Calcineurin feedback inhibition of agonist-evoked cAMP formation.

The effects of immunosuppressant blockers of calcineurin (protein phosphatase 2B) on cAMP formation and hormone release were investigated in mouse pituitary tumor (AtT20) cells. Immunosuppressants enhanced corticotropin-releasing factor- and isoproterenol-evoked cAMP production in proportion with their potency to block calcineurin. Further analysis of cAMP production revealed that intracellular Ca2+ derived through voltage-regulated calcium channels reduces cAMP formation induced by corticotropin releasing-factor or beta 2-adrenergic stimulation and that this effect of Ca2+ is inhibited by blockers of calcineurin. AtT20 cells were found to express at least three species of adenylyl cyclase mRNA-encoding types 1 and 6 as well as a novel isotype, which appeared to be the predominant species. In two cell lines expressing very low or undetectable levels of the novel cyclase mRNA (NCB20 and HEK293 cells respectively), corticotropin-releasing factor-induced cAMP formation was not altered upon blockage of calcineurin activity. These data identify calcineurin as a Ca2+ sensor that mediates the negative feedback effect of intracellular Ca2+ on receptor-stimulated cAMP production. Furthermore, the effect of calcineurin on cAMP synthesis appears to be associated with the expression of a novel adenylyl cyclase isotype, which is highly abundant in AtT20 cells.

The effects of immunosuppressant blockers of calcineurin (protein phosphatase 2B) on cAMP formation and hormone release were investigated in mouse pituitary tumor (AtT20) cells. Immunosuppressants enhanced corticotropin-releasing factor-and isoproterenol-evoked cAMP production in proportion with their potency to block calcineurin. Further analysis of cAMP production revealed that intracellular Ca 2؉ derived through voltage-regulated calcium channels reduces cAMP formation induced by corticotropin releasing-factor or ␤ 2 -adrenergic stimulation and that this effect of Ca 2؉ is inhibited by blockers of calcineurin. AtT20 cells were found to express at least three species of adenylyl cyclase mRNA-encoding types 1 and 6 as well as a novel isotype, which appeared to be the predominant species. In two cell lines expressing very low or undetectable levels of the novel cyclase mRNA (NCB20 and HEK293 cells respectively), corticotropin-releasing factor-induced cAMP formation was not altered upon blockage of calcineurin activity. These data identify calcineurin as a Ca 2؉ sensor that mediates the negative feedback effect of intracellular Ca 2؉ on receptor-stimulated cAMP production. Furthermore, the effect of calcineurin on cAMP synthesis appears to be associated with the expression of a novel adenylyl cyclase isotype, which is highly abundant in AtT20 cells.
Calcineurin (protein phosphatase 2B) is a Ca 2ϩ /calmodulinregulated protein phosphatase first discovered in brain, where it is highly abundant (0.5-1% of total protein) (1). Elucidation of the physiological role of this protein phosphatase has been relatively slow due to the lack of specific inhibitors of its enzymatic activity. It is now well established that the major immunosuppressant compounds cyclosporin A and FK506 are potent and, with appropriate controls, specific blockers of calcineurin in leukocytes (2) and other systems (3,4). This observation has led to the discovery that calcineurin is an essential element of the signal transduction pathway activated by the T-cell receptor (5,6).
In excitable cells, the functions of calcineurin are less well understood. Calcineurin has been implicated in the control of voltage regulated ion channel activity (7), particularly with respect to L-type calcium channels (8). More recent studies applying immunosuppressant blockers of calcineurin have shown that the synaptic vesicular protein dynamin, which is thought to participate in synaptic vesicle recycling in nerve endings, is a prominent substrate for calcineurin (9) and that blockage of calcineurin enhances glutamate release by synaptosomes prepared from rat brain (10). In pituitary corticotrope tumor (AtT20) cells (4,11) immunosuppressants block calcineurin activity and stimulate Ca 2ϩ -dependent hormone release in correlation with their calcineurin blocking potency. In hippocampal brain slices, calcineurin is involved in the induction of long-term synaptic depression (12). Finally, ligand-operated ion channels such as the NMDA receptor (13) or the 5HT 3 receptor (14) are desensitized by calcineurin.
Taken together, these data indicate multiple roles for calcineurin in diverse signal transduction cascades of excitable cells. A common feature of all of these proposed functions is the Ca 2ϩ -dependent inhibition of cellular activation. This is the opposite of what has been observed in nonexcitable cells, such as lymphocytes (5,6) and adrenocortical glomerulosa cells (15), where calcineurin is an intracellular mediator of the action of stimulatory agents.
The effects of immunosuppressants on the cAMP signaling system in excitable cells have not been previously examined. cAMP is a cardinal signaling molecule in pituitary corticotropes (16), where its synthesis is activated by 41-amino acid residue CRF 1 (17). Increased levels of cAMP augment intracellular free Ca 2ϩ concentration ([Ca 2ϩ ] i ) (4); in turn, Ca 2ϩ synergizes with cAMP to trigger the release of adrenocorticotropic hormone (ACTH) (18). As [Ca 2ϩ ] i is also known to inhibit cAMP formation in several systems (19) and because immunosuppressants enhanced CRF-induced ACTH release in AtT20 cells (20), we have examined the effects of immunosuppressants on CRFinduced cAMP production. The results indicate that in AtT20 cells, calcineurin inhibits CRF-induced cAMP formation and that this is associated with the expression of a novel isotype of adenylyl cyclase.

MATERIALS AND METHODS
Cell Culture-AtT20 D16:16 mouse anterior pituitary tumor cells were maintained in culture as described previously (21). NCB20 mouse neuroblastoma ϫ hamster brain hybridoma cells (courtesy of Dr. Beth Hoffman, National Institute of Mental Health, Bethesda, MD) were cultured in 10% newborn calf serum and Dulbecco's minimal essential * A preliminary account of parts of this work has been printed (11). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  For measurements of ACTH release, cAMP production, or calcineurin activity, the cells were plated on 24-well tissue culture plates (5 ϫ 10 4 cells/well) and used 4 -6 days afterwards. ACTH (21) and cAMP (22) were measured by specific radioimmunoassays. Calcineurin protein phosphatase activity was determined by the 32 P-labeled casein assay (23) or the RII phosphopeptide assay (24) adapted to measure calcineurin phosphatase activity in AtT20 cell extracts as described previously (4).
Measurement of cAMP Responses to Agonists-This protocol was established using AtT20 cells and was also applied in experiments with the other two cell lines. Experiments were all carried out in Hank's balanced salt solution containing 2 mM CaCl 2 and 1 mM MgSO 4 buffered with 25 mM HEPES, pH 7.4, and supplemented with 0.25% (w/v) of bovine serum albumin. The cells were preincubated in serum-free medium for 1 h, after which fresh medium containing blockers of phosphodiesterase, 0.5-1 mM IBMX, and/or 0.1 mM rolipram along with various other agents as specified below and in the figure legends were applied for 30 min at 37°C. Subsequently, the cells were cooled to 24°C (5-10 min in a water bath), and agonists were added for 10 min. The reaction was stopped by the addition of 0.2 M HCl to achieve a final concentration of 0.1 M (25).
In the absence of phosphodiesterase blockers, agonist-induced changes of total cAMP content were undetectable at 24°C. In the presence of IBMX, total cAMP content (cells ϩ medium) increased with time up to 20 min after the addition of CRF and remained constant for up to 30 min. In contrast, intracellular cAMP content peaked between 2 and 5 min and subsequently declined to basal levels even in the presence of the phosphodiesterase blockers. Hence, after establishing that immunosuppressant drugs had the same effect on peak cellular and total cAMP content under these conditions, all experiments shown here report total cAMP content.
Depletion and Repletion of Intracellular Calcium Pools-Calcium depletion was initiated during preincubation in the presence of phosphodiesterase blockers, which were applied as described in the preceding section. Cells were preincubated for 30 min in medium containing 2 mM EGTA and no added Ca 2ϩ supplemented with 5 M A23187 and 2.5 M nifedipine in order to deplete rapidly mobilized cellular stores of Ca 2ϩ and to ensure that L-type Ca 2ϩ channels, the principal avenue of voltage-regulated Ca 2ϩ influx in AtT20 cells (26,27,28) were fully blocked. Thus, Ca 2ϩ , added as chloride salt to the extracellular fluid at the time agonist stimulation was initiated, would enter largely through the pores made by the ionophore A23187. The rationale for this pretreatment is that calcineurin reportedly influences L channel activity (7,8), whereas the treatment regimen used here would make Ca 2ϩ entry independent of this regulation.
In some experiments, nifedipine was given alone during preincubation to achieve blockage of Ca 2ϩ entry via L channels; alternatively, the intracellular calcium chelator BAPTA-AM (20 M) was used in the preincubation medium in order to attenuate the rise of intracellular free calcium levels caused by CRF.
Pretreatment with Immunosuppressants-Immunosuppressant analogues (FK506, Fujisawa Ltd, Osaka, Japan) cyclosporin A, and SDZ 220 -384 (MeVal 4 cyclosporin A, a weak agonist) (29), (Sandoz Pharma, Basel, Switzerland), L-685,818 an FK506 antagonist, (Merck) were applied during the 30-min preincubation period, invariably in the presence of phosphodiesterase blockers, and in some cases to cells undergoing the calcium depletion protocol described above. Immunosuppressants were made up in ethanol at 10 Ϫ3 M and diluted with the incubation medium to the desired final concentrations. In some cases, cells were preincubated with L-685,818 a structural analogue of FK506 that binds to FKBP12 and inhibits prolylisomerase activity but is devoid of immunosuppressant activity (30) for 10 min before the addition of FK506. Preincubation was carried out at 37°C, because immunosuppressants are taken up into cells through an active transport process involving the multidrug resistance transporter glycoprotein (31), and the uptake into cells is temperature-dependent (32). Test incubations with agonists were carried out at 24°C, because at this temperature the effects of immunosuppressants on cAMP formation were more pronounced than at 37°C. At present, we have not identified the cause for the temperature-dependent difference in the efficiency of immunosuppressants to influence cAMP formation.
Secretion of ACTH-Incubations for ACTH secretion were carried out as described previously (21) except that the test incubation with CRF was for 30 min at 24°C, to allow comparisons with the conditions of cAMP accumulation experiments. Immunosuppressants were applied as in cAMP experiments; blockers of phosphodiesterase were not used.
Amplification and DNA Sequence Analysis of Adenylyl Cyclases cDNAs in AtT20 Cells-Total RNA was prepared from approximately 10 7 cells using Trizol reagent (Life Technologies, Inc.) according to the manufacturer's instructions. Reverse transcriptase PCR was carried out using an RNA PCR kit (Perkin Elmer). Briefly, 0.8 g of total RNA was denatured at 95°C for 5 min and then annealed with 2.5 M random hexanucleotide primers for first-strand cDNA synthesis, which was carried out for 15 min at 42°C in a 20-l reaction mixture containing 10 mM Tris⅐HCl, pH 8.3, 50 mM KCl, 5 mM MgCl 2 , 1 mM dNTP, 20 units of RNase inhibitor, and 50 units of Moloney murine leukemia virus reverse transcriptase. The reaction was terminated at 99°C for 5 min and then cooled to 4°C and stored on ice. PCR was performed using degenerate oligonucleotides, either pair A or pair B corresponding to highly conserved regions within the first (pair A, 5Ј-CTCATCGATGGI-GAYTGYTAYTAYTG-3Ј; 3Ј-GGCTCGAGCCAIACRTCRTAYTGCCA-5Ј, expected product size 220 bp) and second (pair B, 5Ј-GAAGCTTA-ARATIAARACIATIGGI(T/A)(C/G)IACITAYATGGC-3Ј; 3Ј-GGGATCC-ACRTTIACIGTRTTICCCCAIATRTCRTA-5Ј, expected product size 180 bp) cytoplasmic domains of previously cloned mammalian adenylyl cyclases (33,34,35). For PCR, the reverse transcription reaction (20 l) was expanded to 100 l and contained 10 mM Tris⅐HCl, pH 8.3, 50 mM KCl, 2 mM MgCl 2 , 200 M each dNTP, 35 pmol of each primer, and 2.5 units of Amplitaq DNA polymerase. PCR reactions were overlaid with mineral oil (Sigma) and denatured at 95°C for 3 min followed by 5 cycles (60 s denaturation at 94°C, 60 s annealing/extension at 45°C) and then a further 35 cycles (60 s denaturation at 94°C, 60 s annealing/ extension at 55°C), and finally 7 min of annealing/extension at 55°C. An aliquot (5%) of each reaction was analyzed by agarose gel electrophoresis (3% Nusieve (Flowgen). Products within the expected size range for each primer pair were excised from the gel, purified using a Wizard PCR Prep kit (Promega), and ligated into the vector pGEM-T (Promega). Clones containing an insert of the expected size were identified, and their DNA sequence was determined by the dideoxynucleotide method (Sequenase 2.0 kit, U. S. Biochemical Corp.). Clone JP114, containing 180 bp of cyclase sequence was isolated and used for generating cDNA and cRNA probes for RNA detection.
Detection of mRNA Expression-Northern analysis was performed using standard procedures. Briefly, 10 g of total RNA was separated by formaldehyde gel electrophoresis and transferred by blotting onto positively charged nylon membrane (Appligene) and then fixed by baking at 80°C and prehybridized at 42°C for 2 h in 50% deionized formamide, 5 ϫ saline/sodium phosphate/EDTA, 0.5 ϫ Denhardt's solution, 0.1% (w/v) SDS, 0.2 mg/ml denatured salmon sperm carrier DNA, and 10% Dextran sulfate. Random-primed [ 32 P]dCTP-labeled DNA probe (50 ng; Ͼ10 9 cpm/g) was then added, and hybridization continued overnight at 42°C. The membrane was washed twice for 20 min in 2 ϫ SSC, 0.1% SDS, followed by 20 min in 1 ϫ SSC, 0.1% SDS at 42°C and finally 20 min in 0.5 ϫ SSC, 0.1% SDS at 50°C before wrapping in cling-film and exposing to autoradiographic film at Ϫ70°C or to Molecular Dynamics PhosphorImager cassettes and quantified with the ImageQuant software using the 28 S RNA band as a standard for RNA loading. Division of the integrated volume of pixels of the selected radiolabeled band with the integrated volume of the internal standard band yields the relative hybridization intensity, which was used to compare the intensity of labeled RNA bands within blots.
Ribonuclease protection assays were performed using an RPA II kit (Ambion, AMS Biotechnology, Witney, Oxon, UK) according to the manufacturer's instructions. Briefly, 10 g of total RNA was hybridized overnight at 45°C to 10 5 cpm of radiolabeled JP114 antisense riboprobe. Following hybridization, reactions were digested with single strand-specific RNase, and protected fragments were resolved on a 6% denaturing polyacrylamide gel, which was fixed for 30 min in 15% methanol, 5% acetic acid, dried, and exposed to autoradiographic film at Ϫ70°C.

RESULTS
Enhancement of CRF-stimulated cAMP Production by Immunosuppressants-Blockers of calcineurin activity enhanced CRF-induced cAMP production (Fig. 1). This effect was statistically significant (p Ͻ 0.05 or less) at 5-30 min after the addition of CRF (Fig. 2A). An enhancement of cAMP formation by FK506 was apparent at lower concentrations (0.1-10 nM) of CRF, while the maximal response was unchanged (Fig. 2B).
The effect of FK506 on CRF-induced cAMP production could be antagonized by the nonimmunosuppressant analogue L685,818 (30), which also blocked the inhibitory effect of FK506 on calcineurin-mediated dephosphorylation of phosphocasein (Fig. 3, A and B).
Receptor-evoked Synthesis of cAMP Is under Inhibitory Control by Intracellular Ca 2ϩ and Calcineurin-Lowering of Furthermore, data reported elsewhere (20) showed that the addition of graded amounts of Ca 2ϩ with CRF to cells depleted of Ca 2ϩ and pretreated with the ionophore A23187 produced a concentration-dependent inhibition of CRF-induced cAMP production to levels seen in nondepleted cells incubated in medium containing 2 mM Ca 2ϩ . The effect of exogenous Ca 2ϩ could be inhibited by FK506, which failed to alter cAMP accumulation in the absence of Ca 2ϩ (20).
Site and Specificity of Immunosuppressant Action-The effect of FK506 on CRF-induced cAMP formation was also evident after 16 h of pretreatment of AtT20 cells with pertussis toxin (1 g/ml) (Fig. 4), which strongly suppressed inhibitory G-protein function as assessed by the attenuation of somatosta- tin-mediated inhibition of cAMP formation. Pertussis toxin treatment also had no effect on the suppression of CRF-induced cAMP formation by extracellular Ca 2ϩ in Ca 2ϩ -depleted cells (not shown).
FK506 had no significant effect on cAMP accumulation evoked by 10 or 30 M forskolin, a drug that at these concentrations activates adenylyl cyclase independent of G s . Loading of the cells with BAPTA-AM caused a small (15%), statistically significant (p Ͻ 0.05) enhancement of forskolin-evoked cAMP accumulation (not shown) .
Finally, in contrast to the effects of FK506 and cyclosporin A, pretreatment with other blockers of protein phosphatases such as calyculin A (1-30 nM) and okadaic acid (0.2-5 M), caused a concentration-dependent inhibition (up to 80%) of CRF-induced cAMP accumulation (not shown and Ref. 36).
Enhancement of CRF-stimulated ACTH Release by FK506 -Blockage of calcineurin activity by FK506 enhanced the release of ACTH evoked by CRF (Fig. 5A), and this action was prevented by L-685,818 (Fig. 5B). Note, that the apparent EC 50 values of FK506 to inhibit calcineurin activity in AtT20 cells (4), to stimulate ACTH release, and to augment cAMP accumulation induced by CRF are all approximately 10 nM.
␤-Adrenergic Stimulation Is under Similar Regulation by Calcineurin-Isoproterenol stimulates cAMP production through ␤ 2 -adrenergic receptors in AtT20 cells (37), and this was enhanced by both BAPTA-AM and FK506 (Table I).

Effect of Immunosuppressants on cAMP Accumulation in AtT20 Cells Correlates with the Expression of a Novel Adenylyl
Cyclase mRNA-In order to determine the profile of adenylyl cyclase isoforms present in AtT20 cells, two sets of degenerate oligonucleotide primers were used to analyze AtT20 cell total RNA for adenylyl cyclase-related sequences by means of reverse transcriptase PCR. Using primer set B, a PCR product of approximately 180 bp was obtained. DNA sequence analysis revealed that approximately 8% of the subcloned 180-bp cDNA fragments amplified proved to be identical to type 6 adenylyl cyclase. The majority (Ͼ90%), however, gave a novel sequence that was highly homologous to the primary amino acid sequences of known mammalian adenylyl cyclases found in current data bases but was not identical to any of these (Fig. 6). Type 1 adenylyl cyclase was detected in AtT20 cells using primer set A.
Northern blot analysis of total RNA using the novel adenylyl cyclase 180-bp cDNA fragment as a probe indicated hybridization to an approximately 9-kilobase mRNA expressed in AtT20 cells (Fig. 7), and a single hybridizing species of RNA of similar size was detected in NCB20 and HEK293 cells at much lower intensity (Relative hybridization intensity of 9-kb band (arbitrary units): AtT20, 0.38; NCB20, 0.03; HEK293, 0.07).
As a potentially more sensitive alternative, mRNA expression was also assayed by ribonuclease protection using a radiolabeled antisense riboprobe transcribed from the novel adenylyl cyclase cDNA. An approximately 160-bp ribonucleaseprotected RNA species (Fig. 7) indicates that the novel adenylyl cyclase mRNA is highly abundant in AtT20 cells, whereas much lower levels are present in NCB20 cells, and in HEK293 cells the mRNA was undetectable. 2 Calcineurin protein phosphatase activity (substrate RII sub- showed that there is no effect of either preincubation protocol on basal cAMP levels. Effect of BAPTA-AM and FK506 was significant p Ͻ 0.05 at 0.4 -100 nM isoproterenol when compared with the respective control group, one-way ANOVA followed by Newman-Keuls test. Data were log-transformed before analysis to eliminate heterogeneity of variances. unit peptide (24)) in cell extracts prepared from AtT20, HEK293, and NCB20 cells fell to 25,19, and 42% of the respective control activities after pretreatment with 1 M FK506. Similar to AtT20 cells, CRF-stimulated cAMP formation in HEK293 cells (Fig. 8) as well as NCB20 cells (not shown), and this was enhanced by the depletion of intracellular calcium stores as described for AtT20 cells. However, while in AtT20 cells FK506 consistently enhanced CRF-induced cAMP forma-tion, in NCB20 cells only one out of four experiments gave a statistically significant enhancing effect of 1 M FK506 on CRF-induced cAMP accumulation, and in none out four experiments in the case of HEK293 cells (Fig. 8). No effects of cyclosporin A were found in either system (not shown).

DISCUSSION
These data show that receptor-stimulated cAMP formation may be inhibited by calcineurin and that this regulation is associated with the expression of a novel adenylyl cyclase mRNA.
All studies of cAMP formation reported here were carried out in the presence of blockers of phosphodiesterase, and hence the effects observed relate to changes in the rate of synthesis of cAMP rather than to its degradation.
Evidence for the involvement of calcineurin in the control of cAMP accumulation is provided by the use of immunosuppressant compounds previously (4) shown to block calcineurin activity in AtT20 cells with the same order of potency that they influenced cAMP accumulation (present study). The EC 50 for FK506 to block calcineurin activity is considerably higher in AtT20 cells (ϳ10 nM) than in T lymphocytes (ϳ0.8 nM), which is probably attributable to differences in the respective cellular levels of calcineurin and FKBP12 in these systems. Importantly, L685,818, an analogue of FK506 (30) that binds to the prolylisomerase FKBP-12 in a manner similar to FK506 but does not give rise to a drug-protein complex that inhibits the activity of calcineurin, reversed the effects of FK506 on calcineurin activity and cAMP formation as well as ACTH release. When given alone, L685,818 had no discernible effect on cAMP formation or ACTH secretion, further suggesting that the changes observed upon treatment with FK506 are due to the inhibition of calcineurin and not due to the blockage of the prolylisomerase activity of FKBP12. Finally, neither FK506 nor cyclosporin A were effective in cells deprived of Ca 2ϩ (20).
Taken together, these characteristics justify the conclusion that the effects of immunosuppressants described here are attributable to the inhibition of calcineurin.
The production of cAMP in AtT20 cells is under inhibitory control by [Ca 2ϩ ] i . Stimulation with cAMP is known to elicit a rise of [Ca 2ϩ ] i in these cells, which is largely derived from the extracellular pool by influx through dihydropyridine-sensitive Ca 2ϩ -channels (26,27,28). Thus the [Ca 2ϩ ] i signal is a measure of the electrical activity of the cells and, in addition to triggering hormone release, provides feedback inhibition to the chemical messenger system that generates it. In the case of FIG. 6. Comparison of the sequence amplified from AtT20 cell RNA using primer set B with the corresponding sequences of mammalian adenylyl cyclases found in current data bases (EMBL, GenBank, SwissProt). The numbers relate to the amino acid sequence of rabbit ACtype5 (ocmradcyv). The alignment was generated by the PileUp program of the GCG package. The sequence is annotated by (q) showing amino acid identity of the novel sequence with at least one previously reported AC, (E) shows functionally conservative substitutions (-) denotes nonconservative substitutions as defined by Krupinski et al. (52). The abbreviations used are as follows: Hum9, Human AC type9 (GenBank #D25538); mmu12919, mouse AC type7; cya2_rat, rat ACtype2; cya_rat, rat ACtype4; hsadencyr8, Human AC-type8; ratacviii, rat ACtype8; a46187, human ACtype5; cya6_mouse, mouse ACtype6; a49201, mouse ACtype5; cya6_rat, rat ACtype6; cya6_canfa, dog ACtype6; s29717, rat ACtype5; ocmradcyv, rabbit AC-type5; cya5_canfa, dog ACtype5; cya1_bovin, bovine Actype1; cya3_rat, rat ACtype3; jp114, AtT20 new sequence(mouse ACtype10)  8. Stimulation of cAMP accumulation in HEK293 cells by corticotropin-releasing factor. Cells were pretreated with the calcium depleting medium at 37°C in the presence of 1 mM IBMX plus or minus 1 M FK506 for 30 min before the addition of CRF for 10 min at 24°C. In the case of the control and the FK506-treated group, the CRF solution contained sufficient CaCl 2 in buffered EGTA to bring the medium to nominally 2 mM free extracellular Ca 2ϩ . Unstimulated cAMP content was 0.05 Ϯ 0.006 pmol/well and was unaltered by depletion of calcium or preincubation with FK506. n ϭ 4/group; data are means Ϯ S.E.; *, p Ͻ 0.01 when compared with the control group in 1-way ANOVA followed by orthogonal contrasts.
CRF-induced cAMP formation, this feedback appears to be mediated by calcineurin.
Several possibilities have to be considered with respect to the site of action of Ca 2ϩ /calcineurin in the signal transduction cascade.
An action of calcineurin at the receptor level is conceivable; however, the prevailing concept of G-protein-coupled receptors (38) dictates that receptor down-regulation or uncoupling is largely due to the action of protein kinases while protein phosphatases reverse this process. In contrast, the present data implicate calcineurin as an inhibitor of receptor-stimulated cAMP production.
Dephosphorylation of the coupling protein G s is also a possible site of regulation by calcineurin (39). Once more, current evidence in the literature associates protein phosphorylation with down-regulation of G-protein function and implicates protein phosphatases in the restoration of the cellular response (40,41).
With respect to the effector enzyme adenylyl cyclase, these proteins have lately emerged as dynamic sites of signal integration (42). At least two types of cyclase, types 5 and 6 (43), are inhibited by Ca 2ϩ , but the mechanism of this effect has not been elucidated (33). The inhibition of type 5 and 6 cyclase by Ca 2ϩ is most marked after stimulation by agonists such as isoproterenol in chick heart cells (44), or VIP in GH 4 C 1 pituitary tumor cells (45), but much less prominent after activation with forskolin in GH 4 C 1 cells (45). Overall this is analogous to the observations made here, which in the first instance suggest a prominent action of calcineurin at or before the level of G-protein effector coupling. However, as multiple types of adenylyl cyclase coexist in all cell types analyzed to date (33,46,47), and forskolin appears to activate these by different efficacies and mechanisms, an effect of Ca 2ϩ on catalytic activity as opposed to the interaction with the G s ␣-subunit-coupling site cannot be excluded (43). Reverse transcriptase PCR analysis and sequencing of amplified cDNAs clearly show that at least three types of adenylyl cyclase mRNA (type 1 and 6 as well as a novel isotype) are co-expressed in AtT20 cells, and thus the above considerations also apply to this system.
It is unlikely, that type 1 cyclase is involved in the effects reported here as it is invariably stimulated by Ca 2ϩ , whereas Ca 2ϩ was strongly inhibitory to both CRF and ␤-adrenergic stimulation of cAMP. Type 6 adenylyl cyclase could be implicated as it is inhibited by Ca 2ϩ . However this isotype is abundant in NCB20 (33) as well as HEK293 cells (47) where the stimulation of cAMP accumulation through endogenously expressed receptors for CRF was not altered by immunosuppressants, despite a marked inhibition of calcineurin activity measured using the RII substrate phosphopeptide. Importantly, the novel adenylyl cyclase homologue mRNA was found in very low amounts in NCB20 cells and HEK293 cells, whereas it appears to be the predominant adenylyl cyclase isotype mRNA in AtT20 cells. A partial mammalian sequence that is identical except for a single amino acid to the one reported here has been previously designated as adenylyl cyclase type 10 (48). Results from this laboratory (49) show that the 9-kb mRNA detected in AtT20 cells contains a full-length adenylyl cyclase coding sequence giving rise to an adenylyl cyclase inhibited by calcineurin.
A previous study (15) has reported that calcineurin is stimulatory to cAMP formation; immunosuppressants blocked the enhancement of ACTH-evoked cAMP production by angiotensin II and activators of protein kinase C in bovine adrenal cortical cells as well as transfected COS-7 cells. As adenylyl cyclase type 10 mRNA is undetectable in COS-7 cells and protein kinase C activation is inhibitory to cAMP production in AtT20 cells (36), it is unlikely that adenylyl cyclase type 10 is involved in the enhancement of cAMP production by calcineurin as reported by Baukal and co-workers (15). Another study, using partially purified solubilized bovine brain adenylyl cyclase (50) reported inhibition of cyclase activity by calcineurin; however, this was attributed to the sequestration of endogenous calmodulin in the enzyme preparation by calcineurin, and the consequent inhibition of a calmodulin-stimulated cyclase activity. Thus, whether calcineurin regulates cyclase type 10 directly or through an intermediary phosphoprotein specific to AtT20 cells remains to be determined by future studies.
Taken together, the present data indicate that in AtT20 cells calcineurin is a Ca 2ϩ -operated feedback inhibitor of CRF or ␤ 2 -adrenergic receptor-evoked cAMP responses. As [Ca 2ϩ ] i is largely derived through voltage-regulated Ca 2ϩ channels in AtT20 cells, these cells exemplify a case where calcineurin functions as a link between the cAMP-generating and electrical signaling systems of the cell. The potential functional significance of this mechanism is illustrated by changes of hormone secretion that parallel the enhancement of the cAMP signal. Immunosuppressants augmented CRF-induced ACTH secretion and attenuated the inhibitory effect of adrenal corticosteroids (20). Our findings conform with previous reports (7,10,13,14) in showing that calcineurin is a fundamental negative feedback regulator of cellular responses in excitable cells and extend this function to the cAMP signal transduction cascade. In this latter respect the data also support the earlier notion (51), that calcineurin is a generic antagonist of cAMP-induced stimulatory mechanisms.