Muscarinic Receptor Activation of Arachidonate-mediated Ca 2 1 Entry in HEK293 Cells Is Independent of Phospholipase C*

Receptor-enhanced entry of Ca 2 1 in non-excitable cells is generally ascribed to a capacitative mechanism in which the activation of the entry pathway is specifically dependent on the emptying of agonist-sensitive intracellular Ca 2 1 stores. Although such entry can be clearly demonstrated under conditions of maximal or near-maximal stimulation, it is uncertain whether such a mechanism can operate during the oscillatory [Ca 2 1 ] i signals that are frequently seen following stimulation with low concentrations of agonists. In this study, we report that the stimulation of human m3 muscarinic receptors stably transfected into HEK293 cells results in the appearance of a novel arachidonate-mediated Ca 2 1 entry pathway. We show that the generation of arachidonic acid and the activation of this pathway are specifically associated with stimulation at the low agonist concentrations that typically give rise to oscillatory [Ca 2 1 ] i signals. At such agonist concentrations, how-ever, the generation of arachidonic acid is independent of the simultaneous activation of the phospholipase C-inositol 1,4,5-trisphosphate pathway. We further show that the arachidonate-mediated Ca 2 1 entry demon-strates characteristics that distinguish it from the corresponding capacitative pathway in the same cells and therefore is likely to represent an entirely distinct pathway that is specifically responsible for the receptor-enhanced entry of Ca 2 1 during [Ca 2 1 ] i oscillations. Calcium signaling in non-excitable cells is composed of two components: a release of calcium from intracellular stores and an increased entry of calcium from the extracellular medium. The of inositol (InsP

Receptor-enhanced entry of Ca 2؉ in non-excitable cells is generally ascribed to a capacitative mechanism in which the activation of the entry pathway is specifically dependent on the emptying of agonist-sensitive intracellular Ca 2؉ stores. Although such entry can be clearly demonstrated under conditions of maximal or near-maximal stimulation, it is uncertain whether such a mechanism can operate during the oscillatory [Ca 2؉ ] i signals that are frequently seen following stimulation with low concentrations of agonists. In this study, we report that the stimulation of human m3 muscarinic receptors stably transfected into HEK293 cells results in the appearance of a novel arachidonate-mediated Ca 2؉ entry pathway. We show that the generation of arachidonic acid and the activation of this pathway are specifically associated with stimulation at the low agonist concentrations that typically give rise to oscillatory [Ca 2؉ ] i signals. At such agonist concentrations, however, the generation of arachidonic acid is independent of the simultaneous activation of the phospholipase Cinositol 1,4,5-trisphosphate pathway. We further show that the arachidonate-mediated Ca 2؉ entry demonstrates characteristics that distinguish it from the corresponding capacitative pathway in the same cells and therefore is likely to represent an entirely distinct pathway that is specifically responsible for the receptorenhanced entry of Ca 2؉ during [Ca 2؉ ] i oscillations.
Calcium signaling in non-excitable cells is composed of two components: a release of calcium from intracellular stores and an increased entry of calcium from the extracellular medium. The role of inositol 1,4,5-trisphosphate (InsP 3 ), 1 generated as a result of the receptor activation of phospholipase C, in the release of calcium from specific intracellular stores is well established, but the nature of the calcium entry pathway and its regulation is far from clear. To date, discussion of such calcium entry has generally focused on the so-called "capacitative model," in which calcium entry is activated as a direct consequence of the emptying of the intracellular calcium stores and is independent of how this emptying is actually achieved (1,2). The precise nature of the mechanism for the activation of capacitative entry is, as yet, unclear, but it may involve the release and/or generation of a diffusible signaling molecule within the cell that activates the plasma membrane channels ("store-operated channels") responsible for calcium entry. Alternatively, a more direct molecular coupling between the stores and the plasma membrane channels may occur (3). Although such capacitative entry can be clearly demonstrated in a wide variety of different cells, it is far from certain that such a mechanism is the only one involved in the increase in calcium entry in non-excitable cells following receptor activation (4,5). For example, many cells show an oscillatory [Ca 2ϩ ] i signal when stimulated at low agonist concentrations (6,7), and such signals are associated with an enhanced entry of Ca 2ϩ . However, evidence indicates that the activation of capacitative Ca 2ϩ entry generally requires significantly higher levels of agonist-generated InsP 3 than does the release of Ca 2ϩ from the bulk of the agonist-sensitive internal stores (i.e. at low concentrations of InsP 3 , substantial release of Ca 2ϩ from agonistsensitive stores can occur without any activation of capacitative entry) (8 -10). Consequently, it is far from clear that the transitory (and/or incomplete) nature of calcium store depletion during [Ca 2ϩ ] i oscillations would provide an adequate or appropriate signal for the activation of calcium entry via a capacitative mechanism. Such considerations led us previously to investigate the nature of receptor-activated increases in Ca 2ϩ entry during [Ca 2ϩ ] i oscillations in cells from the exocrine avian nasal gland. In these studies, we showed that such Ca 2ϩ entry was non-capacitative in nature (11) and appeared to involve a novel arachidonate-activated pathway (12).
In the experiments reported here, we extend our earlier findings to another, more widely used and functionally less highly specialized cell type, namely HEK293 cells. The specific cell line chosen had been stably transfected with the human m3 muscarinic receptor (m3-mAChR), thereby avoiding possible complications resulting from the presence of multiple muscarinic receptor subtypes. Using this cell line, we were able to show that the receptor-mediated generation of arachidonic acid is independent of the simultaneous activation of the PLC-InsP 3 pathway, indicating that the m3-mAChR is capable of coupling both to the activation of PLC and the generation of arachidonic acid in a separate but parallel manner. Furthermore, we show that the arachidonate-mediated calcium entry pathway demonstrates characteristics that distinguish it from the more well known capacitative or store-operated entry of calcium. MATERIALS  single individual cells were determined following loading with the fluorescent probe indo-1. Loading was achieved by incubation in 4 M indo-1/AM for 12 min, followed by washing (three times) in saline. Cells were then incubated for a further 30 min at 37°C to allow for complete hydrolysis of the acetoxymethyl ester. [Ca 2ϩ ] i was measured as the ratio of the emitted fluorescence, measured as photon counts, at 405 and 485 nm using excitation at 350 nm as described previously (14,15). The technique utilized for the simultaneous determination of changes in [Ca 2ϩ ] i and Mn 2ϩ quench was as described previously (15). In the experiments designed to eliminate the effects of increases in [Ca 2ϩ ] i , cells were loaded with the Ca 2ϩ chelator BAPTA by preincubation in saline containing 15 M BAPTA/AM for 20 min. Preliminary experiments showed that this was sufficient to completely abolish any detectable [Ca 2ϩ ] i signals in cells subsequently stimulated with concentrations of carbachol up to 5 M.
Determinations of arachidonic acid release were essentially as described previously (12). Briefly, m3-HEK cells were cultured in six-well plates until ϳ50% confluent. 0.5 Ci/ml [ 3 H]arachidonic acid was then added to each well, and culture was continued overnight. The cells were then washed three times with saline containing 0.2% fatty acid-free bovine serum albumin prior to addition of agonist and/or drugs as appropriate. Arachidonic acid released during a 5-min incubation was determined by liquid scintillation counting of the supernatant and normalized to the total counts in the cells, determined following solubilization with 1.5 M NaOH. Total inositol phosphates were determined in a similar manner. Cells cultured in six-well plates were incubated overnight in the presence of 5 Ci/ml myo-[ 3 H]inositol, followed by washing three times in saline. To enhance detection of inositol phosphate generation and turnover, experiments were performed in the presence of lithium (10 mM) following a 10-min preincubation in the same concentration of lithium. At the end of the experimental period (5 min), the saline was removed and replaced with 1 ml of 0.5 M trichloroacetic acid. After incubation for 15 min on ice, the trichloroacetic acid-soluble fraction was extracted with ether and neutralized with sodium bicarbonate, and the extract was applied to Dowex columns. The loaded columns were washed with water, followed by 60 mM sodium formate plus 5 mM borax, before eluting the inositol phosphates with 1.2 M ammonium formate in 100 mM formic acid. Samples of the eluted inositol phosphates were counted by liquid scintillation. Total inositol phosphate generation is expressed as percent total counts in the phosphoinositide pool, determined from counting samples of the trichloroacetic acid-insoluble tissue fractions following their digestion overnight with 1.5 M NaOH.

RESULTS
m3-mAChRs Couple to Arachidonic Acid Generation to Activate Calcium Entry-We first determined whether activation of the transfected m3 muscarinic receptor in HEK cells was coupled to the generation of arachidonic acid, as described previously for the native muscarinic receptor in avian nasal gland cells (12). Fig. 1 shows that addition of the muscarinic receptor agonist carbachol to cells prelabeled with [ 3 H]arachidonic acid resulted in a marked increase in arachidonic acid generation. This increase was dependent on agonist concentration and was clearly apparent even at agonist concentrations close to the threshold for inducing detectable [Ca 2ϩ ] i signals in these cells (ϳ1 M carbachol). Concentrations of carbachol Ͼ5 M were not examined as, based on our earlier studies (12), we believed that the receptor activation of arachidonic acid generation and the associated increase in calcium entry were specifically associated with stimulation at low agonist concentrations, when, typically, oscillatory [Ca 2ϩ ] i signals are seen.
To examine the effect of arachidonic acid on [Ca 2ϩ ] i signaling in m3-HEK cells, low concentrations of exogenous arachidonic acid were added to otherwise unstimulated cells. Addition of as little as 3-8 M arachidonic acid resulted, after a variable lag period of several tens of seconds, in a slow increase in [Ca 2ϩ ] i ( Fig. 2A). Such an increase in [Ca 2ϩ ] i could result from an increase in calcium entry from the extracellular medium or from release of calcium from intracellular stores. To distinguish between these two alternatives, lanthanum was used as a generic blocker of calcium entry channels. In the presence of 100 M La 3ϩ , addition of exogenous arachidonic acid failed to increase [Ca 2ϩ ] i , and only when the La 3ϩ was removed was an increase in [Ca 2ϩ ] i observed (Fig. 2B). The failure to observe any increase in [Ca 2ϩ ] i in the presence of La 3ϩ confirms that such an increase reflects an increase in Ca 2ϩ entry. These data also rule out the additional possibility that arachidonic acid was inducing an increase in [Ca 2ϩ ] i by inhibition of the plasma membrane calcium pump. At high concentrations (ϳ1 mM or higher), La 3ϩ is also known to block the plasma membrane Ca 2ϩ -ATPase, but such an effect would be expected to further increase [Ca 2ϩ ] i , not to block such an increase. To confirm that arachidonic acid was indeed activating calcium entry, simultaneous determinations of changes in [Ca 2ϩ ] i and Mn 2ϩ quench were performed. In these, extracellular Mn 2ϩ (0.2 mM) is used as a surrogate for Ca 2ϩ as, at low external concentrations, it readily passes through many kinds of Ca 2ϩ channels. On entering the cytosol, the Mn 2ϩ binds to and quenches the fluorescence of such probes as indo-1, and the rate of fluorescence quenching can then be used as an indirect measure of the rate of Mn 2ϩ (and hence Ca 2ϩ ) entry, at least through those pathways permeable to Mn 2ϩ . As shown in Fig. 2C, the arachidonate-induced increase in [Ca 2ϩ ] i was associated with a simultaneous increase in the rate of Mn 2ϩ quench, indicating that Ca 2ϩ entry is increased.
To characterize the effects of the carbachol-induced generation of arachidonic acid and the consequent increase in Ca 2ϩ entry on [Ca 2ϩ ] i signals in m3-HEK cells, we examined the effect of the biscoclaurine alkaloid isotetrandrine. This substance acts as an apparently specific and readily reversible inhibitor of agonist-induced arachidonic acid generation (16,17), as demonstrated in our earlier studies on the avian nasal gland (12). Consistent with those studies, isotetrandrine (10 M) had no effect on carbachol-induced inositol phosphate generation in m3-HEK cells, but completely blocked the simultaneous carbachol-induced generation of arachidonic acid (Fig.  3). Addition of 10 M isotetrandrine to cells demonstrating an oscillatory [Ca 2ϩ ] i signal in response to low concentrations of carbachol resulted in an immediate cessation of the oscillations, which subsequently rapidly returned with normal amplitude and frequency on removal of the isotetrandrine (Fig. 4A). This is identical to the effect seen in these cells if Ca 2ϩ entry was inhibited during an oscillatory [Ca 2ϩ ] i response. For example, addition of La 3ϩ to a cell during an oscillatory [Ca 2ϩ ] i response to carbachol resulted in the immediate cessation of the oscillations, which promptly returned on removal of the La 3ϩ (Fig. 4B). To confirm that exposure to isotetrandrine during oscillations was indeed inhibiting Ca 2ϩ entry, we again used the Mn 2ϩ quench approach. As previously reported by us (11,14) and by others (18,19), oscillations in [Ca 2ϩ ] i were associated with a constant enhanced rate of Mn 2ϩ quench, reflecting a constant Ca 2ϩ entry. Fig. 4C shows that addition of isotetrandrine induced an immediate inhibition of this rate of Mn 2ϩ quench, which was rapidly restored on removal of the isotetrandrine. The observed effects on Mn 2ϩ quench (and hence, presumably Ca 2ϩ entry) were temporally directly associated with the inhibition of the [Ca 2ϩ ] i oscillations. It should be noted that the acute sensitivity to inhibition of Ca 2ϩ entry during oscillatory [Ca 2ϩ ] i signals seen here is not a universal feature in all cell types, some of which will continue to oscillate for some time even after complete removal of extracellular Ca 2ϩ (e.g. Xenopus oocytes). Such variation most likely reflects differences in surface-to-volume ratios and in the relative activities of the Ca 2ϩ pumps on the plasma membrane and on the intracellular stores (as well as levels of cytosolic buffers, etc.). However, such differences should not be interpreted as indicating that Ca 2ϩ entry plays no consistent or critical role in oscillatory [Ca 2ϩ ] i signals as, even in Xenopus oocytes, changes in Ca 2ϩ entry profoundly influence the frequency of agonistinduced [Ca 2ϩ ] i oscillations (20).
In intact cells, arachidonic acid frequently undergoes rapid oxidation, resulting in the production of a variety of eicosanoids that are, themselves, known to possess signaling activities. To determine whether the observed effects were due to arachidonic acid itself or to one of the many products of arachidonic acid metabolism, we examined the effects of inhibition of the lipoxygenase pathway, using nordihydroguaiaretic acid, and the cyclooxygenase pathway, using indomethacin. Addition of either of these agents to carbachol-stimulated cells showing oscillating [Ca 2ϩ ] i signals caused an increase in oscillation frequency, which often progressively developed into a sustained elevation of [Ca 2ϩ ] i (Fig. 5, A and B). If the activation of Ca 2ϩ entry observed was dependent on the metabolism of arachidonic acid by these pathways, then their inhibition would be expected to result in the reduction or cessation of the [Ca 2ϩ ] i oscillations, much as seen above when arachidonic acid generation was inhibited by isotetrandrine (Fig. 4). As such, the responses observed are clearly incompatible with the proposal that products of these metabolic pathways for arachidonic acid are responsible for the observed increases in Ca 2ϩ entry during oscillations. The effects seen are, however, consistent with a progressive increase in arachidonic acid levels in stimulated cells subsequent to inhibition of the normal metabolic pathways responsible for its degradation. Such increases might be expected to further increase Ca 2ϩ entry, resulting in a sustained elevation of [Ca 2ϩ ] i , as observed. The direct effect of arachidonic acid was further confirmed in experiments showing that addition of the non-metabolizable arachidonic acid analog ETYA to unstimulated cells caused an increase in [Ca 2ϩ ] i similar to that seen with arachidonic acid (Fig. 5C). In these experiments, it was noted that the effects of ETYA were frequently more rapid in onset than those seen with arachidonic acid, but required somewhat higher concentrations (ϳ20 versus ϳ5 M). Such data are consistent with the response to exogenously added arachidonic acid being muted by metabolism in the intact cell, yet showing some selectivity for arachidonic acid over its analog ETYA. Together, the above data indicate that it is arachidonic acid itself that is the active moiety responsible for the observed effects on Ca 2ϩ entry. They also suggest that both lipoxygenase and cyclooxygenase pathways are active in HEK293 cells and, in vivo, act to limit increases in receptor-stimulated levels of arachidonic acid.
mAChR Activation Stimulates Arachidonic Acid Generation Independently of PLC Activation-The m3-mAChR is known to couple, via a member of the G q family of G proteins, to the activation of phospholipase C and the generation of diacylglycerol and inositol 1,4,5-trisphosphate. The former activates protein kinase C, whereas the latter is integral to the initiation of [Ca 2ϩ ] i signals. To determine whether the observed increased generation of arachidonic acid was a downstream effect of the simultaneous activation of this PLC pathway, various approaches were employed. The possible involvement of InsP 3mediated increases in [Ca 2ϩ ] i was examined by determining arachidonic acid generation in cells loaded with the calcium chelator BAPTA (as described under "Materials and Methods") prior to stimulation with carbachol. As shown in Fig. 6, the carbachol-induced generation of arachidonic acid was completely unimpaired in such BAPTA-loaded cells, indicating that this effect was not dependent on the receptor-activated [Ca 2ϩ ] i signals.
We next examined the possible involvement of the PLC- More significantly, pretreatment with PMA completely abolished the normal mAChR-induced arachidonic acid release seen following addition of carbachol (Fig. 7). These data indicate that the principal effect of PKC activation appears to be to uncouple the stimulation of arachidonic acid release by mAChRs. Although the precise basis for this effect is unknown, these data clearly indicate that the activation of PKC does not result in the stimulation of arachidonic acid generation. Based on the above data from BAPTA-loaded cells and PMA-treated cells, it seems that neither the PLC-mediated increase in [Ca 2ϩ ] i nor the activation of PKC is involved in the observed activation of arachidonic acid release. Nevertheless, we were concerned that the above experiments may not entirely eliminate a potential involvement of the PLC pathway. For example, it has been reported in several cell types that the stimulation of PKC can exert a negative feedback on receptor-activated PLC activity. It was therefore possible that the failure of PMA to stimulate arachidonic acid release may have simply reflected the concomitant inhibition of PLC activity. Experiments on m3-HEK cells confirmed that PMA (0.1 M with preincubation for 5 min) did indeed significantly inhibit the carbachol-induced increase in inositol phosphate generation, an effect consistent with a reduction of agonist-activated PLC activity (data not shown). Alternatively, it is possible that the mAChR activation of arachidonic acid generation requires an increase in both [Ca 2ϩ ] i and PKC activity. We therefore carried out an additional series of experiments examining the effect of the drug U73122 (21) on mAChR activation of arachidonic acid generation. Although not always entirely specific, U73122 has been extensively used as an inhibitor of PLC activity, and consistent with this, we found that in the presence of 10 M U73122, carbachol-induced inositol phosphate generation was completely abolished (Fig. 8A). However, under identical conditions, carbachol-induced arachidonic acid generation was unaffected (Fig. 8B). These data confirm that the observed receptor-activated arachidonic acid generation is not a downstream effect of any PLC activity.
Arachidonate-mediated Calcium Entry Is Distinct from Capacitative Entry-As discussed above, the activation of the capacitative mechanism of Ca 2ϩ entry is solely dependent on the emptying of intracellular stores of calcium no matter how this is achieved. However, under normal circumstances of receptor stimulation, the discharge of the intracellular Ca 2ϩ stores occurs as a result of the PLC-mediated generation of inositol 1,4,5-trisphosphate, which activates Ca 2ϩ -permeable channels on the stores. The demonstration that mAChR activation of arachidonic acid generation is not a downstream effect of the simultaneous activation of PLC (see above) sug-

FIG. 8. Effect of U73122 on carbachol-induced inositol phosphate generation (A) and arachidonic acid release (B).
Total inositol phosphate generation and arachidonic acid release were measured as described under "Materials and Methods." Prior to stimulation with carbachol (CCh) at the concentrations indicated, experimental cells were exposed to U73122 (10 M) for 10 min. Data are presented as the carbachol-induced increase in total inositol phosphate generation or arachidonic acid release over the corresponding controls in the absence (shaded columns) or presence (stippled columns) of U73122. Values are the means Ϯ S.E. (n ϭ 4 (arachidonic acid (AA)) and n ϭ 3 (total inositol phosphates (IPs))). PI, phosphoinositide. gests that the arachidonate-activated Ca 2ϩ entry is unlikely to be part of a capacitative mechanism. This is further supported by the fact that isotetrandrine, which, as shown, is an effective inhibitor of receptor-activated arachidonic acid generation, has no effect on the capacitative entry of Ca 2ϩ induced by thapsigargin (data not shown). Similar data were obtained in our earlier studies on avian nasal gland cells (12) and have been reported for the related compound tetrandrine in adrenal glomerulosa cells (22). However, this does not entirely preclude the possibility that arachidonic acid is involved in some way in the capacitative mechanism. For example, the complete depletion of the stores induced by thapsigargin may result in such a profound stimulation of the capacitative mechanism that isotetrandrine is not able to reverse it significantly. Alternatively, a single type of Ca 2ϩ entry channel (store-operated channel) may be capable of being activated by the capacitative signal, whatever its nature, or by arachidonic acid depending on the circumstances. To preclude these possibilities, we examined the effect of reducing extracellular pH on the Ca 2ϩ entry activated by thapsigargin and by arachidonic acid. Modest reductions in extracellular pH have been shown to have a profound inhibitory effect on capacitative Ca 2ϩ entry in a range of different cell types (23)(24)(25)(26), and in the case of the storeoperated channel I CRAC , this has been shown to reflect a direct action of extracellular protons on the channel (27). 2 As shown in Fig. 9A, reducing the pH of the superfusing saline to 6.7 induced a marked decrease in the sustained [Ca 2ϩ ] i seen in thapsigargin-treated cells, consistent with an inhibition of capacitative entry. This effect was readily reversed on restoration of normal extracellular pH. An identical reduction in extracellular pH was, however, completely without effect on the similar sustained [Ca 2ϩ ] i seen in cells exposed to exogenous arachidonic acid (Fig. 9B). This marked difference in the effect of extracellular pH indicates that the Ca 2ϩ entry channel activated by arachidonic acid is unlikely to be the same as that activated by thapsigargin-induced depletion of intracellular stores.

DISCUSSION
In the studies reported here, we have shown that activation of the m3-mAChR stably transfected in HEK293 cells results in the generation of arachidonic acid and that this specifically activates a Ca 2ϩ entry pathway that is critical to the regulation of the oscillatory [Ca 2ϩ ] i signals generated by low concentrations of appropriate agonists. Although detailed concentrationresponse curves for the generation of arachidonic acid were not determined, it is clear that this response shows a sensitivity to muscarinic agonists that is very similar to that demonstrated by the activation of PLC and is therefore consistent with a potential role in the regulation or generation of [Ca 2ϩ ] i signals. We have also shown that inhibition of the generation of arachidonic acid results in the immediate inhibition of receptor-enhanced Ca 2ϩ entry and the cessation of the oscillatory [Ca 2ϩ ] i response. The data further indicate that this is an effect of arachidonic acid itself and not of a product of the commonly recognized metabolic pathways for arachidonic acid. Regarding our data on the activation of Ca 2ϩ entry by exogenous application of arachidonic acid, such application has been reported to produce a variety of effects on [Ca 2ϩ ] i signals in a many different cells, including effects on Ca 2ϩ release from intracellular stores (28) as well as on Ca 2ϩ entry (29). However, most of these studies involved the use of very high concentrations of arachidonate (Ͼ50 M), raising the possibility of a variety of nonspecific effects. In contrast, the effects on Ca 2ϩ entry we have observed are seen at very low concentrations of exog-enously added arachidonic acid (typically ϳ5 M). Given that extracellular application via the general perfusing medium is probably only a crude mimic of the specific receptor-mediated intracellular generation of arachidonic acid, this is likely to represent a significant underestimate of the true sensitivity of the Ca 2ϩ entry pathway to this substance.
The data we have presented demonstrate the receptor-activated generation of arachidonic acid, together with potential pathways for its rapid removal. We have shown that low concentrations of arachidonic acid activate a Ca 2ϩ entry pathway and that inhibition of its generation produces a parallel inhibition of the receptor-activated entry of Ca 2ϩ seen during [Ca 2ϩ ] i oscillations. All of these features are consistent with the critical experimental criteria widely recognized as defining a substance as a second messenger and lead us to conclude that arachidonic acid is the intracellular signal responsible for the mAChR activation of Ca 2ϩ entry during [Ca 2ϩ ] i oscillations. We have previously reported a similar arachidonate-dependent entry of Ca 2ϩ during oscillatory [Ca 2ϩ ] i responses to muscarinic receptor activation in the exocrine cells of the avian nasal gland (12). The demonstration of this same signaling pathway in these very different cell types suggests that it is likely to have a widespread distribution. This is further supported by recent reports indicating the presence of an arachidonate-activated Ca 2ϩ entry that is involved in maintaining [Ca 2ϩ ] i oscil-2 R. Penner, personal communication. During the periods indicated (solid bars), extracellular pH was reduced to 6.7. B, indo-1-loaded cells were exposed to either 1 M thapsigargin (E) or 8 M arachidonic acid (q) to induce a sustained increase in [Ca 2ϩ ] i . At the point indicated (solid bar), extracellular pH was reduced to 6.7. lations induced by cholecystokinin B receptors transfected into Chinese hamster ovary cells (30). An arachidonate-activated Ca 2ϩ entry that is believed to be involved in basic fibroblast growth factor-induced responses in Balb/c 3T3 fibroblasts has also been recently described (31), and low concentrations of exogenous arachidonic acid have been shown to activate Ca 2ϩpermeable, voltage-insensitive cation channels in chromaffin cells (32). Direct regulation of ion channels by fatty acids, including arachidonic acid, has been shown for a variety of different channel types (33,34) and includes both stimulatory and inhibitory effects. More important, although the full characterization of the arachidonate-mediated pathway we have described must await further studies, we have shown that it possesses characteristics that distinguish it from the more well known capacitative or store-operated pathway. For example, in our previous study of this pathway in avian nasal gland cells (12), we showed that its activation was apparently independent of any InsP 3 -induced release of Ca 2ϩ from intracellular stores. In fact, the data indicated that, at the low agonist concentrations that typically give rise to oscillatory [Ca 2ϩ ] i signals, generated levels of InsP 3 alone were not adequate to induce any detectable release of Ca 2ϩ from the stores, and such release was absolutely dependent on the additional effect of the arachidonate-mediated entry of Ca 2ϩ . However, the possibility remained that the entry we observed was, in fact, dependent on a capacitative mechanism that was activated by the emptying of a small specific subset of the overall agonist-sensitive store whose discharge was undetectable in our experiments. Published evidence suggests this is unlikely because, as noted earlier, data from a variety of different cells indicate that if such a subset exists, its discharge would require higher concentrations of InsP 3 than those required to empty the bulk of the agonist-sensitive stores in the cell (8 -10). Nevertheless, in the study reported here, we have made an additional critical observation that supports our hypothesis that the arachidonate-mediated Ca 2ϩ entry is entirely distinct from the capacitative pathway activated as a result of the depletion of intracellular stores. We showed that, consistent with reports from various different cell types, such capacitative entry in HEK293 cells is acutely sensitive to reductions in extracellular pH. In marked contrast, the Ca 2ϩ entry activated by arachidonic acid in the same cells is entirely insensitive to such changes in extracellular pH. Given that the observed effects on the capacitative entry are reported to reflect a direct effect on the storeoperated Ca 2ϩ channel (at least in the case of I CRAC ), such a clear difference indicates that the arachidonate-activated pathway must be distinct from that activated by store depletion. This conclusion is further supported by our demonstration that the receptor-mediated increase in arachidonic acid generation we have observed is independent of the simultaneous activation of the PLC-InsP 3 pathway (see below). Clearly, such a finding is inconsistent with the idea that the muscarinic receptor stimulation of the arachidonate-mediated pathway is downstream of any InsP 3 -mediated release of stored Ca 2ϩ .
Data from a wide range of different cell types suggest that there are generally two principal sources of receptor-mediated increases in arachidonic acid generation. The first involves the action of diacylglycerol/monoacylglycerol lipases on diacylglycerol. This, in turn, is mainly derived from the hydrolysis of phosphatidylcholine as a result of receptor-induced increases in either PLD activity (via phosphatidic acid) (35,36) or PLC (37,38). In this context, it has previously been reported that the m3-mAChR stably transfected into HEK293 cells is capable of coupling to a PLD (39,40). However, several characteristics of this response in the m3-mAChR-transfected HEK293 cells indicate that this is unlikely to be the basis for the observed carbachol-induced increase in arachidonic acid generation. First, the activation of PLD by the m3-mAChR stably transfected into HEK293 cells has been shown to completely desensitize within 2 min of stimulation, even at the low concentrations of agonist used in this study (40). This desensitization is apparently not a result of the loss of cell-surface receptors, but reflects a rapid and sustained uncoupling of the receptors from the activation of PLD. Second, it has been shown that addition of the phorbol ester PMA induces a pronounced stimulation of PLD activity in m3-HEK cells (40,41) as well as in many other mammalian cells (42,43). Examination of the published data indicate that 0.1 M PMA increased PLD activity to levels ϳ4-fold higher than those seen with a maximal concentration of carbachol (40). This is clearly in marked contrast to our data showing that, at the same concentration, PMA had a marked inhibitory effect on the carbachol-induced arachidonic acid generation. Based on these findings, we conclude that the reported characteristics of the PLD response in HEK293 cells are not consistent with our findings on the mAChR-activated arachidonic acid generation and that the PLD-mediated generation of diacylglycerol is unlikely to account for the observed mAChRinduced arachidonic acid response.
The alternative mechanism for receptor-induced increases in arachidonic acid generation involves the action of a cytosolic PLA 2 releasing arachidonic acid from appropriate phospholipids. Two main types of cytosolic PLA 2 with distinct properties and structures are currently identified: the "classic" type IV Ca 2ϩ -dependent cPLA 2 and the more recently characterized type VI Ca 2ϩ -independent PLA 2 (iPLA 2 ) (44 -46). This latter group is thought to be principally involved in general membrane phospholipid remodeling (47), whereas it is the type IV Ca 2ϩ -dependent cPLA 2 that has usually been shown to be associated with the receptor activation of arachidonic acid release (48). In this respect, a particularly significant finding of the studies reported here is that the increase in arachidonic acid generation we have observed was shown to be independent of the simultaneous activation of PLC. Receptor activation of cPLA 2 has been generally reported to involve two distinct processes: a Ca 2ϩ -dependent translocation of the cPLA 2 to the membrane to allow interaction with its phospholipid substrate and a phosphorylation that is usually mediated via a mitogenactivated protein kinase, whose activity may, in turn, be modulated by PKC (48). Alternatively, PKC may itself directly phosphorylate the cPLA 2 to increase its activity. The relative importance of the Ca 2ϩ -dependent translocation and the phosphorylation steps in the activation of cPLA 2 appears to vary in different cell types and under different circumstances (48). In marked contrast, the studies we have presented here clearly show that neither an increase in [Ca 2ϩ ] i nor the activation of PKC (either individually or together) was required for the observed carbachol-induced generation of arachidonic acid. In this regard, it is perhaps important to point out that most studies investigating the receptor-activation of cPLA 2 have generally utilized maximal (or near-maximal) concentrations of the relevant agonist. Consequently, although the data presented do not exclude the presence of these activation pathways in HEK293 cells, it is clear that neither a Ca 2ϩ -dependent translocation nor a PKC-dependent phosphorylation step can play any significant role in the activation of arachidonic acid generation under the specific conditions employed, i.e. at low concentrations of muscarinic agonists. This was further supported by the data obtained using the drug U73122, which confirmed that the mAChR stimulation of arachidonic acid generation we observed is not downstream of PLC activation. Moreover, because the HEK293 cell line utilized here possesses only a single muscarinic receptor subtype (m3), the data show that this receptor is capable of coupling both to the activation of PLC and the generation of arachidonic acid in a separate but parallel manner. We therefore conclude that the two signaling pathways are independently activated by the m3-mAChR. Given the critical role that such arachidonic acid generation plays in the activation of the Ca 2ϩ entry required to drive receptor-activated oscillatory [Ca 2ϩ ] i signals in these and other cells, the identification of the mechanism responsible for coupling of muscarinic receptors to arachidonic acid generation at these physiologically relevant levels of stimulation is clearly of considerable importance.