INTRODUCTION

The agonist-induced entry of Ca²⁺ from the extracellular medium is a key component in the cytosolic Ca²⁺ signals that link activation of various receptors on the cell surface with the initiation and control of cell function. Despite some continued controversy, models for such Ca²⁺ entry in nonexcitable cells currently focus on the so-called capacitative mechanism where the sustained depletion of agonist-sensitive Ca²⁺ stores is both necessary and sufficient to activate entry (1, 2). This capacitative mechanism of Ca²⁺ entry activation has been shown to be important for the sustained Ca²⁺ signals seen in a variety of nonexcitable cells under conditions of maximal stimulation, and recently, direct electrophysiological measurements of a Ca²⁺-selective current that is specifically activated by depletion of the intracellular Ca²⁺ stores have been made (3, 4, 5, 6, 7). In contrast, relatively little attention has been given to Ca²⁺ entry under more physiological levels of agonist where, for example, oscillating Ca²⁺ responses are common (8, 9). Current models for these phenomena generally emphasize special properties and modifications of inositol 1,4,5-trisphosphate (InsP₃)¹-induced Ca²⁺ release, and Ca²⁺ entry is assumed to play only a relatively minor role. This assumption is contradicted by our recent findings using cells from the exocrine avian nasal gland where we showed that such patterns of [Ca²⁺]_i signaling do not depend exclusively on the generated levels of InsP₃ but also, critically, on the activation of Ca²⁺ entry which acts by regulating the InsP₃-mediated liberation of intracellular Ca²⁺ from agonist-sensitive stores (10, 11). Subsequent studies examining the nature of agonist-activated Ca²⁺ entry under these conditions (12) revealed that it demonstrates properties that are incompatible with current versions of the capacitative model. Specifically, (a) activation of this entry pathway is critically dependent on agonist occupation of the receptor but independent of the status of agonist-sensitive intracellular Ca²⁺ stores; (b) entry (measured as the rate of Mn²⁺ quench) is independent of the cyclical emptying and refilling of the agonist-sensitive Ca²⁺ pool during oscillations; and (c) on initial exposure to low agonist concentrations, activation of Ca²⁺ entry precedes any detectable release of Ca²⁺ from the stores.

The novel finding that agonist-induced Ca²⁺ entry during [Ca²⁺]_i oscillations is noncapacitative raises the question of the nature of the signaling pathway responsible for its activation. Recently, it has been reported that the generation of [Ca²⁺]_i oscillations in pancreatic acini by agonists acting at the high affinity cholecystokinin receptor involves a phospholipase A₂ (PLA₂)-mediated formation of arachidonic acid (13, 14). Although the precise role of the generated arachidonic acid was not determined in these studies, our demonstration that carbachol-induced [Ca²⁺]_i oscillations in the avian nasal gland cells are acutely dependent on the noncapacitative entry of Ca²⁺ (10) suggested that this entry could be a potential target for arachidonic acid generated as a result of agonist action. We now present evidence confirming that indeed the control of the novel noncapacitative Ca²⁺ entry during oscillations involves a PLA₂-mediated generation of arachidonic acid. As carbachol-induced [Ca²⁺]_i oscillations in these cells require both the generation of InsP₃ and the arachidonate-mediated Ca²⁺ entry, we conclude that the successful initiation and maintenance of an oscillatory [Ca²⁺]_i response involves a combined and coordinated activation of a PLC and a PLA₂.

MATERIALS AND METHODS

The methods used for the isolation of cells from the nasal glands of 6-10-day-old domestic ducklings (Anas platyrhynchos), their loading with the fluorescent probe indo-1, and subsequent microfluorometric measurements in individual cells were as described previously (10, 16). Changes in [Ca²⁺]_i in single individual cells were determined as the ratio of the emitted fluorescence, measured as photon counts, at 405 and 485 nm with excitation at 350 nm, corrected for background and autofluorescence online. The simultaneous determination of fluorescence ratio and Mn²⁺ quench in individual cells was essentially as described previously (16). Briefly, this involves the summing of the two emitted fluorescences which are factored such that the resulting adjusted total fluorescence is independent of [Ca²⁺]. Determination of the required factor is performed for each individual cell at the beginning of the experiment by observing the effects of induced oscillatory changes in [Ca²⁺]_i, in the absence of extracellular Mn²⁺.

Determinations of arachidonic acid generation followed standard techniques (17, 18). Briefly, isolated cells were plated into a six-well culture plate and incubated in L-15 culture medium for 1 h, after which nonadherent cells were removed by washing with fresh medium. [³H]Arachidonic acid (0.5 µCi/ml, American Radiochemicals) was added and the cells incubated for 20-24 h at 37 °C in a CO₂ incubator. The wells containing the labeled cells were then washed three times with saline containing 0.2% fatty acid-free bovine serum albumin (to act as a sink for released arachidonic acid). Agonists and/or inhibitors were added as appropriate and the released arachidonic acid determined by liquid scintillation counting of samples of the supernatant medium. Counts released were normalized to the total counts in the cells, determined by solubilization of the cells in each well with 1.5 M NaOH and counting of the resulting solution.

Cellular levels of InsP₃ were determined in preparations of dispersed nasal gland cells using a commercial radiolabeled binding assay (Amersham) as described previously (19).

Arachidonic acid, isotetrandrine, PLA₂-activating protein (PLAP)-peptide, indomethacin, and nordihydroguaiaretic acid were all obtained from Biomol.

RESULTS

Demonstration that muscarinic receptor activation is coupled to the generation of arachidonic acid was carried out in cells loaded overnight with [³H]arachidonic acid. Addition of the muscarinic agonist carbachol to these cells resulted in a marked increase in the release of arachidonic acid to the medium (Fig. 1A). Obviously, the observed increase in arachidonic acid release may simply reflect a secondary or downstream effect of muscarinic receptor activation rather than a direct effect. For example, the addition of carbachol will result in the elevation of [Ca²⁺]_i, and this could activate a PLA₂ resulting in the release of arachidonic acid. To examine whether the release of arachidonic acid was dependent on (i.e. secondary to) the simultaneous rise in [Ca²⁺]_i, the effect of carbachol was determined in cells preloaded with the Ca²⁺ chelator BAPTA. Cells were loaded with BAPTA by incubation with the acetoxymethyl ester form (BAPTA-AM) sufficient to completely abolish all [Ca²⁺]_i responses to the addition of carbachol (Fig. 1B). Examination of the carbachol-induced release of arachidonic acid in such cells revealed that it was unaffected (Fig. 1A). Although such findings do not exclude a possible additional Ca²⁺-dependent activation of arachidonic acid release, they do indicate that an agonist-induced increase in [Ca²⁺]_i is not a necessary prerequisite for the activation induced by carbachol.

In individual avian nasal gland cells loaded with the Ca²⁺-sensitive fluorophore indo-1, simple addition of a low concentration (2-5 µM) of exogenous arachidonic acid to otherwise unstimulated cells induced a marked increase in [Ca²⁺]_i (Fig. 2A). This reflects an arachidonate-activated entry of Ca²⁺ because virtually no increase in [Ca²⁺]_i was seen in a nominally Ca²⁺-free medium (estimated [Ca²⁺] = approximately 10 µM) (Fig. 2B). Furthermore, simultaneous measurement of the rate of Mn²⁺ quenching showed that this increased dramatically on addition of exogenous arachidonic acid (Fig. 2C). This increase in Ca²⁺ entry did not result from a generalized increase in ionic permeability, however, as whole cell perforated patch-clamp measurements (20) in current clamp mode showed that the overall cell membrane potential was maintained in the presence of 5 µM arachidonic acid (data not shown). It should also be noted that the activation of this Ca²⁺ entry appears to be acutely sensitive to arachidonic acid as the concentrations employed here are some 10-fold lower than those commonly used to demonstrate arachidonate-induced effects on, for example, K⁺ channels in smooth muscle cells (21). Importantly, the concentrations used are significantly lower than those reported to induce the mobilization of intracellular Ca²⁺ stores in certain cells (22, 23), and no such mobilization could be detected in our cells at any of the concentrations of arachidonic acid used here.

It has been reported recently that the [Ca²⁺]_i oscillations and amylase secretion induced by agonists acting at high affinity cholecystokinin receptors in pancreatic acinar cells, processes believed to involve the PLA₂-mediated generation of arachidonic acid (13, 14), can be inhibited by the biscoclaurine (bisbenzylisoquinoline) alkaloid isotetrandrine (24). We therefore examined the effect of isotetrandrine on the oscillatory [Ca²⁺]_i signals induced by low concentrations of carbachol in individual indo-loaded nasal gland cells. Addition of isotetrandrine (10 µM) resulted in an immediate, yet readily reversible, inhibition of the carbachol-induced [Ca²⁺]_i oscillations (Fig. 3A). The inhibitory effect we observed on [Ca²⁺]_i oscillations closely resembles that seen when Ca²⁺ entry is reduced by depolarization of the cell in a high K⁺ medium (Fig. 3B), suggesting that isotetrandrine may also be inhibiting Ca²⁺ entry. Consistent with this, we found that the inhibition of oscillations by isotetrandrine was associated with a marked reduction of the simultaneous rate of Mn²⁺ entry measured as the quenching of intracellular indo-1 (Fig. 3C). We have shown previously that the initiation of [Ca²⁺]_i oscillations by carbachol in these cells is associated with an enhanced rate of Mn²⁺ quenching of intracellular indo-1 (10, 12). Although there are several potential problems in the use of the Mn²⁺ technique as an indicator of Ca²⁺ entry, it is clear that an agonist-induced change in the quench rate must reflect at least some component of divalent cation entry (12) and that this entry is inhibited by isotetrandrine.

In marked contrast to its effects on oscillatory [Ca²⁺]_i signals, isotetrandrine was completely without effect on the sustained [Ca²⁺]_i signal generated by higher concentrations of carbachol (Fig. 4A). It is well established that these sustained elevations in [Ca²⁺]_i reflect the capacitative entry of Ca²⁺, and they are abolished readily by agents known to inhibit Ca²⁺ entry such as lanthanum, or SK&F 96365, or simply by reducing the driving force for Ca²⁺ entry (for example by depolarizing the cell membrane by exposure to a high K⁺ medium). Capacitative entry activated as a result of store depletion by thapsigargin (0.5 µM) was similarly unaffected by isotetrandrine (Fig. 4B). As a further test of the inability of isotetrandrine to affect capacitative Ca²⁺ entry, we examined whether the drug had any influence on the the refilling of agonist-sensitive Ca²⁺ stores following their discharge with a high agonist concentration. The ability of the stores to refill was determined by examining the magnitude of the [Ca²⁺]_i spikes induced by paired exposures to 10 µM carbachol spaced approximately 200 s apart. In the continued presence of isotetrandrine, the magnitudes of the two [Ca²⁺]_i spikes induced by the addition of a high concentration of carbachol were virtually identical, indicating that the refilling of the agonist-sensitive stores in the period between the two carbachol exposures was unimpaired (Fig. 4C). As refilling of the agonist-sensitive stores is known to depend on capacitative entry of Ca²⁺, the absence of any effect of isotetrandrine on this process confirms that it is without effect on this pathway. It should also be noted that these data demonstrate that isotetrandrine is without any obvious influence on [Ca²⁺]_i in unstimulated cells, and preexposure to the drug does not affect the InsP₃-induced release of Ca²⁺. We conclude that the observed effects of isotetrandrine on [Ca²⁺]_i oscillations appear to result from specific effects on the noncapacitative entry of Ca²⁺, which we have previously shown to be critical for agonist-induced oscillations (10, 11, 12).

Isotetrandrine and its related biscoclaurine alkaloids are known to produce many different effects in various cell types (see 25 and references therein). However, studies have revealed that at the low micromolar concentrations employed here, these diverse actions can generally be ascribed to either effects on voltage-operated Ca²⁺ channels (specifically by interacting at the benzothiazipine binding site) (26) or to effects on the activation of a PLA₂ and the subsequent generation of arachidonic acid (25, 27). In common with most nonexcitable cells, voltage-operated Ca²⁺ channels are absent in avian nasal gland cells. Nevertheless, it was considered possible that isotetrandrine might be interacting with a benzothiazipine-like site on a putative non-voltage-operated Ca²⁺ channel. However, the classic benzothiazipine receptor antagonist diltiazem was completely without effect on the carbachol-induced oscillations in [Ca²⁺]_i (data not shown). In contrast, an action on PLA₂ activity was indicated by the demonstration that isotetrandrine (10 µM) significantly inhibited the effect of carbachol on the release of arachidonic acid from cells preloaded with [³H]arachidonic acid (Fig. 5A). In contrast to its effects on arachidonic acid release, isotetrandrine was completely without effect on the carbachol-induced increase in InsP₃ levels inside the cell (Fig. 5B). This confirms that the inhibition of [Ca²⁺]_i oscillations by isotetrandrine did not result from effects on carbachol-induced increases in PLC activity and inositol phosphate generation. This finding is consistent with the above noted inability of isotetrandrine to affect either InsP₃-induced Ca²⁺ release (Fig. 4C) or the sustained capacitative [Ca²⁺]_i signals induced by high agonist concentrations (Fig. 4A). A similar conclusion has been reported previously for the related drug tetrandrine in studies on adrenal glomerulosa cells (28). It should also be noted that these data further indicate that the isotetrandrine-sensitive generation of arachidonic acid induced by carbachol is unlikely to originate from a combined action of a PLC-mediated production of diacylglycerol from inositol-containing phospholipids followed by the action of a diacylglycerol lipase to produce arachidonic acid (29). This, again, supports the suggestion that the observed effects of isotetrandrine involve an action on a PLA₂.

As further evidence for a role in the inhibition of PLA₂, we found that the inhibition of [Ca²⁺]_i oscillations induced by isotetrandrine could be reversed by simple addition of low concentrations of exogenous arachidonic acid (2-5 µM) despite the continued presence of isotetrandrine (Fig. 6A). Significantly, the observed range of exogenous arachidonic acid concentrations required to reverse the isotetrandrine-induced inhibition of oscillations is identical to those shown previously to activate Ca²⁺ entry directly in unstimulated cells (Fig. 2) and are 10-40 times less than those reported in other cell types to mobilize intracellular Ca²⁺ pools or to induce a nonspecific permeabilization of the cell membrane. Although the precise manner in which isotetrandrine affects PLA₂ activity is unclear, previous reports have shown that it acts by uncoupling the PLA₂ from an activating G protein rather than any direct effects on the enzyme itself (25). Consistent with this proposed mode of action, we found that the isotetrandrine-induced inhibition of oscillations could also be reversed by a small peptide that forms part of the recently cloned PLA₂-activating protein (30) (Fig. 6B). PLAP was originally isolated based on its sequence homology with the bee venom protein melittin, and the PLAP-peptide has been shown to stimulate PLA₂ activity in a variety of cells including intact pancreatic acinar cells (24) and synovial cells (31). It seems likely that the observed ability of PLAP-peptide to activate PLA₂ in intact cells results from its insertion into the membrane as its sequence includes several hydrophobic residues (30). Whatever its precise mode of action, these data clearly support the hypothesis that the observed action of isotetrandrine on the carbachol-induced [Ca²⁺]_i oscillations involves effects on PLA₂ activity.

Finally, we examined whether the observed effects of arachidonic acid on the activation of the noncapacitative Ca²⁺ entry during [Ca²⁺]_i oscillations were directly dependent on the fatty acid itself rather than any products of arachidonic acid oxidation. We found that addition of the cyclooxygenase inhibitor indomethacin (10 µM) failed to affect the [Ca²⁺]_i oscillations induced by low concentrations of carbachol (Fig. 7A). In contrast, inhibition of the lipoxygenase pathway by nordihydroguaiaretic acid (10 µM) resulted in the gradual transition of oscillations into a sustained elevated [Ca²⁺]_i signal (Fig. 7B). We interpret these data as indicating that the noncapacitative Ca²⁺ entry pathway during oscillations is likely regulated by arachidonic acid itself rather than any products of cyclooxygenase or lipoxygenase pathways. The effect of nordihydroguaiaretic acid can be explained if the principal route for arachidonic acid breakdown is via the lipoxygenase pathway. Inhibition of this pathway by nordihydroguaiaretic acid in a carbachol-stimulated cell would then result in the accumulation of arachidonic acid, enhancing Ca²⁺ entry sufficiently to generate a sustained elevation of [Ca²⁺]_i.

DISCUSSION

Together, these results demonstrate that the agonist-induced, noncapacitative entry of Ca²⁺ during [Ca²⁺]_i oscillations, which we have shown to play a critical role in driving the oscillations, involves the action of arachidonic acid generated as a result of the activation of a PLA₂. Specifically we have shown that (a) muscarinic receptor agonists activate the generation and release of arachidonic acid; (b) this activation was independent of agonist-induced changes in [Ca²⁺]_i and InsP₃ levels; (c) inhibition of this activation immediately stopped the noncapacitative Ca²⁺ entry during oscillations but was entirely without effect on capacitative entry; (d) inhibition of the noncapacitative Ca²⁺ entry during oscillations could be reversed by addition of low concentrations of exogenous arachidonic acid or by direct activation of PLA₂; and (e) addition of the same low concentrations of exogenous arachidonic acid to otherwise unstimulated cells is capable of activating Ca²⁺ entry.

It is now known that there are several types of PLA₂, but they can be classified broadly into two structurally and functionally distinct groups: secretory PLA₂ (sPLA₂) and cytosolic PLA₂ (cPLA₂) (32). If the observed effects of isotetrandrine on [Ca²⁺]_i oscillations result from its influence on an agonist-activated PLA₂, then clearly this is likely to be a member of the cPLA₂ group which includes both Ca²⁺-dependent and Ca²⁺-independent enzymes (32). The nature of the PLA₂ in nasal gland cells is, however, unknown, and the precise mechanism by which it is activated is unclear. The fact that it does not depend secondarily on the agonist-induced increase in [Ca²⁺]_i suggests that it may be coupled fairly directly to receptor activation, and as such it is possible that the activation involves a distinct PLA₂-coupled G protein. However, we have shown previously that the continued oscillatory signal requires the combined action of increases in InsP₃ and Ca²⁺ entry, and we know that the activation of PLC in these cells involves a G_q-type G protein (33). Hence, one way in which such a combined response could be coordinated would be for both actions to be mediated through the same receptor-G protein complex, perhaps via independent actions of the G and G subunits. Such a putative mechanism is not without precedence as G subunits have been shown to have several direct effects in various systems (34), including the activation of PLA₂ (35, 36). As noted above, the isotetrandrine-induced inhibition of PLA₂ is believed to result from the uncoupling of PLA₂ from an activating G protein rather than any direct effects on the enzyme itself (25); and in at least one case (pancreatic acinar cells), it has been proposed that this activation involves the subunits of the PLC-activating G protein G_q (24). In this context, it is particularly intriguing that the PLAP that we have shown to be able to reverse the isotetrandrine-induced inhibition of oscillations shows marked sequence similarity to the  subunits of G proteins (37).

Finally, our studies have shown that the biscoclaurine alkaloid isotetrandrine is an effective and readily reversible inhibitor of the noncapacitative entry of Ca²⁺ seen during agonist-induced [Ca²⁺]_i oscillations, but it is without effect on the capacitative Ca²⁺ entry induced by high concentrations of an agonist or by thapsigargin. Significantly, this further confirms the existence of a distinct noncapacitative mechanism of activation of agonist-enhanced Ca²⁺ entry during [Ca²⁺]_i oscillations and provides a new means of specifically blocking this pathway. However, it must be emphasized that all indications suggest that isotetrandrine is acting by interfering with the mechanism of activation of the entry pathway and not with the pathway itself. It remains possible, therefore, that a single entry pathway may be activated by both capacitative and noncapacitative mechanisms. In this context it is interesting that it has been suggested that capacitative activation of Ca²⁺ entry may involve a GTP-dependent step requiring the hydrolysis of GTP to GDP such as might be mediated by a so-called small G protein (38, 39). The reported participation of arachidonic acid in the regulation of such small G proteins (15, 40) may indicate that they provide the common link for both capacitative and noncapacitative mechanisms of agonist-induced activation of Ca²⁺ entry.