vascular smooth muscle tone is regulated by a number of physiological stimuli, many of them converging on the intracellular messengers Ca²⁺ and cAMP. Intracellular Ca²⁺, as a global signal (18, 19), leads to contraction of vascular smooth muscle cells (VSMCs) by binding to calmodulin and subsequent activation of myosin light chain kinases (36). A prominent mechanism for Ca²⁺ release in smooth muscle cells is receptor-mediated activation of the inositol trisphosphate (IP₃) pathway (2, 5). In contrast, cAMP relaxes VSMCs primarily by activation of protein kinase A (PKA, Ref. 1), thereby leading to a decrease in the Ca²⁺ sensitivity of the contractile apparatus (30). It is well known that β-adrenergic receptors induce vasorelaxation by coupling to the above-mentioned cAMP signaling pathway in vascular tissues (29). In the present study we addressed the question of whether and how these two pathways might be connected. Substantial evidence exists for several cell types that intracellular Ca²⁺ can regulate intracellular cAMP levels via activation or inhibition of certain adenylyl cyclases (ACs) or phosphodiesterases (PDEs). Ca²⁺ concentrations below 50 nM (resting state) are not known to substantially regulate ACs nor PDEs. However, elevations of cytosolic Ca²⁺ to up to 2 μM, for instance by stimulation of the G_q-PLC pathway, potentially lead to direct inhibition of AC5 and 6 (4, 6, 41) and to activation of AC1, 3, and 8 via formation of a Ca²⁺/calmodulin complex (40). In addition Ca²⁺/calmodulin may activate PDE1 resulting in an increased degradation of cAMP (28, 32, 33). Ca²⁺ concentrations > 10 μM, which we consider physiologically not to be relevant, have been demonstrated to directly inhibit all subtypes of ACs (10).

Noting that both Ca²⁺ and cAMP signals are highly dynamic (23, 39), we sought to analyze the crosstalk between Ca²⁺ and cAMP with high temporal resolution in intact cells by combining a fluorescence resonance energy transfer (FRET)-based cAMP sensor Epac1-camps (27) with Ca²⁺ measurements using Fura 2 (31). Intracellular Ca²⁺ concentrations in VSMCs were increased by stimulation with uridine diphosphate (UDP), most likely via activation of endogenous G_q-coupled P2Y₆ receptors (3, 7, 21). We observed a robust decline of cAMP upon exposure to UDP. The results of our study suggest that UDP-activated purinergic receptors antagonize G_s-mediated increases in cAMP primarily via Ca²⁺-mediated inhibition of AC5 and 6.

MATERIALS AND METHODS

VSMC preparation.

VSMCs were isolated from mouse aortas as described by Kühbandner et al. (22) with minor modifications. Whole aortas were excised from 5- to 20-wk-old mice and dissected free from connective tissues in phosphate-buffered saline (PBS). Subsequently, 4 to 7 aortas were digested for 45 min in 0.5 ml of enzyme solution 1, containing 0.7 mg/ml papain (from Papaya Latex, Sigma-Aldrich, St. Louis, MO), 1 mg/ml BSA, and 1 mg/ml DTT. Afterward, the tissue was incubated in 0.5 ml of enzyme solution 2 containing 1 mg/ml hyaluronidase (Applichem, Darmstadt, Germany), 1 mg/ml collagenase (from Clostridium histolyticum, Type H, Sigma-Aldrich), and 1 mg/ml BSA for 20 min. Both solutions were based on a Ca²⁺-free medium containing (in mM) 85 sodium glutamate, 60 sodium chloride, 10 HEPES, 5.6 potassium chloride, and 1 magnesium chloride, pH 7.4. After the second digestion, the resulting cell suspension was incubated in 10 ml of cell culture medium (Dulbecco's modified Eagle's medium, PAN Biotech, Passau, Germany, supplemented with 10% fetal calf serum, 2 mM glutamine, 0.1 mg/ml streptomycin, and 100 U/ml penicillin) and centrifuged for 5 min at 174 g. After removal of the supernatant, cells were resuspended in cell culture medium (1 ml per aorta) and plated onto cell culture dishes. Smooth muscle cells were cultivated for 1 wk. Medium was changed on days 1 and 4 after preparation. All experimental protocols involving animals were approved by the animal care committee of the University of Wuerzburg and are in accordance with the European coummunities council directive (86/609/EEC).

Smooth muscle cell transfection.

For cAMP measurements, cells were transiently transfected with a plasmid encoding Epac1-camps alone or in combination with AC4 or AC6 plasmids using electroporation with the Amaxa Nucleofector and the Basic Kit for Primary Smooth Muscle Cells according to the manufacturer's instructions (Lonza, Basel, Switzerland). After transfection, cells were plated on 0.1% gelatine-coated glass coverslips (24 mm). To promote cell adhesion, one drop of cell suspension was placed on a coverslip for 1 h before 2 ml serum-free DMEM were added. The total amount of transfected DNA was 5 μg per 500,000 cells. Transfection efficiency [determined by yellow fluorescent protein (YFP) fluorescence] was between 30 and 60%. Measurements were performed 24 h after transfection. Thirty minutes before the experiments, 0.5 U/ml adenosine deaminase (Roche Pharma, Grenzach-Wyhlen, Germany) were added to the medium to prevent increases of basal intracellular cAMP levels by activation of adenosine receptors.

Small interfering RNAs (siRNAs) against AC5 and 6 (siGENOME SMARTpools M-051739-00-0010 and M-043463-00-0010, Dharmacon, Lafayette, CO) were transfected using the Dharmafect2 siRNA Transfection Reagent (Dharmacon) according to the manufacturer's instructions. To silence AC5 and 6 in VSMCs, we transfected 60 pmol of an siRNA pool against AC5 together with 60 pmol of a siRNA pool against AC6. Transfection efficiency was monitored by labeling siRNAs with Silencer siRNA Labelling Kit-Cy3 (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions.

VSMCs were transfected with Epac1-camps 24 h after siRNA transfection and used for cAMP measurements after another 24 h. For RT-PCR experiments, cells were used 48 h after siRNA transfection. When indicated, the cytosol of VSMCs was buffered with BAPTA by a 30- to 90-min preincubation with 10 μM BAPTA-AM.

HEK-TsA201 cell transfection.

Transfection of HEK-TsA201 cells was performed as follows: 1,250 μl serum-free DMEM, 1 μg Epac1-camps cDNA, and 25 μg polyethyleneimine (Sigma-Aldrich) were mixed and incubated at room temperature for 30 min. They were then added to fresh serum-containing DMEM in a 10-cm cell culture dish with cells grown to about 70% confluency (16). The medium was changed 16–18 h after transfection, and the measurements were performed 24 h later.

FRET measurements in vitro.

In vitro FRET measurements were performed as described by Nikolaev et al. (27) with minor modifications. Forty-eight hours posttransfection, HEK-TsA201 cells were washed three times with ice-cold PBS, scraped from the plate, and resuspended in 5 mM Tris·HCl, pH 7.3, 2 mM EDTA. After disruption with an Ultraturrax device for 40 s on ice and 20 min centrifugation at 280,000 g, fluorescence emission spectra of the supernatant were measured using a fluorescence spectrometer LS50B (PerkinElmer Life Sciences; excitation at 436 nm, emission range 460–550 nm) before and after adding defined concentrations of cAMP (Sigma). cAMP saturation curves were plotted using Prism 4.0 (GraphPad software, San Diego, CA).

cAMP measurements in intact cells.

Cells were kept in external solution containing (in mM) 137 NaCl, 5.4 KCl, 2 CaCl₂, 1 MgCl₂ and 10 HEPES, pH 7.3. Unless otherwise indicated, cells were continuously superfused with external solution. Agonists were added using a rapid solenoid-valve superfusion device (ALA Scientific Instruments, Westbury, NY), which allowed switching solutions within 5–10 ms.

Fluorescence microscopy was performed using an Axiovert 200 inverted microscope (Carl Zeiss, Jena, Germany) equipped with a ×63 oil immersion objective and a Polychrome IV light source (TILL Photonics, Gräfelfing, Germany). Single cell fluorescence was recorded using a dual-emission photometric system (TILL Photonics). Illumination time was ≤100 ms at a frequency of 1 Hz. The excitation wavelength was set to 436 ± 10 nm (beam splitter dichroic long pass 460 nm, DCLP460), and emission was recorded from single cells at 535 ± 15 (F_YFP) and 480 ± 20 nm (F_CFP; DCLP505). Special care was taken to ensure that fluorescence levels and distribution of Epac1-camps were similar in all examined cells. F_YFP was corrected for direct excitation and bleed through as described (13). FRET ratios were calculated as ratio of corrected YFP over CFP emissions (F_YFP/F_CFP).

Ca²⁺ measurements.

VSMCs, seeded on coverslips, were washed twice with PBS and incubated with 2 μM Fura 2-AM in PBS for 45 min at 37°C. Subsequently, cells were washed with PBS and incubated for another 45 min at room temperature in the dark. Then the buffer was removed and cells were stored up to 90 min at 37°C in serum-free DMEM. For Fura 2 measurements, cells were imaged with a CoolSNAP-HQ CCD camera and an Axiovert 200 inverted microscope (Carl Zeiss). Image acquisition was performed using Metafluor software (Visitron Systems, Puchheim, Germany). Excitation wavelength was switched between 340 nm and 380 nm (beam splitter DLCP400, emission filter 510 ± 20 nm, AHF Analysentechnik, Tübingen, Germany) by using the Polychrome IV monochromator (TILL-Photonics). Illumination time was 60 ms and images were taken every second. Data were corrected for photo bleaching.

Simultaneous measurements of Ca²⁺ and cAMP.

VSMCs were transfected with Epac1-camps as described and labeled with Fura 2-AM 18–24 h later. For fast sequential imaging of Fura 2 fluorescence and FRET ratios of Epac1-camps, a DCLP455 dichroic mirror was used. Emissions of both Fura 2 fluorescence and of CFP were recorded at 480 nm (±15 nm); emission of YFP was recorded at 535 nm (±20 nm). Excitation times of each Fura 2 wavelength were 100 ms; excitation time of cyan fluorescent protein (CFP) was between 100 and 200 ms; switching time between different wavelengths was <10 ms. Background fluorescence was subtracted from F_YFP and F_CFP. YFP intensity was corrected for CFP signal bleed through and direct YFP excitation.

Data processing.

Fluorescence intensities in photometric measurements were acquired using CLAMPEX 9.0 (Axon Instruments, Foster City, CA). Values are presented as means ± SE from n experiments. Statistical analyses and curve fitting were performed using Prism 4.0 (GraphPad software) and CLAMPFIT 9.0 (Axon Instruments) (16). Imaging data were analyzed with MetaMorph 5.0 (Visitron Systems), Microsoft Excel, and Prism 4.0.

Statistical analyses were performed with the unpaired Student's t-test. P values are defined as follows: *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001. In Fig. 8 multiple testing was performed with Bonferroni-corrected t-tests. In this context P-values here are defined as follows: *P ≤ 0.025; **P ≤ 0.005; ***P ≤ 0.0005.

RNA isolation and RT-PCR experiments.

Total RNA from cells treated with either siRNA against AC5 and 6 or with nontargeting siRNA was isolated using peqGold TriFast (peqLab Biotechnologie, Erlangen, Germany) according to the manufacturer's instructions. To convert RNA into cDNA, SuperScript II reverse transcriptase (Invitrogen, Karlsruhe, Germany) was used with 1 μg RNA per 20 μl of each reaction.

For detection of endogenous PDE1, AC5, AC6, and for control of AC4 overexpression, transcribed cDNA was used to perform PCR each with the following specific primer pairs. (PDE1a, FW: 5′-TTT CTC TCC TGA CGG ACT CAA-3′; REV: 5′-GCATAG CTC CCA TCA CAC AC-3′; PDE1b, FW: 5′-ACT TCA TTG TGG AGC CAA CC-3′; REV: 5′-CAC GGA AAC TGA CCA CAT CA-3′; PDE1c, FW: 5′-AAG CTG AAC AAG GCA CAA CC-3′; REV: 5′-CTC ATT TCC GGT GTT GGA GT-3′; AC5, FW: 5′-CCA AGG TGG CGC TGA AGG TGG-3′; REV1: 5′-TCC TGT TTC CCT TAC AGG GCA TTG-3′; REV2: 5′-AGT TCC ATG GAC ACC CCC AGG C-3′; AC6, FW: 5′-TGA TGG GCT GGA CTG CCC AGC-3′; REV: 5′-CCC TTT TGC TGT CAG CAA GAT CCG-3′; AC4, FW: 5′-GGA GCG AGA GGA GAC TGA GA-3′; REV: 5′-GAC GCG AAG AGG ACA CAA AC-3′).

Real-time PCR was performed using SYBR-490 (iQ SYBR Green Supermix; Bio-Rad, Hercules, CA) and the iCycler (Bio-Rad). Primers (AC5 FW: 5′-CAA GGC AGG CAA GAC CCA-3′; AD5 REV: 5′-CTC CAG CCA CTA CAG GTC CAA TG-3′; AC6 FW: 5′-ACA ACG AGG GTG GAG T-3′; AC6 REV: 5′-AGG TGC TAC CGT CTT-3′; Biomers.Net, Ulm, Germany) for mouse AC5 and 6 were designed using the ENSEMBL genome browser. Data analysis was performed using the 2-ΔΔCT method (25). To calibrate this method, RT-PCR of the RNA dilution series was performed for each experiment. GAPDH and cDNA primer were used as a reference. PCR conditions were as follows: 94°C for 2 min followed by 40 cycles of 94°C for 15 s, 57°C for 30 s, and 72°C for 40 s. After each experiment, dissociation curves were monitored to control the purity of the pcr products.

AC activity assays.

Membrane preparation and AC activity assays were conducted as described by Hoffmann et al. (17) with the following minor modifications: 100 μM IBMX was changed to 500 μM RO-201724 (Sigma). Additionally, 0.24 U/reaction adenosine desaminase and 60 μM EGTA were added to each reaction sample. Defined amounts of CaCl₂ were added to achieve 2 or 5 μM free Ca²⁺ according to the online EGTA/Ca²⁺ calculator (http://brneurosci.org/egta.html). AC activity was stimulated with 1 μM isoproterenol (Sigma).

RESULTS

We focused on the question of how transient elevations of intracellular Ca²⁺ concentrations might affect cAMP levels in primary murine aortic VSMCs. To achieve sufficient temporal resolution, we chose to measure cAMP with the FRET-based sensor Epac1-camps (27). This sensor consists of a single cAMP-binding domain from Epac1, COOH, and NH₂ terminally tagged with CFP and YFP, respectively. Upon binding of cAMP, the conformation of the sensor changes, leading to a less efficient FRET between CFP and YFP as reflected by a decrease in YFP (F_YFP) and an increase in CFP emission (F_CFP) upon excitation at 436 nm (27). Changes in FRET were followed over time by monitoring the FRET ratio (F_YFP/F_CFP). This sensor is well suited to measure cAMP concentrations between 0.2 and 20 μM, covering the physiologically most relevant range for the activation of cAMP-regulated proteins such as PKAs and Epacs. As depicted in Fig. 1A, stimulation of endogenous β-adrenoceptors by isoproterenol resulted in concentration-dependent changes in FRET, reflecting graded alterations in cAMP concentrations. The resulting concentration-response curve exhibited half-maximal alterations in FRET at about 40 nM isoproterenol (log EC₅₀ of −7.42 ± 0.12, n = 29, Fig. 1B).

In principle, the use of Epac1-camps should allow the absolute quantification of cAMP concentrations in intact cells. We attempted to translate changes of FRET ratio measured in VSMCs into cAMP concentrations by first establishing a reference concentration-response curve for changes in Epac1-camps FRET measured in vitro (Fig. 1C). Lysates from HEK-TsA201-cells transiently expressing Epac1-camps were incubated with increasing concentrations of cAMP, and changes of FRET ratio were measured using fluorescence spectrometry. The resulting concentration-response curve showed an EC₅₀ value of 1.9 μM (Fig. 1C). To estimate cAMP concentrations in vivo, it was assumed that a stimulation of VSMCs using 100 μM isoproterenol would saturate all the expressed Epac1-camps with cAMP, i.e., the maximal FRET change in vitro would be equivalent to the maximal FRET change observed in VSMCs using 100 μM isoproterenol (ΔR/R₀: −0.50 ± 0.02, n = 26, Fig. 1, C and D). We also assumed that in unstimulated VSMCs (in the absence of isoproterenol), the cAMP concentration was below the detection limit of the sensor (less than ∼200 nM). Saturation of Epac1-camps was in fact indicated by a prolonged delay until the FRET ratio increased after washout of isoproterenol (Fig. 1D). We have shown previously using less sensitive cAMP FRET sensors (27) that during this time cAMP is already degraded but Epac-camps1 can only pick up changes in cAMP concentrations below ∼20 μM.

Effects of Ca²⁺ signals on basal cAMP levels in the absence of AC stimulation.

The purinergic agonist UDP is known to induce elevations of intracellular Ca²⁺ via activation of G_q-coupled purinergic receptors such as P2Y₆ receptors (20). When cells were exposed to UDP, no significant change in FRET of Epac1-camps was observed under basal conditions (Fig. 2, A and B, left). This result indicated that either UDP did not change the concentration of intracellular cAMP or that final cAMP concentrations were below the detection limit of about 200 nM.

Purinergic receptors attenuate β-adrenoceptor-mediated increase in cAMP.

In contrast, when we induced an elevation of intracellular cAMP in Epac1-camps expressing VSMCs by stimulation of β-adrenoceptors with isoproterenol, UDP did affect the cAMP signals as depicted in Fig. 2A (middle). Isoproterenol-induced elevations of intracellular cAMP were rather stable over more than 4 min. Stimulation with UDP subsequent to the isoproterenol exposure led to a clear increase in the FRET ratio, indicating a decrease of intracellular cAMP (Fig. 2A, right). To properly monitor intracellular cAMP levels upon stimulation with UDP, saturation of Epac1-camps had to be excluded. Therefore, we titrated intracellular cAMP concentrations to between 0.5 and 9 μM, which is within the dynamic range of Epac1 and PKA regulation. Because of physiological variations in responsiveness of native cells to isoproterenol, we increased isoproterenol concentrations stepwise (10 nM, 100 nM, or 1 μM), until the isoproterenol-induced decrease in the FRET ratio (F_YFP/F_CFP) in individual cells was larger than 12% (ΔR/R₀ < −0.12), corresponding to ≥0.6 μM cAMP. To avoid artifacts due to saturation of the sensor, we excluded all experiments that exhibited an isoproterenol-induced decrease of the FRET ratio of more than 42%, corresponding to approximately ≥9 μM cAMP. After the FRET ratio had reached a steady state during isoproterenol stimulation, cells were superfused with 100 μM UDP in the presence of isoproterenol (Fig. 2A, right) to stimulate purinergic receptors. This resulted in an increase of FRET, reflecting a decrease of cAMP levels. To quantify the increase of cAMP upon isoproterenol stimulation, we calculated the change of FRET ratio ΔR/R₀ for each experiment (Fig. 2B). On average, using the isoproterenol application protocol as described above, the FRET ratio decreased by 29 ± 1% (n = 29) (Fig. 2B), corresponding to intracellular cAMP concentrations of 2.9 ± 0.3 μM (n = 29) (Fig. 2C). Stimulation of purinergic receptors with UDP (100 μM) led to a reversal of the isoproterenol effect down to cAMP concentrations of 0.8 ± 0.1 μM (Fig. 2C). In the presence of isoproterenol, UDP decreased the cAMP concentrations on average by 67 ± 4%, n = 29 (Fig. 2D). UDP is known not to activate G_i-coupled P2Y receptors (7, 21). Therefore, an UDP-induced inhibition of cAMP synthesis by G_i was unlikely. To prove this hypothesis, we pretreated cells with 100 ng pertussis toxin (PTX) per milliliter overnight. However, PTX treatment markedly affected the overall morphology of the cells and the subcellular distribution of the FRET sensor, which makes an interpretation of the experiment difficult; in agreement with the hypothesis, we still observed a robust inhibition of isoproterenol-prestimulated cAMP levels by UDP (ΔR/R₀ = −0.20 ± 0.02; adenosine 3′,5′-cyclic monophosphate = 1.6 ± 0.3 μM, n = 22).

Decrease of intracellular cAMP upon stimulation of purinergic receptors is mediated by Ca²⁺.

Stimulation of VSMCs with UDP activates the G_q pathway (3) and leads to a robust increase of intracellular Ca²⁺ (20, 35). To monitor intracellular Ca²⁺ concentrations in isoproterenol-prestimulated VSMCs upon stimulation with UDP, cells were labeled with Fura 2-AM. We monitored the Fura 2 emission intensities upon alternating excitation of Ca²⁺-bound Fura 2 at 340 nm and Ca²⁺-free Fura 2 at 380 nm (Fura 2 ratio F₃₄₀/F₃₈₀). Upon stimulation of VSMCs with 100 μM UDP, F₃₄₀/F₃₈₀ increased robustly, reflecting transient increases in cytosolic Ca²⁺ (Fig. 3A), which notably did not completely return to baseline before washout. UDP-evoked Ca²⁺ signals remained at 24 ± 3% (n = 5) of the maximal signal (data not shown). Importantly, pretreatment of VSMCs with isoproterenol, as performed in the experiments depicted in Fig. 2, did not change the UDP-induced Ca²⁺ transients (Fig. 3A).

If Ca²⁺ mediates the observed purinergically induced decrease of cAMP levels, chelation of intracellular Ca²⁺ should antagonize this effect. To test this, VSMCs were incubated with 10 μM BAPTA-AM. This treatment had no effect on isoproterenol-induced cAMP levels as measured by Epac1-camps but significantly blunted the reduction in cAMP levels caused by stimulation of purinergic receptors with UDP (Fig. 3B). In BAPTA-AM-treated cells, the UDP-evoked change in cAMP levels was only a third compared with control conditions (Fig. 3C). These results support the notion that an increase of intracellular Ca²⁺ upon stimulation of purinergic receptors mediates the decrease in cAMP levels in isoproterenol-prestimulated VSMCs.

To test whether these effects could also be elicited by another G_q-coupled receptor and to analyze whether an elevation of intracellular Ca²⁺ actually precedes the decline of cAMP levels in VSMCs, simultaneous recordings of Epac1-camps FRET and Fura 2 fluorescence were performed. Stimulation of ATP-sensitive receptors induced a rapid increase of the Fura 2 ratio (indicating a fast Ca²⁺ transient) followed by a slow increase of the Epac1-camps FRET ratio (Fig. 4A), reflecting a decrease of cAMP concentrations. This supports a causative role of Ca²⁺ increases in lowering cAMP levels. To test directly whether an increase of intracellular Ca²⁺ can decrease preelevated intracellular cAMP levels, VSMCs were treated with the Ca²⁺-ionophore A23187 (1 μM, Sigma-Aldrich) after submaximal isoproterenol stimulation. Indeed upon application of A23187, Epac1-camps-FRET increased as illustrated in Fig. 4B. Note that the kinetics of this experiment were slower because compounds were added by pipette instead of the fast superfusion system due to the hydrophobic “sticky” nature of A23187. These results suggest that an increase in intracellular Ca²⁺ mediated the decrease of cAMP levels upon activation of the G_q pathway by purinergic stimulation.

Ca²⁺-mediated decrease of isoproterenol-prestimulated cAMP levels occurs independently of PDE1.

The observed decrease of isoproterenol-prestimulated cAMP levels by Ca²⁺ may result either from inhibition of cAMP synthesis or from increased degradation of cAMP. Ca²⁺-dependent regulation has been described for PDE1 as well as for several AC isoforms (8, 32, 40). To test the expression of the Ca²⁺-regulated isoforms, we used RT-PCR and found that the PDE1a isoform exhibited strongest expression in this cell type (Fig. 5A).

To assess whether PDE1 is involved in the UDP-induced decrease of cAMP, PDE1 was specifically blocked by incubating VSMCs with the PDE1 inhibitor 8-methoxymethyl-isobutylmethylxanthine (8-MM-IBMX). Treatment with 50 μM 8-MM-IBMX led to a 6 ± 1% increase in the Epac1-camps FRET ratio, indicating that PDE1 activity contributes to basal cAMP levels in VSMCs (data not shown). However, this PDE1 activity did not appear to contribute significantly to the UDP effects on cAMP levels, since the UDP-induced increase in Epac1-camps FRET ratio was similar in the presence and absence of 8-MM-IBMX (Fig. 5, B and C).

Ca²⁺-insensitive AC4 expression prevents the UDP-induced decrease of cAMP levels.

In an attempt to override endogenous ACs, we overexpressed AC4, which is not inhibited by physiological concentrations of Ca²⁺ (9, 10, 15, 40). We confirmed exogenous AC4 mRNA expression by RT-PCR (Fig. 5D) and repeated the UDP stimulation experiments. In AC4-overexpressing cells the UDP-induced increase of the FRET ratio was abolished (Fig. 5, D and E). Whereas neither R₀ values nor isoproterenol-induced changes in FRET were significantly different between AC4-transfected and control cells, the UDP-induced alterations in cAMP levels were significantly different; whereas in control cells, UDP caused a ∼75% reduction, it caused no overall change in AC4-overexpressing cells (Fig. 5E), and in some experiments it even stimulated cAMP levels (Fig. 5F). Thus the ability of purinergic receptors to cause a decrease of cAMP appears to be dependent on the AC subtype by which cAMP is produced.

Knockdown of AC5 and AC6 with siRNA reduces the ability of UDP to decrease cAMP.

ACs 5 and 6 are the sole ACs inhibited by physiological concentrations of Ca²⁺. Endogenous expression of both AC5 and 6 in VSMCs was verified by RT-PCR with specific primers (Fig. 6A). We then overexpressed the Ca²⁺-inhibitable AC6 to test whether we could amplify the UDP-induced decrease of cAMP levels. However, in AC6-transfected cells, the UDP-induced change of cAMP levels was similar compared with control conditions (Fig. 6, B–D). This suggests that endogenous isoproterenol-stimulated ACs exhibit a similar Ca²⁺-mediated pattern of cAMP inhibition compared with AC6.

To test whether the observed Ca²⁺-mediated decline of cAMP indeed resulted from a Ca²⁺-mediated inhibition of these ACs, knockdown of AC5 and 6 was performed by transfection of VSMCs with a pool of siRNAs against these two cyclases. The transfection efficiency was monitored by fluorescence excitation of Cy3-labeled siRNAs (Fig. 7A) and exceeded 75%. To validate the knockdown of AC5 and 6, real-time PCR experiments were performed. When compared with scrambled siRNA-transfected VSMCs, the pool of siRNA against AC5 and 6 reduced their mRNA levels to 13.3 ± 0.8% and 19.5 ± 9.5%, respectively (Fig. 7B). Because of the high similarity of AC5 and 6 sequences, all tested real-time PCR primer pairs (four pairs for each subtype) showed a slight cross reactivity. Therefore, we were not able to clearly distinguish expression of AC5 and 6.

We then tested for UDP-induced alterations in cAMP levels in siRNA-transfected VSMCs. AC5 and 6 siRNA significantly reduced the UDP-evoked increase in FRET compared with cells transfected with scrambled siRNA (Figs. 7, C–E). These results suggest that endogenous AC5 and 6 are substantially involved in the decrease in cAMP levels upon purinergic stimulation of isoproterenol-prestimulated VSMCs.

β-Adrenoceptor-mediated stimulation of ACs is inhibited by Ca²⁺.

To test directly whether Ca²⁺ inhibits isoproterenol-induced AC activity in VSMCs, we performed radioenzymatic assays of AC activity in membranes of VSMCs. These experiments were conducted in the presence of RO 201724 to block PDE activity. In the absence of isoproterenol, basal AC activity was not inhibited by Ca²⁺ (data not shown). In the presence of 1 μM isoproterenol, we found a significant inhibition of cAMP production by 2 μM as well as 5 μM Ca²⁺ compared with conditions of very low free Ca²⁺ (60 μM EGTA, Fig. 8). This result confirmed that isoproterenol-induced cAMP production in VSMC membranes is sensitive to intracellular Ca²⁺, supporting the notion that Ca²⁺-mediated inhibition of AC5/6 may represent a major component of the observed UDP-mediated decrease of cAMP in VSMCs.

DISCUSSION

The present study aimed to investigate the regulation of intracellular cAMP levels by intracellular Ca²⁺ signals in single intact VSMCs. To allow investigations in a temporally resolved fashion, we used Epac1-camps (27) as a FRET probe for intracellular cAMP and Fura 2 to monitor intracellular Ca²⁺ levels. We used UDP as an agonist for endogenous G_q-coupled P2Y₆ receptors to induce elevations of intracellular Ca²⁺ (7, 21). In the absence of AC prestimulation, we did not detect any UDP-dependent FRET changes in VSMCs (Fig. 2A). Based on this observation and the fact that in isolated membranes Ca²⁺ induced no inhibition of basal AC, we conclude that intracellular Ca²⁺ signals will have no relevant impact on basal cAMP concentrations in VSMCs.

In contrast we found robust UDP-induced decreases of cAMP concentrations from ∼3 to 1 μM when AC activity had been prestimulated by moderate activation of β-adrenoceptors. We propose that this effect results from the G_q-mediated increase in intracellular Ca²⁺ based on the following observations. 1) Direct elevation of intracellular Ca²⁺ (by application of the Ca²⁺ ionophore A23187) decreased cAMP with a similar temporal pattern as observed upon purinergic stimulation (Fig. 4B). 2) Chelation of intracellular Ca²⁺ with BAPTA-AM significantly reduced the UDP-induced effect (Fig. 3). 3) Simultaneous measurements of intracellular Ca²⁺ and cAMP levels upon stimulation of purinergic receptors showed that the rapid increase of intracellular Ca²⁺ preceded the slow decline of cAMP (Fig. 4A). The slow time course of the cAMP decline relative to the fast rise of Ca²⁺, but also relative to the fast kinetics of receptors and G proteins (16), most likely reflects the time it requires endogenous PDEs to hydrolyze already produced cAMP.

It has been reported that in different cell types, intracellular Ca²⁺ is a regulator of both cAMP synthesis and cAMP degradation (8, 11, 37). Ca²⁺ inhibits AC5 and 6 (14, 24, 26) and activates PDE1 (32, 33). Activation of other PDE classes by Ca²⁺ has not been reported (28, 34). Two findings argue against a major role for Ca²⁺-mediated activation of PDEs in the UDP-induced decrease of cAMP: 1) Blocking PDE1 by 8-MM-IBMX did not significantly alter the effect of UDP (Fig. 5); and 2) Ca²⁺ at 2 and 5 μM directly inhibited isoproterenol-stimulated AC activity in cell membranes (Fig. 8).

Consequently, Ca²⁺-dependent inhibition of ACs appears to be the most likely mechanism for the UDP-induced cAMP decrease in VSMCs. The observation that overexpression of AC4 prevented UDP to decrease cAMP levels points to an AC subtype-specific inhibition by Ca²⁺.

Indeed, knockdown of endogenous AC5 and 6 by siRNA antagonized the UDP-induced effect. We were not able to significantly potentiate this effect by overexpression of AC6, indicating that β-adrenoceptors may predominantly couple to endogenous AC5 and 6.

Our data are in line with earlier findings (12), which report a Ca²⁺-induced cAMP decline in A7r5 smooth muscle cells. Similarly, Ca²⁺-dependent decreases of cAMP levels in endothelial cells have recently been reported (38).

Taken together, we demonstrate an intricate interplay between Ca²⁺ and cAMP, two important antagonistic regulators of vascular tone. From a physiological point of view, it is important to note that the vasoconstricting messenger Ca²⁺ decreases the levels of the vasodilating mediator cAMP, thereby facilitating effective vasoconstriction. Activation of the G_q-pathway and subsequent Ca²⁺ signals thus further contribute to vasoconstriction. According to our results, UDP-induced Ca²⁺ signals effectively switch off β-adrenergically induced activity of AC5 and 6, leading to a robust decrease of cAMP within 1–2 min, which will contribute to vasoconstriction by eliminating vasodilatory β-adrenergic effects.

This crosstalk may represent an important mechanism in the reversal of cAMP-induced vasorelaxation by stimulation of G_q-coupled receptors.

GRANTS

We thank the Deutsche Forschungsgemeinschaft for financial support of this work (SFB 688, TP B6).

DISCLOSURES

No conflicts of interest are declared by the author(s).