HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 39, 35980-35989, September 27, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/39/35980    most recent M204832200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Le Jeune, I. R. Articles by Hall, I. P. Search for Related Content PubMed PubMed Citation Articles by Le Jeune, I. R. Articles by Hall, I. P. Social Bookmarking                      What's this?

Cyclic AMP-dependent Transcriptional Up-regulation of Phosphodiesterase 4D5 in Human Airway Smooth Muscle Cells

IDENTIFICATION AND CHARACTERIZATION OF A NOVEL PDE4D5 PROMOTER^*

Ivan R. Le Jeune^§, Malcolm Shepherd^§¶, Gino Van Heeke, Miles D. Houslay^**, and Ian P. Hall

From the  Division of Therapeutics and Institute of Cell Signalling, University Hospital, Nottingham NG7 2UH, United Kingdom, the ^¶ Department of Respiratory Medicine, Floor 6, Gartnavel General Hospital, Great Western Road, Glasgow, G12, the ^** Molecular Pharmacology Group, Division of Biochemistry and Molecular Biology, Davidson Building, Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow G12 8QQ, United Kingdom, and the  Novartis Horsham Research Centre, Wimblehurst Road, Horsham RH12 5AB, United Kingdom

Received for publication, May 16, 2002, and in revised form, July 15, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Phosphodiesterase 4D (PDE4D), part of the complex cAMP-specific PDE4 family, plays a pivotal role in the regulation of airway smooth muscle relaxation by catalyzing the hydolysis of cAMP. Its gene on chromosome 5q12 encodes 5 splice variants, which show tissue-dependent expression and regulation. The genomic arrangement of PDE4D was determined using in silico methods, and a putative promoter of one of the protein kinase A-activated, long isoforms, PDE4D5 was identified. Promoter-luciferase constructs, transiently transfected into a ₂ adrenoreceptor-expressing CHO-K1 cell line, were used to demonstrate that the PDE4D5 promoter up-regulated reporter gene expression in response to increased cell cAMP. Site-directed mutagenesis of the cAMP-response element (CRE) at position 201 identified this as the principal component of the mechanism underlying this cAMP responsiveness. In the second part of this study, cAMP-dependent induction of PDE4D5 transcript in primary cultured human airway smooth muscle cells (hASMs) was demonstrated using both qualitative reverse-transcriptase PCR and quantitative real-time PCR. Isolated PDE4D5 isoenzyme activity, measured after selective immunoprecipitation from hASMs, confirmed that this increase in expression led to an up-regulation of functional activity. We present evidence for cAMP-driven PDE4D5 up-regulation in hASMs and suggest a CRE-containing, isoform-specific promoter as the primary mechanism.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The ubiquitous cyclic nucleotide, cAMP, is a key second messenger in the signaling cascades controlling a multitude of vital cellular functions and responses (1, 2). Compartmentalized intracellular cAMP gradients are governed by a balance between its generation from ATP by the action of adenylyl cyclases and the rate of degradation. In contrast to the diverse stimuli responsible for production of cAMP (3), members of the multigenic phosphodiesterase (PDE)¹ family provide the sole means of degradation by catalyzing its hydrolysis to 5'-AMP.

At present the PDE superfamily is known to consist of 11 separate families classified according to substrate selectivity, inhibitor sensitivity, and sequence (4-8). Those involved in the hydrolysis of cAMP include PDE 1, 2, 3, 4, 7, 8, 10, and 11 (9) The expression of these families demonstrate tissue-specific patterns with multiple family members usually being present in each tissue type.

Tissue profiling of human airway smooth muscle cells has revealed the expression of the majority of PDE families. However, functional studies indicate that PDE 3 and 4 essentially co-regulate cAMP levels in this tissue type and are, therefore, important mediators of bronchial tone (10, 11). The rolipram inhibited PDE4 family has received most attention as a therapeutic target in airways disease. It provides the dominant phosphodiesterase activity in inflammatory cells such as neutrophils, eosinophils, and CD4+ lymphocytes, which are thought to be central to the pathogenesis of asthma and chronic obstructive pulmonary disease (12, 13). Recent clinical studies with selective inhibitors of the PDE4 family have demonstrated a broncho-relaxant action, making them an attractive prospective adjunctive therapy (14-19).

The PDE4 family consists of 4 subfamilies, A, B, C, and D, with each family member being encoded on separate genes localized to 3 chromosomes (8, 20-26). Molecular cloning of these genes has revealed a further level of complexity due to the presence of multiple splice variants. In the case of PDE4D, for example, there are 5 splice variations at the 5'-end of the gene, which give rise to 3 long, 1 short, and 1 supershort isoform (7, 27, 28).

There is an increasing body of evidence demonstrating functional differences between the isoforms. For example, differential effects of phosphorylation in the so-called Upstream Conserved Regions (UCRs) of PDE4D appear to confer isoform-specific functional regulation. Investigators have shown that phosphorylation of the serine 54 residue in UCR1 of the long isoforms by protein kinase A produces a 2-3-fold increase in the V_max of the isoform (29-32). ERK phosphorylation of the catalytic region also appears to have different effects depending on the UCRs present: the UCR1-UCR2 module present in the long isoforms directs inhibition, whereas the lone UCR2 of the short isoform directs activation, and the truncated UCR2 of the supershort isoform allows no effect on enzymatic activity of ERK phosphorylation (33, 34).

In addition to these regulatory differences between the isoforms, their distinct, cell type-specific patterns of intracellular distribution further strengthen the case for diverse functional roles (4, 6, 7, 12, 28, 35, 36). Each PDE4 isoform is characterized by its unique N-terminal region, which appears to play a key role in intracellular targeting (9). For example, the N-terminal region of PDE4D5 confers interaction with the signaling scaffold protein RACK1 (37); that of PDE4A4/5 with LYN SH3 domain (38, 39) and that of PDE4D3 with AKAPs (40, 41). These observations imply that distinct PDE4 isoforms may control specific pools of cAMP and hence selectively influence particular cellular processes. Indeed, distinct pools of cAMP have recently been identified in cardiac myocytes (42) where they are regulated by PDE activity. Thus manipulation of the expression profile of PDE4 isoforms provides a physiological and potential pharmacological mechanism of controlling cell cAMP signaling in cells.

One mechanism, which might potentially alter local PDE4 expression patterns, is transcriptional control. However, little is know about the promoters that control the expression of the complex multigenic PDE family, save for work on PDE3 (43), PDE4D1/2 short forms (44), and the PDE4A10 long form (45). Up-regulation of the short PDE4D1/2 isoforms by elevated cAMP levels has been suggested to contribute to the role that PDE4 isoforms play in cellular desensitization processes (44). The other route by which PDE4 isoforms contribute to cellular desensitization being via the PKA-mediated activated of PDE4 long isoforms (46). Because specific PDE4 isoforms can thus be expected to influence the nature of cAMP signaling in cells, it is important to identify and characterize the properties of the promoters that determine the expression of various PDE isoforms in particular cell types.

In this study we describe the identification of the putative PDE4D5 promoter and demonstrate its ability to up-regulate reporter gene expression in response to conditions that increase cell cAMP content. We also present evidence for physiological up-regulation of PDE4D5 isoform expression and activity in human airway smooth muscle cells (hASMs) exposed to equivalent conditions. This is of potential clinical importance as the up-regulation of this isoform by its substrate could provide a negative feedback mechanism functionally downregulating the bronchodilator effect of ₂ agonists in airways disease and thus diminishing the therapeutic efficacy of these agents.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In Silico Work-- Accession numbers for complete cDNA sequences of each of the 5 PDE4D splice variants (PDE4D1, U50157; PDE4D2, U50158; PDE4D3, 50159; PDE4D4, L20969, and PDE4D5, AF012073)² were compared with the entire unfinished human genome using the NCBI BLAST facility. Exact matches were obtained for all of the sequences from a range of contigs representing unfinished, unordered sequences of genomic DNA (AC026095, AC027322, and NT_006665). The verified cDNA sequences were used as templates to re-order the relevant sequences within these contigs, and thus elucidate the intron-exon arrangement of the PDE4D gene. Sequence immediately upstream from the start codons of each of the 5 splice variants was then assessed using on-line transcription factor binding site-based, promoter predictive shareware packages. These included Matinspector Professional (Genomatix), TESS, and TSSG.

Construct Creation-- The region containing the putative promoter region of PDE4D5, identified using in silico techniques, was amplified using standard PCR to generate an 1800 bp band. Amplification reactions were performed using the Expand High Fidelity PCR System (Roche Molecular Biochemicals) with the following primers: (forward, 5'-GCACACACATACCTTGCCACT-3'; reverse, 5'-TCACAAGGTCTTCGTTCAGC-3'). Products were separated by electrophoresis in an ethidium bromide-containing Tris acetate/EDTA 1% agarose gel. DNA was viewed under a low power UV light and bands excised using a sterile blade before purification from using StrataPrep^TM DNA Extraction Kit (Stratagene) as outlined in the manufacturer's instructions. The eluted product was used as a template for subsequent PCRs. At this stage a mismatched reverse primer (5'-TGTCTGCTGAGCCATGGTCCT-3') was designed to introduce a NcoI restriction site around the start codon of PDE4D5, and this was paired with two fully matched forward primers (long, GTTGTGCGCATGCTAACAGG; short, GCAGGGAAGTATTTCTCC) to create two different length amplicons. In these reactions a non-proofreading Taq polymerase was used to allow the creation of mismatches in the sequence. After an identical purification procedure, 7 µl of the eluted products were ligated into 1 µl of Promega PGEM T Easy vector with 1 µl of T4-ligase and its single use 10× buffer at 4 °C overnight. The resultant plasmids were then used to transform 50 µl of competent Escherichia coli K12 strain JM109 bacteria and grown in ampicillin-containing broth. Colonies with successfully ligated vectors were selected using the blue/white screen integral to the PGEM T Easy vector and grown in broths prior to extraction of vector DNA using the Wizard Mini-Prep kit (Promega). This was then digested, in the case of the short construct, using NcoI and a blunt-cutting XmnI (cut site within the forward primer) and the appropriate band separated by electrophoresis. The long construct was initially restricted with NcoI and a blunt cutting AviII again in its forward primer. A further restriction with Ksp632I was required to separate the band of interest.

These two putative promoter fragments of 621 and 1544 bp in length, respectively, were then purified and ligated into PGL3 enhancer luciferase reporter vector cut with NcoI and SmaI to give a complementary staggered cut immediately in front of the luciferase start codon and a blunt end encouraging proper orientation of the putative promoters in the reporter vector. 2 µl of purified DNA was mixed with 1 µl of pre-cut and purified vector DNA in the 10-µl reaction mixture, and ligation was carried out at 15 °C over 3 h. The vectors were again transformed into E. coli strain JM109 and grown as above. This time colonies were selected at random before culturing in broths and extracting vector DNA.

The broths were then streaked out to a single colony to ensure complete purification and regrown. After extraction and purification of vector DNA using Wizard Mini-Prep kits, the complete inserts were carefully sequenced using Promega internal vector primers (RV primer 3 (clockwise) and GL primer 2 (anti-clockwise)) to confirm that the inserted putative promoter sequence was of expected length and sequence and properly orientated. The same broths were then grown on a larger scale to allow extraction and purification of DNA using Wizard Maxi-Prep Kit. This method provided sufficient quantities of DNA for the planned transfection experiments. Prior to use, construct DNA concentration was analyzed by spectrophotometry.

Site-directed Mutagenesis-- Using the wild-type promoter constructs as template DNA a cAMP response element (CRE) site, identified by in silico analysis of the PDE4D5 promoter, was mutated using Stratagene QuikChange^TM site-directed mutagenesis kit. Complementary primers were designed to encode a 4 base pair mutation in the CRE sequence most proximal to the start codon at position 201 in Fig. 1B (5'-gTGACGTTt to gTGTGTCTt-3'; uppercase letters denote CRE sequence and bold letters represent the mutation that was introduced.) The sequence of the 39-bp forward primer used was 5'-GCTTGGGAGAGGAGGGGTGTGTCTTAATACGCTTCCGCG-3', and the reverse primer was exactly complementary.

View larger version (36K): [in this window] [in a new window]   Fig. 1.   Schematic representation of the genomic arrangement of the PDE4D gene in relation to the functional regions of the enzyme and PDE4D5 putative promoter sequence. A, upper line represents the section of genomic DNA of chromosome 5q12, which contains the PDE4D gene. Above the line is a scale indicating the size of the gene with an arbitrary start point. Below the line are labeled marks indicating the position of numbered exons within this genomic framework. Arrows from these numbers link to a schematic functional representation of the PDE4D isoforms depicted in decreasing order of length. The regions in dark gray (UCR1, UCR2, and the C-terminal) regulate the activity of the catalytic region (medium gray). Both intracellular targeting and regulatory properties are attributed to the isoform specific N terminals (light gray). B, is an annotated version of the putative PDE4D5 promoter sequence. 5'-ends of the long (1544 bp) and the short (621 bp) constructs are depicted by arrows above the matching sequence. Two mismatched base pairs (CC) denoted in a gray box were introduced in the reverse primer in order to create an NcoI staggered restriction site at the start codon. The ATG is represented in bold typeface. Black boxed regions with white font within the promoter denote cAMP response elements (CRE) and sequences in open boxes denote CCAAT enhancer binding proteins (C/EBP) predicted by transcription factor binding site analysis software.

From the short and long construct transformation plates produced according to the manufacturer's protocol, four colonies were selected, grown on, and re-streaked to enable selection of a single colony. Purified vector DNA samples were obtained by Wizard Mini-Prep of the broths grown from selected colonies were sequenced across the inserts to confirm both the presence of the desired mutation and conservation of the remaining sequence. At this point larger amounts of construct were prepared using Wizard Maxi-prep, and DNA concentrations were assayed by spectrophotometry.

Culture of CHO K1 Cells-- Chinese hamster ovary cells, stably transfected with both the human ₂ adrenoreceptor gene and the gene encoding human secreted placental alkaline phosphatase driven by a promoter containing six cyclic AMP response elements, were selected for the transfection experiments (47). The stably transfected system enabled a parallel assay of SPAP (proving that the cells were responding to stimulatory conditions and had the necessary cellular apparatus to produce a cAMP-dependent up-regulation of transcription), and luciferase to determine whether the transiently transfected putative PDE4D5 promoter construct could increase transcription of its reporter gene in response to the same conditions. These cells were randomly seeded in 24-well plates at a density created by the resuspension of a confluent 1 × 75 cm² flask in 100 ml of Hams F-12 Dulbecco's modified Eagles medium (DMEM) supplemented with 10% serum and glutamine and a volume of 1 ml per well. The following day the medium was removed, and the cells were serum-starved prior to transient transfection with the relevant constructs.

Transfection-- Transfection was performed in 90% confluent 24-well plates with each condition in triplicate per plate. FuGENE 6 (Roche) was used to transfect cells with a 40:1 ratio of test DNA and Renilla SV40 control vector. Plates were incubated in a humidified atmosphere of 5% CO₂, 95% air at 37 °C for 24 h. The wells were then washed twice with phosphate-buffered saline to remove any trace of the transfection mixture and glutamine, and serum-containing Hams F-12 DMEM was re-applied for 24 h to allow the cells to recover fully. At this point the medium was removed and replaced with 1 ml of serum-free Hams F-12 DMEM per well in preparation for the assay.

SPAP Assay-- Within 18 h of serum starving the transiently transfected CHO K1 cells, serum-free medium was replaced and the drugs added at the requisite dilutions. The conditions used included forskolin (10 µM final concentration), isoprenaline (1 µM), 3-isobutyl-1-methylxanthine (IBMX) (50 µM), actinomycin (4 µM), cycloheximide (50 µM), and 8 bromo-cAMP (1 mM). Any antagonists used were added 30 min prior to the relevant agonists. The plates were returned to the incubator for 5 h. At this point the medium was replaced with 300 µl per well of fresh serum-free medium and left for 1 h. 20 µl of the supernatant from each well was then aliquoted into a flat-bottomed 96-well plate with blank medium as controls. The 96-well plates were then heat-inactivated at 65 °C for exactly 30 min to remove endogenous alkaline phosphatase (SPAP is stable at this temperature).

1 ml of 100 mM p-nitrophenol phosphate (PNPP) was thawed and added to 20 ml of diethanolamine buffer immediately prior to the assay. 200 µl of this reaction mixture were then added to each well, and the plates kept for 35 min at 37 °C in room air in a light proof incubator. The timing of this incubation was determined by optimization experiments, which demonstrated that the maximal optical density below 1.5 (the range within which the relationship between absorbance and SPAP munits/ml remains linear) occurred between 30 and 40 min of incubation. Absorbance was measured on a Dynex Revelation 96-well plate reader at 405 nm and converted to munits/ml using the following equation: (munits/ml) = A/18.5tV; where A = optical density, t = time of incubation with PNPP, and V = volume of sample (20 µl in all experiments).

Luciferase Assay-- Luciferase activity in cell lysates was measured using the Dual-Luciferase Reporter assay system (Promega). After collection of the SPAP assay data, the plates were washed in phosphate-buffered saline and then lysed using 200 µl of 1× passive lysis buffer (supplied reagent). The plates were rocked for 20 min at room temperature prior to storage at 80 °C until it was convenient to perform the assay.

After preparation of the reagents according to the manufacturer's protocol, 20 µl of the thawed lysate from each well was mixed with 100 µl of the luciferase assay reagent II by vigorous pipetting. The firefly luciferase activity measurement (under the influence of the putative promoters) was then measured over 10 s by a luminometer. 100 µl of Stop & Glo reagent were then added and vortexed briefly to mix. This had the dual effect of quenching the initial reaction and catalyzing the bioluminescent reaction of Renilla luciferase. The latter is quantified by a 10-s luminometer measurement and used in the analysis to correct for well-to-well variations in transfection efficiency.

Human Airway Smooth Muscle Cell Culture-- hASMs were prepared from explants of trachealis muscle obtained from individuals free of respiratory disease within 12 h of death. A tracheal segment from immediately above the carina was removed and a strip of trachealis dissected clear of the surrounding tissue. Following initial washes with DMEM containing penicillin (200 units/ml), streptomycin (200 µg/liter), and amphotericin B (0.5 µg) the overlying mucosa was removed and 2-mm² sections of smooth muscle were excised and placed in 6-well plates. Once adherent, the explants were covered in DMEM containing the above anti-microbials, 10% fetal bovine serum, and glutamine (2 mM). The medium was regularly replaced until muscle cell growth occurred. At this point the culture medium was changed to DMEM supplemented with 10% fetal bovine serum and 2 mM glutamine. When the cells around the explant approached confluence, the explant was removed, and 24 h later the cells were harvested by trypsinization and plated out into a 25-cm² flask. Subsequent passages were performed in 75 and 162 cm² flasks using the same culture medium; cells were incubated at 37 °C in 5% CO₂ and 95% air. Kotlikoff and co-workers (48, 49) have extensively characterized the phenotype of these cells.

hASMs cultured in this way were used for both PDE activity assays and reverse-transcriptase PCR (RT-PCR). In order to standardize culture conditions for the two sets of experiments, cells destined for RNA extraction grown in 75 cm² were passaged at the same times as those grown in 162 cm² flasks for subsequent cell lysis. Exposure to the range of conditions described below was also carried out in parallel. Harvesting procedures are described in the relevant sections.

Cyclic AMP Assay-- Tritiated cAMP production was assayed using a previously described [³H]adenine prelabeling assay (50). Briefly, hASMs were grown in 24-well plates to confluence. For each plate 25 ml of Hanks HEPES Buffer (HHB) were warmed and mixed with 50 µl of tritiated adenine. The growth medium was then replaced with 1 ml per well of this mixture (2 µCi per well) and the cells incubated in air at 37 °C for 2 h. At the end of this period cells were washed in HHB before replacing the medium with HHB and adding 10 µl of antagonists in an appropriate concentrations to produce the final concentrations required. Cells were then incubated for 20 min prior to the addition of agonists. After a further 20 min (sufficient time to induce a significant increase in cAMP generation by hASMs in response to  agonist stimulation, Ref. 51) the reaction was stopped by lysing the cells with 50 µl of concentrated HCl and freezing for a minimum of 2 h at 20 °C.

Determination of cAMP formation in a lysed cell sample was achieved via a two-column separation system. Initially a 100-µl aliquot of the lysed sample was removed from each well and [³H]adenine counted to provide a correction factor for cell numbers per well. The lysate was then spiked with the same volume of ¹⁴C-labeled cAMP in medium.

Column chromatography was performed to separate tritiated cAMP from other adenine-based cellular products. In the first column dowex 50 was used to adsorb cAMP. The majority of adenine-based cellular components were washed through with a carefully estimated volume of water. The dowex columns were then placed over alumina columns, and cAMP was washed through with a larger volume of water. Both cAMP and any remaining ATP adsorb onto the neutral alumina, however, as cAMP binds less avidly, it could be washed into a scintillation vial with imidazole buffer. Radiation in elutants was then determined by dual scintillation counting (³H and ¹⁴C). Finally, cAMP concentration was calculated from the raw disintegrations per minute by correction both for cell number, using the [³H]adenine counts, and for variations in column recovery by measurement of the fraction of [¹⁴C]cAMP eluted from each column pair.

Reverse Transcriptase PCR-- Total RNA was extracted from the pelleted hASMs stored at 80 °C using Qiagen RNeasy Minikits. The samples were eluted in 50 µl of RNase-free water and stored at 80 °C. On thawing the samples underwent a 3-stage process of reverse transcriptase-PCR according to the manufacturer's instructions, which comprised: DNase digestion, cDNA synthesis (with a reverse transcriptase negative control), and PCR using previously validated primers specific to PDE4D5 (forward, TGCCAGCTGTACAAAGTTGACC; reverse, TTCTCGGAGAGATCACTGGAGA). Products were separated by electrophoresis in ethidium bromide-containing Tris acetate/EDTA-based 1% agarose gels. Visualization under UV light allowed qualitative analysis of the presence of transcript in each condition.

Real-time PCR-- Determination of the relative concentrations of PDE4D5 mRNA in extracts of total RNA from pelleted hASMs exposed to a range of stimulatory conditions was performed on Applied Biosystems Prism 7700 cycler using TaqMan reagents. Initially cDNA was created using methods described in the previous section. As advised in the manufacturer's protocol, primers and probes were designed using PrimerExpress v1.5 to ensure suitability for the TaqMan system and appropriate relative melting temperatures. Sequence from the PDE4D5-specific N-terminal was used to generate isoform-specific primers and a probe. The primer pair was designed to amplify a short amplicon to which a probe, labeled with FAM reporter dye at the 5'-end and a TAMRA quencher dye at the 3'-end, annealed. (PDE4D5 sense primer, TGAAGTGGATAATCCGCATTGT; antisense, GGTTTTCTCGCAAGGATTTCAC; and probe, CAAACCCGTGGCTGAACGAAGACC). Reactions were carried out in 25-µl volumes using 900 nM primer concentrations and 250 nM probe concentrations. TaqMan Universal Master Mix provided optimal buffer components. Relative quantitation required comparison to a housekeeping gene so each of the sample replicates was also run with Applied Biosystems predeveloped assay reagent mixture for detection of human GAPDH. GAPDH normalized results were then analyzed by the relative standard curve method as described in the ABI Prism User Bulletin. This involved running both test and control primers and probes against a standard sample in a range of dilutions and measuring the threshold cycle (Ct; this is the point at which exponential growth of the PCR product begins) for each point. This was then plotted against a log of the dilution factor to create a linear set of Ct data covering the range of Cts expected from the test samples. Using the equation of the linear trendline plotted through this the relative mRNA concentration in each test sample was calculated. Results were then normalized to GAPDH mRNA concentrations derived by the same technique and expressed as a percentage of the values obtained from the forskolin-stimulated samples. Each sample was assayed in triplicate. Mean results for each condition were then expressed as a percentage of the maximal (forskolin) response. This procedure was repeated for 3 biological samples to produce the mean and S.E. data presented in this study.

Immunoprecipitation and PDE Activity Assays-- After appropriate stimulation over 18 h, hASMs were washed with phosphate-buffered saline then harvested by lysis from the 162 cm² flasks. This was achieved using a disposable cell scraper and 300 µl of lysis buffer made up of 25 mM HEPES pH 7.5, 50 mM NaCl, glycerol to 10% volume, Triton X-100 to 1% volume, 50 mM NaF, 5 mM EDTA, and Complete protease inhibitor mixture (Roche Molecular Biochemicals) to 8% volume. The lysate was stored at 80 °C.

Prior to immunoprecipitation, nonspecific binding in the lysates (500 µg of protein per pull-down) was pre-cleared by incubation with 30 µl per sample of washed protein G-agarose bead slurry (Invitrogen) for 30 min at 4 °C. The beads were then removed by refrigerated centrifugation and the PDE isoforms immunoprecipitated from the precleared lysate by mixing end over end at 4 °C for 1 h with the appropriate antibody.

A 75-µl per sample aliquot of protein G-agarose beads was then prewashed with phosphate-buffered saline twice and with lysis buffer. The immune complexes were then mixed for 30 min at 4 °C with these beads before separation by centrifugation. The bead pellets were then washed with lysis buffer, KHEM (50 mM KCl, 10 mM EGTA, 50 mM HEPES, and 1.92 mM MgCl₂, pH 7.2 with added 0.001 mM dithiothreitol and Complete protease inhibitor mixture to 8% volume prior to use) and finally twice with PDE assay buffer (20 mM Tris, pH 7.4). Immunoprecipitation was performed in duplicate within each experiment.

PDE activity was assayed using a modification of the Thompson and Appleman (52) two-step procedure described previously by Houslay and co-workers (53). Briefly, a mixture of tritiated cAMP (Amersham Biosciences) and non-tritiated cAMP in a 20 mM Tris pH 7.4 + 10 mM MgCl₂ assay buffer was mixed with the bead pellets or lysates by vortexing. The tubes were then incubated at 30 °C for 20 min with frequent agitation. This created optimal conditions for the PDEs to catalyze hydrolysis of the phosphoester bond in cAMP. The reaction was stopped by plunging the tubes into boiling water for 3 min and then placing on ice for 10 min. 25 µl of venom from Crotalus Attrox (Sigma V7000) was then added and the samples incubated at 30 °C for a further 10 min. Dowex (Sigma 1X8-400) was prepared by addition of 1 liter of 1 M NaOH per 100 g of dowex powder, after which it was allowed to settle and the supernatant discarded. The dowex was then washed 30 times with dH₂O over 2 weeks or until pH 7.0 having been left at 4 °C to settle completely between washes. The dowex was then acidified with 1 liter of 1 M HCl and stirred for 15 min producing a color change from orange to yellow. Finally it was washed another five times in dH₂O until pH reached ~3. It was then stored in a 50:50 dowex/water mixture. Immediately prior to the assay 1 volume of 100% ethanol was added to 2 volumes of the dowex/water mixture to create a slurry used in the next step of the assay. At this point 400 µl of the dowex slurry were added to each Eppendorf tube, vortexed, and placed on ice for 15 min. During this period samples were inverted 5 times on four occasions to mix thoroughly. Samples were then vortexed before being spun at 13,000 rpm at 4 °C for 3 min. 150 µl of the supernatant fraction from each tube was then placed in a corresponding scintillation vial containing 10 ml of scintillant. Tritium radioactivity was counted for 1 min per vial on the beta counter.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Initial work was performed using the BLAST facility to define the genomic arrangement of the PDE4D gene situated on chromosome 5q12. This was found to be complex, comprising 17 exons spread over just under 1 Mb of genome (Fig. 1A). Exons encoding the C-terminal, catalytic region, upstream conserved region 2 and the N termini of the short and supershort isoforms (PDE4D1 and PDE4D2) were found to be clustered at one end of the gene. They are separated from a second cluster of 4 exons encoding upstream conserved region 1 by an intron of almost 160 kb. Beyond this are three exons encoding the N termini of the long isoforms (PDE4D3, 4, and 5); these are thought to play a role in subcellular localization (9, 37, 40, 54-57). As indicated in Fig. 1A and Table I these exons are situated a substantial genomic distance 5' to the UCR1 cluster. Comparison with other members of the PDE4 family and Rattus norvegicus PDE4D showed remarkable matching of exon length throughout the catalytic and regulatory regions. Direct comparison of amino acid and coding DNA sequences demonstrated striking homology between subfamilies and species reaching 90-95% in the catalytic and upstream conserved regions.

Having identified the intron-exon arrangement within the gene framework we were able to localize four putative intronic promoters upstream from the start codons for each of the five isoforms. Splice variants PDE4D1 and PDE4D2 share a transcription start site and therefore have the same putative promoter, which has been formerly characterized in the rat as a cAMP-responsive promoter containing no TATA box but a number of GC-rich regions, SP1 and activator protein (AP) 1 and 2 sites (44). Each of the other putative promoter regions was searched for transcription factor binding sites (TFBS) using independent promoter predictive shareware and we found that the putative promoter of PDE4D5 displayed a number of interesting characteristics including two putative cAMP response elements (CRE), illustrated in Fig. 1B, and a number of CCAAT enhancer-binding protein binding sites (C/EBP).

In order to establish the properties of this putative promoter experimentally, two constructs of 621 bp (4D5SPluc) and 1544 bp (4D5LPluc) in length containing intronic DNA immediately 5' to the PDE4D5 start codon, were ligated into the PGL3 enhancer vector. These constructs were then studied in a CHO-K1 cell line, which had been stably transfected with both the human 2 adrenoreceptor (300 fmol/mg of protein), and a secreted placental alkaline phosphatase reporter gene driven by a strongly cAMP responsive promoter containing 6 CREs. 4D5SPluc up-regulated basal expression of luciferase by 8.7 ± 0.2-fold over the vector lacking the promoter element (n = 4; p < 0.001 ANOVA); the long construct by 26.0 ± 0.4-fold (n = 4; p < 0.001 ANOVA) as shown in Fig. 2. Similar data were obtained when constructs were transfected into BEAS-2B, a bronchial epithelial cell line: 4D5SPluc up-regulated basal expression of luciferase by 10.8 ± 0.7-fold over the empty vector (n = 4; p < 0.001 ANOVA); 4D5LPluc by 27.4 ± 2.0-fold (n = 4; p < 0.001 ANOVA).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Unstimulated activity of PDE4D5 promoter constructs transiently transfected into a CHO cell line. Results were expressed as fold over the empty enhancer vector and represent mean ± S.E. of four experiments carried out in triplicate. The short promoter construct (4D5SPluc) increased unstimulated reporter gene expression by 8.7 ± 0.2-fold over the empty vector (n = 4; p < 0.001 ANOVA) and the long construct by 26.0 ± 0.4-fold (n = 4; p < 0.001 ANOVA). Data from the PGL3 control vector, which contains a strong SV40 promoter, are included as a positive control.

Having confirmed the strong basal promoter activity of both of these constructs, we went on to assess their responsiveness to agents that increased intracellular cAMP. Initially we utilized the SPAP reporter under the direct control of a 6-CRE promoter, which had been stably transfected into the CHO-K1 cell line to demonstrate that these cells had the machinery required to effect an up-regulation of gene transcription when stimulated with agents that increase cAMP. The data shown in Fig. 3A demonstrate that SPAP output was unaffected by either the process of transient transfection or the plasmid introduced. Transfection of either the empty enhancer vector, the test constructs or the PGL3 control vector all generated a similar level of SPAP production between 1.0 ± 0.0 and 1.1 ± 0.1-fold changes in comparison to unstimulated, non-transfected control (NTC) cells (n = 4; p > 0.05 ANOVA). There were increases in SPAP output from stimulated cells independent of the transiently transfected construct. The following data relate to SPAP output from all cells exposed to each stimulatory condition regardless of the transiently transfected construct. 10 µM forskolin, a direct adenylyl cyclase activator, up-regulated SPAP output by 2.9 ± 0.1-fold over unstimulated NTC cells; 1 µM isoprenaline, a non-selective  agonist, by 4.0 ± 0.1-fold; 50 µM 3-isobutyl-1-methylxanthine (IBMX), a non-selective PDE inhibitor, by 1.8 ± 0.1-fold; and the combination of IBMX and isoprenaline by 4.1 ± 0.2-fold. All of these increases were highly significant (n = 7-8; p < 0.001 ANOVA).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   The effect of increased cell cAMP and mutation of the putative cAMP response element at position 201 the regulatory influence of the PDE4D5 promoter constructs expressed in a CHO cell line. CHO cells, stably transfected with human 2 adrenoceptors and a SPAP reporter driven by multiple cAMP response elements, were transiently transfected to express a firefly luciferase gene. Luciferase was under the control of either the wild-type or the CRE-mutated PDE4D5 promoter constructs, 4D5SPluc (621 bp) or 4D5LPluc (1544 bp). The cells were challenged with 10 µM forskolin (a direct adenylyl cyclase activator), 1 µM isoprenaline (a non-selective  agonist), 50 µM IBMX (a non-selective PDE inhibitor) or a combination of isoprenaline and IBMX as described in "Experimental Procedures." Both SPAP assays and luciferase assays were performed. The data shown in A demonstrate that all unstimulated cells had consistent SPAP output at the level of the non-transfected control (NTC) cells, independent of the construct with which they were transiently transfected. The degree of up-regulation of SPAP production induced on stimulation was dependent only on the agent used, not the test construct. B, having demonstrated that the cell line was able to effect an up-regulation of gene transcription under stimulatory conditions, which increase cell cAMP, we tested the cAMP responsiveness of the PDE4D5 promoter constructs. Bars on the left represent Renilla-corrected luciferase production from CHO cells transfected with PGL3 Enhancer, PGL3 Control, and then 4D5SPluc under a range of stimulatory conditions marked in the table below all expressed as mean fold ± S.E. over unstimulated 4D5SPluc. On the right hand side are the equivalent data for the longer construct, 4D5LPluc. Identical results for PGL3 Enhancer and PGL3 Control have been expressed as fold over 4D5LPluc to demonstrate that the absolute activity of the 4D5LPluc remained greater than 4D5SPluc. In all cases fold increases reached significance (n = 4; p < 0.001 ANOVA). C, in order to establish the mechanism of this cAMP-dependent up-regulation we compared wild-type constructs with those whose CRE at position 201 had been mutated. In a separate series of experiments, cells transfected with all four constructs were challenged with forskolin or isoprenaline. The left hand bars represent fold over unstimulated wild-type 4D5SPluc. In basal conditions the 201 CRE mutation construct had no significant effect on luciferase production. Forskolin stimulation produced a 1.79 ± 0.03-fold increase from wild-type 4D5SPluc and a 0.99 ± 0.04-fold change from the CRE-mutated 4D5SPluc (n = 4; p < 0.05 ANOVA). Isoprenaline produced a 2.03 ± 0.14-fold increase with the wild-type construct and a 0.93 ± 0.06-fold change with the mutated construct (p < 0.001). The bars on the right hand side follow the same pattern for the long construct with 1.64 ± 0.12-fold (wild-type) compared with 0.99 ± 0.06 (CRE mutated) on forskolin stimulation and 2.00 ± 0.08 (wild-type) compared with 1.05 ± 0.03 (mutated) with isoprenaline (n = 4; p < 0.001 ANOVA).

Having confirmed the capacity of this cell line to up-regulate gene expression in response to stimulatory conditions effecting an increase in cellular cAMP, we proceeded to investigate the potential for identical conditions to regulate the PDE4D5 promoter. After removal of the supernatant for SPAP assay, the cells were lysed, and luciferase expression measured. This was under the control of either the 621 bp (4D5SPluc) or the 1544 bp (4D5LPluc) PDE4D5 promoter constructs. For the short construct, forskolin stimulation produced a 1.63 ± 0.04-fold increase; isoprenaline, 1.85 ± 0.06; IBMX, 1.54 ± 0.03; and the combination of isoprenaline and IBMX, 1.78 ± 0.07-fold, with all results expressed as mean ± S.E. of fold over the unstimulated 4D5SPluc. For the long construct, 4D5LPluc, there were similar increases in luciferase expression upon stimulation when results were expressed as fold over unstimulated construct: forskolin generated a 1.59 ± 0.07-fold increase; isoprenaline, 2.03 ± 0.10; IBMX, 1.51 ± 0.08; and isoprenaline and IBMX, 2.06 ± 0.12. The inherent increased promoter activity of 4D5LPluc when compared with 4D5SPluc, however, remained apparent. This is demonstrated by the relative sizes of the bars representing PGL3 Enhancer and PGL3 Control on either side of Fig. 3B. In all instances the increased expression over unstimulated constructs were statistically significant (n = 4; p < 0.001 ANOVA).

As the short and long constructs responded similarly to cyclic AMP we hypothesized that the CRE nearest the start codon (position 201 in Fig. 2A), which is present in both, might be responsible for mediating the up-regulation, rather than the more distant CRE at position 1390. To evaluate this theory we used site-directed mutagenesis to alter four of the base pairs within the CRE sequence of both 4D5SPluc and 4D5LPluc from the native gTGACGTTt (uppercase letters represent the most consistent CRE sequence from published data) to gTGTGTCTt. The mutation was designed to maintain GC content and thus minimize the effect it would have on secondary structure. Luciferase expression data from transient transfection of both the mutated and wild-type constructs into the same CHO cell line confirmed that the CRE at position 201 was an important mediator of cAMP-inducibility in the PDE4D5 promoter.

Under basal conditions, mutation of the 201 CRE element had no significant effect on reporter gene activity. Forskolin stimulation, however, produced a 1.79 ± 0.03-fold increase from wild-type 4D5SPluc compared with 0.99 ± 0.04-fold change from the CRE-mutated 4D5SPluc, and isoprenaline produced a 2.03 ± 0.14-fold increase with the wild-type construct and a 0.93 ± 0.06-fold change with the mutated construct (n = 4; p < 0.05 and <0.001 respectively ANOVA). Although the absolute activity of the long construct was greater, when results were expressed as fold over unstimulated 4D5LPluc, the pattern of changes was similar. Under basal conditions there was reduced expression of luciferase driven by the mutated construct at 0.79 ± 0.03, which failed to reach significance. Forskolin challenge increased luciferase expression in wild-type 4D5LPluc by 1.64 ± 0.12-fold compared with 0.99 ± 0.06 from its CRE-mutated counterpart; equivalent isoprenaline-stimulated values were 2.00 ± 0.08 (wild-type) and 1.05 ± 0.03 (mutated). In both cases the effect of the mutation was significant (n = 4; p < 0.001 ANOVA). Results for the short and long constructs are presented on the left and right sides of Fig. 3C, respectively.

The relative inability of the constructs whose CRE had been mutated to up-regulate luciferase expression in conditions of raised cAMP, supported the hypothesis that this CRE played an important role in the mechanism of cAMP responsiveness of the PDE4D5 promoter. Having demonstrated cAMP inducible up-regulation of PDE4D5 promoter activity, we aimed to investigate whether this effect was reflected in the major target cells within the airway, hASMs, as this would be of potential clinical significance.

First we designed a set of experiments to quantify the increase in intracellular cAMP induced by equivalent conditions to those used in the transfection experiments. We also studied the effects of 8-bromo-cAMP (1 mM), a cell-permeable analogue of cAMP able to activate protein kinase A, actinomycin (4 µM) as an inhibitor of transcription, and cycloheximide (50 µM) as an inhibitor of protein synthesis. These were then used in later experiments to confirm that any increases seen in PDE4D5 transcript or activity were due to newly synthesized enzyme and not post-translational effects or a direct activation by cAMP.

Fig. 4, shows the effect of the agents studied on [³H]cAMP production in hASMs. As expected, 8-bromo-cAMP, actinomycin, and cycloheximide had no significant effect upon basal or forskolin-stimulated cAMP formation. Forskolin alone produced a 13.0 ± 0.7-fold increase in cAMP production over basal (n = 4; p < 0.001 ANOVA) in keeping with its ability to directly activate adenylyl cyclase. Isoprenaline produced a 5.0 ± 0.3-fold increase over basal (n = 4; p < 0.001 ANOVA). While IBMX alone was responsible for only a 1.4-fold increase in cAMP production that failed to reach statistical significance, it increased the response to isoprenaline to 1.9 ± 0.1-fold (n = 4; p < 0.001 ANOVA) compared with isoprenaline alone. The relatively small response to IBMX alone probably reflects the low levels of basal cAMP turnover in this cell type and has been noted in previous work (50).

View larger version (38K): [in this window] [in a new window]   Fig. 4.   The effect of pharmacological stimulation on [3H]cAMP production in primary cultured human airway smooth muscle cells. Results are expressed as fold over [3H]cAMP production in unstimulated hASMs and represent the mean ± S.E. of n = 4 experiments, each measured in triplicate. 8-bromo-cAMP (a cAMP analogue), actinomycin (used as a transcription inhibitor), and cycloheximide (as an inhibitor of translation) had no significant effect upon basal cAMP formation. Forskolin alone produced a 13.0 ± 0.7 (p < 0.001 ANOVA) fold increase in cAMP production over basal: this was unaffected by actinomycin or cycloheximide. Isoprenaline produced a 5.0 ± 0.3-fold increase over basal (p < 0.001 ANOVA), while IBMX (a non-selective phosphodiesterase inhibitor) produced a 1.4 ± 0.1-fold increase in cAMP production (non-significant). The combination of IBMX and isoprenaline produced a 9.5 ± 0.9-fold increase (p < 0.001 ANOVA).

Using primers designed to amplify a sequence within the PDE4D5 specific N-terminal, we next performed RT-PCR experiments on total RNA extracts from hASMs pretreated with all of the drug combinations listed above for 6 h (antagonists were added 15 min prior to agonists in all cases). Typical gels are depicted in Fig. 5A. Induction of PDE4D5 in hASMs stimulated with forskolin was demonstrated by the appearance of bright bands not present under unstimulated conditions. This forskolin-induced up-regulation was abolished by actinomycin, an inhibitor of transcription, but not cycloheximide, an inhibitor of translation indicating transcriptional up-regulation. Further bands demonstrate that the cell-permeant cAMP analogue, 8-bromo-cAMP, isoprenaline, and IBMX were also able to induce PDE4D5 transcription in hASMs.

View larger version (37K): [in this window] [in a new window]   Fig. 5.   The effects of increased cell cAMP on PDE4D5 transcript expression and functional activity in primary cultured airway smooth muscle cells. A, RT-PCR reactions using primers designed to amplify a region of the PDE4D5 isoform-specific N-terminal were performed on total RNA extracts from primary cultured human airway smooth muscle cells that had been stimulated for 6 h by a range of agents. Resultant cDNA was run adjacent to a reverse transcriptase negative (RT-) control in paired lanes of ethidium bromide-containing Tris acetate/EDTA-based 1% agarose gels and visualized under UV light. All of these experiments were run in triplicate; typical gel photographs are displayed. The left hand gel demonstrates induction of PDE4D5 transcript with 10 µM forskolin, which was abolished by 4 µM actinomycin but not 50 µM cycloheximide. The final lane in the left hand gel and those in the right hand gel indicate that 1 mM 8-bromo-cAMP, 1 µM isoprenaline, and 50 µM IBMX also induce expression of PDE4D5. B, quantitation of transcriptional up-regulation by real-time PCR. Data from n = 3 biological samples assayed in triplicate are shown as cDNA concentration calculated by the relative standard curve method and normalized to GAPDH (see "Experimental Procedures"), expressed as a percentage ± S.E. of the forskolin response. ANOVA with Bonferroni's post-test indicates that all conditions except actinomycin alone and forskolin and actinomycin showed highly significant increases over basal (all p < 0.001). Bars have been plotted in an equivalent position to those in C to allow easy comparison. C, measurement of isoform-specific functional activity for PDE4D5. Data are expressed as fold over basal. The bars represent a mean ± S.E. of n = 3 experiments each assayed in triplicate. 8-bromo-cAMP, forskolin, IBMX, isoprenaline, and a combination of the latter two all produced significant increases over unstimulated hASMs (p < 0.001 ANOVA). The response to forskolin was fully abrogated by actinomycin (p < 0.001) and substantially reduced by cycloheximide (p < 0.001) suggesting that the increase in activity is due to transcriptional up-regulation.

As the RT-PCR data were qualitative, real-time PCR, using the TaqMan system, was performed to quantify the differences in mRNA expression in hASMs stimulated under these different conditions. Mean template levels generated from 3 biological samples representing each of the stimulatory conditions and expressed as a percentage of levels in the forskolin-stimulated samples were as follows: basal hASMs 0.6 ± 0.4 (mean percentage of forskolin ± S.E.); 8-bromo-cAMP, 65.2 ± 4.8; isoprenaline, 70.2 ± 5.6; IBMX, 39.3 ± 8.3; isoprenaline and IBMX, 100.0 ± 6.2; actinomycin, 1.0 ± 0.1; and forskolin and actinomycin had 1.7 ± 0.4 (Fig. 5B). Statistical analysis using ANOVA with Bonferroni's post-test indicated that all except actinomycin and forskolin and actinomycin showed significant increases over basal (p < 0.001 in all cases). Once again forskolin-induced up-regulation of PDE4D5 mRNA expression was fully abrogated by actinomycin.

Having demonstrated up-regulation of PDE4D5 transcripts by agents able to elevate cell cAMP content, we went on to study the potential for this up-regulation to alter the functional PDE4D5 activity. We sought to quantify this effect by direct assay of isoform-specific PDE activity. The PDE4D5 isoform was immunopurified from whole cell lysates by immunoprecipitation using a PDE4D5-specific antiserum. hASMs were pretreated with the same range of agents at doses previously listed, although stimulation for this set of experiments was over an 18-h time course. PDE4D5 was immunopurified and its activity determined over the indicated condition (Fig. 5C). Reflecting the RT-PCR data, all conditions that up-regulated transcript expression also increased functional PDE4D5 isoform-specific activity. Specifically isoprenaline increased PDE4D5 activity by 4.5 ± 0.4-fold over unstimulated hASMs, forskolin by 4.3 ± 0.2-fold, IBMX by 5.0 ± 0.3-fold, and 8-bromo-cAMP by 4.1 ± 0.3 (n = 3: p < 0.001 ANOVA). The forskolin induced up-regulation in PDE4D5 activity was fully abrogated by actinomycin, and partially abrogated by cycloheximide 50 µM. These results, together with the PCR data, suggest not only that increased de novo synthesis of the isoenzyme, rather than its activation by PKA phosphorylation, is responsible for the increased activity, but also that this up-regulation of synthesis occurs at the level of transcription.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We identify here the putative promoter for PDE4D5 and demonstrate for the first time cAMP induction of this long isoform. The genomic arrangement of the human PDE4D gene on chromosome 5q12 is complex, containing 2 major clusters of exons encoding the highly conserved regulatory and catalytic regions of the subfamily, separated by relatively short intronic sequences. Spread over almost 700 kb in the region of genomic DNA 5' to these clusters are the three exons that encode the isoform-specific N-terminals of the long forms responsible for their intracellular targeting and regulation. We investigated the promoter potential of the 1544 bp genomic sequence immediately upstream from the PDE4D5 N-terminal exon.

As with the active, cAMP-responsive promoter controlling the expression of isoforms PDE4D1 and 2 in rat (44), this putative promoter was atypical in that it contained no TATA box. Analysis of the transcription factor binding sites identified within this sequence suggested that it also had the potential to respond to cAMP. We identified two CREs at positions 201 and 1390. Data generated from the transfection of putative promoter constructs demonstrate cAMP-driven up-regulation of reporter gene expression. The magnitude of this effect, relative to basal promoter activity, was equal in both short and long constructs. Deletion of the CRE element situated 201 bp 5' to the start codon of PDE4D5, and thus contained in both constructs, had no significant effect on basal luciferase expression but caused complete abrogation of the up-regulation of expression when stimulated by forskolin or isoprenaline. These data confirm the involvement of the 201 CRE in the mechanism underlying the cAMP-responsiveness of the PDE4D5 promoter.

Having identified the isoform-specific PDE4D5 promoter and defined the mechanism of its cAMP responsiveness in a transformed cell system, in the second part of this article we proceeded to demonstrate cAMP-dependent up-regulation of both mRNA expression and catalytic activity of PDE4D5 in hASMs. Various investigators have demonstrated that prolonged elevation of intracellular cAMP increased the expression of PDE4D short isoenzymes (4D1 and 4D2) in rat Sertoli cells (58), T lymphocytes (59), and in human cultured myometrial cells (60). This is the first demonstration, however, of cAMP-dependent induction of expression of a long form, specifically PDE4D5. Given that the long isoforms have been shown to be activated by PKA (9), this finding has potential importance.

Total abrogation of the cAMP-induced up-regulation of both mRNA expression and functional activity by actinomycin demonstrated that these effects were a direct result of increased message expression and de novo synthesis of the enzyme, rather than the phosphorylation and consequent activation of PDE4D5 by PKA. There are two potential explanations for the apparent lack of effect on enzyme activity of phosphorylation: first, this could be due to the chronicity of the experiment such that over the 18-h time course the transient and acute influence of phosphorylation receded back to basal levels. Alternatively, under resting conditions the short form PDE4 activity may predominate and thus the only PKA activation will occur on induction of the long PDE4D5.

We have, therefore, demonstrated cAMP-driven PDE4D5 up-regulation in hASMs and presented evidence for a CRE-containing, isoform-specific promoter as the primary mechanism of this up-regulation.

The particular physiological consequences of increased PDE4D5 expression depend, in part, upon the subcellular distribution of this isoform. Previous work has demonstrated that it could be found both in cytosolic and particulate fractions of the cell (56) but this diversity betrayed no isoform-specific function. Recently, however, PDE4D5, uniquely among the subfamily, has been shown to bind specifically and with high affinity to the RACK1 WD-repeat scaffold protein via a single region of its unique N terminus (amino acids 22-26) (37). Though the functional significance of this association is unknown, it is tempting to speculate that RACK1 "recruits" PDE4D5 to regulate local cAMP concentrations.

Further work is needed to identify which subcellular pools of cAMP are controlled by PDE4D5 if we are to understand its functional significance more completely. However, given the importance of intracellular cAMP content in the control of airway smooth muscle tone, these findings may have clinical relevance in that this mechanism would provide a negative feedback, which might limit the effect of agents targeted at elevating cell cAMP content (61).

	ACKNOWLEDGEMENT

We thank Professor Stephen Hill (Institute of Cell Signaling, Nottingham University) for the kind donation of the stably transfected CHO-K1 cell line used in this work.

	FOOTNOTES

^* This work was supported by a program grant from the Medical Research Council (MRC) and The Wellcome Trust.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Recipient of a Medical Research Council Clinical Training Fellowship.

To whom correspondence should be addressed: Division of Therapeutics, C Floor South Block, University Hospital, Nottingham NG7 2UH, United Kingdom. Tel.: 44-115-9709905; Fax: 44-115-9422232; E-mail: ian.hall@nottingham.ac.uk.

Published, JBC Papers in Press, July 16, 2002, DOI 10.1074/jbc.M204832200

² Relevant GenBank^TM accession numbers are as follows: PDE4D1, U50157; PDE4D2, U50158; PDE4D3, 50159; PDE4D4, L20969; PDE4D5, AF012073; genomic fragments, AC026095, AC027322, and NT_006665.

	ABBREVIATIONS

The abbreviations used are: PDE(s), cyclic nucleotide phosphodiesterase(s); UCR, upstream-conserved region; RACK1, receptor for activated C-kinase; AKAP, a-kinase anchoring protein; PKA, cAMP-dependent protein kinase; hASMs, primary cultured human airway smooth muscle cells; TFBS, transcription factor binding sites; CRE, cAMP response element; CHO, Chinese hamster ovary; DMEM, Dulbecco's modified Eagle's medium; SPAP, secreted placental alkaline phosphatase; RT-PCR, real-time PCR; HHB, Hank's HEPES buffer; DEPC, diethylene pentaacetic acid; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; C/EBP, CCAAT enhancer-binding protein; ANOVA, analysis of variance; IBMX, 3-isobutyl-1-methylxanthine; ERK, extracellular-regulated kinase.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES