Differential regulation of type I and type VIII Ca2+-stimulated adenylyl cyclases by Gi-coupled receptors in vivo.

Coupling of intracellular Ca2+ to cAMP increases may be important for some forms of synaptic plasticity. The type I adenylyl cyclase (I-AC) is a neural-specific, Ca2+-stimulated enzyme that couples intracellular Ca2+ to cAMP increases. Since optimal cAMP levels may be crucial for some types of synaptic plasticity, mechanisms for inhibition of Ca2+-stimulated adenylyl cyclases may also be important for neuroplasticity. Here we report that Ca2+ stimulation of I-AC is inhibited by activation of Gi-coupled somatostatin and dopamine D2L receptors. This inhibition is due primarily to Giα and not βγ subunits since coexpression of βγ-binding proteins with I-AC did not affect somatostatin inhibition. However, βγ released from Gs did inhibit I-AC, indicating that the enzyme can be inhibited by βγ in vivo. Interestingly, type VIII adenylyl cyclase (VIII-AC), another Ca2+-stimulated adenylyl cyclase, was not inhibited by Gi-coupled receptors. These data indicate that I-AC and VIII-AC are differentially regulated by Gi-coupled receptors and provide distinct mechanisms for interactions between the Ca2+ and cAMP signal transduction systems. We propose that I-AC may be particularly important for synaptic plasticity that depends upon rapid and transient cAMP increases, whereas VIII-AC may contribute to transcriptional-dependent synaptic plasticity that is dependent upon prolonged, Ca2+-stimulated cAMP increases.

defined. Inhibition of the Ca 2ϩ /CaM-stimulated adenylyl cyclases may be important during synaptic plasticity since optimal, not necessarily maximal, cAMP levels are necessary for certain forms of learning (4). Specifically, the sensitivity of Ca 2ϩ -stimulated I-AC or VIII-AC to G i -coupled receptor activation has not been reported. Although I-AC is inhibited by G i␣ in vitro (29,30), regulatory mechanisms determined with purified enzymes or membrane preparations are not necessarily operative in vivo. For example, I-AC is stimulated by activated G s␣ in membranes; however, it is not stimulated by G s -coupled receptor activation in vivo (31). Additionally, G s stimulation of III-AC is stimulated by Ca 2ϩ /CaM in vitro (32); however, in vivo, elevations of intracellular Ca 2ϩ inhibit G s stimulation of III-AC (19,20). Therefore, it was important to determine the responses of I-AC and VIII-AC to G i -coupled receptors in intact cells. In this study, we report that I-AC, but not VIII-AC, is inhibited by activation of somatostatin or dopamine D 2 L receptors in HEK 293 cells. A comparison of I-AC with VIII-AC indicates that these two Ca 2ϩ -stimulated enzymes have distinct regulatory properties.
Cell Culture-Human embryonic kidney 293 (HEK 293) cells were grown at 37°C in HEPES-buffered Dulbecco's modified Eagle's medium (H-DMEM) supplemented with 10% fetal bovine serum in a humidified 95% air, 5% CO 2 incubator. HEK 293 cells stably expressing the long form of the dopamine D 2 receptor (293-D 2 L) were generously provided by Dr. Kim Neve (Veterans Association Medical Center, Portland, OR). 293-D 2 L cells require maintenance in 2 g/ml puromycin (Aldrich). Cell culture materials were from Life Technologies, Inc. unless otherwise indicated.
Expression of I-AC and VIII-AC in HEK 293 Cells-The I-AC cDNA clone was isolated from a bovine brain cDNA library as described (28). The cDNA clone for VIII-AC was generously provided by Dr. John Krupinski (Weis Center for Research, Geisinger Clinic, Danville, PA). Polyclonal populations of hygromycin (Calbiochem; 500 units/ml)-resistant 293 cells or 293-D 2 L cells were obtained by stable transfection of the pCEP4 expression vector (Invitrogen), pCEP4-IAC, or pCEP4-VII-IAC by the calcium phosphate method (33). All stable cell lines were created from the same parental population of HEK 293 cells or 293-D 2 L cells. Expression of transfected adenylyl cyclases was determined by cAMP accumulation assays as described below.
cAMP Accumulation Assay-Changes in intracellular cAMP levels were measured by determining the ratio of [ 3 H]cAMP to a total ATP, ADP, and AMP pool in [ 3 H]adenine-loaded cells as described (34). This assay system allows for rapid and sensitive determination of relative changes in intracellular cAMP levels. While the ratios measured between assays may show some variation, the relative changes in cAMP levels between assays is quite reproducible. Briefly, as cells in six-well plates approached confluency (ϳ90%), they were incubated in H-DMEM ϩ 10% fetal bovine serum containing 2 Ci/ml [ 3 H]adenine (ICN) for 16 -20 h. The next day, cells were aspirated, washed once with 150 mM NaCl, and incubated in buffer A (20 mM HEPES, pH 7.4, 120 mM NaCl, 5 mM KCl, 2.5 mM MgSO 4 , 2 mM CaCl 2 , 10 mM glucose, and * 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. 2 mM sodium phosphate) or in H-DMEM ϩ 1% penicillin/streptomycin containing the indicated effectors (e.g. A23187, carbachol, dopamine, forskolin, isoproterenol, serotonin, and somatostatin) plus 1 mM 3-isobutyl-1-methylxanthine. For extracellular Ca 2ϩ dose responses, buffer A was made with varying concentrations of CaCl 2 . Reactions were terminated by aspiration and the addition of 1 ml of ice-cold 5% trichloroacetic acid, 1 M cAMP. Culture dishes were maintained at 4°C for 1-4 h, and acid-soluble nucleotides were separated by sequential Dowex AG 50WX4 and neutral alumina chromatography as described (35). Reported data are the averages of triplicate determinations. Pertussis toxin (List Biological Laboratories Inc.), when used, was added to cells along with [ 3 H]adenine for 16 -20 h.
Transient Coexpression of I-AC with the Carboxyl Terminus of ␤-Adrenergic Receptor Kinase 1 or Transducin-␣ in HEK 293 Cells-The peptide minigene construct encoding the carboxyl terminus of ␤-adrenergic receptor kinase 1 (␤ARK1-ct) in the pRK5 plasmid (36) was generously provided by Dr. Robert Lefkowitz (Duke University Medical Center, Durham, NC). A cDNA clone for the ␣ subunit of human rod transducin was provided by Dr. Neil M. Nathanson (University of Washington, Seattle). The G s -coupled serotonin receptor 5HT-7 cDNA (37) was a gift from Dr. Mark Hamblin (Veterans Association Medical Center, Seattle). Briefly, the night before transfection, cells were plated in 100-mm plates at 70% density. The following morning, each plate was transfected with 8 g of total DNA (1 g of 5HT-7; 2.5 g of pCEP4, pCEP4-IAC, or pCEP4-IVAC; 4 g of pCDNAIII (Invitrogen), pCDNAIII-transducin-␣, or pRK-␤ARK1-ct; and 0.5 g of RSV-␤-galactosidase) in H-DMEM in the presence of 50 -60 l of Lipofectamine (Life Technologies, Inc.). After 5 h, cells were rinsed with H-DMEM ϩ 1% penicillin/streptomycin ϩ 10% fetal bovine serum and maintained for 24 h under normal conditions. The following day, cells were split, pooled by transfection, and seeded into six-well culture dishes (one transfected plate/six-well dish) for cAMP assays as well as into 12-well plates (two wells/transfection) for ␤-galactosidase assays. The next morning, cells used for cAMP assays were labeled for 4 -6 h with 2-3 Ci/well [ 3 H]adenine. Just prior to the cAMP assay, companion cells for ␤-galactosidase assays were lysed in 500 l of buffer B (100 mM KH 2 PO 4 , pH 7.8, 0.2% Triton X-100, and 1 mM dithiothreitol) and frozen until use. cAMP and ␤-galactosidase assays were carried out as described below, and all raw data were normalized to the measured ␤-galactosidase signal for each transfection.
␤-Galactosidase Assay-Lysates of transiently transfected cells (described above) were thawed and centrifuged at 16,000 ϫ g for 10 min. 20 l of the supernatant was combined with 100 l of reaction buffer (100 mM Na 2 HPO 4 , pH 8.0, 1 mM MgCl 2 , 35 mM Galacton (Tropix Inc., Bedford, MA), and 100 mM D-galactose) and incubated in the dark at room temperature for 60 min. During this incubation period, a 10% solution of Emerald (Tropix Inc.) in 0.2 N NaOH was prepared for subsequent addition to the samples at 5-s intervals by a Berthold luminometer. Each well of lysed cells was assayed in duplicate, and data were used to normalize for transfection efficiency.
Primary Neuron Cultures-Primary cortical and hippocampal neurons were cultured essentially as described by Impey et al. (38) and maintained in NeuroBasal medium supplemented with B27. Briefly, dissected brain regions were minced in dissociation medium (75 mM Na 2 SO 4 , 3.3 mM K 2 SO 4 , 15 mM MgCl 2 , 0.225 mM CaCl 2 , 18 mM glucose, 4 mM kynurenic acid, and 2.0 mM HEPES, pH 7.4) and incubated in dissociation medium containing 100 units/ml activated papain (Worthington) with gentle agitation at 37°C for 30 min. The disrupted tissue was then rinsed with dissociation medium, and cells were dissociated by gentle trituration through a 2-ml serological pipette. Cells were plated at a density of 5 ϫ 10 5 cells/well in 12-well culture dishes coated with poly-D-lysine (50 -66 g/ml).

Inhibition of Ca 2ϩ -stimulated I-AC by G i -coupled Receptors
in Vivo-To examine the effect of G i -coupled receptors on I-AC in vivo, we carried out an initial screen of HEK 293 cells to determine if they express endogenous inhibitory receptors that couple to adenylyl cyclases. cAMP levels were elevated 6-fold by activation of endogenous ␤-adrenergic receptors with isoproterenol (31). Application of 1 M somatostatin inhibited isoproterenol stimulation of the endogenous adenylyl cyclase(s) by ϳ50%, indicating the presence of somatostatin receptors (data not shown). Somatostatin receptors are seven-transmembrane domain proteins that typically couple to the G i /G o class of heterotrimeric G proteins (39,40).
HEK 293 cells do not express Ca 2ϩ -stimulated adenylyl cyclases, and treatment with the Ca 2ϩ ionophore A23187 or the muscarinic receptor agonist carbachol does not elevate intracellular cAMP (41). However, carbachol or A23187 activates I-AC expressed in HEK 293 cells by increasing intracellular Ca 2ϩ (41). A23187 elevated intracellular cAMP levels ϳ25-fold, and somatostatin inhibited Ca 2ϩ -stimulated activity, with an IC 50 of ϳ10 nM (Fig. 1A). Carbachol stimulation of I-AC was also inhibited ϳ50% by 1 M somatostatin (data not shown). Somatostatin inhibition of I-AC was not limited to Ca 2ϩ -stimulated activities; forskolin-stimulated I-AC was inhibited 41% by 500 nM somatostatin (data not shown). To determine if other G i -coupled receptors also couple to inhibition of I-AC, HEK 293 cells expressing the dopamine D 2 L receptor were stably trans- fected with I-AC. In these cells, dopamine maximally inhibited I-AC activity by 75%, with an IC 50 of 50 nM (Fig. 1B).
The effect of somatostatin on Ca 2ϩ -stimulated I-AC activity was examined at several concentrations of A23187 ( Fig. 2A) and extracellular Ca 2ϩ (Fig. 2B) to determine if somatostatin affected the Ca 2ϩ sensitivity of I-AC. Ca 2ϩ -stimulated I-AC activity was inhibited by ϳ50% at all concentrations of A23187 and Ca 2ϩ examined. These data indicate that somatostatin inhibits I-AC without affecting its Ca 2ϩ sensitivity.
Pertussis Toxin Abolishes Inhibition of I-AC by G i -coupled Receptors-To determine whether inhibition of I-AC by somatostatin or dopamine is mediated via the G i /G o class of G proteins, HEK 293 cells expressing I-AC were treated with pertussis toxin. This toxin catalyzes the ADP-ribosylation of the ␣ subunit of G i /G o (42)(43)(44)(45)(46) and uncouples G i /G o from its receptors (47). Somatostatin inhibition of I-AC was completely abolished by pertussis toxin (Fig. 3A). Furthermore, dopaminergic inhibition of Ca 2ϩ -stimulated I-AC was pertussis toxinsensitive (Fig. 3B). Similar results were obtained when carbachol was used to stimulate I-AC (Fig. 4, A and B). These data indicate that somatostatin or dopamine inhibition of I-AC is most likely mediated by G i or G o .
Cellular Expression of ␤␥-binding Proteins Does Not Affect G i -coupled Receptor Inhibition of I-AC-Because Ca 2ϩ stimulation of I-AC was inhibited by G i -coupled receptors in vivo, we were interested in assessing the role of the G protein ␤␥ subunits in hormonally driven inhibition since both G i␣ and ␤␥ inhibit I-AC in vitro (26,30,48,49). To accomplish this, we carried out transient transfections of HEK 293 cells in which the COOH-terminal ␤␥-binding region of ␤-adrenergic receptor kinase 1 (␤ARK1-ct) or the ␣ subunit of human rod transducin (G t␣ ) was coexpressed with I-AC. IV-AC was used as a positive control for the expression of ␤␥-binding proteins. Since G s␣ stimulation of IV-AC is potentiated by ␤␥ (50), effective expression of ␤␥-binding proteins should block this potentiation. Cellular expression of "peptide minigenes" encoding ␤␥-binding pleckstrin homology domains of various proteins attenuates ␤␥ effects on phospholipase C, II-AC, and the mitogen-activated protein kinase pathway (36,51,52). Additionally, G t␣ is an effective scavenger of free ␤␥ subunits in intact cells (53,54).
In transient transfection experiments in which the G s -coupled 5HT-7 receptor was cotransfected, somatostatin potentiated serotonin stimulation of IV-AC, presumably through release of ␤␥ from G i (Fig. 5A). Coexpression of ␤ARK1-ct with IV-AC attenuated somatostatin-mediated potentiation of serotonin-stimulated IV-AC almost entirely (Fig. 5A). Under the same transfection conditions, there was no effect of ␤ARK1-ct expression on somatostatin inhibition of Ca 2ϩ -stimulated I-AC (Fig. 5B). Similarly, coexpression of G t␣ with IV-AC attenuated somatostatin potentiation of serotonin-stimulated IV-AC (Fig.  6A), while somatostatin inhibition of A23187-stimulated I-AC was unaffected by coexpression of G t␣ (Fig. 6B). These results indicate that the primary mechanism for G i -mediated inhibition of I-AC was through G i␣ rather than ␤␥ release from dissociating G i /G o heterotrimers. However, we cannot rule out the possibility that a ␤ARK1-ct⅐␤␥ or G t␣ ⅐␤␥ complex was still capable of modulating I-AC, but not IV-AC.
Coexpression of ␤␥-binding Proteins Elicits a Modest Stimulation of I-AC by G s -coupled Receptor Activation-Activation of G s -coupled receptors does not stimulate I-AC in vivo unless intracellular Ca 2ϩ is simultaneously elevated (31). Since G s␣ stimulation of I-AC in vitro can be inhibited by ␤␥ (26,30,48,49,55) and since receptor activation of G s releases ␤␥, is it possible that the insensitivity of I-AC to G s -coupled receptor activation is due to ␤␥ inhibition? To address this issue, we examined the sensitivity of I-AC to activation of the G s -coupled 5HT-7 receptor in vivo when ␤␥-binding proteins were coexpressed. Coexpression of G t␣ with I-AC elicited a substantial (ϳ4-fold) stimulation of I-AC by serotonin (Fig. 7B). Experiments in which ␤ARK1-ct was used as the ␤␥ scavenger gave similar results (Fig. 7D). In both cases, effective expression of ␤␥ scavengers was determined with IV-AC as described (Figs.  7, A and C). These data suggest that ␤␥ release from dissociating G s heterotrimers inhibits I-AC.
Ca 2ϩ Stimulation of VIII-AC Is Not Inhibited by G i -coupled Receptors-To determine whether Ca 2ϩ stimulation of VIII-AC is also regulated by G i -activating hormones, VIII-AC-expressing cells were treated with 500 nM somatostatin in the presence of increasing concentrations of A23187. VIII-AC was stimulated ϳ7-fold by 10 M A23187 in the presence of 1.8 mM extracellular Ca 2ϩ (Fig. 8). Concentrations of somatostatin as high as 500 nM, which inhibited I-AC, did not inhibit VIII-AC at any concentration of A23187 (Fig. 8) or extracellular Ca 2ϩ (data not shown) examined. Stimulation of VIII-AC in vivo by forskolin was also insensitive to somatostatin (data not shown).
To determine if VIII-AC is insensitive to other G i -coupled receptors, cells stably coexpressing the dopamine D 2 L receptor with VIII-AC were treated with increasing concentrations of dopamine in the presence or absence of A23187. Ca 2ϩ -stimulated VIII-AC was only inhibited 15% by 1 M dopamine (Fig.  9). The insensitivity to somatostatin but slight inhibition by dopamine probably reflect differences in receptor density. The density of endogenous somatostatin receptors in HEK 293 cells is ϳ18 fmol/mg of protein (56), while the D 2 L cells used in this study express ϳ1500 -2000 fmol/mg of protein. 2 The D 2 L receptor density in the cell lines used in this study was ϳ10-fold higher than the D 2 receptor density in the striatum, which expresses the highest levels of D 2 receptors in the brain (57). Since the same stock of HEK 293 cells or dopamine D 2 L receptor-expressing cells was used for expression of I-AC or VIII-AC, variation in the number of somatostatin or D 2 L receptors present in I-AC-or VIII-AC-expressing cells cannot account for the difference in G i sensitivity. Furthermore, somatostatin did not inhibit VIII-AC in four independent HEK 293 cell lines stably expressing VIII-AC. These data strongly suggest that VIII-AC is insensitive to G i -coupled receptor stimulation in vivo. G s and Ca 2ϩ Do Not Synergize to Stimulate VIII-AC-Since I-AC and VIII-AC respond quite differently to G i -coupled receptors in HEK 293 cells, it was of interest to compare the sensitivities of these two enzymes to G s -coupled receptors and Ca 2ϩ in vivo. Both I-AC and VIII-AC were stimulated by Ca 2ϩ alone, but neither was activated by isoproterenol, a ␤-adrenergic agonist (Fig. 10). I-AC was stimulated by isoproterenol when intracellular Ca 2ϩ was elevated with A23187. In contrast, isoproterenol did not produce any additional stimulation of VIII-AC beyond that caused by A23187 alone (Fig. 10). Stimulation of I-AC by coapplication of Ca 2ϩ and isoproterenol was strongly inhibited by somatostatin, whereas it had no effect on VIII-AC activities (data not shown).

Somatostatin Inhibits Ca 2ϩ -stimulated cAMP Accumulation in Primary Cortical and Hippocampal Neuron Cultures-Since
I-AC is neural-specific (25) and expressed in the cortex and hippocampus (28), we were interested in determining whether or not somatostatin inhibits Ca 2ϩ -stimulated cAMP levels in cultured neurons. Primary cultures of cortical and hippocam-2 V. Watts and K. Neve, personal communication.

FIG. 6. Effect of transducin-␣ expression on somatostatin inhibition of I-AC. HEK 293 cells were transiently transfected with both
RSV-␤-galactosidase and the 5HT-7 receptor and either pCDNAIII or G t␣ as well as pCEP4, pCEP4-I-AC, or pCEP4-IV-AC as described under "Experimental Procedures." Cells transfected with G t␣ are denoted as Gt. A, IV-AC-transfected cells with or without G t␣ coexpressed were treated with 10 M serotonin (5-hydroxytryptamine (5-HT)) in the presence or absence of 500 nM somatostatin (SOM). B, I-AC-transfected cells with or without G t␣ coexpressed were treated with 5 M A23187 in the presence or absence of 500 nM somatostatin. Data are expressed as percent cAMP accumulation in the absence of somatostatin, with this level being set as 100%. The -fold stimulation over the basal activities was similar between transfections with or without transducin-␣. The data are the means Ϯ S.D. of triplicate determinations and were subtracted for endogenous (pCEP4 transfectants) cAMP accumulation and corrected for transfection efficiency using ␤-galactosidase expression. pal neurons were prepared from day 1 rat pups. Treatment of these cultures with 1 M somatostatin produced a substantial inhibition of cAMP levels stimulated with 1 M A23187, demonstrating that Ca 2ϩ -stimulated cAMP synthesis in the hippocampus and cortex is sensitive to G i -coupled hormones (Fig. 11). DISCUSSION Although Ca 2ϩ stimulation of specific adenylyl cyclases in the brain may contribute to neuroplasticity (3,(5)(6)(7)(8)28), mechanisms for inhibition of adenylyl cyclases may be equally important. In most cells, increases in cAMP in response to extracellular and intracellular signals are transient. Mechanisms for attenuation of cAMP signals include inhibition of adenylyl cyclases by G i -coupled receptors (18), CaM kinase inhibition of specific adenylyl cyclases (19), and cAMP hydrolysis by phosphodiesterases (58). The objective of this study was to determine if I-AC or VIII-AC is inhibited by G i -coupled receptors in vivo.
In this study, we discovered that I-AC is inhibited in vivo by activation of G i -coupled receptors including somatostatin and dopamine D 2 L receptors. Inhibition was apparently mediated through the G i /G o class of heterotrimeric G proteins since it was blocked by pertussis toxin. In particular, we presume that inhibition was mediated by G i because G i␣1 and G i␣3 are expressed at relatively high levels in HEK 293 cells compared with G o␣2 (56). Furthermore, G o␣ inhibition of I-AC in vitro is 10-fold less potent than that elicited by G i␣ (30). Mechanisms for receptor-coupled inhibition of I-AC in vivo have not been defined, but could include contributions from activated G i␣ and/or ␤␥ since both have been shown to inhibit I-AC in vitro (26,29,30,49,55). Our data suggest that receptor-stimulated inhibition of I-AC is due primarily to activated G i␣ since coexpression of ␤␥-binding proteins did not prevent somatostatin inhibition. Parallel experiments with IV-AC coactivated with serotonin and somatostatin illustrated that ␤␥ was effectively inhibited by coexpression of the same ␤␥-binding proteins, ␤ARK1-ct or G t␣ .
Release of inhibitory ␤␥ from G s may explain why I-AC is not activated by G s -coupled receptors in vivo. In fact, sequestration of ␤␥ released from G s caused stimulation of I-AC by G s -coupled receptor activation in vivo. The differing effects of ␤␥ sequestration following G s versus G i receptor activation could be because of distinct ␤␥ subunit compositions for G s or G i . Interestingly, using the yeast two-hybrid system, Yan and Gautam (59) have shown that there are substantial differences in the interactions between II-AC and ␤1-5. In particular, ␤1 interacts with II-AC three to four times more strongly than ␤4 and at least twice as strongly as ␤3 and ␤5. It may be that different ␤ isoforms determine, in large part, the potency with which a given ␤␥ complex affects adenylyl cyclases in vivo. Alternatively, ␤␥ release from G i may be functionally inconsequential with regard to I-AC since the dominant mechanism for G imediated inhibition is through the release of G i␣ . In any case, these data indicate that ␤␥ and G i␣ can both inhibit I-AC activity in vivo.
Mammalian brain expresses at least two CaM-stimulated adenylyl cyclases, I-AC and VIII-AC, both of which are activated by intracellular Ca 2ϩ . These enzymes are expressed in the hippocampus (15,28), and I-AC knockout mice show residual Ca 2ϩ -stimulated adenylyl cyclase activity consistent with the properties of VIII-AC (7,60). Surprisingly, VIII-AC was not inhibited by activation of somatostatin receptors or high levels of exogenously expressed dopamine D 2 L receptors. In addition, VIII-AC was not stimulated by G s -coupled receptors in vivo in the presence or absence of elevated intracellular Ca 2ϩ . A com- parison of the regulatory properties of I-AC and VIII-AC indicates that they differ in Ca 2ϩ sensitivity, stimulation by G scoupled receptors, and inhibition by G i -coupled receptors or CaM-dependent protein kinase IV (Table I). The differences in Ca 2ϩ sensitivity (60) are particularly interesting since I-AC is stimulated by very low intracellular Ca 2ϩ (150 -200 nM), while VIII-AC is much less sensitive to Ca 2ϩ (700 -800 nM). VIII-AC is apparently regulated only by intracellular Ca 2ϩ in vivo.
The regulatory properties of I-AC suggest that it may produce transient cAMP increases in response to relatively low intracellular Ca 2ϩ signals. VIII-AC may generate more prolonged cAMP signals in response to robust Ca 2ϩ increases since its activity is not inhibited by G i -coupled receptors or CaM kinases. These distinct signaling mechanisms may be important for different forms of synaptic plasticity. For example, it has been proposed that I-AC may play a role in mossy fiber/ CA3 LTP by coupling presynaptic Ca 2ϩ elevations to cAMP increases, which enhances glutamate release (8). In fact, mutant mice lacking I-AC show greatly depressed mossy fiber/CA3 LTP, 3 but normal long-lasting LTP in the CA1 region of the hippocampus. Long-lasting LTP is dependent upon cAMP-dependent protein kinase (61,62) and transcription (63). The stimulus paradigm that generates long-lasting LTP stimulates cAMP response element-mediated transcription in the CA1 area of the hippocampus (64). Since cAMP stimulation of transcription may depend upon prolonged cAMP increases (65,66), VIII-AC may play an important role in long-lasting LTP that depends upon cAMP-stimulated transcription.
In summary, the two Ca 2ϩ -stimulated brain adenylyl cyclases, I-AC and VIII-AC, are not redundant enzyme activities. They show very distinct regulatory properties that may be important for specific forms of synaptic plasticity in the brain. The physiological function of these enzymes and the relationship between their regulatory properties and neuroplasticity may become apparent when mutant mice lacking VIII-AC become available.  11. Somatostatin inhibits Ca 2؉ -stimulated cAMP levels in primary cultures of rat cortical or hippocampal neurons. Primary neuron cultures were prepared as described under "Experimental Procedures." After 9 days in culture, neurons were labeled overnight and then treated with 1 M A23187 in the presence or absence of 1 M somatostatin. Relative cAMP accumulation was determined as described under "Experimental Procedures." The data are the means Ϯ S.D. of triplicate assays.