Specific Uncoupling by Islet-activating Protein, Pertussis Toxin, of Negative Signal Transduction via a-Adrenergic, Cholinergic, and Opiate Receptors in Neuroblastoma X Glioma Hybrid Cells*

Exposure of NGlO8-15 hybrid cells to islet-activating protein (IAP), pertussis toxin, caused strong ADP-ri-bosylation of one of the membrane proteins with a molecular weight of 41,000. This ADP-ribosylation was paralleled by decreases in the inhibition of CAMP ac- cumulation in intact cells or associated with reversal of the inhibition of GTP-dependent membrane adenylate cyclase, via a-adrenergic, cholinergic muscarinic, or opiate receptors. The affinity of these receptors for agonists was lowered by guanyl-5’-yl P-y-imidodiphos-phate (Gpp(NH)p) reflecting their coupling to the gua- nine nucleotide regulatory protein in this cell line. This effect of Gpp(NH)p was lost in membranes of IAP- treated cells; in the absence of Gpp(NH)p, the affinity for agonist was lower in treated than in nontreated cells. In contrast, the function of these receptors to bind antagonists remained unaltered in IAP-treated cells. Thus, IAP treatment of NGlO8-15 cells caused specific uncoupling of negative signal transduction from inhib- itory receptors to the adenylate cyclase catalytic unit via the guanine nucleotide regulatory protein, as a result of ADP-ribosylation of one of the subunits of the regulatory protein.

tors is known to communicate with the catalytic unit of adenylate cyclase conversely in an inhibitory fashion (6). The regulatory protein (Ni) involved in such inhibition of the cyclase is conceivably an entity distinct from N, (l), although much remains to be solved as to how Ni inhibits adenylate cyclase. An agent which is capable of specific modification of Ni function, if available, would contribute greatly to solving the problem in this regard, just as cholera toxin has done for NS.
Islet-activating protein is one of the candidates for such promising agents. IAP2 is an exotoxin produced by Bordetella pertussis bacteria (7,8) and is, like cholera toxin, an oligomeric protein with the A-B structure (9) in which a B-oligomer binds to particular receptors on the surface of the target cell to enable an A-protomer to be inserted gradually into the membrane of the cell, where this active component of the toxin catalyzes transfer of the ADP-ribose moiety of cytosolic NAD to a specific membrane protein3 (9)(10)(11). The definite lag time invariably observed preceding the onset of IAP action on intact cells is accounted for by this gradual access of the active component to the site of its action (12); there was no lag upon direct addition of IAP (or its A-protomer) to the broken cell preparation3 (10,11). IAP was effective in causing characteristic loss of certain receptor-mediated cAMP responses of cells. Stimulation of a*-adrenergic (13) receptors in rat islets (14,15), as well as cholinergic muscarinic and adenosine A, receptors in rat heart cells (16), failed to cause -significant changes in the cellular cAMP content when these cells had been kept in contact with IAP for several hours, despite marked decreases in cAMP via these receptors in the cells not treated with the toxin.
This effect of IAP was c o n f i e d by later studies showing that receptor-mediated inhibition of membrane adenylate cyclase was reversed by IAP treatment of rat islet cells (17) or adipocytes (18) from which membranes were prepared. Since the IAP-induced reversal of the inhibition was associated with IAP-catalyzed ADP-ribosylation of a membrane protein with a molecular weight value of 41,000, a guanine nucleotide binding protein distinct from the cholera toxin substrate (10,II), it was proposed that IAP interferes with functions of Ni as a result of ADP-ribosylation of one of its constituent subunits (18).
Mouse neuroblastoma X rat glioma hybrid cells are considered to afford a cellular or membrane system that is suitable for further approaches to possible modification by IAP of Ni; there are three kinds of receptors, az-adrenergic, cholinergic, and opiate, in this cell line which transduce negative signals, via the GTP regulatory component, to the adenylate cyclase catalytic unit (19)(20)(21). As will be shown in the present communication, LAP uncouples the Ni-mediated communication between these inhibitory receptors and the cyclase catalytic unit in this cell line, as revealed by abolition of receptormediated, GTP-dependent inhibition of adenylate cyclase as well as Gpp(NH)p-insensitive reduction in affinities of the receptors for agonists.

EXPERIMENTAL PROCEDURES
Materials-IAP was purified from the 2-day culture supernatant of Bordetella pertussis (Tohama strain, Phase I) according to the procedure described elsewhere (7). Ro- 20- Growth of Cells and Preparation of Cell Suspension-Stock cultures of mouse neuroblastoma X rat glioma hybrid cells of the NG108-15 clone were grown in Falcon dishes in Dulbecco's modified minimum essential medium, 10% fetal calf serum supplemented with 0.1 mM hypoxanthine, 1 p~ aminopterin, 16 p~ thymidine, 100 units/ml of penicillin G, and 0.1 mg/ml of streptomycin under a humidified atmosphere of 95% air, 5% COZ at 37 "C (21). After reaching confluence, the cells were further cultured in the fresh medium fortified with IAP or the vehicle used for dissolving it (22) for 3 h (unless otherwise specified in Figs. 1 and 2 A ) to prepare IAP-treated or nontreated cells, respectively. Cholera toxin treatment of cells was performed similarly. The cells were then harvested by replacing the medium with the solution consisting of 137 mM NaCl, 5.6 mM glucose, 1 mM EGTA, and 5 mM Hepes (pH 7.4). Cells were washed with the same solution further supplemented with 1 mM MgCIZ and 3 mM CaCL by centrifugation at 100 X g for 2 min at 4 "C. The cell pellet was suspended in this buffer solution and incubated for 30 min at 37 "C, unless otherwise specified, before the following experiments. CAMP Responses of Cell Suspension-The cell suspension thus prepared was incubated for 15 min at 37 "C with additions including receptor agonists and inhibitors of cyclic nucleotide phosphodiesterase as described in the table and figure legends. The incubation was terminated by the addition of HC1 at the final concentration of 0.1 N, followed by immersing the incubation tube into a boiling water bath for 2 min. The cellular cAMP quantitatively transferred to the supernatant during this procedure (14) was determined by a sensitive and specific radioimmunoassay method (23).
Membrane Preparation and Assay for Its Adenylate Cyclase Activity-The cell pellet was washed, suspended, and stored for 10 min in an ice cold solution of 25 mM Tris (pH 7.5), 5 mM MgC12, and 1 mM EGTA. This suspension was then homogenized in a Teflonpestled Potter-Elvehjem glass homogenizer. The homogenate, after being freed of undisrupted cells and nuclei by a 3-min centrifugation at 1 0 0 X g at 4 "C, was centrifuged at 4000 X g for 10 min to obtain the membrane-rich fraction which was stored in liquid nitrogen until use. This preparation was washed once with the solution of 50 mM Tris (pH 7.5), 5 mM MgC12, 1 mM EDTA immediately before use.
The reaction mixture (total volume, 100 pl) for adenylate cyclase assay was 50 ~l l~ Tris (pH 7.5) supplemented with 5 mM MgC12, 1 EGTA, 0.5 mM 3-isobutyl-l-methylxanthine, 1 mg/ml of bovine Serum albumin, 0.5 m~ ATP, 5 mM phosphocreatine, and 50 units/ml of creatine phosphokinase. Further additions are shown in the tables and figures. The reaction was initiated by the addition of the membrane preparation (10-20 p g of protein), continued for 5 min at 37 "C, and stopped by the addition of 0.33 M ZnSOl and 0.5 M Na~C03. Cyclic AMP in the deproteinized supernatant was then assayed by the radioimmunochemical method (23).
Binding of Receptor Agonists or Antagonists to the Membrane Preparation-Membrane preparations (60-170 pg of protein) were incubated for 15 min at 37 "C in the reaction mixture of 50 mM Tris (pH 7.5), 5 m~ MgCIZ, 1 mM EGTA, and 1 mg/ml of bovine serum albumin in a total volume of 1 0 0 pl. The reaction mixture was further supplemented with increasing concentrations of radioactive arltagonists (with constant specific radioactivities) for measurement of kinetic parameters for their binding, whereas it was supplemented with 4 nM [3H]dihydroergocryptine (or 2 nM [3H]quinuclidinyl benzilate) and increasing concentrations of norepinephrine (or carbachol) for measurement of the agonist binding. Gpp(NH)p (100 p~) was added where indicated. The reaction was stopped by the addition of 1 ml of ice-cold 50 mM Tris (pH 7.5), 5 m~ MgC12; it was followed immediately by vacuum filtration through a Whatman GF/C glass fiber filter (24mm diameter). The filter was rapidly washed twice with the ice-cold Tris/MgCL solution (5 ml) and dried at 80 "C for 45 min before being counted for radioactivity in Triton/toluene scintillation fluid. Specific binding was defined as the total binding minus the nonspecific binding not displaced by 10 p~ nonradioactive phentolamine, atropine, or naloxone. The maximum specific binding, which was 50-60, 80-90, and 50-60% of the total binding for [3H]dihydroergocryptine, ['HI quinuclidinyl benzilate, and [3H]naloxone, respectively, was achieved within 5 min. Protein was determined by the method of Lowry et al. (24) with bovine serum albumin as standard. Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis of ADP-ribosylated Membrane Proteins-Membranes (100-200 pg of protein) were incubated with [c~-~'P]NAD in the presence or absence of 25 pg/ml of IAP in the reaction mixture consisting of 50 m~ Tris (pH 7.5),5 mM MgCL, 1 m~ EGTA, 1 mM ATP, and 20 mM thymidine. The radiolabeled membranes were washed twice with 50 mM Tris, 5 mM MgC12, 1 mM EGTA, dissolved in 1% sodium dodecyl sulfate by heating at 1 0 0 "C for 3 min, and submitted to polyacrylamide slab gel electrophoresis as described previously (10,11). The gel was then autoradiographed at -80 "C for 2-7 days using Kodak X-Omat AR film (IO, 11). (25), while the incubation with norepinephrine (25,26), carbachol (a cholinergic agent) (27), or morphine (28, 29) resulted in significant decreases, in the cellular cAMP content in c o n f i a t i o n of the earlier reports ( Table I). The per cent decreases of cAMP elicited by these receptor agonists were essentially the same  ( 8 8 ) .

Cholinergic, or Opiate Receptors and Its Reversal by IAP Pretreatment in NG108-15 Cells-Incubation of NG108-15 cells with PGEl caused tremendous increases
, Data in parentheses show percentages of the "None" value.
in either the presence or absence of PGE,. The effect of norepinephrine was mimicked by az-adrenergic agonists and efficiently blocked by an-antagonists in comparison with alagonists and a,-antagonists, respectively, whereas the effect of carbachol was blocked by atropine (data not shown), in agreement with previous reports (21,27). Thus, NG108-15 cells possess an-adrenergic, muscarinic, and opiate receptors all of which are "inhibitory receptors" in a sense that stimulation of these receptors by their specifk agonists leads to inhibition of cAMP generation. When cells that had been exposed to IAP for 3 h were incubated similarly, none of the agonists of these three kinds of inhibitory receptors suppressed the cAMP generation during incubation, regardless of whether or not the generation was simultaneously stimulated by PGE, (Table I). Thus, receptor-mediated decreases in cellular cAMP were reversed by TAP treatment of the cell as had been observed with rat islet (14, 15) and heart (16) cells. An additional effect of IAP treatment was potentiation of PGE, action; PGEl-induced increases in cAMP attained 100-fold in the IAP-treated cells as compared with a 70-fold increase in nontreated cells (Table   I). Increases in cAMP via glucagon receptors in rat islet cells (14) and ,&adrenergic receptors in rat heart (16) and C6 glioma (22) cells have been reported to be likewise enhanced by IAP treatment of these cells. Despite such unique modulation of receptor-linked changes in the cAMP content, its basal value obtained before or after incubation without any receptor agonist was not affected by IAP. This is again the same as previous findings with other cell types (14-16, 22).
The cAMP data of cholera toxin-treated cells are also recorded in Table I for comparison with those of IAP-treated cells. More cAMP was present in cells after culture with cholera toxin and its spontaneous increase during incubation without receptor agonists was enormous as compared with that in nontreated cells. This high basal rate of cAMP production in cholera toxin-treated cells was further increased by PGE, only slightly; the increment of about 600 pmol/mg of protein was much smaller than the 3500-pmol increment elicited in nontreated cells. The most striking difference from IAP-treated cells was found in that the function of inhibitory receptors to suppress the cAMP production was retained unaltered in cholera toxin-treated cells. When IAP was superimposed upon cholera toxin treatment, it was still effective in not only mostly abolishing the suppression of cAMP generation via inhibitory receptors but also slightly enhancing the stimulatory action of PGE1. Cholera toxin-induced increases in the spontaneous formation of cAMP during incubation of cells was also exaggerated by superimposed IAP.
Kinetics of IAP-induced Reversal of Cellular cAMP Decreases via a-Adrenergic, Cholinergic, or Opiate Receptors- Fig. 1 shows concentration-dependent effects of norepinephrine, carbachol, or morphine in decreasing the cAMP content of IAP-treated and nontreated cells. In this experiment, the submaximal dose of IAP was chosen such that the IAP treatment only partially interfered with the suppression of cAMP generation via inhibitory receptors. It enabled us to estimate kinetics of cAMP changes even in IAP-treated cells. The decreases in cellular cAMP induced by norepinephrine, carbachol, or morphine were much smaller in IAP-treated cells than in nontreated cells in similar degrees at almost all the effective concentrations of these agonists (Fig. 1). Thus, the maximum effects of the agonists were reduced by IAP treatment without significant changes in their half-maximally effective concentrations; IAP was not competitive with inhibitory receptor agonists.
Interference by IAP with inhibitory receptors was a slowly developing process (Fig. 2 A ) . The IAP-induced interference was only partial after a 1-h exposure but it became complete at any time longer than 3 h, when the concentration of IAP was 100 ng/ml. Upon a 3-h exposure, the total recovery from receptor-mediated inhibition was observed at 100 ng/ml of IAP with its concentration to cause the half-maximal recovery being between 1 and 10 n g / d (Fig. 2B). Potency of IAP was dependent on exposure time (Fig. 2C); when cells had been rendered to be in contact with IAP for 8 h, the maximal and half-maximal effects were obtained with 2 and 0.1 ng/ml of IAP, respectively. Shortening the exposure time to 2 h reduced the potency of IAP so markedly that its half-maximally effective dose was as high as 10 ng/ml. A similar relation of potency of IAP to exposure time has been observed as to the enhancing effect of IAP on P-adrenergic accumulation of cAMP in rat C6 glioma cells (22).

Reversal of Receptor-mediated Inhibition of Membrane Adenylate Cyclase by Prior Treatment of Cells with IAP-
Norepinephrine, carbachol, or morphine directly added to membranes caused significant inhibition of adenylate cyclase activity (Table 11). The per cent inhibition via these inhibitory receptors in isolated membranes was, at the best, 30-4076, much smaller than the per cent inhibition of cAMP accumulation via the same receptors in intact cells which usually attained to 80-90% at its maximal value (see Figs. 1 and 2). Such has been the case with the inhibitory receptors in this cell line so far reported (21, 30-32).
Essentially no inhibition of the cyclase was evoked by these inhibitory receptor agonists in membranes from IAP-treated cells (Table 11): Thus, receptor-mediated inhibition of adenylate cyclase was overcome by prior treatment of IAP; the effect of IAP in this regard was highly significant in a statistical sense. In these experiments shown in Table 11, membrane adenylate cyclase was assayed in the presence of 1 p~ GTP, because GTP was an essential factor in supporting receptormediated inhibition of the cyclase (data not shown) and because the cyclase activity assayed with 10 or 100 p~ GTP in the absence of receptor agonists was much higher in membranes from IAP-treated cells than in those from nontreated cells (data not shown), obscuring the difference in sensitivity to inhibitory agonists between these two membrane preparations. In fact, however, the effect of IAP in overcoming the receptor-mediated inhibition of adenylate cyclase was evident at higher GTP concentrations as well (data not shown).

Failure of IAP Treatment to Alter the Numbers of Inhibitory Receptors and Their Affinities for Respective Antago-
nists-Specific binding of [3H]dihydroergocryptine, [3H]quinuclidinyl binzilate, and [3H]naloxone to membranes from IAPtreated or nontreated cells was measured in the presence of increasing concentrations of these antagonists to afford Scatchard plots (not shown). The numbers of binding sites as well as the dissociation constants for binding were then estimated as recorded in Table 111. There was no difference in these kinetic parameters between IAP-treated and nontreated cells. Thus, neither the density in membranes nor affinities for antagonists were affected by IAP treatment as regards aadrenergic, cholinergic, and opiate receptors. Neither GTP nor Gpp(NH)p was effective at all on these antagonist binding to membranes (data not shown).

Reduced Affinities ofAgonist Binding to &-Adrenergic and Cholinergic Receptors and Loss of Gpp(NH)p Sensitivity in Membranes from IAP-treated Cells-Affiiities of a-adrenergic and cholinergic receptors for their agonists were measured based on displacement of [3H]dihydroergocryptine and [3H]
quinuclidinyl benzilate bound to membranes by norepinephrine and carbachol, respectively (Fig. 3). In membranes from cells not treated with IAP, both displacement curves were shifted to the right about 10-fold in the presence of Gpp(NH)p (Fig. 3, A and 0

. Occupancy of GTP binding sites on Ni may
Cholinergic or opiate receptor-mediated inhibition of adenylate cyclase was not totally abolished by the maximally effective concentration (100 ng/ml, see Fig. 2) of IAP in Table 11. But, we have recently found that the inhibition was completely overcome by IAP or its A-protomer (9) directly added to membranes in the presence of NAD and ATP (H. Kurose and M. Ui, manuscript submitted for publication).

Inhibition of membrane adenylate cyclase by norepinephrine, carbachol, or morphine and its reversal by IAP treatment of cells
Membranes were prepared from cells that had been treated with or without 100 ng/ml of IAP. Adenylate cyclase activity of the membranes was assayed with 1 p~ GTP in the presence of 100 p~ norepinephrine, carbachol, or morphine.  < 0.001). lower the affinity of coupled receptors for agonists, as has been reported for a-adrenergic receptors in this cell line (33), platelet (34,35), brain (36), and adipocytes (37), as well as for cholinergic receptors in heart (38-43), brain (44, 45), and smooth muscle (46). In the case of membranes from IAPtreated cells, the affinities for agonist binding to the receptors were lower even in the absence of Gpp(NH)p; they were roughly equal to the affinities observed in the presence of Gpp(NH)p in nontreated cells' membranes (Fig. 3, B and D ) . There was no further reduction of affinities upon addition of the guanosine triphosphate. Thus, the guanine nucleotide binding protein does not appear to be coupled to adrenergic and cholinergic receptors in membranes of IAP-treated cells.

ADP-ribosylation of Membrane Proteins by IAP as Correlated with Its Action to Overcome Receptor-mediated Inhibition of CAMP accumulation-When membranes from
normal cells were incubated with [~u-~'P]NAD, 32P was incorporated into an M, = 41,000 protein band in the presence of IAP as a result of ADP-ribosylation of this membrane protein as evidenced with rat glioma C6 cells (10,11). In Fig. 4, intact NG108-15 cells had been allowed to be in contact with increasing concentrations of IAP. Concentration-dependent reversal of norepinephrine-induced inhibition of cAMP accumulation occurred during subsequent incubation of a batch of these cells (Fig. 4, open circles). Another batch of the IAP-treated cells was disrupted to provide the membrane fraction, which was incubated with [LU-~*P]NAD in the presence of a saturating concentration of IAP and then analyzed for the 32P content in the M, =  nephrine inhibition of cAMP generation (Fig. 4, solid circles), leading to a conclusion that reduction of receptor-mediated inhibition of cAMP generation observed in IAP-treated cells is proportional to the degree of ADP-ribosylation of the M , = 41,000 protein during the IAP treatment of the cells.

DISCUSSION
Mouse neuroblastoma X rat glioma hybrid NG108-15 cells are endowed with devices for transduction of negative signals from three kinds of receptors, cyz-adrenergic, cholinergic muscarinic, and opiate, to adenylate cyclase via the guanine nucleotide regulatory protein (Ni). Stimulation of these receptors by specific agonists caused decreases in the cAMP content in intact cells (Table I and Fig. l), probably reflecting the receptor-mediated inhibition of GTP-dependent adenylate cyclase actually observable with the membrane preparation ( Table  11). The ability of these receptors to inhibit adenylate cyclase, or to lower the cellular cAMP content, was prevented by prior exposure of cells to IAP for several hours (Tables I and 11, Figs. 1 and 2). Similar prevention by IAP of receptor-mediated inhibition of adenylate cyclase or decreases in cellular cAMP has been reported for a-adrenergic receptors in rat islet cells (14,15,17), cholinergic muscarinic and adenosine A, receptors in rat heart cells (16), and adenosine AI and PG receptors in rat adipocytes (18). It is unlikely that IAP interacts directly with such a variety of different receptors. Instead, IAP appears to interact with the common Ni protein in such a manner as to uncouple adenylate cyclase from these inhibitory receptors (18) as will be discussed below.
Evidence was provided in Fig. 3 for uncoupling of cy-adrenergic and muscarinic receptors from the guanine nucleotide binding protein in membranes of IAP-treated cells. The affinity of these receptors for agonist was lowered and rendered insensitive to Gpp(NH)p by IAP treatment of cells. A number of reports have shown that agonists are bound to adenylate cyclase-linked receptors with lower affinities in the presence of GTP (or its nonhydrolyzable analog) than in its absence. This effect of guanosine triphosphate is considered to reflect coupling of the receptors to the guanine nucleotide binding protein, as fist observed for stimulatory receptors such as padrenergic receptors in frog erythrocytes (47) and cloned cell lines (48). The same holds true for inhibitory a-adrenergic and muscarinic receptors in various cell types (36)(37)(38)(39)(40)(41)(42)(43)(44)(45)(46). The cycvariants of S49 lymphoma cells are known to lack N, (49). Comparison of agonist binding between membranes of the c y -and its parental (wild type) S49 line, just like the similar comparison between IAP-treated and nontreated NG108-15 line (Fig. 3), revealed that the affinity of the receptors for agonists was lower in the former than in the latter, with the affinity lowering effect of Gpp(NH)p observable in the latter but not in the former (50). Thus, the guanine nucleotide regulatory protein (Ni) is not coupled to the inhibitory receptor in membranes of IAP-treated NG108-15 cells just as was it the case with cyc-membranes which functionally lack N,.
Exposure of intact cells to IAP resulted in ADP-ribosylation of the membrane M, = 41,000 protein by endogenous NAD.
The ADP-ribosylation was dependent on the concentration of IAP and changed in parallel with the degree of the IAP action to overcome receptor-mediated decreases in the cellular cAMP (Fig. 4). Thus, IAP-catalyzed ADP-ribosylation is likely to be responsible for IAP-induced reversal of receptormediated inhibition of adenylate cyclase, in support of our previous conclusion that the M, = 41,000 protein ADP-ribosylated by IAP would be one of the subunits of the Ni protein (18). The N protein possesses bidirectional functions. The first is directed to the catalytic unit of adenylate cyclase in such a fashion as to activate (N,) or inhibit (Ni) it upon occupancy of the GTP binding site, whereas the second function is directed to coupled receptors the affinity of which for agonists is reduced. The second function of N, was impaired by IAP as discussed above. The inhibition of 3T3 membrane adenylate cyclase caused by Gpp(NH)p binding to N, under certain conditions has been reported to be abolished by lAP treatment ofthe cells (51), indicating that lAP interferes with the first function of N, to inhibit adenylate cyclase in the presence of guanine nucleotides. Thus, our present and previous results, taken together, led us to conclude that both (probably all) of the physiological functions of N, are lost by lAP-catalyzed ADP-ribosylation of one of its constituent subunits, the Mr = 41,000 peptide. In addition to the inhibitory receptors, there are receptors for PGE I which are coupled to adenylate cyclase in a stimulatory fashion in NG108-15 cells. Increases in cellular cAMP via PGE I receptors were larger in lAP-treated cells than in nontreated cells (Table I). Similar potentiation of receptormediated cAMP accumulation by lAP treatment has been reported for ,8-adrenergic receptors in rat heart cells (16) and C6 glioma cells (22) and for glucagon receptors in rat islets (14). This may reflect lAP-induced enhancement of receptormediated activation of adenylate cyclase, as evidenced for ,8adrenergic receptors in C6 glioma cells (22) and 3T3 fibroblasts (51). GTP-dependent adenylate cyclase was also enhanced by lAP treatment of NG108-15 cells in agreement with previous findings with C6 and 3T3 cells (22,51). It is very likely, as has recently been argued (18), that lAP-induced abolition of the N, function is also responsible for enhancement of these adenylate cyclase activations conceivably due to a shift of the Ns/N, balance which would be an important factor in determining GTP-dependent adenylate cyclase activity.
In sharp contrast with lAP-treated cells, NG108-15 cells treated with cholera toxin were as susceptible to agonists of inhibitory receptors as were nontreated cells (Table I). Such was the case with rat heart cells (16). It is evident, therefore, that cholera toxin does not interact with Ni, modification of which appears to be monopolized by lAP. An additional difference between cholera toxin-treated and lAP-treated cells was found in the basal cAMP level before or after incubation of the cells in the absence of any agonist; it was extraordinarily higher in cholera toxin-treated cells, but never so in IAPtreated cells, than in nontreated cells (Table I). Similar results have been reported for heart (16) and C6 (22) cells. Accumulation of cAMP under these conditions in cholera toxin-treated cells would have arisen from the high basal adenylate cyclase activity assayed in the absence of GTP or receptor agonists (22,51) and this high basal activity, in tum, may be accounted for by interaction of cholera toxin with N, (5). Cholera toxininduced accumulation of cAMP in the basal (nonstimulated) state was further enhanced by superimposed lAP treatment, probably due again to a shift ofthe N./N i balance in favor of N. as discussed above (18).
In summary, lAP, pertussis toxin, interfered with physiological functions of N, as a result of specific ADP-ribosylation of the M, = 41,000 protein, one of its subunits, in NG108-15 cells, thereby preventing the negative signal transduction from inhibitory receptors to the adenylate cyclase catalytic unit.