Identification and molecular characterization of a m5 muscarinic receptor in A2058 human melanoma cells. Coupling to inhibition of adenylyl cyclase and stimulation of phospholipase A2.

We report the identification and biochemical characterization of an endogenous m5 muscarinic acetylcholine receptor (mAChR) in the A2058 human melanoma cell line. This is the first demonstration of a m5AChR outside the central nervous system. The unusual effector coupling of this endogenous m5AChR is presented. The coding region amplified by polymerase chain reaction was identical to the known m5AChR sequence. Binding studies indicated a Kd of 99 ± 6 pM and a Bmax of 45 ± 4 fmol/mg membrane protein. This m5AChR coupled to stimulation of arachidonic acid release and to a 50% inhibition of forskolin-stimulated cAMP accumulation. The inhibition of cAMP production was insensitive to pertussis toxin treatment, but was dependent upon extracellular calcium. In contrast to the odd mAChR pattern, no cAMP was produced in response to carbachol (CC) stimulation. Moreover, no release of inositol phosphates could be measured after CC treatment despite the presence of at least 2 phospholipase C isoforms in A2058 cells. CC-stimulated arachidonic acid release (EC50 = 17.8 ± 0.1 μM) was dependent upon external Ca2+, with marked reduction after coincubation with EGTA, Co2+, or high doses of verapamil (IC50 = 166 μM) or diltiazem (IC50 = 243 μM). Brief exposure to phorbol 12-myristate 13-acetate augmented CC-stimulated arachidonic acid release, whereas prolonged phorbol 12-myristate 13-acetate treatment resulted in down-regulation of release. Activation of the m5AChR resulted in Ca2+ influx that was attenuated by muscarinic antagonism and removal of extracellular Ca2+. A2058 cells exposed to CC had no alteration of cell shape or growth potential in monolayer culture, however, a statistically significant reduction in density-independent growth was observed over the range of CC concentrations from 0.1 to 100 μM. This endogenous m5AChR has a novel signal transduction coupling profile and receptor activation reduces clonogenic potential.

The m5 receptor has been found in brain in low abundance, however, endogenous m5 has only been demonstrated in the brain (1,2,4,19,20). All other mAChR receptor subtypes are found in the brain as well as other systemic sites. The m2 receptor is abundant in heart muscle and is found in smooth muscles, such as the small and large intestine, trachea, bladder, and uterus (1,2,4,6,18,20,21). In contrast, the oddnumbered receptors are not found in heart, and only the m3 subtype may be found in other smooth muscle-containing organs. Both m1 and m3 are located in glandular tissues such as lacrimal gland, exocrine pancreas, and salivary gland (22). Functional receptors have been found in a variety of brain cell lines and brain cancer cell lines. An adherent variant small cell lung cancer cell line has been shown to express the m3 subtype receptor. mAChRs have not been reported previously in melanomas, a malignancy of basal layer pigmented neuroepithelial cells.
We now demonstrate and biochemically characterize a m5 AChR in an early passage human melanoma cell line. This m5 receptor is present in extremely low abundance and mediates the inhibition of anchorage independent proliferation. It is functionally associated with a unique combination of signal transduction pathway coupling including calcium mobilization, release of arachidonic acid, inhibition of the generation of cAMP, and lack of inositol polyphosphate production.
Cell Culture-The early passage A2058 human melanoma cell line (Յ passage 20) was maintained in culture in Dulbecco's modified Eagles medium with 10% fetal calf serum, as described, and used for no more than 8 passages (23). Sublines derived by limiting dilution cloning (A2058 -1F5, Յ passage 25) were studied preferentially, unless indicated otherwise. This subline has identical growth and functional characteristics as the parental A2058 cell line. Where indicated, CHO cells stably transfected with the m5AChR (CHOm5) were used as controls for the muscarinic receptor responses; these cells were maintained in ␣-minimal essential medium with 10% fetal calf serum.
Receptor Binding Studies-Plasma membranes were prepared from A2058-1F5 cells as described previously, and competition and saturation binding assays were performed with N-[methyl-3 H]scopolamine as described (24). Assay solutions were incubated for 1 h at 30°C in a final assay volume of 0.5 ml and a final membrane concentration of 40 -400 g of protein/ml. Membranes were rapidly filtered over Whatman GF/B filters pretreated for 3 h with 0.1% polyethyleneimine (v/v, pH 7.4) using an Inotech (Lansing, MI) multi-position cell harvester, and washed 3 times (50 mM Tris, 0.5 mg/ml bovine serum albumin, pH 7.4). Filters containing washed membranes were transferred to scintillation vials and incubated in 0.1% Triton X-100 (v/v) overnight before addition of scintillation mixture and counting. Binding data were analyzed with the GraphPad program (GraphPad Software, San Diego, CA) which performs weighted non-linear least squares curve fitting to the general model of Feldman (25).
Analysis of Signal Transduction Pathways-Carbachol (CC), a stable acetylcholine analog, was used as the agonist for all signaling experiments and data are presented as percent stimulation (mean Ϯ S.E.) of at least three independent experiments. The release of cAMP in response to CC stimulation of A2058 -1F5 cells was measured as described previously using a radioimmunoassay (9,26). cAMP accumulation was measured in the presence of calcium-free Eagle's number 2 media containing either 1 mM added CaCl 2 or with addition of 0.5 mM EGTA. Where indicated, cells were incubated with pertussis toxin for up to 18 h prior to assay.
Arachidonic acid release assays were performed as described (10) with the following changes. A2058-1F5 cells were plated at 100,000 per well in 24-well plates and grown for 24 h to 80% confluence, after which they were labeled with 0.25 Ci of [ 3 H]arachidonic acid for 18 h. Bicarbonate-free Eagle's minimal essential medium supplemented with 1.0 mM CaCl 2 , 0.1 mM MgCl 2 , and 0.1% fatty acid-free bovine serum albumin was used as the reaction buffer (Buffer A). Release reactions were carried out for 15 min unless stated otherwise.
Intracellular calcium measurements were made using single cell fluorescence videoimaging as described (11). Briefly, cells were grown on glass coverslips precoated with 10 ng/ml poly-L-lysine, air dried, and then overcoated with 300 g/ml vitrogen. For imaging, cells were loaded with 1 M FURA-2-acetoxymethyl ester for 30 min at 37°C in growth media, washed once, and stored in Eagle's number 2 media containing 1 mg/ml fatty acid-free bovine serum albumin and 20 mM HEPES buffer (pH 7.4) at 25°C for no more than 45 min prior to use. Cells were bathed in buffer containing reagents and calcium as indicated. FURA-2 fluorescence was measured at an emission wavelength of 510 nm in single cells mounted on a Nikon Diaphot microscope illuminated alternately with 350 and 380 nm light (band pass 4 nm) using a SLM-Aminco DMX-1000 spectrofluorometer with an intensified CCD camera as the emission detector (SLM-Aminco, Urbana, IL). Data are presented as the ratio of fluorescence 340/380, as an index of changes in intracellular calcium concentration.
Several experimental protocols for the quantitation of inositol phosphates were tested using CC stimulation of A2058 cells. These included chloroform/methanol extraction and Dowex AG 1-8 anion exchange chromatography (23), a modification of this procedure (27), and the commercially available radioreceptor assay IP 3 kit (DuPont NEN). Varied cell numbers and reaction times of 0 -60 s and up to 15 min were tested. In all experiments, CHOm5 cells were used as a positive control.
Cell Lysis and Immunoblot-A2058 cell lysates were made from subconfluent cells either untreated or incubated with 20 M CC for either 5 or 15 min using modified RIPA buffer in the presence of protease and phosphatase inhibitors (50 mM Tris, pH 7.5, 1% Nonidet P-40, 0.1% sodium deoxycholate, 150 mM NaCl, 4 mM EDTA, 10 mM sodium pyrophosphate, 2 mM sodium orthovanadate, 10 mg/ml leupeptin, and 10 mg/ml aprotinin). Lysates were incubated on ice for 30 min, centrifuged at 14,000 ϫ g for 10 min at 4°C, and the supernatant recovered. Lysate samples, 50 g, were electrophoresed on 4 -12% gradient polyacrylamide gels, transferred, blocked, and probed with antibody against phospholipase C␤-1, 5 g/ml, or PLC␥-1, 0.1 g/ml. Blots were washed, incubated with 125 I-protein A, and exposed to film. For immunoprecipitations, equal protein quantities were incubated with anti-phosphotyrosine antibody (1 g/100 g of protein) or anti-PLC␥-1 (0.1 g/100 g of protein) for 90 min at 4°C. Immunocomplexes were collected with 50 l of protein A-Sepharose beads for 1 h at 4°C, washed with RIPA buffer, eluted in 2 ϫ Laemmli sample buffer, and then processed as described for the immunoblot. Immunoprecipitated proteins were immunoblotted overnight with anti-PLC␥-1 (0.1 g/ml) and visualized as above (28).
Anchorage Independent Growth-The effect of CC on colony formation in soft agar was determined as described (24,29). Thirty million A2058 cells were plated into a top 3% agar layer in complete media with CC, atropine, both agents, or neither. After culture for 10 days, colonies were counted under phase-contrast microscopy. Data presented are the mean Ϯ S.E. of three independent experiments. Unpaired Student's t test was used for statistical analysis.

PCR Cloning and Sequence Analysis of mAChR in A2058 Cells
Oligonucleotides from conserved transmembrane domains were used initially to PCR amplify a segment of the A2058 mAChR. When the initial fragments were sequenced and confirmed to have high homology to the m5 subtype, further PCR amplification reactions used more selective oligonucleotides. Fig. 1 shows the results of PCR amplification of RNA from CHO cells stably transfected with the human m1-m5AChR (lanes 1-5 in each panel) for comparison to the results for the A2058 -1F5 line (lane 6). In Fig. 1A, one major band with a size of 1200 bp was amplified using primers selected for m5 and was found only in CHOm5 and A2058 -1F5 lanes. This fragment size agreed with the expected size of the amplified region for the m5AChR gene but is not the correct size for the smaller m1 and m3 transcripts, 900 and 983 bp, respectively. No amplified products were observed from the reactions using cDNA from CHOm1, m2, m3, and m4 cells (Fig. 1A, lanes 1-4, respectively). To rule out undetectable products from m1-m4 and to confirm the identify of the band in the m5 and 1F5 lanes, Southern analysis was performed using as a probe a 200-bp PCR product selective for a segment of the third cytoplasmic domain of the m5AChR (Fig. 1A, bottom panel). In order to show that there is only one species of AChR in the A2058 cells, further PCR amplification was done using oligonucleotides selective to the even-numbered AChRs. A band of approximate size of 900 bp was demonstrated only in the CHO m2 and m4 cells after amplification with m2/m4 oligonucleotide pairs. Southern blot hybridization of those gels with an internal and selective m2/4 oligonucleotide confirmed that the 900-bp transcript was present only in the CHOm2 and m4 cells and not in the A2058-1F5 cells. The full coding length of the A2058 AChR has been amplified, subcloned, and sequenced and was found to be homologous to that of the hu-m5AChR. These results demonstrate that the mAChR on the A2058 cells is genetically a m5 muscarinic receptor.

Receptor Characterization
m5 receptor expression level and receptor affinity were determined by radioligand binding experiments. Saturation binding experiments were performed in which N-[methyl-3 H]scopolamine was used to define total binding and unlabeled 10 M atropine was added to saturate specific binding. Fig. 2 shows the binding characteristics of this receptor. Greater than 90% specific binding was observed in A2058 plasma membranes with binding constants of K d ϭ 99 Ϯ 6 pM and B max ϭ 45 Ϯ 4 fmol/mg of membrane protein.

Signal Transduction Coupling of Endogenous mAChR in A2058 Cells
Inhibition of forskolin-stimulated cAMP production-Forskolin treatment of A2058 cells resulted in production of cAMP as shown in Fig. 3A. No production of cAMP was observed when A2058-1F5 cells were exposed to CC in concentrations up to 1 mM (data not shown), as would be expected for an odd-numbered muscarinic receptor signaling profile. In contrast, CC exposure inhibited the stimulation of cAMP production in response to forskolin in a dose dependent fashion, but only by 60%. This was investigated further by exposure to pertussis toxin and by removal of extracellular calcium from the reaction medium. When cells were pretreated with pertussis toxin for up to 18 h prior to activation with forskolin, with or without additional CC, no reversal of the effect of CC on cAMP production was demonstrated. Treatment with pertussis toxin did not effect the basal levels of cAMP produced in response to CC treatment alone but increased the forskolin-stimulated cAMP, suggesting the relief of a tonic inhibition by the pertussis toxin incubation. Removal of extracellular calcium from the reaction mixture resulted in abrogation of 50% of forskolin-stimulated cAMP production. While this effect was observed independently of concomitant exposure to CC, it abrogated the CCsensitive component (Fig. 3B). These results suggest that the CC-mediated inhibition of cAMP production may be due to mAChR-stimulated influx of calcium and that there may be multiple isotypes of adenylyl cyclase activated by forskolin in these cells (described below, Fig. 8).
Lack of Stimulation of Inositol Phosphate Production-The ability of CC to stimulate production of inositol polyphosphates in the A2058 cells was approached with several different experimental designs. These included two different organic solvent separations and an inositol trisphosphate radioreceptor assay. As shown in Fig. 4A, no stimulation of inositol trisphosphate could be detected with the very sensitive radioreceptor assay in the face of a 2-fold stimulation in the CHOm5 control cells. Inositol monophosphate, diphosphate, or total inositol polyphosphates were not detectable using organic extraction and anion exchange chromatography methods in response to CC exposure of A2058 cells. Inositol trisphosphate production has been demonstrated previously in these cells in response to autocrine motility factor activation (23). Immunoblot was used to investigate the expression of phospholipase C isozymes in the A2058 cells. Fig. 4B shows the presence of both phospholipase C-␤1 and phospholipase C-␥ isozymes in lysates of A2058 cells. Since phospholipase C-␥ requires phosphorylation for activation, the ability of CC to stimulate tyrosine phosphorylation was tested with 5-and 15-min exposures to CC (Fig. 4C). Very minimal phosphorylation was observed at either time point, suggesting that this isotype is not a major downstream effector of the m5AChR in these cells.
Stimulation of Arachidonic Acid Release-CC stimulated a concentration-dependent release of arachidonic acid from the A2058 -1F5 cells (EC 50 ϭ 17.8 Ϯ 0.1 M, Fig. 5A). Arachidonic  , lanes 2 and 4). Southern analysis using an internal oligonucleotide probe selective for m2/4 is shown in the lower panel, confirming the lack of product in the A2058 cells. FIG. 2. Specific N-[methyl-3 H]scopolamine ( 3 H[NMS) binding to A2058 cell plasma membranes. Plasma membranes were prepared from A2058 cells. Saturation binding analysis was performed as described under "Experimental Procedures" using N-[methyl-3 H]scopolamine as the labeled ligand and 10 M atropine to define nonspecific binding. Data are mean Ϯ S.E., n ϭ 3. acid release was blocked by co-incubation with atropine, the muscarinic receptor antagonist (Fig. 5B). CC-stimulated arachidonic acid release was first observed within 2 min of agonist addition. A plateau was reached by 10 min and was sustained for 30 -60 min (Fig. 5A). All further arachidonic acid release experiments were allowed to proceed for 15 min.
Effect of Protein Kinase C Activation-A2058 cells were incubated with PMA for 15 min, 1, or 24 h to determine the effect of activation and down-regulation of protein kinase C on release of arachidonic acid (Fig. 6). Inclusion of PMA alone for 15 min resulted in a 200% increase in released arachidonic acid which rose to 400%, equivalent to that seen with CC treatment, at a 1-h exposure. An additive effect was observed when cells were exposed to PMA for 1 h followed by CC treatment, suggesting that arachidonic acid release could be stimulated in A2058 cells by protein kinase C, independently of muscarinic receptor-mediated activation. A 24-h exposure of A2058 cells to PMA to down-regulate cellular protein kinase C activity inhib-ited the CC-stimulated response by approximately 50% and abrogated the PMA-induced arachidonic acid release seen at 1 h, consistent with a requirement for protein kinase C in activation of this response.
Calcium Sensitivity of Arachidonic Acid Release-The mAChR-stimulated release of arachidonic acid from A2058 cells was partially dependent upon extracellular calcium as shown in Fig. 7. Inclusion of the heavy metal cobalt or the Ca 2ϩ chelator EGTA in the reaction media markedly reduced CCinduced arachidonic acid release (Fig. 7A). Mobilization of calcium through calcium influx can occur through voltage-dependent or independent channels. The voltage-sensitivity of the stimulated arachidonic acid release was further investigated using potassium-mediated cellular depolarization. Exposure of A2058 cells to high K ϩ media slightly decreased the release of arachidonic acid without altering the baseline release or ability of atropine to inhibit this response (data not shown). Addition

FIG. 3. Carbachol treatment inhibits forskolin-induced generation of cAMP.
A, inhibition of forskolin-stimulated cAMP accumulation and lack of pertussis toxin sensitivity. When A2058 cells were incubated with increasing concentrations of CC, no production of cAMP was measurable using a sensitive radioimmunoassay for cAMP (data not shown). CC produced a dose-dependent inhibition of forskolin-induced cAMP production. Overnight treatment of A2058 cells with pertussis toxin did not reverse the CC inhibition of forskolin-induced cAMP production but caused a ϳ50% increase in cAMP production (mean Ϯ S.E., n ϭ 3). Control, open circles; pertussis toxin, closed circles. B, removal of extracellular calcium abrogates the m5AChR-mediated inhibition. cAMP accumulation assays were done in the presence of 1 mM extracellular calcium or nominally calcium-free as indicated (mean Ϯ S.E., n ϭ 3). Control, open circles; no extracellular calcium, closed circles.

FIG. 4. Carbachol does not stimulate production of IP 3 from A2058 cells.
A, measurement of IP 3 . A2058 cells were tested for IP 3 production at various levels of confluence using organic extraction and anion exchange chromatography (data not shown) or using an IP 3 radioreceptor assay. CHOm5 cells were used as positive controls. Data are mean Ϯ S.E., n ϭ 3. B, demonstration of phospholipase C isozymes.  (Fig. 7B). Nifedipine (not shown) minimally inhibited arachidonic acid release at concentrations below 300 M. These data suggest that the calcium influx required for arachidonic acid release comes from mobilization of calcium through nonvoltage-gated calcium channels.
Effect of Carbachol on Intracellular Calcium Concentrations in A2058 Cells-A2058-1F5 cells were loaded with the calciumsensitive fluorescent dye, FURA-2, and changes in single cell fluorescence were measured over time (Fig. 8). CC stimulated a rapid increase in intracellular calcium that was attenuated with the addition of a muscarinic receptor antagonist, atropine (Fig. 8A). The relative contribution of extracellular calcium influx was evaluated in Fig. 8B, in which CC was added in the absence of extracellular calcium and induced a transient but rapid rise in intracellular calcium that decayed back to basal levels. After a brief wash period, CC re-exposure resulted in a small rise in intracellular calcium that quickly returned to control levels suggesting that intracellular calcium pools had been depleted almost completely during the first CC application (Fig. 8B). Re-application of calcium-containing media resulted in a large calcium influx that was sustained at a new plateau as long as CC was present. An additional experimental approach was selected to determine if the calcium influx component was down-regulated following a repeat application of CC (Fig. 8C). Intracellular calcium pools were depleted by the addition of CC in calcium-free medium. Sequential addition, removal, readdition, and removal of extracellular calcium in the perifusion buffer caused a concomitant rise and fall of intracellular calcium. At the end of the experiment, a small transient intracellular release of calcium was evoked by CC in calcium-free media, most likely due to partial refilling of cytoplasmic pools during the course of the calcium-containing buffer perifusion. CC induces both a release of calcium from cytoplasmic stores as well as calcium influx in A2058-1F5 cells.

CC Suppresses A2058 Cell Colony Formation in Soft Agar
We have shown previously that activation of the odd-numbered muscarinic receptor species in CHO cells specifically transfected with these receptors results in a morphologic and functional change to a nontumorigenic phenotype indicated by a shift from stellate to a more flattened and fibroblast-like shape (24). The demonstration of this receptor in a human melanoma cell line offered the opportunity to investigate the significance of these observations in an endogenously expressed receptor. When A2058 parental cells or the 1F5 subtype were cultured in monolayer with increasing concentrations of CC, no morphologic changes were detected, nor were changes in monolayer growth potential noted. However, when CC 0.1-100 M was included in the culture media of soft agar cultures, a statistically significant decrease in colony formation (p 2 Յ 0.001) and colony size in soft agar was demonstrated (Fig.  9). The reduction in clonogenic potential suggests a potential suppressor function of this receptor in the melanoma cells. DISCUSSION The discovery of an endogenous m5AChR in the A2058 human melanoma line provides the first demonstration of a novel pattern of m5AChR receptor-signal coupling as well as the demonstration of an endogenous m5AChR outside of the central nervous system. This receptor couples to release of the arachidonic acid and mobilization of intracellular calcium with-out measurable production of inositol phosphates or cAMP. Activation of this m5AChR results in inhibition of an adenylyl cyclase activity that is insensitive to pertussis toxin treatment. FIG. 8. Effect of carbachol on intracellular calcium release and calcium influx. A2058 cells were plated and allowed to adhere overnight on coated coverslips, washed free of serum, and loaded with the cell permeant FURA-2 AM, 1 M. Cells were perifused with medium containing 2 mM calcium or with calcium-free Eagle's number 2 medium containing 0.5 mM EGTA as indicated. Results are expressed as the ratio of 510 nm emission intensity measured following excitation at 350 nm/380 nm wavelengths. Data presented are representative traces from a single cell randomly selected from at least three experiments with at least 5 cells measured during each experiment. A, demonstration of internal calcium release followed by receptor-operated calcium influx. Cells were exposed to CC in calcium-containing media after which atropine, 10 M, was added. B, separation of intracellular release and calcium influx in response to CC. In calcium-free media, cells were exposed to CC, washed free of CC, and re-exposed to CC as indicated. Following emptying of the intracellular stores, cells were perifused with calcium-containing media and calcium influx is observed. C, desensitization of internal release in the absence of refilling of internal stores. Continuous exposure to CC in the absence and presence of calciumcontaining media is associated with desensitization of the internal release response to CC that required calcium influx to allow internal pool refilling in the absence of stimulus prior to second response to CC with internal release of calcium. FIG. 7. Calcium dependence of arachidonic acid release. Release experiments were done according "Experimental Procedures" and are expressed as mean Ϯ S.E., n Ն 3. A, effect of incubation with EGTA and cobalt. Competitive concentrations of EGTA or CoCl 2 were added to the incubation medium and arachidonic acid release measured in the absence of CC (control) or presence as indicated. B, incubation with inhibitors of voltage-gated calcium channels. A2058 cells were preincubated for 30 min with increasing concentrations of diltiazem or verapamil and then exposed to CC. The IC 50 concentrations of diltiazem and verapamil were 243 Ϯ 80 and 166 Ϯ 30 M, respectively. Cell viability was not altered by the short term incubation with any of these agents.
The pharmacologic demonstration of the receptor by saturation binding analysis shows it to be in low abundance and of high affinity for agonist. Molecular studies indicate the presence of a single species homologous to the known m5AChR coding sequence. These results argue that the A2058 human melanoma cells have a single mAChR of the m5 subtype with a new pattern of receptor signal transduction effector coupling. Activation of this endogenous receptor inhibited the anchorageindependent growth potential of the A2058 cells. The m5AChR previously has only been demonstrated in the central nervous system (1,2,19,20). Skin melanomas are of neural crest origin and contain neural markers, but are different from retinal melanomas which are of central nervous system origin. This is the first demonstration of an endogenous m5AChR outside of the central nervous system.
The novel signaling pattern of the A2058 cell mAChR could be interpreted to suggest the presence of two molecular subtypes of the receptor or a normal receptor with altered signal coupling. The possibility of two endogenous receptor species was investigated by binding and molecular studies. The low abundance of receptor protein prevented reliable pharmacologic characterization of multiple receptor subtypes by radioligand binding. Molecular studies using PCR amplification demonstrated the presence of a single mAChR subtype on the A2058 cells. Sequence analysis of PCR fragments covering the coding sequence revealed complete homology with the published human m5 region (19). This was confirmed by specific hybridization under high stringency with a separate PCR fragment amplified from the m5-selective third cytoplasmic domain.
Classical muscarinic receptor signaling has yielded a receptor-specific pattern for the odd-and even-numbered receptors as shown in Table I. Odd-numbered mAChR couple to activation of phospholipase A2, phospholipases C-␤ and C-␥1, adenylyl cyclase, and activation of calcium internal release and influx (5, 8 -10, 12-15, 17, 30 -33). Coupling of even-numbered receptors results in the inhibition of cAMP generation and augmentation of the release of arachidonic acid initially stimulated through receptor cross-talk with a non-muscarinic receptor (16). The pattern of A2058 mAChR signaling in response to CC did not fit this pattern of receptor-signal coupling.
The observed inhibition of production of cAMP by activation of the A2058 m5AChR, instead of the expected stimulation, would not have been predicted on the basis of the receptor subtype. The lack of an even-numbered mAChR or novel muscarinic acetylcholine receptor subtype suggests that the altered signal coupling is a cell-specific process. Mechanisms to explain this different signaling pattern include a different effector protein complement or altered G protein complement in these cells. Even-numbered mAChRs preferentially couple to adenylyl cyclase inhibition through G o and G i , whereas the oddnumbered mAChRs use the G q and G 11 families to couple to phospholipase C-␤ (17,31). The A2058 cells have been shown to have G s , G o , and G ia2 2 and have been shown to couple to inositol trisphosphate production in a pertussis toxin-sensitive fashion (23). The lack of pertussis toxin sensitivity of the inhibition of adenylyl cyclase by receptor activation in the A2058 cells further differentiates the receptor-signal coupling of this endogenous mAChR. The increased cAMP produced with pertussis toxin exposure suggests that there is an adenylyl cyclase for which a tonic inhibition is relieved by pertussis toxin, releasing adenylyl cyclase activity. In a different study, stimulation of m1 receptors expressed at physiologic levels in RAT-1 cells was shown to inhibit adenylyl cyclase activity; however, that inhibition was pertussis toxin-sensitive (15). The reverse phenomenon was observed in which overexpressed m4 receptors shifted from inhibition to stimulation of adenylyl cyclase by pretreatment with pertussis toxin and blockade of G i activity (34). In both cases, the aberrant coupling to adenylyl cyclase was pertussis toxin-sensitive. Site-directed mutagenesis has identified regions of the third cytoplasmic domains of the m2 and m3 receptors which regulate the pertussis toxin sensitivity and are involved in G protein interactions (35). The inhibition of cyclase activity in A2058 cells after activation of the m5AChR was not abrogated by pertussis toxin exposure, despite availability of the requisite G protein complement and lack of mutation in the critical regions of the third cytoplasmic domain.
The partial inhibition of forskolin-stimulated adenylyl cyclase activation by removal of extracellular calcium suggests that there are more than one isozyme of adenylyl cyclase isozyme in the A2058 cells. Adenylyl cyclase isozymes V and VI are characterized by inhibition in function as a result of increase in intracellular calcium concentration (36). A pertussis toxin-independent, calcium-dependent inhibition of substance K receptor-mediated inhibition of cAMP produced in response to isoproterenol was reported in C6-2B glioma cells (37). A similar calcium-dependent inhibition of cAMP accumulation was observed using activation of the bradykinin receptor in NCB-20 plasma membranes (38). Our m5AChR-induced inhibition of adenylyl cyclase activity in A2058 cells is calcium influx-sensitive, suggesting that receptor-mediated calcium entry may be the driving inhibitory force.
Another point of disparity between classical receptor-signal coupling of the mAChR and this m5AChR is the activation of 2 Dr. S. Aznavoorian, personal communication.  the phospholipases C to release inositol trisphosphate. The odd-numbered muscarinic receptors have been shown to couple through G proteins to phospholipase C-␤ (9,13,15,17,35,39). In addition, activation of phospholipase C-␥ in CHOm5 cells has been demonstrated through muscarinic receptor-mediated and calcium-sensitive phosphorylation (12). A previous report also described an even-numbered mAChR coupling to phosphoinositide turnover in a pertussis toxin-sensitive pathway (18). Both organic separation and anion chromatography, and an inositol trisphosphate radioreceptor assay failed to demonstrate a rise in inositol phosphates in response to CC treatment of A2058 cells where CHOm5 cell controls were positive. It is possible that the production of inositol trisphosphate was below the level of detection of these assay systems; however, this is unlikely as the measurement was approached in three independent ways and the radioreceptor assay has picogram sensitivity. Western immunoblot demonstrated the presence of phospholipase C-␤1 and phospholipase C-␥ isozymes in A2058 cells. The relative lack of phosphorylation of phospholipase C-␥ after exposure to CC for 5 or 15 min argues for it not being coupled to the m5AChR in these cells. This further underscores the different coupling effector system in these cells, as activation of phospholipase C-␥ in CHO cells stably transfected with the m5AChR has been demonstrated (12). The observed release of internal calcium under calcium-free conditions argues for the presence of inositol trisphosphate or alternative internal calcium-mobilizing, inositol trisphosphate-independent mechanisms. The A2058 m5AChR coupled to release of arachidonic acid and mobilization of calcium influx. Arachidonic acid release proceeded with an EC 50 similar to that previously described for the transfected m5 receptor (10,24,32). The requirement for extracellular calcium was demonstrated by the inhibition of the CC response by inclusion of EGTA in the reaction buffer and increase in CC-stimulated arachidonic acid release in the presence of calcium. The molecular mechanism for calcium in activation of phospholipase A2 has been identified (30, 40 -42). Calcium is involved in cPLA2 translocation, phosphorylation, and activation by mitogen-activated protein kinase. mAChR have been shown to regulate influx of calcium through nonvoltage-gated calcium channels, specifically, the receptor-operated channels (10 -12, 14, 17, 30, 43). The A2058 m5AChR stimulates receptor-operated calcium influx as shown by the ability of atropine to block influx. Last, the A2058 m5AChR follows previously described patterns in arachidonic acid release in its sensitivity to protein kinase C (10,33). Activation of protein kinase C through phorbol ester exposure results in activation of arachidonic acid release itself and augments the CC-stimulated effects in the A2058 cells.
Activation of the A2058 mAChR with CC resulted in a receptor-specific inhibition in A2058 cell cloning capacity in soft agar assays. Anchorage-independent growth in culture is a marker for tumorigenicity. We have previously demonstrated the phenomenon of receptor-mediated anti-oncogenesis using CHOm5 cells (24). Exposure of CHOm5 cells to CC in soft agar cultures also resulted in inhibition of anchorage-independent growth. The reduction in tumorigenicity in CC-activated CHOm5 cells was confirmed further in nude mouse xenograft experiments. Treatment of the inoculation site with CC markedly reduced the tumor incidence and those tumors that did initiate had an extremely low growth rate. Inclusion of atropine reversed the effects of CC yielding a full tumor incidence and exponential tumor growth. Similarly, an anti-proliferative effect of CC-activation of transfected m1 receptors has been documented in A9 cells (8,44). The parallel results of reduction in soft agar colony formation with activation of the m5 receptor in the A2058 cells suggests that the A2058 m5AChR may have an anti-oncogenic function.
The classical muscarinic receptor signal transduction coupling has followed receptor-specific patterns as described in Table I. These signaling patterns have been confirmed using endogenous systems and transfected receptor models and have yielded identical results. To date, there has been no endogenous m5AChR available to confirm the signaling pattern previously described using stable transfectants (10,11,17). This endogenous A2058 melanoma cell m5AChR is a low abundance, high affinity receptor with novel signal effector coupling. The unique signal coupling demonstrates the importance of the host cell in establishing permissive downstream effector signal transduction and offers new insight into the biology and biochemistry of the muscarinic receptor system.