ARCHIVES
OF RIOCHEMIS’I‘KY
AND
BIOPHYSICS
Vol. 287, No. 2, June, pp. 283-287, 1991
Interaction
of Protein Kinase C with Phosphoinositides
Abha Chauhan,l
Hans Brockerhoff,
New York State Institute Staten Island, Neu: York
for Basic Research 10314
H. M. Wisniewski, in Developmental
Received December 3, 1990, and in revised form February
The possible existence and function of an inositol cycle and an inositide shuttle, and their role in cation transport, were first discussed some 30 years ago (1). (Similar discussions took place in the laboratories of R. M. C. Dawson, Cambridge, and J. N. Hawthorne, Nottingham).
PI = PIP ti PIP2 Inositide
shuttle
It is now held that many agonists, via cell membrane receptors and G-proteins, activate an intracellular phospholipase C which splits phosphatidylinositol 4,5 bisI To whom correspondence
should he addressed.
0003.9861/91 $3.00 c ,opyrlght ~1 1991 by Academic Press, All rights of reproduction in an> form
1050 Fvrest
Hill
Road,
5, 1991
Calciumjphosphatidylserine-dependent protein kinase C (PKC) is activated by phosphatidylinositol 4,5-hisphosphate (PIP,), as well as by diacylglycerol (DG) and phorbol esters. Here we report that PIP2, like DG, increases the affinity of PKC for Ca’+, and causes Ca”‘dependent translocation of the enzyme from the soluble to a particulate fraction (liposomes). Phosphatidylinositol 4-phosphate (PIP) also displaces phorhol ester from PKC and causes Ca”-dependent translocation of the enzyme to liposomes, but is much less efficient than PIP,, and a much weaker activator, with a histone phosphorylation U(PIP)/U(PIP,) of -0.15. Scatchard analysis indicates competitive inhibition between PIP and phorbol ester with K,(PIP) = 0.26 mol% as compared with Ki(PIP2) = 0.043 mol%. No effect of phosphatidylinositol (PI) on phorbol ester binding to PKC, translocation of PKC, or activation of PKC was observed. These results suggest that both PIP and PIP2 can complex with PKC, but full activation of the enzyme takes place only when PIP is converted to PIP2. We suggest that an inositide interconversion shuttle has a role in the regulation of protein phosphorylation. ‘~8 1991 Academic Press, Inc.
PI --) PIP + PIP2 c t I IP:] Inositol cycle
and Ved P. S. Chauhan
Disabilities,
phosphate (PIP,)” into two second messengers, inositol 1,4,5-trisphosphate (IP,) and diacylglycerol (DG) (2, 3). IP3 then releases Ca2+ from an intracellular store, which initiates many cellular responses, centrally among them the activation of protein kinase C (PKC), a phospholipidand calcium-dependent protein phosphorylating enzyme. DG stays in the plasma membrane, from where it causes the translocation of PKC from the cytosol to the membrane and the activation of PKC by increasing its affinity for Ca”+. This picture has recently been somewhat blurred by the discovery that the DG arising in a cell upon stimulation is derived largely from phosphatidylcholine (PC) rather than from PIP2 (4). While the inositol cycle has gained wide attention, the inositide shuttle has been almost completely neglected, despite the facts that the required kinases and hydrolases are ubiquitous and of comparable activity, and that the turnover of the 4- and 5-monophosphate groups is the fastest of any lipid-bound phosphate (5-8). The existence of an active interconversion of the inositides is demonstrated (8) in pig erythrocytes: they rapidly incorporate phosphate into the inositides, although they cannot synthesize PI de IZOUO.This can be explained only by an interconverting phosphate shuttle mechanism. The recent report that PIP, is a potent activator of protein kinase C, but phosphatidylinositol4-phosphate (PIP) and phosphatidylinositol (PI) are not (9), opens the way for the inositide shuttle as a regulator of protein phosphorylation by PKC. This study probes further into this possibility. It describes some parallels between DG. PKC, PIP2. PKC, and PIP * PKC; in particular, their dependence on Ca’+, the binding of the effecters to PKC in competition with phorbol ester, and the “translocation” of the kinase from the aqueous to the membrane fraction caused by the effecters.
’ Abbreviations wed: DG, diacylglycerol; PI, phosphatidylinositol; PIP, phosphatidylinositol 4-phosphate; PIP,, phosphatidylinositol 4,5-hisphosphate; PS, phosphatidylserine; PC, phosphatidylcholine; PKC, protein kinase C; PDI3u, phorhol 12,1X-dihutyrate; IP,, inositol 1,4,5trisphosphate.
283 Inc. reserved.
284
MATERIALS
CHAUHAN
AND METHODS
Materials. Ultrogel ACA 202 was obtained from IBF Biotechnics, phenyl-Sepharose 4B from Pharmacia, DE-52 from Whatman, and Triton X-100 from Aldrich. Calf thymus histone type III-S, polyethylene glycol 8000, phorbol 12,13-dibutyrate (PDBu), leupeptin, phenylmethanesulfonyl fluoride, bovine serum albumin, phosphatidylserine (PS), PC, PI, PIP, PIP2, and 1,2-sn-dioleoylglycerol (DG) were purchased from Sigma. The purity of lipids was confirmed by thin-layer chromatography. BIS and polyacrylamide were from Bio-Rad. [y-a*P]ATP (3000 Ci/mmol) and [3H]PDBu (18.8 Ci/mmol) were procured from New England Nuclear, and hydrofluor was from National Diagnostics. Charles River CD male rats were used for the source of protein kinase C. Special care was taken to ensure that PIP, and PIP were the true PKC effecters and did not degrade into by-products such as DG or fatty acids. First, inositides from three independent sources gave identical results, in hundreds of assays. Furthermore, a mixed micellar assay was carried out with [3H]inositol-labeled PIP2 and nonradioactive ATP. The reaction was quenched with chloroform/methanol (2:l). After centrifugation, radioactivity was counted in both aqueous (IP,) and organic phase (PIP,) to investigate if any hydrolysis of PIP2 into IP? and DG occurred during the reaction. No radioactive counts were observed in the aqueous phase. Even greater confidence in the identity of the polyphosphoinositides as PKC effecters, however, derives from a comparison of the PKC affinity of PIP2 against the affinities of the possible contaminating by-products, DG or fatty acids. Their K’s (mol% at half maximal activity) are much higher (ca. 1 mol% and ca. 30 mol%) (10) than that of PIP, (ca. 0.04 mol%). With an affinity so much smaller than that of PIP, the activities measured with this effector could not possibly be obtained even if all of the PIP, had been converted to byproducts. Purification of protein kinase C. The enzyme was purified from rat brains essentially by the method of Woodgett and Hunter (11) with DE52, phenyl-Sepharose 4B, and PS-acrylamide column chromatography. The final preparation was concentrated by reverse dialysis against solid polyethylene glycol 8000, dialyzed overnight at 4°C into 20 mM TrisHCl, pH 7.5, containing 0.1% (v/v) mercaptoethanol, 100 pM EGTA, and 10% glycerol, and stored at 4°C. The enzyme preparation obtained is a mixture of the closely related isozymes N, &, &, and y (3). The specific activity of PKC was 3-4 Fmol rng- ’ mini at 37°C in a mixed micellar assay system of 0.3% Triton X-100 (12) containing 9 mol% PS, 2 mol% DG, and 50 pM Cazt, with histones III-S as substrate. Freshly prepared enzyme was used for all the experiments. Phorbol ester binding. The effect of phosphoinositides (PI, PIP, or PIPJ on phorbol ester binding to PKC was studied in a mixed micellar assay as described previously (13). Briefly, Triton X-100 mixed micelles containing 20 mol% PS and different concentrations of phosphoinositides or DG were prepared by solubilizing the dried lipids in 3% Triton X-100. Ultrogel ACA 202 columns were prepared by filling 2 ml of gel into silanized Pasteur pipets; the columns were equilibrated with 20 mM Tris-HCl, pH 7.5, 500 pM Ca’+, and 0.015% Triton X-100. The incubation was carried out at room temperature for 5 min in a total volume of 100 ~1 containing 10 ~1 of mixed micelles, 500 FM CaCl,, 20 mM TrisHCl, pH 7.5, [3H]PDBu, and enzyme sample. Fifty microliters of the incubation mixture was then loaded on the Ultrogel columns, and bound [3H]PDBu and free [“H]PDBu were determined by collecting eluates of 0.9 and 1.8 ml of equilibration buffer, respectively (14). Hydrofluor (20 ml) was added to the eluates, which were vortexed and counted. Nonspecific binding was determined in the presence of excess unlabeled PDBU (10 PM), and specific binding for each sample was calculated as the difference between total and nonspecific binding. Translocation of PKC from medium to liposomes. Multilamellar liposomes (10 Kmol) containing 20 mol% PS and 80 mol% PC or PC plus other specified lipids (PI, PIP, PIP,, or DG) were prepared by hydrating the dried lipids in 20 mM Tris-Cl, pH 7.5, containing 0.1% mercaptoethanol (buffer A), followed by centrifugation at 300,OOOgfor 20 min. These liposome pellets were washed twice and then suspended in 0.2 ml of buffer A. PKC partially purified by DE-52 and phenyl-Sepharose column chromatography was incubated in a total volume of 0.2 ml with
ET AL. 1 pm01 of multilamellar liposomes at varying Ca’+ concentrations, 10 MgCl,, and 20 pg leupeptin. The incubation mixture was centrifuged aft,er 5 min at 300,OOOgfor 20 min and the PKC activity in the supernatant measured by mixed micellar assay. The pellet was suspended in 0.2 ml of buffer A containing 2 mM EDTA, and the suspension vortexed for 1 min and centrifuged at 300,OOOgfor 20 min. The supernatant thus obtained (representing membrane-bound PKC) was assayed for PKC activity by mixed micellar assay. The pellet was solubilized again with Triton X-100 and assayed for PKC activity. An average of 5% of the total PKC activity was obtained only after Triton X-100 treatment, and could not be extracted from the pellet with EDTA treatment. mM
RESULTS Efficacies of phosphoinositides and diacylglycerol. Figure 1 repeats part of a previous study (9) with the
L
1
m lll!l! PS
0
,“, MO!
r -TPI 05 40,’
PIP 0, MO,
PIP
O5 Ma’
-
FIG. 1. Effect of phosphoinositides and diacylglycerol on PKC activity. PKC was purified from rat brains with DE-52, phenyl-Sepharose, and PS-acrylamide column chromatography (11). Activity was assayed by measuring the incorporation of ‘*P from [y-“‘P]ATP into histones in a mixed micellar assay (12). Total reaction volume (0.1 ml) contained 20 mM Tris-Cl, pH 7.5, 0.8 mg/ml histone III-S, 50 pg/ml leupeptin, 12.5 pM ATP, 10 mM MgCl,, 50 pM Ca’+, and 10 ~1 PS, PS/phosphoinositides, or PS/DG lipid micelles. The concentration of lipids in the reaction mixture (with respect to Triton X-100) were: PS, 9 mol% or 445 pM; PI or PIP, 0.1 mol% or 4.95 pM and 0.5 mol% or 22.75 pM; PIP,, 0.05 mol% or 2.47 pM and 0.1 mol% or 4.95 pM; DG, 1 mol% or 49.5 PM and 2 mol% or 99 pM.
PROTEIN
KINASE
C AND
285
PHOSPHOINOSITIDES
100 z 2
PI
80
n i
60
y c9 90
40
a
PIP
20
DG PIP*
01 0 0
7
6
5
4
3
I
I
I
I
0.2
0.4
0.6
0.8
Phosphoinositide
4
,/ 1.0
or DG (mol %)
- Log [Ca2+] FIG. 2. Effect of phosphatidylinositol 4,5-hisphosphate and diacylglycerol on reaction velocity of PKC at various concentrations of Ca*+. Protein kinase C activity was measured as described in Fig. 1. At zero Ca’+ concentrations 4 mM EGTA was present instead of CaC&. Mixed micelles contained Triton X-100 (3% solution); and PS, 9 mol% (A); or PIP,, 0.1 mol% (0); or DG, 2 mol% (Cl).
difference that completely pure PKC (4 nmol/pg protein/ min, a mixture of PKC cy,,L&,&, and y) (3) has here been used, and PKC activity has been assessed at two different concentrations of each effector. There is no difference from the previous results obtained with a partially purified PKC. However, it had been stated (9) that “PI and PIP show little, if any, capacity of activating the kinase.” A closer look at the results in the previous and the present study shows that the activity of PI may in fact be entirely negligible; but PIP does have some effect. This has also been found by others (15). From a comparison of the data of Fig. 1, with both lipids at 0.1 mol%, it appears that PIP has roughly 15% the effect of PIP, on histone phosphorylation by PKC. Calcium dependency. Figure 2 shows that PIP2, like DG, stimulates PKC by shifting the Cazf dose-response curve for activation of the enzyme to a lower Ca2+ concentration. The affinity of Ca2+ for PKC rises by several orders of magnitude in the presence of DG as well as of PIP,; when it is considered that V,,, is 2-3 times greater for DG . PKC than for PIP,. PKC (9), the Ca-activation curves for DG * PKC and PIPZ. PKC nearly coincide. The half-maximal binding constant Kc, is around lop7 for both kinase species. PKC is the phorbol Effect on phorbol ester binding. ester receptor of cells and PIP, displaces phorbol ester competitively from PKC, in the manner of DG (13). The effect of the other phosphoinositides, PI and PIP, on phorbol ester binding to PKC was studied in a mixed micellar assay (14), and compared with that by PIP, or DG (Fig. 3). While no effect of PI on phorbol ester binding was observed, PIP inhibited the specific binding of [“H]PDBu. However, the inhibition was much less pronounced with PIP than with PIP2 or DG. A Scatchard
FIG. 3. Inhibition of phorbolester binding by polyphosphoinositides and DG. Binding of [“H]PDBu binding to PKC was measured with Triton X-100 mixed micelles containing PS (20 mol% relative to Triton, 100 mol%) and variable mol% of PI (A), PIP (O), PIP, (Ml, or DG (01 as described under Materials and Methods. Nonspecific binding was determined in the presence of excess of unlabeled PDBu (10 FM) and subtracted.
analysis of binding-inhibition between 0.01 and 0.5 mol% PIP showed no change in maximal binding (B,,,) but a decrease in the binding affinity for [3H]PDBu with increasing concentration of PIP (Fig. 4a), indicating a competitive nature of inhibition between phorbol ester binding a.nd PIP binding. The plot of K&, the apparent dissociation constant for [3H]PDBu/enzyme in the presence of PIP, was a linear function of PIP concentration (Fig. 4b), with a h, (the dissociation constant for PIP binding) of 0.26 mol%. Under the same conditions the Ki for PIP, was 0.043 (13).
It has been reported by others (15) that none of the inositides competes with phorbol ester for binding to PKC.
:-:n0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
[3H] PDBu Bound, pmol FIG. 4a. Competitive inhibition of [3H]PDBu binding by PIP in mixed micelles. A Scatchard analysis of specific binding of [“H]PDBu at variable concentrations of PDBu (lo-40 nM) as a function of PIP concentration in Triton X-100 mixed micelles containing 20 mol% PS (0). PIP in mol% of Triton X-100 is 0 (O), 0.1 (A), 0.25 (ml, and 0.5 (0).
286
CHAUHAN
8
I
I
-0.4
-0.3
-0.2
ET AL
1
-0.1
0
0.1
0.2
0.3
0.4
0.5
PIP, mol% FIG. 4b. Apparent dissociation constants for [“H]PDBu as a function of PIP concentration. Ki, the apparent dissociation constant for [“H]PDBu/Enzyme in presence of PIP, was determined from Fig. 4a for each PIP concentration by Scatchard analysis. K,, the dissociation constant for PIP, represented by the negative of the X intercept, is 0.26 mol%.
We have no explanation for this discrepancy with our results; it will have to be resolved in future studies. Translocation. The translocation of PKC from aqueous to liposomal phase, with the percentage of kinase in the pellet rising from about 30 to about 60%, is achieved by the lipid effecters DG and PIP, at 1 PM Ca concentration (Fig. 5). Figure 6 shows percentage translocation as a function of effector concentration. DG and PIP, are of comparable activity; PIP is much less effective; and PI hardly effective at all.
o!-rz--F-
8
100
Ca’+[p M] FIG. 5. Effect of calcium on translocation of PKC to multilamellar liposomes. Multilamellar liposomes (10 pmol) containing 20 mol% of PS and 80 mol% total of PC and other specified lipids were prepared as described under Materials and Methods. PKC partially purified by DE-52 and phenyl-Sepharose column chromatography was incubated for 5 min in the presence of 1 Fmol of liposomes, 10 mM MgCl*, different concentrations of CaCl*, and 20 ag leupeptin (final volume 0.2 ml) and centrifuged at 300,OOOg for 20 min. PKC activity in the supernatant was measured by mixed micellar assay. The pellet was suspended in 0.2 ml of buffer A containing 2 mM EDTA, vortexed for 1 min, and centrifuged. The supernatant thus obtained (representing membrane-bound PKC activity) was assayed for PKC activity. The concentration of PIP (o), PIP, (w), or DG (0) was 0.5 mol% and the PC/PS ratio was 79.5/ 20. The control sample (A) had a PC/PS ratio of SO/ZO.
can both compounds activate the kinase in a competitive arrangement? The answer may lie in a circumspect interpretation of the data. Competitive activation does not imply that the effecters must all bind to the same site,
DISCUSSION
of DG - PKC and PIP, * PKC. The present Similarity study shows that PIP2 activates PKC in the manner of DG and phorbol esters by lowering the concentration of the Ca”+ required for PKC-activation, and by translocating the kinase from the soluble to the membrane fraction. There is a difference between the two activated forms in their maximal velocity, (V,,/V,,,,) = 2.5, and in their affinity, (KUG/KpIP2) = 50 (9); but apart from this kinetic variation both activated enzymes behave remarkably similar. They both use phosphatidylserine for an acidic lipid membrane matrix; they have an absolute requirement for Ca’+, i.e., they form a ternary effector . Ca . kinase complex, they both compete with phorbol ester for binding to PKC (13), they seem to have similar substrate specificity (15), and they cause calcium-dependent water-membrane translocation of PKC. Questions arise: how can two structurally greatly divergent molecules interact with the kinase to produce activated forms of nearly the same properties? The traditional PKC-effector, DG, can form three H-bonds with the kinase; phorbol PKC has the same H-bond arrangement, superposable on that formed with DG (16); but PIP2 has only two corresponding H-bonding sites, the fatty ester CO groups; the OH group is missing, replaced by a bulky inositol trisphosphate group. PIP2 cannot possibly bond to the DG receiving site in the same isosteric manner. How, then,
100 80 60
PI
al
c
0 .l 2 .3 .4 .5 .6 .7 .8 .9 1 Phosphoinositide or DG (mot%) FIG. 6. Effect of phosphoinositides or DG on translocation of PKC to multilamellar liposomes. Translocation of PKC was measured as described in the legend to Fig. 3 in the presence of 50 pM Ca2+ and varying concentrations of PI, PIP, PIP,, and DG.
PROTEIN
KINASE
C AND
only that the binding of one prevents the binding of another (13). This could be achieved not only by competition for the same site on the enzyme but possibly also by the binding of the effecters at distinct but overlapping enzyme-effector interfaces (17). Translocation. The phenomenon of the translocation of PKC from soluble to membrane fraction has been studied with intact cells (18-20) and liposomes (21). Both DG and phorbol ester elicit this effect, which also requires Ca’+. Peculiarly, translocation is not complete (U-20). This is confirmed in our experiments (Fig. 6). Even without Ca”+ and effector, ca. 30% of PKC is found in the pellet; with Ca2+ and DG or PIP2, this concentration rises, but not above ca. 70%, except at high, unphysiological concentrations of Ca”+ (100 PM) when a pellet PKC concentration of 90% is reached. Thus, about 40% at most of the kinase is translocated even with the most active effecters. It could be speculated that only one or two of the isozymes present (Q~, &, &, y) are translocated; but the general similarity between these enzymatic species in regard to their activation (3) makes such an explanation unlikely. It has recently been reported (22) that DG activates translocation not only of PKC but also of choline phosphate cytidyltransferase. This suggests that enzyme translocation by DG (and PIPZ?) may be of widespread occurrence. The rapid turnover of the inositide 4- and &phosphates (5-8, 23), together with the results of the present study, shows dependence of PKC activation on the state of phosphorylation of the lipids, and inevitably suggests that the inositide shuttle, PI = PIP 2 PIPZ, at least to some extent regulates the phosphorylation of cellular proteins. Most of the phosphoinositide is in the form of PI, which cannot complex with PKC (no effect on phorbol ester binding or translocation or activation of the kinase is observed). PIP, on the other hand, can displace phorbol ester from PKC, and also cause translocation of the kinase, but cannot activate PKC effectively, suggesting that PIP can complex PKC to the membrane, when a threshold Ca2+ concentration is reached; but this PIP. Ca * PKC complex is an inactive form. The enzyme is turned on upon phosphorylation of PIP to PIP, by PIP-Skinase, and this results in the formation of an active CamPIP,. PKC complex. It cannot, at present, be decided if PIP. PKC or PIP by itself is the full substrate for the PIP-Skinase. Conventionally, the PIPZ-phospholipase C is considered the cascade enzyme to be activated first, and though it has been shown that with support of a G-protein (24) activation can take place at lower than micromolar Ca’+ levels (25), it is still peculiar that an enzyme which initiates Ca” release should itself depend on that cation, directly or indirectly. PIP-Skinase and PI-4-kinase, on the other hand, are independent of calcium. We suggest that the activation of PIP-kinase precedes the activation Inositide
shuttle.
287
PHOSPHOINOSITIDES
of PIP,-phospholipase C. This could increase the pool of PIP,, and lead to increased membrane-PKC binding, or to the opening of Ca-cages (23) and the production of the Ca-spike (26, 27). ACKNOWLEDGMENT This work was supported by funds from the New York State Office of Mental Retardation and Developmental Disabilities.
REFERENCES 1. Brockerhoff, 1766.
H., and Ballou, C. E. (1962) J. Biol. Chem. 237,1764-
2. Streb, H., Irvine, R. F., Berridge, M. J., and Schulz, I. (1983) Nature 306,67-69. 3. Nishizuka,
Y. (1988) Nature 334, 661-665.
4. Exton, cJ.H. (1990) J. Biol. Chem. 265, l-4. 5. Brockerhoff, 52.
H., and Ballou, C. E. (1962) d. Biol. Chem. 237, 49-
6. Garrett, N. E., Garrett, R. J. B., and Talwalkar, Physiol. 87, 63-70.
R. T. (1975) J. Cell.
7. Palmer, F. B. St. C. (1985) Can. J. Biochem. CellBiol. 8. Schneider, R. P., and Kirschner, Actn 202, 283-294. 9. Chauhan, V. P. S., and Brockerhoff, Res. Commun. 155,18-23.
63,927-931.
L. B. (1970) Biochim. H. (1988) B&hem.
Biophys. Biophys.
10. Chauhan, V. P. S., Chauhan, A., Deshmukh, D. S., and Brockerhoff, H. (1990) Life Sci. 47, 981-986. 11. Woodgett, 4843.
J. R., and Hunter,
T. (1987) J. Biol. Chem. 262, 4836-
12. Hannun, Y. A., Loomis, C. R., and Bell, R. M. (1985) J. Biol. Chem. 260, 10,039-10,043. 13. Chauhan, A., Chauhan, V. P. S., Deshmukh, D. S., and Brockerhoff, H. (1989) Biochemistry 28, 4952-4956. 14. Hannun, Y. A., and Bell, R. M. (1986) J. Biol. Chem. 261, 93419347. 15. Lee, M-H., and Bell, R. M. (1990) FASEB J. 4, A1912. 16. Brockerhoff, H. (1986) FEBS Lett. 201, 1-4. 17. Segel, I. H. (1976) in Biochemical Calculations, 2nd ed., pp. 246, Wiley, New York. 18. Sacktor, T. C., and Schwartz, J. H. (1990) Proc. N&l. Acad. 5%. USA 87, 2036-2039. 19. Persaud, S. J., Jones, P. M., Sugden, D., and Howell, S. L. (1989) FEBS Lett. 245, 80-84. 20. Vaartjes, W. J., de Haas, C. G. M., and van den Bergh, S. G. (1986) Biochem. Biophys. Res. Commun. 138,1328-1333. 21. Bazzi, M. D., and Nelsestuen, G. L. (1987) Biochemistry 26, 115122. 22. Kolesnick, R. N., and Hemer, M. R. (1990) J. Biol. Chem. 265, 10,900-10,904. 23. Brockerhoff, H. (1986) Chem. Phys. Lipids 39, 83-92. 24. Deckmyn, H., Shi-Ming, T., and Majerus, P. W. (1986) J. Biol. Chem. 261, 16,553-16,558. 25. Rana, R. S., and Hokin, L. E. (1990) Physiol. Reu. 70, 115-164. 26. Albert, P. R., and Tashjian, Jr., A. H. (1984) J. Biol. Chem. 259, 5827-5832. 27. Tashijian, Jr., A. H., Heslop, J. P., and Berridge, M. J. (1987) J. Riochem. 243, 305-308.
OF RIOCHEMIS’I‘KY
AND
BIOPHYSICS
Vol. 287, No. 2, June, pp. 283-287, 1991
Interaction
of Protein Kinase C with Phosphoinositides
Abha Chauhan,l
Hans Brockerhoff,
New York State Institute Staten Island, Neu: York
for Basic Research 10314
H. M. Wisniewski, in Developmental
Received December 3, 1990, and in revised form February
The possible existence and function of an inositol cycle and an inositide shuttle, and their role in cation transport, were first discussed some 30 years ago (1). (Similar discussions took place in the laboratories of R. M. C. Dawson, Cambridge, and J. N. Hawthorne, Nottingham).
PI = PIP ti PIP2 Inositide
shuttle
It is now held that many agonists, via cell membrane receptors and G-proteins, activate an intracellular phospholipase C which splits phosphatidylinositol 4,5 bisI To whom correspondence
should he addressed.
0003.9861/91 $3.00 c ,opyrlght ~1 1991 by Academic Press, All rights of reproduction in an> form
1050 Fvrest
Hill
Road,
5, 1991
Calciumjphosphatidylserine-dependent protein kinase C (PKC) is activated by phosphatidylinositol 4,5-hisphosphate (PIP,), as well as by diacylglycerol (DG) and phorbol esters. Here we report that PIP2, like DG, increases the affinity of PKC for Ca’+, and causes Ca”‘dependent translocation of the enzyme from the soluble to a particulate fraction (liposomes). Phosphatidylinositol 4-phosphate (PIP) also displaces phorhol ester from PKC and causes Ca”-dependent translocation of the enzyme to liposomes, but is much less efficient than PIP,, and a much weaker activator, with a histone phosphorylation U(PIP)/U(PIP,) of -0.15. Scatchard analysis indicates competitive inhibition between PIP and phorbol ester with K,(PIP) = 0.26 mol% as compared with Ki(PIP2) = 0.043 mol%. No effect of phosphatidylinositol (PI) on phorbol ester binding to PKC, translocation of PKC, or activation of PKC was observed. These results suggest that both PIP and PIP2 can complex with PKC, but full activation of the enzyme takes place only when PIP is converted to PIP2. We suggest that an inositide interconversion shuttle has a role in the regulation of protein phosphorylation. ‘~8 1991 Academic Press, Inc.
PI --) PIP + PIP2 c t I IP:] Inositol cycle
and Ved P. S. Chauhan
Disabilities,
phosphate (PIP,)” into two second messengers, inositol 1,4,5-trisphosphate (IP,) and diacylglycerol (DG) (2, 3). IP3 then releases Ca2+ from an intracellular store, which initiates many cellular responses, centrally among them the activation of protein kinase C (PKC), a phospholipidand calcium-dependent protein phosphorylating enzyme. DG stays in the plasma membrane, from where it causes the translocation of PKC from the cytosol to the membrane and the activation of PKC by increasing its affinity for Ca”+. This picture has recently been somewhat blurred by the discovery that the DG arising in a cell upon stimulation is derived largely from phosphatidylcholine (PC) rather than from PIP2 (4). While the inositol cycle has gained wide attention, the inositide shuttle has been almost completely neglected, despite the facts that the required kinases and hydrolases are ubiquitous and of comparable activity, and that the turnover of the 4- and 5-monophosphate groups is the fastest of any lipid-bound phosphate (5-8). The existence of an active interconversion of the inositides is demonstrated (8) in pig erythrocytes: they rapidly incorporate phosphate into the inositides, although they cannot synthesize PI de IZOUO.This can be explained only by an interconverting phosphate shuttle mechanism. The recent report that PIP, is a potent activator of protein kinase C, but phosphatidylinositol4-phosphate (PIP) and phosphatidylinositol (PI) are not (9), opens the way for the inositide shuttle as a regulator of protein phosphorylation by PKC. This study probes further into this possibility. It describes some parallels between DG. PKC, PIP2. PKC, and PIP * PKC; in particular, their dependence on Ca’+, the binding of the effecters to PKC in competition with phorbol ester, and the “translocation” of the kinase from the aqueous to the membrane fraction caused by the effecters.
’ Abbreviations wed: DG, diacylglycerol; PI, phosphatidylinositol; PIP, phosphatidylinositol 4-phosphate; PIP,, phosphatidylinositol 4,5-hisphosphate; PS, phosphatidylserine; PC, phosphatidylcholine; PKC, protein kinase C; PDI3u, phorhol 12,1X-dihutyrate; IP,, inositol 1,4,5trisphosphate.
283 Inc. reserved.
284
MATERIALS
CHAUHAN
AND METHODS
Materials. Ultrogel ACA 202 was obtained from IBF Biotechnics, phenyl-Sepharose 4B from Pharmacia, DE-52 from Whatman, and Triton X-100 from Aldrich. Calf thymus histone type III-S, polyethylene glycol 8000, phorbol 12,13-dibutyrate (PDBu), leupeptin, phenylmethanesulfonyl fluoride, bovine serum albumin, phosphatidylserine (PS), PC, PI, PIP, PIP2, and 1,2-sn-dioleoylglycerol (DG) were purchased from Sigma. The purity of lipids was confirmed by thin-layer chromatography. BIS and polyacrylamide were from Bio-Rad. [y-a*P]ATP (3000 Ci/mmol) and [3H]PDBu (18.8 Ci/mmol) were procured from New England Nuclear, and hydrofluor was from National Diagnostics. Charles River CD male rats were used for the source of protein kinase C. Special care was taken to ensure that PIP, and PIP were the true PKC effecters and did not degrade into by-products such as DG or fatty acids. First, inositides from three independent sources gave identical results, in hundreds of assays. Furthermore, a mixed micellar assay was carried out with [3H]inositol-labeled PIP2 and nonradioactive ATP. The reaction was quenched with chloroform/methanol (2:l). After centrifugation, radioactivity was counted in both aqueous (IP,) and organic phase (PIP,) to investigate if any hydrolysis of PIP2 into IP? and DG occurred during the reaction. No radioactive counts were observed in the aqueous phase. Even greater confidence in the identity of the polyphosphoinositides as PKC effecters, however, derives from a comparison of the PKC affinity of PIP2 against the affinities of the possible contaminating by-products, DG or fatty acids. Their K’s (mol% at half maximal activity) are much higher (ca. 1 mol% and ca. 30 mol%) (10) than that of PIP, (ca. 0.04 mol%). With an affinity so much smaller than that of PIP, the activities measured with this effector could not possibly be obtained even if all of the PIP, had been converted to byproducts. Purification of protein kinase C. The enzyme was purified from rat brains essentially by the method of Woodgett and Hunter (11) with DE52, phenyl-Sepharose 4B, and PS-acrylamide column chromatography. The final preparation was concentrated by reverse dialysis against solid polyethylene glycol 8000, dialyzed overnight at 4°C into 20 mM TrisHCl, pH 7.5, containing 0.1% (v/v) mercaptoethanol, 100 pM EGTA, and 10% glycerol, and stored at 4°C. The enzyme preparation obtained is a mixture of the closely related isozymes N, &, &, and y (3). The specific activity of PKC was 3-4 Fmol rng- ’ mini at 37°C in a mixed micellar assay system of 0.3% Triton X-100 (12) containing 9 mol% PS, 2 mol% DG, and 50 pM Cazt, with histones III-S as substrate. Freshly prepared enzyme was used for all the experiments. Phorbol ester binding. The effect of phosphoinositides (PI, PIP, or PIPJ on phorbol ester binding to PKC was studied in a mixed micellar assay as described previously (13). Briefly, Triton X-100 mixed micelles containing 20 mol% PS and different concentrations of phosphoinositides or DG were prepared by solubilizing the dried lipids in 3% Triton X-100. Ultrogel ACA 202 columns were prepared by filling 2 ml of gel into silanized Pasteur pipets; the columns were equilibrated with 20 mM Tris-HCl, pH 7.5, 500 pM Ca’+, and 0.015% Triton X-100. The incubation was carried out at room temperature for 5 min in a total volume of 100 ~1 containing 10 ~1 of mixed micelles, 500 FM CaCl,, 20 mM TrisHCl, pH 7.5, [3H]PDBu, and enzyme sample. Fifty microliters of the incubation mixture was then loaded on the Ultrogel columns, and bound [3H]PDBu and free [“H]PDBu were determined by collecting eluates of 0.9 and 1.8 ml of equilibration buffer, respectively (14). Hydrofluor (20 ml) was added to the eluates, which were vortexed and counted. Nonspecific binding was determined in the presence of excess unlabeled PDBU (10 PM), and specific binding for each sample was calculated as the difference between total and nonspecific binding. Translocation of PKC from medium to liposomes. Multilamellar liposomes (10 Kmol) containing 20 mol% PS and 80 mol% PC or PC plus other specified lipids (PI, PIP, PIP,, or DG) were prepared by hydrating the dried lipids in 20 mM Tris-Cl, pH 7.5, containing 0.1% mercaptoethanol (buffer A), followed by centrifugation at 300,OOOgfor 20 min. These liposome pellets were washed twice and then suspended in 0.2 ml of buffer A. PKC partially purified by DE-52 and phenyl-Sepharose column chromatography was incubated in a total volume of 0.2 ml with
ET AL. 1 pm01 of multilamellar liposomes at varying Ca’+ concentrations, 10 MgCl,, and 20 pg leupeptin. The incubation mixture was centrifuged aft,er 5 min at 300,OOOgfor 20 min and the PKC activity in the supernatant measured by mixed micellar assay. The pellet was suspended in 0.2 ml of buffer A containing 2 mM EDTA, and the suspension vortexed for 1 min and centrifuged at 300,OOOgfor 20 min. The supernatant thus obtained (representing membrane-bound PKC) was assayed for PKC activity by mixed micellar assay. The pellet was solubilized again with Triton X-100 and assayed for PKC activity. An average of 5% of the total PKC activity was obtained only after Triton X-100 treatment, and could not be extracted from the pellet with EDTA treatment. mM
RESULTS Efficacies of phosphoinositides and diacylglycerol. Figure 1 repeats part of a previous study (9) with the
L
1
m lll!l! PS
0
,“, MO!
r -TPI 05 40,’
PIP 0, MO,
PIP
O5 Ma’
-
FIG. 1. Effect of phosphoinositides and diacylglycerol on PKC activity. PKC was purified from rat brains with DE-52, phenyl-Sepharose, and PS-acrylamide column chromatography (11). Activity was assayed by measuring the incorporation of ‘*P from [y-“‘P]ATP into histones in a mixed micellar assay (12). Total reaction volume (0.1 ml) contained 20 mM Tris-Cl, pH 7.5, 0.8 mg/ml histone III-S, 50 pg/ml leupeptin, 12.5 pM ATP, 10 mM MgCl,, 50 pM Ca’+, and 10 ~1 PS, PS/phosphoinositides, or PS/DG lipid micelles. The concentration of lipids in the reaction mixture (with respect to Triton X-100) were: PS, 9 mol% or 445 pM; PI or PIP, 0.1 mol% or 4.95 pM and 0.5 mol% or 22.75 pM; PIP,, 0.05 mol% or 2.47 pM and 0.1 mol% or 4.95 pM; DG, 1 mol% or 49.5 PM and 2 mol% or 99 pM.
PROTEIN
KINASE
C AND
285
PHOSPHOINOSITIDES
100 z 2
PI
80
n i
60
y c9 90
40
a
PIP
20
DG PIP*
01 0 0
7
6
5
4
3
I
I
I
I
0.2
0.4
0.6
0.8
Phosphoinositide
4
,/ 1.0
or DG (mol %)
- Log [Ca2+] FIG. 2. Effect of phosphatidylinositol 4,5-hisphosphate and diacylglycerol on reaction velocity of PKC at various concentrations of Ca*+. Protein kinase C activity was measured as described in Fig. 1. At zero Ca’+ concentrations 4 mM EGTA was present instead of CaC&. Mixed micelles contained Triton X-100 (3% solution); and PS, 9 mol% (A); or PIP,, 0.1 mol% (0); or DG, 2 mol% (Cl).
difference that completely pure PKC (4 nmol/pg protein/ min, a mixture of PKC cy,,L&,&, and y) (3) has here been used, and PKC activity has been assessed at two different concentrations of each effector. There is no difference from the previous results obtained with a partially purified PKC. However, it had been stated (9) that “PI and PIP show little, if any, capacity of activating the kinase.” A closer look at the results in the previous and the present study shows that the activity of PI may in fact be entirely negligible; but PIP does have some effect. This has also been found by others (15). From a comparison of the data of Fig. 1, with both lipids at 0.1 mol%, it appears that PIP has roughly 15% the effect of PIP, on histone phosphorylation by PKC. Calcium dependency. Figure 2 shows that PIP2, like DG, stimulates PKC by shifting the Cazf dose-response curve for activation of the enzyme to a lower Ca2+ concentration. The affinity of Ca2+ for PKC rises by several orders of magnitude in the presence of DG as well as of PIP,; when it is considered that V,,, is 2-3 times greater for DG . PKC than for PIP,. PKC (9), the Ca-activation curves for DG * PKC and PIPZ. PKC nearly coincide. The half-maximal binding constant Kc, is around lop7 for both kinase species. PKC is the phorbol Effect on phorbol ester binding. ester receptor of cells and PIP, displaces phorbol ester competitively from PKC, in the manner of DG (13). The effect of the other phosphoinositides, PI and PIP, on phorbol ester binding to PKC was studied in a mixed micellar assay (14), and compared with that by PIP, or DG (Fig. 3). While no effect of PI on phorbol ester binding was observed, PIP inhibited the specific binding of [“H]PDBu. However, the inhibition was much less pronounced with PIP than with PIP2 or DG. A Scatchard
FIG. 3. Inhibition of phorbolester binding by polyphosphoinositides and DG. Binding of [“H]PDBu binding to PKC was measured with Triton X-100 mixed micelles containing PS (20 mol% relative to Triton, 100 mol%) and variable mol% of PI (A), PIP (O), PIP, (Ml, or DG (01 as described under Materials and Methods. Nonspecific binding was determined in the presence of excess of unlabeled PDBu (10 FM) and subtracted.
analysis of binding-inhibition between 0.01 and 0.5 mol% PIP showed no change in maximal binding (B,,,) but a decrease in the binding affinity for [3H]PDBu with increasing concentration of PIP (Fig. 4a), indicating a competitive nature of inhibition between phorbol ester binding a.nd PIP binding. The plot of K&, the apparent dissociation constant for [3H]PDBu/enzyme in the presence of PIP, was a linear function of PIP concentration (Fig. 4b), with a h, (the dissociation constant for PIP binding) of 0.26 mol%. Under the same conditions the Ki for PIP, was 0.043 (13).
It has been reported by others (15) that none of the inositides competes with phorbol ester for binding to PKC.
:-:n0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
[3H] PDBu Bound, pmol FIG. 4a. Competitive inhibition of [3H]PDBu binding by PIP in mixed micelles. A Scatchard analysis of specific binding of [“H]PDBu at variable concentrations of PDBu (lo-40 nM) as a function of PIP concentration in Triton X-100 mixed micelles containing 20 mol% PS (0). PIP in mol% of Triton X-100 is 0 (O), 0.1 (A), 0.25 (ml, and 0.5 (0).
286
CHAUHAN
8
I
I
-0.4
-0.3
-0.2
ET AL
1
-0.1
0
0.1
0.2
0.3
0.4
0.5
PIP, mol% FIG. 4b. Apparent dissociation constants for [“H]PDBu as a function of PIP concentration. Ki, the apparent dissociation constant for [“H]PDBu/Enzyme in presence of PIP, was determined from Fig. 4a for each PIP concentration by Scatchard analysis. K,, the dissociation constant for PIP, represented by the negative of the X intercept, is 0.26 mol%.
We have no explanation for this discrepancy with our results; it will have to be resolved in future studies. Translocation. The translocation of PKC from aqueous to liposomal phase, with the percentage of kinase in the pellet rising from about 30 to about 60%, is achieved by the lipid effecters DG and PIP, at 1 PM Ca concentration (Fig. 5). Figure 6 shows percentage translocation as a function of effector concentration. DG and PIP, are of comparable activity; PIP is much less effective; and PI hardly effective at all.
o!-rz--F-
8
100
Ca’+[p M] FIG. 5. Effect of calcium on translocation of PKC to multilamellar liposomes. Multilamellar liposomes (10 pmol) containing 20 mol% of PS and 80 mol% total of PC and other specified lipids were prepared as described under Materials and Methods. PKC partially purified by DE-52 and phenyl-Sepharose column chromatography was incubated for 5 min in the presence of 1 Fmol of liposomes, 10 mM MgCl*, different concentrations of CaCl*, and 20 ag leupeptin (final volume 0.2 ml) and centrifuged at 300,OOOg for 20 min. PKC activity in the supernatant was measured by mixed micellar assay. The pellet was suspended in 0.2 ml of buffer A containing 2 mM EDTA, vortexed for 1 min, and centrifuged. The supernatant thus obtained (representing membrane-bound PKC activity) was assayed for PKC activity. The concentration of PIP (o), PIP, (w), or DG (0) was 0.5 mol% and the PC/PS ratio was 79.5/ 20. The control sample (A) had a PC/PS ratio of SO/ZO.
can both compounds activate the kinase in a competitive arrangement? The answer may lie in a circumspect interpretation of the data. Competitive activation does not imply that the effecters must all bind to the same site,
DISCUSSION
of DG - PKC and PIP, * PKC. The present Similarity study shows that PIP2 activates PKC in the manner of DG and phorbol esters by lowering the concentration of the Ca”+ required for PKC-activation, and by translocating the kinase from the soluble to the membrane fraction. There is a difference between the two activated forms in their maximal velocity, (V,,/V,,,,) = 2.5, and in their affinity, (KUG/KpIP2) = 50 (9); but apart from this kinetic variation both activated enzymes behave remarkably similar. They both use phosphatidylserine for an acidic lipid membrane matrix; they have an absolute requirement for Ca’+, i.e., they form a ternary effector . Ca . kinase complex, they both compete with phorbol ester for binding to PKC (13), they seem to have similar substrate specificity (15), and they cause calcium-dependent water-membrane translocation of PKC. Questions arise: how can two structurally greatly divergent molecules interact with the kinase to produce activated forms of nearly the same properties? The traditional PKC-effector, DG, can form three H-bonds with the kinase; phorbol PKC has the same H-bond arrangement, superposable on that formed with DG (16); but PIP2 has only two corresponding H-bonding sites, the fatty ester CO groups; the OH group is missing, replaced by a bulky inositol trisphosphate group. PIP2 cannot possibly bond to the DG receiving site in the same isosteric manner. How, then,
100 80 60
PI
al
c
0 .l 2 .3 .4 .5 .6 .7 .8 .9 1 Phosphoinositide or DG (mot%) FIG. 6. Effect of phosphoinositides or DG on translocation of PKC to multilamellar liposomes. Translocation of PKC was measured as described in the legend to Fig. 3 in the presence of 50 pM Ca2+ and varying concentrations of PI, PIP, PIP,, and DG.
PROTEIN
KINASE
C AND
only that the binding of one prevents the binding of another (13). This could be achieved not only by competition for the same site on the enzyme but possibly also by the binding of the effecters at distinct but overlapping enzyme-effector interfaces (17). Translocation. The phenomenon of the translocation of PKC from soluble to membrane fraction has been studied with intact cells (18-20) and liposomes (21). Both DG and phorbol ester elicit this effect, which also requires Ca’+. Peculiarly, translocation is not complete (U-20). This is confirmed in our experiments (Fig. 6). Even without Ca”+ and effector, ca. 30% of PKC is found in the pellet; with Ca2+ and DG or PIP2, this concentration rises, but not above ca. 70%, except at high, unphysiological concentrations of Ca”+ (100 PM) when a pellet PKC concentration of 90% is reached. Thus, about 40% at most of the kinase is translocated even with the most active effecters. It could be speculated that only one or two of the isozymes present (Q~, &, &, y) are translocated; but the general similarity between these enzymatic species in regard to their activation (3) makes such an explanation unlikely. It has recently been reported (22) that DG activates translocation not only of PKC but also of choline phosphate cytidyltransferase. This suggests that enzyme translocation by DG (and PIPZ?) may be of widespread occurrence. The rapid turnover of the inositide 4- and &phosphates (5-8, 23), together with the results of the present study, shows dependence of PKC activation on the state of phosphorylation of the lipids, and inevitably suggests that the inositide shuttle, PI = PIP 2 PIPZ, at least to some extent regulates the phosphorylation of cellular proteins. Most of the phosphoinositide is in the form of PI, which cannot complex with PKC (no effect on phorbol ester binding or translocation or activation of the kinase is observed). PIP, on the other hand, can displace phorbol ester from PKC, and also cause translocation of the kinase, but cannot activate PKC effectively, suggesting that PIP can complex PKC to the membrane, when a threshold Ca2+ concentration is reached; but this PIP. Ca * PKC complex is an inactive form. The enzyme is turned on upon phosphorylation of PIP to PIP, by PIP-Skinase, and this results in the formation of an active CamPIP,. PKC complex. It cannot, at present, be decided if PIP. PKC or PIP by itself is the full substrate for the PIP-Skinase. Conventionally, the PIPZ-phospholipase C is considered the cascade enzyme to be activated first, and though it has been shown that with support of a G-protein (24) activation can take place at lower than micromolar Ca’+ levels (25), it is still peculiar that an enzyme which initiates Ca” release should itself depend on that cation, directly or indirectly. PIP-Skinase and PI-4-kinase, on the other hand, are independent of calcium. We suggest that the activation of PIP-kinase precedes the activation Inositide
shuttle.
287
PHOSPHOINOSITIDES
of PIP,-phospholipase C. This could increase the pool of PIP,, and lead to increased membrane-PKC binding, or to the opening of Ca-cages (23) and the production of the Ca-spike (26, 27). ACKNOWLEDGMENT This work was supported by funds from the New York State Office of Mental Retardation and Developmental Disabilities.
REFERENCES 1. Brockerhoff, 1766.
H., and Ballou, C. E. (1962) J. Biol. Chem. 237,1764-
2. Streb, H., Irvine, R. F., Berridge, M. J., and Schulz, I. (1983) Nature 306,67-69. 3. Nishizuka,
Y. (1988) Nature 334, 661-665.
4. Exton, cJ.H. (1990) J. Biol. Chem. 265, l-4. 5. Brockerhoff, 52.
H., and Ballou, C. E. (1962) d. Biol. Chem. 237, 49-
6. Garrett, N. E., Garrett, R. J. B., and Talwalkar, Physiol. 87, 63-70.
R. T. (1975) J. Cell.
7. Palmer, F. B. St. C. (1985) Can. J. Biochem. CellBiol. 8. Schneider, R. P., and Kirschner, Actn 202, 283-294. 9. Chauhan, V. P. S., and Brockerhoff, Res. Commun. 155,18-23.
63,927-931.
L. B. (1970) Biochim. H. (1988) B&hem.
Biophys. Biophys.
10. Chauhan, V. P. S., Chauhan, A., Deshmukh, D. S., and Brockerhoff, H. (1990) Life Sci. 47, 981-986. 11. Woodgett, 4843.
J. R., and Hunter,
T. (1987) J. Biol. Chem. 262, 4836-
12. Hannun, Y. A., Loomis, C. R., and Bell, R. M. (1985) J. Biol. Chem. 260, 10,039-10,043. 13. Chauhan, A., Chauhan, V. P. S., Deshmukh, D. S., and Brockerhoff, H. (1989) Biochemistry 28, 4952-4956. 14. Hannun, Y. A., and Bell, R. M. (1986) J. Biol. Chem. 261, 93419347. 15. Lee, M-H., and Bell, R. M. (1990) FASEB J. 4, A1912. 16. Brockerhoff, H. (1986) FEBS Lett. 201, 1-4. 17. Segel, I. H. (1976) in Biochemical Calculations, 2nd ed., pp. 246, Wiley, New York. 18. Sacktor, T. C., and Schwartz, J. H. (1990) Proc. N&l. Acad. 5%. USA 87, 2036-2039. 19. Persaud, S. J., Jones, P. M., Sugden, D., and Howell, S. L. (1989) FEBS Lett. 245, 80-84. 20. Vaartjes, W. J., de Haas, C. G. M., and van den Bergh, S. G. (1986) Biochem. Biophys. Res. Commun. 138,1328-1333. 21. Bazzi, M. D., and Nelsestuen, G. L. (1987) Biochemistry 26, 115122. 22. Kolesnick, R. N., and Hemer, M. R. (1990) J. Biol. Chem. 265, 10,900-10,904. 23. Brockerhoff, H. (1986) Chem. Phys. Lipids 39, 83-92. 24. Deckmyn, H., Shi-Ming, T., and Majerus, P. W. (1986) J. Biol. Chem. 261, 16,553-16,558. 25. Rana, R. S., and Hokin, L. E. (1990) Physiol. Reu. 70, 115-164. 26. Albert, P. R., and Tashjian, Jr., A. H. (1984) J. Biol. Chem. 259, 5827-5832. 27. Tashijian, Jr., A. H., Heslop, J. P., and Berridge, M. J. (1987) J. Riochem. 243, 305-308.
