisks indicate significant differences at P < 0.01 in fold induction (A and C) or reporter activity (normalized to pCMV-Renilla), relative to both other conditions, unless a specific comparison is indicated.
To better understand the coactivation of the MR N-terminal part, we tested coactivation of the receptors on the TAT-3 reporter by SRC-1e variants in which the known receptor-interacting domains of SRC-1 were deleted, namely the three NR boxes [M123; (22)] and the Q-rich domain, which can interact with the androgen receptor [Q construct; (12)]. In case of the full-length receptors as well as for the AF-2-driven N-terminal truncations of the MR and GR, deletion of the NR boxes abolished coactivation by SRC-1e (Fig. 5, A and B). In contrast, SRC-1e that lacked its NR boxes or the Q-rich domain could still coactivate the C-terminally deleted MR but not GR (Fig. 5C). SRC-1e in which the CBP-binding AD-1 (amino acids 900–950) was deleted could not stimulate transcription in this setting (Fig. 5E), which argues for the specificity of the observed stimulating interactions. These data demonstrate that SRC-1 coactivation effects on AF-2-containing receptors depend on the already well-characterized NR box interaction motifs but that effects on the MR N terminus (likely mediated by AF-1) depend on an as-yet-unidentified SRC domain.
FIG. 5. Coactivation of the AF-1-containing N-terminal domains of MRs by SRC-1e at the TAT-3 reporter in CV-1 cells does not depend on NR boxes or the Q-rich domain (amino acids 1053–1123) but involves the region 988-1240 of the SRC protein. A, The full-length MR and GR cannot be coactivated when the three central NR boxes are deleted (M123 construct), whereas the Q-rich domain of SRC-1e (absent in Q) is dispensable for coactivation. Coactivation at the TAT-3 reporter is stronger for MR than GR (see also Fig. 1). B, The M123 mutant SRC cannot coactivate AF-2 of the N-terminal deletions of MR and GR. C, The MR N terminus was coactivated by SRC-1e independent of the Q-rich domain or the NR boxes, whereas the GG construct was not coactivated by any SRC variant. D, 1-hybrid assays using the TAT-3 reporter. Direct interactions between the reporter-bound receptor and SRC fragments fused to the VP16 activation domain lead to increased reporter activity. Values represent induction by 10–7 M corticosterone, relative to the condition of steroid receptor cotransfected with the VP16 AD alone. The MR but not the GR N terminus showed a specific interaction with SRC fragment 988-1240 fused to the VP16 AD. No interaction was observed for the NR box containing SRC-1 fragments. E, Deletion of activation domain 1 (amino acids 900–950) from SRC-1e abolishes the coactivation of the MM protein. Asterisks indicate significant differences in reporter activity, compared with the no-coactivator (for D: fusion) condition (P < 0.01).
To further substantiate the coactivation of the MR N terminus by SRC-1e, we evaluated the molecular interactions between the truncated receptors bound to the TAT-3 reporter and SRC fragments (that are shown in Fig. 3A) fused to the VP16 activation domain. As depicted in Fig. 5D, SRC-1988–1240 induced reporter activity over VP16 background for the MR but not the GR N terminus. No interactions with other fragments, such as those containing the NR boxes, were observed (Fig. 5D and data not shown). Whereas deletion of the Q-rich domain (1053–1123) does not impair coactivation, other residues in the region between amino acids 988 and 1240 are likely to be responsible for SRC-1e interaction with, and subsequent potentiation of. the MM protein.
Thus, besides pronounced reporter specificity, the SRC-1 splice variants also showed a degree of receptor specificity, apparent from selective coactivation of MR AF-1 by SRC-1e and the interaction observed in the 1-hybrid assay. In the context of the full-length receptor (Fig. 3C), no interaction between the MR and SRC-1988–1240 fragment is apparent. However, the observed lower reporter induction for full-length GR at the TAT-3 reporter in Fig. 3 does suggest a degree of receptor specificity for full-length receptors in their interaction with SRC-1988–1240, which may be related to the apparent stronger coactivation of MR at multiple response elements.
Ligand dependence
We further tested in CV-1 cells whether, besides the strong promoter and modest receptor specificity for SRC-1a and 1e, different ligands may also determine coactivation by these splice variants. Because the effect of antagonists is thought to depend on relative coactivator to corepressor ratio, we tested the coactivation by SRC-1a and -1e for the partial agonistic effects of the GR antagonist RU486 (Fig. 6). The transactivation on the TAT3 Luc reporter induced by 10–7 M RU486 was consistently around 15%, compared with corticosterone in absence of overexpressed coactivator (16 ± 4% over all experiments). The coactivation of RU486-activated GR by SRC-1 splice variants was strikingly different from corticosterone-activated GR: similar to our earlier experiments, SRC-1e but not -1a coactivated in the presence of 10–7 M corticosterone (Fig. 6B), but in presence of 10–7 M RU486, SRC-1a rather than -1e coactivated GR at the TAT3 reporter (Fig. 6A). When the MR was incubated with 10–7 M of the antagonist spironolactone, a modest partial agonism was also observed, but in this case the coactivation by SRC-1a and -1e was similar to that of corticosterone-activated MR (data not shown). Also, the coactivation of MR in the presence of 5 x 10–9 M of the full agonist aldosterone was not different from what was observed with corticosterone (not shown).
FIG. 6. Coactivation of GRs on the TAT-3-luc reporter in CV-1 cells is ligand dependent. A, Partial agonism of 10–7 M RU486 at the GR is higher when SRC-1a, rather than SRC-1e, is overexpressed. B, With 10–7 M corticosterone as a ligand, overexpression of SRC-1e, but not SRC-1a, leads to higher transactivation. Hormone induction relative to the no-coactivator condition is indicated above the bars. Asterisks indicate significant a difference in fold induction, compared with the no-coactivator condition (P < 0.01).
Because antagonists are typically used in the presence of agonists, we constructed a dose-response curve for corticosterone on the GR in the presence or absence of 10–7 M RU486 and determined the antagonist/partial agonistic effect of RU486 as a function of overexpressed SRC-1a or SRC-1e (Fig. 7). The dose-response curve for the GR-mediated corticosterone effect on the TAT-3 reporter in absence and presence of RU486 is shown in Fig. 7A for the condition in which no SRC was overexpressed. At low concentrations of corticosterone, the RU486 showed its partial agonistic effect. At higher concentrations of the full agonist corticosterone, the antagonism of RU486 became clear, although at 10–7 M corticosterone (equimolar with RU486), the antagonist could no longer effectively compete.
FIG. 7. SRC-1a and SRC-1e overexpression differentially affect the dose-response curves for GRs in absence and presence of 10–7 M RU486 in CV1-cells. A, No-coactivator (no Co-A) overexpression. B, Overexpression of SRC-1a. Due to higher partial agonism of RU486 and lower transactivation by corticosterone, there is hardly any antagonistic effect of RU486. C, Overexpression of SRC-1e. Due to the low partial agonism of RU486 (not different from no-coactivator) and particularly the strong coactivation of corticosterone-activated GRs (note the scale of the y-axis), RU486 has strong antagonistic effects at corticosterone concentrations higher than 1 nM.
The relative agonism/antagonism varies significantly between the conditions in which SRC-1 splice variants have different abundances. When SRC-1a was overexpressed (Fig. 7B), the partial agonistic effect of RU486 was modestly (1.3- to 1.5-fold) increased at low corticosterone concentrations (as in Fig. 6B), whereas the corticosterone-occupied GR was not coactivated by the overexpressed SRC-1a. As a consequence, less antagonism was observed at all corticosterone concentrations: transactivation was never blocked more than 2-fold by the presence of the antagonist. In contrast, when SRC-1e was brought to overexpression, corticosterone-activated GR was potently coactivated, but 10–7 M RU486 had a strong antagonistic effect at higher corticosterone concentrations (>10–9 M) and a weak partial agonistic effect at low corticosterone concentrations (<10–9 M; Fig. 7C).
The relative partial agonistic/antagonistic effect of 10–7 M RU486 in the presence of different concentrations of corticosterone is summarized in Table 2. SRC-1a overexpression leads to stronger partial agonistic effects of RU486 at low corticosterone levels and much weaker antagonist effect at higher levels of corticosterone, when compared with SRC-1e. Thus, variations of SRC-1 splice variant abundance may lead to differences in transactivation on binding of full agonists but also to different efficacy of some antagonists.
TABLE 2. The effect of SRC-1a/1e overexpression on the partial agonism/antagonism of RU486 at increasing concentrations of corticosterone
Differential regulation of SRC-1a/1e mRNA abundance
Our findings suggest that the ratio of SRC-1 splice variants may influence cellular responses to glucocorticoids. Although transcriptional control may not be the most common control mechanism of SRC regulation in differentiated tissues (33), SRC-1 mRNA levels in pituitary have been shown to be subject to clear regulation by estrogens and thyroid hormone (34). We therefore addressed the question whether mRNA of SRC-1 splice variants can be regulated by corticosterone and if so, whether this occurs in a uniform manner or whether there is splice variant-specific regulation. We used in situ hybridization with oligonucleotide probes that specifically detect SRC-1 splice variants (22) (Fig. 8). Chronic exposure of adrenalectomized rats to moderate to high levels of corticosterone (plasma levels of 6.7 ± 1.5 μg/dl at the time of decapitation), compared with adrenalectomized rats (<0.5 μg/dl) led to a 17% down-regulation of SRC-1e in the anterior pituitary but not to changes in SRC-1a mRNA levels (Fig. 8, Table 3). In the hippocampus of the rat brain, in which both MRs and GRs are expressed, we observed no effect on expression of SRC-1a or -1e mRNA (data not shown). These data demonstrate that the ratio of SRC-1 splice variant mRNAs not only differs between cell types but may also be dynamically regulated in vivo in some physiological conditions.
FIG. 8. In vivo differential regulation of SRC-1a (A and B) and SRC-1e (C and D) mRNA abundance in the anterior pituitary in adrenalectomized (Adx) rats implanted with a cholesterol (Chol, A and C) or 100 mg corticosterone (Cort, B and D) pellet. SRC-1 mRNA is expressed in the anterior pituitary (AP) but not in the posterior pituitary (PP). Chronic exposure to moderately elevated levels of corticosterone significantly down-regulated SRC-1e (D) but not SRC-1a mRNA (see also Table 3). On the right-hand side, [14C]microscales are shown, indicating a 2-fold difference in concentration of radioactivity in the lower concentration range, which are relevant for the pituitary sections.
TABLE 3. Expression of SRC-1a and 1e mRNA in the pituitary as a function of circulating corticosterone (Cort) levels
Discussion
We investigated what functional consequences for corticosteroid signaling may follow from the substantial differences in expression that can occur for two splice variants of SRC-1, SRC-1a and -1e. Both splice variants contain AD-1 (CBP/p300 binding), AD-2 [methyl transferase binding (35)] and the weak intrinsic histone (histone 3) acetyl transferase domains (36) and differ only in their C-terminal parts. The relatively small structural difference between SRC-1a and -1e resulted in large differences in coactivation of MRs and GRs in transient transfection studies. There was a pronounced promoter dependence: whereas SRC-1a tended to be somewhat stronger at single GREs, SRC-1e consistently coactivated much more strongly at reporters containing multiple response elements. We observed modest receptor-specific effects: at the TAT-3 Luc reporter the differences between the splice variants tended to be larger for the MR than for the GR. The AF-1 of the MR but not the GR could be coactivated by SRC-1e in absence of the receptor LBD via an as-yet-unidentified SRC-input domain likely localized between amino acids 988 and 1240. We observed a ligand-specific effect for the GR: RU 486-activated GR was coactivated by SRC-1a and -1e in an opposite way, compared with corticosterone. This was reflected in a pronounced difference in the efficacy of RU486 as an antagonist, depending on the relative abundance of the SRC-1 splice variants. Finally, we demonstrated that mRNA abundances of SRC-1a and -1e are differentially regulated by corticosterone in the pituitary.
The CV-1 cells in which we performed most of our experiments have their pros and cons. MR AF-1 is inherently weak in CV-1 cells, which may affect our results. On the other hand, they lack endogenous MRs and GRs and therefore allow the corticosteroid receptor status to be defined. Transactivation via the endogenous GR of A549 cells showed effects of SRC-1a and -1e that were very similar to those observed in CV-1 cells. Whereas the choice of these cells is to some extent arbitrary, we can conclude that the specific interactions are not the consequence of overexpression of the receptor protein, although we did not test cell lines expressing endogenous MRs. CV-1 cells do express endogenous p160 coactivators (14), but for lack of slice variant-specific antibodies, we do not know the relative abundance of endogenous SRC-1a and -1e protein. Because the observed effects in these cells are all in the context of overexpressed coactivators, the in vivo importance of the outcome of the differential interactions remains to be determined. However, differential interaction with nuclear receptors between SRC-1 and SRC-2 have been shown to have large functional consequences for fat tissue (37), despite a degree of functional redundancy between these coactivators (38). Similarly, the differences between SRC-1 splice variants may well bear functional relevance because pronounced differences in SRC-1a and -1e mRNA expression occurs among different glucocorticoid target cells in vivo, in some instances in an all-or-nothing fashion (23). In addition, the differential down-regulation of mRNA we observed in the pituitary (Fig. 8) suggests that the ratio between the splice variants may even be physiologically regulated in certain cell types.
One of the most striking differences we observed between the SRC-1 splice variants was the specific strong coactivation by SRC-1e at multiple GRE-containing regulatory regions in the context of different minimal promoters. It is important to note that this dichotomy did not apply to the often-used MMTV-Luc reporter, which was coactivated by both SRC-1a and -1e (data not shown), but was found for two 5' regulatory regions involved in regulation of endogenous genes, tyrosine aminotransferase and PNMT. Recent data show that multiple GRE-containing promoters are also highly interesting because DNA binding of GRs for such promoters does not require the dimerization interface localized in the second Zn finger of the GR DNA-binding domain (DBD) (25, 26). We consider the MMTV promoter as atypical, in that it does contain several GR binding sites, but in the context of the DBD dimer mutants does not behave as containing multiple GREs (26, 39). Thus, it is difficult to interpret in terms of the single vs. multiple GRE phenotype that we observed for the other promoters that were tested.
The promoter-specific effects seem to be unrelated to recruitment of the coactivators by the receptors but rather to differences in output of SRC-1a and SRC-1e, once bound to the single or multiple receptor dimers. The preferential coactivation of synergistic transactivation may point to a role of small ubiquitin-like modifier (SUMO)-ylation sites, which have been implicated in synergistic activation via NRs (4), particularly in the MR (40), and which are also present in p160 coactivators (41).
Differential effects of SRC-1a and -1e overexpression was observed in the context of induction of the endogenous PNMT mRNA in neuroscreen cells. Because PNMT is a direct target for GRs, this likely reflects differential action on an endogenous promoter. The decreased rather than increased expression that was observed for both splice variants relative to the no-coactivator condition may be caused by competition between overexpressed SRC-1 and other coactivators that may be preferred by this promoter. However, as hypothesized based on analysis of the GR-responsive fragment of this gene (Fig. 2B), SRC-1e led to higher PNMT mRNA levels than SRC-1a. Few other endogenous promoters that function as multiple GRE containing are known. Identification of the different types of GRE constellations of corticosteroid-responsive promoters is a relevant challenge for future studies, i.e. to predict those genes for which differential coactivator expression as studied here can bear in vivo relevance.
In most reporter studies, MRs typically stimulate transcription substantially less than GRs at synergizing promoters (3, 42). At the TAT-3 promoter, SRC-1e seemed to coactivate MRs stronger than GRs. This is probably not due to saturation of transactivation for GRs because the hybrid receptor MMG (5) was strongly coactivated despite significantly higher basal transcription levels than GRs (data not shown). The only other experimental setting we are aware of that causes such a pronounced MR-mediated transcriptional synergy is the phenotype observed after the above-mentioned disruption of the DBD dimer interface of the receptor (25). However, when we tested such dimer mutants, they were coactivated in a manner similar to that of wild-type receptors, suggesting distinct mechanisms (data not shown). We would argue that the high transactivation via MRs we observed after overexpression of an endogenous factor may be indicative of cellular settings in which MRs can act, at least at some promoters, more potently than often assumed (irrespective of the agonist that is used).
The specific coactivation of MR AF-1 may contribute to the somewhat larger stimulatory effect of SRC-1e on MRs, when compared with GRs on the TAT-3 reporter. It has been shown earlier to be potentially coactivated via glucocorticoid receptor-interacting protein/SRC-2 (13) and, in a ligand-dependent way, a complex containing CBP and RNA Helicase A (43). Our data also suggest that SRC-1e can interact, directly or indirectly, with the MR N terminus, leading to coactivation of MR AF-1. Our SRC-1-VP16 hybrid data suggest that this interaction depends on SRC-1 amino acids 988-1240 but not on the Q-rich domain (residues 1053–1123) that has been shown to mediate the interaction between SRCs and the AR N terminus (12) because deletion of this domain did not lead to loss of coactivation. The lack of effect of mutation of the LxxLL motifs in SRC-1 is consistent with SRC-1988–1240 as the dominant interaction surface between SRC-1 and MR-N terminus (Fig. 5). The MR AF-1 coactivation did not occur when the SRC-1e AD-1 was deleted, which suggests that recruitment of CBP/p300 mediates the increased transcription activation that was observed.
Ligand-dependent coactivation by p160 SRCs was earlier shown for the vitamin D receptor, which can engage in preferential interaction in SRC-2/glucocorticoid interacting protein-1 (44) but to our knowledge not for splice variants. We saw no differences in coactivation between aldosterone- and corticosterone-activated MRs, despite the fact that these ligands induce different conformations and molecular interactions of the receptor (43, 45). We tested the effect of the SRC-1 splice variants on antagonists action because the ratio of coactivators and corepressors is thought to determine the extent of antagonism/partial agonism of this antagonist (46, 47, 48). Coactivation of RU486-activated GRs by SRC-1a and -1e was reciprocal to that of corticosterone-activated GRs. The potential functional relevance of this finding is underscored by the different dose-response curves for corticosterone in the absence and presence of 10–7 M RU486: in cells that mainly express SRC-1a [as are present in hypothalamus of the rat brain (23) or to a lesser extent in the pituitary after exposure to elevated concentrations of corticosterone (present study and Fig. 8)], certain promoters may be less potently transactivated, but RU486 treatment is expected to have relatively little effect in such a setting. This is all the more relevant because RU486 has recently been shown to have powerful antidepressant properties in a group of psychotic depressed patients (49). Regional differences for RU486 action may be part of the mechanism by which this drug acts in a clinical setting.
The observed differential expression of SRC-1a and -1e may have implications for cell-specific effects of corticosteroid and other NRs. Also, the functional differences that we observed in cell lines may also be relevant in the context of differential regulation of SRC-1 splice variant activity by posttranslation modifications. Further relevance of the functional differences between SRC-1 splice variants also depends on whether activity or expression of the splice variants is differentially regulated, i.e. whether the ratio between the variants may change. In pituitary, strong hormonal regulation of SRC-1 mRNA has been reported after estrogen and thyroid hormone treatment (34), but no distinction between splice variants has been made. Our data from the anterior pituitary show that differential regulation of the splice variants can take place. The magnitude of down-regulation of SRC-1e mRNA after modest elevations of circulating plasma corticosterone was 17%. This is similar to what we have observed for total SRC-1 mRNA in estradiol-treated ovariectomized female rats (conform Ref. 27 ; our unpublished data). Whether the splice variant-specific regulation reflects effects at the level of transcription, splicing, or mRNA stability is at present not clear. The data do suggest that the ratio between SRC-1a and -1e may not be a constant in some tissues, which may affect the corticosteroid responsiveness of such cells in a promoter-specific fashion. However, it is not a general phenomenon because we did not observe regulation of SRC-1 mRNA in hippocampus after manipulation of corticosteroid levels. We do not know in which cell type(s) of the pituitary this effect takes place; if in all cell types, then the effect would qualify as modest. We also lack important information at the protein level because there are no splice variant-specific antibodies.
In summary, we have observed a number of functional differences between SRC-1a and SRC-1e in their interactions with GRs and MRs, which involve and may in part explain promoter-, receptor-, and ligand-specific effects. We have also shown that the mRNA abundance of SRC-1a and -1e can be differentially regulated in vivo. It is obvious that other promoter elements and cellular proteins constitute a superimposed context that determines the extent to which the SRC-1a/1e differences become manifest. Further studies should address the mechanisms that cause the functional differences we have seen, e.g. dependence on histone acetylation activity, recruitment of cointegrators CBP/p300, histone methyl transferases, or other factors that finally lead to increased transactivation through modification of histone structure and/or recruitment of transcription machinery (19). A major challenge for future studies is to address the relevance of these effects for the transcriptional regulation of endogenous promoters in relevant target tissues in vivo. This should involve specific manipulation of expression of the splice variants by knockdown methods and the identification of relevant GRE-containing regulatory regions in the genome.
Acknowledgments
We thank Drs. C. Bevan, G. Jenster, and K. Yamamoto for kind donation of the expression plasmids for SRC-1 mutants, PRE-tk-Luc reporters, and eTAT-Luc reporter, respectively. We thank Dr. Erno Vreugdenhil for his scientific and technical advice. Diana Rien, Servane Lachize, and Anette Verhoeven are kindly acknowledged for technical assistance.
References
Hollenberg SM, Weinberger C, Ong ES, Cerelli G, Oro A, Lebo R, Thompson EB, Rosenfeld MG, Evans RM 1985 Primary structure and expression of a functional glucocorticoid receptor cDNA. Nature 318:635–641
Arriza JL, Weinberger C, Cerelli G, Glaser TM, Handelin BL, Housmann DE, Evans RM 1987 Cloning of human mineralocorticoid receptor complementary DNA: structural and functional kinship with the glucocorticoid receptor. Science 237:268–275
Rupprecht R, Arriza JL, Spengler D, Reul JMHM, Evans RM, Holsboer F, Damm K 1993 Transactivation and synergistic properties of the mineralocorticoid receptor: relationship to the glucocorticoid receptor. Mol Endocrinol 7:597–603
Iniguez-Lluhi JA, Pearce D 2000 A common motif within the negative regulatory regions of multiple factors inhibits their transcriptional synergy. Mol Cell Biol 20:6040–6050
Pearce D, Yamamoto KR 1993 Mineralocorticoid and glucocorticoid receptor activities distinguished by nonreceptor factors at a composite response element. Science 259:1161–1165
Meijer OC, Williamson A, Dallman MF, Pearce D 2000 Transcriptional repression of the 5-HT1A receptor promoter by corticosterone via mineralocorticoid receptors depends on the cellular context. J Neuroendocrinol 12:245–254
De Kloet ER, Vreugdenhil E, Oitzl MS, Joels M 1998 Brain corticosteroid receptor balance in health and disease. Endocr Rev 19:269–301
Makino S, Gold PW, Schulkin J 1994 Effects of corticosterone on CRH mRNA and content in the bed nucleus of the stria terminalis; comparison with the effects in the central nucleus of the amygdala and the paraventricular nucleus of the hypothalamus. Brain Res 657:141–149
McKenna NJ, Lanz RB, O’Malley BW 1999 Nuclear receptor coregulators: cellular and molecular biology. Endocr Rev 20:321–344
Rosenfeld MG, Glass CK 2001 Coregulator codes of transcriptional regulation by nuclear receptors. J Biol Chem 276:36865–36868[Free Full Text]
Heery DM, Kalkhoven E, Hoare S, Parker MG 1997 A signature motif in transcriptional co-activators mediates binding to nuclear receptors. Nature 387:733–736
Bevan CL, Hoare S, Claessens F, Heery DM, Parker MG 1999 The AF1 and AF2 domains of the androgen receptor interact with distinct regions of SRC1. Mol Cell Biol 19:8383–8392
Fuse H, Kitagawa H, Kato S 2000 Characterization of transactivational property and coactivator mediation of rat mineralocorticoid receptor activation function-1 (AF-1). Mol Endocrinol 14:889–899
Torchia J, Rose DW, Inostroza J, Kamei Y, Westin S, Glass CK, Rosenfeld MG 1997 The transcriptional co-activator p/CIP binds CBP and mediates nuclear-receptor function. Nature 387:677–684
Chen D, Ma H, Hong H, Koh SS, Huang SM, Schurter BT, Aswad DW, Stallcup MR 1999 Regulation of transcription by a protein methyltransferase. Science 284:2174–2177
Hong H, Kohli K, Trivedi A, Johnson DL, Stallcup MR 1996 GRIP1, a novel mouse protein that serves as a transcriptional coactivator in yeast for the hormone binding domains of steroid receptors. Proc Natl Acad Sci USA 93:4948–4952
Xu J, Liao L, Ning G, Yoshida-Komiya H, Deng C, O’Malley BW 2000 The steroid receptor coactivator SRC-3 (p/CIP/RAC3/AIB1/ACTR/TRAM-1) is required for normal growth, puberty, female reproductive function, and mammary gland development. Proc Natl Acad Sci USA 97:6379–6384
Liu Z, Wong J, Tsai SY, Tsai MJ, O’Malley BW 2001 Sequential recruitment of steroid receptor coactivator-1 (SRC-1) and p300 enhances progesterone receptor-dependent initiation and reinitiation of transcription from chromatin. Proc Natl Acad Sci USA 98:12426–12431
Metivier R, Penot G, Hubner MR, Reid G, Brand H, Kos M, Gannon F 2003 Estrogen receptor- directs ordered, cyclical, and combinatorial recruitment of cofactors on a natural target promoter. Cell 115:751–763
Kamei Y, Xu L, Heinzel T, Torchia J, Kurokawa R, Gloss B, Lin SC, Heyman RA, Rose DW, Glass CK, Rosenfeld MG 1996 A CBP integrator complex mediates transcriptional activation and AP-1 inhibition by nuclear receptors. Cell 85:403–414
Ding XF, Anderson CM, Ma H, Hong H, Uht RM, Kushner PJ, Stallcup MR 1998 Nuclear receptor-binding sites of coactivators glucocorticoid interacting protein 1 (GRIP1) and steroid receptor coactivator 1 (SRC-1): multiple motifs with different binding specificities. Mol Endocrinol 12:302–313
Kalkhoven E, Valentine JE, Heery DM, Parker MG 1998 Isoforms of steroid receptor co-activator 1 differ in their ability to potentiate transcription by the oestrogen receptor. EMBO J 17:232–243
Meijer OC, Steenbergen PJ, De Kloet ER 2000 Differential expression and regional distribution of steroid receptor coactivators SRC-1 and SRC-2 in brain and pituitary. Endocrinology 141:2192–2199
Miesfeld R, Godowski PJ, Maler BA, Yamamoto KR 1987 Glucocorticoid receptor mutants that define a small region sufficient for enhancer activation. Science 236:423–427
Liu W, Wang J, Yu GQ, Pearce D 1996 Steroid receptor transcriptional synergy is potentiated by disruption of the DNA binding domain dimer interface. Mol Endocrinol 10:1399–1406
Adams M, Meijer OC, Wang J, Bhargava A, Pearce D 2003 Homodimerization of the glucocorticoid receptor is not essential for response element binding: activation of the phenylethanolamine N-methyltransferase gene by dimerization-defective mutants. Mol Endocrinol 17:2583–2592
Nordeen SK 1988 Luciferase reporter gene vectors for analysis of promoters and enhancers. Biotechniques 6:454–458
Cavailles V, Dauvois S, L’Horset F, Lopez G, Hoare S, Kushner PJ, Parker MG 1995 Nuclear factor RIP140 modulates transcriptional activation by the estrogen receptor. EMBO J 14:3741–3751
Kalkhoven E, Teunissen H, Houweling A, Verrijzer CP, Zantema A 2002 The PHD type zinc finger is an integral part of the CBP acetyltransferase domain. Mol Cell Biol 22:1961–1970
Strahle U, Schmid W, Schutz G 1988 Synergistic action of the glucocorticoid receptor with transcription factors. EMBO J 7:3389–3395
Schmid W, Strahle U, Schutz G, Schmitt J, Stunnenberg H 1989 Glucocorticoid receptor binds cooperatively to adjacent recognition sites. EMBO J 8:2257–2263
Ebert SN, Lindley SE, Bengoechea TG, Bain D, Wong DL 1997 Adrenergic differentiation potential in PC12 cells: influence of sodium butyrate and dexamethasone. Brain Res Mol Brain Res 47:24–30
Smith CL, O’Malley BW 2004 Coregulator function: a key to understanding tissue specificity of selective receptor modulators. Endocr Rev 25:45–71
Misiti S, Schomburg L, Yen PM, Chin WW 1998 Expression and hormonal regulation of coactivator and corepressor genes. Endocrinology 139:2493–2500
Koh SS, Chen D, Lee YH, Stallcup MR 2001 Synergistic enhancement of nuclear receptor function by p160 coactivators and two coactivators with protein methyltransferase activities. J Biol Chem 276:1089–1098
Spencer TE, Jenster G, Burcin MM, Allis CD, Zhou J, Mizzen CA, McKenna NJ, Onate SA, Tsai SY, Tsai MJ, O’Malley BW 1997 Steroid receptor coactivator-1 is a histone acetyltransferase. Nature 389:194–198
Picard F, Gehin M, Annicotte JS, Rocchi S, Champy MF, O’Malley BW, Chambon P, Auwerx J 2002 SRC-1 and TIF2 control energy balance between white and brown adipose tissues. Cell 111:931–941
Xu J, Qiu Y, DeMayo FJ, Tsai SY, Tsai MJ, O’Malley BW 1998 Partial hormone resistance in mice with disruption of the steroid receptor coactivator-1 (SRC-1) gene. Science 279:1922–1925
Reichardt HM, Kaestner KH, Tuckermann J, Kretz O, Wessely O, Bock R, Gass P, Schmid W, Herrlich P, Angel P, Schutz G 1998 DNA binding of the glucocorticoid receptor is not essential for survival. Cell 93:531–541
Tallec LP, Kirsh O, Lecomte MC, Viengchareun S, Zennaro MC, Dejean A, Lombes M 2003 Protein inhibitor of activated signal transducer and activator of transcription 1 interacts with the N-terminal domain of mineralocorticoid receptor and represses its transcriptional activity: implication of small ubiquitin-related modifier 1 modification. Mol Endocrinol 17:2529–2542
Kotaja N, Karvonen U, Janne OA, Palvimo JJ 2002 The nuclear receptor interaction domain of GRIP1 is modulated by covalent attachment of SUMO-1. J Biol Chem 277:30283–30288
Arriza JL, Simerly RB, Swanson LW, Evans RM 1988 The neuronal mineralocorticoid receptor as a mediator of glucocorticoid response. Neuron 1:887–900
Kitagawa H, Yanagisawa J, Fuse H, Ogawa S, Yogiashi Y, Okuno A, Nagasawa H, Nakajima T, Matsumoto T, Kato S 2002 Ligand-selective potentiation of rat mineralocorticoid receptor activation function 1 by a CBP-containing histone acetyltransferase complex. Mol Cell Biol 22:3698–3706
Takeyama K, Masuhiro Y, Fuse H, Endoh H, Murayama A, Kitanaka S, Suzawa M, Yanagisawa J, Kato S 1999 Selective interaction of vitamin D receptor with transcriptional coactivators by a vitamin D analog. Mol Cell Biol 19:1049–1055
Rogerson FM, Brennan FE, Fuller PJ 2004 Mineralocorticoid receptor binding, structure and function. Mol Cell Endocrinol 217:203–212
Zhang X, Jeyakumar M, Petukhov S, Bagchi MK 1998 A nuclear receptor corepressor modulates transcriptional activity of antagonist-occupied steroid hormone receptor. Mol Endocrinol 12:513–524
Schulz M, Eggert M, Baniahmad A, Dostert A, Heinzel T, Renkawitz R 2002 RU486 induced glucocorticoid receptor agonism is controlled by the receptor N terminus and by corepressor binding. J Biol Chem 277:26238–26243
Liu Z, Auboeuf D, Wong J, Chen JD, Tsai SY, Tsai MJ, O’Malley BW 2002 Coactivator/corepressor ratios modulate PR-mediated transcription by the selective receptor modulator RU486. Proc Natl Acad Sci USA 99:7940–7944
Belanoff J, Rothschild A, Cassidy F, DeBattista C, Baulieu E, Schold C, Schatzberg A 2002 An open label trial of C-1073 (mifepristone) for psychotic major depression. Biol Psychiatry 52:386–392(O. C. Meijer, E. Kalkhove)
Address all correspondence and requests for reprints to: O. C. Meijer, Division of Medical Pharmacology, Leiden/Amsterdam Center for Drug Research, P.O. Box 9503, 2300 RA Leiden, The Netherlands. E-mail: o.meijer@lacdr.leidenuniv.nl.
Abstract
The mechanisms of receptor- and cell-specific effects of the adrenal corticosteroid hormones via mineralo- (MRs) and glucocorticoid receptors (GRs) are still poorly understood. Because the expression levels of two splice variants of the steroid receptor coactivator-1 (SRC-1) 1a and 1e, can differ significantly in certain cell populations, we tested the hypothesis that their relative abundance could determine cell- and receptor-specific effects of corticosteroid receptor-mediated transcription. In transient transfections, we demonstrate three novel types of SRC-1a- and SRC-1e-specific effects for corticosteroid receptors. One is promoter dependence: SRC-1e much more potently coactivated transcription from several multiple response element-containing promoters. Mammalian 1-hydrid studies indicated that this likely does not involve promoter-specific coactivator recruitment. Endogenous phenylethanolamine-N-methyltransferase mRNA induction via GRs was also differentially affected by the splice variants. Another type is receptor specificity: responses mediated by the N-terminal part of the MR, but not the GR, were augmented by SRC-1e at synergizing response elements. SRC fragment SRC988–1240 by the MR but not the GR N-terminal fragment in a 1-hybrid assay. The last type, for GRs, is ligand dependence. Due to effects on partial agonism of RU486-activated GRs, different ratios of SRC-1a and 1e can lead to large differences in the extent of antagonism of RU486 on GR-mediated transcription. Furthermore, we show that SRC-1e but not SRC-1a mRNA expression was regulated in the pituitary by corticosterone. We conclude that the cellular differences in SRC-1a to SRC-1e ratio demonstrated in vivo might be involved in cell-specific responses to corticosteroids in a promoter- and ligand-dependent way.
Introduction
THE NUMEROUS EFFECTS of adrenal corticosteroid hormones in the body are mediated by glucocorticoid (GRs) (1) and mineralocorticoid receptors (MRs) (2). Both receptors are members of the large nuclear receptor family and accordingly act as transcription factors to transactivate or transrepress specific target genes. The MRs and GRs have a high degree of homology in their DNA binding and ligand binding domain but differ considerably in the N-terminal part of the receptor (2). Thus, these receptors recognize the same response elements on the DNA, but they differ in their transactivational (3, 4) and transrepressive (5, 6) properties.
In some tissues the two receptor types mediate in a coordinate fashion signaling by cortisol and corticosterone, the most important glucocorticoids in man and rodents, respectively. In cells expressing both receptor types, MRs and GRs mediate different and at times opposite effects on cellular physiology, underscoring the importance of differential transcriptional properties of these two receptors (7). Besides receptor specificity, there are many instances of cell-specific gene regulation by corticosteroids, unexplained by receptor expression. For instance, in the rat brain, CRH mRNA is down-regulated by corticosterone via GRs in the hypothalamus but up-regulated in the amygdala nucleus, presumably via cell specific transcriptional effects (8). The mechanisms underlying such cell and receptor specificity are as yet largely unknown.
Over the last several years, a large number of coregulatory proteins that influence transcriptional responses of nuclear receptors has been discovered (9). The family of p160 steroid receptor coactivators (SRCs) consists of three genes (10), and each of their products can bind to the activation function (AF)-2 of nuclear receptors through interactions with LxxLL motifs or nuclear receptor (NR) boxes (11). In addition, SRCs may also interact with the less conserved N-terminal domains of steroid receptors (12, 13). The actual stimulation of transcription depends on direct histone acetyl transferase activity and recruitment of cointegrators such as cAMP response element-binding protein (CBP)/p300 (14) or the methyl transferase CARM-1 (15). The interaction with at least one of the SRCs, which are expressed in a cell-specific manner (16, 17), is thought to be necessary for transcriptional stimulation to occur (18, 19). The interaction between the SRC variant and the specific NR is thought to determine the nature of the interaction and magnitude of coactivation.
Five different splice variants of SRC-1 were originally reported (20), of which SRC-1a and 1e have been consistently found (21, 22). These differ only at their carboxy terminus, which is shorter in SRC-1e. The SRC-1a-specific 56 amino acids contain an extra NR box and a potential suppressor domain. These splice variants were shown to differ in their interactions with the (isolated) ligand-binding domains (LBDs) of NRs (21) and their functional interactions with the estrogen receptor- (22). In addition, the expression of SRC-1a and 1e mRNA is cell specific. In crucial corticosteroid-sensitive cell populations in the brain, considerable differences in the expression of the splice variants were observed (23).
In view of the observed differences between SRC-1a and -1e, we tested, using cultured cells as a model, the hypothesis that there are differential interactions between SRC-1 splice variants and corticosteroid receptors, which may contribute to cell- and receptor-specific corticosteroid effects. Hence, we tested the functional and physical interactions of SRC-1a and 1e with MRs and GRs, using transient reporter assays and mammalian 1-hybrid studies. We also tested the hypothesis that the abundance of SRC1a and SRC-1e mRNA is differentially regulated. We find that these splice variants differentially affect transcription in a receptor-, promoter-, and ligand-dependent fashion and that there can be specific physical interactions between SRC-1 and the MR N-terminal fragments. In addition, we show that the relative abundance of SRC-1a and SRC-1e mRNA is subject to regulation in the pituitary in vivo.
Materials and Methods
Transient transfections
CV-1 cells were grown in DMEM supplemented with 5% fetal calf serum (GIBCO, UK), A549 cells in DMEM/HAM F12 mix with 10% fetal calf serum (GIBCO). For transfections, cells were plated in 24-well plates (Greiner Bio-One, Alphen aan den Rijn, The Netherlands) at 3 x 104 cells/well, and charcoal-stripped serum was used. The cells were transfected using SuperFect (Promega, Madison, WI) at a DNA to superfect ratio of 1:2 (CV-1) or 1:4 (A549). For CV-1 cells steroid receptor expression plasmids were used at 100 ng/well. In CV-1 cells 100 ng of SRC-1 expression plasmid and reporters were used per well, based on a number of initial titration studies (showing qualitatively similar but quantitatively increasing effects from 50 to 400 ng plasmid/well), whereas for A549 cells, we used 200 ng of SRC expression vector and reporter plasmid per well. The transcriptionally inert plasmid pSP65 was used to bring the total amount of DNA to 1 μg/well. As control plasmid we used 1 ng of pCMV-R (Promega) coding for Renilla luciferase controlled by cytomegalovirus (CMV) promoter. We found that expression of this promoter is not influenced by activation of either MR or GR in the cell, in contrast to the pTK-R reporter (Promega), which was consistently repressed by activated GR (data not shown). One day after transfection, the cells were treated with corticosterone (CV-1 cells/rat receptors), cortisol (endogenous human receptor in A549 cells), aldosterone and/or antagonists RU486 (GR), and spironolactone (MR). Unless indicated otherwise, all ligands were given at a dose of 10–7 M, which maximally activates both MRs and GRs, except for aldosterone, which was used at the saturating concentration of 5 x 10–9 M. After 24 h the cells were harvested and firefly and Renilla luciferase activity was determined using the Promega dual label reporter assay and a luminometer (LUMAT LB 9507, Berthold, Bad Wildbad, Germany).
All conditions were tested at least in triplicate, and each experiment was repeated three times with essentially similar results. Data are expressed as average ± SEM. All results were analyzed by ANOVA and statistically tested with post hoc Tukey Kramer test. For the ligand-activated receptors, statistical significance (P < 0.05) was attributed based on differences in fold induction by hormone (i.e. all luciferase values from glucocorticoid response element (GRE)-containing reporters were normalized to Renilla luciferase. Induction was calculated for each steroid-treated sample as signal in the presence of hormone divided by the average signal from the same condition in absence of hormone).
Plasmids
Most plasmids we used have been described previously. Human MR and GR expression plasmids (3) were kindly provided by Dr D. Spengler, with permission of Dr. R. Evans (Salk Institute, San Diego, CA). For rat MRs and GRs, and truncations and chimeras thereof, we used the 6RMR and 6RGR-based plasmids (5, 24). These and the pODLO-based reporter plasmids TAT-1-TATA and TAT3-TATA were described earlier (25), as were the reporters driven by rat phenylethanolamine-N-methyltransferase (PNMT) promoter fragment containing multiple GREs and endogenous TAT 5' regulatory region (26). pSG-5 based expression vectors for full-length and mutant SRC-1a and SRC-1e were also described (22). The SRC-1 plasmids used in this study contain a 2x FLAG-tag at the 5' end of the cDNA. The Q plasmid (SRC-1e 1053–1123) was kindly provided by Dr. C. Bevan (Imperial College, London, UK). The pPRE1-tk-Luc and pPRE2-tk-Luc, and mouse mammary tumor virus (MMTV)-luciferase (LUC) plasmids were kindly provided by Dr. G. Jenster (Rotterdam University Medical Center, Rotterdam, The Netherlands). The pPRE2-tk-Luc reporter was constructed by inserting the XhoI-digested (PRE)2TK fragment from pBL2(PRE)2TK-CAT plasmid (27) into the XhoI-digested pGL3basic vector (Promega). The p(PRE)1-tk-Luc reporter was generated by deleting one progesterone response element (PRE) (AGAACAtccTGTACA) from the p(PRE)2-tk-Luc plasmid using the two flanking BsrGI sites. The expression vectors for SRC-1 fragments fused to the VP16-AD were generated by excision of EcoRI fragments from the pSG424 constructs (21) and cloning them into pSG5-VP16 (28).
Western blots
Cells from wells parallel to luciferase experiments were lysed in 0.5 x radioimmunoprecipitation assay and subjected to SDS-PAGE. Proteins were transferred onto Immobilon membranes (Millipore Corp., Bedford, MA) and processed as described (29). Blots were blocked in 10 mM Tris-HCl (pH 8.0), 150 mM NaCl, and 0.05% Tween 20 containing 5% nonfat dried milk powder and subsequently incubated with the M2 primary anti-FLAG antibody (1:4000, Sigma, St. Louis, MO), the M-20 GR antibody (1:500, Santa-Cruz Biotechnology, Santa Cruz, CA), or the monoclonal anti--tubulin antibody (Sigma, 1:1000). After a wash, blots were incubated with peroxidase-conjugated antibodies (1:10,000; Jackson ImmunoResearch Laboratories, West Grove, PA). Blots were washed again, and immunoreactive bands were visualized by enhanced chemiluminescence.
Quantitative PCR quantification of PNMT mRNA
Neuroscreen-1 cells (Cellomics Europe) were grown in RPMI 1640 medium with L-glutamine (Life Technologies, Inc., Grand Island, NY), supplemented with 10% horse serum, and 5% fetal bovine serum. Cells were transfected with 3 μg of SRC-1A or 1E plasmid per 2 million cells or mock transfected, using a Nucleofector and electroporation buffer V (Amaxa Biosystems). Cells were plated at 106 per well in collagen-coated 6-well plates and after 1 d were exposed to dexamethasone (10–6 M) or vehicle for 18 h. Total RNA was isolated using an RNeasy kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. During RNA extraction, samples were treated on-column with RNase-free DNase I (Qiagen). Total RNA required for standard curves was obtained from the rat adrenal and isolated with Trizol reagent (GIBCO BRL). Potential genomic DNA was degraded with DNase I (Invitrogen Corp., Carlsbad, CA). The purity and concentration of isolated total RNA was measured by the ND-1000 spectrophotometer (NanoDrop Technology) where the ratio of absorption at 260 and 280 nm for all samples was at least 1.9. First-strand cDNA was synthesized with SuperScript II reverse transcriptase (Invitrogen) and Oligo(dT)12–18 (Invitrogen) in accordance with the manufacturer’s specifications. Real-time PCR was performed on a LightCycler 2.0 (Roche Diagnostics, Indianapolis, IN) using LightCycler FastStart DNA Masterplus SYBR Green I (Roche Diagnostics). The expression of PNMT was normalized against ?-actin. The sequences of the used primers are: Actb forward, tgaccgagcgtggctaca; Actb reverse, cagcttctctttaatgtcacgca; PNMT forward, gcgagggtgaagcgagtcttgcc; and PNMT reverse, tacccccgggcctcagcagc. Statistical significance of differences in normalized mRNA levels (four wells per condition) was determined by the Kruskal-Wallis test followed by the Mann-Whitney U test.
In situ hybridization
For determination of pituitary mRNA levels of SRC-1a and 1e, male rats (240 g Wistar rats, Charles River, Germany) were adrenalectomized or sham operated (n = 6) under gas anesthesia using the dorsal approach. Six animals per group were sc implanted with a 100-mg pellet containing 100% cholesterol, 20% corticosterone, or 100% corticosterone. All adrenalectomized animals had access to saline. After 1 wk the animals were decapitated, and pituitaries were collected and frozen at –80 C. The animal experiments were performed in accordance with the European Communities Council Directive 86/609/EEC and with approval from the animal care committee of the Faculty of Medicine, Leiden University (UDEC no. 99062). The tissue was cut at 14 μm on a cryostat and thaw mounted on poly-L-lysine-coated slides, each slide containing two sections from four rats, each rat belonging to a different treatment group. In situ hybridization with SRC-1a- and 1e-specific end-labeled oligonucleotide probes was performed as described (23). The signal was quantified from film (X-OMAT AR, Kodak, Rochester, NY) using an automated image analysis system (Paes Nederland, Zoeterwoude). The signals were in the linear range of gray values according to the [14C] RPA 504 microscales (Amserham, Aylesbury, UK). Four sections per animal, i.e. 24 sections per treatment group, were analyzed by ANOVA and post hoc Tukey-Kramer test.
Results
Promoter-specific coactivation
As a basic approach to study possible differential coactivation of rat MRs and GRs by SRC-1a and SRC-1e, we chose to use transient transfection of CV-1 cells. These cells contain no endogenous corticosteroid receptors and therefore allow good definition of the type of steroid hormone receptors that are present. No induction of any reporter was observed in the absence of transfected receptor in these cells (data not shown). At the TAT-1-Luc reporter, which harbors a single GRE in front of the Drosophila alcohol dehydrogenase promoter TATA box, corticosterone-activated MRs (10–7 M) transactivated very modestly in the absence of overexpressed coactivator. Statistical analysis indicated significant difference in hormone induction among the three coactivator conditions. Overexpression of SRC-1a led to a significantly enhanced reporter activity. Transactivation in the presence of hormone was increased in the presence of SRC-1e, but this did not reach significance relative to the no-coactivator condition when calculated as fold induction by hormone (Fig. 1A). Transactivation via GR was similarly different between conditions, i.e. significantly enhanced in the presence of overexpressed SRC-1a but not significantly by SRC-1e (Fig. 1B).
FIG. 1. Promoter-specific coactivation by SRC-1a and -1e. SRC-1a or -1e was overexpressed in CV-1 cells containing MRs (A, C, and E) and GRs (B, D, and F), without hormone (open bars) and in the presence of 10–7 M corticosterone (closed bars) using the TAT-1 (A and B), TAT-3 reporter (C and D), and reporter backbone ODLO (E and F). Reporter activity is expressed as luciferase normalized by CMV promoter-driven Renilla luciferase. Fold induction is indicated in numbers above the bars. Asterisks indicate significant differences in fold induction by corticosterone relative to both other conditions, unless a specific comparison is indicated (P < 0.05). A and B, At the TAT-1 reporter, SRC-1a overexpression leads to modestly stronger coactivation for both receptors. C and D, At the TAT-3 reporter, SRC-1a has no potentiating effect for either receptor, but SRC-1e shows very strong coactivation. In boxes are the fold coactivation by SRC-1a and -1e and the ratio between these two values. G, Expression of FLAG-tagged SRC-1A and SRC-1E on Western blot from transfections in which the differential coactivation was observed, indicating equal expression levels of the splice variants. H, Expression of the steroid receptor expression plasmid as measured by GR immunoreactivity on Western blot is not affected by hormone treatment or coactivator expression in samples run in parallel to luciferase assays. Bottom row shows tubulin expression as control for loading.
On the TAT-3-Luc reporter, containing a tandem repeat of three consensus GREs in front of the alcohol dehydrogenase promoter TATA box, transactivation in absence of overexpressed SRC-1 is much stronger, notably for GRs, which are known to act synergistically at multiple response elements (30, 31). At this reporter coactivation by SRC-1 splice variants showed large differences in absolute transactivation and fold induction relative to the no-hormone condition. SRC-1a overexpression did not lead to any potentiation of the transcriptional response for either the MR or GR. In contrast, SRC-1e strongly potentiated the MR-mediated response; absolute levels of reporter activity were of the same order as those observed for GRs in the absence of overexpressed coactivator (Fig. 1C). Also, GR-mediated action was enhanced by overexpression of SRC-1e (Fig. 1D). MRs and GRs differed qualitatively in their coactivation by the SRC-1 splice variants. Whereas in both cases SRC-1e is the stronger coactivator on the TAT-3-Luc reporter, the SRC-1e to 1a ratio was consistently higher for the MR.
The pODLO backbone of the TAT-1 and TAT-3 plasmids was not influenced by corticosterone activation of either corticosteroid receptor and/or overexpression of SRC-1a and SRC-1e (Fig. 1, E and F). Furthermore, SRC-1a and SRC-1e were expressed at similar levels under our experimental conditions (Fig. 1G), Also, the expression vector of the receptors was not affected by coactivator expression or hormone treatment (shown for GR in Fig. 1H). Human MRs and GRs showed the same promoter dependence in coactivation as the rat receptors in transient transfection studies in CV-1 cells (data not shown).
The results suggested that the number of GREs may be a critical parameter that determines differential coactivation by SRC-1a and -1e. Hence, we tested in CV-1 cells, in parallel to the TAT-1 and TAT-3 reporters, a number of other reporters that differ only in the number of GREs (Table 1). PRE-1-tk-Luc and PRE-2-tk-Luc contain one and two GREs upstream of a minimal tk promoter, respectively. The rat PNMT has a multiple GRE-containing fragment of the rat PNMT gene 5' flanking region upstream of a SV40 basal promoter. The PNMT 3 reporter contains mutations in three GRE half-sites and behaves like a single GRE-containing promoter (26). As shown in Table 1, in all cases the coactivation by SRC-1a was stronger on the single response element-containing reporter, whereas SRC-1e coactivated more strongly at the multiple GREs. Also shown is the ratio of coactivation by SRC-1a and SRC-1e. For the TAT-3 reporter, this ratio is consistently higher for coactivation of the MR, compared with the GR. Coactivation of GR at the often-used GRE-containing MMTV reporter was not different between SRC-1a and SRC-1e (data not shown).
TABLE 1. The fold coactivation of 10–7 M corticosterone-activated MRs and GRs brought about by SRC-1a and -1e on related promoters that contain single or multiple GREs, and the ratio between the SRC-1e and 1a effect
Next, we were interested to see whether these promoter-specific SRC-1a/1e effects would be reproducible in a different cell line, preferably one that is commonly used, endogenously expresses one of the receptor types, and is of human origin (given that we expressed human SRC-1 variants). The human A549 lung carcinoma cell line fulfills these criteria, and we used it to study coactivation of the endogenous human GR (Fig. 2). Transcription synergy at multiple GREs was notably less strong than in CV-1 cells. Coactivation by SRC-1a and -1e of the simple TAT-1 and TAT-3 reporters was comparable to that in CV-1 cells (Fig. 2A): also in this setting SRC-1a overexpression leads to coactivation at the single GRE-driven TAT-1 reporter, whereas at the TAT-3 reporter, it does not coactivate. SRC-1e coactivation did not reach statistical significance on the TAT-1 reporter (calculated as fold induction by hormone), whereas at the TAT-3 reporter, SRC-1e potently coactivated. Induction of transcription was also achieved by treatment of the cells with 1 nM dexamethasone but not with 1 nM of aldosterone (not shown), pointing to GR rather than MR as the mediator of effects of the mixed agonist corticosterone.
FIG. 2. Coactivation of endogenous GRs in A549 cells by SRC-1a and SRC-1e at synthetic and endogenous promoter fragments depends on the number of response elements. A, Coactivation of the TAT-1 and TAT-3 reporters. Open bars, no hormone; closed bars, 10–7 M cortisol. Reporter activity is normalized with pCMV-Renilla. Fold induction by hormone is indicated above the bars. Asterisks indicate statistically significant differences from the no-coactivator condition at P < 0.05. At the TAT-1-luc reporter by SRC-1a is a modestly stronger coactivator; at the TAT-3-Luc reporter, SRC-1e is much stronger. There are no effects of hormone or SRC expression on the pODLO empty vector. B, Fold induction by 10–7 M cortisol of two endogenous regulatory fragments that harbor a single (eTAT) or multiple (PNMT) response elements again shows preferential coactivation of multiple response elements by SRC-1e. Values represent normalized reporter activity in the presence of hormone divided by those in the absence of hormone. C, Induction of PNMT mRNA in neuroscreen cells via GR activation is lower in the presence of overexpressed SRC-1, but SRC-1e coactivates stronger than SRC-1a. Asterisks indicate statistically significant difference from both other groups by Mann-Whitney U test (P < 0.05).
We also tested coactivation of GRE-containing regulatory fragments from two glucocorticoid-regulated genes, which both were activated via endogenous GR in A549 cells (shown in Fig. 2B as fold induction by hormone because of large differences in absolute luciferase levels from the used reporters). Glucocorticoid responsiveness of the eTAT-luc reporter depends on the regulatory 5' steroid-responsive fragment of the tyrosine aminotranferase promoter, containing a single GRE (26). SRC-1a but not SRC-1e overexpression led to significantly increased transactivation of the eTAT-luc reporter. In contrast, the glucocorticoid-responsive multiple GRE-containing fragment from the PNMT 5' regulatory region showed coactivation only in the presence of overexpressed SRC-1e.
We next studied whether differential coactivation by the splice variants would hold in the context of an endogenous gene. As a model we took induction of the PNMT gene via GR by dexamethasone in neuroscreen-1 cells, a subclone of the PC-12 cell line for which the phenomenon was reported (32). Without addition of dexamethasone to the medium, no PNMT mRNA could be measured by quantitative PCR (data not shown). After treatment with 1 μM dexamethasone, there was a clear induction of PNMT mRNA in mock-transfected cells (Fig. 2C). Overexpression of SRC-1A and 1E both led to lower levels of PNMT mRNA, but expression was modestly, but significantly, higher in cells overexpressing SRC-1E than in cells overexpressing SRC-1A, as hypothesized based on the results with the PNMT GRE-containing region in the context of a reporter plasmid.
Thus, the promoter-dependent coactivation in relation to the presence of multiple GREs was observed using overexpressed as well as (for GR) endogenous receptors, in the contexts of simple and more complex regulatory regions, and held for induction of the endogenous PNMT gene.
Input domains of the SRCs
The promoter-dependent coactivation by SRC-1 splice variants may reflect differential recruitment of the coactivators by the receptors. To test this possibility, we studied the interactions between MR and GR with different domains of the SRC-1 protein at the TAT-1 and TAT-3 reporter, coexpressing in CV-1 cells MR or GR with hybrids of SRC-1 fragments and the VP16 activation domain (AD) as shown in Figure 3A. The hypothesis of preferential coactivator recruitment as the basis for promoter-specific effects would be supported by, for example, a higher degree of receptor interaction of the SRC-1a-specific fragment at the TAT-1 reporter, compared with the TAT-3 reporter. At both reporters the expression of the central LxxLL (NR box) containing fragment SRC-1570–780 fused to VP16-AD led to increased reporter activity after activation of either MR and GR by 10–7 M corticosterone, compared with expression of the VP16-AD alone (Fig. 3). Also, the other NR box-containing fragment SRC-1A1241–1441 led to increased reporter activity at both reporters. The SRC-1e C-terminal region (1241–1399) or that of SRC-1a from which the NR box was deleted (1241–1385) showed no sign of interaction with the steroid receptors. These data provide no evidence for promoter-specific recruitment of SRC-1 splice variants as the cause of promoter-specific coactivation because there was no difference in recruitment of the splice variant-specific carboxy-terminal parts of SRC-1. The only other notable difference was a decreased activation for the fragment SRC-1988–1240 at the TAT-3 reporter, which was specific for GR, but did not reach significance in the ANOVA post hoc test (Fig. 3C; also see below).
FIG. 3. Mammalian 1-hybrid experiments suggest that the promoter-dependent coactivation is not due to promoter-dependent interactions between steroid receptors and SRC-1 NR box-containing fragments. A, Schematic representation of SRC-1 protein. The underlined areas were fused to the VP16 activation domain and used to probe interactions with MRs and GRs. Both at the TAT-1 reporter (B) and TAT-3 reporter (C), the LxxLL (or NR box)-containing fragments of SRC-1 led to increased reporter activity, compared with the condition of steroid receptor incubated in the presence of the VP16 AD alone in CV-1 cells. The SRC-1a-specific C-terminal NR box is present in the 1241–1440 fragment but not in the truncated 1241–1385 or SRC-1e-specific fragment 1241–1399. It clearly interacts with MRs and GRs, irrespective of the promoter context. Asterisks indicate significantly increased induction, compared with receptor in the presence of nonfused VP16 AD at P < 0.01.
Because we found that, at least on the TAT-3 reporter, the differences between splice variants were larger for MRs by SRC-1e than for GRs, we studied coactivation of N- or C-terminally truncated receptors with this reporter. We used CV-1 cells to allow distinction between effects on the two activation functions AF-1 (located in the N terminus) and AF-2 (located in the C-terminal LBD) of MRs and GRs. Members of the p160 SRC family principally coactivate the AF-2 of nuclear receptors but may also act at the N-terminally located AF-1 (13). As in the earlier experiments, full-length GRs and MRs were coactivated strongly by SRC-1e but not by SRC-1a on the TAT3 reporter (not shown). In line with the original discovery of SRCs as LBD-interacting proteins, the N-terminally deleted GG and MM were strongly and comparably coactivated by SRC-1e and not by SRC-1a at this reporter (Fig. 4, A and C). The C-terminal truncation MM and GG are constitutively active transcription factors (Fig. 4, B and D). Transactivation by GG was not influenced by overexpression of SRC-1 splice variants (Fig. 4D). For MM we observed a modest but consistent (approximately 2-fold) coactivation by SRC-1e (Fig. 4B).
FIG. 4. The MR, but not the GR N terminus, can be coactivated by SRC-1e on the TAT-3-Luc reporter. SRC-1a and SRC-1e were overexpressed in CV-1 cells with truncations of the corticosteroid receptors containing the ligand-dependent AF-2 (NN; A and C) or ligand-independent AF-1 (NN; B and D). Open bars, no hormone, closed bars, 10–7 M corticosterone. The AF-2 of both receptors shows coactivation similar to full-length receptors. Fold induction by hormone is indicated above the bars. The AF-1 containing MR fragment (C) but not the GR fragment (D) is consistently coactivated by SRC-1e in this setting. Asterisks indicate significant differences at P < 0.01 in fold induction (A and C) or reporter activity (normalized to pCMV-Renilla), relative to both other conditions, unless a specific comparison is indicated.
To better understand the coactivation of the MR N-terminal part, we tested coactivation of the receptors on the TAT-3 reporter by SRC-1e variants in which the known receptor-interacting domains of SRC-1 were deleted, namely the three NR boxes [M123; (22)] and the Q-rich domain, which can interact with the androgen receptor [Q construct; (12)]. In case of the full-length receptors as well as for the AF-2-driven N-terminal truncations of the MR and GR, deletion of the NR boxes abolished coactivation by SRC-1e (Fig. 5, A and B). In contrast, SRC-1e that lacked its NR boxes or the Q-rich domain could still coactivate the C-terminally deleted MR but not GR (Fig. 5C). SRC-1e in which the CBP-binding AD-1 (amino acids 900–950) was deleted could not stimulate transcription in this setting (Fig. 5E), which argues for the specificity of the observed stimulating interactions. These data demonstrate that SRC-1 coactivation effects on AF-2-containing receptors depend on the already well-characterized NR box interaction motifs but that effects on the MR N terminus (likely mediated by AF-1) depend on an as-yet-unidentified SRC domain.
FIG. 5. Coactivation of the AF-1-containing N-terminal domains of MRs by SRC-1e at the TAT-3 reporter in CV-1 cells does not depend on NR boxes or the Q-rich domain (amino acids 1053–1123) but involves the region 988-1240 of the SRC protein. A, The full-length MR and GR cannot be coactivated when the three central NR boxes are deleted (M123 construct), whereas the Q-rich domain of SRC-1e (absent in Q) is dispensable for coactivation. Coactivation at the TAT-3 reporter is stronger for MR than GR (see also Fig. 1). B, The M123 mutant SRC cannot coactivate AF-2 of the N-terminal deletions of MR and GR. C, The MR N terminus was coactivated by SRC-1e independent of the Q-rich domain or the NR boxes, whereas the GG construct was not coactivated by any SRC variant. D, 1-hybrid assays using the TAT-3 reporter. Direct interactions between the reporter-bound receptor and SRC fragments fused to the VP16 activation domain lead to increased reporter activity. Values represent induction by 10–7 M corticosterone, relative to the condition of steroid receptor cotransfected with the VP16 AD alone. The MR but not the GR N terminus showed a specific interaction with SRC fragment 988-1240 fused to the VP16 AD. No interaction was observed for the NR box containing SRC-1 fragments. E, Deletion of activation domain 1 (amino acids 900–950) from SRC-1e abolishes the coactivation of the MM protein. Asterisks indicate significant differences in reporter activity, compared with the no-coactivator (for D: fusion) condition (P < 0.01).
To further substantiate the coactivation of the MR N terminus by SRC-1e, we evaluated the molecular interactions between the truncated receptors bound to the TAT-3 reporter and SRC fragments (that are shown in Fig. 3A) fused to the VP16 activation domain. As depicted in Fig. 5D, SRC-1988–1240 induced reporter activity over VP16 background for the MR but not the GR N terminus. No interactions with other fragments, such as those containing the NR boxes, were observed (Fig. 5D and data not shown). Whereas deletion of the Q-rich domain (1053–1123) does not impair coactivation, other residues in the region between amino acids 988 and 1240 are likely to be responsible for SRC-1e interaction with, and subsequent potentiation of. the MM protein.
Thus, besides pronounced reporter specificity, the SRC-1 splice variants also showed a degree of receptor specificity, apparent from selective coactivation of MR AF-1 by SRC-1e and the interaction observed in the 1-hybrid assay. In the context of the full-length receptor (Fig. 3C), no interaction between the MR and SRC-1988–1240 fragment is apparent. However, the observed lower reporter induction for full-length GR at the TAT-3 reporter in Fig. 3 does suggest a degree of receptor specificity for full-length receptors in their interaction with SRC-1988–1240, which may be related to the apparent stronger coactivation of MR at multiple response elements.
Ligand dependence
We further tested in CV-1 cells whether, besides the strong promoter and modest receptor specificity for SRC-1a and 1e, different ligands may also determine coactivation by these splice variants. Because the effect of antagonists is thought to depend on relative coactivator to corepressor ratio, we tested the coactivation by SRC-1a and -1e for the partial agonistic effects of the GR antagonist RU486 (Fig. 6). The transactivation on the TAT3 Luc reporter induced by 10–7 M RU486 was consistently around 15%, compared with corticosterone in absence of overexpressed coactivator (16 ± 4% over all experiments). The coactivation of RU486-activated GR by SRC-1 splice variants was strikingly different from corticosterone-activated GR: similar to our earlier experiments, SRC-1e but not -1a coactivated in the presence of 10–7 M corticosterone (Fig. 6B), but in presence of 10–7 M RU486, SRC-1a rather than -1e coactivated GR at the TAT3 reporter (Fig. 6A). When the MR was incubated with 10–7 M of the antagonist spironolactone, a modest partial agonism was also observed, but in this case the coactivation by SRC-1a and -1e was similar to that of corticosterone-activated MR (data not shown). Also, the coactivation of MR in the presence of 5 x 10–9 M of the full agonist aldosterone was not different from what was observed with corticosterone (not shown).
FIG. 6. Coactivation of GRs on the TAT-3-luc reporter in CV-1 cells is ligand dependent. A, Partial agonism of 10–7 M RU486 at the GR is higher when SRC-1a, rather than SRC-1e, is overexpressed. B, With 10–7 M corticosterone as a ligand, overexpression of SRC-1e, but not SRC-1a, leads to higher transactivation. Hormone induction relative to the no-coactivator condition is indicated above the bars. Asterisks indicate significant a difference in fold induction, compared with the no-coactivator condition (P < 0.01).
Because antagonists are typically used in the presence of agonists, we constructed a dose-response curve for corticosterone on the GR in the presence or absence of 10–7 M RU486 and determined the antagonist/partial agonistic effect of RU486 as a function of overexpressed SRC-1a or SRC-1e (Fig. 7). The dose-response curve for the GR-mediated corticosterone effect on the TAT-3 reporter in absence and presence of RU486 is shown in Fig. 7A for the condition in which no SRC was overexpressed. At low concentrations of corticosterone, the RU486 showed its partial agonistic effect. At higher concentrations of the full agonist corticosterone, the antagonism of RU486 became clear, although at 10–7 M corticosterone (equimolar with RU486), the antagonist could no longer effectively compete.
FIG. 7. SRC-1a and SRC-1e overexpression differentially affect the dose-response curves for GRs in absence and presence of 10–7 M RU486 in CV1-cells. A, No-coactivator (no Co-A) overexpression. B, Overexpression of SRC-1a. Due to higher partial agonism of RU486 and lower transactivation by corticosterone, there is hardly any antagonistic effect of RU486. C, Overexpression of SRC-1e. Due to the low partial agonism of RU486 (not different from no-coactivator) and particularly the strong coactivation of corticosterone-activated GRs (note the scale of the y-axis), RU486 has strong antagonistic effects at corticosterone concentrations higher than 1 nM.
The relative agonism/antagonism varies significantly between the conditions in which SRC-1 splice variants have different abundances. When SRC-1a was overexpressed (Fig. 7B), the partial agonistic effect of RU486 was modestly (1.3- to 1.5-fold) increased at low corticosterone concentrations (as in Fig. 6B), whereas the corticosterone-occupied GR was not coactivated by the overexpressed SRC-1a. As a consequence, less antagonism was observed at all corticosterone concentrations: transactivation was never blocked more than 2-fold by the presence of the antagonist. In contrast, when SRC-1e was brought to overexpression, corticosterone-activated GR was potently coactivated, but 10–7 M RU486 had a strong antagonistic effect at higher corticosterone concentrations (>10–9 M) and a weak partial agonistic effect at low corticosterone concentrations (<10–9 M; Fig. 7C).
The relative partial agonistic/antagonistic effect of 10–7 M RU486 in the presence of different concentrations of corticosterone is summarized in Table 2. SRC-1a overexpression leads to stronger partial agonistic effects of RU486 at low corticosterone levels and much weaker antagonist effect at higher levels of corticosterone, when compared with SRC-1e. Thus, variations of SRC-1 splice variant abundance may lead to differences in transactivation on binding of full agonists but also to different efficacy of some antagonists.
TABLE 2. The effect of SRC-1a/1e overexpression on the partial agonism/antagonism of RU486 at increasing concentrations of corticosterone
Differential regulation of SRC-1a/1e mRNA abundance
Our findings suggest that the ratio of SRC-1 splice variants may influence cellular responses to glucocorticoids. Although transcriptional control may not be the most common control mechanism of SRC regulation in differentiated tissues (33), SRC-1 mRNA levels in pituitary have been shown to be subject to clear regulation by estrogens and thyroid hormone (34). We therefore addressed the question whether mRNA of SRC-1 splice variants can be regulated by corticosterone and if so, whether this occurs in a uniform manner or whether there is splice variant-specific regulation. We used in situ hybridization with oligonucleotide probes that specifically detect SRC-1 splice variants (22) (Fig. 8). Chronic exposure of adrenalectomized rats to moderate to high levels of corticosterone (plasma levels of 6.7 ± 1.5 μg/dl at the time of decapitation), compared with adrenalectomized rats (<0.5 μg/dl) led to a 17% down-regulation of SRC-1e in the anterior pituitary but not to changes in SRC-1a mRNA levels (Fig. 8, Table 3). In the hippocampus of the rat brain, in which both MRs and GRs are expressed, we observed no effect on expression of SRC-1a or -1e mRNA (data not shown). These data demonstrate that the ratio of SRC-1 splice variant mRNAs not only differs between cell types but may also be dynamically regulated in vivo in some physiological conditions.
FIG. 8. In vivo differential regulation of SRC-1a (A and B) and SRC-1e (C and D) mRNA abundance in the anterior pituitary in adrenalectomized (Adx) rats implanted with a cholesterol (Chol, A and C) or 100 mg corticosterone (Cort, B and D) pellet. SRC-1 mRNA is expressed in the anterior pituitary (AP) but not in the posterior pituitary (PP). Chronic exposure to moderately elevated levels of corticosterone significantly down-regulated SRC-1e (D) but not SRC-1a mRNA (see also Table 3). On the right-hand side, [14C]microscales are shown, indicating a 2-fold difference in concentration of radioactivity in the lower concentration range, which are relevant for the pituitary sections.
TABLE 3. Expression of SRC-1a and 1e mRNA in the pituitary as a function of circulating corticosterone (Cort) levels
Discussion
We investigated what functional consequences for corticosteroid signaling may follow from the substantial differences in expression that can occur for two splice variants of SRC-1, SRC-1a and -1e. Both splice variants contain AD-1 (CBP/p300 binding), AD-2 [methyl transferase binding (35)] and the weak intrinsic histone (histone 3) acetyl transferase domains (36) and differ only in their C-terminal parts. The relatively small structural difference between SRC-1a and -1e resulted in large differences in coactivation of MRs and GRs in transient transfection studies. There was a pronounced promoter dependence: whereas SRC-1a tended to be somewhat stronger at single GREs, SRC-1e consistently coactivated much more strongly at reporters containing multiple response elements. We observed modest receptor-specific effects: at the TAT-3 Luc reporter the differences between the splice variants tended to be larger for the MR than for the GR. The AF-1 of the MR but not the GR could be coactivated by SRC-1e in absence of the receptor LBD via an as-yet-unidentified SRC-input domain likely localized between amino acids 988 and 1240. We observed a ligand-specific effect for the GR: RU 486-activated GR was coactivated by SRC-1a and -1e in an opposite way, compared with corticosterone. This was reflected in a pronounced difference in the efficacy of RU486 as an antagonist, depending on the relative abundance of the SRC-1 splice variants. Finally, we demonstrated that mRNA abundances of SRC-1a and -1e are differentially regulated by corticosterone in the pituitary.
The CV-1 cells in which we performed most of our experiments have their pros and cons. MR AF-1 is inherently weak in CV-1 cells, which may affect our results. On the other hand, they lack endogenous MRs and GRs and therefore allow the corticosteroid receptor status to be defined. Transactivation via the endogenous GR of A549 cells showed effects of SRC-1a and -1e that were very similar to those observed in CV-1 cells. Whereas the choice of these cells is to some extent arbitrary, we can conclude that the specific interactions are not the consequence of overexpression of the receptor protein, although we did not test cell lines expressing endogenous MRs. CV-1 cells do express endogenous p160 coactivators (14), but for lack of slice variant-specific antibodies, we do not know the relative abundance of endogenous SRC-1a and -1e protein. Because the observed effects in these cells are all in the context of overexpressed coactivators, the in vivo importance of the outcome of the differential interactions remains to be determined. However, differential interaction with nuclear receptors between SRC-1 and SRC-2 have been shown to have large functional consequences for fat tissue (37), despite a degree of functional redundancy between these coactivators (38). Similarly, the differences between SRC-1 splice variants may well bear functional relevance because pronounced differences in SRC-1a and -1e mRNA expression occurs among different glucocorticoid target cells in vivo, in some instances in an all-or-nothing fashion (23). In addition, the differential down-regulation of mRNA we observed in the pituitary (Fig. 8) suggests that the ratio between the splice variants may even be physiologically regulated in certain cell types.
One of the most striking differences we observed between the SRC-1 splice variants was the specific strong coactivation by SRC-1e at multiple GRE-containing regulatory regions in the context of different minimal promoters. It is important to note that this dichotomy did not apply to the often-used MMTV-Luc reporter, which was coactivated by both SRC-1a and -1e (data not shown), but was found for two 5' regulatory regions involved in regulation of endogenous genes, tyrosine aminotransferase and PNMT. Recent data show that multiple GRE-containing promoters are also highly interesting because DNA binding of GRs for such promoters does not require the dimerization interface localized in the second Zn finger of the GR DNA-binding domain (DBD) (25, 26). We consider the MMTV promoter as atypical, in that it does contain several GR binding sites, but in the context of the DBD dimer mutants does not behave as containing multiple GREs (26, 39). Thus, it is difficult to interpret in terms of the single vs. multiple GRE phenotype that we observed for the other promoters that were tested.
The promoter-specific effects seem to be unrelated to recruitment of the coactivators by the receptors but rather to differences in output of SRC-1a and SRC-1e, once bound to the single or multiple receptor dimers. The preferential coactivation of synergistic transactivation may point to a role of small ubiquitin-like modifier (SUMO)-ylation sites, which have been implicated in synergistic activation via NRs (4), particularly in the MR (40), and which are also present in p160 coactivators (41).
Differential effects of SRC-1a and -1e overexpression was observed in the context of induction of the endogenous PNMT mRNA in neuroscreen cells. Because PNMT is a direct target for GRs, this likely reflects differential action on an endogenous promoter. The decreased rather than increased expression that was observed for both splice variants relative to the no-coactivator condition may be caused by competition between overexpressed SRC-1 and other coactivators that may be preferred by this promoter. However, as hypothesized based on analysis of the GR-responsive fragment of this gene (Fig. 2B), SRC-1e led to higher PNMT mRNA levels than SRC-1a. Few other endogenous promoters that function as multiple GRE containing are known. Identification of the different types of GRE constellations of corticosteroid-responsive promoters is a relevant challenge for future studies, i.e. to predict those genes for which differential coactivator expression as studied here can bear in vivo relevance.
In most reporter studies, MRs typically stimulate transcription substantially less than GRs at synergizing promoters (3, 42). At the TAT-3 promoter, SRC-1e seemed to coactivate MRs stronger than GRs. This is probably not due to saturation of transactivation for GRs because the hybrid receptor MMG (5) was strongly coactivated despite significantly higher basal transcription levels than GRs (data not shown). The only other experimental setting we are aware of that causes such a pronounced MR-mediated transcriptional synergy is the phenotype observed after the above-mentioned disruption of the DBD dimer interface of the receptor (25). However, when we tested such dimer mutants, they were coactivated in a manner similar to that of wild-type receptors, suggesting distinct mechanisms (data not shown). We would argue that the high transactivation via MRs we observed after overexpression of an endogenous factor may be indicative of cellular settings in which MRs can act, at least at some promoters, more potently than often assumed (irrespective of the agonist that is used).
The specific coactivation of MR AF-1 may contribute to the somewhat larger stimulatory effect of SRC-1e on MRs, when compared with GRs on the TAT-3 reporter. It has been shown earlier to be potentially coactivated via glucocorticoid receptor-interacting protein/SRC-2 (13) and, in a ligand-dependent way, a complex containing CBP and RNA Helicase A (43). Our data also suggest that SRC-1e can interact, directly or indirectly, with the MR N terminus, leading to coactivation of MR AF-1. Our SRC-1-VP16 hybrid data suggest that this interaction depends on SRC-1 amino acids 988-1240 but not on the Q-rich domain (residues 1053–1123) that has been shown to mediate the interaction between SRCs and the AR N terminus (12) because deletion of this domain did not lead to loss of coactivation. The lack of effect of mutation of the LxxLL motifs in SRC-1 is consistent with SRC-1988–1240 as the dominant interaction surface between SRC-1 and MR-N terminus (Fig. 5). The MR AF-1 coactivation did not occur when the SRC-1e AD-1 was deleted, which suggests that recruitment of CBP/p300 mediates the increased transcription activation that was observed.
Ligand-dependent coactivation by p160 SRCs was earlier shown for the vitamin D receptor, which can engage in preferential interaction in SRC-2/glucocorticoid interacting protein-1 (44) but to our knowledge not for splice variants. We saw no differences in coactivation between aldosterone- and corticosterone-activated MRs, despite the fact that these ligands induce different conformations and molecular interactions of the receptor (43, 45). We tested the effect of the SRC-1 splice variants on antagonists action because the ratio of coactivators and corepressors is thought to determine the extent of antagonism/partial agonism of this antagonist (46, 47, 48). Coactivation of RU486-activated GRs by SRC-1a and -1e was reciprocal to that of corticosterone-activated GRs. The potential functional relevance of this finding is underscored by the different dose-response curves for corticosterone in the absence and presence of 10–7 M RU486: in cells that mainly express SRC-1a [as are present in hypothalamus of the rat brain (23) or to a lesser extent in the pituitary after exposure to elevated concentrations of corticosterone (present study and Fig. 8)], certain promoters may be less potently transactivated, but RU486 treatment is expected to have relatively little effect in such a setting. This is all the more relevant because RU486 has recently been shown to have powerful antidepressant properties in a group of psychotic depressed patients (49). Regional differences for RU486 action may be part of the mechanism by which this drug acts in a clinical setting.
The observed differential expression of SRC-1a and -1e may have implications for cell-specific effects of corticosteroid and other NRs. Also, the functional differences that we observed in cell lines may also be relevant in the context of differential regulation of SRC-1 splice variant activity by posttranslation modifications. Further relevance of the functional differences between SRC-1 splice variants also depends on whether activity or expression of the splice variants is differentially regulated, i.e. whether the ratio between the variants may change. In pituitary, strong hormonal regulation of SRC-1 mRNA has been reported after estrogen and thyroid hormone treatment (34), but no distinction between splice variants has been made. Our data from the anterior pituitary show that differential regulation of the splice variants can take place. The magnitude of down-regulation of SRC-1e mRNA after modest elevations of circulating plasma corticosterone was 17%. This is similar to what we have observed for total SRC-1 mRNA in estradiol-treated ovariectomized female rats (conform Ref. 27 ; our unpublished data). Whether the splice variant-specific regulation reflects effects at the level of transcription, splicing, or mRNA stability is at present not clear. The data do suggest that the ratio between SRC-1a and -1e may not be a constant in some tissues, which may affect the corticosteroid responsiveness of such cells in a promoter-specific fashion. However, it is not a general phenomenon because we did not observe regulation of SRC-1 mRNA in hippocampus after manipulation of corticosteroid levels. We do not know in which cell type(s) of the pituitary this effect takes place; if in all cell types, then the effect would qualify as modest. We also lack important information at the protein level because there are no splice variant-specific antibodies.
In summary, we have observed a number of functional differences between SRC-1a and SRC-1e in their interactions with GRs and MRs, which involve and may in part explain promoter-, receptor-, and ligand-specific effects. We have also shown that the mRNA abundance of SRC-1a and -1e can be differentially regulated in vivo. It is obvious that other promoter elements and cellular proteins constitute a superimposed context that determines the extent to which the SRC-1a/1e differences become manifest. Further studies should address the mechanisms that cause the functional differences we have seen, e.g. dependence on histone acetylation activity, recruitment of cointegrators CBP/p300, histone methyl transferases, or other factors that finally lead to increased transactivation through modification of histone structure and/or recruitment of transcription machinery (19). A major challenge for future studies is to address the relevance of these effects for the transcriptional regulation of endogenous promoters in relevant target tissues in vivo. This should involve specific manipulation of expression of the splice variants by knockdown methods and the identification of relevant GRE-containing regulatory regions in the genome.
Acknowledgments
We thank Drs. C. Bevan, G. Jenster, and K. Yamamoto for kind donation of the expression plasmids for SRC-1 mutants, PRE-tk-Luc reporters, and eTAT-Luc reporter, respectively. We thank Dr. Erno Vreugdenhil for his scientific and technical advice. Diana Rien, Servane Lachize, and Anette Verhoeven are kindly acknowledged for technical assistance.
References
Hollenberg SM, Weinberger C, Ong ES, Cerelli G, Oro A, Lebo R, Thompson EB, Rosenfeld MG, Evans RM 1985 Primary structure and expression of a functional glucocorticoid receptor cDNA. Nature 318:635–641
Arriza JL, Weinberger C, Cerelli G, Glaser TM, Handelin BL, Housmann DE, Evans RM 1987 Cloning of human mineralocorticoid receptor complementary DNA: structural and functional kinship with the glucocorticoid receptor. Science 237:268–275
Rupprecht R, Arriza JL, Spengler D, Reul JMHM, Evans RM, Holsboer F, Damm K 1993 Transactivation and synergistic properties of the mineralocorticoid receptor: relationship to the glucocorticoid receptor. Mol Endocrinol 7:597–603
Iniguez-Lluhi JA, Pearce D 2000 A common motif within the negative regulatory regions of multiple factors inhibits their transcriptional synergy. Mol Cell Biol 20:6040–6050
Pearce D, Yamamoto KR 1993 Mineralocorticoid and glucocorticoid receptor activities distinguished by nonreceptor factors at a composite response element. Science 259:1161–1165
Meijer OC, Williamson A, Dallman MF, Pearce D 2000 Transcriptional repression of the 5-HT1A receptor promoter by corticosterone via mineralocorticoid receptors depends on the cellular context. J Neuroendocrinol 12:245–254
De Kloet ER, Vreugdenhil E, Oitzl MS, Joels M 1998 Brain corticosteroid receptor balance in health and disease. Endocr Rev 19:269–301
Makino S, Gold PW, Schulkin J 1994 Effects of corticosterone on CRH mRNA and content in the bed nucleus of the stria terminalis; comparison with the effects in the central nucleus of the amygdala and the paraventricular nucleus of the hypothalamus. Brain Res 657:141–149
McKenna NJ, Lanz RB, O’Malley BW 1999 Nuclear receptor coregulators: cellular and molecular biology. Endocr Rev 20:321–344
Rosenfeld MG, Glass CK 2001 Coregulator codes of transcriptional regulation by nuclear receptors. J Biol Chem 276:36865–36868[Free Full Text]
Heery DM, Kalkhoven E, Hoare S, Parker MG 1997 A signature motif in transcriptional co-activators mediates binding to nuclear receptors. Nature 387:733–736
Bevan CL, Hoare S, Claessens F, Heery DM, Parker MG 1999 The AF1 and AF2 domains of the androgen receptor interact with distinct regions of SRC1. Mol Cell Biol 19:8383–8392
Fuse H, Kitagawa H, Kato S 2000 Characterization of transactivational property and coactivator mediation of rat mineralocorticoid receptor activation function-1 (AF-1). Mol Endocrinol 14:889–899
Torchia J, Rose DW, Inostroza J, Kamei Y, Westin S, Glass CK, Rosenfeld MG 1997 The transcriptional co-activator p/CIP binds CBP and mediates nuclear-receptor function. Nature 387:677–684
Chen D, Ma H, Hong H, Koh SS, Huang SM, Schurter BT, Aswad DW, Stallcup MR 1999 Regulation of transcription by a protein methyltransferase. Science 284:2174–2177
Hong H, Kohli K, Trivedi A, Johnson DL, Stallcup MR 1996 GRIP1, a novel mouse protein that serves as a transcriptional coactivator in yeast for the hormone binding domains of steroid receptors. Proc Natl Acad Sci USA 93:4948–4952
Xu J, Liao L, Ning G, Yoshida-Komiya H, Deng C, O’Malley BW 2000 The steroid receptor coactivator SRC-3 (p/CIP/RAC3/AIB1/ACTR/TRAM-1) is required for normal growth, puberty, female reproductive function, and mammary gland development. Proc Natl Acad Sci USA 97:6379–6384
Liu Z, Wong J, Tsai SY, Tsai MJ, O’Malley BW 2001 Sequential recruitment of steroid receptor coactivator-1 (SRC-1) and p300 enhances progesterone receptor-dependent initiation and reinitiation of transcription from chromatin. Proc Natl Acad Sci USA 98:12426–12431
Metivier R, Penot G, Hubner MR, Reid G, Brand H, Kos M, Gannon F 2003 Estrogen receptor- directs ordered, cyclical, and combinatorial recruitment of cofactors on a natural target promoter. Cell 115:751–763
Kamei Y, Xu L, Heinzel T, Torchia J, Kurokawa R, Gloss B, Lin SC, Heyman RA, Rose DW, Glass CK, Rosenfeld MG 1996 A CBP integrator complex mediates transcriptional activation and AP-1 inhibition by nuclear receptors. Cell 85:403–414
Ding XF, Anderson CM, Ma H, Hong H, Uht RM, Kushner PJ, Stallcup MR 1998 Nuclear receptor-binding sites of coactivators glucocorticoid interacting protein 1 (GRIP1) and steroid receptor coactivator 1 (SRC-1): multiple motifs with different binding specificities. Mol Endocrinol 12:302–313
Kalkhoven E, Valentine JE, Heery DM, Parker MG 1998 Isoforms of steroid receptor co-activator 1 differ in their ability to potentiate transcription by the oestrogen receptor. EMBO J 17:232–243
Meijer OC, Steenbergen PJ, De Kloet ER 2000 Differential expression and regional distribution of steroid receptor coactivators SRC-1 and SRC-2 in brain and pituitary. Endocrinology 141:2192–2199
Miesfeld R, Godowski PJ, Maler BA, Yamamoto KR 1987 Glucocorticoid receptor mutants that define a small region sufficient for enhancer activation. Science 236:423–427
Liu W, Wang J, Yu GQ, Pearce D 1996 Steroid receptor transcriptional synergy is potentiated by disruption of the DNA binding domain dimer interface. Mol Endocrinol 10:1399–1406
Adams M, Meijer OC, Wang J, Bhargava A, Pearce D 2003 Homodimerization of the glucocorticoid receptor is not essential for response element binding: activation of the phenylethanolamine N-methyltransferase gene by dimerization-defective mutants. Mol Endocrinol 17:2583–2592
Nordeen SK 1988 Luciferase reporter gene vectors for analysis of promoters and enhancers. Biotechniques 6:454–458
Cavailles V, Dauvois S, L’Horset F, Lopez G, Hoare S, Kushner PJ, Parker MG 1995 Nuclear factor RIP140 modulates transcriptional activation by the estrogen receptor. EMBO J 14:3741–3751
Kalkhoven E, Teunissen H, Houweling A, Verrijzer CP, Zantema A 2002 The PHD type zinc finger is an integral part of the CBP acetyltransferase domain. Mol Cell Biol 22:1961–1970
Strahle U, Schmid W, Schutz G 1988 Synergistic action of the glucocorticoid receptor with transcription factors. EMBO J 7:3389–3395
Schmid W, Strahle U, Schutz G, Schmitt J, Stunnenberg H 1989 Glucocorticoid receptor binds cooperatively to adjacent recognition sites. EMBO J 8:2257–2263
Ebert SN, Lindley SE, Bengoechea TG, Bain D, Wong DL 1997 Adrenergic differentiation potential in PC12 cells: influence of sodium butyrate and dexamethasone. Brain Res Mol Brain Res 47:24–30
Smith CL, O’Malley BW 2004 Coregulator function: a key to understanding tissue specificity of selective receptor modulators. Endocr Rev 25:45–71
Misiti S, Schomburg L, Yen PM, Chin WW 1998 Expression and hormonal regulation of coactivator and corepressor genes. Endocrinology 139:2493–2500
Koh SS, Chen D, Lee YH, Stallcup MR 2001 Synergistic enhancement of nuclear receptor function by p160 coactivators and two coactivators with protein methyltransferase activities. J Biol Chem 276:1089–1098
Spencer TE, Jenster G, Burcin MM, Allis CD, Zhou J, Mizzen CA, McKenna NJ, Onate SA, Tsai SY, Tsai MJ, O’Malley BW 1997 Steroid receptor coactivator-1 is a histone acetyltransferase. Nature 389:194–198
Picard F, Gehin M, Annicotte JS, Rocchi S, Champy MF, O’Malley BW, Chambon P, Auwerx J 2002 SRC-1 and TIF2 control energy balance between white and brown adipose tissues. Cell 111:931–941
Xu J, Qiu Y, DeMayo FJ, Tsai SY, Tsai MJ, O’Malley BW 1998 Partial hormone resistance in mice with disruption of the steroid receptor coactivator-1 (SRC-1) gene. Science 279:1922–1925
Reichardt HM, Kaestner KH, Tuckermann J, Kretz O, Wessely O, Bock R, Gass P, Schmid W, Herrlich P, Angel P, Schutz G 1998 DNA binding of the glucocorticoid receptor is not essential for survival. Cell 93:531–541
Tallec LP, Kirsh O, Lecomte MC, Viengchareun S, Zennaro MC, Dejean A, Lombes M 2003 Protein inhibitor of activated signal transducer and activator of transcription 1 interacts with the N-terminal domain of mineralocorticoid receptor and represses its transcriptional activity: implication of small ubiquitin-related modifier 1 modification. Mol Endocrinol 17:2529–2542
Kotaja N, Karvonen U, Janne OA, Palvimo JJ 2002 The nuclear receptor interaction domain of GRIP1 is modulated by covalent attachment of SUMO-1. J Biol Chem 277:30283–30288
Arriza JL, Simerly RB, Swanson LW, Evans RM 1988 The neuronal mineralocorticoid receptor as a mediator of glucocorticoid response. Neuron 1:887–900
Kitagawa H, Yanagisawa J, Fuse H, Ogawa S, Yogiashi Y, Okuno A, Nagasawa H, Nakajima T, Matsumoto T, Kato S 2002 Ligand-selective potentiation of rat mineralocorticoid receptor activation function 1 by a CBP-containing histone acetyltransferase complex. Mol Cell Biol 22:3698–3706
Takeyama K, Masuhiro Y, Fuse H, Endoh H, Murayama A, Kitanaka S, Suzawa M, Yanagisawa J, Kato S 1999 Selective interaction of vitamin D receptor with transcriptional coactivators by a vitamin D analog. Mol Cell Biol 19:1049–1055
Rogerson FM, Brennan FE, Fuller PJ 2004 Mineralocorticoid receptor binding, structure and function. Mol Cell Endocrinol 217:203–212
Zhang X, Jeyakumar M, Petukhov S, Bagchi MK 1998 A nuclear receptor corepressor modulates transcriptional activity of antagonist-occupied steroid hormone receptor. Mol Endocrinol 12:513–524
Schulz M, Eggert M, Baniahmad A, Dostert A, Heinzel T, Renkawitz R 2002 RU486 induced glucocorticoid receptor agonism is controlled by the receptor N terminus and by corepressor binding. J Biol Chem 277:26238–26243
Liu Z, Auboeuf D, Wong J, Chen JD, Tsai SY, Tsai MJ, O’Malley BW 2002 Coactivator/corepressor ratios modulate PR-mediated transcription by the selective receptor modulator RU486. Proc Natl Acad Sci USA 99:7940–7944
Belanoff J, Rothschild A, Cassidy F, DeBattista C, Baulieu E, Schold C, Schatzberg A 2002 An open label trial of C-1073 (mifepristone) for psychotic major depression. Biol Psychiatry 52:386–392(O. C. Meijer, E. Kalkhove)