The Role of Calcium in Phospholipid Turnover following Glucose Stimulation in Neonatal Rat Cultured Islets*

Phospholipid turnover was studied in cultured neonatal rat pancreatic islets. In islets prelabeled with [""PIPi, 15-min stimulation with glucose (16.7 mM) caused increased labeling of phosphatidic acid (93%) and phosphatidylinositol (94%) and decreased labeling of the polyphosphoinositides (20%). Omission of cal- cium ion during the period of glucose stimulation did not modify the changes in inositol phospholipids. In islets equilibrated with [""PIPi in the presence and absence of stimulatory glucose concentrations (1 1.1 and 1.7 mM, respectively), chelation of calcium by ethylene glycol his(@-aminoethyl ether)-N,N,N',N'- tetraacetic acid prevented the increase in phosphatidic acid and phosphatidylinositol labeling. However, the decrease in polyphosphoinositide labeling was inhibited by the chelator only in islets labeled in the absence of stimulatory glucose concentrations, the decrease persisting in islets labeled in the presence of glucose. This suggests that a specific pool of polyphosphoinositides is labeled in the presence of agonist and decreases in response to acute glucose stimulation irrespective of availability of external calcium. In the absence of calcium, the addition of [y-""Pi]ATP to a membrane preparation of cultured islets yielded three lipid phosphorylation products (phosphatidic acid, phosphatidylinositol 4-monophosphate, and phos- phatidylinositol

The Role of Calcium in Phospholipid Turnover following Glucose Stimulation in Neonatal Rat Cultured Islets* (Received for publication, February 21,1984) Marjorie E. Dunlop  Phospholipid turnover was studied in cultured neonatal rat pancreatic islets. In islets prelabeled with [""PIPi, 15-min stimulation with glucose (16.7 mM) caused increased labeling of phosphatidic acid (93%) and phosphatidylinositol (94%) and decreased labeling of the polyphosphoinositides (20%). Omission of calcium ion during the period of glucose stimulation did not modify the changes in inositol phospholipids. In islets equilibrated with [""PIPi in the presence and absence of stimulatory glucose concentrations (1 1.1 and 1.7 mM, respectively), chelation of calcium by ethylene glycol his(@-aminoethyl ether)-N,N,N',N'tetraacetic acid prevented the increase in phosphatidic acid and phosphatidylinositol labeling. However, the decrease in polyphosphoinositide labeling was inhibited by the chelator only in islets labeled in the absence of stimulatory glucose concentrations, the decrease persisting in islets labeled in the presence of glucose. This suggests that a specific pool of polyphosphoinositides is labeled in the presence of agonist and decreases in response to acute glucose stimulation irrespective of availability of external calcium.
In the absence of calcium, the addition of [y-""Pi]ATP to a membrane preparation of cultured islets yielded three lipid phosphorylation products (phosphatidic acid, phosphatidylinositol 4-monophosphate, and phosphatidylinositol 4,5-bisphosphate). In broken cell preparations, [32P]Pi-labeled phosphatidylinositol was also detected. The extent of all these phosphorylations was decreased by the presence of free calcium ion (40 CIM) .
These data indicate that polyphosphoinositide turnover takes place after glucose stimulation independent of extracellular calcium and support the possibility that this may play a primary role in altering cell calcium availability.
Extensive investigations have established that glucose-induced insulin release requires an increase of calcium ions (Ca") within the pancreatic p cell (review in Refs. 1 and 2) and that intracellular and extracellular sources may contribute to raised free cytosolic Ca2+ on stimulation (3)(4)(5). While the inter-relationship of islet calcium sources is complex, it is possible that both intra-and extracellular Ca2+ availability may be affected by glucose-induced changes in islet phospho-Iipids and accompanying changes in the plasma membrane microenvironment. The precedent is seen in a number of * This project was supported by the National Health and Medical Research Council of Australia. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. tissues (reviewed in Refs. [6][7][8] in which hormones which exert their effects through mobilization of Ca2+ show coincident changes in phosphatidylinositol metabolism. In these diverse systems, a role for inositol phospholipids in the maintenance and disposition of cellular calcium seems likely. Furthermore, the polyphosphoinositides phosphatidylinositol 4-monophosphate and phosphatidylinositol 4,5-bisphosphate formed in the plasma membrane from phosphatidylinositol by the action of specific kinases have been shown to be rapidly degraded by a number of calcium-mobilizing stimuli (9-11). Similarly, insulin secretagogues glucose (12-15), leucine, and arginine (16) stimulate the metabolism of various 4 cell phospholipids with enhanced PI' metabolism and catabolism of its polyphosphorylated derivates shown in response to glucose. The calcium dependency of agonist-induced breakdown of PI, located in most membrane systems of the cell, and the polyphosphoinositides, located primarily in the plasma membrane (6), has been studied in many systems (8). In most of these, interpretation of calcium dependence remains equivocal. A requirement for calcium has been inferred in the adult islet as no phosphoinositide turnover was seen in the absence of Ca2+ and the presence of EGTA (13,14). In studies employing chelators, it is necessary to consider the capacity of EGTA to deplete intramembrane stores from which Ca2+ may be released by agonists (17). We have demonstrated that glucose stimulation of neonatal islet cells affects a plasma membrane complement of Ca2+ ionophoretic lipids (18), which may be an indication of a capacity for glucose to alter intramembrane calcium.
The following study was undertaken to determine the changes in neonatal islet inositol phospholipids following glucose stimulation and to establish whether these changes are a primary response independent of Ca2+ or whether they require the presence of Ca2+, which could indicate a response secondary to calcium entry.

Materials-[32P]
Pi and [Y-~'P,]ATP were obtained from The Radiochemical Centre, Amersham, England. RPMI 1640 medium and HEPES were from Flow Laboratories, Inc. All phospholipids were from Sigma. Precoated silica gel plates were from Merck. All other chemicals and solvents were from BDH Chemicals Ltd. (AnalaR grade).

8407
Glucose-induced Phospholipid Turnover transferred to RPMI 1640 medium supplemented with the salt concentration of HEPES-buffered Krebs-Ringer bicarbonate buffer (pH 7.4) containing glucose (1.7 RIM) for 20 min. For stimulation, glucose was added to this medium. In calcium-free medium, CaClz was omitted and replaced by NaCl. When present, EGTA was added to this medium to a final concentration of 5 mM. Insulin concentration of supernatant medium was determined by radioimmunoassay (20).
Measurement of p2P]P,-labeled Phospholipids in Glucose-stimulated Islets-Islets free of supernatant medium were extracted in ch1oroform:methanol:concentrated HCI (2001001, v/v). The organic phase generated following the addition of 0.6 volume of 1 M KC1 was dried under N,. Acidic phospholipids were separated by solid phase chromatography in a column of neomycin coupled to a glycophase support as described by Schacht (21). On these columns, PS, PA, PI, PI-4-P, and PI-4,5-P2 are obtained as discrete fractions. The identities of these phospholipids were verified by thin layer chromatography using Silica Gel 60-precoated plates prerun in 1% methanolic potassium oxalate as described by Shaikh and Palmer (22). The initial fractions from the solid phase system contained the nonacidic phospholipids PC and PE which were separated by thin layer chromatography (23).
Phospholipid Phosphorylation in Broken Cell and Islet Membrane Suspension-Islets were washed with phosphate-buffered saline and homogenized in 10 volumes of sucrose (0.35 M) in a hand-held TenBroeck homogenizer. An aliquot of this suspension served as a broken cell preparation. Nuclei and cell debris were removed from a second aliquot by centrifugation (600 X g, 5 min). Following centrifugation of the supernatant (20,000 X g, 20 rnin), a membrane particulate pellet was resuspended in sucrose. Protein content was determined by the method of Bradford (24). Phospholipid phosphorylation was determined using 5 pl of cell extract preincubated in Na acetate buffer (pH 6.8, 50 mM) containing 10 mM Mg acetate, 1 mM EGTA, and 0.2-2.0 mM CaCl, (final volume 60 pl). The reaction was started by the addition of 40 p~ [y-32Pi]ATP (10 pCi/ml).
In additional experiments, a sonicate of PI or diolein (0.5 pg in Na acetate buffer) was added to the preincubated cell extract prior to the addition of [y-32Pi]ATP.
After incubation at 25 "C, the reaction was terminated by addition of acidic chloroform:methanol, two phases were generated as described above, and phosphorylated products were separated by thin layer chromatography (22). Dried plates were autoradiographed to localize PI-4,5-P2, PI-4-P, PI, and PA, separated with RF values of 0.19,0.27, 0.46, and 0.78, respectively. The gel was removed in 2-mm bands and extracted into scintillation mixture for determination of "P, content.

RESULTS
The time course of incorporation of ["PIP, into islet phospholipids is shown in Fig. 1 Over the 15-min stimulation, PC, PS, and PI labeling occurs gradually, while the increased labeling of P A and the decrease in polyphosphoinositides occur more rapidly (within the first 5 min of stimulation). These phosphoinositide changes in islets equilibrated with ["PIP, in the presence of stimulatory (11.1 mM) or nonstimulatory (1.7 mM) glucose prior to the acute stimulation are further shown in Table 11. A difference in the basal labeling of the phosphoinositides is seen. In those islets equilibrated in the presence of glucose ( the basal labeling is seen as decreased PI labeling and increased PA labeling of glucose-equilibrated islets. In both these equilibration conditions, however, acute glucose stimulation increased PI and PA labeling and decreased the labeling of PI-4-P and PI-4,5-P2, as seen for islets equilibrated with glucose for 24 h.
Omission of Ca2+ during acute glucose stimulation had no effect on the enhancement of [32P]PA formation or the relabeling of PI, and the decrease in both polyphosphoinositides was still apparent. When in addition to the omission of Ca2+ 5 mM EGTA was present, PA and PI labeling were not increased by acute glucose stimulation, but a decrease in the polyphosphoinosites was still apparent in islets equilibrated in the presence of stimulatory glucose concentration but not in those islets equilibrated with [32P]Pi at nonstimulatory glucose concentrations. In both these calcium-free conditions, glucose failed to increase insulin release significantly above that seen in the absence of stimulatory glucose concentrations.
In vitro phosphorylation from [y-32Pi]ATP in broken cell and membrane preparations (Fig. 2)

TABLE I1
Effect of calcium removal on [32P]Pi labeling of inositol phospholipids and phosphatidic acid in response to acute glucose stirnulation following prelabeling in the presence of stimulatory and nonstimulatory glucose concentrations Islets were incubated in RPMI 1640 medium modified to bicarbonate-buffered Krebs solution containing glucose (1.7 or 16.7 mM) for 15 min. Ca2+ was omitted in the presence and absence of EGTA (5 mM). Inositol phospholipids were determined following thin layer chromatography. Values shown are mean +. S.E. (n = five to eight observations). For each prelabeling condition, the statistical significance of the difference from control (1.7 mM glucose) is indicated by an asterisk ( p < 0.05), a double asterisk (p < 0.01), and a triple asterisk ( p < 0.005).

DISCUSSION
This study has shown that glucose induces a sequence of events in cultured neonatal rat islets which indicates phosphatidylinositide phosphodiesteratic cleavage to form diacylglycerol, its phosphorylation to form phosphatidic acid, and the resynthesis of PI through cytidylphosphointermedlates. This confirms previous findings in mature islets. The time course studies also support the finding of Laychock (14) that polyphosphoinositide hydrolysis is an early event in glucoseinduced insulin secretion. The role of Ca2+ in PI and polyphosphoinositide turnover in different tissues has been controversial. In adult rat islets, it has been reported that PI and polyphosphoinositide turnover measured as [32P]Pi labeling (14) or inositol phosphate release (15) is markedly inhibited by the removal of Ca2+ with the addition of EGTA to the extracellular medium. It was therefore inferred that this phosphoinositide turnover is dependent on an influx of calcium. Using similar conditions of labeling with ["PIPi as employed by Laychock (14), similar results were obtained in the present study. However, when phosphoinosityde pools were labeled in the presence of stimulatory concentrations of glucose, quite different findings resulted. Even in the absence of extracellular Ca2+, and the presence of a high concentration of EGTA, polyphosphoinositide loss was induced by acute exposure to glucose. This suggests that when labeling is carried out in the presence of agonist (stimulatory glucose concentrations), a specific pool of phosphoinositides, inaccessible to EGTA and possibly at an inner membrane leaflet site, is labeled. This may be analogous t o the situation in other tissues where agonist labeling reveals a specific pool of PI which is hormone-responsive (29-31) and Ca2+-independent (31). In the islet, PI turnover itself remains dependent on extracellular Ca2+ even with agonist labeling, but polyphosphoinositide breakdown does not.
The role of Ca2+ in regulating phosphoinositide turnover was investigated further by looking at the incorporation of [32P]Pi from [Y-~'P~]ATP into phospholipids in membrane preparations and homogenates to assess endogenous kinase activity. Phosphorylation of PI to form PI-4-P and PI-4,5-P2 end of diacylglycerol to form PA was demonstrated in membrane preparations, with resynthesis of PI demonstrable only in homogenates. The net formation of the polyphosphoinosi-tides and of PA was shown in membrane preparations to be inhibited by Ca2+. This may reflect the sensitivity to ea2+ of islet phosphodiesterases and phospholipases as described for liver (321, brain (33), lymphocytes (34), smooth muscle (35), whole pancreas (36), and platelets (37). However, Ca2+ inhibition of the kinases involved may also contribute as microsomal diacylglycerol kinase of rat liver is inhibited by elevated Ca2+ (38). The findings described in the present report carry the following implications. As the polyphosphoinositides are chelators of both Ca2+ and Mg2+ (39), a change in their amount relative to other phospholipids located at internal cell membranes may change the amount of ea2+ bound to the plasma membrane. However, it must be remembered that using the techniques currently employed to measure phosphoinositide turnover, it is not possible to establish whether the pool size is sufficient to effect changes in membrane ea2+ availability and/or disposition. The finding of an absolute dependence on extracellular Ca2+ of PA and PI turnover but not of polyphosphoinositide breakdown may indicate that the latter glucoseinduced phospholipid turnover may be an initial event following glucose stimulation, which precedes Ca2+ entry into the islet. The sensitivity of polyphosphoinositide reaccumulation to Ca2+ demonstrated in the membrane preparations using [y-32Pi]ATP would indicate that while intracellular ea2+ levels remain high, polyphosphoinositide reaccumulation is prevented. In support of the sequence described are the ultrastructural studies of the mature pancreatic islet which show an accumulation of calcium closely associated with the plasma membrane which is depleted following glucose stimulation (40).
The current study emphasises the complexity of the membrane-associated inositol phospholipid pools and the importance of agonist labeling in revealing a specific, glucoseresponsive, and extracellular Ca2+-independent pool of polyphosphoinositides in the pancreatic islet. In thrombin-stimulated platelets, a loss of PI-4,5-P2 has been shown to precede Ca'+ mobilization, phospholipase activation, the formation of arachidonate metabolites, and alteration in polymerization of cytoskeletal elements (41). By analogy, the breakdown of polyphosphoinositides in the agonist-labeled pool in the neonatal islet may be an initiating step integral to glucoseinduced insulin release.