- Email: [email protected]
Phosphoinositide phosphorylation and hydrolysis in pancreatic islet cell membrane
ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 244, No. 2, February 1, pp. 421-429, 1986
Phosphoinositide Phosphorylation and Hydrolysis in Pancreatic Islet Cell Membrane MARJORIE
E. DUNLOP’
AND
WILLY
J. MALAISSE2
Labwratcny of Experimental Medicine, Brussels Free University, 115 Boulevard de Waterloo, Brussels 1000, Belgium Received
September
4, 1985
Membranes were isolated from dispersed rat pancreatic islet cells by attachment to Sephadex beads. When these membranes were exposed to [T-~~P]ATP, formation of 32P-labeled phosphatidate, phosphatidylinositol 4-phosphate, and phosphatidylinositol 4,5-bisphosphate was observed. Carbamylcholine, added 10 s prior to lipid extraction, caused a dose-related fall in 32P-labeled phospholipids. The effect of the cholinergic agent was suppressed by atropine, ethylene glycol bis(&aminoethyl ether)-N,N’-tetraacetic acid, and verapamil, and simulated, in part, by an increase in Ca2+ concentration. When the membranes were derived from islet cells prelabeled with [U-‘4C]arachidonate, carbamylcholine stimulation, in addition to decreasing labeled polyphosphoinositides, was accompanied by an increased production of labeled diacylglycerol, without a concomitant increase in labeled phosphatidylinositol. These results indicate that activation of a plasma membrane-associated phospholipase C directed against polyphosphoinositides represents a primary event in the functional response of the pancreactic /3 cell to cholinergic agents. o 1986 Academic press, I~~. Increasing attention is currently being paid to the possible participation of phosphoinositide hydrolysis in the process of insulin release evoked in the pancreatic ,8 cell by nutrient and cholinergic secretagogues (1, 2). The activity of certain enzymes involved in the synthesis of polyphosphoinositides (phosphatidylinositol kinase, phosphatidylinositol 4-phosphate kinase) or their breakdown (phospholipase C) was recently demonstrated in homogenates of neonatal or adult rat islets (3-5). Using a rat P-cell tumor, Tooke et aZ. (6) observed phosphatidylinositol kinase activity in both the plasma membrane and the secretory-granule membrane, as well as phosphatidylinositol4-phosphate phosphatase activity associated with the ‘Present address: University of Melbourne, Department of Medicine, Royal Melbourne Hospital, Parkville, Victoria, Australia. * To whom correspondence should be addressed. 421
plasma membrane. The present work was undertaken to gain further insight into the metabolism of phosphoinositides in membranes isolated from normal rat islets, in particular the enzymatic determinants of changes provoked in such a preparation by cholinergic stimulation. MATERIALS
AND
METHODS
Materials. [U-“ClArachidonic acid (390 mCi/mmol) was obtained from New England Nuclear, Boston, Massachusetts; adenosine 5’-[y-“Pltriphosphate, triethylammonium salt (>5000 Ci/mmol) from Amersham International plc, Amersham, Bucks, U. K.; RPM1 medium from Flow Laboratories, Ayrshire, U. K.; verapamil HCl from Knoll A.G., Ludwigshafen, FRG; and all other chemicals and phospholipids from Sigma Chemical Company, Ltd. Preparation and incubation of islet cells. Pancreatic islets isolated by the collagenase method from fed albino rats were further disrupted into isolated cells by mechanical agitation for 10 min in Gas+-free Hanks’ solution containing EGTA (1.0 mM), bovine albumin 0003-9861/86 Copyright All rights
$3.00
0 1986 by Academic Press. Inc. of reproduction in any form reserved.
422
DUNLOP AND MALAISSE
(5 mg/ml), and collagenase (2 mg/ml) (7). After filtration through a nylon screen the islet cells were pelleted by centrifugation for 5 min at 600s and, in one series of experiments, resuspended in bicarbonatebuffered medium RPM1 1640, containing glucose (11.1 mM) and [U-‘“Clarachidonic acid (12.8 pM) at a cell concentration of 8 X 106/ml. The cells were incubated with constant oxygenation (02:COz, 95:5, v:v) for 120 min at 37°C. Preparation of id& cell membranes. Islet cells either freshly isolated or following [U-“Clarachidonic acid labeling were pelleted by centrifugation for 5 min at 6OOg and resuspended in a Tris-HCl buffer (10 mM, pH 5.2) containing 217 mM sucrose and 46.5 mM sodium acetate at a concentration of 8 X lo6 cells/ml. The cells (8 X 106/sample) were attached to a positively charged, beaded derivative of Sephadex (Cytodex 1, Pharmacia, Uppsala, Sweden) as described previously (8). The cells were lysed with hypotonic buffer (TrisHCl, 10 mM, pH 8.0,4.0 ml) for 20 min. At this time, where indicated, preparations of beads and attached membranes were sonicated at low amperage. The beads and attached membranes (sonicated or unsonicated) were again centrifuged and the supernatant buffer was removed entirely. On occasion islet cells were homogenized in Potter-Elvehjem tubes in TrisHCl buffer (10 mM, pH 8). Phosphwglutim of islet cell hmmgendes and idated membranes. A homogenate of islet cells (8 X lo6 cells/ 0.05 ml) or beads and attached membranes from an equivalent number of islet cells were phosphorylated using adenosine 5’-triphosphate. Phosphorylation was started by the addition of 200 pl of sodium acetate buffer (50 mM, pH 6.8) containing MgCla (2 mM) and ATP (0.2 mM). In those experiments in which islet membranes prelabeled with arachidonic acid were used, phosphorylation was with the unlabeled nucleotide only. In the remaining experiments using islet cell homogenates and isolated membranes from non[U-‘“Clarachidonate-labeled cells, the ATP (final concentration: 0.2 mM) included [y-“P]ATP (4 pCi/sample). In additional experiments a sonicate of phosphatidylinositol (brain extract type I, Sigma Chemical Company Ltd.), confirmed by thin-layer chromatography (see below) to contain phosphatidylinositol as the sole inositide phospholipid, was prepared in Na acetate buffer, pH 6.8, and added to the sonicated beads and attached membranes to give a final concentration of lo-’ M, immediately prior to the addition of [y-aaP]ATP. After 5 min at 21°C and 10 s prior to lipid extraction, 20 ~1 of carbamylcholine or Ca’+/ EGTA-buffered solutions was added to the incubation medium. Likewise, buffered solutions of the potential antagonists (EGTA, verapamil, and atropine) were added 5 s prior to carbamylcholine or Ca’+/EGTA.
The incubation was terminated by the addition of CHCla:CHsOH:13NHCl, 200:100:1, v:v:v (1.0 ml), and the lipids were extracted for 30 min with constant mixing. Two phases were generated by the addition of KC1 (1 M) and CHCla. The chloroform phase was dried under vacuum at 20°C and the lipid resuspended in 30 ~1 CHCls for application (lo-r1 aliquots) to silica gel 60 plates (Merck 5721, Darmstadt, FRG) preloaded with potassium oxalate (1% in CHIOH:H20, 23, v:v) and air dried at 110°C for 1 h before use. Phospholipids were separated in a solvent system of chloroform: acetone:methanol:acetic acid:H*O (80:34:30:24:16, v:v:v:v:v) under saturation conditions. Of duplicate plates, one was autoradiographed using Kodak X-OMat AR film, and the other scraped, according to the position of authentic standards (after iodine staining) and the gel suspended in 2.5 ml methanol followed by the addition of 10 ml Aquasol(New England Nuclear) for scintillation counting. Under these conditions authentic standards gave Rf values of 1.0, 0.70, 0.51, 0.44, 0.36, 0.33, 0.26, and 0.20 for palmitic acid, phosphatidic acid, lysophosphatidic acid, phosphatidylserine, phosphatidylcholine, phosphatidylethanolamine, phosphatidylinositol, phosphatidylinositol 4-phosphate, and phosphatidylinositol 4,5-bisphosphate, respectively. In addition [U-%]arachidonic acid-labeled lipids were separated following thin-layer chromatography in a solvent system of petroleum ether:diethyl ether:acetic acid (70:30:1, v:v:v) in which the Rfvalues for authentic standards were 0.79,0.40, 0.30, 0.29, 0.27, and 0.08 for triacylglycerol, palmitic acid, cholesterol, 1,3-diolein, 1,2-diolein, and monoacylglycerol, respectively, phospholipids remaining at the origin. In this manner phospholipids and neutral lipids were separated from [U-“Clarachidonic acidlabeled homogenates and islet cell membranes prior to and following phosphorylation and from homogenates and islet cell membranes following identical phosphorylation using [y-azP]ATP. Presentation of results. In the phosphorylation experiments employing [y-**P]ATP, the results are expressed as femtomoles of =P transferred from ATP over 5 min incubation into the membranes isolated from lo6 isolated islet cells. Following exposure to [U-%]arachidonic acid the results are expressed as picomoles of arachidonate incorporated into lo6 cells or membrane equivalent. All results are presented as means; followed by either the range of individual variations when only two observations were made or, otherwise, the SEM together with the number of individual observations (n). Student’s t test was used to determine the levels of statistical significance. RESULTS
32PLabeling of Membrane Lipids ’ Abbreviation used: EGTA, aminoethyl ether)-N,N’-tetraacetic
ethylene glycol acid.
bis(@-
Over 5 min incubation in the presence of nonsonicated mem-
[T-~~P]ATP (0.2 mM),
PHOSPHOINOSITIDE
METABOLISM
branes incorporated 200 f 14 and 258 + 20 fmol of 32P per lo6 cells (n = 5-11) in phosphatidate and phosphatidylinositol 4phosphate, respectively (Fig. 1). The incorporation of radioactivity into phosphatidylinositol 4,5-bisphosphate was close to the lower limit of detection (~10 fmol/106 cells). With sonicated membranes, the incorporation of 32P amounted to 230 + 22 fmol/106 cells (n = 15) in phosphatidylinositol 4-phosphate, whereas labeled phosphatidate and phosphatidylinositol 4,5bisphosphate were not detected over this short period of incubation. If a brain extract containing phosphatidylinositol as virtually the sole inositide was added to the sonicated membranes, the labeling of polyphosphoinositides increased to 634 + 48 fmol/106 cells (n = 5), of which 73.1 t 0.5 and 26.9 + 0.5% (n = 4) appeared
Phosphatidate
Lysophosphatidate
PtdIns4
P
PtdIns4,5P,
Origin
1
2
3
FIG. 1. Autoradiograph of TLC of lipid extracts from a crude islet homogenate (lane l), nonsonicated membranes (lane 2), and sonicated membranes exposed to a brain extract (lane 3) after 5 min incubation with [Y-~P]ATP. The position of authentic standards is indicated (PtdInslP, phosphatidylinositol I-phosphate; PtdIns4,5Pa, phosphatidylinositol4,5-bisphosphate). The arrow points to an unidentified lipid-extractable phosphorylation product also encountered in prior studies (15, 26,27).
IN
RAT
ISLET
423
MEMBRANE
as phosphatidylinositol 4-phosphate and phosphatidylinositol4,5-bisphosphate, respectively (Fig. 1). In the presence of the brain extract, no labeled phosphatidate was detected. By comparison, in a crude islet homogenate, the labeling of phosphatidate averaged 102 + 9, that of phosphatidylinositol 4-phosphate 840 + 133, and that of phosphatidylinositol 4,5-bisphosphate 109 + 4 fmol/106 cells (n = 2-4). In this case, the addition of the brain extract failed to cause an obvious change in the labeling of either phosphatidate or phosphatidylinositol 4,5-bisphosphate, but almost doubled that of phosphatidylinositol 4-phosphate. Hydrolysis of ‘2P-Labeled Polyphosphoinositides
Membrane
When the membrane lipids were labeled with [y-32P]ATP and carbamylcholine added to the medium 10 s prior to extraction of lipids, a dose-related decrease in the amount of labeled phosphatidylinositol 4-phosphate was observed (Table I). Essentially the same results were obtained in nonsonicated (Table I) and sonicated membranes. In the latter case, the relative fall in [32P]phosphatidylinositol 4-phosphate increased from 36.6 + 2.3% to 79.1 k 1.2% and 93.0 + 0.1% as the concentration of carbamylcholine was raised from 10 PM to 500 I.LM and 2.0 InM, respectively. As little as 10 I.LM carbamylcholine was sufficient to cause a one-third decrease (P < 0.001) in [32P]phosphatidylinositol 4-phosphate, virtually complete disappearance of this compound being seen in response to 2.0 mM carbamylcholine. Atropine (10 PM) and verapamil(lO0 PM), when added 5 s before carbamylcholine, antagonized the effect of the cholinergic agent, although failing to affect significantly the recovery of labeled phosphatidylinositol 4-phosphate in the absence of carbamylcholine (Table I). In the presence of verapamil, the carbamylcholine-induced decrease in [32P]phosphatidylinositol 4-phosphate represented no more than 27.7 k 7.5% (n = 4; P < 0.005) of the paired value found in the absence of the calcium antagonist. When EGTA (100 PM) was added 5 s prior to carbamylcholine, the decrease in labeled phosphatidylino-
424
DUNLOP TABLE
AND
I
EFFECT OF CAaB.4MncuoLINE, VERAPAMIL, AND ATROPINE UPONTHE LEVELOF "P-LABELED PHOSPHATIDYLINOSITOL d-PHOSPHATE IN NONSONICATED ISLET CELL MEMBRANES Nil Carbamylcholine (10 ELM) Carbamylcholine (100 PM) Carbamylcholine (500 PM) Carbamylcholine (2000 PM) Verapamil (0.1 mM) Verapamil (0.1 mM) + carbamylcholine (10 pM) Verapamil (0.1 mM) + carbamylcholine (500 pM) Atropine (10 pM) Atropine (10 PM) + carbamylcholine (500 pM)
100.0 67.7 66.6 20.8 11.2 110.2
f + f f + f
1.3 2.7 1.5 0.8 1.8 3.6
(9) (4) (3) (6) (5) (2)
93.8 r?I 2.8 (2) 71.3 + 8.5 (2) 98.9 + 0.3 (2) 87.7 f 3.5 (2)
Note. All values are expressed as percentages of the mean control value found within the same experiment(s). Such a control value averaged 271 f 26 fmol/ lo6 cells.
sitol 4-phosphate was prevented at a 100 concentration of the cholinergic agent and was inhibited by 71% at a 2.0 mM concentration of carbamylcholine. In view of these findings, the effect of Ca2+ itself upon the hydrolysis of phosphatidylinositol 4phosphate was examined. When a fixed amount of EGTA (10 PM) and increasing amounts of CaC12 were added to the medium 10 s prior to lipid extraction, a dose-related fall in labeled phosphatidylinositol 4-phosphate was observed in sonicated membranes (Fig. 2), with an apparent K, for Ca2+ close to 5 PM. Verapamil (0.2 mM) inhibited, by no more than 26.2%, the response to 100 PM Ca2+. Incidentally, EGTA alone slightly increased the amount of labeled phosphatidylinositol 4-phosphate to 113.1 f 3.1% (n = 4; P < 0.025) of the paired control value found in the absence of EGTA. In the experiments so far presented, the effect of carbamylcholine or Ca2+ upon phosphatidylinositol 4,5-bisphosphate could not be assessed, in view of the minimal formation of labeled phosphatidylinositol4,5-bisphosphate in the islet membranes. Therefore, a series of experiments was conducted in the presence of the brain PM
MALAISSE
extract and sonicated membranes, in which case sizable amounts of labeled phosphatidylinositol 4,bbisphosphate are formed (see above). Under these conditions, carbamylcholine (100 pM) decreased the amount of labeled phosphatidylinositol4,5bisphosphate from a control value of 142 to 17 fmol/106 cells. Likewise, increasing concentrations of Ca2+ caused a dose-related decrease in labeled phosphatidylinositol4,5-bisphosphate (Fig. 2), with an apparent K, for Ca2+ close to ‘7 PM. In these experiments, the cholinergic agent and the Ca2+ ion also caused the disappearance of labeled phosphatidylinositol 4-phosphate. However, as illustrated in Fig. 2, higher concentrations of Ca2+ were required to cause a maximal fall in [32P]phosphatidylinositol 4-phosphate in the presence than in the absence of the brain extract. This difference could be due, in part at least, to the fact that, in the presence of the brain extract, the hydrolysis of labeled phosphatidylinositol4-phosphate may coincide with its generation from labeled phosphatidylinositol 4,5-bisphosphate through the action of a calcium-sensitive phosphomonoesterase. It should be noted that, in the presence of the brain extract, EGTA alone (10 PM) increased the amount of labeled phosphatidylinositol 4,5-bisphosphate to 117.1 + 4.4% (n = 4; P < 0.05) of the paired control value found in the absence of EGTA. The responsiveness to both EGTA and CaC12 was thus qualitatively similar in the case of phosphatidylinositol 4-phosphate and phosphatidylinositol4,5bisphosphate, respectively. When glucose (16.7 mM) was added to the medium 10 s prior to extraction of lipids, no change in labeled phosphatidylinositol 4-phosphate was observed in nonsonicated membranes. Hydrolysis of s2P-Labeled Membrane Phosphatidate
The addition of carbamylcholine and, to a lesser extent, that of Ca2+ 10 s prior to lipid extraction also decreased the amount of labeled phosphatidate formed by nonsonicated membranes exposed to [y=P]ATP. Thus, relative to the control value
PHOSPHOINOSITIDE
METABOLISM
IN
RAT
ISLET
425
MEMBRANE
500-
0 400, = u ID”
0
< z E 300. 0 ” ” ,?I .-2 2!
Lo,
200.
.----,
E
X... ......,
T IL 100,
O-
. o-
/
“,1 q. \ :.\ : \..\: i “,*. .*- .----- -- --.- ------1--, .-.*..... .......... ..a.............+..... 10
5 Added
25 CaC12
50
-c--e c....*( t 103
(Jo M)
FIG. 2. Effect of increasing concentrations of CaCla (logarithmic scale) added together (10 PM) 10 s prior to lipid extraction upon the recovery of q-labeled phosphatidylinositol (circles) and phosphatidylinositol 4,5-bisphosphate (crosses) in sonicated membranes the absence (closed circles) or presence (open circles and crosses) of a brain extract. refer to two individual determinations.
(200 f 14 fmol/106 cells), carbamylcholine (10, 100, 500, and 2000 PM) caused a doserelated decrease in [32P]phosphatidate (n = 6, P < 0.001). The carbamylcholine-induced fall in [32P]phosphatidate averaged 37 +- 1 and 74 + 1% at the lowest and highest concentration of the cholinergic agent, respectively. At a 100 I.~M concentration, Ca2+ caused a 22 + 2% decrease in [32P]phosphatidate. Verapamil(lO0 PM) and atropine (10 PM) added 5 s prior to the introduction of carbamylcholine abolished the effect of the cholinergic agent. Thus, in the presence of verapamil and atropine, respectively, the readings obtained after addition of carbamylcholine (10 to 500 PM) averaged 98 * 6 and 94 + 1% of the control value. Neither verapamil nor atropine affected the amount of [32P]phosphatidate measured in the absence of carbamylcholine, the experimental values averaging 96 + 6% of the corresponding control value.
[lJ-14CjArachicbnate Membrane Lipids
Labeling
with EGTA I-phosphate incubated in Mean values
of
In the experiments so far presented, the liberation of diacylglycerol from polyphosphoinositides could obviously not be monitored. As this information is required to distinguish between the action of a phosphomonoesterase and a phosphodiesterase, the fate of membrane lipids labeled with [U-14C]arachidonate was examined. For this purpose, the isolated islet cells were exposed for 120 min to [U-‘“Clarachidonate prior to membrane isolation. For the sake of comparison, the isolated membranes were then treated in a manner comparable to that used in the first series of experiments, being exposed to unlabeled ATP for 5 min prior to lipid extraction. The composition of the [U-14C]arachidonate-labeled lipids extracted from the membranes after the 5-min incubation in
426
DUNLOP
AND
the presence of ATP is illustrated in Table II. The amount of labeled phosphatidylinositol was low relative to either the total amount of labeled phospholipids (3.9 + 0.8%) or the total amount of inositides (53.4 f 12.1%). The labeled phospholipids represented 54.8 + 9.2% of the total amount of labeled lipids (n = 3 in all cases). This pattern differs in several respects from that found in the crude islet homogenate treated under identical conditions. Thus, in the crude homogenate, labeled phospholipids represented only 28.8 f 5.1% of total labeled lipids. Moreover, the relative amount of labeled phosphatidylinositol was higher in the crude homogenate, representing respectively 17.4 + 1.2% of the labeled phospholipids and 73.7 f 2.0% of the labeled phosphoinositides. Incidentally, the total amount of labeled lipids in the crude homogenate was about twice higher than that in the isolated membranes. Over the 5-min period of exposure to ATP (0.2 mM), no obvious change was seen in the amount of labeled phosphatidylethanolamine and phosphatidylcholine or arachidonate. The amount of labeled polyphosphoinositides increased, this being matched by a decrease in labeled phosphatidylinositol, with a mean value 1.11 + 0.01 pmol/106 cells. There was a minor increase in labeled phosphatidate (0.40 pmol/106 cell) of the same order of magnitude as that observed for the 32P labeling of this lipid in nonsonicated membranes. Labeled diacylglycerol decreased over the same 5-min period, coinciding with a comparable inTABLE EFFECTOF
CARBAMYLCHOLINE LIPIDS
MALAISSE
crease in monoacylglycerol. These changes (2.31 k 0.25 pmol/106 cells) were accentuated two- to threefold when Ca2’ (100 PM) was added 10 s prior to lipid extraction, suggesting that they may reflect, in part at least, the activity of a Me-dependent and Ca2+-responsive diacyiglycerol lipase acting selectively at the sn-1 position (9710). Eflect of Carbamylcholine [U-14C]Arachidonate-Labled Membrane Lipids
When added 10 s prior to lipid extraction carbamylcholine (0.5 mM) failed to affect significantly the recovery of the labeled phosphatidate, phosphatidylinositol, the combined fraction of phosphatidylethanolamine and phosphatidylcholine, or arachidonate (Table II). The polyphosphoinositides were decreased by 1.81 f 0.44 pmol/106 cells, whereas diacylglycerol was increased by 3.18 f 0.93 pmol/106 cells. When either atropine (10 PM), verapamil (0.2 mM), or EGTA (0.1 mM) was added 5 s prior to the introduction of carbamylcholine (0.5 mM), the response to the cholinergic agent was prevented, the mean values for the polyphosphoinositides and diacylglycerol now averaging 119.5 k 13.1 and 104.8 f 8.7%, respectively, of the mean control value (n = 5 in both cases). If the present data are compared to those obtained in the 32P-labeling experiments, it becomes obvious that the relative magnitude of the carbamylcholine-induced fall II
(0.5 ~M)UPONTHELEVELOF[U-"C&~RACHIDONATE-LABELED IN ISLET CELL NONSONICATED MEMBRANES Control (pmol/lOs cells)
Phosphatidate Phosphatidylinositol Phosphatidylethanolamine Polyphosphoinositides Arachidonate Diacylglycerol
Note. Mean with
values (? SEM) the statistical significance
upon
+ phosphatidylcholine
16.35 f 1.91 10.36 f 1.45 207.68 f 21.80 6.64 k 0.64 144.65 f 10.57 7.94 f 0.57
refer to three individual determinations of changes evoked by carbamylcholine
Carbamylcholine (% of control) 105.5 2 100.8 f 103.3 f 72.8 f 109.3 f 152.2 +
5.4 6.5
1.2 5x** 4.1
12.6*
in all cases, and are shown (*P < 0.06; **P < 0.05).
together
PHOSPHOINOSITIDE
in labeled polyphosphoinositides greater in the latter case.
METABOLISM
was much
DISCUSSION
With the present method of membrane isolation, the material attached to the Sephadex beads corresponds to a plasma membrane-enriched fraction. Sonication of the beads with attached membranes aims at further purification of the plasma membrane-rich material (10-12). The present data demonstrate the presence in islet cell membranes of the kinase activities catalyzing the phosphorylation of diacylglycerol, phosphatidylinositol, and phosphatidylinositol 4-phosphate. The presence of the latter enzyme is substantiated by the data obtained in the concomitant presence of sonicated membranes and brain extract. The absence of phosphatidate labeling in sonicated, as distinct from nonsonicated, membranes probably reflects mainly a loss of substrate rather than enzyme activity. It is indeed possible to restore phosphatidate labeling in sonicated membranes in the presence of brain extract and 12-diolein (data not shown). Moreover, cellular lysis is known to be accompanied by breakdown of inositides and phosphatidic acid, unless Ca2+ is rigorously excluded from the lysing buffer (13). Carbamylcholine caused a dose-related fall in the membrane content of 32P-labeled phosphatidylinositol4-phosphate. This effect was antagonized by atropine and verapamil, suggesting that it corresponded to a calcium-sensitive process mediated by muscarinic receptors. In this respect, it should be stressed that these experiments were conducted in the absence of any Cachelating agent and, hence, in the presence of some contaminating Ca2+ (~5 PM). The effect of the cholinergic agent being assessed over a period of 10 s, it seems most likely that the decrease in 32P-labeled phosphatidylinositol4-phosphate reflects a process of rapid hydrolysis rather than inhibition of phosphorylation. The loss of labeled polyphosphosinositide was not accompanied by the appearance of any 32Plabeled phospholipid, suggesting that it was attributable to activation of Ca2+-de-
IN
RAT
ISLET
MEMBRANE
427
pendent phosphomonoesterase or phosphodiesterase. Whatever its nature, the enzyme was apparently activated by Ca2+ itself as judged from the effect of both EGTA and increasing concentrations of added CaC12. In this respect, it should be kept in mind that the hydrolysis of phosphoinositides was here examined in the absence of detergent, thus maintaining the physical state of the membrane and, hence, optimal sensitivity to Ca2+ (14). The finding that the calcium antagonist verapamil inhibited the hydrolytic response to carbamylcholine and, to a lesser extent, calcium could be related to competition between Ca2+ and verapamil for a membrane site, although an unspecific action of this lipophilic drug cannot be ruled out. A further series of experiments was conducted to gain information on the precise modality of polyphosphoinositide hydrolysis in the islet membrane preparation. Thus by exposing the islet cells to [U-‘*C]arachidonate prior to membrane isolation, it was possible to demonstrate the formation of diacylglycerol in response to carbamylcholine. This finding supports the participation of a phosphodiesterase in the hydrolysis of polyphosphoinositides in good agreement with data obtained in intact islet cells (15). Since the cholinergic agent failed to affect the recovery of [U-‘*C]arachidonate-labeled phosphatidylinositol, it would appear that the phospholipase C activated by carbamylcholine acted preferentially on polyphosphoinositides, as distinct from phosphatidylinositol. Incidentally, the experiments conducted in the presence of the brain extract indicate that the carbamylcholine-responsive enzyme(s) catalyzes the hydrolysis of both phosphatidylinositol 4-phosphate and phosphatidylinositol4,5-bisphosphate. It should be stressed that the hydrolysis of polyphosphoinositides needs not to be attributed solely to a phospholipase C mechanism. The hydrolysis of phosphatidylinositol 4,5-bisphosphate by a membrane-bound phosphomonoesterase has been reported in several tissues (16-19). However, at variance with the findings reported by Tooke et al. (6), no evidence of phosphomonoesterase acting on phosphatidylinositol 4-phosphate was obtained in
428
DUNLOP
AND
the present study, as neither carbamylcholine (Table II) nor Ca” (data not shown) increased the recovery of [U-‘“Clarachidonate-labeled phosphatidylinositol. An unexpected finding consisted in the rapid hydrolysis of 32P-labeled phosphatidic acid evoked by carbamylcholine in nonsonicated membranes. The rapidity of such a response argues against any significant participation of an inhibitory effect of the cholinergic agent upon diacylglycerol kinase. Activation of a phospholipase A2 seems also improbable, since no %P-labeled lysophosphatidate was detected after exposure of the nonsonicated membranes to carbamylcholine (Fig. 1). This leaves the possibility of activation of a phosphatidate phosphatase, such as a Mga+-dependent speciSc or nonspecific phosphohydrolase producing diacylglycerol and Pi (20) or, as reported in rat lung, a microsomal lipase producing glycerol phosphate and fatty acid (21). The knowledge that the pH optimum of the specific phosphatidate phosphohydrolase is met by the condition of the present experiments (22) and the fact that the carbamylcholine-stimulated formation of [U-‘4C]arachidonate-labeled diacylglycerol somewhat exceeded that accounted for by the fall in labeled polyphosphoinositides would support the former possibility. Whatever its precise nature, the carbamylcholine-responsive enzyme appeared to act preferentially upon the pool of phosphatidate generated through the action of a membrane-associated kinase and not on the much larger pool of [U-‘“Clarachidonate-labeled phosphatidate. An effect of the cholinergic agent upon a restricted pool of phosphatidate in the isolated membrane is not necessarily incompatible, therefore, with the knowledge that, in intact islet cells, cholinergic agents provoke a rapid increase in the 32P labeling of phosphatidate (23). Indeed, in intact cells, the removal from the membrane of diacylglycerol-derived phosphatidate by a putative transfer protein (24) mediating the rapid transfer of phosphatidate to intracellular sites may channel the latter phospholipid away from a membrane-associated phosphohydrolase. Even in intact cells, the phosphatidate formed through the hydrolysis of phos-
MALAISSE
phoinositides may not be the sole precursor for phosphatidylinositol found in the membrane. As a matter of fact, the [U-r4C]arachidonate labeling of membraneassociated phosphatidate and phosphatidylinositol strongly suggests de novo synthesis of phosphatidate (25). In the present experiments, the membrane-associated pool of [U-‘4C]arachidonate-labeled phosphatidylinositol, which may only represent a fraction of the total membrane-associated phosphatidylinositol, seems to act as a preferential substrate for phosphorylation to polyphosphoinositides. Indeed, over 5 min incubation in the presence of unlabeled ATP, the conversion of [U“C]arachidonate-labeled phosphatidylinositol to polyphosphoinositides was of the same order of magnitude as the 32P labeling of the polyphosphoinositides in membranes exposed for the same time to [T-~~P]ATP. In conclusion, the present data afford direct support to the view that activation of a plasma membrane-associated phospholipase C directed against polyphosphoinositides represents the primary event in the cholinergic pathway for stimulation of insulin release. In the isolated membrane, the modulation by calcium chelator and calcium antagonist of the response to carbamylcholine points to a key role for membrane-bound Ca2+ in the coupling of muscarinic receptor occupation to polyphosphoinositide hydrolysis. ACKNOWLEDGMENTS This work was supported in part by grants from the Belgian Foundation for Scientific Medical Research and the Belgian Ministry of Scientific Policy (Brussels, Belgium). M.E.D. was supported by the National Health and Medical Research Council of Australia. We thank F. Malaisse-Lagae and M. Mahy for assistance and C. Demesmaeker for secretarial help. REFERENCES 1. BEST, L., AND MALAISSE, W. J. (1983) Diabetologia 25.299-305. 2. BEST, L., DUNLOP, M., AND MALAISSE, W. J. (1984) Eqxtientia 40,1085-1091. 3. SCHREY, M. P., AND MONTAGUE, W. (1983) Biochm. J. 216,433-441. 4. DUNLOP, M. E., AND LARKINS, R. G. (1984) J. Bid them. 259,8407-8411.
PHOSPHOINOSITIDE
R. G. (1984) Bio~hem 120,820-827. TOOKE, N. E., HALES, C. N., AND HUWON, J. C. (1984) Bio~h~m J. 219,471-480. DELEERS, M., RWSSCHAERT, J.-M., AND MALAISSE, W. J. (1981) Biochem. Biophys. RQS. Cmnmun 98,255-260. DUNLOP, M. E., ANDLARKINS, R. G. (1984) Bio&em. Biophys. RQS. Commun. 118,601-608. MAJERUS, P. W., AND PRESCOTT, S. M. (1982) in Methods in Enzymology (Lands, W. E. M., and Smith, W. L., eds.), Vol. 86, pp. 11-17, Academic Press, New York. MAUCO, G., FAUVEL, J., CHAP, H., AND DOUSTEBLAZY, L. (1984) Biochim. Biophys. Ada 796, 169-177. GOTLIB, L. J. (1982) Bbchim Biophys. Acta 685, 21-26. WRIGHT, J. T., ELMER, W. A., AND DUNLOP, A. T. (1982) Anal. Biochem 125,100-104. GARRETT, J. B., TALWALKER, R. T., AND LESTER, R. L. (1976) J. CQlL Physiol 87,65-70. IRVINE, R. F., LETCHER, A. J., AND DAWSON, R. M. C. (1984) Biochemistry 218,177-185. BEST, L., AND MALAISSE, W. J. (1984) Endocrinology 115,1814-1820.
5. DUNLOP,
M. E., AND LARKINS,
Biophys. RQS Commun
6. 7.
8. 9.
10.
11. 12. 13. 14. 15.
METABOLISM
IN
RAT
ISLET
16. GRIFFIN,
H.
429
MEMBRANE
D., AND HAWTHORNE,
J. N. (1978)
Biochem. J. 176,541-552. 17. Kou~ouzov, 144,16-20. 18. MAUCQ
S., AND MARCHE,
P.
(1982) FEBS Lett.
G., CHAP, H., ANDDOUSTE-BLAZY,
L. (1983)
FEBS Lett. 152,361-364. 19. KOLESNICK,
R. N., AND GERSHENGORN,
M. C. (1984)
J. Biol Chem 259.9514-9519. 20. BELL,
R. M.,
AND COLEMAN,
Rev. B&hem.
R. A.
(1980) Annu.
49,459-480.
21. YEUNG, A., CASOLA, P. G., WONG, C., FELLOWS, J. F., AND POSSMAYER, F. (1979) B&him. Bie
phys. Acta 574,226-239. 22. CRECELIUS,
Biochim
C. A., AND LONGMORE,
W. J. (1983)
Biophys. Acta 750,447-456.
23. BEST, L., AND MALAISSE,
W. J. (1983)
Biochim
Biophys. Acta 750,157-163. 24. PUTNEY, J. W. (1981) Life Sci. 29,1183-1194. 25. BEST, L., AND MALAISSE, W. J. (1984)
Ad.
Biochem. Biophys. 234,253-257. 26. YOSHIOKA,
T., INOUE,
H., AND HOTTA,
Y. (1983)
Biochem. Biophys. Res. Commun. 111,567-573. 27. VOLPI, M., SHA’AFI,
Commun
YASSIN, R., NACCACHE, R. I. (1983) Biochem. 112,957-964.
P. H.,
AND
Biophys. Res.
Phosphoinositide Phosphorylation and Hydrolysis in Pancreatic Islet Cell Membrane MARJORIE
E. DUNLOP’
AND
WILLY
J. MALAISSE2
Labwratcny of Experimental Medicine, Brussels Free University, 115 Boulevard de Waterloo, Brussels 1000, Belgium Received
September
4, 1985
Membranes were isolated from dispersed rat pancreatic islet cells by attachment to Sephadex beads. When these membranes were exposed to [T-~~P]ATP, formation of 32P-labeled phosphatidate, phosphatidylinositol 4-phosphate, and phosphatidylinositol 4,5-bisphosphate was observed. Carbamylcholine, added 10 s prior to lipid extraction, caused a dose-related fall in 32P-labeled phospholipids. The effect of the cholinergic agent was suppressed by atropine, ethylene glycol bis(&aminoethyl ether)-N,N’-tetraacetic acid, and verapamil, and simulated, in part, by an increase in Ca2+ concentration. When the membranes were derived from islet cells prelabeled with [U-‘4C]arachidonate, carbamylcholine stimulation, in addition to decreasing labeled polyphosphoinositides, was accompanied by an increased production of labeled diacylglycerol, without a concomitant increase in labeled phosphatidylinositol. These results indicate that activation of a plasma membrane-associated phospholipase C directed against polyphosphoinositides represents a primary event in the functional response of the pancreactic /3 cell to cholinergic agents. o 1986 Academic press, I~~. Increasing attention is currently being paid to the possible participation of phosphoinositide hydrolysis in the process of insulin release evoked in the pancreatic ,8 cell by nutrient and cholinergic secretagogues (1, 2). The activity of certain enzymes involved in the synthesis of polyphosphoinositides (phosphatidylinositol kinase, phosphatidylinositol 4-phosphate kinase) or their breakdown (phospholipase C) was recently demonstrated in homogenates of neonatal or adult rat islets (3-5). Using a rat P-cell tumor, Tooke et aZ. (6) observed phosphatidylinositol kinase activity in both the plasma membrane and the secretory-granule membrane, as well as phosphatidylinositol4-phosphate phosphatase activity associated with the ‘Present address: University of Melbourne, Department of Medicine, Royal Melbourne Hospital, Parkville, Victoria, Australia. * To whom correspondence should be addressed. 421
plasma membrane. The present work was undertaken to gain further insight into the metabolism of phosphoinositides in membranes isolated from normal rat islets, in particular the enzymatic determinants of changes provoked in such a preparation by cholinergic stimulation. MATERIALS
AND
METHODS
Materials. [U-“ClArachidonic acid (390 mCi/mmol) was obtained from New England Nuclear, Boston, Massachusetts; adenosine 5’-[y-“Pltriphosphate, triethylammonium salt (>5000 Ci/mmol) from Amersham International plc, Amersham, Bucks, U. K.; RPM1 medium from Flow Laboratories, Ayrshire, U. K.; verapamil HCl from Knoll A.G., Ludwigshafen, FRG; and all other chemicals and phospholipids from Sigma Chemical Company, Ltd. Preparation and incubation of islet cells. Pancreatic islets isolated by the collagenase method from fed albino rats were further disrupted into isolated cells by mechanical agitation for 10 min in Gas+-free Hanks’ solution containing EGTA (1.0 mM), bovine albumin 0003-9861/86 Copyright All rights
$3.00
0 1986 by Academic Press. Inc. of reproduction in any form reserved.
422
DUNLOP AND MALAISSE
(5 mg/ml), and collagenase (2 mg/ml) (7). After filtration through a nylon screen the islet cells were pelleted by centrifugation for 5 min at 600s and, in one series of experiments, resuspended in bicarbonatebuffered medium RPM1 1640, containing glucose (11.1 mM) and [U-‘“Clarachidonic acid (12.8 pM) at a cell concentration of 8 X 106/ml. The cells were incubated with constant oxygenation (02:COz, 95:5, v:v) for 120 min at 37°C. Preparation of id& cell membranes. Islet cells either freshly isolated or following [U-“Clarachidonic acid labeling were pelleted by centrifugation for 5 min at 6OOg and resuspended in a Tris-HCl buffer (10 mM, pH 5.2) containing 217 mM sucrose and 46.5 mM sodium acetate at a concentration of 8 X lo6 cells/ml. The cells (8 X 106/sample) were attached to a positively charged, beaded derivative of Sephadex (Cytodex 1, Pharmacia, Uppsala, Sweden) as described previously (8). The cells were lysed with hypotonic buffer (TrisHCl, 10 mM, pH 8.0,4.0 ml) for 20 min. At this time, where indicated, preparations of beads and attached membranes were sonicated at low amperage. The beads and attached membranes (sonicated or unsonicated) were again centrifuged and the supernatant buffer was removed entirely. On occasion islet cells were homogenized in Potter-Elvehjem tubes in TrisHCl buffer (10 mM, pH 8). Phosphwglutim of islet cell hmmgendes and idated membranes. A homogenate of islet cells (8 X lo6 cells/ 0.05 ml) or beads and attached membranes from an equivalent number of islet cells were phosphorylated using adenosine 5’-triphosphate. Phosphorylation was started by the addition of 200 pl of sodium acetate buffer (50 mM, pH 6.8) containing MgCla (2 mM) and ATP (0.2 mM). In those experiments in which islet membranes prelabeled with arachidonic acid were used, phosphorylation was with the unlabeled nucleotide only. In the remaining experiments using islet cell homogenates and isolated membranes from non[U-‘“Clarachidonate-labeled cells, the ATP (final concentration: 0.2 mM) included [y-“P]ATP (4 pCi/sample). In additional experiments a sonicate of phosphatidylinositol (brain extract type I, Sigma Chemical Company Ltd.), confirmed by thin-layer chromatography (see below) to contain phosphatidylinositol as the sole inositide phospholipid, was prepared in Na acetate buffer, pH 6.8, and added to the sonicated beads and attached membranes to give a final concentration of lo-’ M, immediately prior to the addition of [y-aaP]ATP. After 5 min at 21°C and 10 s prior to lipid extraction, 20 ~1 of carbamylcholine or Ca’+/ EGTA-buffered solutions was added to the incubation medium. Likewise, buffered solutions of the potential antagonists (EGTA, verapamil, and atropine) were added 5 s prior to carbamylcholine or Ca’+/EGTA.
The incubation was terminated by the addition of CHCla:CHsOH:13NHCl, 200:100:1, v:v:v (1.0 ml), and the lipids were extracted for 30 min with constant mixing. Two phases were generated by the addition of KC1 (1 M) and CHCla. The chloroform phase was dried under vacuum at 20°C and the lipid resuspended in 30 ~1 CHCls for application (lo-r1 aliquots) to silica gel 60 plates (Merck 5721, Darmstadt, FRG) preloaded with potassium oxalate (1% in CHIOH:H20, 23, v:v) and air dried at 110°C for 1 h before use. Phospholipids were separated in a solvent system of chloroform: acetone:methanol:acetic acid:H*O (80:34:30:24:16, v:v:v:v:v) under saturation conditions. Of duplicate plates, one was autoradiographed using Kodak X-OMat AR film, and the other scraped, according to the position of authentic standards (after iodine staining) and the gel suspended in 2.5 ml methanol followed by the addition of 10 ml Aquasol(New England Nuclear) for scintillation counting. Under these conditions authentic standards gave Rf values of 1.0, 0.70, 0.51, 0.44, 0.36, 0.33, 0.26, and 0.20 for palmitic acid, phosphatidic acid, lysophosphatidic acid, phosphatidylserine, phosphatidylcholine, phosphatidylethanolamine, phosphatidylinositol, phosphatidylinositol 4-phosphate, and phosphatidylinositol 4,5-bisphosphate, respectively. In addition [U-%]arachidonic acid-labeled lipids were separated following thin-layer chromatography in a solvent system of petroleum ether:diethyl ether:acetic acid (70:30:1, v:v:v) in which the Rfvalues for authentic standards were 0.79,0.40, 0.30, 0.29, 0.27, and 0.08 for triacylglycerol, palmitic acid, cholesterol, 1,3-diolein, 1,2-diolein, and monoacylglycerol, respectively, phospholipids remaining at the origin. In this manner phospholipids and neutral lipids were separated from [U-“Clarachidonic acidlabeled homogenates and islet cell membranes prior to and following phosphorylation and from homogenates and islet cell membranes following identical phosphorylation using [y-azP]ATP. Presentation of results. In the phosphorylation experiments employing [y-**P]ATP, the results are expressed as femtomoles of =P transferred from ATP over 5 min incubation into the membranes isolated from lo6 isolated islet cells. Following exposure to [U-%]arachidonic acid the results are expressed as picomoles of arachidonate incorporated into lo6 cells or membrane equivalent. All results are presented as means; followed by either the range of individual variations when only two observations were made or, otherwise, the SEM together with the number of individual observations (n). Student’s t test was used to determine the levels of statistical significance. RESULTS
32PLabeling of Membrane Lipids ’ Abbreviation used: EGTA, aminoethyl ether)-N,N’-tetraacetic
ethylene glycol acid.
bis(@-
Over 5 min incubation in the presence of nonsonicated mem-
[T-~~P]ATP (0.2 mM),
PHOSPHOINOSITIDE
METABOLISM
branes incorporated 200 f 14 and 258 + 20 fmol of 32P per lo6 cells (n = 5-11) in phosphatidate and phosphatidylinositol 4phosphate, respectively (Fig. 1). The incorporation of radioactivity into phosphatidylinositol 4,5-bisphosphate was close to the lower limit of detection (~10 fmol/106 cells). With sonicated membranes, the incorporation of 32P amounted to 230 + 22 fmol/106 cells (n = 15) in phosphatidylinositol 4-phosphate, whereas labeled phosphatidate and phosphatidylinositol 4,5bisphosphate were not detected over this short period of incubation. If a brain extract containing phosphatidylinositol as virtually the sole inositide was added to the sonicated membranes, the labeling of polyphosphoinositides increased to 634 + 48 fmol/106 cells (n = 5), of which 73.1 t 0.5 and 26.9 + 0.5% (n = 4) appeared
Phosphatidate
Lysophosphatidate
PtdIns4
P
PtdIns4,5P,
Origin
1
2
3
FIG. 1. Autoradiograph of TLC of lipid extracts from a crude islet homogenate (lane l), nonsonicated membranes (lane 2), and sonicated membranes exposed to a brain extract (lane 3) after 5 min incubation with [Y-~P]ATP. The position of authentic standards is indicated (PtdInslP, phosphatidylinositol I-phosphate; PtdIns4,5Pa, phosphatidylinositol4,5-bisphosphate). The arrow points to an unidentified lipid-extractable phosphorylation product also encountered in prior studies (15, 26,27).
IN
RAT
ISLET
423
MEMBRANE
as phosphatidylinositol 4-phosphate and phosphatidylinositol4,5-bisphosphate, respectively (Fig. 1). In the presence of the brain extract, no labeled phosphatidate was detected. By comparison, in a crude islet homogenate, the labeling of phosphatidate averaged 102 + 9, that of phosphatidylinositol 4-phosphate 840 + 133, and that of phosphatidylinositol 4,5-bisphosphate 109 + 4 fmol/106 cells (n = 2-4). In this case, the addition of the brain extract failed to cause an obvious change in the labeling of either phosphatidate or phosphatidylinositol 4,5-bisphosphate, but almost doubled that of phosphatidylinositol 4-phosphate. Hydrolysis of ‘2P-Labeled Polyphosphoinositides
Membrane
When the membrane lipids were labeled with [y-32P]ATP and carbamylcholine added to the medium 10 s prior to extraction of lipids, a dose-related decrease in the amount of labeled phosphatidylinositol 4-phosphate was observed (Table I). Essentially the same results were obtained in nonsonicated (Table I) and sonicated membranes. In the latter case, the relative fall in [32P]phosphatidylinositol 4-phosphate increased from 36.6 + 2.3% to 79.1 k 1.2% and 93.0 + 0.1% as the concentration of carbamylcholine was raised from 10 PM to 500 I.LM and 2.0 InM, respectively. As little as 10 I.LM carbamylcholine was sufficient to cause a one-third decrease (P < 0.001) in [32P]phosphatidylinositol 4-phosphate, virtually complete disappearance of this compound being seen in response to 2.0 mM carbamylcholine. Atropine (10 PM) and verapamil(lO0 PM), when added 5 s before carbamylcholine, antagonized the effect of the cholinergic agent, although failing to affect significantly the recovery of labeled phosphatidylinositol 4-phosphate in the absence of carbamylcholine (Table I). In the presence of verapamil, the carbamylcholine-induced decrease in [32P]phosphatidylinositol 4-phosphate represented no more than 27.7 k 7.5% (n = 4; P < 0.005) of the paired value found in the absence of the calcium antagonist. When EGTA (100 PM) was added 5 s prior to carbamylcholine, the decrease in labeled phosphatidylino-
424
DUNLOP TABLE
AND
I
EFFECT OF CAaB.4MncuoLINE, VERAPAMIL, AND ATROPINE UPONTHE LEVELOF "P-LABELED PHOSPHATIDYLINOSITOL d-PHOSPHATE IN NONSONICATED ISLET CELL MEMBRANES Nil Carbamylcholine (10 ELM) Carbamylcholine (100 PM) Carbamylcholine (500 PM) Carbamylcholine (2000 PM) Verapamil (0.1 mM) Verapamil (0.1 mM) + carbamylcholine (10 pM) Verapamil (0.1 mM) + carbamylcholine (500 pM) Atropine (10 pM) Atropine (10 PM) + carbamylcholine (500 pM)
100.0 67.7 66.6 20.8 11.2 110.2
f + f f + f
1.3 2.7 1.5 0.8 1.8 3.6
(9) (4) (3) (6) (5) (2)
93.8 r?I 2.8 (2) 71.3 + 8.5 (2) 98.9 + 0.3 (2) 87.7 f 3.5 (2)
Note. All values are expressed as percentages of the mean control value found within the same experiment(s). Such a control value averaged 271 f 26 fmol/ lo6 cells.
sitol 4-phosphate was prevented at a 100 concentration of the cholinergic agent and was inhibited by 71% at a 2.0 mM concentration of carbamylcholine. In view of these findings, the effect of Ca2+ itself upon the hydrolysis of phosphatidylinositol 4phosphate was examined. When a fixed amount of EGTA (10 PM) and increasing amounts of CaC12 were added to the medium 10 s prior to lipid extraction, a dose-related fall in labeled phosphatidylinositol 4-phosphate was observed in sonicated membranes (Fig. 2), with an apparent K, for Ca2+ close to 5 PM. Verapamil (0.2 mM) inhibited, by no more than 26.2%, the response to 100 PM Ca2+. Incidentally, EGTA alone slightly increased the amount of labeled phosphatidylinositol 4-phosphate to 113.1 f 3.1% (n = 4; P < 0.025) of the paired control value found in the absence of EGTA. In the experiments so far presented, the effect of carbamylcholine or Ca2+ upon phosphatidylinositol 4,5-bisphosphate could not be assessed, in view of the minimal formation of labeled phosphatidylinositol4,5-bisphosphate in the islet membranes. Therefore, a series of experiments was conducted in the presence of the brain PM
MALAISSE
extract and sonicated membranes, in which case sizable amounts of labeled phosphatidylinositol 4,bbisphosphate are formed (see above). Under these conditions, carbamylcholine (100 pM) decreased the amount of labeled phosphatidylinositol4,5bisphosphate from a control value of 142 to 17 fmol/106 cells. Likewise, increasing concentrations of Ca2+ caused a dose-related decrease in labeled phosphatidylinositol4,5-bisphosphate (Fig. 2), with an apparent K, for Ca2+ close to ‘7 PM. In these experiments, the cholinergic agent and the Ca2+ ion also caused the disappearance of labeled phosphatidylinositol 4-phosphate. However, as illustrated in Fig. 2, higher concentrations of Ca2+ were required to cause a maximal fall in [32P]phosphatidylinositol 4-phosphate in the presence than in the absence of the brain extract. This difference could be due, in part at least, to the fact that, in the presence of the brain extract, the hydrolysis of labeled phosphatidylinositol4-phosphate may coincide with its generation from labeled phosphatidylinositol 4,5-bisphosphate through the action of a calcium-sensitive phosphomonoesterase. It should be noted that, in the presence of the brain extract, EGTA alone (10 PM) increased the amount of labeled phosphatidylinositol 4,5-bisphosphate to 117.1 + 4.4% (n = 4; P < 0.05) of the paired control value found in the absence of EGTA. The responsiveness to both EGTA and CaC12 was thus qualitatively similar in the case of phosphatidylinositol 4-phosphate and phosphatidylinositol4,5bisphosphate, respectively. When glucose (16.7 mM) was added to the medium 10 s prior to extraction of lipids, no change in labeled phosphatidylinositol 4-phosphate was observed in nonsonicated membranes. Hydrolysis of s2P-Labeled Membrane Phosphatidate
The addition of carbamylcholine and, to a lesser extent, that of Ca2+ 10 s prior to lipid extraction also decreased the amount of labeled phosphatidate formed by nonsonicated membranes exposed to [y=P]ATP. Thus, relative to the control value
PHOSPHOINOSITIDE
METABOLISM
IN
RAT
ISLET
425
MEMBRANE
500-
0 400, = u ID”
0
< z E 300. 0 ” ” ,?I .-2 2!
Lo,
200.
.----,
E
X... ......,
T IL 100,
O-
. o-
/
“,1 q. \ :.\ : \..\: i “,*. .*- .----- -- --.- ------1--, .-.*..... .......... ..a.............+..... 10
5 Added
25 CaC12
50
-c--e c....*( t 103
(Jo M)
FIG. 2. Effect of increasing concentrations of CaCla (logarithmic scale) added together (10 PM) 10 s prior to lipid extraction upon the recovery of q-labeled phosphatidylinositol (circles) and phosphatidylinositol 4,5-bisphosphate (crosses) in sonicated membranes the absence (closed circles) or presence (open circles and crosses) of a brain extract. refer to two individual determinations.
(200 f 14 fmol/106 cells), carbamylcholine (10, 100, 500, and 2000 PM) caused a doserelated decrease in [32P]phosphatidate (n = 6, P < 0.001). The carbamylcholine-induced fall in [32P]phosphatidate averaged 37 +- 1 and 74 + 1% at the lowest and highest concentration of the cholinergic agent, respectively. At a 100 I.~M concentration, Ca2+ caused a 22 + 2% decrease in [32P]phosphatidate. Verapamil(lO0 PM) and atropine (10 PM) added 5 s prior to the introduction of carbamylcholine abolished the effect of the cholinergic agent. Thus, in the presence of verapamil and atropine, respectively, the readings obtained after addition of carbamylcholine (10 to 500 PM) averaged 98 * 6 and 94 + 1% of the control value. Neither verapamil nor atropine affected the amount of [32P]phosphatidate measured in the absence of carbamylcholine, the experimental values averaging 96 + 6% of the corresponding control value.
[lJ-14CjArachicbnate Membrane Lipids
Labeling
with EGTA I-phosphate incubated in Mean values
of
In the experiments so far presented, the liberation of diacylglycerol from polyphosphoinositides could obviously not be monitored. As this information is required to distinguish between the action of a phosphomonoesterase and a phosphodiesterase, the fate of membrane lipids labeled with [U-14C]arachidonate was examined. For this purpose, the isolated islet cells were exposed for 120 min to [U-‘“Clarachidonate prior to membrane isolation. For the sake of comparison, the isolated membranes were then treated in a manner comparable to that used in the first series of experiments, being exposed to unlabeled ATP for 5 min prior to lipid extraction. The composition of the [U-14C]arachidonate-labeled lipids extracted from the membranes after the 5-min incubation in
426
DUNLOP
AND
the presence of ATP is illustrated in Table II. The amount of labeled phosphatidylinositol was low relative to either the total amount of labeled phospholipids (3.9 + 0.8%) or the total amount of inositides (53.4 f 12.1%). The labeled phospholipids represented 54.8 + 9.2% of the total amount of labeled lipids (n = 3 in all cases). This pattern differs in several respects from that found in the crude islet homogenate treated under identical conditions. Thus, in the crude homogenate, labeled phospholipids represented only 28.8 f 5.1% of total labeled lipids. Moreover, the relative amount of labeled phosphatidylinositol was higher in the crude homogenate, representing respectively 17.4 + 1.2% of the labeled phospholipids and 73.7 f 2.0% of the labeled phosphoinositides. Incidentally, the total amount of labeled lipids in the crude homogenate was about twice higher than that in the isolated membranes. Over the 5-min period of exposure to ATP (0.2 mM), no obvious change was seen in the amount of labeled phosphatidylethanolamine and phosphatidylcholine or arachidonate. The amount of labeled polyphosphoinositides increased, this being matched by a decrease in labeled phosphatidylinositol, with a mean value 1.11 + 0.01 pmol/106 cells. There was a minor increase in labeled phosphatidate (0.40 pmol/106 cell) of the same order of magnitude as that observed for the 32P labeling of this lipid in nonsonicated membranes. Labeled diacylglycerol decreased over the same 5-min period, coinciding with a comparable inTABLE EFFECTOF
CARBAMYLCHOLINE LIPIDS
MALAISSE
crease in monoacylglycerol. These changes (2.31 k 0.25 pmol/106 cells) were accentuated two- to threefold when Ca2’ (100 PM) was added 10 s prior to lipid extraction, suggesting that they may reflect, in part at least, the activity of a Me-dependent and Ca2+-responsive diacyiglycerol lipase acting selectively at the sn-1 position (9710). Eflect of Carbamylcholine [U-14C]Arachidonate-Labled Membrane Lipids
When added 10 s prior to lipid extraction carbamylcholine (0.5 mM) failed to affect significantly the recovery of the labeled phosphatidate, phosphatidylinositol, the combined fraction of phosphatidylethanolamine and phosphatidylcholine, or arachidonate (Table II). The polyphosphoinositides were decreased by 1.81 f 0.44 pmol/106 cells, whereas diacylglycerol was increased by 3.18 f 0.93 pmol/106 cells. When either atropine (10 PM), verapamil (0.2 mM), or EGTA (0.1 mM) was added 5 s prior to the introduction of carbamylcholine (0.5 mM), the response to the cholinergic agent was prevented, the mean values for the polyphosphoinositides and diacylglycerol now averaging 119.5 k 13.1 and 104.8 f 8.7%, respectively, of the mean control value (n = 5 in both cases). If the present data are compared to those obtained in the 32P-labeling experiments, it becomes obvious that the relative magnitude of the carbamylcholine-induced fall II
(0.5 ~M)UPONTHELEVELOF[U-"C&~RACHIDONATE-LABELED IN ISLET CELL NONSONICATED MEMBRANES Control (pmol/lOs cells)
Phosphatidate Phosphatidylinositol Phosphatidylethanolamine Polyphosphoinositides Arachidonate Diacylglycerol
Note. Mean with
values (? SEM) the statistical significance
upon
+ phosphatidylcholine
16.35 f 1.91 10.36 f 1.45 207.68 f 21.80 6.64 k 0.64 144.65 f 10.57 7.94 f 0.57
refer to three individual determinations of changes evoked by carbamylcholine
Carbamylcholine (% of control) 105.5 2 100.8 f 103.3 f 72.8 f 109.3 f 152.2 +
5.4 6.5
1.2 5x** 4.1
12.6*
in all cases, and are shown (*P < 0.06; **P < 0.05).
together
PHOSPHOINOSITIDE
in labeled polyphosphoinositides greater in the latter case.
METABOLISM
was much
DISCUSSION
With the present method of membrane isolation, the material attached to the Sephadex beads corresponds to a plasma membrane-enriched fraction. Sonication of the beads with attached membranes aims at further purification of the plasma membrane-rich material (10-12). The present data demonstrate the presence in islet cell membranes of the kinase activities catalyzing the phosphorylation of diacylglycerol, phosphatidylinositol, and phosphatidylinositol 4-phosphate. The presence of the latter enzyme is substantiated by the data obtained in the concomitant presence of sonicated membranes and brain extract. The absence of phosphatidate labeling in sonicated, as distinct from nonsonicated, membranes probably reflects mainly a loss of substrate rather than enzyme activity. It is indeed possible to restore phosphatidate labeling in sonicated membranes in the presence of brain extract and 12-diolein (data not shown). Moreover, cellular lysis is known to be accompanied by breakdown of inositides and phosphatidic acid, unless Ca2+ is rigorously excluded from the lysing buffer (13). Carbamylcholine caused a dose-related fall in the membrane content of 32P-labeled phosphatidylinositol4-phosphate. This effect was antagonized by atropine and verapamil, suggesting that it corresponded to a calcium-sensitive process mediated by muscarinic receptors. In this respect, it should be stressed that these experiments were conducted in the absence of any Cachelating agent and, hence, in the presence of some contaminating Ca2+ (~5 PM). The effect of the cholinergic agent being assessed over a period of 10 s, it seems most likely that the decrease in 32P-labeled phosphatidylinositol4-phosphate reflects a process of rapid hydrolysis rather than inhibition of phosphorylation. The loss of labeled polyphosphosinositide was not accompanied by the appearance of any 32Plabeled phospholipid, suggesting that it was attributable to activation of Ca2+-de-
IN
RAT
ISLET
MEMBRANE
427
pendent phosphomonoesterase or phosphodiesterase. Whatever its nature, the enzyme was apparently activated by Ca2+ itself as judged from the effect of both EGTA and increasing concentrations of added CaC12. In this respect, it should be kept in mind that the hydrolysis of phosphoinositides was here examined in the absence of detergent, thus maintaining the physical state of the membrane and, hence, optimal sensitivity to Ca2+ (14). The finding that the calcium antagonist verapamil inhibited the hydrolytic response to carbamylcholine and, to a lesser extent, calcium could be related to competition between Ca2+ and verapamil for a membrane site, although an unspecific action of this lipophilic drug cannot be ruled out. A further series of experiments was conducted to gain information on the precise modality of polyphosphoinositide hydrolysis in the islet membrane preparation. Thus by exposing the islet cells to [U-‘*C]arachidonate prior to membrane isolation, it was possible to demonstrate the formation of diacylglycerol in response to carbamylcholine. This finding supports the participation of a phosphodiesterase in the hydrolysis of polyphosphoinositides in good agreement with data obtained in intact islet cells (15). Since the cholinergic agent failed to affect the recovery of [U-‘*C]arachidonate-labeled phosphatidylinositol, it would appear that the phospholipase C activated by carbamylcholine acted preferentially on polyphosphoinositides, as distinct from phosphatidylinositol. Incidentally, the experiments conducted in the presence of the brain extract indicate that the carbamylcholine-responsive enzyme(s) catalyzes the hydrolysis of both phosphatidylinositol 4-phosphate and phosphatidylinositol4,5-bisphosphate. It should be stressed that the hydrolysis of polyphosphoinositides needs not to be attributed solely to a phospholipase C mechanism. The hydrolysis of phosphatidylinositol 4,5-bisphosphate by a membrane-bound phosphomonoesterase has been reported in several tissues (16-19). However, at variance with the findings reported by Tooke et al. (6), no evidence of phosphomonoesterase acting on phosphatidylinositol 4-phosphate was obtained in
428
DUNLOP
AND
the present study, as neither carbamylcholine (Table II) nor Ca” (data not shown) increased the recovery of [U-‘“Clarachidonate-labeled phosphatidylinositol. An unexpected finding consisted in the rapid hydrolysis of 32P-labeled phosphatidic acid evoked by carbamylcholine in nonsonicated membranes. The rapidity of such a response argues against any significant participation of an inhibitory effect of the cholinergic agent upon diacylglycerol kinase. Activation of a phospholipase A2 seems also improbable, since no %P-labeled lysophosphatidate was detected after exposure of the nonsonicated membranes to carbamylcholine (Fig. 1). This leaves the possibility of activation of a phosphatidate phosphatase, such as a Mga+-dependent speciSc or nonspecific phosphohydrolase producing diacylglycerol and Pi (20) or, as reported in rat lung, a microsomal lipase producing glycerol phosphate and fatty acid (21). The knowledge that the pH optimum of the specific phosphatidate phosphohydrolase is met by the condition of the present experiments (22) and the fact that the carbamylcholine-stimulated formation of [U-‘4C]arachidonate-labeled diacylglycerol somewhat exceeded that accounted for by the fall in labeled polyphosphoinositides would support the former possibility. Whatever its precise nature, the carbamylcholine-responsive enzyme appeared to act preferentially upon the pool of phosphatidate generated through the action of a membrane-associated kinase and not on the much larger pool of [U-‘“Clarachidonate-labeled phosphatidate. An effect of the cholinergic agent upon a restricted pool of phosphatidate in the isolated membrane is not necessarily incompatible, therefore, with the knowledge that, in intact islet cells, cholinergic agents provoke a rapid increase in the 32P labeling of phosphatidate (23). Indeed, in intact cells, the removal from the membrane of diacylglycerol-derived phosphatidate by a putative transfer protein (24) mediating the rapid transfer of phosphatidate to intracellular sites may channel the latter phospholipid away from a membrane-associated phosphohydrolase. Even in intact cells, the phosphatidate formed through the hydrolysis of phos-
MALAISSE
phoinositides may not be the sole precursor for phosphatidylinositol found in the membrane. As a matter of fact, the [U-r4C]arachidonate labeling of membraneassociated phosphatidate and phosphatidylinositol strongly suggests de novo synthesis of phosphatidate (25). In the present experiments, the membrane-associated pool of [U-‘4C]arachidonate-labeled phosphatidylinositol, which may only represent a fraction of the total membrane-associated phosphatidylinositol, seems to act as a preferential substrate for phosphorylation to polyphosphoinositides. Indeed, over 5 min incubation in the presence of unlabeled ATP, the conversion of [U“C]arachidonate-labeled phosphatidylinositol to polyphosphoinositides was of the same order of magnitude as the 32P labeling of the polyphosphoinositides in membranes exposed for the same time to [T-~~P]ATP. In conclusion, the present data afford direct support to the view that activation of a plasma membrane-associated phospholipase C directed against polyphosphoinositides represents the primary event in the cholinergic pathway for stimulation of insulin release. In the isolated membrane, the modulation by calcium chelator and calcium antagonist of the response to carbamylcholine points to a key role for membrane-bound Ca2+ in the coupling of muscarinic receptor occupation to polyphosphoinositide hydrolysis. ACKNOWLEDGMENTS This work was supported in part by grants from the Belgian Foundation for Scientific Medical Research and the Belgian Ministry of Scientific Policy (Brussels, Belgium). M.E.D. was supported by the National Health and Medical Research Council of Australia. We thank F. Malaisse-Lagae and M. Mahy for assistance and C. Demesmaeker for secretarial help. REFERENCES 1. BEST, L., AND MALAISSE, W. J. (1983) Diabetologia 25.299-305. 2. BEST, L., DUNLOP, M., AND MALAISSE, W. J. (1984) Eqxtientia 40,1085-1091. 3. SCHREY, M. P., AND MONTAGUE, W. (1983) Biochm. J. 216,433-441. 4. DUNLOP, M. E., AND LARKINS, R. G. (1984) J. Bid them. 259,8407-8411.
PHOSPHOINOSITIDE
R. G. (1984) Bio~hem 120,820-827. TOOKE, N. E., HALES, C. N., AND HUWON, J. C. (1984) Bio~h~m J. 219,471-480. DELEERS, M., RWSSCHAERT, J.-M., AND MALAISSE, W. J. (1981) Biochem. Biophys. RQS. Cmnmun 98,255-260. DUNLOP, M. E., ANDLARKINS, R. G. (1984) Bio&em. Biophys. RQS. Commun. 118,601-608. MAJERUS, P. W., AND PRESCOTT, S. M. (1982) in Methods in Enzymology (Lands, W. E. M., and Smith, W. L., eds.), Vol. 86, pp. 11-17, Academic Press, New York. MAUCO, G., FAUVEL, J., CHAP, H., AND DOUSTEBLAZY, L. (1984) Biochim. Biophys. Ada 796, 169-177. GOTLIB, L. J. (1982) Bbchim Biophys. Acta 685, 21-26. WRIGHT, J. T., ELMER, W. A., AND DUNLOP, A. T. (1982) Anal. Biochem 125,100-104. GARRETT, J. B., TALWALKER, R. T., AND LESTER, R. L. (1976) J. CQlL Physiol 87,65-70. IRVINE, R. F., LETCHER, A. J., AND DAWSON, R. M. C. (1984) Biochemistry 218,177-185. BEST, L., AND MALAISSE, W. J. (1984) Endocrinology 115,1814-1820.
5. DUNLOP,
M. E., AND LARKINS,
Biophys. RQS Commun
6. 7.
8. 9.
10.
11. 12. 13. 14. 15.
METABOLISM
IN
RAT
ISLET
16. GRIFFIN,
H.
429
MEMBRANE
D., AND HAWTHORNE,
J. N. (1978)
Biochem. J. 176,541-552. 17. Kou~ouzov, 144,16-20. 18. MAUCQ
S., AND MARCHE,
P.
(1982) FEBS Lett.
G., CHAP, H., ANDDOUSTE-BLAZY,
L. (1983)
FEBS Lett. 152,361-364. 19. KOLESNICK,
R. N., AND GERSHENGORN,
M. C. (1984)
J. Biol Chem 259.9514-9519. 20. BELL,
R. M.,
AND COLEMAN,
Rev. B&hem.
R. A.
(1980) Annu.
49,459-480.
21. YEUNG, A., CASOLA, P. G., WONG, C., FELLOWS, J. F., AND POSSMAYER, F. (1979) B&him. Bie
phys. Acta 574,226-239. 22. CRECELIUS,
Biochim
C. A., AND LONGMORE,
W. J. (1983)
Biophys. Acta 750,447-456.
23. BEST, L., AND MALAISSE,
W. J. (1983)
Biochim
Biophys. Acta 750,157-163. 24. PUTNEY, J. W. (1981) Life Sci. 29,1183-1194. 25. BEST, L., AND MALAISSE, W. J. (1984)
Ad.
Biochem. Biophys. 234,253-257. 26. YOSHIOKA,
T., INOUE,
H., AND HOTTA,
Y. (1983)
Biochem. Biophys. Res. Commun. 111,567-573. 27. VOLPI, M., SHA’AFI,
Commun
YASSIN, R., NACCACHE, R. I. (1983) Biochem. 112,957-964.
P. H.,
AND
Biophys. Res.