Differential regulation of muscarinic acetylcholine receptor-sensitive polyphosphoinositide pools and consequences for signaling in human neuroblastoma cells.

In this study we have quantitatively assessed the basal turnover of phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2) and M3-muscarinic receptor-mediated changes in phosphoinositides in the human neuroblastoma cell line, SH-SY5Y. We demonstrate that the polyphosphoinositides represent a minor fraction of the total cellular phosphoinositide pool and that in addition to rapid, sustained increases in [3H]inositol phosphates dependent upon the extent of receptor activation by carbachol, there are equally rapid and sustained reductions in the levels of polyphosphoinositides. Compared with phosphatidylinositol 4-phosphate (PtdIns(4)P), PtdIns(4,5)P2 was reduced with less potency by carbachol and recovered faster following agonist removal suggesting protection of PtdIns(4,5)P2 at the expense of PtdIns(4)P and indicating specific regulatory mechanism(s). This does not involve a pertussis toxin-sensitive G-protein regulation of PtdIns(4)P 5-kinase. Using wortmannin to inhibit PtdIns 4-kinase activity, we demonstrate that the immediate consequence of blocking the supply of PtdIns(4)P (and therefore PtdIns(4,5)P2) is a failure of agonist-mediated phosphoinositide and Ca2+ signaling. The use of wortmannin also indicated that PtdIns is not a substrate for receptor-activated phospholipase C and that 15% of the basal level of PtdIns(4,5)P2 is in an agonist-insensitive pool. We estimate that the agonist-sensitive pool of PtdIns(4,5)P2 turns over every 5 s (0.23 fmol/cell/min) during sustained receptor activation by a maximally effective concentration of carbachol. Immediately following agonist addition, PtdIns(4,5)P2 is consumed >3 times faster (0.76 fmol/cell/min) than during sustained receptor activation which represents, therefore, utilization by a partially desensitized receptor. These data indicate that resynthesis of PtdIns(4,5)P2 is required to allow full early and sustained phases of receptor signaling. Despite the critical dependence of phosphoinositide and Ca2+ signaling on PtdIns(4,5)P2 resynthesis, we find no evidence that this rate resynthesis is limiting for agonist-mediated responses.

Phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P 2 ) 1 is a minor membrane-associated phospholipid that is a substrate for enzymes involved in important cellular signal transduction pathways (1). Thus, PtdIns(4,5)P 2 is a substrate for both phospholipase C (PLC) and phosphoinositide 3-kinase (PI3-K) activities. The importance of the signaling pathway initiated by PI3-K and resulting in the generation of phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P 3 ) is rapidly emerging (2,3). However, all estimates so far suggest that the PI3-K pathway utilizes only a small fraction of the PtdIns(4,5)P 2 pool compared with that hydrolyzed by PLC in a signaling cascade (4,5). The latter enzyme liberates both inositol 1,4,5-trisphosphate (Ins(1,4,5)P 3 ) and 1,2-diacylglycerol, which mobilize Ca 2ϩ from intracellular stores and activate several isoforms of protein kinase C, respectively (6). Both Ins(1,4,5)P 3 and diacylglycerol are recycled to provide the substrates (myo-inositol and CMPphosphatidate) necessary for phosphatidylinositol (PtdIns) resynthesis. PtdIns(4,5)P 2 can then be regenerated by the sequential phosphorylation of PtdIns and phosphatidylinositol 4-phosphate (PtdIns(4)P) by phosphatidylinositol 4-kinase (PtdIns 4-kinase) and phosphatidylinositol 4-phosphate 5-kinase (PtdIns(4)P 5-kinase), respectively (6). This resynthesis is vital in maintaining an agonist-sensitive PtdIns(4,5)P 2 pool, as the cellular content of PtdIns(4,5)P 2 is small in comparison to the rate at which it may be consumed during receptor-mediated activation of PLC (7). Although it is possible that changes in substrate/product concentrations, particularly the dramatic changes that might occur in the immediate vicinity of activated PLC, may play a role in regulating PtdIns(4,5)P 2 supply, the pathway is likely to possess more sophisticated regulatory features that enable supply to be matched to demand under agonist-stimulated conditions. In this context a number of potential regulatory mechanisms for PtdIns 4-kinase and/or PtdIns(4)P 5-kinase activities have been proposed (8 -15), although a true understanding of the mechanisms and roles of such regulation remains elusive (2). Furthermore, although the resupply of PtdIns(4,5)P 2 must occur to allow sustained or repetitive phosphoinositide signaling, there is little information to indicate the extent to which its resynthesis contributes to regulatory aspects of signaling mediated by PLC-coupled receptors. It is unclear, for example, whether resynthesis of PtdIns(4,5)P 2 is required during acute agonist-mediated responses and indeed whether depletion of this substrate can contribute to the rapid receptor desensitization that is an almost universal feature of receptors activating this signal transduction pathway (16,17).
Cells of the human SH-SY5Y neuroblastoma cell line have many features characteristic of fetal sympathetic ganglion cells and have been used extensively in studies from signal transduction to neurotransmitter release (18 -22). Our laboratory has used this neuroblastoma extensively to examine mechanistic and regulatory aspects of muscarinic receptor-mediated phosphoinositide and Ca 2ϩ signaling (19 -22), and these have proved to be representative of other PLC-linked receptor types in differing cellular backgrounds (16,17). In the current study we have, therefore, used SH-SY5Y cells to examine quantitatively the regulation of PtdIns(4,5)P 2 hydrolysis and resynthesis during stimulation of their muscarinic acetylcholine receptors which are predominantly of the M 3 subtype (23). In addition we have sought to assess the impact of reduced PtdIns-(4,5)P 2 supply on transmembrane signaling via PLC and to determine whether agonist-mediated depletion of PtdIns(4,5)P 2 contributes to the rapid desensitization of muscarinic receptormediated phosphoinositide signaling.

EXPERIMENTAL PROCEDURES
Cell Culture-Experiments were performed on SH-SY5Y cells (originally a gift from Dr. J. Biedler, Sloan-Kettering Institute, New York) between passages 70 and 90. Cells were maintained in minimum essential medium supplemented with 50 IU ml Ϫ1 penicillin, 50 g ml Ϫ1 streptomycin, 2.5 g ml Ϫ1 amphotericin B, 2 mM L-glutamine, and 10% (v/v) newborn calf serum. Cultures were maintained at 37°C in 5% CO 2  Buffer was then aspirated, and the cells were challenged with 200 l of buffer (Ϯ agonist). Where carbachol was removed to examine recovery, the agonist-containing buffer was aspirated, each well washed (2 ϫ 1 ml Krebs/HEPES), and the incubation continued in the presence of 200 l of buffer. In antagonist reversal experiments, agonist incubation was in 100 l followed by addition of 100 l of atropine (final concentration, 10 M). Reactions were terminated with an equal volume of ice-cold 1 M trichloroacetic acid. After 15 min on ice the aqueous phase was removed and 100 l of 10 mM EDTA added. After vortexing with 0.5 ml of a 1:1 (v/v) freshly prepared mixture of tri-n-octylamine and 1,1,2-trichlorotrifluoroethane, 20 l of 250 mM NaHCO 3 was added to a 300-l aliquot of the aqueous phase. This was applied to a Dowex (AG1-X8) formate column which was then washed with 20 ml of water and 10 ml of 25  H]GroPIns(4,5)P 2 ) as indices of PtdIns, PtdIns(4)P, and PtdIns(4,5)P 2 , respectively, were prepared from cell monolayers based upon previously described methods (24). After removal of the acidified aqueous phase for the determination of [ 3 H]InsP x as described above, lipids were extracted into 0.94 ml of acidified chloroform/methanol (40:80:1 v/v, 10 M HCl). Chloroform (0.31 ml) and 0.1 M HCl (0.56 ml) were then added to induce phase partition. A sample of the lower phase (400 l) was removed, dried in a stream of N 2 , and stored at Ϫ20°C prior to further processing. These samples were dissolved in 1 ml of chloroform and 0.2 ml of methanol and hydrolyzed by addition of 0.4 ml of 0.5 M NaOH in methanol/water (19:1, v/v). Samples were vortex mixed at regular intervals during a 20-min incubation at room temperature. Chloroform (1 ml), methanol (0.6 ml), and water (0.6 ml) were then added, and the samples were mixed and centrifuged (3,000 ϫ g, 10 min). A 1-ml aliquot of the upper phase was neutralized using 1-ml bed volume Dowex-50 (H ϩ form) columns that were washed with 2 ϫ 2 ml of water. The pooled eluate was brought to pH 7 by addition of NaHCO 3 and applied to a Dowex (AG1-X8) formate anion exchange column. The [ 3 H]GroPIns, [ 3 H]GroPIns(4)P, and [ 3 H]GroPIns(4,5)P 2 were then eluted as described elsewhere (24) and quantified by liquid scintillation spectrometry.
Measurement of PtdIns(4,5)P 2 Mass-PtdIns(4,5)P 2 mass was determined by assay of Ins(1,4,5)P 3 released by alkaline hydrolysis following a previously described protocol (25). Briefly, dried lipid extracts, prepared as described above from cells not labeled with [ 3 H]inositol, were dissolved in 0.25 ml of 1 M KOH and heated to 100°C for 15 min during which time they were vortex mixed at regular intervals. Tubes were then placed on ice for 15 min and then samples added to 0.5-ml bed volume Dowex-50 (H ϩ form) columns. Columns were washed (3 ϫ 0.25 ml) with water. NaHCO 3 (100 l, 60 mM) and EDTA (100 l, 30 mM) were then added to the pooled column eluates that were stored at 4°C. The Ins(1,4,5)P 3 which had been released from the PtdIns(4,5)P 2 was assayed as described below within 48 h. Recoveries from each processing step were assessed (25) to allow levels of Ins(1,4,5)P 3 determined to be extrapolated to the amount of PtdIns(4,5)P 2 .
Generation and Measurement of Ins(1,4,5)P 3 -Cell monolayers were preincubated in Krebs/HEPES, challenged, and the reaction terminated as described above. Experiments examining recovery following termination of carbachol action by addition of atropine were performed as described above. A series of experiments were also designed to examine the potential desensitization and resensitization of the peak (10 s) Ins(1,4,5)P 3 response. Cells were treated as described above for experiments examining the recovery of inositol phosphates and phosphoinositides. However, following aspiration of the initial carbachol challenge and washing of the monolayer (2 ϫ with 1 ml Krebs/HEPES), incubation was continued for the required recovery time in 1 ml of buffer before aspiration and rechallenge with carbachol (200 l) for 10 s. Reactions were again stopped by the addition of an equal volume of 1 M trichloroacetic acid. A 160-l aliquot of the acidified aqueous phase was removed, processed, and assayed for Ins(1,4,5)P 3 by a radioreceptor assay (26).

Measurement of Intracellular [Ca 2ϩ ]-The intracellular [Ca 2ϩ ] ([Ca 2ϩ
] i ) was determined in suspensions of fura-2-loaded cells. Briefly, confluent cells were harvested, washed with Krebs/HEPES, and resuspended in 2.5 ml of the same buffer. A 0.5-ml aliquot was removed and manipulated as below but with the exclusion of the acetoxymethyl ester of fura-2 (fura-2-AM) thereby allowing the determination of cellular autofluorescence. Fura-2-AM was added to the remaining cells at 5 M and the cells left with gentle stirring for 40 -60 min at room temperature. Supernatant containing extracellular fura-2-AM was removed following gentle centrifugation of 0.5-ml aliquots. Cells were resuspended in 1 ml of buffer and incubated at 37°C for 10 min prior to further centrifugation and resuspension in 3 ml of buffer at 37°C. With emission at 509 nm, the 340/380 nm excitation ratio was recorded every 3.8 s as an index of [Ca 2ϩ ] i . Cells were challenged by the addition of 50 l of carbachol to give a final concentration of 1 mM. The 340:380 ratio was converted to [Ca 2ϩ ] i as reported previously (27) using 0.1% Triton X-100 in the presence of 1.3 mM Ca 2ϩ to determine R max followed by the addition of 6.7 mM EGTA to determine R min .
Effects of Li ϩ on Carbachol-mediated Phosphoinositide Signaling-The effects of Li ϩ on muscarinic receptor-mediated phosphoinositide signaling was determined as described above with the exception that the preincubation and incubation buffers were inositol-free and contained 10 mM Li ϩ .
Effects of Wortmannin on Carbachol-mediated Phosphoinositide Signaling-Although wortmannin is better known for its ability to inhibit PI3-K (28), this fungal metabolite has recently been demonstrated to inhibit some isoforms of PtdIns 4-kinase (29,30). We have, therefore, used wortmannin in an alternative strategy to Li ϩ to examine the immediate consequences of blocking the provision of PtdIns(4)P and PtdIns(4,5)P 2 under basal and agonist-stimulated conditions. For experiments in which cells were pretreated with wortmannin, this was added during the final 10 min of incubation and incorporated with agonist additions. In experiments investigating the time course of the effects of wortmannin on inositol phosphates and phosphoinositides, wortmannin was added in a 10-l volume.
Materials-Reagents of analytical grade were obtained from suppliers listed previously (20,21,24 (1,4,5)P 3 were averaged to give a single value representative of one experiment, whereas for [ 3 H]phosphoinositides and PtdIns(4,5)P 2 mass, duplicates were pooled at the lipid extraction stage. All data are presented as mean Ϯ S.E. with the number of experiments given in parentheses. Concentration-response curves were fitted by GraphPad Prism (GraphPad Software, Inc., San Diego, CA) using a standard four-parameter logistic equation. EC 50 and IC 50 values are presented as log 10 M. Statistical comparisons were by Student's two-tailed paired or unpaired t test or, where multiple comparisons were required, by one-way analysis of variance followed by Duncan's multiple range test at p Ͻ 0.05 and p Ͻ 0.01.

Agonist-stimulated Changes in Polyphosphoinositides
Incubation of SH-SY5Y human neuroblastoma cells with [ 3 H]inositol resulted in marked changes in the absolute and relative amounts of radioactivity incorporated into the phosphoinositides over the first 20 h (data not shown). However, there were little or no differences between cells labeled for either 44 or 48 h under basal or agonist-stimulated (1 mM carbachol, 5 min) conditions (data not shown), and the phosphoinositide pools were judged to be in equilibrium. In all subsequent experiments, cells were therefore labeled for 48 h and under these conditions [ 3 H]PtdIns comprised the major fraction (94.3 Ϯ 0.3% (n ϭ 4)) of the inositol phospholipid pool, whereas [ 3 H]PtdIns(4)P and [ 3 H]PtdIns(4,5)P 2 represented only minor fractions (2.5 Ϯ 0.1% (n ϭ 4) and 3.2 Ϯ 0.2% (n ϭ 4) respectively). Upon addition of a concentration of carbachol that is maximal for the muscarinic receptor-mediated phosphoinositide-linked responses in these cells (1 mM), there was a rapid and marked accumulation of [ 3 H]InsP x (Fig. 1a). We emphasize that this accumulation of [ 3 H]InsP x is in the absence of a Li ϩ block of inositol monophosphatase activity. Accumulation therefore represents the net result of both generation and metabolism to free inositol. Over the first 60 s of stimulation with carbachol there was an increase in [ 3 H]InsP x to 291 Ϯ 23% (n ϭ 4) of basal levels that was sustained throughout the remaining period of agonist stimulation. In the same cell monolayers, there were rapid and marked decreases in the levels of [ 3 H]PtdIns(4)P and [ 3 H]PtdIns(4,5)P 2 upon agonist challenge to 32.4 Ϯ 2.9% (n ϭ 4) and 24.7 Ϯ 2.8% (n ϭ 4) of basal levels, respectively, by 60 s (Fig. 1b). These reductions were again sustained throughout the period of agonist challenge. In contrast, the level of [ 3 H]PtdIns decreased to only 85.8 Ϯ 4.2% (n ϭ 4) of basal levels over the experimental period (900 s) (Fig. 1b). Given this relatively small reduction of [ 3 H]Pt-dIns, it is unlikely that the specific activities of the 3 H-polyphosphoinositides change greatly during this period of stimulation, and therefore the changes in radioactivity are likely to accurately reflect the changes in mass.
The ability of carbachol to mediate the accumulation of [ 3 H]InsP x and depletion of the phosphoinositides was concentration-related (Fig. 2). EC 50 values (log 10 M, determined at 60 s following agonist addition) were Ϫ5. 43
Reversal of carbachol-mediated changes occurring after washing the cells indicated a dependence of the sustained alterations on persistent muscarinic receptor occupation. This was confirmed by the ability of the muscarinic antagonist, atropine, to reverse these changes. Addition of 10 M atropine, 1 min following addition of a maximally effective concentration of carbachol (1 mM) reduced [ 3 H]InsP x accumulation from 378 Ϯ 20% (n ϭ 4) to 229 Ϯ 11% (n ϭ 4) of basal over a 100-s period. During this time [ 3 H]PtdIns(4,5)P 2 levels were restored from 23.6 Ϯ 2.4% (n ϭ 4) to 81.8 Ϯ 4.1% (n ϭ 4) of basal. However, the rate and extent of recovery of [ 3 H]PtdIns(4)P was less, returning from 25.7 Ϯ 4.4% (n ϭ 4) to 49.5 Ϯ 5.4% (n ϭ 4) of basal over the same period. The addition of atropine during challenge with carbachol had no effect upon the small agonistinduced reduction of [ 3 H]PtdIns (data not shown). That changes in [ 3 H]inositol-labeled phospholipids reflect changes in mass levels was further indicated by the measurement of PtdIns(4,5)P 2 mass during carbachol stimulation and following addition of atropine. Thus, challenge with carbachol for 60 s resulted in an approximately 80% reduction in the mass of PtdIns(4,5)P 2 which was approximately restored within 100 s of atropine addition (Table I). These experiments also allowed the measurement of Ins(1,4,5)P 3 mass from the same cells. This demonstrated the characteristic "peak and plateau" response characteristic of muscarinic receptor activation in these cells (19 -22) and the return of Ins(1,4,5)P 3 levels to basal by 100 s following atropine addition (Table I,  Extending the period of exposure to carbachol from 1 min (as above) to 5 min before washing had no effect on the rate or extent of recovery of either the [ 3 H]inositol-labeled inositol phosphates or phosphoinositides (data not shown and Fig. 8).

Effect of Pertussis Toxin on the Extent and Rate of Recovery of Inositol Phosphates and Phosphoinositides
Pretreatment of cells for 24 h with 100 ng/ml pertussis toxin had no effect on either the extent of depletion or the rate and extent of recovery of the [ 3 H]inositol-labeled inositol phosphates or phosphoinositides following 5 min exposure to 1 mM carbachol ( Fig. 4 and data not shown).

Effects of Li ϩ on Agonist-stimulated Inositol Phosphate and Phosphoinositide Levels
Incubation of cells with carbachol in the presence of 10 mM Li ϩ resulted in an accumulation of [ 3 H]InsP x that was rapid over the 1st min of stimulation (increase of 90% of the basal value which was 13,877 Ϯ 1495 (n ϭ 4) dpm/well). Following this there was a slower but sustained accumulation over a further 19 min of carbachol stimulation (increase of 50% of basal/min). The sustained linear increase between 1 and 20 min of stimulation demonstrate an effective block of inositol monophosphatase activity that will prevent the recycling of inositol back into the phospholipids. Despite this the carbacholmediated depletion of the phospholipids was identical in the presence and absence of Li ϩ (data not shown). We were, therefore, unable to use Li ϩ to manipulate cellular levels of the phosphoinositides, and an alternative strategy using inhibitors of PtdIns 4-kinase (wortmannin and LY-294002) was employed.

Effects of Wortmannin and LY-294002 on Basal and Agonist-stimulated Inositol Phosphate and Polyphosphoinositide Levels
Basal Levels-Addition of 10 M wortmannin to unstimulated cells produced a decrease in [ 3 H]InsP x over a 10-min time course to 81.6 Ϯ 20.8% (n ϭ 3) of initial values (Fig. 5a). Over this time frame there was no effect of wortmannin on [ 3 H]Pt-dIns levels (data not shown), but [ 3 H]PtdIns(4,5)P 2 fell to 84.9 Ϯ 7.3% (n ϭ 3) of basal levels (Fig. 5c). In contrast there was a more dramatic decrease in [ 3 H]PtdIns(4)P to 25.4 Ϯ 1.1% (n ϭ 3) of basal levels (Fig. 5b). The ability of wortmannin to  induce these changes over a 10-min period was concentrationrelated ( Fig. 6, a and  Agonist-stimulated Levels-When wortmannin was added 60 s after 1 mM carbachol there was an immediate reduction in the agonist-mediated accumulation of [ 3 H]InsP x which approached unstimulated levels by 10 min (Fig. 5a), whereas the carbachol-mediated decreases in [ 3 H]PtdIns(4)P and [ 3 H]Ptd-Ins(4,5)P 2 levels were further decreased below levels seen in the presence of agonist alone (Fig. 5, b and c). There was no significant effect on the carbachol-mediated depletion of [ 3 H]PtdIns levels (data not shown). We also employed LY-294002 in an identical experimental protocol on the basis that this compound has been reported to inhibit PI3-K but not PtdIns 4-kinase (31) and would, therefore, enable distinction between the inhibitory effects of wortmannin on PtdIns 4-kinase and PI3-K. However, it has emerged that this compound also inhibits the wortmannin-sensitive PtdIns 4-kinase (30). Indeed the addition of 100 M LY-294002 in the presence or absence of 1 mM carbachol produced identical patterns of change to those caused by wortmannin (data not shown). The ability of wortmannin to influence carbachol-mediated changes in [ 3 H]InsP x and 3 H-polyphosphoinositides was concentrationrelated (Fig. 6, a and c), and wortmannin appeared to be equipotent with respect to all responses (IC 50  Wortmannin has been shown previously to inhibit partially purified PtdIns 4-kinase but not PtdIns(4)P 5-kinase isolated from bovine adrenal cortical cells (29). This suggests that its ability to deplete cellular levels of PtdIns(4,5)P 2 in the current study was dependent upon a reduced supply of PtdIns(4)P and not an inhibition of PtdIns(4)P 5-kinase. This was supported by experiments in ␤-escin permeabilized SH-SY5Y cells in which we demonstrated that 10 M wortmannin reduced the incorporation of 32 P from [ 32 P]ATP into PtdIns(4)P but not PtdIns(4,5)P 2 over a 10-min period under basal conditions (data not shown).
Addition of 10 M wortmannin 10 min prior to challenge with 1 mM carbachol markedly attenuated the transient peak of Ins(1,4,5)P 3 accumulation and abolished the sustained component of the response (Fig. 7a). Challenge of cells with 1 mM carbachol also resulted in a biphasic elevation of [Ca 2ϩ ] i consisting of a rapid transient peak (901 Ϯ 59 nM (n ϭ 4)) followed by a lower but sustained elevation (367 Ϯ 43 nM (n ϭ 4)) (Fig.  7b). Addition of 10 M wortmannin, 10 min prior to carbachol challenge, markedly attenuated the transient peak of [Ca 2ϩ ] i elevation (500 Ϯ 12 nM (n ϭ 4)) and abolished the sustained phase (Fig. 7b). This sustained phase of [Ca 2ϩ ] i elevation is dependent upon the influx of extracellular Ca 2ϩ via an as yet undefined mechanism but which most likely involves capacita-  tive entry (20). Addition of 1 M thapsigargin to these cells also resulted in capacitative Ca 2ϩ entry and a sustained elevation of [Ca 2ϩ ] i of similar magnitude to that mediated by 1 mM carbachol, but this was unaffected by 10 M wortmannin (data not shown). Thus, a direct block of Ca 2ϩ entry does not underlie the ability of wortmannin to inhibit Ca 2ϩ signaling (and potentially also phosphoinositide signaling through a reduction in the Ca 2ϩ feed-forward activation/facilitation of PLC).

Desensitization and Recovery of Carbachol-mediated Ins(1,4,5)P 3 Responses
By using a similar experimental protocol to that outlined above in which cells underwent a potentially desensitizing challenge with 1 mM carbachol for 5 min, we sought to examine recovery of the carbachol-mediated peak (10 s) Ins(1,4,5)P 3 responses. This peak response showed desensitization with washout and recovery periods of less than 2 min which approximately followed the time course of recovery of PtdIns(4,5)P 2 levels. By 2 min the response was fully restored (Fig. 8). Desensitization is often reflected in a reduction in agonist potency rather than a reduction in the maximal response, and we therefore conducted an identical series of experiments in which cells were rechallenged with a concentration of carbachol (25 M) equivalent to the EC 50 for the peak Ins(1,4,5)P 3 response in these cells (20). Compared with rechallenge with 1 mM carba- chol, these experiments demonstrated a more prolonged delay in the recovery of the peak Ins(1,4,5)P 3 responses which took greater than 5 min to fully recover and clearly lagged behind the recovery of PtdIns(4,5)P 2 (Fig. 8). Experiments in which the initial carbachol challenge was for 1 min rather than 5 min gave similar results (data not shown).

Derivation of Quantitative Aspects of PtdIns(4,5)P 2 Turnover during Sustained Agonist Activation
Under the pseudo-steady-state conditions that exist during sustained muscarinic acetylcholine receptor activation, the rate of PtdIns(4,5)P 2 synthesis must be equivalent to the rate of its breakdown. The initial rate of [ 3 H]PtdIns(4,5)P 2 recovery following removal of carbachol therefore provides an index of the rate of PtdIns(4,5)P 2 breakdown which pertained during sustained receptor activation. From the initial recovery rate (10,962 dpm/15 s/well ϭ 43,848 dpm/min/well, see Fig. 3), we calculated that the PtdIns(4,5)P 2 present during sustained receptor activation (8877 dpm/well) would be completely turned over every 12 s. This assumes that breakdown of PtdIns(4,5)P 2 (by either PLC or PtdIns(4,5)P 2 5-phosphatase activities) is negligible compared with synthesis immediately following agonist removal (any contribution of breakdown resulting in an underestimation of utilization) and that all of the PtdIns(4,5)P 2 exists within an agonist-sensitive pool. However, a fraction of PtdIns(4,5)P 2 (approximately 15% of basal) is resistant to depletion by wortmannin during carbachol stimulation (Fig. 5c) and is unable to support [ 3 H]InsP x accumulation (Fig. 5a). This fraction may not, therefore, be accessible to agonist-stimulated PLC and constitutes a temporary or permanent agonist-insensitive pool. If 15% (5177 dpm/well) of basal (34,510 dpm/well) PtdIns(4,5)P 2 represents an agonist-insensitive fraction, then the remaining pool (8877-5177 ϭ 3700 dpm/well) must turn over every 5 s under maximal sustained receptor activation.

Quantitative Estimates of PtdIns(4,5)P 2 Utilization Immediately following Receptor Activation
A number of studies have demonstrated that the M 3 -muscarinic receptor which is primarily responsible for carbachol-mediated activation of PLC in these cells (23) undergoes a rapid, but partial desensitization within seconds of agonist exposure. This is indicated by a biphasic accumulation of [ 3 H]InsP x in cells in which inositol monophosphatase activity has been blocked with Li ϩ to trap all products of PLC-mediated phosphoinositide hydrolysis (16,20,22,32). In addition, such rapid desensitization is reflected in the biphasic profile of Ins(1,4,5)P 3 accumulation following carbachol challenge as seen in this (Fig. 7a) and other studies (19 -22). It should be noted, therefore, that the utilization of PtdIns(4,5)P 2 during sustained receptor activation, as calculated above, represents that of a partially desensitized receptor. The rate of PtdIns-(4,5)P 2 utilization by non-desensitized receptors immediately upon agonist addition may well be far greater than this. Indeed the initial rate of decrease of PtdIns(4,5)P 2 following addition of carbachol (24,363 dpm/10 s/well ϭ 146,176 dpm/min/well) is 3.3-fold greater than during the sustained phase. This is in agreement with the 2-4-fold difference in the rates of [ 3 H]InsP x accumulation over the 1st min of agonist stimulation and the subsequent sustained phase accumulation in Li ϩblocked SH-SY5Y cells (22,32). However, our calculation assumes that the contribution of resynthesis is minimal during the initial fall of PtdIns(4,5)P 2 . As we demonstrate that resynthesis of PtdIns(4,5)P 2 is required even for the first 10 s of PLC activity (see below), this initial rate, and the difference between acute and sustained phases, is likely to be in excess of that calculated.

Quantitative Estimates of Basal PtdIns(4,5)P 2 Utilization and Relative Increases during Receptor Activation
PtdIns(4,5)P 2 is hydrolyzed in the absence of added agonist, and we have estimated this basal turnover using wortmannin. Addition of wortmannin under basal conditions caused an immediate reduction in [ 3 H]PtdIns(4)P (Fig. 5b) but no reduction in [ 3 H]PtdIns(4,5)P 2 for at least 2 min (Fig. 5c). Thus, during this 2-min period, although synthesis of [ 3 H]PtdIns(4)P was blocked due to inhibition of a PtdIns 4-kinase by wortmannin, there was sufficient [ 3 H]PtdIns(4)P to maintain the levels of [ 3 H]PtdIns(4,5)P 2 . The initial rate of decrease of [ 3 H]PtdIns-(4)P may, therefore, provide an index of the basal consumption of PtdIns(4,5)P 2 . This rate of 6839 dpm/min/well infers that an amount of PtdIns(4,5)P 2 equivalent to the total cellular pool (34,510 dpm/well) turns over every 5 min under basal (unstimulated) conditions. The rate of basal PtdIns(4,5)P 2 utilization is 6.4-fold less than the estimated rate of consumption during sustained receptor activation and at least 21.4-fold less than that occurring immediately upon agonist addition (see above). These determinations assume that recycling of PtdIns(4)P to PtdIns (by a PtdIns(4)P 4-phosphatase) and hydrolysis of PtdIns(4)P by PLC are negligible following addition of wortmannin. Any contribution would lead to an overestimation of basal activity and an underestimation, therefore, of the relative increase in the consumption of PtdIns(4,5)P 2 following agonist addition. As a consequence of the potential overestimation of basal hydrolysis, our values of 21-and 6-fold over basal stimulation of PtdIns(4,5)P 2 hydrolysis during the immediate and sustained components of maximal muscarinic-receptor activation, respectively, are likely to be underestimates. Indeed these values contrast with 150-and 60-fold stimulations reported previously (32) in these cells in which basal PtdIns(4,5)P 2 hydrolysis was assessed by the determination of [ 3 H]InsP x accumulation in Li ϩ -blocked cells. However, due to the uncompetitive nature of Li ϩ action, complete block may not occur under conditions of low flux through the enzyme (33) resulting in an underestimation of basal activity. It is likely that the true extent of stimulation lies between these values.

Absolute Changes in Levels of PtdIns(4,5)P 2
The measured basal level of PtdIns(4,5)P 2 was 87.7 pmol/ well, whereas cells labeled to equilibrium with [ 3 H]inositol had [ 3 H]PtdIns(4,5)P 2 present at 34,510 dpm/well. This relationship has enabled us to derive estimates of PtdIns(4,5)P 2 levels and utilization in mass units. Furthermore, our measurement of cell number and protein content of a typical well (4.9 ϫ 10 5 cells, 250 g of protein) has allowed conversion of these estimates to a per cell or per unit cellular protein basis. Thus, the basal level of PtdIns(4,5)P 2 is 0.18 fmol/cell (1.1 ϫ 10 8 molecules/cell) of which 0.15 fmol/cell exists in an agonist-sensitive pool. During sustained receptor activation with a maximally effective concentration of carbachol, PtdIns(4,5)P 2 levels fall to 0.05 fmol/cell of which 0.02 fmol/cell exists in an agonist-sensitive pool. During such sustained receptor activation, PtdIns(4,5)P 2 turns over at a rate of 0.23 fmol/cell/min (1.4 ϫ 10 8 molecules/cell/min), while during the acute phase of receptor activation this rate is at least 0.76 fmol/cell/min (4.6 ϫ 10 8 molecules/cell/min). From previous estimates of the catalytic activity of PLC-␤ (200 mol/min/mg protein (34)), we can estimate that the equivalent of approximately 15,000 molecules of PLC would require to be maximally activated to achieve this hydrolytic rate. DISCUSSION PtdIns(4,5)P 2 is considered to be the physiological substrate for both PLC and PI3-K activities. Despite the dependence upon PtdIns(4,5)P 2 as the substrate for Ins(1,4,5)P 3 , diacylglycerol, and PtdIns(3,4,5)P 3 synthesis, this polyphosphoinositide constitutes a minor fraction (ϳ5%) of the cellular inositol phospholipid content under basal conditions. The increasing realization that PtdIns(4,5)P 2 , and perhaps PtdIns(4)P, can modulate other aspects of cellular homeostasis (1, 35-37) might provide one explanation for the relatively low cellular levels of the polyphosphoinositides. Irrespective of what sets the cellular PtdIns(4,5)P 2 concentration, it is clear that efficient regulatory mechanisms must exist if supply and demand are to be matched, particularly during agonist-stimulated PLC and/or PI3-K activation.
The current study demonstrates that marked and sustained depletions of both PtdIns(4)P and PtdIns(4,5)P 2 are caused by a maximally effective concentration of carbachol. In SH-SY5Y cells PtdIns(4,5)P 2 is present at 360 Ϯ 27 (n ϭ 4) pmol/mg protein, and as shown here, this level can be decreased by Ն70% for prolonged periods under conditions of maximal agonist activation of M 3 -muscarinic acetylcholine receptors. The precipitous fall in the cellular levels of these phospholipids corresponds temporally with the peak of Ins(1,4,5)P 3 accumulation in these cells and indicates that conversion of PtdIns(4)P to PtdIns(4,5)P 2 is required to support both the initial and sustained components of transmembrane signaling via PLC.
The sustained depletions of the polyphosphoinositides are related to the extent of agonist stimulation although carbachol was more potent at depleting PtdIns(4)P than PtdIns(4,5)P 2 . This infers the presence of mechanisms that serve to protect the pool of PtdIns(4,5)P 2 at the expense of PtdIns(4)P. Such protection is also indicated by the markedly different rates of recovery of PtdIns(4)P and PtdIns(4,5)P 2 following abrupt termination of receptor activation. Thus, PtdIns(4,5)P 2 not only recovers to basal levels with a shorter half-time but also recovers by Ͼ60% before PtdIns(4)P recovery commences. Such protection of PtdIns(4,5)P 2 is in accord with studies in other systems (24,38,39) and suggests that substrate availability and product inhibition are unlikely to be the only factors governing the supply of PtdIns(4,5)P 2 . Thus, mechanisms exist that coordinate polyphosphoinositide synthesis in accordance with de-mand, and the current data indicate that these mechanisms are of relevance in intact cells.
Regulation of polyphosphoinositide supply could potentially occur via an activation of the kinases responsible for their generation and/or a reduction in the activities of the phosphatases that appear to be involved in substrate cycling (13). However, it seems unlikely that the massive increase in demand for PtdIns(4,5)P 2 could be met by changes in polyphosphoinositide-phosphatase activities alone as this would require a persistently high level of substrate or "futile" cycling (13). While coordinated changes in phosphatase and kinase activities may occur, there is little information on the regulation of phosphatase activities. There is, however, a variety of experimental evidence to implicate regulation of both PtdIns 4-kinase and PtdIns(4)P 5-kinase (8 -15). Of particular relevance to the current study is the suggestion that a G i -mediated activation of rho is responsible for the regulation of PtdIns(4)P 5-kinase (15,40). Thus, in HEK cells expressing recombinant muscarinic M 3 receptors, PtdIns(4,5)P 2 recovers to above basal levels in a pertussis toxin-sensitive fashion following removal of agonist, and it has been argued that this allows an enhanced phosphoinositide response upon rechallenge (40). Such results are in contrast to those of the current study in which we observed a desensitization of the Ins(1,4,5)P 3 response and a complete lack of effect of pertussis toxin on the rate or extent of recovery of PtdIns(4,5)P 2 following muscarinic receptor activation. Whether the difference in these studies reflects cell specificity or a promiscuous coupling of M 3 receptors to G i at high expression levels in HEK cells is unclear. However, the current study demonstrates that at physiologically relevant levels of muscarinic receptor expression (300 fmol/mg protein (32)), pertussis toxin-sensitive G-proteins play no significant role in regulating resynthesis of PtdIns(4,5)P 2 or setting the absolute level of this substrate in SH-SY5Y cells indicating that this is not a universal phenomenon.
By the use of wortmannin we have demonstrated that approximately 15% (of basal levels) of PtdIns(4,5)P 2 is unavailable to agonist-stimulated PLC. Similar values have been obtained using Li ϩ block in Chinese hamster ovary cells expressing recombinant muscarinic M 1 receptors (24). The reason for the inaccessibility of this fraction is unclear but may well relate to localization within the cell (41,42), and this has the potential to impart regulatory control on the signaling pathway. Taking the agonist-insensitive pool into account, we calculate that the agonist-sensitive pool of PtdIns(4,5)P 2 turns over at a rate of approximately 12 times per min under sustained stimulation with a maximally effective concentration of carbachol. In the absence of resynthesis, the existing substrate pool would be consumed in less than 2 s of agonist addition. This contrasts somewhat with a previous estimate for PtdIns-(4,5)P 2 turnover in these cells of 13-19 s (32) and emphasizes the critical dependence of peak (at 10 s) Ins(1,4,5)P 3 responses on PtdIns(4,5)P 2 resynthesis. Small changes in the absolute basal level of PtdIns(4,5)P 2 will, therefore, have little impact on phosphoinositide responses.
Wortmannin treatment of SH-SY5Y cells results in the loss of muscarinic receptor-mediated phosphoinositide and [Ca 2ϩ ] i responses demonstrating the critical dependence of this signaling pathway on a wortmannin-(and LY-294002-) sensitive PtdIns 4-kinase as reported in bovine adrenal glomerulosa cells (29,43). A recent report has indicated that phosphorylation of PtdIns(5)P at the D-4 position by the enzyme, previously identified as type II PtdInsP-5-OH kinase, provides another potential pathway for the generation of PtdIns(4,5)P 2 (44). Unless this route of synthesis also relies upon a wortmannin-sensitive enzyme, the current data suggest that this pathway contrib-utes little or not at all to the provision of PtdIns(4,5)P 2 in these cells under basal or agonist-stimulated conditions. The ability of wortmannin to further deplete PtdIns(4,5)P 2 during sustained receptor activation is also paralleled by a loss of [ 3 H]InsP x generation. This suggests that despite the ability of most PLC isoforms to hydrolyze PtdIns, PtdIns(4)P, and PtdIns(4,5)P 2 (45), the large cellular pool of PtdIns does not provide a substrate for PLC during muscarinic-receptor activation in SH-SY5Y cells. The issue of the substrate specificity of receptor-activated PLC is difficult to address in situ, and the ability of agents such as wortmannin to manipulate the phosphoinositide pools represents a novel approach to this problem. Furthermore, the ability of wortmannin to abolish the sustained [Ca 2ϩ ] i elevation during muscarinic receptor stimulation suggests that there is no Ca 2ϩ entry via a receptoroperated Ca 2ϩ channel (rather than a capacitative mechanism) which has been inferred in other systems (46,47).
Although both wortmannin and LY-294002 are better known for their ability to block PI3-K, inhibition of this enzyme cannot account for the results of the present study. Thus, the effect of wortmannin on both basal and carbachol-stimulated (poly)phosphoinositide levels is minimal at concentrations that have been reported to abolish totally PI3-K-dependent responses (30 -100 nM (29,43)). Wortmannin (and LY-294002) appears to inhibit one isozyme of the multiple PtdIns 4-kinase activities that are present in many cell types (3), possibly the recently cloned type III PtdIns 4-kinase (48,49). Although we acknowledge that wortmannin is not a specific inhibitor of PI-4 kinase (28,29,43), the data in the present study are totally consistent with an inhibition of this enzyme and indeed are in agreement with studies in which Li ϩ has been used to manipulate cellular levels of the polyphosphoinositides in Chinese hamster ovary cells during stimulation of muscarinic M 1 receptors (24). The current study demonstrates that while Li ϩ may be used to manipulate phosphoinositide pools in an agonist-dependent manner in some cell lines, this is not the case in SH-SY5Y cells. This is probably as a consequence of a large intracellular pool of free inositol, and a chronic inositol depletion strategy is required to render the cells sensitive to Li ϩ (50). Given the similar temporal profiles and susceptibility to desensitization of PLC signaling mediated by muscarinic M 3 receptors expressed in Chinese hamster ovary cells (Li ϩ -sensitive) and SH-SY5Y cells (Li ϩ -insensitive) (16,20,22,32), it is unlikely that the size of the inositol pool has a major impact on these aspects of phosphoinositide metabolism.
The present data emphasize the dynamic nature of the agonist-sensitive PtdIns(4,5)P 2 pool, and the use of wortmannin demonstrates that the immediate consequence of blocking the synthesis of PtdIns(4)P (and also therefore PtdIns(4,5)P 2 ) is the failure of agonist-induced Ins(1,4,5)P 3 generation and Ca 2ϩ mobilization. These data clearly indicate that the rate of supply of PtdIns(4,5)P 2 must be tightly and adequately matched to demand to prevent a full or partial failure of agonist-mediated phosphoinositide signaling. Indeed, a reduction in the availability of PtdIns(4,5)P 2 is a possible mechanism underlying the rapid, although often partial, desensitization of PLC activity that occurs within seconds of agonist occupation of many types of PLC-linked receptors including muscarinic receptors of SH-SY5Y cells (20,22,32). Whether depletion of PtdIns(4,5)P 2 contributes to or underlies such desensitization has proved difficult to resolve. However, we demonstrate here that the sustained reduction in the level of PtdIns(4,5)P 2 is related to the concentration of carbachol. This suggests that at submaximal agonist concentrations, there is PtdIns(4,5)P 2 available for hydrolysis and implies that limited substrate supply is not responsible for the desensitization phenomenon unless each molecule of PLC has access to a limited amount of PtdIns(4,5)P 2 and that this is replaced at a rate related to the extent of receptor activation.
The Ins(1,4,5)P 3 response to activation of muscarinic M 3 receptors in SH-SY5Y cells and indeed other cells consists of a rapid transient peak followed by a lower but sustained plateau phase (19 -22). We have argued previously that these two phases of Ins(1,4,5)P 3 accumulation represent desensitizationsensitive and desensitization-resistant phases, respectively (16). In the current study we also demonstrate desensitization of the peak Ins(1,4,5)P 3 response using a rechallenge protocol. These studies indicate that prestimulation with a maximal concentration of carbachol results in desensitization of the Ins(1,4,5)P 3 peak response upon rechallenge with either a maximal or submaximal concentration of carbachol. Although recovery of the response to a maximal concentration of carbachol paralleled the recovery of PtdIns(4,5)P 2 , recovery of the response to a submaximal concentration was slower. Thus, there are instances when there is sufficient PtdIns(4,5)P 2 to support a maximal Ins(1,4,5)P 3 response, and yet responses to a submaximal agonist concentration remain partially desensitized. This indicates that there are events unrelated to the recovery of PtdIns(4,5)P 2 (and other constituents of the signal transduction pathway) that are involved in the desensitization phenomenon. We have argued previously that this may be related to phosphorylation of the muscarinic M 3 receptor in a manner analogous to the phosphorylation and desensitization of the ␤ 2 -adrenoreceptor (16).
The current data indicate a critical dependence upon a wortmannin-and LY 294002-sensitive PtdIns 4-kinase for the synthesis of PtdIns(4,5)P 2 and indicate that the immediate consequence of blocking the synthesis of PtdIns(4,5)P 2 is the failure of receptor-mediated phosphoinositide and Ca 2ϩ signaling. Mechanisms clearly exist that enable the supply of PtdIns-(4,5)P 2 to be matched to demand, and these mechanisms are effective even under conditions of maximal muscarinic receptor activation in SH-SY5Y cells. Thus, despite the critical dependence upon PtdIns(4,5)P 2 synthesis, our data indicate that agonist-mediated depletion of this lipid substrate is unlikely to account for acute receptor desensitization in SH-SY5Y cells. The pattern of PLC activation and extent of PtdIns(4,5)P 2 depletion directed by muscarinic receptor activation in this cell line is consistent with that mediated by many other receptor types in a variety of cellular systems (5,16,17,51) suggesting that the current findings are likely to have general applicability to PLC signaling systems.