Identification of a Glucocorticoid Response Element in the Rat β2-Adrenergic Receptor Gene
- 1Division of Critical and Pulmonary Care Medicine, Department of Medicine (L.E.C., F.C.H., S.E.J., D.W.M.), and 2Department of Physiology and Biophysics (L.E.C., W.C.), University of Arkansas for Medical Sciences, Little Rock, Arkansas 72205
Abstract
Regulation of β2-adrenergic receptor (β2AR) levels by glucocorticoids is a physiologically important mechanism for altering β2AR responsiveness. Glucocorticoids increase β2AR density by increasing the rate of β2AR gene transcription, but the cis-elements involved have not been well characterized. We now show that one of six potential glucocorticoid response elements (GREs) in the 5′-flanking region of the rat β2AR gene is necessary for glucocorticoid-dependent stimulation of receptor gene expression. Using a nested set of deletion fragments of the rat β2AR gene 5′-flanking region fused to a luciferase reporter gene, glucocorticoid-dependent induction of reporter gene expression in HepG2 cells was localized to a region between positions −643 and −152, relative to the transcription initiation site. In electrophoretic mobility shift assays, a double-stranded oligonucleotide incorporating a near-consensus GRE from this region (positions −379 to −365) formed complexes with the human recombinant glucocorticoid receptor, as well as with nuclear protein from dexamethasone-treated HepG2 cells. Mutation of a single base within this GRE sequence greatly diminished interaction of the mutated oligonucleotide with the human recombinant glucocorticoid receptor. The functional activity of the GRE was characterized using a luciferase reporter construct driven by a minimal thymidine kinase promoter. In HepG2 cells transfected with constructs containing the GRE, dexamethasone increased reporter gene expression approximately 3-fold, whereas a dexamethasone effect was not observed with constructs lacking the GRE. Taken together, these findings show that a GRE located at positions −379 to −365 in the 5′-flanking region of the rat β2AR gene mediates glucocorticoid stimulation of β2AR gene transcription.
The β2AR is a member of a large superfamily of membrane-associated receptors that are coupled to G proteins and produce their effects by activating intracellular signal transduction pathways (Strader et al., 1995). For many G protein-coupled receptors, modulation of receptor number is an established mechanism controlling responsiveness to hormones and neurotransmitters. Heterologous regulation of β2AR levels by glucocorticoids is a physiologically important example of such control (Collins et al., 1988). Numerous in vitro andin vivo studies have shown that β2AR levels and β-agonist-stimulated adenylyl cyclase activity are increased by glucocorticoids (Cheng et al., 1980; Norriset al., 1987; Collins et al., 1988; Takahashi and Iizuka, 1991; Dangel et al., 1996). The increase in β2AR number results from an increase in the rate of synthesis of new receptors (Norris et al., 1987), which in turn is preceded by increased steady state levels of β2AR mRNA (Collins et al., 1988;Hadcock and Malbon, 1988; Zhong and Minneman, 1993; Mak et al., 1995; Dangel et al., 1996) and increased rates of transcription of the β2AR gene (Collinset al., 1988; Mak et al., 1995). Taken together, these findings suggest that enhanced β2AR gene transcription is a principal mechanism underlying glucocorticoid-mediated increases in β2AR levels.
GREs mediate transcriptional activation of numerous eukaryotic genes by glucocorticoid receptors (Beato et al., 1989). Comparison of these response elements has revealed a consensus, 15-nucleotide, nearly palindromic sequence that binds the glucocorticoid receptor (Nordeenet al., 1990). The β2AR genes from several mammalian species contain GRE-like sequences in both coding and noncoding regions (Emorine et al., 1987; Kobilka et al., 1987; Buckland et al., 1990; Jiang and Kunos, 1995; McGraw et al., 1996). Although there is indirect evidence to suggest that regions in the 5′-noncoding region of the β2AR gene are involved (Malbon and Hadcock, 1988), the specific genetic elements responsible for the functional effect of glucocorticoids on β2AR gene transcription have yet to be identified.
Recently, we extended the sequence of 5′-flanking DNA of the rat β2AR gene to approximately 3.7 kilobases upstream from the start of translation (McGraw et al., 1996). Sequence analysis indicated that this segment of the gene contains six putative GREs that could potentially mediate the effect of glucocorticoids on β2AR gene expression. Using a combination of transient transfection assays with β2AR-luciferase fusion genes and EMSAs, we now show that a GRE-like element located at positions −379 to −365 (relative to the transcription start site) confers glucocorticoid inducibility to the rat β2AR gene. Our data also suggest that, within the context of the β2AR promoter, the activity of this GRE may be regulated by additional unidentified genetic elements.
Experimental Procedures
Materials.
Dexamethasone was purchased from Sigma Chemical (St. Louis, MO) and was dissolved in ethanol to a stock concentration of 200 μm. The restriction enzymes AvrII andNheI were purchased from Boehringer Mannheim (Indianapolis, IN). All other restriction enzymes were purchased from Promega Corp. (Madison, WI). Plasmids pRShGRα and N-600 prATLUC were kindly provided by Robert McGehee (Arkansas Children’s Hospital, Little Rock, AR).
Cell culture.
HepG2 cells were grown in Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum (Gibco BRL, Grand Island, NY) that had been depleted of steroid hormones. Serum was stripped of steroids by the addition of 1 g of activated charcoal (Sigma)/100 ml of serum and incubation at 55° for 1 hr. The charcoal-treated serum was centrifuged at 23,000 × g for 10 min at 4°. The supernatant was filter-sterilized and stored at −20° before use.
Plasmid construction.
We have extended the known sequence of the rat β2AR gene to position −3491 (McGrawet al., 1996). The nucleotide numbering system that we used in this study is based on the assignment of +1 to the transcription start site (nucleotide −220 relative to the start of translation, as previously determined) (Jiang et al., 1996). Chimeric gene constructs were made by fusing β2AR gene fragments with pGL3-Basic (Promega), a promoterless luciferase expression vector. Initially, pβ2AR(−3129/+126), pβ2AR(−1115/+126), and pβ2AR(−152/+126) were prepared by subcloning restriction fragments from the 5′-flanking region of the rat β2AR gene into the available sites of pGL3-Basic, as previously reported (McGraw et al., 1996). pβ2AR(−2552/+126), pβ2AR(−643/+126), and pβ2AR(−62/+126) were prepared by linearization of β2AR(−3129/+126) with MluI andNheI and unidirectional digestion of the linearized plasmid with exonuclease III/mung bean nuclease (Stratagene, La Jolla, CA). Short segments of β2AR DNA containing putative GREs were subcloned into pT81LUC (Nordeen, 1988), a luciferase expression vector driven by a minimal TK promoter, to test their ability to enhance expression of a heterologous promoter. For these constructs, double-stranded oligonucleotides containing either GRE1 or GRE5 and the polymerase chain reaction-generated β2AR gene fragment containing GRE2, GRE3, and GRE4 were cloned into the XmaI site of pT81LUC. The sequence and orientation of all constructs were confirmed by dideoxy sequencing (Sanger et al., 1977) using Sequenase version II (United States Biochemical Corp., Cleveland, OH). All plasmids used in transfections were prepared using Bigger Prep plasmid preparation kits (5 Prime-3 Prime, Boulder, CO).
Mutagenesis.
GRE5 was mutated in plasmid pβ2AR(−3129/+126) using the oligonucleotide gaaagaataagctcacccggacacgc and the GeneEditor in vitro site-directed mutagenesis system (Promega). The resulting mutant, pβ2ARm1(−3129/+126), replaced the guanine at position +6 of GRE5 with an adenine. The mutation was confirmed by dideoxy sequencing (Sangeret al., 1977). The identity between the remainder of pβ2AR(−3129/+126) and pβ2ARm1(−3129/+126) was confirmed by restriction site analysis.
Polymerase chain reaction.
The relatively short segment (−831 to −708) in the β2AR gene containing three putative GREs was amplified from rat genomic DNA using a Perkin-Elmer model 480 thermal cycler (Norwalk, CT) and β2AR gene-specific primers. The primers were 5′-catatacccgggcgaagttactgccttggtgcggttg-3′ (sense) and 5′-catatacccgggggcaagaacacaggaggtgactc-3′ (antisense). Each primer was synthesized with a XmaI site to facilitate cloning of the amplified product into pT81LUC. The reaction mixture included 0.5 μg of rat genomic DNA, 0.5 pmol of each primer, 0.2 mmdeoxynucleoside triphosphates, 1.5 mmMgCl2, and 2.5 units of Thermus aquaticus DNA polymerase (Promega). A 5-min hot start at 94° was used, followed by 30 cycles of 94° for 1 min and 66° for 1 min and then a final single cycle of 72° for 7 min.
Transient transfections.
HepG2 cells were transfected by calcium phosphate coprecipitation in 60-mm dishes, as described previously (Chen and Okayama, 1987). For experiments in which dexamethasone-stimulated promoter activity in different 5′-deletion constructs was tested, subconfluent cells in Dulbecco’s modified Eagle medium with 10% fetal bovine serum that had been stripped of steroids were transfected with 0.38 pmol of the β2AR-luciferase fusion genes, 2 μg of pRSVβgal, 1 μg of pRShGRα, and pGEM-7Zf(−) to adjust the amount of total DNA in each dish to 8.33 μg. HepG2 cells were co-transfected with pRShGRα, a human glucocorticoid receptor expression vector (Brasier et al., 1990), because replicating cells can be deficient in trans-acting factors such as the glucocorticoid receptor (Johnson, 1990). Cells were exposed to precipitates overnight and the medium was changed. Either vehicle or dexamethasone (final concentration, 0.1 μm) was added 24 hr after transfection, and cells were harvested after an 8-hr incubation. After harvesting, cell lysates were prepared and assayed for luciferase using the Promega luciferase assay system or for β-galactosidase using the Galacto-Light system (TROPIX, Bedford, MA), with a Monolight model 2010 luminometer (Analytical Luminescence Laboratory, Ann Arbor, MI). For experiments in which the ability of putative β2AR GREs to enhance the activity of a heterologous promoter in the presence of dexamethasone was assessed, the dual-luciferase reporter assay system (Promega) was used as described above, except that pRSVβgal was omitted from the transfection mixture. In its place, pRL-SV40 encodingRenilla luciferase was used as a measure of transfection efficiency.
Human recombinant glucocorticoid receptor and nuclear extract preparation.
Human recombinant glucocorticoid receptor was obtained from Affinity Bioreagents (Golden, CO). Nuclear extracts were prepared essentially as previously described (Andrews and Faller, 1991). Except where otherwise noted, centrifugations were performed in an Eppendorf model 5415C microfuge at maximum speed, at room temperature. Approximately 106 to 107 HepG2 cells that had been transiently transfected with pRShGRα were scraped into 1.5 ml of cold phosphate-buffered saline, pH 7.4, and pelleted. Cells were resuspended in 400 μl of buffer A (10 mm HEPES-KOH, pH 7.9, 1.5 mm MgCl2, 10 mm KCl, 0.5 mm dithiothreitol, 0.2 mm phenylmethylsulfonyl fluoride) at 4°. Cells were allowed to swell for 10 min, they were vortex-mixed for 10 sec and then centrifuged for 10 sec, and the supernatant was discarded. The pellet was resuspended in 20–50 μl of buffer C (20 mm HEPES-KOH, pH 7.9, 25% glycerol, 420 mm NaCl, 1.5 mmMgCl2, 0.2 mm EDTA, 0.5 mm dithiothreitol, 0.2 mm phenylmethylsulfonyl fluoride) at 4° and incubated for 20 min. Cellular debris was removed by centrifugation for 2 min at 4°, and the supernatant containing DNA-binding proteins was stored at −70°. Nuclear extract protein concentrations were determined (Bradford, 1976) using bovine serum albumin as the standard.
EMSAs.
Oligonucleotides were either commercially synthesized (Bio-Synthesis, Inc., Lewisville, TX) or, in the case of oligonucleotides containing the human tyrosine aminotransferase gene GRE, were obtained from Affinity Bioreagents. The sequences of the sense oligonucleotides only are shown in Table1. Previously, glucocorticoid receptor bound to a functional GRE in MMTV was shown to occupy no more than 30 bp of DNA (Nordeen et al., 1990). Therefore, the oligonucleotides that we used in EMSAs included 35 nucleotides of β2AR gene sequence plus an additional five nucleotides comprising a restriction site for subcloning of the fragment. Complementary oligonucleotides in equimolar amounts were heated to 100°, cooled overnight to 25°, divided into aliquots, and stored at −20° before use. Double-stranded oligonucleotide probes were end-labeled using T4 polynucleotide kinase and [γ-32P]ATP. Binding reactions were performed in a 20-μl volume containing approximately 20,000 cpm of labeled probe, 6–12 μg of nuclear extract or human recombinant glucocorticoid receptor (Affinity Bioreagents), 20 mmHEPES, pH 7.9, 60 mm KCl, 5 mmMgCl2, 2 mm dithiothreitol, 10% glycerol, 200 ng of poly(dI·dC), 1 μg of bovine serum albumin, and unlabeled competitor oligonucleotides. The instructions provided by the manufacturer were followed when the human recombinant glucocorticoid receptor was used in EMSAs. The binding reaction mixtures were incubated at 25° for 30 min and then loaded onto 6% nondenaturing polyacrylamide gels in 25 mm Tris, pH 8.3, 25 mm boric acid, 0.5 mm EDTA. For supershift experiments, incubations with polyclonal anti-human glucocorticoid receptor antibody (Affinity Bioreagents) or preimmune antiserum (diluted 1:500) were performed for 30 min at room temperature after addition of radiolabeled GRE5 and nuclear extract. Gels were dried and autoradiographed overnight at −70°, using Fujifilm RX-Fuji medical X-ray film and intensifying screens.
Oligonucleotides used in EMSAs and luciferase assays with pT81LUC
Results
The β2AR gene.
In an earlier study, in addition to correcting an error in the previously reported sequence of the rat β2AR gene (Buckland et al., 1990), we cloned an additional 1400 bp of 5′-flanking DNA (McGraw et al., 1996). Analysis of the known sequence of the rat β2AR gene yielded seven potential glucocorticoid regulatory elements (Fig.1A). Six of the potential GREs are located upstream from the receptor open reading frame, whereas the seventh GRE is located in the 3′-flanking region of the gene. Sequence comparisons were made between the putative GREs in the β2AR gene and the consensus sequence for previously demonstrated, positively modulated GREs. This consensus GRE sequence arose from analysis of GRE-like elements in the following genes: MMTV, Moloney murine sarcoma provirus, metallothionein IIa, lysozyme, vitellogenin, growth hormone, uteroglobin, tyrosine aminotransferase, tryptophan oxygenase, and acidic glycoprotein (Nordeen et al., 1990) (Fig. 1B). In preliminary studies, the putative GRE downstream from the receptor open reading frame was found to be nonfunctional (data not shown); therefore, we focused our attention on the six GRE-like elements in the 5′-flanking region of the gene. We have numbered these GRE sequences 1 through 6, with GRE6 being the most proximal.
Schematic diagram of the rat β2AR gene and location of putative GREs. A, The coding region is indicated by a solid rectangle. The first nucleotide in the ATG codon that codes for the initiator methionine is designated +1. The approximate locations of putative GREs are indicated by open ellipses and are numbered 1 through 7. B, For each putative GRE, the exact location, sequence, and identity with the consensus GRE sequence (Nordeen et al., 1990) are shown.Underlined nucleotides, identity with the consensus GRE.
Transient transfections using β2AR promoter truncations.
To first determine which regions of the β2AR 5′-flanking region are necessary for glucocorticoid-mediated transcriptional activation, a series of six 5′-deletion fragments were fused to a luciferase reporter gene. Among these fragments, pβ2AR(−3129/+126) and pβ2AR(−2552/+126) contain all six putative GREs, pβ2AR(−1115/+126) contains the proximal five putative GREs, pβ2AR(−643/+126) contains GRE5 and GRE6, and pβ2AR(−152/+126) and pβ2AR(−62/+126) contain only the most proximal GRE. To test the effect of dexamethasone on the expression of the β2AR-luciferase fusion genes, luciferase activity was determined in transfected HepG2 cells after incubation with either vehicle or 0.1 μm dexamethasone for 8 hr. Results from preliminary experiments indicated that 8 hr was the optimal time to observe dexamethasone responsiveness, because cell viability decreased with longer exposures to dexamethasone (data not shown). Fig. 2 depicts the results of experiments in which progressively truncated β2AR-luciferase fusion genes transiently transfected into HepG2 cells were tested for dexamethasone responsiveness. Approximately 2-fold induction with dexamethasone (compared with levels in the absence of added glucocorticoid) was observed with pβ2AR(−3129/+126), pβ2AR(−2552/+126), and pβ2AR(−1115/+126). This level of induction of luciferase activity is similar to the 2–4-fold induction in β2AR levels that has been observed in the lung after injection of rats with dexamethasone (McGraw et al., 1995) or after addition of glucocorticoids to cultured cells (Collinset al., 1988; Takahashi and Iizuka, 1991; Dangel et al., 1996). With pβ2AR(−643/+126), the dexamethasone induction was approximately 1.3-fold. In contrast, no dexamethasone effect was observed with pβ2AR(−152/+126) or pβ2AR(−62/+126). As a positive control, we used N-600 prATLUC, a fusion gene containing a segment of the rat angiotensinogen gene with two functional GREs coupled to a luciferase-encoding gene. In the presence of dexamethasone, expression of N-600 prATLUC was increased approximately 8–10-fold, consistent with the level of glucocorticoid induction previously demonstrated with this fusion gene in HepG2 cells (Brasier et al., 1989). These results indicated that the region between positions −643 and −152 was necessary for dexamethasone induction of β2AR gene expression. This region contains a single GRE (GRE5), which then became the focus of additional experiments.
Expression of β2AR-luciferase fusion genes in HepG2 cells incubated in the absence or presence of dexamethasone. Human HepG2 cells were transiently transfected with the human glucocorticoid receptor expression vector pRShGRα (Brasieret al., 1990), the β-galactosidase expression vector pRSVβgal, and β2AR-luciferase fusion genes containing 3254, 2677, 1240, 768, 277, or 187 bp of 5′-flanking DNA from the β2AR gene linked to a firefly luciferase-encoding gene (LUC) in pGL3-Basic (Promega), as described in the text. After transfection, the cells were incubated for 8 hr in either the absence or presence of 0.1 μm dexamethasone and were harvested. Values are means ± standard errors of data from nine independent experiments, each performed in triplicate.Arrow, primary transcription start site in the β2AR gene. *, Significant (p < 0.05) difference in luciferase activities in lysates from untreated and dexamethasone-treated cells, as determined by Student’st test.
EMSAs using nuclear extracts.
To determine whether nuclear transcription factors could indeed bind to GRE5, we performed EMSAs using a 35-bp double-stranded oligonucleotide that includes GRE5 and nuclear extracts prepared from HepG2 cells that had been treated with 0.1 μmdexamethasone for 8 hr. In preliminary experiments in which increasing amounts of HepG2 cell nuclear extracts were added to radiolabeled GRE5 probe, we determined that 6 μg of nuclear extract resulted in optimal levels of shifted product (data not shown). Incubation of radiolabeled GRE5 with HepG2 nuclear extracts resulted in a prominent shifted band (Fig.3). The specificity of binding was confirmed by observing that the addition of increasing concentrations of unlabeled GRE5 displaced radiolabeled GRE5 in a dose-dependent manner, whereas the unlabeled random oligonucleotide did not suppress radiolabeled GRE5 binding (Fig. 3).
EMSA characterization of HepG2 cell nuclear proteins that interact with GRE5. Nuclear extracts (6 μg) prepared from HepG2 cells transfected with pRShGRα were incubated with radiolabeled GRE5 in the presence of increasing concentrations of the indicated double-stranded oligonucleotide, as described in the text. Lane 1, radiolabeled GRE5 alone; lanes 2 and8, radiolabeled GRE5 incubated with nuclear extract; lanes 3–7, radiolabeled GRE5 incubated with nuclear extract and a 10-, 50-, 100-, 250-, or 500-fold molar excess of GRE5, respectively;lanes 9–13, radiolabeled GRE5 incubated with nuclear extract and a 10-, 50-, 100-, 250-, or 500-fold molar excess of the random oligonucleotide, respectively. The sequences of the oligonucleotides are shown in Table 1. Labeled complexes were separated on polyacrylamide gels and visualized by autoradiography.
To assess the immunological identity of the proteins in the band shift complex produced with HepG2 cell nuclear extracts and GRE5, supershift experiments were carried out with polyclonal anti-human glucocorticoid receptor antibody. The anti-human glucocorticoid receptor antibody interacted with the GRE5-protein complex, as evidenced by the appearance of a more slowly migrating band (Fig.4). Nonspecific antiserum had no effect on the mobility of the GRE5-protein complex, nor did the anti-human glucocorticoid receptor antibody interact directly with GRE5 (Fig. 4).
Immunological characterization of the protein-GRE5 complexes. Supershift assays were performed as described for EMSAs except that, after the addition of either anti-human glucocorticoid receptor antibody or preimmune antiserum to the mixture of radiolabeled GRE5 and nuclear extract, the incubation was continued for 30 min at room temperature. Lane 1, radiolabeled GRE5 alone; lane 2, radiolabeled GRE5 incubated with nuclear extract;lane 3, radiolabeled GRE5 incubated with nuclear extract and anti-human glucocorticoid receptor (Anti-GR) antibody; lane 4, radiolabeled GRE5 incubated with nuclear extract and preimmune antiserum (diluted 1:500); lane 5, radiolabeled GRE5incubated with anti-human glucocorticoid receptor antibody. Labeled complexes were separated on polyacrylamide gels and visualized by autoradiography.
EMSAs using the human recombinant glucocorticoid receptor.
We further characterized the ability of GRE5 to bind the glucocorticoid receptor in vitro using EMSAs with human recombinant glucocorticoid receptor and serial dilutions of competitor oligonucleotides. Radiolabeled GRE5 incubated with human recombinant glucocorticoid receptor resulted in a single shifted band (Fig.5). Addition of increasing concentrations of unlabeled GRE5 displaced radiolabeled GRE5 in a dose-dependent manner (Fig. 5). In contrast, increasing concentrations of either unlabeled GRE1 or unlabeled mutant m1GRE5 (with a single nucleotide change in the core sequence of GRE5) showed decreased ability to compete with radiolabeled GRE5 for binding to human recombinant glucocorticoid receptor (Fig. 5).
Specificity of the interaction between GRE5 and the human recombinant glucocorticoid receptor, as determined in EMSAs. Human recombinant glucocorticoid receptor (6 μg), in the absence of added glucocorticoids, was incubated with radiolabeled GRE5 in the presence of increasing concentrations of the indicated double-stranded oligonucleotide, as described in the text. Lane 1, radiolabeled GRE5 alone; lane 2, radiolabeled GRE5 incubated with human recombinant glucocorticoid receptor; lanes 3–5, radiolabeled GRE5incubated with human recombinant glucocorticoid receptor and a 25-, 100-, or 250-fold molar excess of GRE5, respectively;lanes 6–8, radiolabeled GRE5 incubated with human recombinant receptor and a 25-, 100-, or 250-fold molar excess of m1GRE5, respectively; lanes 9–11, radiolabeled GRE5 incubated with human recombinant receptor and a 25-, 100-, or 250-fold molar excess of GRE1, respectively. The sequences of the oligonucleotides are shown in Table1. Labeled complexes were separated on polyacrylamide gels and visualized by autoradiography.
Incubation of radiolabeled GRETAT (the GRE from the human tyrosine aminotransferase gene) with the human recombinant glucocorticoid receptor resulted in a single shifted band (Fig.6). Addition of increasing concentrations of unlabeled GRETAT, GRE5, or GRECON (a 15-bp consensus GRE flanked by sequence surrounding GRE5 in the β2AR gene) displaced radiolabeled GRETAT from the human recombinant glucocorticoid receptor in a dose-dependent manner (Fig. 6).
Specificity of the interaction between GRE5 and the human recombinant glucocorticoid receptor, as determined in EMSAs. The human recombinant glucocorticoid receptor (6 μg), in the absence of added glucocorticoid, was incubated with radiolabeled GRETAT in the presence of increasing concentrations of the indicated double-stranded oligonucleotide, as described in the text. Lanes 1 and 13, radiolabeled GRETAT alone; lanes 2 and 12, radiolabeled GRETAT incubated with human recombinant glucocorticoid receptor; lanes 3–5, radiolabeled GRETATincubated with human recombinant glucocorticoid receptor and a 25-, 100-, or 250-fold molar excess of GRETAT, respectively;lanes 6–8, radiolabeled GRETAT incubated with human recombinant glucocorticoid receptor and a 25-, 100-, or 250-fold molar excess of GRE5, respectively; lanes 9–11, radiolabeled GRETAT incubated with human recombinant glucocorticoid receptor and a 25-, 100-, or 250-fold molar excess of GRECON, respectively. The sequences of the oligonucleotides are shown in Table 1. Labeled complexes were separated on polyacrylamide gels and visualized by autoradiography.
Transient transfections using a β2AR-luciferase fusion gene with mutated GRE5.
To further test the involvement of GRE5 in glucocorticoid regulation of β2AR expression, a plasmid [pβ2ARm1(−3129/+126)] was constructed that had been mutated at position +6 of GRE5(gggtgagctgttct to gggtgagctattct). This mutation, the same base change as in oligonucleotide m1GRE5 (Table 1), is essential for glucocorticoid inducibility of a MMTV GRE (Nordeen et al., 1990). Our results demonstrated loss of glucocorticoid inducibility using pβ2ARm1(−3129/+126) (Fig.7). Interestingly, in the absence of added dexamethasone, the activity of pβ2ARm1(−3129/+126) was markedly lower than that of pβ2AR(−3129/+126) (Fig. 7). A possible explanation is that basal expression of pβ2AR(−3129/+126) in HepG2 cells that overexpress the glucocorticoid receptor is relatively high, despite removal of glucocorticoids from the serum by charcoal stripping. Alternatively, GRE5 contributes to the basal activity of the β2AR gene promoter.
Effect of a single point mutation in GRE5 on the dexamethasone inducibility of a β2AR-luciferase fusion gene. Human HepG2 cells were transiently transfected with pRShGRα, pRSVβgal, and either pβ2AR(−3129/+126) or pβ2ARm1(−3129/+126), as described in the text. After transfection, the cells were incubated for 8 hr in either the absence or presence of 0.1 μmdexamethasone and were harvested. Values are means ± standard errors of data from five independent experiments, each performed in triplicate. ∗, Significant (p < 0.05) difference in luciferase activity, compared with untreated and dexamethasone-treated cells, as determined by Student’st test.
Transient transfections using putative GREs fused to a heterologous promoter.
To further examine the functionality of the putative GREs, fragments that contained either GRE1, GRE5, or GRE2 plus GRE3 and GRE4 together were fused to a luciferase expression plasmid driven by a minimal TK promoter. An MMTV fragment containing a functional GRE (Majors and Varmus, 1983) cloned into pT81LUC was used as a positive control. Approximately 4-fold induction was observed with dexamethasone using the MMTV-pT81LUC fusion gene (Fig. 8). Activity of GRE5-pT81LUC was induced 3.2-fold in the presence of dexamethasone (Fig. 8), a value that was higher than that observed with any of the β2AR-luciferase fusion genes that included GRE5. Activity of GRE1-pT81LUC was not induced by dexamethasone (Fig. 8). Transfection of HepG2 cells with the segment −831 to −708 (which contains GRE2, GRE3, and GRE4) fused to pT81LUC resulted in luciferase activity in either vehicle- or dexamethasone-treated cells that was below the level of detection in the assay (Fig. 8). This result suggests the presence of an element in the segment −831 to −708 of the β2AR gene that negatively affects the activity of the minimal TK promoter.
Expression of GRE-luciferase fusion genes in HepG2 cells incubated in the absence or presence of dexamethasone. HepG2 cells were transiently transfected with the human glucocorticoid receptor expression vector pRShGRα (Brasier et al., 1990), pRL-SV40, and luciferase fusion genes consisting of either GRE5, GRE1, or MMTV double-stranded oligonucleotides linked to a firefly luciferase-encoding gene in pT81LUC (Nordeen, 1988), as described in the text. The sequences of the double-stranded oligonucleotides are shown in Table 1. After transfection, the cells were incubated for 8 hr in either the absence or presence of 0.1 μm dexamethasone and were harvested. Values are means ± standard errors of data from five independent experiments, each performed in triplicate. ∗, Significant (p < 0.05) difference in luciferase activities in lysates from untreated and dexamethasone-treated cells, as determined by Student’s t test.
Discussion
Glucocorticoids increase β2AR expression and β2-agonist-stimulated adenylyl cyclase activity in a variety of tissues and cell types (Cheng et al., 1980; Norris et al., 1987; Collins et al., 1988; Takahashi and Iizuka, 1991; Zhong and Minneman, 1993;Dangel et al., 1996). For most steroid hormone-responsive genes, including those regulated by glucocorticoids, changes in the level of expression induced by hormone are principally controlled by changes in the rate of transcription of the target gene (Beato et al., 1989). These trans-activation effects are a result of the glucocorticoid hormone-receptor complex interacting with GREs within the target gene and thus acting as a transcription factor (Beatoet al., 1989). In the case of the β2AR gene, the results of numerous studies suggest transcriptional regulation by glucocorticoids. Steady state levels of β2AR mRNA in several cell lines and tissues have been shown to be increased in the presence of glucocorticoids (Collins et al., 1988; Hadcock and Malbon, 1988; Takahashi and Iizuka, 1991; Zhong and Minneman, 1993; Maket al., 1995; Dangel et al., 1996). The rate of transcription of the β2AR gene has been shown, by nuclear run-off transcription assays, to be increased by glucocorticoids in both DDT1 MF-2 cells (Collinset al., 1988) and rat lung (Mak et al., 1995).
In this study, the 5′-flanking region of the rat β2AR gene was isolated and used to prepare β2AR-luciferase fusion genes, as a means to identify cis-acting elements that mediate the stimulatory effects of glucocorticoids on β2AR gene expression. The HepG2 human liver cell line was chosen for these studies because human hepatocytes are known to express the β2AR (Bevilacqua et al., 1987) and HepG2 cells were previously used to study glucocorticoid regulation of angiotensinogen gene expression (Brasier et al., 1989,1990). HepG2 cells are deficient in functional glucocorticoid receptors (Brasier et al., 1990); however, this deficiency enabled us to investigate the role of glucocorticoids in the regulation of β2AR gene expression by performing co-transfection with a glucocorticoid receptor-encoding expression plasmid.
Early reports demonstrated that, in rats, glucocorticoids decrease β2AR number and β-agonist-stimulated adenylyl cyclase activity (Wolfe et al., 1976; Chan et al., 1979). More recent evidence suggests that glucocorticoid regulation of rat liver β2AR expression is age dependent. Hepatic β2AR number, although decreased in young rats, is actually increased in aged rats by glucocorticoids (Slotkin et al., 1996). To the best of our knowledge, glucocorticoid regulation of the β2AR has not been studied in human liver. However, our data clearly demonstrate that the HepG2 cell line is a valid model system in which to study glucocorticoid regulation of β2AR gene expression.
Cloning of the rat β2AR gene (Buckland et al., 1990; Jiang and Kunos, 1995; McGraw et al., 1996) has allowed investigation of the genetic elements involved in glucocorticoid regulation of β2AR gene expression. Indirect evidence obtained in an early study suggested that the putative cis-acting elements were located in the 5′-flanking region of the β2AR gene (Malbon and Hadcock, 1988). Examination of the rat β2AR gene sequence for known transcription factor binding sites indicated that six potential GREs are located within the proximal 3.7 kilobases of 5′-flanking DNA. In the transfection analyses using various β2AR-luciferase fusion genes, we were able to demonstrate that dexamethasone responsiveness was mediated by a region spanning nucleotides −643 to −152. Within this segment is a single putative GRE, with the sequence gggtgagcttgttct (Fig. 1). This GRE, designated GRE5, displays reasonable similarity to the consensus GRE sequence ggtacannntgttct (Beato et al., 1989). Approximately 2-fold induction with dexamethasone was observed with the longer β2AR-luciferase fusion gene constructs that contained GRE5. This level of induction was less than that observed with N-600 prATLUC. However, it should be noted that numerous investigators have obtained similar levels of induction (2–4-fold) with glucocorticoid administration, in measurements of either receptor numbers, steady state mRNA levels, or β2AR gene transcription rates, in a variety of cells and tissues (Norris et al., 1987; Collins et al., 1988; Hadcock and Malbon, 1988; Takahashi and Iizuka, 1991;Zhong and Minneman, 1993; Mak et al., 1995; Dangel et al., 1996). Therefore, the GRE that we have identified as lending glucocorticoid responsiveness to β2AR gene expression seems to be physiologically relevant.
The reason for the diminished responsiveness of the β2AR gene to glucocorticoid induction, compared with that of other glucocorticoid-inducible genes such as N-600 prATLUC (Brasier et al., 1989), is unclear at this time. The limited similarity of GRE5 to the consensus GRE (Fig. 1B) suggests that the biological response of the β2AR gene to dexamethasone stimulation may be related to the extent of homology to the consensus GRE. Of interest was the observation that GRE5, when fused to a luciferase expression vector driven by the basal TK promoter and transiently transfected into HepG2 cells, resulted in 3.2-fold induction of activity. This induction was similar to that obtained with MMTV sequences cloned into pT81LUC. These results suggest either that the β2AR promoter is a relatively weak substrate for GRE enhancer activity, that the activity of GRE5 is constrained by sequences in the 5′-flanking region of the β2AR gene, or that other segments of the β2AR gene that are missing from our reporter gene constructs are necessary for greater induction by glucocorticoids. In many genes, GREs have been demonstrated to synergize with other transcription factor-binding elements to increase gene activity. For example, a direct interaction between interleukin-1-inducible NFκB binding to an acute-phase response element and two flanking GREs has been demonstrated to underlie the acute-phase activation of the angiotensinogen gene in rat liver (Ron et al., 1990). Through recruitment of hepatic nuclear factor 3, protein kinase A has been shown to modulate the activity of a GRE in the rat tyrosine aminotransferase gene (Espinaset al., 1995) and phosphoenolpyruvate carboxykinase genes (Wang et al., 1996). Finally, widely spaced GREs have been shown to be additive in stimulating glucocorticoid induction of the tryptophan oxygenase gene in rat liver (Danesch et al., 1987).
The results from the EMSAs confirmed that GRE5 is capable of binding nuclear proteins isolated from glucocorticoid-treated HepG2 cells, as well as the human recombinant glucocorticoid receptor. A double-stranded oligonucleotide containing GRE5 and surrounding sequence bound a protein in nuclear extracts prepared from glucocorticoid-treated HepG2 cells that was immunologically identified as a glucocorticoid receptor. Moreover, the same double-stranded oligonucleotide bound the human recombinant glucocorticoid receptor. Interestingly, a single base change in GRE5 greatly reduced the ability of the mutant GRE (m1GRE5) to compete with radiolabeled GRE5 for binding to the human recombinant glucocorticoid receptor. This change (guanine to adenine in position +6 of the GRE) was previously shown to result in the complete loss of glucocorticoid inducibility of a MMTV GRE fused to a luciferase reporter gene (Nordeen et al., 1990).
Glucocorticoids are important therapeutic agents used in the treatment of asthma, but a small proportion of asthmatic patients are resistant to the beneficial effects of glucocorticoids (Cypcar and Busse, 1993). The molecular mechanisms underlying this form of steroid resistance remain unclear, although one of the beneficial effects of using glucocorticoids to treat asthma is increased β2AR expression (Barnes, 1996). Our EMSA results demonstrated that a single nucleotide change in GRE5 of the rat β2AR gene resulted in a greatly diminished ability to bind the human recombinant glucocorticoid receptor. This same nucleotide change in pβ2AR(−3129/+126) resulted in a fusion gene, pβ2ARm1(−3123/+126), that was no longer inducible by dexamethasone, as determined in transient transfection assays. Interestingly, mutations in the regulatory DNA of the human apolipoprotein (Suzuki et al., 1997) and 5-lipoxygenase (Inet al., 1997) genes have been shown to alter their expression, although the functional significance is currently not clear. The cis-acting elements in the human β2AR gene that mediate glucocorticoid responsiveness have not yet been identified. Nevertheless, the demonstration of steroid resistance in human patients, together with our present findings regarding glucocorticoid regulation of β2AR gene expression, suggests that the molecular basis for steroid resistance in a subset of asthmatic patients could involve mutations in GREs in the β2AR gene.
Acknowledgments
We thank Dr. Peter Buckland (University of Wales, Cardiff, UK) for contribution of the β2AR genomic clone, Dr. K. Cameron Falkner for helpful advice on performing supershift assays, and Dr. Patricia Wight for critical reading of the manuscript.
Footnotes
-
Send reprint requests to: Dr. Lawrence E. Cornett, Department of Physiology and Biophysics, Slot 750, University of Arkansas for Medical Sciences, 4301 West Markham Street, Little Rock, AR 72205. E-mail:cornettlawrencee{at}exchange.uams.edu
-
This work was supported in part by National Institutes of Health Grant R01-GM30669.
- Abbreviations:
- AR
- adrenergic receptor
- GRE
- glucocorticoid response element
- EMSA
- electrophoretic mobility shift assay
- bp
- base pair(s)
- MMTV
- mouse mammary tumor virus
- TK
- thymidine kinase
- HEPES
- 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
-
- Received August 6, 1998.
- Accepted September 11, 1998.
- The American Society for Pharmacology and Experimental Therapeutics
