Introduction

Top Abstract Introduction Methods Results Discussion References

NO has been shown to inhibit platelet aggregation and adherence, mediate vascular smooth muscle relaxation and regulate neurotransmitter release (reviewed in Bredt and Snyder, 1994; Garthwaite, 1991; Stamler, 1994). Within the nervous system, NO is believed to exert a regulatory role (e.g., in smooth muscle relaxation on stimulation of peripheral nonadrenergic-noncholinergic neurons in the autonomic nervous system; in cerebral blood flow in response to activity-dependent activation of neuronal NOS; and in synaptic plasticity such as the development of long-term potentiation in hippocampal neurons).

Recent studies have focused on identifying the specific processes and target molecules that are altered by NO and specifying the underlying mechanism or mechanisms by which NO elicits its effects on those targets (Bredt and Snyder, 1994; Garthwaite, 1991; Stamler, 1994). One mechanism by which NO functions as a signalling molecule is to interact with iron-heme groups of proteins, as exemplified by its action on soluble guanylate cyclase. It is the subsequent increase in cGMP levels that is believed to mediate NO-induced vasodilation. However, the functional significance of cGMP in the neuronal cells remains unclear. The failure of cGMP and NOS, as determined by immunocytochemistry, to be colocalized in the nervous system argues that NO acts in ways distinct from changes in cGMP (Bredt and Snyder, 1994; Bredt et al., 1990). Recently, a number of other actions of NO have been proposed For example, NO-mediated modification of glyceraldehyde-3'-phosphate dehydrogenase or poly(ADP-ribose) synthetase by NAD has been proposed to result in the depletion of energy stores leading to cell death (Dimmeler and Brune, 1992; Zhang and Snyder, 1992; Zhang et al., 1994). NO binding to iron-sulfur centers of enzymes (e.g., cis-aconitase, is believed to participate in the cytotoxic actions of macrophages; Drapier et al., 1993, 1994). S-Nitrosylation, via the transfer of NO⁺ from nitrosothiols, is another mechanism by which NO has been shown to modify proteins (Stamler, 1994).

Cellular components generally have been implicated as targets of NO based on the ability of NO or NO-releasing compounds to alter their activities in vitro. Thus, whether these target proteins are altered in intact cells in response to NO, or what the possible physiological consequences may be, remains speculative. Our approach to discerning how NO functions is to examine its effects on intact cells. We have demonstrated that the addition of NO donor compounds or NO gas to Dictyostelium discoideum cells inhibits their aggregation via cAMP pulses, and does so independently of any changes in cGMP or guanylyl cyclase activity (Tao et al., 1992, 1996, 1997). The sum of the data indicated that NO specifically alters either a regulatory domain of the adenylyl cyclase itself or a distinct regulatory moiety. To our knowledge, this was the first demonstration that the effects of NO in intact cells could be mediated by its effect on cAMP production. We were therefore interested in testing the hypothesis that mammalian cells could respond to NO by altering their adenylyl cyclase activity. In this article, we report the effects of NO on cAMP production in the N18GT2 neuroblastoma cell line.

	Methods

Top Abstract Introduction Methods Results Discussion References

Cells and incubation conditions. The N18TG2 neuroblastoma cell is a clone derived from the C1300 spontaneous mouse tumor (Augusti-Tocco and Sato, 1969; Minna and Gilman, 1973). N18TG2 neuroblastoma cells were grown in Dulbecco's modified Eagle's/Ham's F12 (1:1) medium (Biowhittaker, Walkersville, MD) containing 10% heat-inactivated donor bovine calf serum (JHR Biosciences, Lenexa, KS) and penicillin/streptomycin (Sigma Chemical, St. Louis, MO). Media was changed 16 to 24 hr before experimentation, and cells were confluent at the time of experimentation. Cells were dissociated from flasks by gentle trituration with phosphate buffered-saline containing 0.6 mM EDTA, and resuspended to 2 × 10⁶ cells/ml in GBSS containing 0.1 mg/ml fatty acid-free bovine serum albumin, 10 mM NaHEPES, pH 7.4, and 0.1 mM Ro20-1724 (a PDE inhibitor). Cells were incubated for 30 min at 37°C in the presence or absence of indicated additives, and then 400 µl was pipetted into test tubes containing 100 µl of indicated reagents for stimulated production of cAMP. The incubation proceeded for 4 min at 37°C. This time period was chosen to ensure we were monitoring cells when cAMP accumulation was linear with time (Walz et al., 1987). The incubation was terminated by the addition of 50 µl of 500 mM Na acetate, pH 4.5, and boiling for 4 min, and cAMP content was determined in the supernatant fractions after sedimentation.

cAMP determinations. For most of the experiments reported here, cAMP levels were quantified using a modification of the Gilman assay (Brostrom and Kon, 1974). An alternative modification of the Gilman assay was developed in our laboratory that uses the filtration technology of the Tomtech cell harvester (Turku, Finland) and the Wallach Betaplate scintillation spectrometer, which quantifies radioactivity by liquid scintillation counting of the filtermats. The assay mixture [containing 20 mM sodium acetate, pH 4.5, 0.25 mg/ml calf thymus histone, 0.6 pmol [³H]cAMP and standards (0.5-15 pmol) or unknown samples] was incubated in a modified-flat 96-well ELISA plate (Corning, NY) in a total volume of 150 µl. The incubation was initiated by the addition of 1.2 µg of bovine cardiac protein kinase A plus 2 µg of bovine serum albumin into each well and incubated for 1 hr on ice. Incubations were terminated by the addition of 50 µl of 1% -globulin and gently swirled for 10 min, followed by the addition of 75 µl of 30% polyethylene glycol-8000 and swirling for an additional 15 min. Using a Tomtec 96-well cell harvester, samples were filtered onto a grid-printed Fiberglas filtermat (Wallach type B) and rinsed five times with 10 mM potassium phosphate, pH 6.0, at 4°C. The filtermat was dried for 3.5 min by microwave, and then 50 µl of Wallach Betaplate Scintillant was applied to both sides of each grid site on the filter. The filtermat was then placed into a sealing bag, heat sealed and positioned in a cassette to be counted by a Wallach Betaplate liquid scintillation spectrometer.

Protein determinations. An aliquot of the cells was dissolved in 1 M NaOH for determination of the protein by the method of Bradford (1976).

Data analysis. Each determination was performed in triplicate and assayed in duplicate. The average coefficient of variation for the data values within each experiment was <10%.

Materials. NO gas was obtained from Acetylene Gas (St. Louis, MO). SNP was from Sigma Chemical. SIN-1 was kindly provided by Casella AG (Pharmaforschung Galenik, Frankfurt, Germany). PTX was purchased from List Biological Laboratories (Campell, CA.). [³H]cAMP was from Dupont-NEN (Boston, MA). Forskolin, cAMP, protein kinase A, calf thymus histone, -globulin, secretin, PGE₁ and polyethylene glycol-8000 were from Sigma Chemical. All other reagents were obtained from Sigma Chemical or other standard sources.

	Results

Top Abstract Introduction Methods Results Discussion References

NO inhibits hormone- and forskolin-stimulated cAMP production. N18TG2 adenylyl cyclase is known to be stimulated by G_s in response to stimulation of prostanoid (PGE₁ and prostacyclin), secretin/vasoactive intestinal peptide and adenosine A₂ receptor ligands (Gilman and Nirenberg, 1971; Brunton et al., 1976; Howlett, 1982; Roth et al., 1984). We examined the hormone-stimulated adenylyl cyclase in these cells to determine whether NO was able to alter signal transduction. Initial studies used the NO-generating agent SNP as a means of supplying NO to the incubation media. Cells were preincubated in medium containing Ro20-1724, a PDE inhibitor (Walz et al., 1987), in the absence or presence of SNP for 30 min. Cells were then stimulated with either PGE₁ or secretin, and the levels of cAMP that accumulated after 4 min were determined. As shown in figure 1, pretreatment of cells with SNP resulted in the attenuation of cAMP accumulation in response to hormone. Approximately 50% of the activity was lost in each case when cells were preincubated for 30 min with SNP. In this experiment, similar degrees of inhibition were obtained using 0.1 mM and 1.0 mM SNP, suggesting that the maximum effect had been reached in this concentration range. Basal accumulation of cAMP, a measure of cAMP generated in the absence of hormonal stimulation, was unaffected by pretreatment with SNP.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   SNP inhibits hormone- and forskolin-stimulated cAMP production. N18TG2 cells were preincubated for 30 min in GBSS plus PDE inhibitor (filled bars) plus 0.1 mM SNP (open bars) plus 1.0 mM SNP (hatched bars). Cells were then treated with either 10 µM PGE1, 0.3 µM secretin, 10 µM forskolin or without any additions (basal), as indicated. After an additional 4 min, cAMP levels were determined as described in Materials and Methods. The data are representative of two experiments in which all of the conditions for adenylyl cyclase activation were examined simultaneously. Each condition, however, has been examined individually a minimum of four additional times with the same results.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Both SNP and SIN-1 inhibit forskolin-stimulated cAMP production in a concentration-dependent manner. N18TG2 cells were preincubated in GBSS plus PDE inhibitor and the indicated concentrations of (A) SNP or (B) SIN-1. After 30 min, cells were incubated for an additional 4 min with no additions (hatched bars) or with 10 µM forskolin (filled bars). cAMP levels were then determined; 100% represents the amount of cAMP produced by forskolin-stimulated cells that had not been preincubated with NO-donor compounds (634 pmol/mg in A and 720 pmol/mg). The data are representative of two experiments.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   NO gas inhibits forskolin-stimulated cAMP production. N18TG2 cells were preincubated for 30 min in GBSS plus PDE inhibitor. Subsequently, aliquots of a saturated solution of NO were added to achieve the indicated final concentrations. The addition was made simultaneously with 10 µM forskolin (). Alternatively, no forskolin was present (). After an additional 4 min, cAMP levels were determined. The level of cAMP in cells treated with forskolin alone (650 pmol/mg) was assigned the value of 100%. All other levels are reported relative to that maximum value. The data represent the average of two experiments.

NO does not potentiate the action of G_i. We wanted to determine whether the attenuating effect of NO on adenylyl cyclase is mediated through a G_i protein. Hormonal inhibition of adenylyl cyclase has been shown to occur in N18TG2 cells in response to alpha-2 adrenergic, muscarinic cholinergic, delta opiod and CB1-cannabinoid receptors (Matsuzawa and Nirenberg, 1975; Sharma et al., 1977; Nathanson et al., 1978; Sabol and Nirenberg, 1979; Howlett et al., 1986; Devane et al., 1986; Bidaut-Russell and Howlett, 1987). PTX is used as a technique to determine involvement of G protein or proteins in receptor-mediated activities. It acts on the heterotrimeric form of the G_i family by catalyzing ADP-ribosylation of a cysteine residue near the carboxyl terminus of the alpha subunit and precludes the signal for inhibition of adenylate cyclase. As a control, data shown in figure 4 are consistent with cannabinoid receptor-mediated inhibition of adenylyl cyclase involving a G_i protein. The inhibition by the cannabinoid receptor agonist was abolished in the PTX-pretreated group compared with the control. SNP inhibited cAMP accumulation to the same extent both before and after pretreatment with PTX. This failure of PTX to prevent the NO-mediated inhibition contrasts with the ability of the toxin to abolish the G_i-mediated inhibition of cAMP accumulation. This study provides evidence against the involvement of a membrane-bound G protein in the action of NO.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   PTX pretreatment does not alter the effect of NO on N18TG2 cells. N18TG2 cells were pretreated for 18 hr with or without 50 ng/ml PTX as indicated. At the end of this incubation period, cells were harvested and preincubated for 30 min in GBSS plus PDE inhibitor in the presence or absence of 1 mM SNP, as indicated. Aliquots of cells were subsequently stimulated with 10 µM forskolin alone (filled bars), forskolin in combination with 1 µM desacetyllevonantradol (hatched bars) or unstimulated (clear bars). Samples were then assayed for cAMP accumulation. The data are representative of two experiments.

cAMP accumulation in N18TG2 cells is not CaM regulated. It has been reported that the Ca⁺⁺/CaM regulation of recombinant type I adenylyl cyclase in membranes from Sf9 cells is inhibited by high concentrations of NO, although basal enzyme activity is not altered (Duhe et al., 1994). To examine whether the adenylyl cyclase in N18TG2 cells is regulated by CaM, we incubated cells in the presence of the Ca⁺⁺ ionophore A23187 to cause an influx of Ca⁺⁺ from the media. Under these conditions, CaM-regulated adenylyl cyclases show dramatic increases in cAMP production, increases that are particularly evident when the response to forskolin or hormone activation is monitored (Choi et al., 1992). As shown in figure 5A, this was not the case for N18TG2 cells. The stimulation of cAMP accumulation by forskolin and PGE₁ showed a minor increase (<10%) on addition of the ionophore. As an alternative means of examining the possible role of CaM as a regulator of cAMP synthesis in N18TG2 cells, we incubated cells with the CaM inhibitor calmidazolium for 30 min and then determined their responses to forskolin and hormone activation. As shown in figure 5B, preincubation with calmidazolium also did not significantly alter the ability of either forskolin or PGE₁ to stimulate cAMP production. Basal cAMP accumulation was also unaffected. From these data, we conclude that the predominant isoform of adenylyl cyclase in N18TG2 cells is not activated by CaM. Thus, the dramatic inhibition of NO on forskolin or hormone activation of the enzyme is not mediated by CaM.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   cAMP accumulation in N18TG2 cells is not regulated by CaM. A, Cells were incubated in GBSS plus PDE inhibitor for 30 min. Basal, forskolin (10 µM) and PGE1 (10 µM) stimulated cAMP accumulation were determined in the absence (filled bars) or presence (hatched bars) of 1 µM A23187. The data are representative of five experiments. B, Cells were incubated for 30 min in GBSS plus PDE inhibitor in the absence (filled bars) or presence (hatched bars) of calmidazolium (1 µM). After that time, basal and forskolin- (10 µM) and PGE1- (10 µM) stimulated cAMP levels were determined. Levels are presented relative to that achieved in forskolin-stimulated cells, which was 625 pmol/mg. The data are representative of two experiments.

The inhibition of cAMP production by NO is readily reversible. We determined whether NO-treated cells could recover the ability to produce cAMP in response to forskolin when NO was eliminated from the media. In the experiment shown in figure 6A, cells were preincubated for 30 min with 0.1 mM SNP, washed free of the donor compound with a single 1-min sedimentation and resuspended in fresh media without SNP. After the indicated periods of time, forskolin-stimulated cAMP levels were determined. When determined immediately on resuspension of the cells in fresh media, forskolin-stimulated cAMP accumulation was inhibited by 40% to 45% by the SNP treatment. At 5 min after the removal of SNP, forskolin-stimulated cAMP accumulation remained inhibited, but the inhibition was attenuated (reaching ~30%). At 20 min after the removal of SNP, cells appeared to have recovered completely from the effects of SNP such that cAMP levels in response to forskolin were comparable to those of untreated cells.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   The inhibition of forskolin-stimulated cAMP production by SNP and NO is readily reversible. A, Cells were preincubated for 30 min in GBSS plus PDE inhibitor plus (clear bars) or minus (filled bars) 0.1 mM SNP. Cells were then washed free of SNP and assayed for forskolin-stimulated (10 µM) cAMP accumulation immediately after washing (time 0) or after the indicated time periods. The level of cAMP in forskolin-stimulated cells that had not been treated with NO donor (804 pmol/mg) was assigned the maximum value of 100%. B, Cells were preincubated for 30 min in GBSS containing a PDE inhibitor, after which aliquots of a saturated NO solution were added to a final concentration of 0.5 µM (clear bars) or 5 µM (hatched bars) NO. Filled bars represent cells that were not treated with NO. Immediately after the addition of NO (time 0) or after an additional 5 or 10 min, forskolin-stimulated (10 µM) cAMP levels were determined. The level of cAMP in forskolin-stimulated cells that had not been treated with NO (750 pmol/mg) was assigned the maximum value of 100%. In each case, the data are representative of two experiments.

cGMP does not mediate the effects of NO on cAMP production. Many of the effects of NO have been explained by its ability to stimulate a soluble guanylate cyclase, with the subsequent rise in cellular cGMP mediating the physiological response. cGMP has not been demonstrated to be an effector of adenylyl cyclase, making it unlikely that cGMP is a direct participant in the inhibition of cAMP production by NO. cGMP has been shown to regulate a family of cAMP PDEs (reviewed in Beavo, 1995) and thus could potentially cause a decrease in intracellular cAMP levels by stimulating its degradation. This is also an unlikely explanation for our results. If decreased cAMP levels resulting from NO treatment reflected enhanced PDE activity, we would also expect to see a decrease in basal cAMP accumulation in NO-treated cells. Furthermore, in previous studies it was shown that cGMP does not regulate PDE activity in N18TG2 cells (Walz et al., 1987). However, to address the possibility that cGMP, via some as-yet-undescribed mechanism, mediates the inhibition of forskolin- and hormone-stimulated cAMP production, we incubated cells with 8-bromo-cGMP to determine whether that treatment mimicked the effects of NO. As seen in figure 7, preincubation of cells with 0.5 mM 8-bromo-cGMP for 30 min failed to alter either basal or forskolin-stimulated accumulation of cAMP. The sum of the data argues that cGMP does not play a role in the NO-mediated inhibition of cAMP production.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   8-Bromo-cGMP does not alter cAMP production. N18TG2 cells were preincubated for 30 min in GBSS plus PDE inhibitor. Where indicated, 0.5 mM 8-bromo-cGMP was also present. After this time, basal (clear bars) and forskolin-stimulated (10 µM) (filled bars) cAMP levels were determined. The data are representative of two experiments.

The reactive species appears to be NO. In some cases, it is believed that the effects of NO reflect its conversion to peroxynitrite, whereas in other cases, NO may need to function in conjunction with H₂O₂ to exert its maximum effects (Stamler, 1994). The finding that SIN-1, which favors the production of peroxynitrite (Wink and Ford, 1995), was not as effective as SNP in inhibiting cAMP accumulation in N18TG2 cells would argue that peroxynitrite is not likely to be a critical component. To examine this further, we incubated cells with either SOD, catalase or the combination of both enzymes in the extracellular media at the time of addition of SNP (fig. 8). SOD catalyzes the conversion of O₂ present in the media to H₂O₂, thereby diminishing the possibility that NO would be converted to peroxynitrite. Catalase converts any H₂O₂ present into H₂O plus O₂. After a 30-min incubation with the indicated additions, cAMP production in response to forskolin was determined. As shown, none of the conditions tested altered the effectiveness of NO to inhibit forskolin-stimulated cAMP production. These data support the premise that NO mediates the inhibition of hormone- and forskolin-stimulated cAMP production.

View larger version (28K): [in this window] [in a new window]   Fig. 8.   The effects of SNP are not altered by SOD or catalase. Cells were preincubated for 30 min in GBSS containing a PDE inhibitor (control) or 750 U/ml SOD and 200 U/ml catalase. Alternatively, cells were preincubated for 30 min in GBSS containing a PDE inhibitor plus either 0.2 mM SNP, SNP plus catalase, SNP plus SOD or SNP plus SOD and catalase, as indicated. After a 30-min incubation, forskolin-stimulated (10 µM) cAMP levels were determined. The data are representative of two experiments.

	Discussion

Top Abstract Introduction Methods Results Discussion References

We demonstrated that the addition of NO to N18TG2 cells, in the form of either NO donor compounds or NO gas, inhibits the accumulation of cAMP in response to both hormone and forskolin stimulation. It is unlikely that this inhibition reflects an enhanced degradation of the cyclic nucleotide because a phosphodiesterase inhibitor is present during these incubations and very short incubation times are used for the measurements of cAMP production. In addition, if NO stimulated the activity of a PDE, we would expect to see a decrease in basal cAMP accumulation. That was not observed. Fisch et al. (1995) observed a synergistic effect of adenylyl cyclase activators and SIN-1 on platelet cAMP levels, an effect that was eliminated when cells were incubated with piroximone, an inhibitor of the type III PDE. The authors conclude that an NO-mediated increase in cGMP leads to an inhibition of type III PDE and thus an apparent increase in cAMP levels. This mechanism would not apply to our system; N18TG2 cells do not express a cGMP-regulated PDE (Walz et al., 1987), and we observe a decrease, not an increase, in cAMP levels. It is also unlikely that the inhibition reflects a decreased ATP pool. We have observed that NO donor compounds inhibit adenylyl cyclase activity in isolated plasma membranes, where ATP is not a limiting factor (Howlett et al., unpublished results). Our experiments also indicated that cGMP does not play a role in this response.

Currently, eight adenylyl cyclase isoforms have been identified, each of which is characterized by distinct and relatively complex levels of regulation (reviewed in Iyengar, 1993; Mons and Cooper, 1995; Sunahara et al., 1996). The isoform or isoforms of adenylyl cyclase expressed in N18TG2 cells have not yet been cloned, but the regulatory properties indicate the presence of a type V or VI isoform. The enzyme is under effective regulation by hormone receptors that stimulate via G_s and inhibit via G_i. Furthermore, basal and forskolin-stimulated cAMP accumulation by N18TG2 cells is not significantly modified by the addition of Ca⁺⁺ ionophores or CaM antagonists. Similar experiments using membranes to characterize the adenylyl cyclase activity in N18TG2 cells confirm its lack of regulation by CaM (Howlett et al., unpublished observations).

We do not believe that the inhibitory action of NO on adenylyl cyclase is mediated through a G protein. G_s is not the target of NO in N18TG2 cells because forskolin-stimulated adenylyl cyclase activity was also inhibited. If the inhibition resulted from an activation of the G_i protein, we would have expected this effect to be attenuated by prior treatment with PTX. This was not the case. Duhe et al. (1994) have shown that the Ca⁺⁺/CaM regulation of recombinant type I adenylyl cyclase in membranes from Sf9 cells is inhibited by saturating concentrations of NO, although basal enzyme activity is unaffected. The authors proposed that NO may function by oxidizing a cysteine residue or residues at the CaM binding site of the adenylyl cyclase. Although NO inhibited the CaM regulation of type I adenylyl cyclase in Sf9 membranes, it did not alter the ability of forskolin to stimulate the enzyme. This contrasts with the effects that we have observed in N18TG2 cells, in which NO did inhibit forskolin-stimulated cAMP accumulation. These observations also suggest that although different families of adenylyl cyclases may be targets for NO, the underlying means by which their activities are affected are isoform specific.

We have shown that the NO-mediated inhibition is readily reversible, requiring <20 min of cell incubation in the presence of reduced levels of NO to regain complete activity. This rapid reversibility indicates that NO-mediated modifications such as ADP-ribosylation or tyrosine nitrosation, which have not been demonstrated to be readily reversible in intact cells, would not underlie the effects of NO on cAMP production. Our studies also indicated that peroxynitrite is not the perpetrator of the inhibition of cAMP accumulation. Peroxynitrite is a highly reactive species that can oxidize lipids, proteins and DNA and whose presence is generally associated with neurotoxicity (Stamler, 1994). The action of NO on the heme-iron of guanylyl cyclase is rapidly reversible on removal of the source of NO. However, none of the adenylyl cyclases identified to date have been shown to be regulated by heme iron or iron-sulfur centers. The most likely means by which NO alters adenylyl cyclase activation in N18TG2 cells is via S-nitrosylation. NO, in particular the NO⁺ form, has been shown to readily react with cysteine residues in proteins, forming S-nitrosothiols. This modification is readily reversible depending on the redox state of the cells (Stamler, 1994).

Adenylyl cyclases have been proposed to serve as "coincidence detectors," which are to be acted on simultaneously by various neuromodulators, and then submit the results of these influences to the cell via cAMP changes (Bourne and Nicoll, 1993). The response to two different signals could result in a synergistic response. For example, type II adenylyl cyclase exhibits a response to G_s stimulators that can be significantly amplified by the simultaneous presence of beta gamma generated by G_i stimulators or by protein kinase C activation in response to G_q stimulators. Alternatively, "discordant" coincidence detection (Mons and Cooper, 1995) could occur by which the response to one signal is attenuated by the concurrent presence of the second signal. For example, neurotransmitters that produce a flux of Ca⁺⁺ would result in significant stimulation of type I adenylyl cyclase that would then be optimally inhibited by G_i stimulation. As another example, type V adenylyl cyclase responds to G_s stimulation but is inhibited by Ca⁺⁺. Our studies describe a novel coincidence detection system for adenylyl cyclase; the inactivation by NO of the enzyme's responses to G_s-stimulatory neuromodulators. Removal or decreased concentrations of the stimulus leads to a rapid reversal of its effects, optimizing cell sensitivity to additional or repeated input.

Given the select nature of the cells that express NOS in the nervous system, it would appear that the NO necessary to influence cell response to adenylyl cyclase activators may frequently be supplied by neighboring cell types (Bredt et al.1990, 1994). In the cerebellum, the proximity of NOS-containing cells to Purkinje cells, which do not make NO, would be consistent with a role of NO as a transcellular regulator of the soluble G-cyclase in Purkinje cells. In the peripheral nervous system, evidence also supports the role of neuronal NO as an transcellular communicator (e.g., in targeting smooth muscle as the responding cells). Alternatively, it is possible that N18TG2 cells produce NO and do so in a manner that regulates the activation of its adenylyl cyclase. This scenario would simulate that of cerebellar granule cells that contain NOS and respond to their own production of NO with enhanced cGMP synthesis (Malcolm et al., 1996). In preliminary experiments, using N18TG2 cells incubated at high density to maximize localized concentrations of endogenously produced NO, we have observed that hormone-stimulated cAMP production is attenuated. The addition of L-N^G-methylarginine, a NOS inhibitor, appeared to alleviate this inhibition, suggesting that under these conditions, endogenously produced NO may regulate the adenylyl cyclase in N18TG2 cells (Shipley S, unpublished observations). We are currently pursuing these observations to determine the levels of NO produced by N18TG2 cells, the physiological effectors of its synthesis and their effects on cAMP production.