Regulation of a Ca2+-sensitive adenylyl cyclase in an excitable cell. Role of voltage-gated versus capacitative Ca2+ entry.

In nonexcitable cells, we had previously established that Ca(2+)-sensitive adenylyl cyclases, whether expressed endogenously or heterologously, were regulated exclusively by capacitative Ca(2+) entry (Fagan, K. A., Mahey, R. and Cooper, D. M. F. (1996) J. Biol. Chem. 271, 12438-12444; Fagan, K. A., Mons, N., and Cooper, D. M. F. (1998) J. Biol. Chem. 273, 9297-9305). Relatively little is known about how these enzymes are regulated by Ca(2+) in excitable cells, where they predominate. Furthermore, no effort has been made to determine whether the prominent voltage-gated Ca(2+) entry, which typifies excitable cells, overwhelms the effect of any capacitative Ca(2+) entry that may occur. In the present study, we placed the Ca(2+)-stimulable, adenylyl cyclase type VIII in an adenovirus vector to optimize its expression in the pituitary-derived GH(4)C(1) cell line. In these cells, a modest degree of capacitative Ca(2+) entry could be discerned in the face of a dramatic voltage-gated Ca(2+) entry. Nevertheless, both modes of Ca(2+) entry were equally efficacious at stimulating adenylyl cyclase. A striking release of Ca(2+) from intracellular stores, triggered either by ionophore or thyrotrophin-releasing hormone, was incapable of stimulating the adenylyl cyclase. It thus appears as though the intimate colocalization of adenylyl cyclase with capacitative Ca(2+) entry channels is an intrinsic property of these molecules, regardless of whether they are expressed in excitable or nonexcitable cells.

whether they were expressed endogenously or heterologously (2,3). In particular, Ca 2ϩ released from internal stores by any mechanism was unable to regulate adenylyl cyclase activity (4), whereas Ca 2ϩ entering via CCE modulated cAMP synthesis positively or negatively, depending on the adenylyl cyclase species expressed (2,3). These findings have now been extended to other nonexcitable cell systems (5)(6)(7). Although the endogenous Ca 2ϩ -stimulable adenylyl cyclase of cerebellar granule cells (8) and of hippocampal slices (9) is stimulated by Ca 2ϩ influx through voltage-gated calcium channels (VGCCs), it is not known whether adenylyl cyclases in excitable cells are as discriminating as those expressed in nonexcitable cells for the nature of the Ca 2ϩ rise to which they respond. In the present study, the Ca 2ϩ -stimulable ACVIII was placed in an adenovirus vector to provide efficient expression in the excitable, anterior pituitary-derived tumor line, GH 4 C 1 . GH 4 C 1 cells are spontaneously electrically active and express VGCCs that give rise to prominent intracellular rises in [Ca 2ϩ ] i upon membrane depolarization (10). They also express TRH receptors coupled to phospholipase C that elevate [Ca 2ϩ ] i both by inositol 1,4,5-trisphosphate-linked mechanisms and by modifying the activity of VGCCs (11). It seemed possible that GH 4 C 1 cells, like most neuronal cells, might not display prominent CCE; however, if CCE was detectable, the opportunity would be provided to determine whether any selectivity was displayed in the regulation of the adenylyl cyclase for either type of [Ca 2ϩ ] i rise in the same cell type.
Cells and Viruses-Viruses were constructed and propagated using HEK 293 cells, a human embryonic kidney cell line transformed by and expressing high levels of Ad5 E1A and E1B proteins (12). The virus used for recombination was Ad5dl327 Bst ␤-gal, which encodes LacZ in place of the E1A and E1B genes and permits color screening for recombinant viruses (13). Ad5dl327 Bst ␤-gal was purified after infection of HEK 293 cells at maximal cytopathic effect, releasing the virus from the concentrated cell pellet by three cycles of rapid freezing and thawing; the cell debris was pelleted and re-extracted twice by resuspension in a small volume of phosphate-buffered saline followed by rapid freezing and thawing and pelleting of the cell debris. The supernatants were combined and banded for 50 min at 36,000 rpm using a CsCl step gradient consisting of 1 ml of 1.4 g/ml CsCl in phosphate-buffered saline and 1.5 ml of 1.25 g/ml CsCl in phosphate-buffered saline in a SW40 rotor. The virion band was collected by side puncture, diluted in 1.35 g/ml CsCl in phosphate-buffered saline, and rebanded for 3 h at 65,000 rpm in a VTi65 rotor. The virion band was again collected by side puncture. To prepare Ad5dl327 Bst ␤-gal-terminal protein (TP) complex, the sample was diluted in 4 M guanidine-HCl and 2.8 M CsCl (14) and banded overnight at 65,000 rpm using a VTi65 rotor. The gradient was fractionated by dripping and fractions containing DNA were identified by agarose gel electrophoresis of aliquots. DNA-containing fractions were combined and dialyzed versus 6 changes of 2 liters each of 10 mM Tris-HCl, pH 8.0, 10 mM NaCl, and 0.2 mM EDTA at 4°C. The presence of TP on the DNA greatly increases the infectivity after transfection (15), presumably due to both increased nuclear uptake facilitated by the nuclear localization signal in TP and protection of the DNA from degradation. Ad5dl327 Bst ␤-gal-TP was digested with BstBI. For analysis of the completion of digestion, an aliquot was digested with proteinase K before agarose gel electrophoresis. BstBI-digested Ad5dl327 Bst ␤-gal-TP was aliquoted and stored at Ϫ20°C.
Plasmid Construction-A cDNA encoding ACVIII was cloned into pACCMV under the control of the cytomegalovirus major immediate early promoter (CMV promoter; Ref. 16) to yield pACCMV-ACVIII, which places expression under the control of the CMV promoter, and into the plasmid pXC15E1A#12 to yield pXC15E1A-ACVIII, which places expression under the control of the adenovirus type 5 E1A promoter. pXC15E1A-ACVIII was further modified by insertion of a BstBI adaptor in the SalI site between the 3Ј end of the ACVIII coding sequence and the sequences encoding the intron and poly(A) sites (which are contributed by the 3Ј end of the E1B gene) to generate pXC15E1A-ACVIII Bst .
Transfections-HEK 293 cells at approximately 70% confluence were transfected using a modified calcium phosphate procedure (17) using 6 g of plasmid DNA and approximately 0.2 g of Ad5dl327 Bst ␤-gal-TP complex. Cells were incubated with the transfection solution overnight before being fed with fresh medium.
Construction of Adenovirus Transducing Vector Encoding ACVIII-Because cAMP is a prominent regulator of cell growth and numerous cellular processes, it would be expected that adenylyl cyclase expression would be tightly regulated and constrained to low levels (18). Thus, whereas attempts were made to construct transducing viruses placing ACVIII expression under the control of both the strong CMV promoter and weaker E1A promoters, it was expected that the use of the E1A promoter would yield a virus that directed expression at a level closer to the normal level. The use of the E1A promoter was also expected to make construction and growth of the virus easier because high-level expression of the cyclase might be expected to alter expression of a variety of genes and interfere specifically with regulation of adenovirus gene expression (e.g., Refs. 19 -22). In addition to being a generally weaker promoter than the CMV promoter, the activity of the E1A promoter is inhibited by the high level of E1A protein expressed in HEK 293 cells (23 and data not shown). Attempts to introduce pXC15E1A-ACVIII into BstBI-digested Ad5dl327 Bst ␤-gal-TP by standard overlap recombination (13,24) were unsuccessful. As an alternative, ligation of the modified plasmid, pXC15E1A-ACVIII Bst , with the large, right arm of the adenovirus chromosome-TP complex was used. 6 g of pXC15E1A#12-ACVIII Bst was digested with BstBI to generate a ligation site for the viral arm and with XmnI, which cleaves in the ␤-lactamase coding sequence within the plasmid vector backbone to leave a blunt end, to inhibit recircularization of the plasmid as well as ligation to form concatamers. The restriction enzyme-digested plasmid was ligated with Ad5dl327 Bst ␤-gal-TP complex that had been digested with BstBI. The ligation mix was used to transfect HEK 293 cells to generate the virus. It was expected that expression of ACVIII during the transfection would be significantly reduced because of the reduction in the number of ACVIII templates that contained exon and poly(A) sequences. Furthermore, direct ligation with the large arm of Ad5dl327 Bst ␤-gal-TP complex was expected to efficiently introduce the E1A-ACVIII cassette into the virus. Ligation to the large arm of adenovirus restores the intron and a poly(A) site provided by the 3Ј end of the E1B gene, thus ACVIII expression should be directed by the recombinant virus. Plaques were purified from the transfection stock after serial dilution in HEK 293 cells and overlaying with Noble agar-containing medium and serum. Plates were stained with neutral red and 5-bromo-4-chloro-3-indolyl-␤-D-pyranogalactoside 7 days after infection, and clear plaques were picked on day 8. Plaques were grown in HEK 293 cells and tested for the presence of the ACVIII gene by polymerase chain reaction. A positive clone was grown in large stock, purified by banding consecutively on CsCl step and isopycnic gradients as indicated above, and dialyzed versus three changes of 1 liter each of 135 mM NaCl, 10 mM Tris-HCl, pH. 8.0, 1 mM MgCl 2 , and 50% (v/v) glycerol at 4°C. The virus particle concentration was determined by reading absorbance at 260 nm, with 1 A 260 unit considered equivalent to 10 12 particles. The particle:plaque-forming unit ratio was approximately 100:1 for all preparations.
Cell Culture and Infection-Rat anterior pituitary GH 4 C 1 cells were maintained in 13 ml of Ham's F-10 medium (Life Technologies, Inc.) with 15% (v/v) horse serum and 2.5% fetal bovine serum (Gemini) in 75-cm 2 flasks at 37°C in a humidified atmosphere of 95% air and 5% Measurement of cAMP Accumulation-cAMP accumulation in intact cells was measured according to the method of Evans et al. (25) as described previously (26), with some modifications. GH 4 C 1 cells on 100-mm culture dishes were incubated in Ham's F-10 medium (90 min at 37°C) containing [2-3 H]adenine (20.0 Ci/dish) to label the ATP pool. The cells were washed once and detached using phosphate-buffered saline containing EDTA (0.03%). The cells were then resuspended in a nominally Ca 2ϩ -free Krebs buffer containing 120 mM NaCl, 4.75 mM KCl, 1.44 mM MgSO 4 , 11 mM glucose, 25 mM HEPES, and 0.1% bovine serum albumin (fraction V) adjusted to pH 7.4 with 2 M Tris base. The resuspended cells were aliquoted (approximately 3 ϫ 10 5 cells/tube) and used for cAMP determination in triplicate assays. Experiments were carried out at 30°C in the presence of the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (IBMX, 100 M), which was preincubated with the cells for 10 min before a 1-min assay. Unless indicated otherwise, forskolin (10 M) was included in each assay to increase the cAMP signal. Assays were terminated by the addition of 10% (w/v, final concentration) trichloroacetic acid. The [ 3 H]ATP and [ 3 H]cAMP content of the supernatants was quantified according to the standard Dowex/ alumina methodology (27) as described previously (26). Accumulation of cAMP is expressed as the percentage of conversion of [ 3 H]ATP into [ 3 H]cAMP; means Ϯ S.D. of triplicate determinations are indicated.
[Ca 2ϩ ] i Measurements-GH 4 C 1 cells (10 ϫ 10 6 /ml) were loaded with Fura-2/AM (2 M) plus 0.02% pluronic acid in Ham's F-10 media (serum-free media containing 20 mM HEPES and 0.1% bovine serum albumin, pH 7.4) for 45 min at room temperature. The cells were then washed twice with the same media and resuspended at a concentration of 10 ϫ 10 6 cells/ml. Aliquots (400 l; 4 ϫ 10 6 cells) were pelleted, resuspended in nominally Ca 2ϩ -free Krebs buffer, and used for [Ca 2ϩ ] i measurements in a Perkin-Elmer LS50B spectrofluorometer. No differences were seen in [Ca 2ϩ ] i -traces from infected and uninfected cells. Cell pretreatment with EGTA (0.1 mM) is referred to as the Ca 2ϩ -free condition. The 340/380 nm emission ratios were converted to [Ca 2ϩ ] using the standard formula (28).

RESULTS
The functional expression of ACVIII was determined by biochemical assay. Because adenylyl cyclase type II, a Ca 2ϩ -insensitive isoform, is the predominant mRNA in GH 4 C 1 cells (29), the expression of ACVIII activity could be demonstrated conclusively by an increase in cAMP accumulation in response to an elevation in [Ca 2ϩ ] i . The most robust means of elevating [Ca 2ϩ ] i in GH 4 C 1 cells is via VGCCs (30). Consequently, VGCCs were activated by membrane depolarization by increasing [K ϩ ] o in the presence or absence of extracellular Ca 2ϩ ([Ca 2ϩ ] o ). When assayed with forskolin, control cells showed no response to VGCC-mediated Ca 2ϩ entry, whereas ACVIII-infected cells exhibited a robust stimulation in cAMP accumulation upon Ca 2ϩ entry, yielding an approximately 3.5-fold stimulation compared with the Ca 2ϩ -free condition (Fig. 1A). Vasoactive intestinal peptide stimulation of the cyclase via ␣ s activation could also be augmented approximately 2.5-fold by VGCC-mediated Ca 2ϩ entry (Fig. 1B). Therefore, the use of an adenovirus construct to express ACVIII provided a simple and efficient means of heterologously expressing this protein in this excitable cell type.
Little is known about the regulation of adenylyl cyclases by VGCC-mediated Ca 2ϩ entry in excitable cells. In nonexcitable cells, we have previously established that Ca 2ϩ stimulation of heterologously expressed ACVIII occurs exclusively via CCE. Therefore, we wanted to characterize more fully the ability of VGCCs to regulate ACVIII expressed heterologously in an excitable cell. The activity of VGCCs relies on the membrane potential, which is dictated, in part, by the K ϩ concentration in the medium bathing the cells. The amount of Ca 2ϩ entry after membrane depolarization with varying [K ϩ ] o was assessed ( Fig. 2A). Populations of Fura-2-loaded GH 4 Fig. 2A).
The above-mentioned findings showed that a Ca 2ϩ -sensitive adenylyl cyclase could be regulated by Ca 2ϩ entry through VGCCs in excitable cells. Furthermore, the magnitude of the stimulation mirrored the extent of Ca 2ϩ entry. We have seen similar results with CCE in nonexcitable cells. In excitable cells, the role of CCE has been explored only sparingly in the face of the much more pronounced [Ca 2ϩ ] i rise generated by VGCCs. However, we wondered whether CCE was present in excitable cells, if it could also regulate ACVIII, and how this might compare with the regulation by VGCC-mediated Ca 2ϩ entry. CCE is activated by depletion of intracellular Ca 2ϩ stores using the sarcoplasmic/endoplasmic Ca 2ϩ -ATPase inhibitor thapsigargin (TG) (31). Treatment of the cells with TG resulted in a modest [Ca 2ϩ ] i rise (approximately 130 nM) that returned toward baseline because the cells were in Ca 2ϩ -free media (Fig. 3A). Addition of [Ca 2ϩ ] o , either 0.5 or 2 mM, resulted in a rapid [Ca 2ϩ ] i rise that reached a peak of approximately 230 or 380 nM, respectively, within the time course of the cAMP measurements. Although the [Ca 2ϩ ] i rise due to CCE was rather robust, it was considerably smaller than the [Ca 2ϩ ] i rise generated by VGCC-mediated Ca 2ϩ entry (Fig. 3B). Depolarization of the cells with KCl (10 mM) in the presence of either 0.5 or 2 mM [Ca 2ϩ ] o resulted in peak [Ca 2ϩ ] i rises of approximately 400 and 550 nM, respectively. Therefore, triggering of CCE by intracellular Ca 2ϩ store depletion resulted in a modest [Ca 2ϩ ] i rise compared with that arising from VGCC-mediated Ca 2ϩ entry. Next, the ability of CCE and VGCC-mediated Ca 2ϩ entry to stimulate ACVIII was compared. Cells maintained in Ca 2ϩ -free Krebs buffer were treated with TG 4 min before the addition of varying [Ca 2ϩ ] o (Fig. 3C). The stimulation of ACVIII   Fig. 3, A and B).
The clear demonstration of CCE in GH 4 C 1 cells can be difficult because individual GH 4 C 1 cells demonstrate spontaneous Ca 2ϩ oscillations due to intermittent activation of VGCCs, even in "resting" conditions (10). Therefore, in a population of GH 4 C 1 cells, considerable VGCC activity may underlie the magnitude of the Ca 2ϩ entry ascribed to CCE. The extent of VGCC-mediated Ca 2ϩ entry in resting conditions was explored by comparing Ca 2ϩ entry in untreated and TG-treated cells. Untreated cells reveal Ca 2ϩ entry due to spontaneously active VGCCs along with CCE from passive store depletion due to the cells being in Ca 2ϩ -free buffer. The L-type VGCC blocker, nimodipine, can then be used to verify the contribution of the predominant L-type VGCC occurring in GH 4 C 1 cells. Populations of Fura-2 loaded GH 4 C 1 cells were either untreated or treated with TG 4 min before the addition of [Ca 2ϩ ] o (2 mM). Prior treatment of the cells with TG may have augmented the [Ca 2ϩ ] i rise, giving a peak [Ca 2ϩ ] i rise of approximately 250 nM (Fig.  4A, trace a), compared with a peak [Ca 2ϩ ] i rise of approximately 220 nM (Fig. 4A, trace b) in the untreated cells (to illustrate the two traces more clearly, a running average (5-s intervals) of each trace has been overlaid on the actual trace). The magnitude of the [Ca 2ϩ ] i rise generated by CCE was clearly discerned in the presence of the VGCC blocker nimodipine. Addition of nimodipine (1 M) along with [Ca 2ϩ ] o resulted in a greatly reduced [Ca 2ϩ ] i rise in the untreated cells (approximately 125 nM as compared with 220 nM without nimodipine (Fig. 4A, cf. b versus d)). The effect of nimodipine on TG-treated cells was much less drastic, decreasing the peak [Ca 2ϩ ] i rise from 250 nM without nimodipine to 180 nM in the presence of the VGCC blocker (Fig. 4A, cf. a versus c). The ability of nimodipine to greatly attenuate the [Ca 2ϩ ] i rise generated by Ca 2ϩ addition alone illustrated the prominent, spontaneous VGCC-mediated Ca 2ϩ entry occurring in these cells. Accordingly, stimulation of ACVIII by the above-mentioned conditions also revealed the presence of spontaneous VGCC activity. Addition of [Ca 2ϩ ] o to GH 4 C 1 cells maintained in Ca 2ϩ -free Krebs buffer resulted in an approximately 1.5-fold stimulation of ACVIII as compared with the control, Ca 2ϩ -free condition (Fig. 4A, inset, open bar versus right-hatched bar). Pretreatment of the cells with TG increased the amount of ACVIII stimulation, resulting in a 2.4-fold increase in activity (Fig. 4, inset, cross-hatched bar). In the presence of the VGCC blocker nimodipine, the ability of [Ca 2ϩ ] o addition alone to stimulate the cyclase was completely abolished (Fig. 4A, inset, left-hatched bar), whereas cells pretreated with TG still showed a robust regulation of ACVIII (approximately 1.9-fold stimulation; Fig. 4A, inset, horizontal striped bar). Thus, basal VGCC activity in resting cells did contribute significantly to the stimulation of ACVIII. Conversely, there was no indication of CCE in resting, untreated cells, because the Ca 2ϩ entry generated by Ca 2ϩ addition was completely blocked by nimodipine.
Another method to verify the presence of CCE is to determine whether the combination of the two Ca 2ϩ entry mechanisms (CCE and VGCC) yields augmented [Ca 2ϩ ] i rises and/or regulation of ACVIII. Populations of GH 4 C 1 cells were treated with TG (Fig. 4B, traces a and c) regulation of ACVIII (Fig. 4B, inset). GH 4 C 1 cells that were pretreated with TG before [Ca 2ϩ ] o addition showed a stimulation of ACVIII by approximately 2.4-fold as compared with control, untreated cells (Fig. 4B, inset,

open bar versus righthatched bar). Depolarization of the cells by the addition of [K ϩ ] o along with [Ca 2ϩ
] o also resulted in a stimulation of the cyclase, giving an approximately 2.2-fold increase (Fig. 4B,  inset, cross-hatched bar). Combination of the two modes of Ca 2ϩ entry resulted in an augmented regulation of ACVIII, yielding an approximately 2.8-fold stimulation (Fig. 4B, inset, lefthatched bar). In the same experiment as Fig. 4B, the ability of VGCC-mediated Ca 2ϩ entry activated by KCl addition to regulate the cyclase was also blocked by nimodipine (Fig. 4C). Furthermore, treatment of the cells with both TG and KCl in the presence of nimodipine resulted in a stimulation of ACVIII that was identical to that seen with TG treatment alone (Fig.  4C, cross-hatched versus vertical-striped bars). Therefore, the two Ca 2ϩ entry mechanisms, CCE and VGCC-mediated Ca 2ϩ entry, in combination, led to an augmentation in both the [Ca 2ϩ ] i rise and the magnitude of the stimulation of ACVIII produced by either Ca 2ϩ entry mechanism alone, indicating they are separate Ca 2ϩ entry processes.
The previous results revealed the presence of CCE with the use of nimodipine to block the spontaneously active VGCCs. It was also shown that the two Ca 2ϩ entry mechanisms are quite similar in their ability to regulate ACVIII when the underlying VGCC-mediated Ca 2ϩ entry is removed from the CCE (cf. Fig.  4, A versus B, cross-hatched bars). Although the effects of VGCC-mediated Ca 2ϩ entry and "pure" CCE on the regulation of the cyclase are similar, the corresponding [Ca 2ϩ ] i rises are quite different (Fig. 5, A and B). A robust [Ca 2ϩ ] i rise was generated by VGCC-mediated Ca 2ϩ entry triggered by [K ϩ ] o (Fig. 5A, 10 mM; trace a), reaching a peak of approximately 700 nM, whereas TG-mediated Ca 2ϩ entry reached a peak of approximately 250 nM (Fig. 5B, trace a). In the presence of nimodipine (1 M), the [Ca 2ϩ ] i rise generated by depolarization was almost eliminated (Fig. 5A, trace b), dropping to approximately 120 nM, whereas TG-mediated Ca 2ϩ entry was decreased to approximately 170 nM (Fig. 5B, trace b). This relatively small [Ca 2ϩ ] i rise generated by TG treatment in the presence of nimodipine, "pure" CCE, could still effectively stimulate AC-VIII (cf. Fig. 4C). Therefore, a [Ca 2ϩ ] i rise of approximately 170 nM generated by CCE was as efficacious as the approximately 700 nM [Ca 2ϩ ] i rise produced by VGCC-mediated Ca 2ϩ entry in regulating ACVIII.
The observation that a modest amount of Ca 2ϩ entry via CCE was able to stimulate ACVIII to a similar extent as a much more robust VGCC-mediated Ca 2ϩ entry prompted us to investigate whether Ca 2ϩ release from intracellular stores could also regulate the exogenously expressed ACVIII. Although TG releases intracellular Ca 2ϩ , the small release occurs over a prolonged period. Phospholipase C-coupled agonists produce a much more rapid and significant [Ca 2ϩ ] i rise. Addition of TRH (100 nM) to populations of GH 4 C 1 cells maintained in Ca 2ϩ -free Krebs buffer resulted in a very rapid, large [Ca 2ϩ ] i rise of approximately 500 nM (Fig. 6A). Intracellular Ca 2ϩ can also be released by Ca 2ϩ ionophores. Treatment of GH4C1 cells incubated in Ca 2ϩ -free Krebs buffer with ionomycin (IM; 2 M) yielded a rapid and robust [Ca 2ϩ ] i rise that reached approximately 900 nM (Fig. 6B). The ability of these large [Ca 2ϩ ] i rises generated by releasing intracellular Ca 2ϩ with either TRH or IM to stimulate ACVIII was explored. cAMP accumulation was measured in GH 4 C 1 cells that were either uninfected (open bars) or infected with the ACVIII-containing virus (hatched bars) and maintained in Ca 2ϩ -free Krebs buffer before the addition of forskolin along with TRH (100 nM; Fig. 6A, inset) or IM (2 M ; Fig. 6B, inset). Neither TRH or IM stimulated ACVIII activity as compared with the uninfected control cells. Therefore, the triggering of a large [Ca 2ϩ ] i rise due to the release of Ca 2ϩ from intracellular stores via either inositol 1,4,5-trisphosphase generation or ionophore-mediated release did not affect ACVIII. DISCUSSION One of the more surprising findings to emerge from the study of Ca 2ϩ -sensitive adenylyl cyclases-whether they are expressed heterologously or endogenously-is their strict dependence on CCE for their regulation in nonexcitable cells (2,3). Release from internal stores (4) or Ca 2ϩ entry via ionophore (2,3) or triggered by arachidonic acid (7) is without effect. Extremely little is known about the regulation of Ca 2ϩ -sensitive adenylyl cyclases in excitable cells. In the present study, we have investigated the regulation of ACVIII by Ca 2ϩ in an excitable cell line, rat anterior pituitary tumor-derived GH 4 C 1 cells. The study addressed four major issues: (a) whether Ca 2ϩ entry via VGCCs could regulate a Ca 2ϩ -stimulable adenylyl cyclase, (b) whether the modest CCE expressed in these cells could regulate the adenylyl cyclase, (c) whether Ca 2ϩ release from internal stores could regulate adenylyl cyclase in this excitable cell line, and (d) whether the relative magnitudes of the [Ca 2ϩ ] i rises generated by CCE and VGCCs could predict the degree of stimulation of the adenylyl cyclase.
Does Ca 2ϩ entry via VGCCs regulate a Ca 2ϩ -stimulable adenylyl cyclase? Ca 2ϩ -sensitive adenylyl cyclases are expressed mainly in excitable tissues, with the Ca 2ϩ -stimulable ACI and ACVIII found exclusively in neuronal cells (32)(33)(34), whereas ACV and ACVI, the Ca 2ϩ -inhibitable isoforms, predominate in cardiac tissue (35)(36)(37). Given the expression pattern of Ca 2ϩsensitive adenylyl cyclases and the prevalence of VGCCs in those tissues, we were surprised at the paucity in the literature on the ability of VGCC-mediated Ca 2ϩ entry to regulate cAMP accumulation. Although there is abundant literature on the ability of the cAMP-signaling cascade to regulate L-type VGCCs (reviewed in Ref. 38), little has been done to directly address the effect and the potential feedback of Ca 2ϩ entry by such channels on adenylyl cyclase activity. An exception was a study showing that Ca 2ϩ entry through VGCCs inhibited adenylyl cyclase activity in embryonic chick ventricle myocytes (39). In the case of Ca 2ϩ /calmodulin stimulation of adenylyl cyclase activity, depolarization of hippocampal CA1 slices by 50 mM KCl increased cAMP accumulation, which was blocked by calmodulin antagonists (9). Furthermore, in cultured cerebellar granular cells, depolarization-induced Ca 2ϩ entry stimulated cAMP accumulation, which was blocked by nimodipine (8). The current study has extended these findings in showing that heterologous expression of a specific Ca 2ϩ -stimulable adenylyl cyclase, ACVIII, is regulated by Ca 2ϩ entry through VGCCs in a clonal cell line. The potent block of this effect by nimodipine establishes that the effect is mediated by L-type VGCCs (40).
Does CCE regulate a Ca 2ϩ -stimulable adenylyl cyclase in an excitable cell? The potential role of CCE in excitable cells has been overshadowed by the more robust Ca 2ϩ entry generated by VGCCs. Also hindering the study of CCE not only in excitable cells, but in all cell types, is the uncertainty of the molecular nature of the channels responsible for CCE. The Drosophila transient receptor potential protein, which is involved in insect phototransduction, is increasingly being viewed as a putative CCE channel (41). A family of mammalian transient receptor potential protein homologues have now been described, each possessing different activation and conductance characteristics when heterologously expressed (reviewed in Refs. 42 and 43). Additionally, several of the mammalian transient receptor potential protein isoforms have been found in various brain regions, further suggesting a potential role for CCE in excitable cells (44,45). One such role for CCE in excitable cells emerged from the work of Koizumi and Inoue (46), who showed in PC12 cells that caffeine, TG, and cyclopiazoic acid, all agents that release intracellular Ca 2ϩ and therefore stimulate CCE, evoked dopamine release. Exocytosis has also been shown to be regulated by CCE in adrenal chromaffin cells (47). Our findings show that CCE, which gives rise to a modest increase in [Ca 2ϩ ] i , was very effective at stimulating ACVIII. Furthermore, the findings that CCE augmented the stimulation of ACVIII by VGCC-mediated Ca 2ϩ entry in a manner that was insensitive to nimodipine establishes the fact that CCE can regulate the cyclase.
Does Ca 2ϩ release from intracellular Ca 2ϩ stores regulate ACVIII in an excitable cell? We have previously shown an inability of Ca 2ϩ release to regulate Ca 2ϩ -sensitive adenylyl cyclases in nonexcitable cells (4), and furthermore, we have shown that Ca 2ϩ regulation of adenylyl cyclases relies totally on Ca 2ϩ entry through CCE channels (3). In the current study, releasing intracellular Ca 2ϩ with either phospholipase C-coupled agonists or ionophore was without effect on ACVIII activity. Similar to our findings, dopamine release in PC12 cells was insensitive to Ca 2ϩ release from intracellular stores and in fact depended on Ca 2ϩ entry (46,48). Thus we are now inclined to generalize that Ca 2ϩ release from intracellular stores will not regulate Ca 2ϩ -sensitive adenylyl cyclases, regardless of cell type.
Does the magnitude of the [Ca 2ϩ ] i rise generated by various means predict the amount of stimulation of ACVIII? We have compared the ability of three modes of raising [Ca 2ϩ ] i , VGCCmediated Ca 2ϩ entry, CCE, and Ca 2ϩ release from internal stores, to stimulate ACVIII. The [Ca 2ϩ ] i values achieved by VGCC-mediated Ca 2ϩ entry and ionophore-mediated Ca 2ϩ release were both substantial (peak values of approximately 700 and 900 nM, respectively), with CCE being much more modest (approximately 250 nmM). However, the magnitude of these [Ca 2ϩ ] i rises in no way predicts the amount of stimulation of ACVIII. The large [Ca 2ϩ ] i rise promoted by Ca 2ϩ release was totally without effect, whereas a similar peak [Ca 2ϩ ] i rise generated by VGCC-mediated Ca 2ϩ entry was very potent at stimulating ACVIII. CCE, the least effective in terms of producing a large [Ca 2ϩ ] i rise, was as efficacious as VGCC-mediated Ca 2ϩ entry in stimulating ACVIII.
Obviously the inability of the magnitude of different forms of [Ca 2ϩ ] i rise to predict subsequent effects on ACVIII is due to the unresolved spatial information provided by Fura-2 in population measurements of [Ca 2ϩ ] i . One approach to addressing this issue directly would be to measure the [Ca 2ϩ ] in the vicinity of the cyclase by an adenylyl cyclase/aequorin chimera (49), with the prediction that the chimera would report similar [Ca 2ϩ ] in response to VGCC and CCE and far less in response to release from stores. It is becoming increasingly obvious that the plasma membrane is not uniform in lipid composition or in the distribution of regulatory elements. Recent findings indicate that adenylyl cyclases must occur in cholesterol-rich domains to be susceptible to CCE in nonexcitable cells (50). In this context, it is also relevant that transient receptor potential protein 1 has recently been reported in rafts (51). It would be very interesting to determine whether the same residence in cholesterol-rich domains would apply to the adenylyl cyclase or to any of these channels in excitable cells. The nonequivalence in the ability of VGCC-mediated Ca 2ϩ entry and CCE to regulate ACVIII in GH 4 C 1 cells may indicate that the adenylyl cyclase is closer to the CCE channel than the VGCC. It would be expected, and indeed it has been predicted, that [Ca 2ϩ ] of 1 M would be found up to approximately 50 nm from a VGCC (52), whereas an equal [Ca 2ϩ ] would be found approximately 5-10 nm from the mouth of a CCE channel (53), given the conductances of L-type and I crac channels reported in the literature. Furthermore, the moderately high affinity for Ca 2ϩ of Fura-2 makes it likely that the [Ca 2ϩ ] i achieved by VGCC in the cytosol is underestimated (28). Consequently, the VGCC may be even more distant and in quite a different domain from the adenylyl cyclase.
It seems fair to conclude from these studies that the intimate relationship that was first demonstrated between CCE channels and adenylyl cyclase in nonexcitable cells is maintained in excitable cells, even though the adenylyl cyclase is also susceptible to the very robust VGCC-mediated Ca 2ϩ entry present in these cells. Whether this association is maintained by colocalization within cholesterol-rich domains of the plasma membrane, as in nonexcitable cells (50), or by some additional process remains to be determined.