A Reassessment of Guanine Nucleotide Effects on Catecholamine Secretion from Permeabilized Adrenal Chromaffin Cells*

The role of guanine nucleotides in catecholamine secretion was investigated in a-toxin-permeabilized chromaffin cells. The stable GTP analogues, GTP-7-S (guanosine 5’-(y-thio)triphosphate) and GMP-PNP (guanosine 5’-(B,y-imido)triphosphate), potentiated calcium-evoked catecholamine release in a dose-de-pendent manner. This effect was reversed by GDP-B- S (guanosine 5’-(&thio)d1phosphate) indicating that a GTP-binding protein plays a modulatory role in the calcium-dependent secretory process in chromaffin cells. Calcium and the phosphorylating nucleotide ATP were both necessary for secretion, even in the presence of GTP analogues, suggesting that the activation of a GTP-regulatory protein alone does not trigger exocytosis in these cells. TPA (12-0-tetradecanoylphorbol- 13-acetate), a direct activator of protein kinase C, was found to mimic the effects of the GTP analogues, in- ducing a dose-dependent potentiation of the calcium-evoked release in a-toxin-permeabilized cells. Treat- ment of the concentration on the secretory response to GTP-y-S and TPA Chromaffin cells labeled with [3H]noradrenaline were permeabilized for 30 min with a-toxin in calcium-free KG medium. Cells were then stimulated for 10 min with KG medium containing either 5 or 15 mM NTA and the appropriate amount of CaC12 to give the indicated free calcium concentration, in the presence or absence of 200 nM TPA o:r 100 pM GTP-y-S. Increasing NTA concentration does not alter the ability of TPA and GTP-y-S to enhance the release of catecholamines when the calcium concentration is buffered at 20 UM bv NTA.

The role of guanine nucleotides in catecholamine secretion was investigated in a-toxin-permeabilized chromaffin cells. The stable GTP analogues, GTP-7-S (guanosine 5'-(y-thio)triphosphate) and GMP-PNP (guanosine 5'-(B,y-imido)triphosphate), potentiated calcium-evoked catecholamine release in a dose-dependent manner. This effect was reversed by GDP-B-S (guanosine 5'-(&thio)d1phosphate) indicating that a GTP-binding protein plays a modulatory role in the calcium-dependent secretory process in chromaffin cells. Calcium and the phosphorylating nucleotide ATP were both necessary for secretion, even in the presence of GTP analogues, suggesting that the activation of a GTP-regulatory protein alone does not trigger exocytosis in these cells. TPA (12-0-tetradecanoylphorbol- 13-acetate), a direct activator of protein kinase C, was found to mimic the effects of the GTP analogues, inducing a dose-dependent potentiation of the calciumevoked release in a-toxin-permeabilized cells. Treatment of the permeabilized cells with sphingosine, a potent inhibitor of protein kinase C, completely abolished the stimulatory effects of both TPA and GTP-7-S. Moreover, long term incubation of chromaffin cells with TPA, a treatment which depletes cells of protein kinase C activity, suppressed the stimulatory effects of GTP-7-S. Protein kinase C is activated when it becomes membrane-bound in the presence of calcium and diacylglycerol; here, GTP-7-S was found to enhance the calcium-induced translocation of protein kinase C to membranes in a-toxin-permeabilized cells. These results suggest that guanine nucleotides modulate secretion by activating protein kinase C-linked events in chromaffin cells. Furthermore, the potentiation of calcium-induced secretion in a-toxin-permeabilized cells following activation of protein kinase C either directly with TPA or indirectly with GTP analogues provides additional support for the concept that protein kinase C may exert a positive control directly on the intracellular exocytotic machinery.
One approach to elucidate the intracellular mechanisms underlying exocytosis in secretory cells involves permeabilization of the plasma membrane, which permits manipulation of the intracellular environment. A number of procedures for membrane permeabilization have been developed recently. In these systems, cells release their secretory products by exocytosis simply in response to an elevation in calcium to micromolar concentrations (1)(2)(3)(4)(5). From such studies, a possible regulatory function of GTP analogues has been proposed in exocytosis of many different cell types but the manner in which this control is expressed varies. To date, two separate mechanisms have emerged. In platelets (6) and mast cells (7), GTP and GTP analogues act synergistically with calcium, significantly reducing the quantity of this divalent cation required to induce secretion. In other systems such as rabbit neutrophils (8) or insulinoma RINm 5F cells (9), GTP analogues appear to both potentiate the response to calcium and activate exocytotic secretion independently of calcium.
For several experimental reasons, adrenal medullary chromaffin cells provide another excellent tool to study the molecular mechanisms underlying release of secretory products (10,ll). Postreceptor events controlling exocytosis have been extensively studied in these cells as they have proven amenable to cell permeabilization techniques (12)(13)(14)(15). Unexpectedly, certain results reported for chromaffin cells contradict data obtained from other secretory systems. In bovine chromaffin cells permeabilized by exposure to high voltage electric fields, guanine nucleotides inhibit or potentiate calciumevoked exocytosis, depending on the GTP analogue used (16), whereas in cells permeabilized with detergents, GTP analogues have no effect on calcium-induced secretion but slightly enhance calcium-independent catecholamine release (17). Since permeabilization with a high voltage discharge can only be used for cells in suspension (2), this contradiction may reflect structural and functional differences between freshly isolated cells and cells maintained in culture. On the other hand, treatment of chromaffin cells with detergents induces a prompt leakage of cytoplasmic proteins (18) suggesting that the lack of effect of GTP analogues on calcium-evoked secretion may be due to the release of some GTP-binding proteins or target proteins controlled by them.
Previously we have described the use of a-toxin from Staphylococcus aureus as a tool for selectively permeabilizing the plasma membrane of chromaffin cell (19). The lesions generated by the toxin have a dimension which permits flux of ions and small metabolites but prevents the escape of larger molecules like cytoplasmic proteins (19)(20)(21)(22)(23)(24). Thus, a-toxin preserves the general cytoplasmic organization (25) and is certainly advantageous to study the requirements of exocytosis for ions, nucleotides, and modulators of similar size. We therefore have undertaken a detailed characterization of the effects of guanine nucleotides on the exocytotic process using a-toxin-permeabilized chromaffin cells. We demonstrate that GTP analogues potentiate calcium-dependent catecholamine secretion but do not trigger secretion in the absence of cal-cium. In addition, our results provide evidence that GTP acts on a putative G-protein which may regulate exocytosis by controlling protein kinacre C activation. (16 Ci/mmol) was supplied by Amersham (Les Ulis, France). a-Toxin was from Institut Pasteur Production (Paris, France). 12-0-Tetradecanoylphorbol-13-acetate (TPA)' and forskolin were purchased from Sigma. Digitonin was from Merck (Darmstadt, West Germany). All nucleotides were purchased from Boehringer Gmbh (Edannheim, West Germany). Sphingosine was a generous gift from Dr. N. Neskovic (Centre de Neurochimie, Strasbourg, France).
[3H]Noradrenaline Relame from Permeabilized Cells-Chromaffin cells were loaded for 30 rnin with [3H]noradrenaline, washed four times with Locke's solution (19), two times with calcium-free Locke's solution (containing 1 mNI EGTA) and once with calcium-free KG medium (150 mM glutamate, potassium salt, 10 mM PIPES, 5 mM nitrilotriacetic acid, 0.5 mM EGTA, 5 mM MgATP, 4.5 mM magnesium acetate, 0.2% bovine serum albumin, pH 7.0). Washing intervals were set constant at 10 min. Cells were subsequently permeahilized for 30 min with 40 units/ml a-toxin in 200 pl of calcium-free KG medium. Extracellular flui'ds were then removed and cells were stimulated with 200 pl of KG! medium containing various amounts of calcium chloride to yield the indicated free Ca2+ concentrations and magnesium acetate to maintain a final free M$+ concentration of 1 mM. Release of [3H]noradrenaline is expressed as the percentage of total radioactivity present in the cells prior to calcium-induced stimulation. The amount of [3H]noradrenaline released during permeabilization remained close t o 5% of the total radioactivity present in the cells prior to permeabilization.
S. aureus a-Toxin-S. aureus a-toxin was purified by ammonium sulfate precipitation at 80'% saturation and stored as a suspension in an 80% saturated ammonium sulfate solution at 4 "C. The hemolytic activity of the toxin solutions was determined with 2.5% rabbit erythrocytes (19). Cultured chromaffin cells (2.5 X 10' cells) were permeabilized for 30 min.  Calcium Concentration-The exact free calcium concentration in KG medium was calculated as described (29) using the stability constants given by Sillen and Martell (30). The free calcium concentration was also measured using a selective calcium minielectrode operating with a neutral carrier incorporated into a polyvinyl chloride membrane (31).
Presentation of Data-All experiments described were carried out on at least two different cell preparations. In the figures which are representative of a typical experiment, data are given as the mean of triplicate determinations on the same cell preparation * S.E. Fig. 1 shows the effect of GTP-7-S, a nonhydrolyzable GTP analogue, on catecholamine secretion from a-toxin-permeabilized chromaffin cells. GTP-7-S promoted a dose-dependent rise in calcium-evoked secretion: for example, the enhanced release induced by 50 p~ GTP-7-S was 70% greater than that induced by calcium alone. Similarly, GMP-PNP potentiated calcium-evoked release but was less effective than GTP-7-S since at 200 p~ GMP-PNP increased the calcium-induced release by only 60% (data not shown). Thus, GTP-7-S was relatively more potent, an observation that was also reported for permeabilized platelets (6) and mast cells (7). Stimulation of secretion was specific for nonhydrolyzable guanine nucleotides since we could not reproduce these effects using GTP. In addition, neither GTP-7-S ( Fig. 1) nor GMP-PNP (data not shown) were able to trigger secretion in the absence of calcium, at any of the tested concentrations (Table I).

Effects of Guanine Nucleotides on Secretion from a-Toxinpermeabilized Chromaffin Cells-
As expected, the effect of GTP-7-S could be inhibited by the stable GDP analogue GDP-@-S (Fig. 2). Simultaneous [3H]Noradrenaline release is expressed as a percentage of the total radioactivity present in the cells prior to stimulation. GTP-7-S potentiates calcium-dependent secretion in a dose-dependent manner but does not affect the basal calcium-independent release.

Nucleotide dependence of calcium-induced catecholamine secretion from a-toxin-permeabilized chromaffin cells Chromaffin cells were permeabilized for 30 min with a-toxin in a
MgATP-and calcium-free KG medium in order to deplete cells of their cytosolic ATP. Cells were then stimulated for 10 min with KG medium in the presence or absence of 20 pM free calcium and with the indicated concentrations of MgATP and GTP-y-S. Under these conditions, calcium-dependent [3H]noradrenaline release is observed only when MgATP is present during the stimulation period. Guanine nucleotides are unable to induce catecholamine secretion in the absence of exogenous ATP. ND = not determined. incubation of 500 p~ GDP-@-S with GTP-7-S completely blocked the secretory response induced by 20 PM GTP-7-53 and significantly reduced the effect of higher GTP-7-S concentrations (50 p~ GTP-y-S enhanced secretion by 15% in these conditions compared to 60% in the absence of GDP-@-S). However, the inhibitory GDP analogue alone had no effect on calcium-dependent release from a-toxin-permeabilized cells, even at concentrations as high as 1 mM (data not shown).
In order to further characterize guanine nucleotides activity, we examined the effects of GTP-7-S on the calcium doseresponse curve for secretion. The presence of 50 p~ GTP-7-S during the stimulation period produced a leftward shift in the calcium activation curve for secretion since the relative stimulatory effect was more pronounced at low calcium concentrations (Fig. 3). This observation can be interpreted as a guanine nucleotides-mediated increase in the apparent affinity of exocytosis for calcium. A particular feature of the use of a-toxin as a permeabilizing agent compared to high voltage discharge or detergent lies in the possibility of inducing the release of high levels of catecholamines at higher calcium concentrations. However, free calcium levels of 100 p~ induce a maximal secretory response, and the addition of more calcium has no further effect on the extent of secretion (Fig. 3). In contrast, the combined presence of 50 pM GTP-7-S and 100 p~ free calcium during the stimulation period increased the level of the secretory response (Fig. 3), indicating that guanine nucleotides are also able to enhance the maximal effects of the Ca2+-activated processes.
MgATP Dependence of Calcium-induced Release from a-Toxin-permeabilized Celk-In the preceding experiments, MgATP was present during the permeabilization and the stimulation period. We have previously shown that MgATP is essential for calcium-stimulated release from a-toxin-permeabilized chromaffin cells and cannot be replaced by other nucleotides (19). This requirement for ATP is observable only when cells are depleted of their cytoplasmic ATP by permeabilizing in ATP-free medium. In view of a recent report showing that calcium combined with nonhydrolyzable analogues of GTP causes exocytotic secretion from permeabilized mast cells (7), we examined whether GTP-7-S supported calcium-induced release from a-toxin-treated cells depleted of endogenous ATP prior to stimulation. Cells were first permeabilized in calcium-and ATP-free medium and subsequently stimulated with permeabilizing medium containing 100 p~ free calcium. As shown in Table I, catecholamine release did not occur when ATP was absent during stimulation even in the presence of 100 p~ GTP-$3. Thus, in contrast to mast cells, guanine nucleotides cannot synergize with calcium to trigger secretion in the absence of phosphorylating nucleotides in permeabilized chromaffin cells.
Effect of Protein Kinase C Modulators on. Catecholamine Secretion from a-Toxin-treated Cells-In several systems, guanine nucleotides can activate phospholipase C, consequently generating inositol trisphosphate (IP3) and diacylglycerol (32)(33)(34)(35). Although one of the functions of IPS is to mobilize intracellular calcium, such a contribution to the secretory response would probably be masked by our experimental conditions due to the presence of calcium chelators in the permeabilizing medium. Nevertheless, we tested the possibility that guanine nucleotides may increase the secretory response by releasing calcium from intracellular stores. Micromolar concentrations of IP3 trigger an immediate and dosedependent release of calcium from intracellular sites in permeabilized chromaffin cells (36, 37). We therefore examined the effect of IP, on tlne calcium-evoked secretion from atoxin-treated cells (Table 11). However, unlike guanine nucleotides, IP3 at concentrations up to 100 p~ did not increase the extent of secretion at any of the calcium concentrations tested. Moreover if guanine nucleotides would enhance secretion by releasing intracellular calcium, then increasing the buffering conditions for calcium in the incubation media should reduce their stimulatory effect on secretion. We thus examined the effect of increasing the NTA concentration in the KG medium on the ability of GTP-7-S to potentiate catecholamine release when the calcium concentration was buffered at 20 PM by NTA. As shown in Table 111, increasing the NTA concentration during the stimulation period neither modified the secretory response evoked by 20 PM free calcium nor altered the stimulatory effects of GTP-7-S in a-toxinpermeabilized cells. Similarly, we observed that increasing EGTA concentration from 0.5 to 5 mM during the stimulation period had no effect on the ability of GTP-7-S to enhance the calcium-evoked secretion (data not shown). Thus it seems unlikely that guanine nucleotides effects on secretion are

TABLE I1 Effect of inositol trisphosphate (IPS) on calcium-evoked secretion from a-toxin-permeabilized chromaffin cells
Cells were treated for 30 min with a-toxin in calcium-free KG medium. Following removal of extracellular fluids, cells were stimulated for 10 min with KG medium containing the indicated free calcium concentrations, in the absence or presence of IP3. IP3, which is known to release calcium from intracellular stores, has no effect on the calcium-evoked secretion of catecholamines in permeabilized cells. Alternatively, guanine nucleotides may act through phospholipase C thereby producing diacylglycerol which in turn stimulates protein kinase C and potentiates catecholamine release. This possibility was examined in the experiments described in Figs. 4A and 5 using TPA, a potent activator of protein kinase C. The presence of 100 nM TPA during a 10min incubation period before stimulation induced a dosedependent potentiation of calcium-evoked catecholamine release (Fig. 4A). In addition, TPA was found to mimic the effects of GTP analogues on the calcium dose-response curve for secretion increasing the apparent sensitivity to calcium and the extent of catecholamine release without significantly reducing the minimum free calcium concentration required to trigger secretion (compare Fig. 5 with Fig. 3).
Sphingosine has been recently described as a potent physiological inhibitor of protein kinase C activity (38)(39)(40)(41). The effect of sphingosine on secretion from a-toxin-treated cells is shown in Fig. 4B. Permeabilized cells were incubated for 10 min with various concentrations of sphingosine in calciumfree KG medium and subsequently stimulated in the absence of sphingosine with KG medium containing 20 or 100 pM free calcium to determine the calcium-evoked release or with either calcium-free KG medium or MgATP-free KG medium containing calcium to determine the basal release. As seen in Fig. 4B, sphingosine increased the basal release determined in the absence of calcium but had little effect on the basal release measured in the presence of calcium but absence of MgATP, an observation which indicates that sphingosine may cause some intracellular perturbations in permeabilized cells when applied in calcium-free media for prolonged incubation periods (over 10 min). Sphingosine did not interfere with the catecholamine release evoked by calcium and MgATP, an observation which may be related with the fact that the inhibitory GDP analogue GDP-(3-S had no effect on secretion from permeabilized cells. Thus, the effects of guanine nucleotides on calcium-evoked secretion parallel those of TPA and sphingosine which directly modulate protein kinase C activation.
Next we addressed the question of whether TPA and guanine nucleotides could act at the same site on the secretory process. We examined the effect of a submaximal concentration of TPA on both submaximal and maximal concentrations of GTP"& As illustrated in Fig. 6, 50 nM TPA increased the effect produced by 25 PM GTP-7-S but was unable to enhance further secretion in presence of 50 p~ GTP-7-S. This is consistent with the possibility that TPA and GTP-7-S may be involved by a common mechanism in the control machinery of exocytosis.

TO strengthen the hypothesis that guanine nucleotides may TABLE I11
Effect of NTA concentration on the secretory response to GTP-y-S and TPA Chromaffin cells labeled with [3H]noradrenaline were permeabilized for 30 min with a-toxin in calcium-free KG medium. Cells were then stimulated for 10 min with KG medium containing either 5 or 15 mM NTA and the appropriate amount of CaC12 to give the indicated free calcium concentration, in the presence or absence of 200 nM TPA o:r 100 p M GTP-y-S. Increasing NTA concentration does not alter the ability of TPA and GTP-y-S to enhance the release of catecholamines when the calcium concentration is buffered at 20 U M bv NTA. Effect of TPA and sphingosine on secretion from atoxin-permeabilized chromaffin cells. A, cells were exposed to 40 units/ml a-toxin in calcium-free KG medium and then incubated for 10 min in calcium-free KG medium containing various concentrations of TPA. The medium was subsequently removed and cells were stimulated with KG medium containing 20 p~ free calcium (closed symbols). Basal release (open symbols) was determined in calciumfree KG medium. B, a-toxin-permeabilized chromaffin cells were incubated for 10 min in calcium-free KG medium containing the indicated concentrations of sphingosine. Cells were subsequently stimulated with KG medium containing 20 or 100 p~ free calcium (closed symbols) in the absence of sphingosine. Basal release was determined in either calcium-free KG medium (open squares) or MgATP-free KG medium containing 20 p~ free calcium (open circles). Following 10 min stimulation, extracellular fluids were collected and [3H]noradrenaline (3H-NA) was assayed in the solutions and remaining cells.
[3H]Noradrenaline release is expressed as a percentage of the total radioactivity present in the cells prior to stimulation. modulate secretion by activating protein kinase C, we examined the effect of GTP--/-S and TPA on a-toxin-permeabilized chromaffin cells preincubated with sphingosine in order to block protein kinase C activity. As shown in Fig. 7A, treatment of permeabilized cells with sphingosine completely suppressed the stimulatory effects of GTP-7-S and TPA on calcium-dependent secretion. In many cell types, chronic exposure to active phorbol esters results in down-regulation of

TPA and guanine nucleotides on membrane-associated protein kinase C actiuity in a-toxin-permeabilized chromaffin cells
a-Toxin-permeabilized cells were stimulated for 10 min with KG medium containing the indicated concentrations of calcium, in the presence of 50 p~ GTP-7-S or 100 nM TPA. Cells were then immediately scraped in ice-cold homogenizing buffer, and the suspension was sonicated and centrifuged to separate membrane from soluble fractions. Data show membrane-bound protein kinase C activity. The results are from three experiments on three different cell preparations with four wells/group in each experiment.  Fig. 7B, calcium-evoked catecholamine release was not modified but neither GTP-y-S, nor TPA were able to potentiate the secretory response from cells pretreated with TPA. Taken together, these findings suggest that guanine nucleotides may modulate secretion by a mechanism involving the activation of protein kinase C. In support of the specificity of this effect, neither sphingosine nor long term treatment with TPA affected the response to forskolin, a well established adenylate-cyclase activator, which enhances catecholamine release from intact cells (45,46) and also from a-toxin-permeabilized cells (Fig. 7). Effect of Guanine Nucleotides on Membrane-associated Protein Kinase C Activity in a-Toxin-permeabilized Chromaffin Cells-Activation of protein kinase C is known to be associated with a translocation of the enzyme from a soluble to a membrane-bound compartment. Indeed, TPA produced a large increase in membrane-bound protein kinase C activity in a-toxin-permeabilized cells (Table IV), which represented a translocation of approximately 85% of the total cellular protein kinase C activity. To confirm that guanine nucleotides may modulate secretion by activating protein kinase C, we examined the effect of GTP-y-S on the membrane-associated protein kinase C activity. The fraction of protein kinase C that was particulate under resting conditions, i.e. in the absence of calcium (Table IV), represented 20-30% of the total cellular protein kinase C activity. As shown in Table IV, ~ the cells prior to stimulation. Results are expressed relative to the response obtained from control cells (control; open column), stimulated with KG medium containing 100 p M free calcium (34.8 & 0.3% of [3H]noradrenaline cell content). Sphingosine, a potent inhibitor of protein kinase C, and long term incubation with TPA, a treatment known to cause a progressive decline of protein kinase C activity, reverse the stimulatory effect that TPA and guanine nucleotides are able to exert on the calcium-dependent secretory response.
of Secretion in Chromaffin Cells stimulation of a-toxin-permeabilized cells for 10 min with 20 pM Caz+ increased the membrane-bound protein kinase C activity. The presence of 50 PM GTP-7-S during the stimulation period caused a moderate but significant increase in the shift of protein kinase C to membranes in the presence of calcium (Table IV). This shift was reproducible from one chromaffin cell preparation to the other (Table IV; n = 3). Thus, guanine nucleotides promote the association of protein kinase C with chromaffin cell membranes, an observation which is consistent with the effect of guanine nucleotides on catecholamine release being mediated by protein kinase C activation.
The apparent discrepancy between the similar enhancement of catecholamine release caused by GTP-7-S and TPA and the relative small shift of protein kinase C to membranes induced by GTP-7-S compared to TPA may suggest that either (i) guanine nucleotides do not entirely modulate secretion by activating protein kinase C or (ii) there is not a complete linear relationship between the amount of membrane-bound protein kinase C and catecholamine secretion. Support for the latter hypothesis comes from recent observations on permeabilized chromaffin cells (60), indicating that a shift in membrane-bound protein kinase C from 20 to 60% did not cause an additional increase in subsequent calciumdependent secretion.

DISCUSSION
The experiments presented here were designed to determine whether GTP analogues can modulate calcium-dependent exocytosis in adrenal medullary chromaffin cells and to determine the cellular mechanisms underlying this control. In platelets (6) and mast cells (7) guanine nucleotides synergize with calcium to trigger the exocytotic reaction. However, the results concerning adrenal chromaffin cells are conflicting. Exposure of electrically permeable bovine and chicken chromaffin cells to guanine nucleotides has different effects: calcium-dependent exocytosis in bovine cells is inhibited by GTP-7-S but activated by GMP-PNP, whereas that in chicken cells is stimulated by both molecules (16). In detergent-permeabilized bovine chromaffin cells, guanine nucleotides cause a small increase in the calcium-independent basal release and the effects of guanine nucleotides and calcium on catecholamine secretion are subadditive (17).
A simple interpretation cannot be easily offered to the various observations obtained with electropermeabilized cells. Concerning the effect of GTP analogues in detergent-treated cells, one plausible explanation may be the leakage of cytoplasmic proteins induced by permeabilization with digitonin (47-49). We have recently presented evidence that one or more of the released cytoplasmic proteins are responsible for the loss of calcium-evoked secretory activity observed following digitonin treatment (18). In view of these observations, we suggest that the inability of guanine nucleotides to increase calcium-dependent secretion in digitonin-treated chromaffin cells is due to the loss of some GTP-binding proteins and/or proteins controlled by them. Additional support for this idea comes from the observation that phorbol esters only affect secretion when cells are preincubated with these compounds prior to digitonin permeabilization, a treatment which results in the retention of protein kinase C by favoring the membrane-bound form (47, 49).
We have previously characterized catecholamine release from a-toxin-permeabilized chromaffin cells (19) and observed that the minimum free calcium concentration required to trigger secretion is 4 p M . In contrast, detergent-treated cells and electropermeabilized cells show a greater sensitivity for calcium since secretion occurs at 0 .5 p~ calcium (4, 13). The particularity of a-toxin-treated cells is that the extent of secretion increases in presence of high levels of calcium reaching a maximal release of 40% of the total catecholamines at 100 p M free calcium (25). A similar level of secretion is never observed in detergent-treated cells (25) or in electropermeabilized cells (50) since increasing the calcium concentration above 20 pM results in a progressive inhibition of the secretory response in these permeable cells (25,50). Interestingly, atoxin-and digitonin-permeabilized cells release comparable amounts of catecholamines at 20 ~L M free calcium (25), a cytosolic calcium concentration which is close to the value measured in intact cells following nicotinic stimulation (51, 52).
The integrity of the cytoplasmic composition is better preserved following a-toxin treatment (19, 25). Moreover, atoxin-permeabilized cells are able to secrete catecholamines over long period of times (60 min) compared to digitoninpermeabilized cells which lose their responsiveness to calcium within 10 min (25). Thus, by using the a-toxin experimental model, we expected to be in the position of identifying with better precision the actual effects of GTP analogues on the overall calcium-dependent exocytotic process. Our data give evidence that in a-toxin-permeabilized chromaffin cells GTP analogues are not able to trigger exocytosis by themselves but potentiate calcium-evoked secretion. In addition, the effects of GTP analogues on secretion seem to be linked to protein kinase C-mediated events, suggesting that guanine nucleotides may act on catecholamine secretion through a putative GTP-binding protein (Gp) which may control phospholipase C activity, thereby generating diacylglycerol and activating protein kinase C (53). These results do not exclude the possibility that other GTP-binding proteins may also be essential components of the transducing machinery related to the secretory process. Indeed the presence of three substrates for pertussis toxin which specifically ADP-ribosylates GTP-binding proteins, has been described recently in both plasma and secretory granule membranes of chromaffin cells (54).
It is of interest to mention that GTP analogues cannot substitute for MgATP to support calcium-evoked exocytosis implying that chromaffin cells do not seem to possess a GTPbinding protein (G,) that directly stimulates exocytosis as recently described in mast cells (7, 55) and neutrophils (8). Thus it appears that the specific requirements of the exocytotic release varies from cell to cell rendering pertinent the question of whether an unique mechanism controls the secretory process. In certain cells containing different granule populations, each type of granule can be released independently following a rise in intracellular calcium concentration which supports the view that more than one secretory control mechanism exists, the nature of which depends on the secretory products to be released. For example, platelets are able to release selectively serotonin or N-acetyl-P-glucosaminidase from distinct granules (56), suggesting that the exocytotic process for each granule population is regulated independently. In neutrophils, secretion from two distinct granule populations is modulated differently by calcium (57), but only one class of secretory granules release their contents in response to protein kinase C activation (38). Taken together these observations reinforce the concept that a wide range of control mechanisms underlie the secretory process.
The activation of protein kinase C by phorbol esters has been of central importance in implicating this enzyme in various cellular events. Here we show that the major effect of the activation of protein kinase C either directly with TPA or indirectly with GTP analogues is to increase the extent of the calcium-dependent catecholamine release. Since both TPA and GTP-7-S have no effect on exocytosis in the absence of calcium it is likely that protein kinase C may be involved either as a modulator of the calcium-dependent processes or as an integral part of 1;he machinery controlling exocytosis. In order to further define the involvement of protein kinase C in the secretory procless, specific antagonists are necessary to inhibit the enzyme. Hannun et al. (39) have previously shown that sphingosine is a potent inhibitor of purified protein kinase C and specifically inhibits the protein kinase Cinduced phosphorylaticln in platelets (39). The mode of inhibition is competitive with respect to phorbol esters, which activate the kinase by interacting at the same site as the natural activator diacylglycerol. Sphingosine has been used to investigate the role of protein kinase C in differentiation of HL-60 cells (40) and to demonstrate that in neutrophils, protein kinase C participates in the secretion of specific granule proteins but not in that of azurophilic granule proteins (38). Sphingosine does not apparently affect other important intracellular signals, including calcium levels (411, calmodulin-dependent phosphorylation (39), and cyclic AMPdependent processes (41).
Recently, Burgoyne et al. (58) showed that sphingosine inhibits calcium-activated exocytosis in digitonin-permeabilized cells and concluded that endogenous activation of protein kinase C is a major requirement for secretion in chromaffin cells. In contrast, Holz and Senter (59) reported that mild proteolysis with trypsin inhibited the TPA-induced enhancement of secretion without affecting the calcium-evoked release from digitonin-treated cells, implying that protein kinase C activation is not necessary for secretion. In the present study, we have reduced protein kinase C activity in chromaffin cells by using two diffelrent approaches: treatment with sphingosine and chronic exposure to active phorbol esters which results in down-regulation of protein kinase C with a loss of enzyme activity. Both treatments remove the effect of TPA on secretion without affecting the calcium-dependent secretory response in a-toxin-permeabilized cells. Thus the activation of protein kinase C may not be an obligatory step in the machinery of exocytosis in permeabilized cells. Consistent with this conclusion is the finding that H7 and sphingosine which reduce by 80% the protein kinase C activity in electropermeabilized chromaffin cells have little effect on the calcium-dependent release (50).
To conclude, we propose that the secretory response from a-toxin-permeabilized chromaffin cells consists of a primary response, which is not affected by sphingosine or by long term treatment with TPA, and a secondary response triggered by TPA or guanine nucleotides and inhibited by sphingosine or by long term treatment with TPA. Both primary and secondary responses require calcium and MgATP. Our data suggest that the secondary response could be a protein kinase Cdependent event requiring the activation of the enzyme by a mechanism implicating a GTP-binding protein.