Transcriptional Regulation of the Human α1a-Adrenergic Receptor Gene

We recently cloned cDNAs encoding three subtypes of human α1-adrenergic receptors (α1ARs), α1a, α1b, and α1d (Schwinn, D. A., Johnston, G. L., Page, S. O., Mosley, M. J., Wilson, K. H., Worman, N. P., Campbell, S., Fidock, M. D., Furness, L. M., Parry-Smith, D. J., Peter, B., and Bailey, D. S. (1995) J. Pharmacol. Exp. Ther. 272, 134–142) and demonstrated predominance of α1aARs in many human tissues (Price, D. T., Lefkowitz, R. J., Caron, M. G., Berkowitz, D., and Schwinn, D. A. (1994) Mol. Pharmacol. 45, 171–175). Several lines of evidence indicate that α1aARs are important in clinical diseases such as myocardial hypertrophy and benign prostatic hyperplasia. Therefore, we initiated studies to understand mechanisms underlying regulation of α1aAR gene transcription. A genomic clone containing 6.2 kb of 5′-untranslated region of the human α1aAR gene was recently isolated. Ribonuclease protection and primer extension assays indicate that α1aAR gene transcription occurs at multiple initiation sites with the major site located 696 base pairs upstream of the ATG, where a classic initiator sequence is located. Transfection of luciferase reporter constructs containing varying amounts of 5′-untranslated region into human SK-N-MC neuroblastoma cells indicate that a region extending 125 base pairs upstream from the main transcription initiation site contains full α1aAR promoter activity. Furthermore, distinct activator and suppressor elements lie 2–3 and 3–5 kilobase pairs upstream, respectively. Although the α1aAR promoter contains neither TATA or CAAT elements, gel shift mobility assays targeting three GC boxes immediately upstream of the main transcription initiation site confirm binding of Sp1. Activity of the α1aAR promoter is cell-specific, demonstrating highest activity in cells endogenously expressing α1aARs. The human α1aAR gene also contains several cis regulatory elements, including several insulin and cAMP response elements. Consistent with these observations, we provide the first evidence that treatment of SK-N-MC cells with insulin and cAMP elevating agents leads to an increase in α1aAR expression. In conclusion, these data represent the first characterization of the α1aAR gene; our findings should facilitate further studies designed to understand mechanisms regulating α1AR subtype-specific expression in healthy and diseased human tissue.

␣ 1 -Adrenergic receptors (␣ 1 ARs) 1 are members of the larger family of G protein-coupled receptors, mediating various sympathetic nervous system responses such as smooth muscle contraction and myocardial inotropy (1). ␣ 1 ARs couple predominantly via G q to activation of phospholipase C-␤ and hydrolysis of membrane phospholipids. Resultant formation of inositol triphosphate leads to release of calcium from intracellular stores and ultimately to muscle contraction (2), while diacylglycerol formation results in activation of protein kinase C. cDNAs encoding three subtypes of ␣ 1 ARs (␣ 1a , ␣ 1b , and ␣ 1d ; see Ref. 3 for new IUPHAR ␣ 1 AR subtype nomenclature) have been cloned, expressed in cells, and characterized pharmacologically (4 -9). We have previously demonstrated species heterogeneity (rat, rabbit, and human) and distinct tissue distribution for each ␣ 1 AR subtype, with ␣ 1a ARs predominating in many human tissues (10 -12). Clinically, ␣ 1a ARs have been shown to be important in the dynamic component of benign prostatic hyperplasia, with ␣ 1a AR-mediated prostate contractions correlating with urinary symptoms (13,14), and in the development of myocardial hypertrophy (15). ␣ 1 AR stimulation is known to have trophic and long term hypertrophic effects on cardiac and vascular smooth muscle structure and function (16 -18). In cardiac myocytes, ␣ 1 AR activation induces selective transcriptional activation of a "hypertrophic gene program," including expression of embryonic genes (such as atrial natriuretic factor (ANF), c-myc, c-fos, and c-jun), up-regulation, accumulation, and assembly of constitutively expressed contractile proteins (including skeletal actin, ␤-myosin heavy chain, and smooth muscle ␣-actin), activation of Raf-1 kinase/mitogen-activated protein kinase pathways, and increases in cell size without cellular proliferation. While the precise signaling mechanisms through which ␣ 1 AR-stimulated transcriptional activation occurs are not known, multiple pathways involving G q and several protein kinases have been implicated (16, 17, 19 -21). These enzymes presumably alter gene transcription via activation of transcription factors such as serum response factor, AP-1, and perhaps some cell-specific nuclear proteins, which in turn bind to distinct regulatory sequences in the 5Ј-untranslated region (5Ј-UTR) of ANF and other ␣ 1 AR-responsive genes such as ␣-actin, and ␤-myosin heavy chain. While some of these pathways are common to other receptors and agents that activate protein kinase C, at least one pathway is specific for ␣ 1 ARs. This is demonstrated by characterization of the ␣ 1 AR-sensitive phenylephrine response element (PERE) found in the 5Ј-UTR of the ANF gene, a cis regulatory element that is not responsive to hypertrophic agents that activate only the protein kinase C signaling pathway (e.g. phorbol esters) (22). This finding emphasizes that ␣ 1 AR stimulation provides a unique pathway for inducing muscle hypertrophy.
Recent studies demonstrate that ␣ 1 ARs themselves are regulated by agonist. Furthermore, regulation by agonist appears to be subtype-specific. Chronic norepinephrine stimulation of ␣ 1 ARs in neonatal rat ventricular primary myocyte cultures leads to repression of ␣ 1b and ␣ 1d AR mRNA expression, while concurrently transcription of ␣ 1a AR mRNA is induced. These changes are not explained by altered mRNA stability, and are accompanied by increased ␣ 1a AR subtype-specific protein expression, as well as myocardial hypertrophy (15). This phenomenon of differential up-regulation of cardiac ␣ 1 AR subtypes is also exhibited by other hypertrophic agonists such as endothelin-1, prostaglandin F2␣, and phorbol esters, which increase ␣ 1a AR transcription while repressing ␣ 1b and ␣ 1d AR mRNA expression (15). Hence, ␣ 1a ARs appear to play a critical role in the induction of myocardial hypertrophy, and are themselves transcriptionally induced by catecholamines and other agonists commonly found in clinical settings of myocardial hypertrophy.
Despite the importance of ␣ 1a ARs in pathophysiologic states, no studies have examined mechanisms underlying ␣ 1a AR transcriptional regulation. Understanding mechanisms involved in transcriptional regulation of human ␣ 1a ARs is important, given that these receptors play a role in prostate disease and cardiac hypertrophy. Therefore, we cloned 6.2 kb of 5Ј-UTR of the human ␣ 1a AR gene and, in the present study, we delineate various elements involved in its transcription. Our data indicate that the human ␣ 1a AR has multiple transcription initiation sites and contains a TATA-less promoter within 125 bp upstream of the main transcription initiation site in a region where we document Sp1 binding. The ␣ 1a AR promoter is cellspecific, and the receptor is up-regulated by cAMP and insulin. Finally, we identify the presence of two PERE consensus sequences upstream of transcription initiation sites, a finding that may explain the observed agonist-induced up-regulation of ␣ 1a ARs in rat heart (15).

EXPERIMENTAL PROCEDURES
Cloning Human ␣ 1a AR Genomic DNA To isolate 5Ј-UTR sequence in the human ␣ 1a AR gene, a human genomic library (female peripheral blood leukocytes in EMBL3 SP6/T7, average insert size 15 kb, CLONTECH, Palo Alto, CA) was screened. A probe containing 450 bp of 5Ј-UTR (DNA immediately upstream from the initiator ATG) was generated from EcoRI/NcoI digestion of a human ␣ 1a AR cDNA (5) clone generously provided by Dr. Julie Tseng-Crank (Glaxo-Wellcome Inc., Research Triangle Park, NC), and radiolabeled with [␣-32 P]dCTP via random priming; approximately 2 ϫ 10 6 phage were screened (100 plates of 150 mm size, containing about 2 ϫ 10 4 phage clones/plate). Positive clones were initially characterized by restriction analysis, followed by southern hybridization (23). DNA sequencing of the entire 5Ј-regulatory region of each distinct clone was performed using Sanger dideoxy chain termination methods (fmol™ DNA sequencing system; Promega, Madison, WI). Ligation of overlapping clones produced a final genomic construct containing 6.2 kb of 5Ј-UTR, which was subcloned into pGEM-5Zf(ϩ) (Promega) at SacI/PstI sites using standard molecular biology methods. This final construct provided a template for further 5Ј-deletion reporter constructs used in examining transcriptional regulation of the human ␣ 1a AR gene.

RNase Protection Assays
RNA Isolation-Total RNA was extracted from cells (SK-N-MC, Chang liver, DU145, rat-1 fibroblast cells (wild type and stably expressing each human ␣ 1 AR subtype)) using the RNazol method (Teltest Inc., Friendswood, TX). Each RNA sample was quantitated spectrophotometrically at 260 and 280 nm and stored at Ϫ70°C as an ethanol precipitate.
RNase Protection Assay Methods and Quantitation of Final Product-RNase protection assays were conducted as described previously (27) with a few modifications using the RPA II kit (Ambion). To determine the ␣ 1a AR transcription initiation site, 30 g of total RNA from SK-N-MC cells was dissolved in 20 l of RPA II kit hybridization buffer containing Ͼ20-fold excess of radiolabeled probe (Ϫ485, ϩ4) or probe (Ϫ898, Ϫ481) (3 ϫ 10 5 cpm/reaction), and incubated overnight at 42-65°C. In experiments designed to quantitate ␣ 1 AR subtype mRNA expression in cells, 30 g of total RNA from each cell line was dissolved in 20 l of RPA II kit hybridization buffer containing excess ␣ 1a AR probe (3 ϫ 10 5 cpm/reaction) and control cyclophilin probe (1 ϫ 10 5 cpm/10 g of total RNA), and incubated at 55°C (␣ 1a AR, ␣ 1b AR) and 65°C (␣ 1d AR) overnight. After digestion with RNase A (50 g/ml) and RNase T1 (700 units), protected RNA fragments were separated on a 6% polyacrylamide gel; radiolabeled RNA size markers and a DNA sequencing ladder using the same primer and genomic DNA were loaded on the same gel. The dried gel was exposed to X-Omat AR film (Eastman Kodak Co.) for 24 -36 h. Each RNase protection assay gel was further analyzed on a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). Volume integration of protected radiolabeled bands was corrected for background and total number of radiolabeled uracils contained in each RNA probe, normalized for cyclophilin signal, and expressed as arbitrary density units using ImageQuant gel image analysis software (Molecular Dynamics).

Primer Extension Analysis
A 36-bp oligonucleotide (5Ј-GCAGTTACCTACATTTTGAGCTGC-CCCACCGAAGGC-3Ј) was made complementary to the region (Ϫ530 to Ϫ566 upstream from the translation initiation site (ATG)) of the human ␣ 1a AR gene. The oligonucleotide was end-labeled with [␥-32 P]ATP using T4 polynucleotide kinase and primer extension assays performed as described previously (28) with 100 g of total RNA from SK-N-MC cells. Final products were separated electrophoretically using an 8% polyacrylamide gel; a sequencing reaction using human ␣ 1a AR genomic DNA as a template and identical primer utilized for primer extension were run side-by-side on the gel.

Transient Cell Transfections
Transient transfection into human SK-N-MC cells was used to measure reporter gene expression. Prior to transfection, cells were harvested by trypsinization, followed by centrifugation (5 min, 4°C, 640 ϫ g). Cells were washed twice with ice-cold phosphate-buffered saline (PBS) and then resuspended in PBS at a density of 1 ϫ 10 7 /ml cell suspension. Cells (0.5 ml; 5 ϫ 10 6 total cells) were incubated with 8 g of human ␣ 1a AR 5Ј-UTR deletion construct and 8 g of control pCAT-control vector (chloramphenicol acetyltransferase (CAT) gene placed under the control of SV40 promoter and enhancer sequences; Promega) in electroporation cuvettes for 10 min at room temperature. Transfection was accomplished using electroporation (Gene Pulser II, Bio-Rad; optimal conditions were 250 -290 V and 550 -950 microfarad capacity, resulting in a time constant of 12-15 ms). After electroporation, SK-N-MC cells were incubated in 100-mm plates for 96 h. Cells were then washed twice with PBS, followed by a 5-min incubation in 0.8 ml of lysis buffer (0.25 M Tris-Cl, pH 7.8, 0.5% (v/v) Triton X-100, 10 mM dithiothreitol) at room temperature, and scraped into tubes. Cellular debris was pelleted by centrifugation (5 min, 4°C, 640 ϫ g); 100 l of cell extract (supernatant) was used for luciferase assays, 50 l for CAT assays, and 10 l to determine protein concentration with BCA reagents (Pierce).

CAT and Luciferase Activity
CAT assays were performed by the phase-extraction method as described previously (23). The final xylene phase was placed in a scintillation vial and resultant tritium activity counted by a scintillation counter. Luciferase activity was measured as described previously using the Luciferase Assay System (Promega) (26,29). Luciferase reaction buffer (44 l; 0.125 M MgSO 4 , 42 mM ATP, 4.5 mg/ml bovine serum albumin) and 156 l of 50 mM glycine (pH 7.8) were added to 100 l of cell extracts and incubated at room temperature for 30 min. Luminescence resulting from luciferase gene expression was measured by a luminometer (Lumat LB 9501; Berthold Systems, Inc., Aliquippa, PA). The reaction was initiated upon the addition of 100 l of a 200 M luciferin solution; measured luminescence was stable within 20 -30 s. Final luciferase activity was normalized to CAT activity to control for transfection efficiency, and results were reported as -fold over basal control (pGL2-Basic or pGL2-Enhancer; Promega).

Electrophoretic Gel Mobility Shift Assay
Gel mobility shift assays were performed using two double-stranded oligonucleotides corresponding to DNA sequence immediately upstream of the human ␣ 1a AR main transcription initiation site (Ϫ74 to Ϫ3 (probe 1), and Ϫ146 to Ϫ89 (probe 2)) in buffer consisting of 12 mM Tris, pH 7.9, 12% glycerol, 35 mM KCl, 0.07 mM EDTA, 1 mM dithiothreitol, 7.5 mM MgCl 2 , with 2 g of poly(dI-dC) (to inhibit nonspecific protein binding). Probes 1 and 2 were created with 4 -8 bp of 5Јoverhanging sequence at the terminal ends; each probe was radiolabeled by filling in the 5Ј-overhanging DNA sequence using Klenow in the presence of [␣-32 P]dGTP and other nonradiolabeled nucleotides. Radiolabeled probes 1 and 2 (0.2 ng each) were incubated with purified Sp1 (1 footprinting unit; Promega, Milwaukee, WI) in 20 l of gel mobility shift assay buffer and incubated for 15 min at room temperature. We synthesized the oligonucleotide from the dihydrofolate reductase (DHFR) gene known to bind Sp1 (5Ј-AGTTGATCGGGGCGGGGC-GATCTA-3Ј) (30); a 50 -100-fold excess was used to compete for binding to probe 1 and probe 2 (added at the same time as labeled probe). To test for specificity, 50-and 100-fold molar excesses of a mutant Sp1 DHFR oligonucleotide (5Ј-AGTTGATCGGAAAGGGGCGATCT-3Ј) (30) were also utilized. Resultant oligonucleotides (and associated proteins) were separated according to size by electrophoresis on a 4% nondenaturing gel (30:1 acrylamide/bisacrylamide ratio, containing 2% glycerol) in Tris glycine buffer (0.025 M Trizma (Tris base), 0.19 M glycine) at 4°C. The gel was dried and exposed to Kodak X-Omat AR film for 4 -12 h.

Effect of Insulin and cAMP on Regulation of ␣ 1a AR Expression
SK-N-MC cells were grown in monolayers (described above) using heat-inactivated fetal bovine serum to 90% confluence. After initial time-course studies, insulin (1 g/ml in water) was incubated with SK-N-MC cells for 12 or 24 h. In cAMP experiments, forskolin (25 M, activator of adenylyl cyclase) and 3-isobutyl-1-methylxanthine (IBMX; 0.2 mM, phosphodiesterase inhibitor) were incubated with SK-N-MC cells for 12 or 24 h. Control cells were incubated with the same concentration of solvent as treated cells (water for insulin experiments, ethanol and dimethyl formamide for cAMP experiments). At the end of the treatment period, total RNA was isolated from SK-N-MC cells and utilized for RNase protection assays (using human ␣ 1a AR and control human cyclophilin probes) as described above. To confirm whether mRNA changes in ␣ 1 AR expression were present at a protein level, saturation binding was performed using the radiolabeled ␣ 1 AR antagonist 125 I-HEAT (300 pM) as described previously (5); prazosin 1 M was used for nonspecific binding. Analysis of ␣ 1 AR subtypes was determined using competition analysis with a K d concentration of 125 I-HEAT (130 pM) and non-radiolabeled ␣ 1a AR subtype selective ligand 5-methylurapidil as described previously (5,31).

RESULTS
Human ␣ 1a AR Genomic Clone-Four genomic clones (16.4 kb) were isolated with identical restriction maps (Fig. 1). Each clone contained a 6.0-kb BamHI fragment (consisting of 899 bp of 5Ј-UTR, the entire 0.88 kb of the first exon encoding the human ␣ 1a AR, and 4.3 kb of intron sequence; Fig. 1A), a 5.4-kb SacI fragment (consisting exclusively of ␣ 1a AR 5Ј-UTR; Fig.  1B), and a 11.0-kb fragment (consisting of 823 bp of 5Ј-UTR, the entire 0.88-kb first exon of the human ␣ 1a AR, and 9.29 kb of intron sequence; Fig. 1C). A SacI/PstI fragment containing the first exon of the gene (Fig. 1C) was ligated with the 5.4-kb SacI fragment (Fig. 1B) to produce the final genomic clone in pGEM-5Zf(ϩ) (Promega) (Fig. 1D). Correct orientation of fragments was confirmed by comparison with overlapping sequence from the BamHI genomic fragment (Fig. 1, A and D). The final SacI/PstI genomic construct consists of the following: 6.2 kb of 5Ј-UTR, the entire 0.88-kb human ␣ 1a AR first exon, and 3.0 kb of intron sequence (Fig. 1D). These data confirm the presence of an intron in the human ␣ 1a AR gene (Ͼ9 kb; located in the encoded protein at the junction of the sixth transmembrane domain and the third extracellular loop), similar to our original bovine ␣ 1a AR clone, and reported more recently in human and rat ␣ 1b AR genes (32,33).
Identification of the Transcription Initiation Site-As recently described by our laboratory (24), sequence analysis of the human ␣ 1a AR 5Ј-UTR (GenBank™ U72653) reveals a TATA-less promoter. Therefore, location of the putative transcription initiation site was investigated using RNase protection and primer extension assays. RNase protection assays were performed with total RNA isolated from SK-N-MC cells (the only human cell line expressing endogenous ␣ 1a ARs); the location of RNA probes used (relative to the translation initiation site ATG) is shown in Fig. 2A. As shown in Fig. 2B, probe (Ϫ485, ϩ4) is generally fully protected (lanes 3-5, 489-bp fragment), indicating that the main transcription initiation site is Ͼ489 bp upstream of the translation initiation site. In addition, a partially protected fragment (110 bp) is located 106 bp upstream of the translation initiation site (Fig. 2B, lanes 7-9), representing a minor transcription initiation site. Note there are several other fragments, which can be attributed to probe self-protection since they are present in the tRNA lane in multiple (n ϭ 6) experiments (despite rigorous prehybridization denaturation at 95°C, and hybridization at temperatures up to 65°C). To further define the main transcription initiation site, a second RNase protection assay probe was utilized. Experiments with probe (Ϫ898, Ϫ481) show four partially protected fragments (lanes [12][13][14]; the largest and most prominent is 200 bp in size (corresponding to Ϫ681 bp upstream from the ATG; also denoted by an arrow in Fig. 2A) as well as minor fragments (195, 144, and 131 bp in size, corresponding to Ϫ676, Ϫ625, and Ϫ612). Taken together, these experiments demonstrate one major and four minor transcription initiation sites for the human ␣ 1a AR gene in SK-N-MC cells. Of note, identical results are seen in RNA isolated from human prostate (data not shown).
To confirm that the major transcription initiation site for the Experiments with probe (Ϫ485, ϩ4) show undigested probe (lane 1; note the undigested probe is larger than protected fragments since additional polylinker sequence is present), control tRNA hybridized with probe (Ϫ485, ϩ4) (lanes 2 and 6; note that some self-protected fragments are seen even after rigorous prehybridization denaturation at 95°C and hybridization at temperatures up to 65°C in these tRNA lanes), and protected full-length probe (489 bp) from three different RNA samples (lanes [3][4][5]. Only one smaller 110-bp band (106 bp upstream from the ATG, lanes 7-9) demonstrates specific hybridization since most other bands are also present in tRNA control lanes (repeatedly seen in various experiments). Hence, while one minor transcription initiation site lies 106 bp upstream from the ATG, the major transcriptional initiation site(s) appear to lie upstream Ͼ489 bp. Experiments with probe (Ϫ898, Ϫ481) show undigested probe (Ϫ898, Ϫ481) (lane 10), control tRNA hybridized with probe (Ϫ898, Ϫ481) (lane 11), identical partially protected fragments from three different RNA samples ( lanes  12-14, arrows), and a DNA sequence ladder using probe (Ϫ898, Ϫ481) as a template (lanes 15-18). The approximate location of protected fragments hybridized with probe (Ϫ898, Ϫ481) relative to the human ␣ 1a AR translation initiation site is indicated by numbers at the far right. One predominant initiation site is confirmed with RNase protection analysis, located 681 upstream from the translational initiation site (Ϫ681 in Fig. 2B, arrow in Fig. 2A). Three other transcription initiation sites are noted (Ϫ676, Ϫ625, Ϫ612) for this probe. Overall, the data demonstrate one major and four minor transcription initiation sites for the human ␣ 1a AR gene in SK-N-MC cells.
human ␣ 1a AR gene occurs approximately 680 bp upstream from the ATG, a 36-bp oligonucleotide complementary to the region Ϫ530 to Ϫ566 was used in primer extension assays with total RNA isolated from SK-N-MC cells. Primer extension assays reveal three transcription initiation sites; Fig. 3 shows two of the transcription initiation sites furthest upstream from the ATG, one major site (Ϫ696, larger upper arrow) contains an initiation sequence of CAA in agreement with other initiators (34 -36), while a second site contains the sequence GCC (Ϫ689 bp, smaller lower arrow). The third initiation site was found further downstream (Ϫ625, data not shown).
Both primer extension and RNase protection assays demonstrate a major transcription initiation site for the human ␣ 1a AR gene between Ϫ680 and Ϫ700, and minor transcription initiation sites 5-7 bp and 69 -71 bp further downstream; however, initiation sites identified with these two methods are not identical. Although discrepancies as large as 17 bp between primer extension and RNase protection assays have been reported previously for other genes (37), to ensure the discrepancy was not due to the presence of a small intron, alternative splicing in the 5Ј-UTR, or RNA tertiary conformation, we se- Constructs containing varying amounts of human ␣ 1a AR 5Ј-UTR fused to the promoterless luciferase gene were transiently transfected into SK-N-MC cells and resultant luciferase activity measured (n ϭ 5-12 independent transfections (each performed in duplicate) for each construct). The left half of the figure schematizes each construct (white box ϭ promoterless luciferase), while the right half presents luciferase results. Luciferase activity is normalized for cotransfected CAT activity, and expressed as -fold over control (black box, relative to pGL2-Basic in A and B; gray box, relative to pGL2-Enhancer in B) (mean Ϯ S.E.). The main site of transcription initiation is defined as ϩ1. Restriction enzymes are as follow: S, SacI; H, HindIII; A, AvrII. A, constructs in A utilize a pGL2-Basic vector (which does not contain a SV40 enhancer); luciferase results reveal the presence of an endogenous activator in human ␣ 1a AR 5Ј-UTR between Ϫ1927 and Ϫ2869, and a suppressor of basal transcription located upstream of Ϫ2869. B, constructs in B utilize pGL2-Enhancer (which contains a SV40 enhancer in the vector). These results demonstrate that the first 125 bp upstream from the transcriptional initiation site is sufficient for basal transcription (contains entire promoter sequence). See "Results" for details.

FIG. 5. Sp1 gel mobility shift experiments demonstrating purified Sp1 binds to GC boxes in the human ␣ 1a AR promoter.
A, location of human ␣ 1a AR promoter sequence corresponding to gel mobility shift probes 1 and 2. Probe 1 contains one GC box (striped box), while probe 2 contains two GC boxes; of note, both probes contain other GC-rich regions that might also bind Sp1. The major transcription initiation site is labeled ϩ1 and the solid box represents human ␣ 1a AR translated sequence. B, gel mobility shift assay performed using probe 1 (lanes 1-6). C, gel mobility shift assay using probe 2 (lanes 7-12). B and C, lanes 1 and 7 contain 0.2 g of probe alone; lanes 2-6 and 8 -12 contain 0.2 g of probe, 2 g of poly(dI-dC), and 1 footprinting unit of purified Sp1; lanes 3 and 4 and lanes 9 and 10 contain 50-and 100-fold molar excess of an oligonucleotide corresponding to the DHFR sequence known to bind Sp1. To test for specificity, lanes 5 and 6 and lanes 11 and 12 include 50-and 100-fold molar excess of a mutant DHFR Sp1 oligonucleotide. Specific binding of Sp1 to both human ␣ 1a AR probes is observed with multiple bands (arrows), suggesting binding of more than one protein simultaneously. See "Results" for details. quenced cDNA made from total RNA isolated from SK-N-MC cells and compared it to sequence from our genomic clone; the sequences are identical and contain DNA up through base Ϫ696 (data not shown). Hence, we localize the major transcription initiation site for the human ␣ 1a AR gene in SK-N-MC cells to 696 bp upstream of the ATG, with four minor initiation sites occurring further downstream; minor transcription initiation sites are located the following distances from the ATG: 689 (derived from primer extension), 640 (derived from RNase protection plus a 15-bp discrepancy between this probe and primer extension; 625 ϩ 15 ϭ 640), 625 (derived from primer extension; confirmed by RNase protection 612 ϩ 15 ϭ 627 Ϸ 625), and 106 bp (derived from RNase protection; this probe does not require correction)). The major transcription initiation site (Ϫ696) is denoted ϩ1 in all subsequent experiments.
To further confirm the site of transcription initiation, two reporter luciferase constructs were created (one containing 5 bp of 5Ј-UTR upstream of the main start site (pLS⌬Ϫ5)) and another beginning 39 bp downstream from the main transcription initiation site (pLS⌬ϩ39). Each construct was transfected into SK-N-MC cells and resultant luciferase activity assessed. The pLS⌬Ϫ5 vector had 3.4-fold more luciferase activity than the pLS⌬ϩ39 vector (36 Ϯ 9.9 versus 11 Ϯ 3.2% maximal luciferase activity, respectively), suggesting that elimination of the main transcription initiation site significantly decreases RNA transcription in human SK-N-MC cells.
Delineation of the Promoter Region and Sequences Involved in Modulating Basal Transcription-Serial deletions of the human ␣ 1a AR 5Ј-UTR ligated to a luciferase reporter gene were used to determine the general location of the promoter and investigate regulatory regions involved in ␣ 1a AR expression in human SK-N-MC cells. Each construct was transfected into SK-N-MC cells, and luciferase activity was measured (normalized for cotransfected CAT activity) and expressed as -fold increase over control plasmid (pGL2-Basic for panel A and pGL2-Enhancer for panel B). As shown in Fig. 4A, in the absence of vector SV40 enhancer sequences (pGL2-Basic vector), transcription occurs with as little as 125 bp of 5Ј-UTR immediately upstream from the transcriptional initiation site. Sequences located between Ϫ1927 and Ϫ2869 (relative to the transcription initiation site ϩ1) enhance basal transcription. Further upstream, DNA sequences between Ϫ2869 and Ϫ5498 repress basal transcription, indicating the possible presence of a suppressor or silencer (38). The presence of SV40 enhancer sequence in the vector (pGL2-Enhancer vector; Fig. 4B) increases luciferase activity in pLS⌬Ϫ125 and pLS⌬Ϫ1927 to the level of pLS⌬Ϫ2869. These results confirm that DNA sequences present in the first 125 bp upstream from the transcription initiation site are sufficient for basal transcription of the human ␣ 1a AR gene, and thus contain the basic promoter.
Sp1 binds to GC Boxes in Close Proximity to the Human ␣ 1a AR Transcription Initiation Site-Since the human ␣ 1a AR promoter contains neither TATA nor CAAT box consensus sequences (24), the presence of three GC boxes ((G/T)(G/A)G-GCG(G/T)(G/A)(G/A)(C/T)) (39) located immediately upstream (Ϫ139, Ϫ100, Ϫ12 bp) from the predominant transcription initiation site suggests that binding of Sp1 in these regions might play a role in basal and/or activated transcription in this promoter. To investigate this possibility, we performed gel mobility shift assays utilizing radiolabeled oligonucleotides corre- sponding to DNA sequences in the human ␣ 1a AR promoter. Probe 1 contains one GC box, while probe 2 contains two GC boxes (Fig. 5A); of note, both probes contain additional GC-rich sequence, which might also bind Sp1. Fig. 5 (B and C) shows results from gel mobility shift assays using probes 1 and 2, respectively. The addition of purified Sp1 shifts the band shown with probe alone (lanes 1 and 7) to higher size (lanes 2 and 8); note that two bands are seen (arrows) suggesting multiple proteins bind to each probe. The addition of purified DHFR oligonucleotide sequence known to bind Sp1 (30) competes away Sp1 binding to the probe (lanes 3 and 4 and lanes 9  and 10), confirming specificity. Specific Sp1 binding is further confirmed by the absence of competition with a mutant DHFR Sp1 oligonucleotide containing AAA in its core sequence instead of GGC (30). Since each probe contains at least one GC box, Sp1 appears to bind to at least two (and possibly more) sites immediately upstream of the human ␣ 1a AR transcription initiation site. Two of these targeted Sp1 binding sites (Ϫ100 and Ϫ12 bp) are present within the 125 bp shown above to contain full promoter activity.
Cell-specific Transcription of the Human ␣ 1a AR Gene-Once the promoter region was delineated, we next examined whether the ␣ 1a AR promoter exhibits cell-specific transcription. Although ␣ 1a ARs are expressed in many native human tissues, expression in human cell lines is restricted. Three human cell lines known to express ␣ 1 ARs (SK-N-MC (neuroblastoma), DU145 (prostate cancer), Chang liver (hepatoma)), as well as a cell line that does not contain endogenous ␣ 1 ARs (rat-1 fibroblasts), were utilized in RNase protection assays to determine ␣ 1 AR subtype mRNA expression (Fig. 6, A and B). SK-N-MC cells contain ␣ 1d Ͼ ␣ 1a with minimal/no ␣ 1b AR mRNA (confirming previous reports by our laboratory) (10). DU145 and Chang liver cell lines contain only the ␣ 1b AR subtype, while rat-1 fibroblasts contain no endogenous ␣ 1 ARs. These cells were transiently transfected with the smallest ␣ 1a AR reporter construct shown previously to give full promoter activity (pLS⌬Ϫ125; see Fig. 4B), and resultant luciferase activity was examined. As shown in Fig. 6C, SK-N-MC cells, which are known to express ␣ 1a ARs, have the highest luciferase activity suggesting the presence of proteins/factors important in transcription. In contrast, much lower luciferase activity was present in Chang liver Ͼ DU145 cells. The rat-1 cells, which do not contain any endogenous ␣ 1a ARs, are indistinguishable from the control vector which contains no promoter. Taken together, these results demonstrate the occurrence of cell-specific transcription of the human ␣ 1a AR gene.
Effect of Insulin and cAMP on ␣ 1a AR Transcription and Expression-As schematized in Fig. 7, multiple putative cis regulatory elements for binding of insulin and cAMP response element-binding protein (CREB) are present in the human ␣ 1a AR gene. We therefore investigated whether these agents affected human ␣ 1a AR gene transcription. Fig. 8 shows results from an RNase protection assay using total RNA isolated from SK-N-MC cells after incubation with insulin (Fig. 8A, lanes  1-4) or forskolin/IBMX (which increase intracellular cAMP, Fig. 8B, lanes 5-8). No significant differences occur between control and treated cells for either drug at 12 h. However, after 24 h, insulin increases human ␣ 1a AR gene transcription modestly (20 -30%, Fig. 8A, lanes 3 and 4), while forskolin/IBMX increases transcription significantly (2.2-fold over control, Fig.  8B, lanes 7 and 8). Since the magnitude of induction of human ␣ 1a AR transcription is largest with cAMP, we examined ␣ 1 AR protein expression in membranes isolated from SK-N-MC cells after 24 h of forskolin/IBMX. The results of these experiments indicate a 2.6-fold increase in ␣ 1 AR protein. Furthermore, competition analysis with the ␣ 1a AR subtype selective ligand 5-  (ϩ1, arrow). A second GC-rich region (consistent with Sp1 binding; rectangle with thin rightward diagonal lines, ϩ550 to ϩ690) surrounds the downstream minor transcription initiation site. Putative insulin response elements (IRE, gray rectangles) and cAMP response elements (CRE, black ovals) are highlighted in addition to other known transcription factor binding sequences (Ap1, white ovals; Ap2, rectangles with leftward diagonal lines). Two PERE (checkered rectangle) consensus sequences are present in the human ␣ 1a AR gene (1405 bp upstream of the main transcription initiation site in SK-N-MC cells (ϩ1) and 194 bp upstream of a minor transcription initiation site located 106 bp 5Ј to the translation initiation site). This raises the possibility that the PERE could act as a cis regulatory element enhancing ␣ 1a AR transcription in the presence of agonist. For more details and identification of other consensus sequences present, see Ref. 24.
FIG. 8. Induction of human ␣ 1a AR gene transcription by cAMP and insulin. Results from an RNase protection assay examining human ␣ 1a AR and control cyclophilin mRNA concentrations in total RNA (30 g) isolated from SK-N-MC cells treated with insulin (1 g/ml, A) or cAMP (25 M forskolin and 0.2 mM IBMX, B). Although no significant changes in ␣ 1a AR mRNA expression occur at 12 h for either drug, modest increases in ␣ 1a AR mRNA expression with insulin (20 -30% increase) and significant increases with forskolin/IBMX (2.2-fold increase) occur at 24 h. methylurapidil reveals that increased ␣ 1 AR binding is due to increases in the ␣ 1a AR subtype (Table I). DISCUSSION The ␣ 1a AR is the predominant ␣ 1 AR in most human tissues including prostate and heart and appears to play an important role in benign prostatic hyperplasia and cardiac hypertrophy, yet nothing is known about transcription of the ␣ 1a AR gene. In the present study we report cloning of the human ␣ 1a AR gene, determine its overall genomic structure including the presence of an intron (Ͼ9 kb) in the encoded protein at the junction of the sixth transmembrane and third intracellular loop (identical location as bovine ␣ 1a AR (4), rat/human ␣ 1b (32,33)), and identify several transcription initiation sites, the most prominent located 696 bp upstream from the translation initiation codon in human SK-N-MC cells. Furthermore, we demonstrate that 125 bp immediately upstream of the transcription initiation site serves as the functional promoter, show Sp1 binding in this GC-rich region, identify several 5Ј-UTR regions expressing activator and suppressor activity, demonstrate cell-specific transcriptional regulation of human ␣ 1a ARs, and identify subtype-specific increased transcription and protein expression of human ␣ 1a ARs in the presence of insulin and cAMP in SK-N-MC cells. This represents the first identification and characterization of mechanisms underlying transcriptional regulation of ␣ 1a ARs.
One of the most important ways gene expression is regulated is through transcription. To examine basal and inducible transcription in the human ␣ 1a AR gene, we began by defining the transcription initiation site in SK-N-MC cells. Both RNase protection and primer extension assays identified several transcription initiation sites, with the main site located 680 -700 bp upstream from the translation start site. Since there was a 15-bp discrepancy in the exact transcription initiation site between these two methods, sequencing of cDNA made from ␣ 1a AR mRNA in this region was used to resolve this conflict. Although normally primer extension assays are less accurate than RNase protection assays, in the present study sequencing confirmed primer extension results, locating the main transcription initiation site 696 bp upstream from the ATG. One possible explanation for the discrepancy between RNase protection and primer extension assays is the presence of potential palindromic sequence in this region. A tertiary structure loop of DNA can be formed by CAAA (Ϫ1 to ϩ3) complementing with TTTG (ϩ16 to ϩ19); a similar palindromic motif has been reported in a plant gene to give a shortened (hence incorrect) RNase protection assay fragment, similar to our findings (37). Taking this information into account, five transcription initiation sites in the human ␣ 1a AR gene are located 696 (main), 689, 640, 625, and 106 bp upstream from the ATG. Precedence for multiple transcription initiation sites seen in the human ␣ 1a AR gene exists in several catecholamine receptor genes, including the rat ␣ 1b AR (which contains three distinct promoters) (33,40,41), human ␣ 1b AR (32), rat ␤ 1 AR (42), and the human dopamine 1A receptor (43).
The absence of TATA and/or a CAAT box sequence in the human ␣ 1a AR gene is interesting. Although TATA-less promoters were initially thought to be associated with housekeeping genes, over the last 5 years it has become apparent that many membrane receptor genes contain TATA-less promoters (35). In the adrenergic receptor family, using strict criteria (TATA(A/ T)A(A/T); 7/7 matches required), only human ␣ 2a , rat ␣ 2a/d , and turkey ␤ARs contain a TATA box. CAAT box consensus sequence (GG(C/T)CAATCT; 7/9 matches required with CAATC intact) is present only in human/rat ␣ 1b and mouse ␣ 2b AR genes. In the absence of a TATA box, transcription begins at initiator (Inr) consensus sequences and is often associated with Sp1 binding to GC-rich regions further upstream in the promoter (34 -36, 44, 45). Several Inr sequences (34,35) are present in the human ␣ 1a AR gene (24), with a classic Inr consensus sequence ( Ϫ3 CTCA ϩ1 ) located exactly at the major site of transcription initiation (696 bp upstream from the translation start site); this sequence has been associated with strong promoters in TATA-less genes. In fact, this Inr is found in the terminal deoxynucleotidyl transferase gene and adenovirus IVa2 and major late promoters, where (in the absence of a TATA motif) the sequence appears to be loosely recognized by RNA polymerase II in the context of a preinitiation complex consisting of TATA-binding protein (TBP), TFIIB, or TFIIF (45). Sp1 is a transcription factor that binds to at least two GC-rich regions on TATA-less promoters and (with other factors) is necessary for basal transcription (35). Sp1 also binds to TATA containing promoters to stimulate transcription to higher levels; activation of transcription by Sp1 involves coactivation with the transcription factor TAF 110 as well as involvement of TBP, TAF 250 and TAF 150 (46). Using gel mobility shift assays, we demonstrate specific Sp1 binding to at least two (and possibly more) GC-rich regions located immediately upstream of the main transcription initiation site in the human ␣ 1a AR gene. Since only 125 bp is required to obtain full promoter activity in this gene, Sp1 binding to at least two of these sites (Ϫ100, Ϫ12) may be important in ␣ 1a AR gene transcription.
Human ␣ 1a AR promoter activity is cell-specific, with highest activity demonstrated in SK-N-MC cells. Although trans-proteins bound to cis-regulatory elements are known to modulate transcription, recent evidence suggests that varying arrangements of basal promoter elements in the transcription complex can modulate the effect of a given activator (47), providing another mechanism for cell-specific differential regulation of transcription. Indeed, while increased basal human ␣ 1a AR promoter activity in Chang liver cells may be due to the existence SK-N-MC cells were incubated with forskolin (25 M) and IBMX (0.2 mM), compounds known to increase intracellular cAMP, for 24 h (n ϭ 5-6 independent experiments, each performed in triplicate). Rat-1 cells stably transfected with each human ␣ 1 AR subtype display the following affinities for 5-methylurapidil (pK i ):␣ 1a (8.5) Ͼ ␣ 1d (7.2) Ն ␣ 1b (6.6) (see Ref. 5). The high affinity site clearly corresponds to the ␣ 1a AR subtype. Since human SK-N-MC cells contain ␣ 1d ARs Ͼ Ͼ ␣ 1a ARs, with minimal/no ␣ 1b AR expression (11), and 5-methylurapidil does not reliably distinguish between ␣ 1b and ␣ 1d AR subtypes, the low affinity site probably represents the ␣ 1d AR. Increases in ␣ 1a ARs with forskolin/IBMX (21-40% total ␣ 1 ARs) represent a 5-fold increase in absolute ␣ 1a AR expression (4.0 -20 fmol/mg total protein). of multiple cis-regulatory consensus sequences for liver-specific proteins present in the 5Ј-UTR of this gene, differences between other cell lines may be due to varying compliments of proteins required in the transcription complex. In genes with multiple transcription initiation sites, cell-specific utilization of alternative transcription initiation sites can also affect expression; the human ␣ 1b AR provides an example in this regard, since a single transcript occurs in kidney, whereas multiple mRNA species (with size corresponding to alternative transcription initiation sites) are present in other tissues (32). The presence of GC-rich sequence 5Ј to a minor transcription initiation site (located 106 bp upstream of the translation start site), raises the possibility of alternative promoter activity in different cells/tissues in the human ␣ 1a AR gene; however, thus far the human ␣ 1a AR gene appears to have identical initiation sites in every tissue tested (SK-N-MC, prostate, vessels; data not shown). In addition to cell-specific promoter activity, results from experiments utilizing varying amounts of human ␣ 1a AR 5Ј-UTR fused to reporter luciferase sequence demonstrate modulation of basal promoter activity in SK-N-MC cells due to the presence of activator sequence 2-3 kb upstream from the transcription initiation site, as well as a strong suppressor/ negative modulator sequence Ͼ3 kb upstream. Sequence analysis reveals the presence of several putative positive and negative response elements in each corresponding region, although the exact activator and suppressor remain to be determined. Of note, a novel transcription factor, ␣-adrenergic receptor transcription factor (␣ARTF), has recently been shown to be essential for transcription of the rat ␣ 1b AR gene in most tissues (48). Consensus sequences for ␣ARTF binding are present in the human ␣ 1a AR gene, but are not present in the promoter region; hence, ␣ARTF does not appear to be essential for human ␣ 1a AR transcription.
An important finding of the present study is that the ␣ 1a AR is up-regulated by insulin and cAMP. Interestingly, both of these agents have been shown to affect other subtypes of ␣ 1 ARs, although the effect of these agents appears to be subtype-specific. For example, a recent study showed that insulin and insulin-like growth factor increased expression of the ␣ 1d AR in rat vascular smooth muscle up to 3-fold without affecting expression of the ␣ 1b (49). The increase in ␣ 1a AR expression by insulin in SK-N-MC cells in our study is more modest, despite the presence of several putative insulin response elements present in 5Ј-UTR of ␣ 1a AR gene. It is possible that insulin-mediated effects on ␣ 1a AR expression are cell typespecific and a more pronounced effect may occur in other tissues. In contrast to insulin, we demonstrate subtype-specific up-regulation (2.2-fold) of ␣ 1a AR mRNA and protein expression occurs with cAMP in SK-N-MC. This is an important finding and suggests that ␣ 1a AR may be cross-regulated by receptors that stimulate cAMP formation (e.g. ␤ARs, thyroid stimulatory hormone receptors). Since many tissues contain both ␣ 1 AR and ␤ARs (e.g. heart, prostate, and adipocytes), regulation of ␣ 1a AR by cAMP has important clinical implications. Indeed, a recent study in brown adipose tissue suggests that neural stress (cold) and direct ␤ 3 AR stimulation both result in specific ␣ 1a AR upregulation in this tissue (50). The exact mechanism by which cAMP up-regulates ␣ 1a AR mRNA is not clear; however, it is known that cAMP increases transcription of a number of genes via activation of the CREB family of transcription factors which interact with cAMP response element (CRE) sequence present in the 5Ј-UTR. Although the main CRE is characterized by the palindromic sequence 5Ј-TGACGTA-3Ј, several variations of this sequence also bind CREB and mediate transcription (51). Examination of the human ␣ 1a AR gene identifies the presence of several putative CREs (24), hence it is conceivable that cAMP increases ␣ 1a AR expression directly by increasing transcription through CRE sequences; this mechanism is involved in cAMP-mediated up-regulation of the ␣ 1b AR (52). In addition to directly acting through CREs, cAMP can also increase transcription by interacting with other regulatory sequences such as AP-1 (53), AP-2 (54), Sp1 (55), inverted CCAAT motif (56), and the estrogen response element (57). Since many of these regulatory sequences are present in human ␣ 1a AR gene (24), it is possible that cis-acting elements other than the CRE are responsible for observed cAMP-mediated up-regulation of ␣ 1a ARs in SK-N-MC cells. Alternatively, cAMP may up-regulate ␣ 1a ARs by increasing the stability of its mRNA, as has been observed for phosphoenolpyruvate carboxykinase (58). The exact mechanism by which cAMP up-regulates ␣ 1a AR will be the subject of our future studies.
An interesting property of the ␣ 1a AR, not shared by ␣ 1b or ␣ 1d , is that agonist exposure in myocytes leads to ␣ 1a AR upregulation (15). In the current study, we find two PERE consensus sequences (22) in the 5Ј-UTR of the human ␣ 1a AR gene located upstream of transcription initiation sites in SK-N-MC cells (Ϫ1405 from the main (ϩ1) transcription initiation site; Ϫ194 from a minor transcription initiation site located 106 bp upstream of the translation initiation site) (Fig. 7). It is intriguing to speculate that subtype-specific up-regulation of ␣ 1a ARs in the presence of agonist in rat neonatal myocyte cultures might be mediated by the binding of a cardiac specific transcription factor to the PERE cis regulatory element, resulting in increased ␣ 1a AR transcription in myocytes. Whether agonist-induced up-regulation of ␣ 1a ARs is specific to myocardium or more generalizable to other non-cardiovascular tissues (such as prostate in benign prostatic hyperplasia) remains to be determined.
In summary, we present the first description of the human ␣ 1a AR gene, initial characterization of its promoter, and upregulation by cAMP. Based on the finding of two PERE consensus sequences in the human ␣ 1a AR gene, we also suggest the tantalizing possibility that PERE-mediated induction of ␣ 1a AR expression provides a novel mechanism underlying cellspecific up-regulation of ␣ 1a ARs in myocardial hypertrophy.