Characterization of phospholipase A2 and acyltransferase activities in purified zymogen granule membranes

Biochimica Elsevier

et Biophysics

245

Acra, 1045 (1990) 245-251

BBALIP 53453

Characterization of phospholipase A, and acyltransferase activities in purified zymogen granule membranes Ronald

P. Rubin

*, Raiford

Division of Cellular Pharmacology,

H. Thompson

and Suzanne

Medical College of Virgrnia, Richmond,

Key words: Phospholipase A,; Acyl-CoA:acyltransferase;

G. Laychock

*

VA (U.S.A.)

Arachidonic acid turnover; Pancreas; (Rat)

Phospholipase A, and acyltransferase activities were identified in membranes associated with purified pancreatic zymogen granules. In homogenate and granule membranes, phospholipase activity was linearly related to protein concentration and was Ca2+-dependent with an alkaline pH optimum. The Ca” sensitivity was observed over the range of concentrations through which intracellular ionic Ca” is elevated by physiological stimuli in intact cells. Intact zymogen granules and granule membranes also demonstrated reacylating activity in the presence and absence of an exogenous acceptor. Reacylating activity was related to the concentration of lyosphospholipid added and was optimally activated at alkaline pH. A more rapid rate of reacylation was observed when [ 14C]arachidonoyl CoA was employed as the donor molecule rather than [3H]arachidonate (plus coenzyme A); this suggests the absence of acyl-CoA synthetase in the purified granule membranes. We conclude that granule membrane phospholipase A, and acyltransferases may be involved in arachidonic acid turnover in exocrine pancreas and perhaps in membrane fusion events associated with exocytosis.

Introduction Turnover of the arachidonoyl moiety of membrane phospholipids appears to play a key role in ligand-receptor interactions in many cell types, including pancreatic acinar cells [l]. Arachidonic acid and other fatty acids are incorporated into phospholipids either by the de novo pathway (Kennedy pathway) or by deacylation-reacylation through phospholipase and acyltransferase activities (Lands pathway) [2]. The deacylationreacylation cycle serves to regulate the levels of arachidonic acid and lysophospholipids which may participate in various cellular functions, including secretion. Acyltransferase reactions also represent an important mechanism for regulating the fatty acid profile of pancreatic phospholipids in that phospholipids gen-

* Present address: Department of Pharmacology & Therapeutics; School of Medicine & Biomedical Sciences, State University of New York at Buffalo, Buffalo, New York 14214, U.S.A. Abbreviations: PMSF, phenylmethylsulfonyl fluoride; Mes, (2[ Nmorpholino]ethanesulfonic acid; Mops, morpholinopropanesulfonic acid; EGTA, ethyleneglycol bis(/3amincethyl ether)N, N, N’, N ‘-tetraacetic acid; PtdIns, phosphatidylinositol; PtdCho, phosphatidylcholine; PLA,, phospholipase A,. Correspondence: R.P. Rubin,Department of Pharmacology & Therapeutics, School of Medicine State University of New York at Buffalo, 102 Farber Hall Buffalo, New York 14214, U.S.A.

erally exhibit a non-random distribution of the acyl constitutents, with saturated fatty acids esterified predominantly in the C-l position and unsaturated fatty acids at the C-2 position [3]. Cellular membranes are involved in a variety of cellular functions such as signal transduction and exocytosis. The zymogen granule is a key component of the secretory process since it is involved in the packaging, storage and release of granule contents by fusing with the apical membrane. Enzymes such as protein kinase and ATP diphosphohydrolase activities, which have been identified in purified pancreatic zymogen granule preparations [4,5], can therefore be expected to be involved in various phases of the secretory response. In a previous study, we demonstrated the conversion of 2-lysophospholipids into corresponding phospholipids via acyl-CoA acyltransferase in homogenates of rat pancreatic acini [6]. Arachidonic acid was greatly preferred over stearic acid as the acyl donor and lysophosphatidylinositol (1ysoPtdIns) and lysophosphatidylcholine (1ysoPtdCho) were the preferred acceptors. Acylating activity observed with lysophosphatidylserine and lysophosphatidylethanolamine was only lo-20% of that observed with IysoPtdIns or 1ysoPtdCho [6]. Moreover, in intact pancreatic acini the glycerophospholipids undergo deacylation-reacylation during ‘activation by secretagogues [7]. Although phospholipase A, (PLA,) has been purified from rat pancreas and its complete amino acid

246 sequence determined from complementary DNA [8], multiple forms of the enzyme may exist in this tissue. The present study presents evidence for the existence of the deacylation-reacylation system in secretory granules by demonstrating PLA, activity and the reacylation of 1-acyl-sn-glycero-3-phospholipids by arachidonic acid in purified intact granules and granule membranes. We also describe the properties of these enzyme systems. Materials and Methods Materials Reagent concentrate for protein determinations was obtained from Bio-Rad Laboratories (Richmond, CA, U.S.A.). [5,6,8,9,11,12,14,15(n)3H]arachidonic acid (190 Ci/mmol) and [ arachidonyl-l-‘4C]arachidonoylcoenzyme A (53 mCi/mmol) were purchased from New England Nuclear (Wilmington, DE, U.S.A.). 1-Stearoyl-2-[l-‘4C]arachidonoylphosphatidylcholine (58 mCi/mmol) was obtained from Amersham Research Products (Arlington Heights, IL, U.S.A). The lysophospholipids (1ysoPtdCho and lysoPtdIns), which were purchased from Avanti Polar Lipids, were prepared by PLA 2 hydrolysis, purified by high-performance liquid chromatography and lyophilized. All other chemicals used were at least of reagent quality. Preparation of zymogen granule fraction Zymogen granule membranes were prepared from male Sprague-Dawley rat pancreatic homogenates by the method of DeLisle et al. [9], with the following modifications. Two rats were sacrificed by decapitation per experiment. The crude granule pellet was mixed with 40% Percoll, 250 mM sucrose, 50 mM Mes (pH 5.5), 0.1 mM MgSO,, 0.1 mM PMSF and 2 mM EGTA. The Percoll gradient was established by spinning at 50000 X g in a Beckman L8-80 centrifuge in a type 50.2 Ti rotor (fixed angle). Zymogen granules were washed twice to remove the Percoll by pelleting at 8000 x g for 10 min and resuspending the pellet in buffered sucrose containing (mM): sucrose 300, MgSO, 0.1, EGTA 1, Mops 2 (pH 6.5). This solution, as well as all others, contained 10 mg% soybean trypsin inhibitor and 1 mM benzamidine to inhibit tissue proteinases. There was a greater than 3-fold enrichment of amylase in the washed granule fraction relative to the crude homogenate (75 k 5 versus 249 + 18 mg maltose/mg protein) (n = 30). These results are comparable to the 4-5-fold enrichment obtained by others [9]. The granule fraction was devoid of mitochondria as assessed by electron-microscopy and very scant vestiges of membrane fragments could be observed. Mitochondria are the most likely contaminant in the zymogen granule membrane fraction; however, we were unable to detect cytochrome c reductase activity (a mitochondrial marker enzyme) in this fraction, as compared

to the crude homogenate fraction which contained 2 + 0.3 units/mg protein (n = 11). The purity of the granule fraction was also verified by the lack of nuclei as determined by electron-microscopy and by the markedly reduced DNA content (0.25 + 0.1 mg, as compared with 19.5 + 1.8 mg in the crude homogenate) (n = 3-8). Preparation of granule membrane fraction Membranes were prepared from the washed granule fraction by lysing the granule pellet in 1 mM EGTA and 30 mM Tris buffer (pH 8.3) and incubating for 10 min at 37°C with shaking. The solution was then spun for 60 min at 100000 X g and the pellet resuspended in ice-cold homogenization buffer. Enzyme and chemical assays PLA, activity was analyzed as previously described, using l-stearoyl-2-[1-i4C]arachidonoyl phosphatidylcholine as the substrate (0.2 pCi/ml) [lo]. The assay was carried out using 70-140 pg of protein in a total volume of 240 ~1. The final incubation medium (pH 8.0) contained (mM): sucrose 105; Tris-HCl 210; Mops 2; and CaCl, 10 mM unless otherwise noted. PLA, activity is expressed as the percent hydrolysis of radiolabeled arachidonic acid from phospholipid, which was calculated from the radioactivity (cpm) in free arachidonic acid divided by total phospholipid radioactivity (which averaged 36458 cpm). Values were corrected for blank hydrolysis (less than 1%) which was determined in the absence of enzyme. To verify that arachidonate release was mediated by phospholipase activities were assessed using PLA,, [‘4C]dioleoylPtdCho as the substrate and a one-dimensional thin-layer system that separates radiolabelled fatty acid, lysoPtdCho, monoacylglycerol and diacylglycerol [ll]. When granule membranes were incubated with exogenous substrate, in the presence of zero or 10 mM Ca2+, the increase in radiolabeled fatty acid release elicited by Ca2+ was accompanied by a corresponding increase in [‘4C]lysoPtdCho, a product of PLA, activation. By contrast, radiolabeled mono- and diacylglycerol were not detected, thus excluding the possibility that our observations were the result of the sequential actions of phospholipase C and diacylglycerol lipase. Acyltransferase activity was determined by adding 200 ~1 of tissue suspension to 400 ~1 of cocktail containing 0.4 PCi [3H]arachidonic acid or 0.06 PCi [14C]arachidonoyl-CoA. The cocktail contained (mM): Tris-HCl 45; MgCl, 15; EGTA 1.5; ATP 9; Coenzyme A 0.15; Cleland’s Reagent 1.5. Incubations were carried out, in the presence and absence of 10 PM 1ysoPtdIns or lysoPtdCho, for various intervals and terminated by adding 3 ml chloroform/methanol (1 : 2, v/v). The lysophospholipids were dried down, resuspended in buffer and sonicated prior to addition to the reaction mixture. Our previous study showed that lo-20 PM

247 lysophospholipid gave optimal rates of arachidonate incorporation [6]. In certain experiments the Ca2+ or protein concentration and/or pH of the incubation medium was modified. The radiolabeled phospholipids were extracted and separated by thin-layer chromatography as previously described [lo]; the plates were scraped and samples counted by liquid scintillation spectrometry. Amylase and cytochrome c reductase assays, as well as total protein and DNA, were determined as previously described [6,12,13]. Results Phospholipase A, activity in granule membrane and ceilfree homogenates Initial experiments in determining the subcellular localization of PLA, activity was measured using [i4C]phosphatidylcholine (PtdCho). A pH of 8.0 was initially chosen to determine enzyme activity at our standard incubation conditions. When isolated membranes derived from a purified zymogen granule fraction were incubated with radioactive substrate, there was a time-dependent release of [i4C]arachidonate indicative of PLA, activity (Fig. 1). Enzyme activity increased linearly for up to 90 min. Increasing the membrane concentration in the range of 70-140 pg protein/ml resulted in a linear increase in phospholipase activity. Further increases in the protein concentration produced streaking of radioactivity on the TLC plates, making accurate determinations of enzyme activity difficult (unpublished data). When the whole cell-free homogenate was examined, a linear increase in enzyme activity was observed for 20 min in the same range of protein concentrations (70-140 pg) used for determination of membrane associated enzyme activity (Fig. 1). The basis for the more rapid termination of the linearity of PLA, activity in the total

25

b

Homogenate

o

Membranes

Time

( min)

Fig. 1. Time-course for the hydrolysis of [r4C]arachidonoyl PtdCho in total homogenate and purified zymogen granule membrane fraction of exocrine pancreas. Total homogenate or zymogen granule membrane fraction was incubated with radioactive substrate (pH 8.0) for indicated intervals as described in Materials and Methods.

5

0

L 0

I 2

I 1 Ca”+

(PM

/+5

)

of PLA, activity in granule membrane Fig. 2. Ca 2+-dependence fraction. Membranes were incubated with radioactive substrate for 30 min in the presence of varied amounts of Ca*+ added to a constant EGTA concentration (pH 8.0) Each point represents the mean value (+ SE.) obtained from four separate preparations.

homogenate, relative to the membrane fraction, may be attributed to the presence of an endogenous inhibitor (fatty acid) present in the homogenate [14]. Effect of Ca2 + and pH on phospholipase A, activity Phospholipases can be categorized on the basis of their Ca*+ and pH requirements [15]. To accurately determine the Ca*+ sensitivity of the membranous experiments were performed utilizing an PLA,, EGTA/CaCl, buffering system previously employed in our laboratory [16]. Significant activity was observed in the absence of added Ca2+ plus EGTA (Fig. 2). However, the enzyme was sensitive in a linear manner to increments of ionic Ca2+ in the submicromolar range. Enzyme activity began to reach a plateau level at about 5 PM. Thus, granule membrane PLA, activity is sensitive to physiological concentrations of Ca*+. Further increases in PLA, activity could be observed at millimolar concentrations of ionic Ca2+, although these increments were more variable when compared to the responsivity to micromolar concentrations of Ca2+ (data not shown).

I

I

I

6

6

10

PH

Fig. 3. Effect of varying pH on PLA, activity of purified granule membranes. PLA, activity was determined after a 30 mm incubation over the pH range of 4.5-9.5. Each point represents the mean value obtained from three different experiments.

248 Fig. 3 shows that PLA, activity in granule membranes was pH-dependent, with highest measured activity at neutral pH. At pH 9.5, the highest pH value tested, enzyme activity was comparable to that obtained at pH 7.5. In addition, the sensitivity of PLA, to Ca*+ was unaltered when the pH was increased from 7.5 to 9.0 (unpublished data). Reacylation of IysoPtdIns Our previous study demonstrated that 1ysoPtdIns and 1ysoPtdCho were preferred acceptors for reacylation of phosphoglycerides in homogenates and crude subcellular fractions of pancreatic acini [6]. In the present study, reacylation of Ptdlns was detected in both intact granules and the granule membrane fraction. Reacylating activity was dependent upon the amount of protein added (lo-50 /.tg) and required the presence of ATP and CoA (unpublished data). The reacylating activity of intact granules and homogenate was linearly related to the concentration of 1ysoPtdIns up to at least 10 PM (Fig. 4). The acylating activity detected in the granule membrane fraction could be markedly enhanced by the addition of 10 /.LM 1ysoPtdIns (Fig. 5). Reacylating activity in granule membranes, like that of homogenate, was also pH-dependent, with the greatest measured activity at pH 9.5, the highest pH value examined (Fig. 5). The relative rates of the two reactions involved in the reacylation process, arachidonic acid acetylation and esterification of arachidonoyl-CoA to lysophospholipid, was assessed by comparing the esterification of [14C]arachidonoyl-CoA with that of [3H]arachidonic acid. In homogenate, when [r4C]arachidonoyl-CoA was used as the donor molecule, there was a rapid esterification of PtdIns which reached completion within 5 min (Fig. 6A). The rapid attainment of completion of the reaction was associated with substrate depletion. By contrast, when [3H]arachidonate was employed as the

35 r

PH

Fig. 5. Effect of pH on acylating activity of granule membranes. Incubations were carried out with [3H]arachidonic acid for 30 min as described in Materials and Methods in the presence or absence of 10 PM lysophosphatidylinositol (PI). Values shown are expressed as cpm.103/mg protein and are representative of two separate experiments.

donor molecule, reacylation was linear for up to at least 30 min (Fig. 6A). In the granule membrane fraction, esterification also proceeded much more rapidly when [ “C]arachidonoyl-CoA was employed as the donor molecule; the maximum reaction rate was exhibited

70 rA

A

1

10

l Homogenate

0

20

10

Time

0

2

I

6 1M

8

10

12

IysoPl

Fig. 4. The effect of the lysophosphatidylinositol concentration on acid incorporation into phosphatidylinositol (PI) of total homogenate and intact zymogen granules. Incubations were carried out for 30 min (pH 7.4). Values shown are expressed as cpm.103/mg protein and are representative of three independent experiments.

[ ‘H]arachidonic

30

LO

(mln)

Fig. 6. A comparison of the time-course incorporation of [‘Hlarachidonic acid and [‘4C]arachidonoyl-CoA into phosphatidylinositol (PI) of (A) total homogenate and (B) granule membranes. Incubations were carried out for indicated intervals with either [i4C]arachidonoyl CoA or [3H]arachidonic acid as described in Materials and Methods, using 0.05 pg protein. The results depicted in each panel are expressed as cpm.103 and are representative of three separate experiments.

249

I

-

02 b Homogenate

01

0 Memtrones

25

50 /1g

75

100

00 125

z D 0

2

L? 0 0 s

protein

Fig. 7. Effect of protein concentration on reacylation of lysophosphatidylinositol (PI) by total homogenate and granule membranes. Incubations were carried out for 2 min with [‘4C]arachidonoyl-CoA and 10 PM 1ysoPI as described in Materials and Methods (pH 8.3). Results shown are expressed as cpm. 10’ and are representative of three separate experiments.

within the first 5 min (Fig. 6B). During the 5-30 min time interval, the rate fell to values comparable to that observed using [ 3H]arachidonate. However, after 30 min, esterification was still maintained at a much higher level (Fig. 6B). A comparison of the relative abilities of homogenate and granule membrane fractions to reacylate lysoPtdins, using [i4C]arachidonoyl-CoA as the donor, over a range of protein concentrations is depicted in Fig. 7. The kinetics of the two curves were disparate in that the peak effect observed with homogenate was obtained with 25 pg protein and gradually diminished at higher protein concentrations. Reacylating activity of the granule membrane fraction increased precipitously with the addition of up to 25 pg protein and continued a less striking increase with up to 100 pg protein (Fig. 7). Reacylation of IysoPtdCho The ability of 1ysoPtdCho to serve as an acceptor molecule was also examined. Again, over a range of protein concentrations the homogenate possessed a

1"" -2.0

$

- 1.5

2

- 1.0

2

E

2

z

h Homogenate -05

0 Membranes

0

25

50

pg

75

100

0.0 125

2 _o 0 B m

protein

Fig. 8. Effect of protein concentration on reacylation of 1ysoPtdCho by total homogenate and granule membranes. Subcellular fractions were incubated for 2 min with [i4C]arachidonoyl acid-CoA and 10 CM 1ysoPtdCho (pH 8.3). Values shown are expressed as cpm.103and are representative of three separate experiments.

greater ability to promote reacylation of lysoPtdCho, using [ “C]arachidonoyl-CoA, than did the granule membrane fraction (Fig. 8). Moreover, the kinetics of the two curves were similar to those obtained when 1ysoPtdIns was employed as the acceptor molecule (Fig. 7), in that esterification of the 1ysoPtdCho by the homogenate reached a peak with 25 pg protein and then gradually declined (Fig. 8). With the granule membrane fraction, esterification continued to increase with protein concentrations as high as 100 pg (Fig. 8). Again, the increase in reacylation was more striking at the lower protein concentrations. It should be noted that reacylation of PtdCho averaged 39 900 cpm/mg protein in the absence of added acceptor, as opposed to 189 100 cpm/mg protein in the presence of 10 PM lysoPtdCho, which represents a 4-5-fold increase in reacylating activity produced by the addition of acceptor (n = 2). The reacylation of choline phosphoacylglycerol paralleled that of PtdIns, in that reacylating activity of the membrane fraction in the presence or absence of 10 PM 1ysoPtdCho was about two orders of magnitude higher at pH 8.3 than 6.5. Discussion

The esterification and de-esterification of arachidonic acid into and from membrane phospholipids has been the focus of numerous studies. Yet, a great deal still remains unknown as to the exact cellular sources of arachidonic acid which are involved in these pivotal reactions. The basic contributions of this study were: (a) to localize to the zymogen granule membrane the enzymes involved in deacylation-reacylation in exocrine pancreas; and (b) to partially characterize the properties of these enzyme systems. The biochemical characterization of PLA z activity associated with granule membranes is similar to that reported for phospholipases found in various subcellular membrane fractions, in that enzyme activity is dependent upon the presence of Ca*+ and has an alkaline pH optimum [17,18]. Although some Ca*+-independent activity of the PLA, was detected, the enzyme could be activated in the physiological range of Ca*+ concentrations. In light of the critical role of Ca2+ in stimulussecretion coupling, the finding that physiological increments of ionic Ca*+ are able to activate PLA, of zymogen granule membranes prompts speculation that Ca2+ activation of PLA, plays an important role in the membrane fusion processes associated with exocytosis. The PLA, activity of the granule membrane fraction was comparable to that observed in the total homogenate. Moreover, the PLA, activity of the membrane fraction was linear up to at least 90 min, in contrast to the PLA, activity of the homogenate which exhibited linearity for only up to 30 min. These findings suggest that a functionally significant portion of the PLA,

250 activity of the acinar cell may reside in the zymogen granule. Purified chromaffin granule membranes have also been found to contain PLA, activity [19]. The differential localization of optimal acylating activities of IysoPtdIns and 1ysoPtdCho as described in our previous study suggests the existence of two distinct acyltransferases [6]. Additionally, the microsomal fraction was virtually devoid of reacylating activity when 1ysoPtdIns was the exogenously added acceptor, but did reacylate 1ysoPtdCho [6]. The present findings revealed that the kinetics of the protein concentration curves for the reacylation of 1ysoPtdCho in homogenate and membrane fraction paralleled that for the reacylation of 1ysoPtdIns. Both reactions expressed an alkaline pH optimum. These results suggest that if indeed there are two reacylating systems for 1ysoPtdIns and 1ysoPtdCho in exocrine pancreas, they share some common properties (see also Ref. 20). It should also be noted that the kinetic behavior of the granule membrane fraction was quite different from that of the total homogenate, in that reacylating activity was linear over a wide range of protein concentrations. In homogenate, the decrease in acyltransferase activity with increasing amounts of protein could be a consequence of the action of endogenous inhibitors or proteolytic enzymes present in this relatively crude fraction. The incorporation of fatty acids into phospholipids by the deacylation-reacylation cycle requires the formation of the corresponding acyl-CoA which is catalyzed by an acyl-CoA synthetase, and a subsequent transfer of the fatty acid by an acyl-CoA transferase. Thus, the use of [‘4C]arachidonoyl-CoA makes the analysis of arachidonate incorporation independent of acyl-CoA synthetase and allows a more direct assessment of acyltransferase activity. The much slower reaction rate when [ 3H]arachidonic acid, rather than [‘4C]arachidonoylCoA, was employed as the donor molecule implies that the granule fraction is virtually devoid of CoA synthetase activity. Indeed, in other test systems, acyl-CoA synthetase is primarily associated with the microsomal and mitochondrial fractions [21]. The synthetase-mediated step appears to be fatty acid-selective [22,23] and thus together with the acyltransferase reaction, confers fatty acid specificity of phospholipids. Phosphoinositides of exocrine pancreas are enriched in arachidonic acid in the C-2 position [3] and the liberation of arachidonate from this site has been proposed to be mediated through the action of PLA, [l]. However, evidence also exists for the participation of other putative pathways leading to arachidonic acid release [24]. Demonstration of Ca2+-dependent PLA, activity of zymogen granule membrane proteins is concordant with previous findings suggesting a link between the deacylation-reacylation reaction and the secretory response of exocrine pancreas [6,7]. The elevation in cytosolic Ca2+, together with requisite co-factors such as guanine

nucleotide binding proteins [24] might lead to the activation of PLA, in the secretory granule. The transient rise in the lysophospholipid levels might promote granule and plasma membrane fusion reactions, before being reacylated. Indeed, the fact that the granule fraction possessed only a limited ability to reacylate lysophospholipid relative to the total homogenate may underlie an important functional role for the lysophospholipids in exocytotic amylase secretion. However, we were unable to determine whether activation of the deacylating and reacylating enzymes occurs in situ in response to pancreatic secretagogues. Stimulation of pancreata prior to granule isolation alters the sedimentation properties of the granules making their isolation difficult. However, it is of interest to note that the rates of deacylation and reacylation are stimulated by agonists in exocrine and glioma cells [25,26]. In conclusion, this study provides evidence for the existence of the deacylating-reacylating enzyme system in zymogen granule membranes. However, the precise physiological role(s) of these enzymic reactions is not yet known. Further proof of the direct involvement of the granule membrane PLA, and acyltransferases in arachidonate turnover in exocrine pancreas requires the demonstration of the in situ activation of the deacylating-reacylating enzymes, as well as the correlation of effects of enzyme inhibitors in intact cells with those obtained in broken cell preparations. Acknowledgements This work was supported by PHS Grant AM-28029. We thank Dr. Archana Chaudhry and Ms. Maureen Adolf for their invaluable assistance in various phases of this study. References

4 5 6 7 8

9

Rubin, R.P. (1986) in Phosphoinositide and Receptor Mechanisms (Putney, J.W., Jr., ed.), Receptor Biochemistry and Methodology, Vol. 7, pp. 1499162, A.R. Liss, New York. Lands, W.E.M. and Crawford, C.G. (1976) in The Enzymes of Biological Membranes (Martonosi A., ed.), Vol. 2, pp. 3-85, Plenum Press, New York. Geison, R.L., Banschbach, M.W., Sadeghian K. and HokinNewerson, M. (1976) Biochem. Biophys. Res. Commun. 68. 3433 349. LeBel, D. and Beattie, M. (1984) Biochim. Biophys. Acta 769, 611-621. Bumham, D.B., Munowitz, P., Thorn, N. and Williams, J.A. (1985) B&hem. J. 227, 7433751. Rubin, R.P. (1983) B&hem. Biophys. Res. Commun. 112, 502507. Halenda, S.P. and Rubin, R.P. (1982) Biochem. J. 208, 713-721. Ohara, 0.. Tamaki, M., Nakamura, E., Tsumta, Y., Fujii, Y., Shin, M., Teraoka, H. and Okamoto, M. (1986) J. Biochem (Tokyo) 99, 733-739. DeLisle, R.C., Schulz, I., Tyrakowski, T., Haase, W. and Hopfer, U. (1984) Am. J. Physiol. 246. G411-418.

251 10 Laychock, S.G., Hoffman, J.M., Meisel, E. and Bilgin, S. (1986) B&hem. Pharmacol. 35, 2003-2008. 11 Matsuzawa, Y. and Hostetler, K.Y. (1980) J. Biol. Chem. 255, 5190-5194. 12 Bradford, M.M. (1976) Anal. B&hem. 72, 248-254. 13 Laychock, S.G. (1987) Endocrinology 120, 517-524. 14 Ballou, L.R. and Cheung, W.Y. (1983) Proc. Natl. Acad. Sci. USA 80, 5203-5207. 15 Van den Bosch, H.A. (1980) Biochim. Biophys. Acta 604.191-246. 16 Taylor, C.W., Merritt, J.E., Putney, J.W., Jr. and Rubin, R.P. (1986) Biochem. J. 238, 765-772. 17 Kramer, R.M., Checani, G.C., Deykin, A., Pritzker, CR. and Deykin, D. (1986) Biochim. Biophys. Acta 878, 3944403. 18 Pfeilschifter, J., Pignat, W., Marki, F. and Wiesenberg, I. (1989) Eur. J. Biochem. 181, 237-242.

19 Husebye, E.S. and Flatmark, T. (1987) Biochim. Biophys. Acta 920, 120-130. 20 Sanjanwala, M., Sun, G.Y. and MacQuarrie (1989) Archiv. Biothem. Biophys. 271, 407-413. 21 Smith, P.B., Reitz, R.C. and Kelley, D. (1982) Biochim. Biophys. Acta 713, 128-135. 22 Neufeld, E.J., Sprecher, H., Evans, R.W. and Majerus, P.W. (1984) J. Lipid. Res. 25, 288-293. 23 Morand, O., Carre, J.B., Homayoun, P., Niel, E., Baumann, N. and Bourre, J.M. (1987) J. Neurochem. 48, 1150-1156. 24 Rana, R.S. and Hokin, L.E. (1990) Physiol. Rev. 70, 115-164. 25 Soling, H-D, Fest, W., Schmidt, T., Esselmann, H. and Bachmann, V. (1989) J. Biol. Chem. 264, 10643-10648. 26 DeGeorge, J., Ousley, A.H., McCarthy, K.D., Lapetina, E.G. and Morell, P. (1987) J. Biol. Chem. 262, 807778083.