Ethanol-induced mobilization of calcium by activation of phosphoinositide-specific phospholipase C in intact hepatocytes.

The short-term effects of ethanol on calcium homeostasis were studied in isolated hepatocytes. Ethanol caused a rapid transient activation of phosphorylase not associated with changes in cAMP levels which peaked after 20-30 s and declined slowly over a period of 5-10 min. Maximal activation was found with 200 mM ethanol, and a significant effect was observed at 25 mM ethanol. Similar effects were induced by other organic solvents and by halothane, with more hydrophobic agents being effective at lower concentrations. In hepatocytes loaded with the intracellular calcium indicator quin2, the addition of ethanol caused a transient increase in cytosolic free calcium, with a kinetic pattern compatible with its involvement in the activation of phosphorylase. Pretreatment of the hepatocytes with phenylephrine or vasopressin to deplete the hormone-sensitive calcium pools in the cells prevented the ethanol-induced calcium mobilization. In 32P-labeled hepatocytes addition of ethanol caused a small (5-7%) decrease in the level of [32P]phosphatidylinositol 4,5-bisphosphate and a 10-15% increase in [32P]phosphatidylinositol 4-phosphate and [32P]phosphatidic acid. In hepatocytes labeled with myo-[3H]inositol, ethanol induced a 50-100% increase in the levels of inositol 1,4,5-trisphosphate, inositol 1,3,4-trisphosphate, and inositol bisphosphate. The changes in the inositol 1,4,5-trisphosphate level due to ethanol paralleled the time course of the elevation of cytosolic free calcium levels and activation of phosphorylase a. The effects of ethanol were comparable to those of a physiologic (1 nM) dose of vasopressin; however, unlike with vasopressin, the inositol phosphates and cytosolic calcium levels declined to basal levels 2 min after the addition of ethanol. These results indicate that ethanol, in common with calcium-mobilizing hormones, activates hormone-sensitive phosphoinositide-specific phospholipase C. The resulting changes in inositol 1,4,5-trisphosphate can account for the mobilization of intracellular calcium and the consequent activation of phosphorylase by ethanol.

The short-term effects of ethanol on calcium homeostasis were studied in isolated hepatocytes. Ethanol caused a rapid transient activation of phosphorylase not associated with changes in CAMP levels which peaked after 20-30 s and declined slowly over a period of 5-10 min. Maximal activation was found with 200 mM ethanol, and a significant effect was observed at 25 mM ethanol. Similar effects were induced by other organic solvents and by halothane, with more hydrophobic agents being effective at lower concentrations. In hepatocytes loaded with the intracellular calcium indicator quina, the addition of ethanol caused a transient increase in cytosolic free calcium, with a kinetic pattern compatible with its involvement in the activation of phosphorylase. Pretreatment of the hepatocytes with phenylephrine or vasopressin to deplete the hormone-sensitive calcium pools in the cells prevented the ethanol-induced calcium mobilization.
In 32P-labeled hepatocytes addition of ethanol caused a small (5-7%) decrease in the level of [32P]phosphati-dylinositol4,5-bisphosphate and a 10-15% increase in [s2P]phosphatidylinositol 4-phosphate and [32P]phosphatidic acid. In hepatocytes labeled with r n y~- [~H ] inositol, ethanol induced a 50-100% increase in the levels of inositol l14,5-trisphosphate, inositol 1,3,4trisphosphate, and inositol bisphosphate. The changes in the inositol l,.i,S-trisphosphate level due to ethanol paralleled the time course of the elevation of cytosolic free calcium levels and activation of phosphorylase a. The effects of ethanol were comparable to those of a physiologic (1 nM) dose of vasopressin; however, unlike with vasopressin, the inositol phosphates and cytosolic calcium levels declined to basal levels 2 min after the addition of ethanol.
These results indicate that ethanol, in common with calcium-mobilizing hormones, activates hormone-sensitive phosphoinositide-specific phospholipase C. The resulting changes in inositol 1,4,5-trisphosphate can account for the mobilization of intracellular calcium and the consequent activation of phosphorylase by ethanol.
Ethanol interacts with biological membranes and affects their physical and chemical properties (see Ref. 1 for a review). The consequences of these interactions for cellular functions have not yet been well defined. A variety of cells responds to the continued presence of ethanol by inducing structural * This work was supported by United States Public Health Service Grants AM38461, AA07186, AA07211, and AA07215. 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. changes in their membranes so as to make them more resistant to the physical effects of ethanol. Both the acute effects of ethanol and the adaptations occurring in cells chronically exposed to ethanol can influence the activity of membranebound enzymes, including ion transport systems (1).
In an earlier report (2), we demonstrated that the Ca2+ pump activity in microsomal preparations from rat liver is affected by ethanol treatment. Other studies (3-5) reported an effect of ethanol on the activity of the Ca2+-ATPase in brain and erythrocyte membranes and on voltage-operated Ca2+ channels in neurons. These observations suggest the possibility that ethanol interferes with normal mechanisms of calcium homeostasis in liver and other tissues.
Calcium serves an important function as an intracellular messenger (6). The steady state level of cytosolic free Caz+ in liver cells is maintained in the range of 0.1-0.2 PM (6-9) by the action of calcium transport systems in the plasma membrane, endoplasmic reticulum, and mitochondria, and by binding to intracellular constituents. Certain hormones and other agonists (nl-adrenergic agonists, vasopressin, angiotensin 11, ATP) can increase cytosolic calcium levels by mobilizing intracellular calcium stores (6-10). A primary step in the action of such agonists is the activation of a phosphoinositidespecific phospholipase C in the plasma membrane, resulting in the intracellular release of inositol 1,4,5-trisphosphate (Ins (1,4,5)P3)' and diacylglycerol (see Refs. 11-13 for reviews). Ins (1,4,5)P3 causes the release of calcium from specific nonmitochondrial pools (14). The resultant increase in cytosolic free Ca2+ activates Ca2+-dependent protein kinases and other Ca2+-dependent enzymes (11)(12)(13). Diacylglycerol, in the presence of basal levels of calcium, activates protein kinase C (15), which can phosphorylate and presumably regulate the function of several proteins (12); most of these proteins have not yet been functionally characterized. Thus, receptor-mediated activation of phospholipase C leads to a complex network of intracellular signals resulting in an increased phosphorylation of various soluble and membrane-bound proteins (16).

Preparations and Incubation Conditions
Isolated hepatocytes were prepared from male Sprague-Dawley rats, starved overnight. The collagenase perfusion method of Berry and Friend (23) was used, essentially as described by Meyer et al. (24). The washed cell suspension (30-50 mg of protein/ml) was kept at room temperature until use. Viability of the cells was more than 90%, as measured by trypan blue exclusion.
Incubations were carried out in 25-ml flasks in a shaking water bath at 37 "C in Krebs-Ringer bicarbonate buffer, containing 2% bovine serum albumin and either 10 D M lactate plus 1 mM pyruvate or 15 mM glucose. The pH was 7.4 and the gas phase contained 95% 0,. 5% C02. The incubations contained 3-5 mg of cell protein/ml (1.8-2.8 X lo6 cells/ml) in a final volume of 5 ml. For the determination of phosphorylase a activity, hepatocytes were preincubated for 20-30 min to obtain a stable basal activity. Reactions were stopped by rapidly mixing a 0.1-ml sample of the incubation medium into 0.3 ml of ice-cold stopping medium containing 50 mM NaF, 25 mM pglycerophosphate, 10 mM MES, 2 mM EDTA, and 0.2 mM digitonin or by transferring 0.1 ml of the cell suspension into a test tube containing liquid Nz, followed by 0.3 ml of the stopping medium. Samples for phosphorylase kinase activity were stopped by the same procedure, except that EDTA was omitted from the stopping medium.
For isotopic labeling, the cells were preincubated for 90 min in medium containing either 10 pCi/ml 32Pi and 5 mg of cell protein/ml or 50 pCi/ml my0- [2-~H]inositol and 10 mg of cell protein/ml. For the measurement of 32P-labeled phospholipids, 0.15 ml of cell suspension was quenched with 0.56 ml of chloroform/methanoi/HCl (200:100:5).
[3H]Inositol phosphates were extracted from cells prelabeled with [3H]inositol by quenching 0.6 ml of cell suspension into 0.2 ml of 14% perchloric acid. Quina-loaded cells were prepared essentially as described by Tsien et al. (25,26) and by Charest et al. (9) by incubating the hepatocytes (6-8 mg of protein/ml) for 20-30 min in a medium containing 120 mM NaC1,2O mM Hepes, 5 mM glucose, 1.2 mM MgC12, 1.0 mM CaC12, 0.1 mM EDTA, 1.2 mM potassium phosphate, 4.8 mM KCl, 0.15% Ficoll with 80 p M quin2-tetraacetoxymethylester (quin2-AM). Quin2 fluorescence was measured with a MPF-44B Perkin-Elmer spectrofluorometer (ratio mode) at an excitation wavelength of 339 nm (slit width, 10 nm) and an emission wavelength of 492 nm (slit width, 20 nm). The temperature was maintained at 37 "C, and the cells were kept in suspension in the cuvette by gentle stirring. Maximal fluorescence levels were obtained by permeabilizing the cells with a minimal amount of digitonin (2-3 nmol/mg of protein). In many incubations a gradual oxidation of cellular NAD(P)H was initiated by this digitonin treatment, causing a drift in the fluorescence which interfered with the estimation of the maximal and minimal quin2 fluorescence levels. Therefore, before estimating the total quin2-related fluorescence by the addition of EGTA/Tris (final pH 8.5) (26) an excess of digitonin (50 nmol/mg of protein) was added, causing a rapid and complete oxidation of NAD(P)H. Corrections for extracellular quin2 (27) were made by adding EGTA/Tris to the cuvette before digitonin, and the total quin2 content was obtained after back titration with CaC12. Concentrations of intracellular quina were in the range of 0.5-0.8 mM. The concentration of cytosolic free calcium was calculated assuming a Kd for the quin2-calcium complex of 115 nM as described by Tsien et al. (26), after correction for extracellular quin2.
For measurement of Ca" efflux using arsenazo 111, cells (5-6 mg of proteinlml) were incubated in a Ca2+-free modified Hanks' medium (7, 8) containing 20 p~ arsenazo 111. The calcium-dependent change in absorbance was measured in an Aminco DW2a spectrophotometer at a wavelength pair of 675-685 nm. Calibration was carried out by the addition of small quantities of a standard calcium solution.

Assays
Phosphoinositides-For the analysis of the 32P-labeled lipids, 0.188 ml of 2 M KC1,lO mM EDTA and 0.188 ml of chloroform were added to the chloroform/methanol/HCl extract? The lower chloroform layer was taken and evaporated in a Rotovac evaporator. The lipids were redissolved in chloroform/methanol (21) and separated on oxalatetreated Silica Gel 60 plastic-backed thin layer chromatography plates (20 X 20) (EM Reagents) using a solvent system composed of chloroform/acetone/methanol/acetic acid/H20 (160:60:52:48:32) (28). 32P-Labeled lipids were located by autoradiography using X-Omat AR x-ray film (Kodak) and identified by using lipid standards stained with 1, vapor. The radioactivity in each spot was quantitated by cutting out the indicated areas and liquid scintillation counting in Hz0 (Cerenkov counting).
Inositol Phosphates-Perchloric acid extracts of 3H-inositol-labeled hepatocytes were centrifuged to remove precipitated protein and lipid and neutralized on ice with KOH to pH 7.0. The extract was stored frozen for up to 3 days prior to separation of inositol phosphates by HPLC. The HPLC procedure was based on that described by Irvine et al. (29) for the separation of Ins(1,4,5)P3 from Ins (1,3,4)P3 with modifications similar to those described by Batty et al. (30) to elute InsP4. Briefly, the sample was loaded onto a 250 X 4.6-mm Partisil SAX-10 column in 0.8 ml and washed through with H20 (0-7 min of elution). A constant flow rate of 1.6 ml/min was employed throughout the separation. Subsequently, a linear gradient from HZO to 0.8 M HCOONH, (pH adjusted to 3.7 with H,PO,) was run at a rate of 35 mM HCOONH,/min (7-30 min of elution), at which point the gradient was reduced to 5 mM HCOONHJmin. This slow gradient was continued over the elution period (30-38 min) during which time ATP, Ins(1,3,4)P3, and Ins(1,4,5)P3 were eluted from the column. Finally, the gradient was increased to a rate of 150 mM HCOONHJmin and run to a maximum value of 2.2 M HCOONH, where it was held constant for 4 min (38-51 min of elution). Fractions were collected at 5-min intervals over the period 7-30 min, at 0.4min intervals from 30-38 min, and at 1-min intervals from 38-51 min. The 3H content of these fractions was determined by liquid scintillation counting in Budget-Solve (Research Products Inc.) with 20% methanol added to ensure complete solubilization of the samples (29). The recovery of [2-3H]Ins (1,4,5)P3 added to unlabeled perchloric acid extracts of hepatocytes was approximately 90% using this procedure. Using standards, the following retention times were observed AMP, 14 min; ADP, 21 min; ATP, 31 min; InsP, 16 min; Ins(1,4,5)P3, 35 min. In addition, based on identification criteria discussed under "Results," other elution times were: InsP2, 23 min; Ins(1,3,4)P3, 32 min; InsP4, 43 min.
Phosphorylase a Activity-This was assayed by the method of Gilboe et al. (31), with the modification that the incubations were carried out at 37 "C for 30 min. Total (a + b) phosphorylase activity, measured as described by Uhing et al. (321,was

Calcium Mobilization and
Phospholipase C Activation by Ethanol resin (formate form), 200-400 mesh (Bio-Rad). Ionophore A23187 was obtained from Behring Diagnostics. Arginine vasopressin was obtained from Vega Biochemicals. Lipid standards, quina, quin2-AM, and collagenase were purchased from Sigma. Other chemicals and biochemicals of the highest purity commercially available were purchased from Sigma or from Fisher.

RESULTS
Characteristics of Ethanol-induced Phosphorylase Activation-The addition of ethanol (100 mM) to a suspension of isolated hepatocytes caused a rapid conversion of glycogen phosphorylase from its inactive (b) form to the active (a) form, indicating an activation of phosphorylase kinase. The phosphorylase a activity peaked approximately 30 s after the addition of ethanol and declined gradually over a period of 5-10 min (Fig. 1). This kinetic pattern of phosphorylase activation by ethanol was comparable to that brought about by calcium-mobilizing hormones such as vasopressin or &*-adrenergic agonists (7-9, 34), but the subsequent decline in activity was somewhat faster after ethanol addition.  dfferences from control incubations were obtained with ethanol concentrations of 25 mM and higher, and maximal activation of phosphorylase (80-90% of total phosphorylase activity) was found above 200 mM ethanol. Ethanol in this concentration range had no effect on phosphorylase activity when added after disruption of the cells; thus, the activation of phosphorylase was not due to a direct activation of the enzyme by ethanol. Phosphorylase kinase activity, measured in broken cell preparations with commercial muscle phosphorylase b as substrate, was consistently activated by ethanol at concentrations of 500 mM or more, but ethanol had no effect at concentrations up to 200 mM (data not shown). These observations are in agreement with the data reported by Singh and Wang (20) on purified muscle phosphorylase kinase. These authors found substantial activation of phosphorylase kinase by ethanol at concentrations in excess of 1 M but little effect below 300 mM ethanol. In other control experiments, ethanol in this concentration range had no significant effect on the cellular level of CAMP (0.37 k 0.05 and 0.43 k 0.07 pmol/mg of protein before and 30 s after 200 mM ethanol, respectively, in three separate experiments (see also Ref. 35)). The activation of phosphorylase was not affected by 4-methylpyrazole, a potent inhibitor of alcohol dehydrogenase (data not shown). After the phosphorylase response to the addition of 100 mM ethanol had decayed to basal levels, a similar transient activation of phosphorylase could be induced by a second addition of 100 mM ethanol (Fig. 3). This repeated response occurred even in the presence of 4-methylpyrazole and despite the fact that the ethanol concentration had not substantially decreased during the 15-min incubation time.3 Ethanol oxidation, which occurs in isolated hepatocytes at a rate of approximately 10 nmol/min/mg of protein at 37 "C (36) and which is at least 80% inhibited by 4-methylpyrazole (37), can account for the disappearance of less than 0.2% of the ethanol added in the experiment of Fig. 3.
Thus, neither the onset of phosphorylase activation nor its decay is dependent on the oxidation of ethanol by alcohol dehydrogenase. When 200 mM ethanol was added, either at zero time or after incubating the cells under control conditions for 15 min, the response of phosphorylase was slightly higher and its decay slower than with the 100 mM ethanol additions, even though the final concentration of the agent was the same.
The activation of phosphorylase was not specific for ethanol; other organic solvents had similar effects (Table I).
In a series of short-chain n-alcohols, the degree of phosphorylase activation increased with chain length. This finding suggests that the effect is a function of the hydrophobic properties of the solvent. The general anesthetic halothane, which shares many of the physical effects of ethanol on membranes due to its hydrophobic nature (I), also increased phosphorylase activity in intact hepatocytes. Solvents with higher dielectric constants than the straight-chain alcohols (dimethyl sulfoxide, dimethylformamide) tended to have a qualitatively similar, but less pronounced effect in activating phosphorylase. In the concentrations used here (<2%, v/v), all of the solvents used would have only minor effects on the dielectric constant of the medium.
Role of Calcium in Phosphorylase Kinase Activation-The contribution of changes in cytosolic calcium to the ethanolinduced phosphorylase kinase activation was tested in hepatocytes loaded with the calcium indicator quin2 (25, 26). At the fluorescence wavelength pair used for measuring the calcium-quin2 complex (339 and 492 nm), there is a marked contribution of NAD(P)H fluorescence to the signal. Since ethanol addition causes extensive reduction of NAD(P) in hepatocytes, the contribution of quin2-related fluorescence has to be separated from NAD(P)H fluorescence. In Fig. 4, fluorescence changes in quin2-loaded cells and e the hepatocyte suspension was pretreated with a low concentration (7 mM) of ethanol. This concentration is well above the K,,, of alcohol dehydrogenase for ethanol (37) and gives a maximal reduction of NAD(P) but has little effect on phosphorylase activity (see Fig. 2). A subsequent addition of 300 mM ethanol to unloaded cells (Fig. 4e) induced a partial reoxidation of NAD(P)H, presumably because of substrate inhibition of alcohol dehydrogenase (38). In quin2-loaded cells (Fig. 4a) the same concentration of ethanol caused a rapid transient increase in fluorescence superimposed on the decrease in NAD(P)H fluorescence. After permeabilization with a low level of digitonin, quin2 fluorescence was maximized, while a downward drift was found in the fluorescence of unloaded cells; addition of a high level of digitonin caused a rapid oxidation of NAD(P)H of similar magnitude in both incubations. Ethanol-induced changes in NAD(P)H fluorescence can largely be prevented by pretreatment of the cells with 4-methylpyrazole (15 mM) (Fig, 4, b andf); high concentrations of ethanol now caused a minor increase in NAD(P)H level (Fig. 4f). In quin2-loaded cells, addition of I-methylpyrazole gave a small transient increase in fluorescence which decayed to base line after 2-3 min. Ethanol (300 mM) again caused a rapid transient increase in cytosolic calcium concentration, which was now clearly distinct from the ethanolinduced effects on the NAD(P) redox state.
Corrections for the contribution of NAD(P)H fluorescence changes can also be made by taking parallel measurements at an excitation wavelength of 357 nm. This wavelength is an isosbestic point for the fluorescence of free quin2 and the quin2-calcium complex (26). Alterations in the free Ca2+ concentration, therefore, have no effect on the quin2 fluorescence at that wavelength. At the same time, the fluorescence of NAD(P)H is approximately similar when recorded at an excitation wavelength of 339 and 357 nm. These points are illustrated in Fig. 4, c and g, where the solid line represents 339-nm fluorescence and the dotted line 357-nm fluorescence. The difference in fluorescence for 339-357 nm (Fig. 4, d and  h) gives an estimate of the changes in fluorescence specifically associated with the quin2-calcium complex. It is apparent from Fig. 4, d and h, that this method also corrects for the NAD(P)H changes induced by digitonin addition. The effect of ethanol (300 mM) on quin2-calcium fluorescence can now be recorded independent of NAD(P)H fluorescence changes (Fig. 4d). The comparison of Fig. 4, b and d, illustrates that the increase in cytosolic free Ca2+ after ethanol addition was unaffected by pretreatment of the cells with 4-methylpyrazole, in agreement with the data on phosphorylase activation.
The kinetics of Ca2+ mobilization in quin2-loaded cells were parallel to the activation of phosphorylase, reaching a maximum after about 30 s and decaying to a new steady state after 2-4 min. The peak free calcium level was dose dependent and reached up to 0.65 p M after the addition of 470 mM ethanol to quin2-loaded cells (data not shown). At higher levels of ethanol the response was variable and often blunted, possibly due to cell damage induced by ethanol addition. The steady state free Ca2+ in the hepatocytes after decay of the Ca2+ burst was not consistently affected by ethanol treatment, although in some experiments a slight increase was observed (less than 0.1 ~L M above base line). The concentration of quin2 accumulated in the hepatocytes under our standard loading conditions (varying from 1.0 to 1.5 nmol/mg) can introduce a significant Ca2+ buffering capacity in the cytosol. In Table 11, the effect of varying quin2 loads on the Ca2+ response was assessed, after stimulation The peak of the calcium response occurred after about 20-30 s. The steady state readings were taken 3 min after the addition of the stimulus. Total quin2 content was determined after permeabilization with digitonin. Calibration of the quin2 signal was done by adding known amounts of quin2 to a cuvette containing with ethanol (300 mM) or with the a-adrenergic agonist phenylephrine (2 p~) . The concentration of phenylephrine was selected to give an extent of calcium mobilization comparable to that found with ethanol. The data in Table I1 demonstrate that the peak free cytosolic Ca2+ concentration is depressed markedly with increasing concentration of quin2. The effect of Ca2+ buffering introduced by the quin2 in the cytosol is independent of whether ethanol or phenylephrine was used to generate the Ca2+ increase. The steady state free cytosolic Ca2+ level measured 3 min after the addition of the stimulus was relatively unaffected by the quin2 loading.
Intracellular Origin of Calcium Mobilized by Ethanol-The ethanol-induced change in cytosolic Ca2+ level did not require the presence of calcium in the incubation medium. As shown in Fig. 5A, the addition of excess EGTA immediately before ethanol had no effect on the time course of phosphorylase activation and its decay. Similarly, the fluorescence response in quin2-loaded cells was not affected by the addition of excess EGTA (not shown). Preincubation of the cells for 30 min in a calcium-free medium with 1 mM EGTA decreased the ethanol-induced increase in phosphorylase a activity (Fig.  5B); presumably this is due to the depletion of calcium from intracellular pools over time. Similar effects of pretreatment with 1 mM EGTA were observed on phenylephrine-induced activation of phosphorylase; however, in contrast to ethanol, the decay of the phenylephrine-induced calcium mobilization was much faster in the absence of extracellular calcium than in the standard calcium-containing medium (35).
In further experiments the hepatocytes were washed in a calcium-free medium and incubated in a buffer containing the calcium indicator, arsenazo I11 (Fig. 6). Under these conditions, different calcium-mobilizing agents (A23187, phenylephrine, vasopressin) caused a rapid efflux of calcium from the cells, in agreement with the intracellular origin of the calcium mobilized by these agents. As shown in Fig. 6, ethanol (300 mM) similarly caused a small degree of calcium efflux from the cells and decreased the A23187-mobilizable Ca2+ pool.
Relationship with Hormone-induced Calcium Mobilization-In liver cells, vasopressin and al-adrenergic agonists mobilize calcium from an intracellular storage site, presumably located in the endoplasmic reticulum (14, 39). In the experiment of hepatocytes pretreated with 4-methylpyrazole induced a Ca2+ mobilization response with characteristics comparable to those reported by others (e.g. Refs. 9 and 40). In the presence of extracellular calcium, a transient peak (up to 1 pM cytosolic Ca"+) was followed by a gradual decline to a new steady state of about 0.3 FM free Ca2+. When cells were pretreated with ethanol (250 mM), the fluorescence response to hormone addition was qualitatively similar and the peak calcium level was comparable to that found in cells that had not received ethanol (Fig. 7). By contrast, pretreatment of the cells with phenylephrine completely prevented the calcium mobilization in response to 250 mM ethanol. This observation is confirmed by the Ca2+ efflux experiment shown in Fig. 6; pretreatment of the cells with phenylephrine inhibited the ethanol-induced efflux of calcium. Thus, the ethanol-induced calcium mobilization appears to be completely dependent on the presence of calcium in the hormone-sensitive calcium pool in the cell. Similar results were obtained when calcium mobilization was measured by phosphorylase activation (data not shown). The ethanol-induced response could also be prevented by pretreating the cells with a saturating concentration of vasopressin (351.4 Phenylephrine-induced calcium mobilization in hepatocytes is mediated by Ins (1,4,5)P3 and the question arises whether ethanol acts directly on the Ins(l,4,5)P3-sensitive calcium stores. In hepatocytes in which the plasma membrane has been made permeable by treatment with digitonin, the hormone-sensitive calcium pools can be mobilized directly by the addition of Ins (1,4,5)P3 (14). In the experiment of Fig. 8 the cells were preincubated in the presence of digitonin and ATP and allowed to reach a steady state free Ca2+ concentration (approximately 200 nM). The calcium level was measured by the fluorescence of quin2 which was included in the medium to both monitor and buffer the free Ca2+ concentration.
Under these conditions, addition of Ins (1,4,5)P3 (1.0 PM) caused a rapid release of calcium from nonmitochondrial stores in the permeabilized cells, in agreement with results reported elsewhere (14). Ethanol (300 mM) caused a minor release of calcium into the medium but did not significantly affect either the rate or the extent of Ins(l,4,5)P3-induced calcium release. The calcium release induced by ethanol stabilized at a slightly higher steady-state calcium level and probably represents some inhibition of the ATP-driven calcium pump in the endoplasmic reticulum (2). Both the change in this calcium set-point and the rate of calcium release after ethanol were insufficient to account for the calcium burst induced by ethanol in intact hepatocytes.
Stimulation  FIG. 8. Effect of ethanol on calcium pools in digitonin-permeabilized hepatocytes. Isolated hepatocytes were washed free of calcium and suspended at about 3 mg of protein/ml in medium containing 120 mM KCl, 10 mM NaCI, 20 mM Hepes, 1 mM KH~POI, 0.2 mM MgC12, 75 pM quina, 1 pg/ml rotenone, 10 mM creatine phosphate, 1 unit/ml creatine kinase, 0.1% bovine serum albumin at pH 7.2. The cells were permeabilized using 20 pg/ml digitonin, and then CaC12 was added to give a free Caz+ concentration of about 200 nM (buffered by the quina). Ca2+ movements were followed fluorometrically using the fluorescence of the quin2-calcium complex as described previously (14,42). Uptake of Ca2+ into intracellular vesicular pools was initiated with 2 mM MgATP (not shown). After a steady state of Ca" uptake had been achieved, additions were made as indicated: HzO, 12 pl/m& ETOH, 12 pl/ml ethanol to give 200 m M final concentration; ZP3, 1 p M Ins (1,4,5)P3 which was a maximally effective dose for Ca2+ release. Ca2+ release to the medium is indicated by an increase in the fluorescence signal.
In initial experiments it was found that saturating doses of vasopressin (100 nM) caused an increase of nearly 20-fold in Ins (1,4,5)P3 levels within 15 s (mean radioactivity in Ins (1,4,5)P3 in controls was 126 f 8 dpm increasing to 2498 & 186 dpm after 15 s with 100 nM vasopressin; n = 3). In the same series of experiments a near-saturating dose of ethanol (300 mM) caused only a 70% increase in Ins (1,4,5)P3 (209 2 65 dpm). Thomas et al. (42) have shown, however, that such supraphysiological doses of vasopressin elevate the InsP3 level far in excess of that required for a maximal Ca2+ response. For this reason, the effects of ethanol on Ins(1,4,5)P3 levels were compared to a vasopressin concentration which induces a similar rate of increase in cytosolic free Ca".
In the experiment of Fig. 9 the cytosolic calcium changes were compared after 300 mM ethanol and after 1 or 40 nM vasopressin. Ethanol and 1 nM vasopressin induced a similar time course and extent of change in cytosolic free Ca", from a basal level of about 120 nM to a peak of 300 nM by 20-30 s (Fig. 9, A and B ) . The  nM (C) was added to a suspension of quin2-loaded hepatocytes prepared and incubated as described under "Experimental Procedures." Before the addition of ethanol, the cells received 4-methylpyrazole (15 mM) to inhibit changes in NAD(P)H level due to alcohol dehydrogenase.
vasopressin were greater in magnitude (note the nonlinear Ca2+ scale) and very much faster than those observed with 1 nM vasopressin (Fig. 9C). This is in keeping with the much greater potency of higher vasopressin concentrations to increase Ins (1,4,5)P3 levels (42). Fig. 10 shows time courses for the changes of inositol phosphates in hepatocytes treated with either 300 mM ethanol or 1 nM vasopressin. At these concentrations, ethanol and vasopressin act to increase the cellular levels of Ins (1,4,5)P3 at a similar rate and to a similar extent (Fig. 1OA). As a percentage of the control value before agonist addition, 300 mM ethanol increased Ins(1,4,5)P3 by 79.8 f 8.7% (n = 5) after 20 s, and with 1 nM vasopressin a maximal increase of 94.2 f 19.3% (n = 4) was achieved within 1 min. In contrast to the similar rate of formation of Ins (1,4,5)P3 with 300 mM ethanol or 1 nM vasopressin, the subsequent decline of this compound was very different for the two agonists. With vasopressin Ins( 1,4,5)P3 levels remained elevated throughout the period of the experiment, but with ethanol the level of Ins (1,4,5)P3 had decayed back to the basal value by 2 min of ethanol treatment. This decay is also observed in the cytosolic free Ca2+ changes induced by ethanol (Figs. 4,7, and 9). With vasopressin the elevation of cytosolic free Ca2+ declined more slowly, in keeping with the sustained elevation of Ins (1,4,5)P3.
The similarity between the kinetics of the formation as well as the decay of ethanol-induced Ins (1,4,5)P3 accumulation and the increase and decline of the cytosolic free Ca" level brought about by ethanol provides strong evidence that ethanol, like vasopressin, elevates cytosolic free Ca2+ by increasing the production of Ins(1,4,5)P3. The transient nature of the Ins(1,4,5)P3 increase with ethanol raises the question of its further metabolism. Ins (1,4,5)P3 can be successively dephosphorylated through InsP,, InsP, and finally to inositol (11, 13). It has recently been shown, however, that Ins(1,4,5)P3 may be further phosphorylated to Ins(1,3,4,5)P4, and it is possible that the other InsP, isomer, Ins(1,3,4)P3, is derived from this InsPl (45). The HPLC elution profile contains a small peak which we tentatively identify as InsP4 based on the similarity in ionic strength at which it is eluted compared to the results reported by others (30, 45); this peak contains only about 20% of the ,H-cpm which we find in InsP,. Because this low level of radioactivity makes quantitation of the InsPI rather difficult, we have not made a detailed study of its changes in response to the concentrations of agonists used in the present study.
In addition to Ins(1,4,5)P3 (identified using a [,H]Ins-( 1,4,5)P3 standard), samples extracted from cells prelabeled with [3H]inositol contained a second 3H-labeled peak on HPLC of similar magnitude to Ins (1,4,5)P3, but with a slightly shorter elution time. Based on its proximity to the ATP peak, this peak presumably represents Ins(1,3,4)P3 as identified by Irvine et al. (29). Confirmation of this interpretation comes from the kinetics of changes in this peak relative to Ins (1,4,5)P3. It increases more slowly than Ins (1,4,5)P3 after stimulation with vasopressin (see below), in agreement with results obtained by several authors in other studies of agonistinduced InsP3 formation (29,46). Moreover, the accumulation of radioactivity in this fraction in the presence of 100 nM vasopressin continues over a prolonged time period (10-15 min) and is specifically potentiated more than 10-fold in the presence of 10 mM Licl (47) in a manner similar to that reported by Burgess et al. (46). Fig. 10B shows the changes of Ins (1,3,4)P3 brought about by treatment of hepatocytes with 300 mM ethanol or 1 nM vasopressin. For vasopressin there is a distinct lag of 5-10 s, and then Ins (1,3,4)P3 increases over a period of about 1 min to a plateau value of 250% above control. Thus, Ins(1,3,4)P3 increases somewhat more slowly but to a greater extent than Ins (1,4,5)P3 with 1 nM vasopressin. Similar results were obtained for 100 nM vasopressin, although the accumulation of both isomers of InsP, was very much greater (data not shown). With ethanol, the lag in Ins (1,3,4)P3 production was much less pronounced than with 1 nM vasopressin (Fig. lOB), and after the first 20 s of stimulation this isomer declined toward basal in a manner similar to that observed for the ethanolinduced Ins (1,4,5)P3 elevation (Fig. 1OA). InsPZ (identified by its proximity to the ADP peak on HPLC (29)) also showed a small transient increase with ethanol (Fig. lOC), although its decay was slower than for the InsP, isomers. With vasopressin the InsP, elevation was more pronounced and did not decay during the time course of the experiment. InsP, which was present at very much higher basal levels, did not change significantly during the first 2 min of stimulation even with 100 nM vasopressin, as noted previously (42).
The treatment of hepatocytes with ethanol is also accompanied by changes in the inositol phospholipid precursors of the inositol phosphates. Table I summarizes the results of our experiments with hepatocytes preincubated for 90 min with 32Pi to label the monoester phosphate groups on the inositol ring of PtdIns(4)P and PtdIns(4,5)P,. Ethanol (200 mM) caused a small but significant decline in PtdIns(4,5)P2 within 60 s. A t the same time, however, PtdIns(4)P increased markedly, presumably due to the simultaneous ethanol-dependent activation of PtdIns kinases that generate the polyphosphoinositides from PtdIns. There is also a significant increase in phosphatidic acid, presumably generated from diacylglycerol by diacylglycerol kinase. No significant changes in the levels of PtdIns and phosphatidylcholine occurred during the short time periods of this experiment.

DISCUSSION
Several lines of evidence indicate that the ethanol-induced increase in phosphorylase a activity in hepatocytes is due to the mobilization of intracellular calcium, and not to other known effectors of phosphorylase kinase. In the first place, ethanol does not induce changes in CAMP level and has no effect on the dose-response curve of glucagon for phosphorylase activation? These data indicate that CAMP-dependent protein kinase is not activated by ethanol under these conditions. Second, a direct activation of phosphorylase kinase by ethanol, as reported by Singh and Wang (20) for the purified skeletal muscle enzyme, required higher concentrations than were used in the present study in intact hepatocytes. No significant effect of ethanol is observed in broken hepatocytes at concentrations up to 200 mM. In intact hepatocytes halfmaximal activation of phosphorylase is obtained at 60-70 mM ethanol, and a detectable degree of activation requires as little as 25 mM ethanol. Third, there is direct evidence for a transient ethanol-induced mobilization of calcium in quin2loaded hepatocytes, with a time course compatible with its involvement in the activation of phosphorylase kinase.
Ethanol concentrations that cause maximal activation of phosphorylase in unloaded hepatocytes do not give maximal calcium mobilization in quin2-loaded cells. Different factors may contribute to this phenomenon. At 200 mM ethanol, more than 80% of the phosphorylase is converted to the active a form, an effect similar to that seen with maximally activating levels of phenylephrine or vasopressin. A further increase in cytosolic Ca2+ does not lead to further activation of phosphorylase. Furthermore, loading hepatocytes with quin2 introduces a significant Ca2+ buffering capacity in the cytosol, and the ethanol-induced changes in cytosolic Ca2+ are partly suppressed in the quin2-loaded cells. Thus, in unloaded cells, ethanol can probably induce a significantly higher increase in the cytosolic free Ca2+ level than the estimates reported here indicate. A direct comparison of the ethanol-induced activation of phosphorylase in quin2-loaded and unloaded cells is difficult, because the loading procedure itself results in a significant increase in phosphorylase a activity (9).
The effects of ethanol on calcium mobilization are probably a reflection of its interaction with the cellular membranes. The concentration range in which ethanol activates phosphorylase has been demonstrated to affect structural and functional characteristics of different biological membranes (1). Moreover, the effectiveness of the different organic solvents increases with greater hydrophobicity (Table I). However, the same solvents also affect the dielectric constant of the water to a small degree, and it cannot be excluded that part of the effects of these solvents are due to an alteration of the water structure, for instance at the membrane-water interface.
The source of the calcium mobilized by ethanol is clearly intracellular, and the data strongly suggest that the calcium comes from the same pool as that mobilized by vasopressin and a-adrenergic hormones in liver. This finding suggests the possibility that ethanol interacts with the signal transduction pathway used by these hormones to elevate cytosolic calcium. Our data do not indicate, however, that the effects of ethanol on intact cells can be attributed to a direct action on the membranes of the endoplasmic reticulum to activate Ca2+ efflux and/or inhibit the reuptake. In digitonin-permeabilized hepatocytes, the addition of ethanol caused only a minor and sluggish adjustment of the Ca2+ set-point (see Fig. 8), presumably as a consequence of the inhibition of the endoplasmic reticular Ca2+ pump by ethanol (2). Moreover, ethanol affected neither the rate nor the extent of Ins(l,4,5)P3-induced calcium mobilization. These data indicate, therefore, that ethanol interacts with the signal transduction pathway at a point before the mobilization of calcium from the endoplasmic reticular stores.
The measurement of inositol phosphates and phosphoinositides ( Fig. 10 and Table 111) provided direct evidence that ethanol mobilizes intracellular calcium by activating the hormone-sensitive phosphoinositide-specific phospholipase C. Several arguments support the notion that the rapid accumulation of Ins (1,4,5)P3 after addition of ethanol is the direct cause of calcium mobilization. First, the change in Ins (1,4,5)P3 preceded the rise in calcium concentration. Second, the amount of Ins (1,4,5)P3 generated by ethanol correlated well with its potency as a calcium-releasing agent; this was demonstrated by the similar degree of Ins (1,4,5)P3 production induced by a low dose of vasopressin which was equivalent to 300 mM ethanol in its effects on calcium mobilization. The Further support for a phospholipase C-mediated mechanism of ethanol action comes from the analysis of changes in substrate concentration, i.e. inositol lipids. A small but significant decrease in PtdIns(4,5)Pz occurred in response to ethanol addition, although the magnitude of the response varied and was often difficult to quantitate. This is not surprising in view of the low levels of Ins(1,4,5)P3 formed. Moreover, the steady state levels of the phospholipid intermediates in this pathway are affected by the activity of other enzymes, e.g. phosphomonoesterases, kinases, and phosphodiesterases. Receptor-mediated stimulation of phospholipase C is generally accompanied by an activation of the phosphoinositide kinases which regenerates the PtdIns(4,5)P2 from its precursor lipids (11-13). The elevation of the level of PtdIns(4)P in response to ethanol addition is an indication that the phosphoinositide kinases are also stimulated after ethanol addition.
Further evidence supporting an ethanol-induced activation of phospholipase C activity is the formation of phosphatidic acid, which can readily be generated from diacylglycerol in hepatocytes by diacylglycerol kinase. Thus, the data presented here provide strong evidence that ethanol initiates the activation of the same pathway employed by vasopressin and other calcium-mediated hormones in hepatocytes and that it generates the same second messenger signals.
A major difference between the ethanol-induced and the hormone-receptor-mediated responses is the transient nature of the former signal. The ethanol-induced calcium increase decayed to basal levels over a 2-4-min period, despite the continued presence of ethanol and even when its oxidation through alcohol dehydrogenase was inhibited. The decay of the calcium signal is preceded by a decline of both isomers of InsP, to basal levels (Fig. 10) and hence represents a deactivation of phospholipase C. The calcium mobilized by Ins (1,4,5)P3 is reaccumulated in the same stores, where it is available for a renewed challenge, either by a subsequent addition of ethanol or by a hormonal stimulus (see Figs. 3  and 7). This situation is markedly different from that induced by hormones, where a partially elevated level of Ins (1,4,5)P3 persists and the calcium stores are not replenished until the hormone is removed or an antagonist is added.
Further studies will be required to identify possible mechanisms of the activation of phosphoinositide-specific phospholipase C by ethanol. The hormonal activation of this enzyme involves the transmembrane interaction with a hormone receptor, possibly mediated by a GTP-binding coupling protein similar to, but not identical with, the G, and Gi proteins (stimulatory and inhibitory guanyl nucleotide-binding regulatory proteins of adenylate cyclase) involved in receptor-adenylate cyclase coupling (48). It is conceivable, therefore, that the disordering effects of ethanol on the membrane (1) affect either the interaction of the phospholipase C with the substrate PtdIns(4,5)Pz or the intramembrane receptor-phospholipase C coupling mechanisms. There is some evidence in different cells that ethanol can affect the receptormediated activation of adenylate cyclase by interacting with the coupling mechanism involving GTP-binding proteins (49-

51).
The transient nature of the ethanol-induced calcium mobilization, despite the continued presence of ethanol in the system, suggests a mechanism more complex than a simple activation of phospholipase C by ethanol-induced disordering of the membrane structure. There could be a feedback mechanism leading to a desensitization of phospholipase C in the presence of ethanol. It has been proposed (52, 53) that activation of protein kinase C could cause a feedback inhibition of the receptor-phospholipase C interaction. A further addition of more ethanol, however, generates a renewed calcium burst, and there is no indication of a decreased effectivity of the response to ethanol. The factors involved in phospholipase C regulation have not yet been sufficiently characterized, and further studies are required to identify possible sites of interaction of ethanol and other membrane disordering agents with this process.
The physiological consequences of this effect of ethanol also remain to be identified. In other experiments, we have observed that short-term pretreatment of hepatocytes with ethanol may affect their sensitivity to submaximal concentrations of va~opressin.~ It is conceivable that the response of cells to a variety of stimuli is altered by a repeated transient activation of the polyphosphoinositide-dependent signal transduction pathway. Such effects could be mediated by a redistribution of calcium in intracellular pools or by the phosphorylation of critical proteins in the cells that persist well beyond the transient elevation of calcium. Currently, studies are underway in our laboratory to further explore these possibilities.