Adenovirus-mediated Expression of an Olfactory Cyclic Nucleotide-gated Channel Regulates the Endogenous Ca2+-inhibitable Adenylyl Cyclase in C6-2B Glioma Cells*

Previous studies have established that Ca2+-sensitive adenylyl cyclases, whether endogenously or heterologously expressed, are preferentially regulated by capacitative Ca2+ entry, compared with other means of elevating cytosolic Ca2+ (Chiono, M., Mahey, R., Tate, G., and Cooper, D. M. F. (1995) J. Biol. Chem.270, 1149–1155; 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). These findings led to the suggestion that adenylyl cyclases and capacitative Ca2+entry channels were localized in the same functional domain of the plasma membrane. In the present study, we have asked whether a heterologously expressed Ca2+-permeable channel could regulate the Ca2+-inhibitable adenylyl cyclase of C6-2B glioma cells. The cDNA coding for the rat olfactory cyclic nucleotide-gated channel was inserted into an adenovirus construct to achieve high levels of expression. Electrophysiological measurements confirmed the preservation of the properties of the expressed olfactory channel. Stimulation of the channel with cGMP analogs yielded a robust elevation in cytosolic Ca2+, which was associated with an inhibition of cAMP accumulation, comparable with that elicited by capacitative Ca2+ entry. These findings not only extend the means whereby Ca2+-sensitive adenylyl cyclases may be regulated, they also suggest that in tissues where they co-exist, cyclic nucleotide-gated channels and Ca2+-sensitive adenylyl cyclases may reciprocally modulate each other’s activity.

is gated by cAMP, and is thought to lead to activation of Ca 2ϩ -activated Cl Ϫ currents and membrane depolarization (3). Increasingly, however, a more widespread function for these channels in cell physiology has been envisioned, partially due to the finding that the channels are expressed in a wide range of tissues and cell types. For instance, proteins homologous to the CNG channel have been cloned from such diverse tissues as heart, kidney, and testis, as well as from liver and skeletal muscle (4 -6). CNG channels have also been found in various brain regions, namely, the hippocampus, cortex, and Purkinje cells of the cerebellum and other neural derived tissues such as pineal and pituitary gland (5,(7)(8)(9)(10). The observation that these channels are widely expressed prompts a reevaluation of their role in signal transduction. Functionally, the CNG channels belong to the family of ligand-gated channels, but, structurally, they are similar to voltage-gated channels. CNG channels also share the important feature of Ca 2ϩ permeation with voltagegated Ca 2ϩ channels. At physiological [Ca 2ϩ ], an expressed, homomeric version of the olfactory CNG channel exhibits a nearly "pure" Ca 2ϩ current (11). In comparison, only ϳ5% of the current through the NMDA channel is carried by Ca 2ϩ (12). Therefore, these channels provide a second messenger-regulated form of Ca 2ϩ entry into the cell whose primary function may be to elevate [Ca 2ϩ ] i . Adenylyl cyclases are regulated by physiological transitions in [Ca 2ϩ ] i (reviewed in Refs. 13 and 14)). In fact, of the nine currently described isoforms of adenylyl cyclase, Ca 2ϩ directly regulates four. Adenylyl cyclase types I and VIII are stimulated, while types V and VI are inhibited by submicromolar [Ca 2ϩ ]. We have previously shown that Ca 2ϩ -sensitive adenylyl cyclases are regulated by capacitative Ca 2ϩ entry (CCE) while they are refractory to [Ca 2ϩ ] i rises produced by other means, such as release from internal stores or entry mediated by ionophore in nonexcitable cells (15,16). The dependence of these adenylyl cyclases on Ca 2ϩ entering through CCE channels suggested a functional colocalization of CCE channels and Ca 2ϩ -sensitive adenylyl cyclases. Therefore, it was of interest to determine whether Ca 2ϩ entry through heterologously expressed CNG channels might regulate these enzymes. C6-2B cells, which endogenously express a Ca 2ϩ -inhibitable adenylyl cyclase (type VI) (17), were used to determine whether Ca 2ϩ entry through an olfactory CNG channel could regulate cAMP accumulation. Expression of the rat olfactory CNG channel (18) was accomplished by creating an adenovirus construct containing the channel. Infection with the adenovirus/CNG channel permits efficient expression in a large majority of the cells. The expression of the channel was evaluated by both [Ca 2ϩ ] i measurements in cell populations and electrophysiological methods. Activation of the CNG channel with the cell-permeant cGMP analog, CPT-cGMP, generated a [Ca 2ϩ ] i rise that was depend-ent on [CPT-cGMP], time of exposure to CPT-cGMP, and [Ca 2ϩ ] ex . Furthermore, activation of the channel with CPT-cGMP and its associated [Ca 2ϩ ] i rise produced a substantial inhibition of cAMP accumulation. The magnitude of the global [Ca 2ϩ ] i rise generated by the CNG channel was modest in comparison with capacitative Ca 2ϩ entry, but both modes of Ca 2ϩ entry were equally efficacious in their ability to reduce cAMP levels. Therefore, Ca 2ϩ entry through a heterologously expressed CNG channel can modulate endogenous cAMP levels. These data not only show that a Ca 2ϩ -sensitive adenylyl cyclase can be regulated by a heterologously expressed Ca 2ϩ channel, but also, that CNG channels may play a role in modulating cAMP accumulation in tissues where channels and adenylyl cyclases are co-expressed.  1)-A fragment encoding the rat olfactory CNG channel ␣-subunit cDNA (18) was ligated between the BamHI and SalI sites in the plasmid pACCMV, which encodes the left end of the adenovirus chromosome with the E1A gene and the 5Ј-half of the E1B gene replaced by the cytomegalovirus major immediate early promoter, a multiple cloning site, and intron and polyadenylation sequences from SV40 (19) to yield the plasmid pACCMV-CNGC. pACCMV-CNGC was digested with SalI and ligated with a BstBI adaptor in order to create pACCMV-CNGC Bst , such that sequences encoding CNGC and the left end of the adenovirus chromosome could be ligated directly to the right arm of the adenovirus chromosome to create a transducing vector using a newly developed protocol. 2 pACCMV-CNGC Bst was digested with BstBI (to provide an end to ligate with adenovirus DNA) and XmnI (to provide a blunt end that would inhibit recircularization of the plasmid as well as the formation of concatamers). The digested plasmid DNA was ligated with BstBI-digested Ad5dl327 Bst ␤-Gal-TP complex (20). Ad5dl327 Bst ␤-Gal has a deletion of the fragment between the XbaI sites at 28,593 and 30,471 base pairs and therefore does not encode any of the products of the E3 region. The ligated DNA was used to transfect 293 cells using Ca 3 (PO 4 ) 2 precipitation (21). The transfected cells were incubated for 7 days. A freeze-thaw lysate was prepared from the cells, and dilutions were used to infect 293 plates for plaque purification. The infected 293 plates were overlaid with medium in Noble agar, fed after 4 days, and stained with 5-bromo-4-chloro-3-indolyl ␤-D-galactopyranoside and neutral red (20). Clear plaques, which are derived from recombination that results in deletion of the lacZ gene present in the parental viral chromosome, were amplified and analyzed by polymerase chain reaction and restriction digestion for the presence of the CNGC cDNA. Plaques that proved positive by polymerase chain reaction and restriction digestion analysis were tested for the ability to direct expression of CNGC. This procedure resulted in efficient production of viruses encoding CNGC. The virus, termed Ad5dl327CMV-CNGC, was grown in large scale, purified by successive banding on step and isopycnic CsCl gradients, and dialyzed versus three changes of 10 mM Tris-HCl, pH. 7.9, 135 mM NaCl, 1 mM MgCl 2 , 50% glycerol at 4°C. Virus particle concentration was quantitated by determination of the absorbance at 260 nm. Virus was stored at Ϫ20°until use.

Materials
Measurement of cAMP Accumulation-cAMP accumulation in intact cells was measured according to the method of Evans et al. (22) as described previously (16) with some modifications. C6-2B cells on 100-mm culture dishes were incubated in F-10 medium (60 min at 37°C) containing [2-3 H]adenine (20.0 Ci/dish) to label the ATP pool. The cells were then 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 then aliquotted (approximately 3 ϫ 10 5 cells/tube) and used for cAMP determination assays in triplicate. All experiments were carried out at 30°C in the presence of phosphodiesterase inhibitors, 3-isobutyl-1-methylxanthine (500 M) and Ro 20-1724 2 J. Schaack, unpublished data. (100 M), which were preincubated with the cells for 10 min prior to a 1-min assay. Assays were terminated by the addition of 10% (w/v, final concentration) trichloroacetic acid. Unlabeled cAMP (100 l, 10 mM), ATP (10 l, 65 mM), and [␣- 32  Electrical Recording-Currents through CNG channels were measured using the whole-cell patch clamp technique and an Axopatch-200A patch clamp amplifier (Axon Instruments Inc., Foster City, CA). Pipettes were pulled from borosilicate glass and heat-polished. To ensure adequate voltage control in the whole cell configuration pipette, resistance was limited to 3.5 megaohms and averaged 2.8 Ϯ 0.1 megaohms (n ϭ 39). Voltage offsets were zeroed with the pipette in the bath solution. Pipettes were then lowered onto the cells, and gigaohms seals were formed by applying light suction (12.8 Ϯ 0.9 gigaohms). After achieving whole cell configuration, capacitive transients were elicited by applying 20-mV steps from the holding potential (0 mV), filtered at 10 kHz, and recorded at 40 kHz for calculation of access resistance and input impedance. In all experiments, the voltage error due to series resistance was less than 5 mV. Current records were filtered at 1 kHz, sampled at 5 kHz, and analyzed on an IBM-compatible computer using Pclamp6 software (Axon Instruments). The intracellular pipette filling solution contained 145 mM KCl, 4 mM NaCl, 0.5 mM MgCl 2 , 10 mM HEPES, and either 0 or 1 mM cGMP, and pH was adjusted to 7.4 with KOH. The bath solution contained 145 mM NaCl, 4 mM KCl, 10 mM HEPES, 11 mM glucose, and either 10 mM MgCl 2 or 1 mM EGTA, and pH was adjusted to 7.4 with NaOH.
Statistics-Analyses were performed using the PRISM statistical software package (version 2.00, GraphPad Software, Inc.) Electrophysiological Determination of CNG Channel Expression-The effectiveness of using an adenovirus construct to heterologously express the olfactory CNG channel was assessed by monitoring currents in the whole-cell patch clamp configuration. Currents were elicited by 250-ms steps from the holding potential, 0 mV, to membrane potentials between Ϫ80 and ϩ60 mV in 10 mV increments, followed by a 100-ms step to Ϫ40 mV. To determine if CNG channels were present in the infected cells, the pipette solution contained either 0 or 1 mM cGMP, which is a saturating cGMP concentration (18). The bath solution initially contained 10 mM Mg 2ϩ , which blocks Ͼ95% of inward current and Ͼ80% of outward current through the olfactory CNG channel (11). Thus, if CNG channels are present and cGMP is in the patch pipette, only a small outwardly rectifying current should be observed in the presence of 10 mM external Mg 2ϩ . Removal of Mg 2ϩ from the bath solution would be expected to reveal a substantially larger, nonrectifying current.

Effect of Varying CNG Channel Construct Multiplicity of
When recording from cells infected with the adenovirus en- coding the CNG channel, only small leak currents (Ͻ͉6 pA͉ at Ϯ 40 mV) were observed in 10 mM external Mg 2ϩ when the patch pipette did not contain cGMP (Fig. 3A). Removal of Mg 2ϩ from the bath solution caused a reversible, 2-3-fold increase in leak (Fig. 3, B and C). The addition of the membrane-permeant cGMP analogue, CPT-cGMP (100 M), to the bath solution induced a large current that was subsequently blocked by 10 mM external Mg 2ϩ (data not shown). However, when the patch pipette contained 1 mM cGMP, a small outward current was observed in the presence of 10 mM external Mg 2ϩ (Fig. 3D). Removal of external Mg 2ϩ revealed a substantially larger, nonrectifying current that could be blocked by 10 mM Mg 2ϩ (Fig. 3, E and F). The collected results from 39 cells are shown in Fig.  4. Uninfected cells in the presence (n ϭ 9) or absence (n ϭ 10) of cGMP and infected cells in the absence of internal cGMP (n ϭ 9) displayed small inward leak currents, Ͻ15 pA at Ϫ40 mV, in 0 mM external Mg 2ϩ . Infected cells that gave a measurable response in the presence of internal cGMP (n ϭ 11) displayed large inward currents, Ͼ700 pA at Ϫ40 mV in 0 mM external Mg 2ϩ . The cGMP-dependent current was observed in Ͼ70% (13 of 18, including 2 of 3 cells exposed to CPT-cGMP) of infected cells and no (0 of 9) uninfected cells.
Ability of Ca 2ϩ Entry through the CNG Channel to Inhibit ACVI-The next experiments aimed to determine whether the Ca 2ϩ entry through the expressed CNG channel could regulate a Ca 2ϩ -inhibitable adenylyl cyclase that is endogenously expressed in C6-2B cells. The effect of CPT-cGMP pretreatment followed by the addition of varying [Ca 2ϩ ] ex on cAMP accumulation in uninfected versus infected cells was examined (Fig. 5). cAMP accumulation was measured over a 1-min period following the addition of [Ca 2ϩ ] ex along with forskolin and isoproterenol to stimulate adenylyl cyclase activity (see "Experimental Procedures"). All cells were pretreated with the phosphodiesterase inhibitors 3-isobutyl-1-methylxanthine (500 M) and Ro 20-1724 (100 M) 10 min prior to the 1-min assay. In cells expressing the CNG channel, pretreatment with CPT-cGMP (300 M) caused steadily increasing inhibition in cAMP accumulation as a function of the [Ca 2ϩ ] ex . [Ca 2ϩ ] ex of 1, 2, and 4 mM inhibited cAMP accumulation by 20, 25, and 34%, respectively. This was in contrast to uninfected cells, also pretreated with CPT-cGMP (300 M), which gave a maximal inhibition of 20% at a [Ca 2ϩ ] ex of 4 mM. The modest degree of inhibition of cAMP accumulation seen with increasing [Ca 2ϩ ] ex in the cells not infected with the CNG channel construct was the result of limited capacitative Ca 2ϩ entry (see Fig. 2). The above data support the idea that Ca 2ϩ entry through the CNG channel can be sensed by the Ca 2ϩ -sensitive adenylyl cyclase. To further understand the functional relationship between the CNG channel and the Ca 2ϩ -sensitive adenylyl cyclase, detailed manipulations of the CPT-cGMP concentration and exposure time, as well as the [Ca 2ϩ ] ex , were carried out.
Effect of Varying [CPT-cGMP] and Exposure Time on Ca 2ϩ Entry and Inhibition of ACVI-The ability of the cGMP analog, CPT-cGMP, to activate the CNG channel partly depends on its ability to cross the plasma membrane and reach an effective concentration at the CNG channel. Permeation of CPT-cGMP was examined by varying the amount of time the cells were exposed to the cGMP analog prior to the addition of [Ca 2ϩ ] ex . Fura-2-loaded C6-2B cells were pretreated with varying amounts of CPT-cGMP for either 2 or 5 min prior to the addition of [Ca 2ϩ ] ex . The 2-min exposure to CPT-cGMP (Fig. 6A) In the absence of CPT-cGMP pretreatment, a very modest [Ca 2ϩ ] i rise was observed, due to the fact that the cells were being maintained in a nominally Ca 2ϩ -free medium. When the time of exposure to CPT-cGMP was increased to 5 min, the rate of the [Ca 2ϩ ] i rise following the addition of [Ca 2ϩ ] ex was considerably faster, reaching a plateau within the course of the experiment (Fig. 6B). Although the rates of the [Ca 2ϩ ] i rise increased with longer exposure to CPT-cGMP, the peak values reached were very similar to the 2-min exposure (see Fig. 6A) with 20, 50, and 100 M CPT-cGMP treatments reaching peaks of approximately 200, 320, and 360 nM, respectively. These results showed that permeation of the cGMP analog across the plasma membrane is rather slow, but once maximal activation of the channel has been reached, the peak [Ca 2ϩ ] i rises are very similar for a given [CPT-cGMP].
In order to consolidate the regulatory consequence of Ca 2ϩ entry through the CNG channel on the adenylyl cyclase, different CPT-cGMP exposure times were compared in terms of their effect on cAMP accumulation and Ca 2ϩ entry. Fig. 7 shows the effect of varying both the CPT-cGMP concentration and the exposure time to the cGMP analog on the cAMP accumulation in C6-2B cells infected with the CNG channel construct. Following a 2-min exposure to CPT-cGMP, a combination of [Ca 2ϩ ] ex as well as forskolin and isoproterenol were added to the cells with cAMP accumulation measured over the subsequent minute. With increasing [CPT-cGMP], there was a stepwise increase in the inhibition of cAMP accumulation ranging from 7% inhibition with 20 M CPT-cGMP to 32% with 300 M CPT-cGMP (Fig. 7A). It should be noted that the extent of inhibition observed with increasing [CPT-cGMP] agrees well with the extent of Ca 2ϩ entry (Fig. 6A). Without CPT-cGMP pretreatment of the cells, a minimal inhibition of cAMP accumulation (8%) was observed, which is very similar to the 20 M CPT-cGMP condition. As seen in Fig. 6A, 0 and 20 M CPT-cGMP produce a similar [Ca 2ϩ ] i rise within the first minute, the period over which cAMP accumulation is measured. Therefore, the similarities in the extent of the inhibition seen with 0 and 20 M CPT-cGMP are consistent with the Ca 2ϩ data. In Fig. 7B, the effects of a 5-min exposure to varying CPT-cGMP concentrations on cAMP accumulation are shown. Again, increasing the CPT-cGMP concentration produced further inhibition of cAMP accumulation following the addition of [Ca 2ϩ ] ex , with maximal inhibition (31%) observed at 300 M CPT-cGMP. The amount of inhibition observed in the absence of CPT-cGMP was again 8%, which, following a 5-min exposure to CPT-cGMP, differs greatly from the 20 M CPT-cGMP condition (22%). This result was also in good agreement with the corresponding Ca 2ϩ data (Fig. 6B), where the longer pretreatment with CPT-cGMP resulted in a faster [Ca 2ϩ ] i rise and, therefore, a higher [Ca 2ϩ ] i level achieved within the 1-min assay period. It is also noteworthy that the extent of inhibition in cAMP accumulation appears to reach "maximal" levels at lower CPT-cGMP concentrations with these longer exposure times. In other words, the dose-response curve has been shifted to the left, indicating an increased efficacy in the Ca 2ϩ entry pro- Next, the effect on cAMP accumulation of the incremental increases in [Ca 2ϩ ] i caused by increasing [Ca 2ϩ ] ex following CPT-cGMP treatment was examined. Fig. 9 depicts the effect of Ca 2ϩ entry through CNG channels promoted by treatment with either 50 or 100 M CPT-cGMP on cAMP accumulation in C6-2B cells infected with the CNG channel construct. Increasing the [Ca 2ϩ ] ex from 0 to 4 mM resulted in an increased inhibition in cAMP accumulation with both [CPT-cGMP]. The cells pretreated with 100 M CPT-cGMP yielded the largest inhibition, maximally 32% (Fig. 9). Once again, experimental conditions that alter the [Ca 2ϩ ] i rise produced by Ca 2ϩ entry through the CNG channel achieve corresponding changes in the inhibition of cAMP accumulation. The data above clearly show that Ca 2ϩ entry through an expressed CNG channel regulates the endogenously expressed Ca 2ϩ -inhibitable adenylyl cyclase. Finally, it was of interest to compare regulation of cAMP accumulation by the CNG channel with the normal physiological mode of Ca 2ϩ regulation of cAMP accumulation in these cells, i.e. capacitative Ca 2ϩ entry.
Comparison of the Efficacy of CNG Channel-promoted Ca 2ϩ Entry Versus Capacitative Ca 2ϩ Entry-We had previously established the exclusive ability of capacitative Ca 2ϩ entry to regulate the endogenously expressed Ca 2ϩ -inhibitable adenylyl cyclase in C6-2B cells (16). Other modes of inducing [Ca 2ϩ ] i rises, such as release from intracellular stores or an extremely robust [Ca 2ϩ ] i rise produced by ionophore treatment, were ineffective (16). Therefore, the ability of a [Ca 2ϩ ] i rise emanating from expressed CNG channels to regulate the cyclase was somewhat unexpected. The next set of experiments was designed to examine the relative efficacy of these two forms of Ca 2ϩ entry (endogenous CCE versus heterologously expressed, CNG channel-promoted Ca 2ϩ entry) to regulate cAMP accumulation in C6-2B cells.  (Fig. 10B). Next, the effects of these two modes of Ca 2ϩ entry were compared with respect to their ability to regulate cAMP accumulation (Fig. 11). In both Ca 2ϩ entry protocols, an increase in the amount of inhibition in cAMP accumulation was observed with increasing [Ca 2ϩ ] ex . CCE produced greater inhibition of the cyclase, with a maximal inhibition of 40%, using a [Ca 2ϩ ] ex of 4 mM. In comparison, CNG channel-promoted Ca 2ϩ entry inhibited cAMP accumulation by 32% at the same [Ca 2ϩ ] ex . At first glance, it may appear that CCE is more effective at regulating cAMP accumulation, but when the [Ca 2ϩ ] i rise produced by these two Ca 2ϩ entry methods is compared (see Fig. 10 (Fig. 10). Therefore, there is a rough correlation between the [Ca 2ϩ ] i levels reached and the amount of inhibition of cAMP accumulation observed. This finding argues that both of these Ca 2ϩ entry methods are equally efficacious in regulating adenylyl cyclase activity. DISCUSSION The present study has established that Ca 2ϩ entry through a heterologously expressed CNG channel can regulate the endogenous Ca 2ϩ -inhibitable adenylyl cyclase of C6-2B cells. Electrophysiological measurements showed that infection using the novel adenovirus construct coding for the olfactory CNG channel ␣-subunit achieved expression in more than 70% of the cells, which is remarkable, given the refractory nature of these cells to transient transfection. The expressed channel behaved normally, based on cyclic nucleotide dependence, conductance, and Mg 2ϩ block. Subsequently, substantial Ca 2ϩ entry was observed in populations of cells, which was dependent not only on the [CPT-cGMP] and [Ca 2ϩ ] ex but also on the time of exposure to CPT-cGMP. The Ca 2ϩ entry through the CNG channel inhibited the endogenous adenylyl cyclase activity of C6-2B cells. The degree of inhibition mirrored the magnitude of the [Ca 2ϩ ] i rise generated by the various experimental conditions. For instance, a relatively small [Ca 2ϩ ] i rise generated either by low [CPT-cGMP] or short pretreatment times caused a relatively small inhibition of the adenylyl cyclase. When the [CPT-cGMP] or exposure time was increased, the degree of inhibition of the cyclase was increased.
We had previously shown that capacitative Ca 2ϩ entry regulates the Ca 2ϩ -sensitive adenylyl cyclase in nonexcitable cells, whether the cyclase was endogenously or heterologously expressed (15,16,25). Even extremely high [Ca 2ϩ ] i levels, achieved as a consequence of ionophore treatment, were unable to regulate the adenylyl cyclase activity. These and other data (15,16,25) led us to suggest that the adenylyl cyclase and Ca 2ϩ entry channels must be located in similar microdomains in the cell. Therefore, the present findings, that Ca 2ϩ entry through a heterologously expressed CNG channel regulates adenylyl cyclase, were somewhat unexpected. Indeed, when CCE and CNG channel-promoted Ca 2ϩ entry were compared, it was quite evident that they were equally efficacious at modulating adenylyl cyclase activity. Although CCE could achieve slightly greater inhibition of cAMP accumulation compared with CNG channel-promoted Ca 2ϩ entry (40 versus 32%, respectively), the [Ca 2ϩ ] i rise produced by CCE was substantially larger at a given [Ca 2ϩ ] ex . Extending the rationale that led us to conclude that the CCE channel is functionally colocalized with the Ca 2ϩsensitive adenylyl cyclase, it can also be asserted that the CNG channel must also allow Ca 2ϩ entry in the vicinity of the cyclase and, therefore, be targeted to this same domain.
The present findings not only point to a functional colocalization between CNG channels and the Ca 2ϩ -sensitive adenylyl cyclase, they also strengthen the notion that CNG channels may function as a pathway for Ca 2ϩ entry that is not dependent on Ca 2ϩ store depletion or membrane depolarization. It has been clear for some time that Ca 2ϩ entry through CNG channels plays an important role in transduction and adaptation in visual and olfactory receptors (Refs. 26 -28; reviewed in Refs. 29 and 30). In the cone synapse, it has been shown that CNG channels, as well as voltage-gated Ca 2ϩ channels, are involved in exocytosis of synaptic vesicles (31). Furthermore, it has been shown that exocytosis in cone synapses can be modulated by NO, by affecting cGMP production and altering CNG channel activity (32).
CNG channels have also been postulated to play a role in synaptic plasticity, a process that is dependent on Ca 2ϩ . In the hippocampus, an olfactory-like CNG channel has been found in cell bodies and processes of CA1 and CA3 neurons (8), which express high levels of two Ca 2ϩ -stimulable adenylyl cyclases, types I and VIII (33). Based on these observations, it has been suggested that modulation of adenylyl cyclase activity by Ca 2ϩ entry through the CNG channel in CA1 neurons may participate in maintenance of long term potentiation (8). Evidence in support of this proposal is that hippocampi isolated from an olfactory CNG channel null mouse were impaired in their ability to produce long term potentiation in response to -burst stimulation (34). Another tissue in which CNG channels have been detected is the heart (4,9). The heart is also one of the most abundant sources of Ca 2ϩ -inhibitable adenylyl cyclases, types V and VI (35,36). We had earlier proposed that the existence of feedback loops between cAMP-controlled Ca 2ϩ entry and Ca 2ϩ -inhibitable adenylyl cyclases could give rise to oscillations in both [cAMP] and [Ca 2ϩ ] i (14). The present finding that Ca 2ϩ entry through a CNG channel can inhibit a Ca 2ϩ -inhibitable adenylyl cyclase may provide a molecular basis for such a proposal.
For the present, the ability of Ca 2ϩ entry through a heterologously expressed CNG channel to regulate a Ca 2ϩ -sensitive adenylyl cyclase extends earlier observations that endogenous CCE mechanisms could regulate heterologously expressed adenylyl cyclases (15). This finding may suggest that Ca 2ϩ -sensitive adenylyl cyclases and Ca 2ϩ entry mechanisms are endowed with common characteristics, such as preferential solubility in cholesterol-rich domains (37), that ensure their coincidence in microdomains of the plasma membrane. FIG. 11. Comparison of the efficacy of CNGC-promoted versus capacitative Ca 2؉ entry in regulating cAMP accumulation. Cells expressing the CNG channel were pretreated with CPT-cGMP (100 M, circles) or TG (100 nM, triangles) in nominally Ca 2ϩ -free Krebs buffer 4 min prior to cAMP determination. cAMP accumulation was measured over a 1-min period in the presence of forskolin (10 M), isoproterenol (10 M), and added [Ca 2ϩ ] ex (0, 1, 2, or 4 mM, as indicated). Values are expressed as the percentage of cAMP accumulation compared with the calcium-free condition (channel-infected control, 2.66; TG-treated control, 2.47). All calcium-containing conditions differ significantly from the relevant calcium-free conditions, as judged by Student's t test (p Ͻ 0.005). The CPT-cGMP-treated data set does not differ significantly from the TG-treated data set, as judged by two-way analysis of variance.