Factors Determining the Specificity of Signal Transduction by Guanine Nucleotide-binding Protein-coupled Receptors

To define the integration of multiple signals by different types of adenylyl cyclase (AC) within the cell, we altered the population of enzymes expressed in the cell and determined the subsequent processing of stimulatory and inhibitory input. DDT1-MF2 cells expressed AC VI–IX and were stably transfected with AC II, III, or IV. Enzyme expression was confirmed by RNA blot analysis and functional assays. Basal enzyme activity was only increased in AC II transfectants (6-fold). Maximum stimulation of enzyme activity was increased in each of the AC transfectants to varying extents. α2A/D-AR activation elicited enzyme type-specific responses. α2-AR activation inhibited the effect of isoproterenol in control transfectants, and this action was magnified in AC III transfectants. In AC II and AC IV transfectants, α2-AR activation initiated both positive (Gβγ) and negative signals (Giα) to the Gsα-stimulated enzyme, and both types of signals were blocked by cell pretreatment with pertussis toxin. The negative input to AC II from the α2-AR was blocked by protein kinase C activation in AC II transfectants, but it was the positive input to AC IV that was compromised by protein kinase C activation. These data indicate that the integration of multiple signals by adenylyl cyclases is a dynamic process depending upon the enzyme type and phosphorylation status.

as isoforms with different regulatory properties allowing complex signal integration, and this capability presents opportunities for the cell to engineer highly specific responses to an external stimulus. Such appears to be the case for signals involving the enzyme adenylyl cyclase. There are nine types of adenylyl cyclase (AC) 1 that differ in their regulatory mechanisms and are expressed in a tissue-specific manner (1). All enzyme types are activated by G s ␣. AC I, III, and VIII are positively regulated by calcium/calmodulin, whereas types V and VI are directly inhibited by calcium. AC II and IV are stimulated by G␤␥ in the presence of activated G s ␣. The activities of types I-III and V-VII are also influenced by phosphorylation with protein kinase C and/or A (2)(3)(4)(5)(6)(7)(8)(9).
The ability of various AC types to respond to activated G s ␣ and G i ␣, G␤␥, calcium, and/or phosphorylation places the enzyme at a central point for cross-talk between different signaling systems. The Ca 2ϩ /calmodulin-regulated enzymes (types I, III, and VIII) are suggested to serve as coincidence detectors for signal processing in the hippocampus (types I and VIII) and in vascular smooth muscle cells (type III) (10 -13). 2 Similarly, the multiple regulatory inputs to the type II and IV enzymes would also allow the cell to process signals from diverse receptor systems. Many regulatory properties of the enzymes were determined following purification of the enzyme expressed in Sf9 insect cells, and it is not clear how the regulatory properties observed with the purified enzyme are integrated in the intact cell to produce a defined response to external stimuli. As an initial approach to this issue, we examined the integration of multiple signals to AC (G s ␣, G i ␣, G␤␥, and phosphorylation) following stable expression of enzyme types II, III, and IV in DDT 1 -MF2 cells derived from hamster vas deferens smooth muscle. (33.5 Ci/mmol) and [ 32 P]ATP (30 Ci/mmol) were purchased from NEN Life Science Products. Tissue culture supplies were obtained from JRH Bioscience (Lenexa, KS). Hygromycin B, isoproterenol, and forskolin were from Sigma. Phorbol 12-myristate 13-acetate (PMA) and UK14304 were provided by Research Biochemicals International (Natick, MA). GTP␥S was obtained from Boehringer Mannheim. Dowex 50W-X4 (100 -200 mesh, hydrogen form) was from Bio-Rad, and aluminiumoxid (90 active neutral) was purchased from EM Science (Cherry Hill, NJ). The cDNAs encoding for AC types II and III were kindly provided by Dr. Randall Reed (Department of Molecular Biology and Genetics, Johns Hopkins School of Medicine, Baltimore, MD), and the AC IV cDNA was kindly provided by Dr. Alfred Gilman * This work was supported by National Institutes of Health Grant NS24821 (to S. M. L.) and Council for Tobacco Research Grant 2235 (to S. M. L.). This paper is the fifth in the series ''Factors Determining Specificity of Signal Transduction by G-protein-coupled Receptors.'' 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 (15)(16)(17).

Materials-[ 3 H]cAMP
Cell Culture and Transfection-Cells were grown as monolayers on Falcon Primeria dishes at 37°C (5% CO 2 ) in Dulbecco's modified Eagle's medium with high glucose (4.5 g/liter), supplemented with 10% bovine calf serum (NIH-3T3), 2.5% horse serum plus 2.5% bovine calf serum (DDT 1 -MF2) or 5% horse serum plus 10% fetal bovine serum (PC-12), penicillin (100 units/ml), streptomycin (100 g/ml), and fungizone (0.25 g/ml). In some experiments, cells were pretreated with 100 ng/ml pertussis toxin for 18 h (37°C) prior to membrane preparation or initiation of experiments in intact cells. DDT 1 -MF2 cells were stably transfected with the ␣ 2A/D -AR (RG-20) as described previously (16,17), and clones expressing receptors at a density of ϳ3-4 pmol/mg of mem-brane protein were then used for manipulation of the population of adenylyl cyclases in the cell. The coding sequences of rat AC isoforms II and IV were inserted into the pMSV expression vector (17), and the gene coding for the rat AC type III was subcloned into the pcDNA3 (Invitrogen, San Diego, CA) expression plasmid. AC II (nucleotides 1-4008, coding sequence 70 -3342) in pBS.KS was digested with EcoRI and blunt-ended, and HindIII linkers were added to the 5Ј-and 3Ј-ends of the purified AC II insert. The modified insert was then cloned into the HindIII site of the pMSV expression vector downstream of the viral long terminal repeat promoter. AC III (nucleotides 1-4533, coding sequence 367-3801) in pBS.KS was digested with EcoRI and subcloned into the EcoRI site in pcDNA3 downstream of the cytomegalovirus promoter element. AC IV (nucleotides 1-3358, coding sequence 110 -3304) in pBS.SK was digested with XhoI/BamHI and cloned into the pMSV vector at the HindIII site by blunt end ligation. DDT 1 -MF2 cells expressing the ␣ 2A/D -AR were cotransfected with the AC II, AC III, or AC IV cDNA expression vector (16 g) and the drug resistance cassette pHyg (4 g) by calcium phosphate coprecipitation (17). Transfected cells were selected by their resistance to hygromycin B (750 g/ml), and hygromycin B-resistant clones were screened for expression of the AC isoforms by RNA blot analysis and functional evaluation of enzyme activity. Subsequent to isolation of AC transfectants, we performed a series of preliminary experiments to characterize the enzyme activity in different clonal cell lines. These studies used three AC II, two AC III, and two AC IV clonal cell lines. The results of these experiments indicated that the cell lines for a particular AC transfectant behaved similarly relative to the interaction between the various signaling interventions, and one cell line was then used for detailed experimental analysis.
RNA Analysis-Polyadenylated RNA from control and AC transfectants was isolated using an oligo(dT) cellulose matrix (FastTrack mRNA Isolation Kit, Invitrogen). Five to ten g of mRNA was subjected to electrophoresis on 1% agarose, 3% formaldehyde gels followed by transfer to nylon membranes. The membrane was dried in a vacuum oven at 80°C for 2 h, processed, and hybridized with radiolabeled probes as described previously (15)(16)(17)(18). The population of AC expressed by DDT 1 -MF2, NIH-3T3, and PC12 cells was also determined by DNA amplification following reverse transcription of cellular mRNA. For amplification of AC mRNA, first strand cDNA was synthesized from 2 g of mRNA using Superscript II RNase H reverse transcriptase (Life Technologies, Inc.) and random hexamer (50-ng) primers. The reaction mixture was incubated with RNase H and used for DNA amplification by the polymerase chain reaction. To allow detection of multiple types of AC transcripts, the amplification utilized degenerate oligonucleotide primers (5 mol/liter) as described by Yoshimura and Cooper (19). The primers corresponded to amino acid sequences in the C2a domain of AC that are highly conserved in the different enzyme isoforms Polyadenylated RNA was isolated from brain and three cell lines, electrophoresed, and transferred to nylon membranes as described under "Experimental Procedures." Each lane contained 10 g of polyadenylated RNA. Random-primed probes labeled with 32 P were generated from cDNAs encoding adenylyl cyclases type I-IX and were hybridized to the blot as described previously (27). Hybridization probes were as follows: type I (bovine), nucleotides 736 -3191; type II (rat), full-length (4 kb); type III (rat), full-length (4.5 kb); type IV (rat), full-length (3.3 kb); type V (rat), nucleotides 745-1204, type VI (rat), full-length (3.7 kb); type VII (DDT 1 -MF2), nucleotides 3522-3762; type VIII (rat), fulllength ϳ(4 kb); type IX (mouse), nucleotides 1-957.
ND ND a Polyadenylated RNA from each cell line was isolated and reverse transcribed using random hexamers and subsequently amplified with degenerate oligonucleotide primers as described under "Experimental Procedures." Products were isolated and subcloned, and the AC type amplified was determined by sequencing individual clones. The values in the table indicate the number of clones for each AC. The total numbers of clones sequenced were 17 (DDT1-MF2), 11 (NIH-3T3), and 14 (PC-12).
b ND, not detected.
5Ј-end to facilitate subcloning. The reaction mixture was treated with proteinase K, phenol/chloroform-extracted, and digested with BamHI and HindIII. The restricted amplified DNA of appropriate size was purified on a 2% agarose gel, subcloned into pGEM-7Zf(ϩ) (Promega), and characterized by DNA sequencing with dideoxy chain terminators and modified T7 DNA polymerase (Sequenase, U.S. Biochemical Corp.). Measurement of Cellular cAMP-Intact cell assays utilized confluent cells in six-well Falcon Primeria plates. One hour before the assay, the medium was aspirated and replaced with 2 ml of serum-free Dulbecco's modified Eagle's medium containing 20 mM NaHEPES-HCl, pH 7.5. Cells were preincubated with a phosphodiesterase inhibitor (350 M) isobutyl-L-methylxanthine at 37°C for 30 min before the addition of forskolin, isoproterenol, and/or UK14304. Reactions were terminated following a 10-min incubation by aspiration of medium and the addition of 1 ml of ice-cold 0.1 N HCl/well. In some experiments, cells were pretreated with 100 nM PMA in the presence of isobutyl-L-methylxanthine for 10 min at 37°C before adding isoproterenol and/or UK14304. cAMP was acetylated and assayed by radioimmunoassay as described by Brooker et al. (20).

RESULTS
Adenylyl Cyclase Expression in DDT 1 -MF2 Cells-We previously reported the coupling of rat ␣ 2 -AR subtypes to adenylyl cyclase following stable expression of the receptor in three cell types: DDT 1 -MF2, NIH3T3, and PC-12 (16,22). These model systems were further characterized in terms of the type of adenylyl cyclase transcripts expressed by each cell line. Poly(A)-RNA from each cell line was evaluated by RNA blot analysis using probes for enzyme types I-IX and also by RT-PCR using degenerate oligonucleotides corresponding to regions of conserved amino acid sequence in the nine enzyme types. RNA blot analysis indicated differential expression of enzyme types by the three cell lines (DDT 1 -MF2, types VI, VII, VIII, and IX; NIH-3T3, type VI; 4 and PC-12, types III, VI, VII, and IX) (Fig. 1). Brain RNA was used as a positive control for hybridization, since this tissue expressed each enzyme type. None of the cell lines expressed AC types I, II, IV, or V, whereas each cell line expressed the type VI enzyme. A similar profile of enzyme types was also observed by cDNA amplification following reverse transcriptase of mRNA (RT-PCR) from DDT 1 -MF2 and NIH-3T3 fibroblasts (Table I, Fig. 1). RT-PCR using PC-12 mRNA failed to identify the AC VII and AC IX transcript  (Table I, Fig. 1). 5 DDT 1 -MF2 cells were selected to evaluate signal integration by AC before and after altering the population of enzymes expressed in the cell. AC types II, III, and IV were introduced into the cell by gene transfection. Stimulation of AC by G s ␣ was achieved by activation of the ␤-AR expressed in DDT 1 -MF2 cells (23). To generate the inhibitory input to the enzyme, we stably expressed the AC cDNA in subclones previously transfected with the ␣ 2A/D -AR (16). The influence of G s -and G iprotein coupled systems on AC function for each enzyme type was determined before and after activation of protein kinase C with phorbol 12-myristate 13-acetate.
AC types II, III, and IV were stably expressed in DDT 1 -MF2 cells as indicated by analysis of transfectant RNA and functional properties of the transfected cells (Fig. 2). RNA blot analysis indicated that mRNA of expected size was detected in the AC transfectants but not in control cells transfected with vector alone (Fig. 2A). Basal and maximally stimulated enzyme activity in the DDT 1 -MF2 AC transfectants was compared with that of control vector-transfected cells. Relative to control cells, basal enzyme activity was not altered in AC III or AC IV transfectants, but it was increased 6-fold in AC II transfectants (Fig. 2B). Enzyme activity determined in the presence of forskolin, Mn 2ϩ , and GTP␥S was increased ϳ2-fold in AC III and AC IV transfectants and ϳ6-fold in AC II transfectants relative to the stimulation observed in control transfectants (Fig. 2B). These data indicated that each of the transfected genes was expressed in the DDT 1 -MF2 cell lines.
Influence of Forskolin and Receptors Coupled to G s on Adenylyl Cyclase Activity in DDT 1 -MF2 AC Transfectants-The AC transfectants were further characterized for their responses to the diterpene forskolin and the ␤-AR agonist isoproterenol. Basal enzyme activity in the intact cell was similar in control, AC III, and AC IV transfectants but was slightly increased in AC II transfectants (Table II). In the intact cell, the effect of forskolin on cellular cAMP levels was not altered in AC IV transfectants but was increased in AC II (1.5-fold) and AC III (1.8-fold) transfectants relative to its actions in control transfectants (Fig. 3A). The EC 50 of forskolin in the AC transfectants was similar to the control value. Stimulation of adenylyl cyclase via G s ␣ in DDT 1 -MF2 transfectants was achieved with the ␤-AR agonist isoproterenol. The maximal effect of isoproterenol in the intact cell was only increased in AC II transfectants (229 Ϯ 0.3 versus 168 Ϯ 0.6 ng of cAMP/mg of protein) (Fig. 3B). The EC 50 of isoproterenol for enzyme stimulation in control, AC II, AC III, and AC IV transfectants was 99 Ϯ 6, 146 Ϯ 1, 24 Ϯ 1, and 90 Ϯ 2 nM, respectively.
The G s ␣-mediated stimulation of adenylyl cyclase in the AC transfectants was also investigated after the activation of protein kinase C. Protein kinase C activation by PMA did not alter basal cellular cAMP levels in control transfectants and pro-duced small but consistent increases in cAMP in each AC transfectant (Table II). In membrane preparations, only the AC II transfectant exhibited an elevation in basal enzyme activity (relative to control values) following PMA treatment of the cells (Table III). Protein kinase C activation also enhanced the increase in cellular cAMP elicited by the ␤-AR agonist isoproterenol in each of the AC transfectants (Table II). Similar results were observed in enzyme activity assays using membrane preparations (Table III). These data indicate that the receptor has 5 M. Sato and S. M. Lanier, AC III, VI, VII, and IX were identified by RNA blot analysis of PC-12 mRNA, whereas only AC III and VI were identified by RT-PCR.  access to the transfected enzymes and that the sensitivity of the enzymes to G s ␣ is increased following enzyme phosphorylation by protein kinase C. To gain insight as to how the different enzyme types process multiple stimulatory and inhibitory inputs, we determined the interaction between the G s -coupled ␤-AR and the G i /G o -coupled ␣ 2A/D -AR in the regulation of AC activity.

Influence of Receptors Coupled to G i /G o on Adenylyl Cyclase Activity in DDT 1 -MF2 AC Transfectants-Receptors coupled to
G i -or G o -proteins are capable of sending both positive and negative signals to adenylyl cyclases, as is the case with the ␣ 2A/D -AR. Negative input is relayed through G i ␣ or perhaps G o ␣ and is commonly observed following activation of the ␣ 2A/ D-AR. Positive input to adenylyl cyclases following ␣ 2A/D -AR activation is cell type-specific and may be achieved by 1) increased intracellular calcium concentration and subsequent stimulation of Ca 2ϩ /calmodulin-sensitive adenylyl cyclases (16), 2) receptor coupling to G s ␣ (24), 3) altered phosphorylation status of the enzyme via activation of protein kinase C, or 4) the action of G␤␥ when the enzyme is stimulated by G s ␣ (18,25,26). The results of experiments in the AC transfectants indicated that the effects of ␣ 2A/D -AR receptor activation on cellular cAMP levels were dependent upon the enzyme type expressed in the cell. Basal enzyme activity in control and AC III was either not altered (intact cell) or inhibited (membranes) by the selective ␣ 2 -AR agonist UK14304 (Table IV). In AC II and AC IV transfectants, AC activity was either not altered (AC IV) or increased (AC II) by ␣ 2A/D -AR activation (Table IV). ␣ 2A/D -AR activation inhibited forskolin-stimulated enzyme activity in control and each of the AC transfectants (Fig. 4A). The effects of ␣ 2A/D -AR activation on basal and forskolin-stimulated enzyme activity in control and each AC transfectant were blocked by prior treatment of the cells with pertussis toxin 6 and thus involved receptor coupling to G i 2 or G i 3, members of the pertussis toxin-sensitive family of G-proteins expressed in DDT 1 -MF2 cells (16).
Stimulation of the three types of AC via the ␤-AR agonist isoproterenol was differentially altered by activation of ␣ 2A/D -AR. In AC III transfectants, ␣ 2A/D -AR-mediated inhibition of isoproterenol-induced increases in cAMP was enhanced rela-tive to control (80 versus 50% inhibition) (Fig. 4B). AC III is most closely related to AC I and AC VIII in terms of sequence homology and each of these enzymes are regulated by calcium/ calmodulin (27)(28)(29). These data suggest that G s ␣-stimulated AC III is more sensitive to G i ␣-mediated inhibition compared with the enzymes coupled to ␣ 2A/D -AR in control transfectants or that perhaps there is an additional inhibitory input from G␤␥ (30,31). In AC II and AC IV transfectants, ␣ 2A/D -AR activation elicited a biphasic effect on isoproterenol-induced increases in cellular cAMP (Fig. 4B). At lower concentrations of the ␣ 2 -AR agonist, the effect of isoproterenol on enzyme activity was inhibited, whereas at higher concentrations of UK14304, the response to the ␤-AR agonist was either unaltered or enhanced. Thus, it appears as if ␣ 2A/D -AR activation initiated both a positive and negative signal to AC II and AC IV and that the final cell response was a balance between the two types of input. Both types of input emanating from the receptor required G i 2/G i 3, since no effect of receptor activation was observed in cells pretreated with pertussis toxin. 6 The inhibitory component represented the influence of G i ␣, whereas the stimulatory component probably involves G␤␥ further activating the G s ␣-primed enzyme. Indeed, the stimulatory component was blocked when membranes were preincubated with transducin ␣, which ties up the G␤␥ released by receptor activation (Fig. 5). In control transfectants, the inhibition of AC by ␣ 2A/ D-AR activation was not altered by the addition of transducin ␣. In AC III transfectants, the inhibitory input mediated through the ␣ 2A/D -AR was actually diminished. However, in AC II and AC IV transfectants, the augmentation of isoproterenol-induced increases in AC following ␣ 2A/D -AR activation was lost in the presence of G t ␣, and the inhibition of the enzyme by this system was revealed (Fig. 5).

Influence of Receptors Coupled to G i /G o on Adenylyl Cyclase Activity in DDT 1 -MF2 AC Transfectants following Activation of
Protein Kinase C-The preceding series of experiments indicated that the AC III and AC II/IV enzymes process information from G i /G o -coupled receptors in an enzyme type-specific manner. Additional differences in the integrative properties of the three enzymes were observed after activation of protein kinase C. The consequences of ␣ 2A/D -AR activation on cellular cAMP were not altered by cell treatment with PMA in either control or AC III transfectants (Fig. 6). However, although AC  II and AC IV exhibit sequence homology and share several biochemical properties, the stimulatory/inhibitory input to the enzymes generated by ␣ 2A/D -AR activation were differentially modified by PMA. Treatment of DDT 1 -MF2 AC II transfectants with PMA eliminated the inhibitory input and enhanced the stimulatory input to the enzyme initiated by ␣ 2A/D -AR activation (Fig. 6). PMA treatment of AC IV transfectants actually modified the effects of ␣ 2A/D -AR activation in a manner that was directly opposite to that observed in AC II transfectants. Treatment of DDT 1 -MF2 AC IV transfectants with PMA eliminated the stimulatory component and enhanced the inhibitory effect of ␣ 2A/D -AR activation (Fig. 6). Similar alterations in the input to AC II and AC IV mediated by the ␣ 2A/D -AR were also observed in membranes prepared from cells pretreated with PMA (Fig. 7). Since treatment of control transfectants with PMA did not alter the effect of ␣ 2A/D -AR activation on AC activity, the effects observed in the AC II and AC IV transfectants probably reflect phosphorylation of the transfected enzyme type. These data further suggest that the AC II and AC IV enzymes received two types of input from the activated ␣ 2A/D -AR. The results observed after PMA treatment in AC II transfectants could reflect a complete elimination of the inhibitory input or a sensitization of the stimulatory component, whereas protein kinase C activation in AC IV transfectants would result in an opposite modification either enhancing the inhibitory input or preventing the stimulatory influence of G␤␥ resulting from ␣ 2A/D -AR activation. DISCUSSION Most cells probably express multiple types of adenylyl cyclases. Indeed, RNA blot analysis indicated that DDT 1 -MF2, NIH-3T3, and PC-12 cells express transcripts encoding distinct populations of adenylyl cyclases. None of the cell lines expressed detectable mRNA for AC types I, II, IV, or V. The absence of type I in the three cell lines is consistent with the neural specific expression of this enzyme, whereas type V is enriched in the myocardium and brain (1,32). Types II and IV were also not detected in the three cell lines, although these enzymes are fairly widely distributed in peripheral tissues (1). AC III mRNA was identified in the PC-12 cell line, and the presence of this enzyme may explain previous observations indicating that ␣ 2A/ D-AR activation in these cells augmented forskolin-stimulated enzyme activity in a calcium-dependent manner (16). Although the type III mRNA is enriched in the olfactory system, this enzyme is also expressed in several other cell types including vascular smooth muscle cells. 2 DDT 1 -MF2 cells expressed AC VI-IX. The type VI enzyme is widely expressed, and mRNA encoding this enzyme type was also present in NIH-3T3 and PC-12 cells. The type VII enzyme exhibits highest homology to AC II/IV (9). The expression of the type VIII transcript in DDT 1 -MF2 cells, which is related to AC I in terms of sequence homology and its regulation by calcium/calmodulin, was surprising, since it was thought that this enzyme was restricted to the central nervous system (33). The type IX cDNA was recently isolated and characterized by Premont et al. (34). It is not known if all of the receptors in a given cell that initiate stimulatory and/or inhibitory input to AC communicate with the same population of adenylyl cyclases nor how the different enzymes process the multiple types of input they receive. As an initial approach to these issues, we explored the function of receptors coupled to adenylyl cyclase before and after altering the population of effector enzymes in the cell. In the present study, we focused on the enzyme types II, III, and IV.
The type II and IV enzymes exhibit 76 -79% amino acid homology in their conserved cytoplasmic domains, lack the C2b domain, and are both stimulated by G␤␥ in the presence of activated G s ␣. Of the two enzymes, the type II enzyme is the best characterized, and the interaction of G␤␥ with AC II involves a defined segment of the C2a domain (35). ␣ 2A/D -AR activation resulted in both stimulatory and inhibitory input to the AC II and AC IV enzymes. In AC II transfectants the stimulatory input was enhanced, and the inhibitory input was lost following treatment of the cells with PMA. Protein kinase C phosphorylation of the enzyme apparently alters its interaction with the inhibitory component, i.e. G i ␣. These data and the observations of Chen and Iyengar (36) are in contrast to the failure of G i ␣ to inhibit purified AC II enzyme (37), suggesting that there are unique regulatory properties of the AC enzymes in the intact cell. In contrast to the effects of PMA on ␣ 2A/D -AR coupling to AC II, it is apparently the stimulatory input (mediated by G␤␥) to the AC IV enzyme from the ␣ 2A/D -AR that is modified by protein kinase C. This effect of protein kinase C activation is likely achieved by direct phosphorylation of the  enzyme, since signaling events initiated by both ␣ 2 and ␤-AR were not altered by PMA in control transfectants. Although phosphorylation of AC II and AC IV by protein kinase C sensitized the enzymes to activated G s ␣, the input to the enzyme from G i ␣ and from G␤␥ were differentially affected by protein kinase C activation. These data support the idea that there are distinct sites on the enzyme for the action of the three types of input and suggests that these sites are independently regulated by phosphorylation.
␣ 2A/D -AR activation initiates both positive and negative signals to AC II and AC IV, and the final cell response is apparently a balance between the two types of input. The significance of the differential regulation of these two types of inputs from the receptor to AC II and AC IV by protein kinase C would be particularly evident in cells expressing multiple types of catecholamine receptors. Thus, in cells expressing the type II enzyme, ␤-AR, and ␣ 2 -AR, the adrenergic agonist epinephrine could exert precisely opposite effects on cellular cAMP, depending on the phosphorylation state of the enzyme. When the enzyme is not phosphorylated by protein kinase C, ␣ 2 -AR activation would inhibit enzyme activity, but following enzyme phosphorylation, receptor activation would actually increase enzyme activity. The stimulation in the latter situation reflects synergistic activation of the enzyme by G s ␣ (via ␤-AR stimulation) and G␤␥ (via ␣ 2 -AR activation) as well as the loss of G i ␣ inhibition subsequent to protein kinase C phosphorylation of the enzyme. A switch in the stimulatory versus inhibitory input to AC from the ␣ 2 -AR is precisely the situation that occurs in the rat myometrium over the course of pregnancy (18). Indeed, mRNA blot analysis indicates the expression of AC II in the longitudinal layer of the rat myometrium throughout the time course of pregnancy (14). At early stages of pregnancy, when it is important to maintain a relaxed state of the myometrium, ␣ 2 -AR activation augments the effects of isoproterenol on cellular cAMP, promoting smooth muscle relaxation, and this effect of ␣ 2 -AR agonists is mediated through G␤␥. At later stages of pregnancy, when contraction is important at term, ␣ 2 -AR activation inhibits stimulation of AC by the ␤-AR agonist. Such a switch in the consequences of ␣ 2 -AR activation could be explained by a change in the phosphorylation status of the AC II enzyme. Thus, the integration of multiple inputs by AC is dynamic, and this ability of the enzyme probably plays a key role in the signaling plasticity observed in development and a wide range of physiological processes.