A Novel Mechanism of Diglyceride Formation STIMULATES THE CYCLIC BREAKDOWN AND RESYNTHESIS OF PHOSPHATIDYLCHOLINE*

12-O-Tetradecanoylphorbol-13-acetate (TPA) treat- ment of Madin-Darby canine kidney cells resulted in an increased incorporation of ”Pi and [methyl-’Hlcho- line into choline-containing phosphoglycerides (PC). In pulse-chase experiments, TPA treatment caused an increased release of [methyZ-SH]choline from the PC fraction of prelabeled cells. When cells were prelabeled with [‘Hlarachidonic acid and [14C]palmitic acid, TPA treatment resulted in an increased synthesis of 14C, ‘H- diglycerides. Further studies were done to determine the relationship between PC breakdown and diglycer- ide synthesis. Cells were preincubated with ether-linked l-O-[’H]hexadecyl-2-lyso-sn-glycero-3-phos- phocholine which was acylated to form l-O-[’Hlhex-adecyl-2-acyl-sn-glycero-3-phosphocholine. Subse- quent treatment of these cells with TPA resulted in an increased synthesis of l-O-[3H]hexadecyl-2-acyl-sn- glycerol compared to cells not stimulated with TPA. These findings demonstrate that TPA stimulates PC turnover in Madin-Darby canine kidney cells and pro-vide evidence for a novel mechanism of diglyceride

may activate a phospholipase C or phospholipase D. However, the lipid products of these possible pathways (diglyceride or phosphatidic acid) were not identified. Grove and Schimmel (11) demonstrated that TPA causes increased diacylglycerol levels in chick embryo myoblasts and suggested PC may be the source of the diacylglycerol, since the acyl composition of the diacylglycerol is similar to that of the cellular PC.
Recent studies have shown that TPA stimulates a Ca2+activated, phospholipid-dependent protein kinase (protein kinase C) which is stimulated by diacylglycerol (12, 13). Thus, it has been suggested that TPA exerts its biological effects by mimicking the actions of diacylglycerol which is generated by phospholipase C degradation of phosphatidylinositol (14). Because of the increasingly appreciated importance of diglycerides as regulatory molecules, we have examined the effects of TPA on diglyceride synthesis from PC.
The MDCK cell line was chosen for these studies because it is highly sensitive to stimulation by TPA (2) and has been used in many earlier studies of the effects of TPA on lipid metabolism (1-4). Alkylacylglycerides, in addition to diacylglycerides, were investigated as a model system, because recent studies demonstrate that, although the alkylacylglycerides do not stimulate protein kinase C (15), they, like TPA, stimulate HL-60 promyelocytic leukemia cells to differentiate to macrophage-like cells (16).
Cell Culture-MDCK cells were cultured in Dulbecco's modification of Eagle's medium supplemented with 10% (v/v) fetal bovine serum, 100 units of penicillin/ml, 100 pg of streptomycin/ml, 0.22% Na2HC03, and 2 mM L-glutamine. The cells were maintained in 75cm* flasks and were subcultured at 3-5-day intervals at a ratio of 15.
For prelabeling experiments, the cells were plated in 35-mm dishes at 5 x IO6 cells/plate and allowed to attach to the plates (2-3 h) before the medium was replaced with fresh medium containing the radiolabeled compounds.
Lipid Extraction and Anulysis-The cells were harvested by scraping directly from the cell culture dishes into methanol. The lipids were then extracted by a modification of the method of Bligh and Dyer (18) as described previously (1). After extraction, the lipids were dried under a gentle stream of N2 and resuspended in chloroform. Neutral lipids were separated on layers of Silica Gel 60 developed in a solvent system consisting of hexane/ethyl ether/formic acid (90:60:4, v/v) or a solvent system consisting of the organic phase of ethyl acetate/isooctane/acetic acid/water (55:25:1050, v/v). Diglycerides were identified by their migration with known standards in both systems (R, values of 0.33 and 0.93, respectively). Phospholipids were separated by TLC on a layer of Silica Gel 60 developed in a solvent system consisting of chloroform/methanol/acetic acid/water (75:4812.5:4.5, v/v) or (50:25:84, v/v). Following chromatography, the lipids were located by fluorography (3H and "C) using ENHANCE spray (New England Nuclear) or by autoradiography (32P), both using Kodak SB-5 film. The radiolabeled lipids were then scraped from the plates and quantitated by scintillation counting using a Packard liquid scintillation counter.
Water-soluble products of [3H]choline were determined by evaporating the aqueous phase of the Bligh and Dyer lipid extraction mixture under reduced pressure, redissolving the extract in water, and separating the products by TLC in a solvent system consisting of H,O with 0.9% NaC1/CH30H/NH40H (50:50:5, v/v) as described by Vance et al. (19).
The content of radiolabeled plasmalogens was determined by twodimensional TLC as previously described by Daniel et al. (1). The products obtained after incubation of l-O-[3H]hexadecyl-2-lyso-GPC were resolved on Silica Gel 60 plates developed in chloroform/methanol/acetic acid/water (5025:84, v/v), then the alk-1-enyl groups of the plasmalogens were cleaved by exposure to HCI fumes. The resulting products were next separated in a second dimension using the same solvent system. The radiolabeled products were located by fluorography and identified by their migration in relation to known phospholipid standards run in both dimensions. The radiolabeled products were then scraped from the plate and quantitated by liquid scintillation counting. The products obtained after incubation of 1-O-[3H]hexadecyl-2-lyso-GPC with MDCK cells were further characterized as follows. The PC and PE fractions were isolated by TLC as described and extracted from the silica gel by the method of Bligh and Dyer (18). The phospholipids were then converted to diradylglycerols by phospholipase C treatment and converted to benzoate derivatives as described by Blank et al. (20). The reaction mixture was applied to small Silicar columns and the radiolabeled benzoyl diradylglycerols were quantitatively recovered by eluting the columns with >5 column volumes of chloroform. This procedure removed the majority of unreacted benzoic anhydride which interfered with further analysis of the products. The 3H-labeled benzoyl diradylglycerols were then separated into three subclasses (l-O-alk-l-enyl-2-acyl-, 1-0alkyl-2-acyl-, and 1,2-diacyl-) by TLC using Silica Gel 60 plates developed in a solvent system consisting of benzene/hexane/diethyl ether (50:45:4, v/v). The radiolabeled products were then located by fluorography and identified by their migration with known standards prepared as described by Blank et al. (20). The radiolabeled products were scraped from the plates and quantitated by liquid scintillation counting. In this experiment and others using medium containing fetal calf serum, the concentration of TPA was increased to 100 nM, because our previous studies found that serum bound TPA, requiring higher concentrations of TPA for use for stimulation (21,22). In these cases TPA treatment of the cells caused an increased incorporation of [3H]choline into PC, but not into sphingomyelin. We found that TPA stimulation of the prelabeled cells resulted in the synthesis of diglycerides containing both 3H and 14C (Fig. 1). The time course for the release of both 3H and 14C was similar and reached a maximum at 3-4 h. The subsequent decrease in labeled diglyceride probably results from conversion into PC or phosphatidic acid. Although these experiments demonstrate that TPA causes an increased synthesis of diglycerides, they do not prove that the precursor is PC.

Release of
In the next studies, we introduced radiolabeled 1-0- labeled products recovered f S. D. n = 6). The products were then further characterized. The product tentatively identified as PE was shown to co-migrate with authentic PE in a basic solvent system consisting of chloroform/methanol/ammonium hydroxide (65:35:8, v/v). The products were analyzed by two-dimensional TLC and HCl exposure to determine the content of 1-alk-1'-enyl-linked species (see "Experimental Procedures"). We found that 82.4% f 7.6% ( n = 3) of the radiolabel in P E was in an acid-labile species indicating that the PE was predominantly PE-plasmalogen. The major portion (>95%) of the product identified as PC was resistant to acid hydrolysis, indicating that the PC species were predominantly either 1-alkyl-2-acyl-linked or 1, 2-diacyl-linked. We further characterized the PE and PC products by treating the isolated phospholipids with phospholipase C and converting the resulting diglycerides to benzoyl derivatives as described under "Experimental Procedures." The labeled benzoyl diglycerides were then separated by TLC and their migration was compared t o the migration of known standards. This procedure confirmed that the labeled PE was predominantly l-alk-l'-enyl-2-acyl-sn-glycero-3-phosphoethanolamine and that the labeled PC was predominantly 1-alkyl-2-acyl-GPC. Both compounds had (3% of the total radiolabel in diacyl species.

The cells thus labeled with l-O-[3H]hexadecyl-2-lyso-GPC
were then stimulated with TPA and the radiolabeled products were analyzed by TLC with two different solvent systems.
One solvent system was used to resolve diglycerides, phosphatidic acid, and triglycerides. Using this technique we found that TPA treatment caused an increase in 3H-diglyceride formation (Fig. 2). There was also an increase in 3H-labeled phosphatidic acid and triglycerides (Fig. 2); however, the lag in formation of the latter products and the lower amounts produced indicate that they are secondary to formation of diglycerides.
The products were also separated by TLC using a solvent system consisting of chloroform/methanol/acetic acid/water (50:25:8:4, v/v) to resolve the PC and PE fractions. These studies revealed that TPA stimulated a loss of radiolabel from the PC fraction (Fig. 3); there was no decrease in radiolabel in the PE fraction. In fact, there appeared to be a small increase in radiolabel in PE (Fig. 3). This increase was observed in the control cultures and was not significantly different in the TPA-treated cultures. The mechanism of this transfer is unknown; however, a possible mechanism includes base exchange and desaturation of the alkyl chain of the PE to yield plasmalogen PE.

DISCUSSION
The present results indicate that TPA stimulates the degradation of PC to yield phosphocholine and diglycerides. In  addition, there is a stimulated incorporation of both 32Pi and [3H]choline into PC. Therefore, TPA appears to stimulate a cyclical degradation and resynthesis of PC analogous to the well known "phosphatidylinositol cycle" (24). The first step in this "PC cycle" (Fig. 4) requires the action of a PC-specific phospholipase C. However, no such enzyme was known in mammalian systems until a recent report by Wolf and Gross (25). They described a partially purified phospholipase C from canine myocardium which is active, utilizing either plasmalenylcholine or phosphatidylcholine as substrate. This enzyme does not utilize phosphatidylinositol as a substrate. The myocardial cells were also found to contain an inhibitor of the phospholipase that was removed by partial purification. This inhibitor may account for previous failures to observe the enzymatic activity.
The second step in the proposed cycle is catalyzed by CDPcholine:1,2-diacylglycerol cholinephosphotransferase; however, the rate-controlling step in PC synthesis by the CDPcholine pathway is catalyzed by CTP:choline-phosphate cytidylyltransferase (26). The activity of CTP:choline-phosphate cytidylyltransferase has previously been shown to be stimulated by TPA (5).
We earlier observed the formation of a product that appeared to be identical with 1-0-[3H]alkyl-2-acylglycerol in human neutrophils stimulated with ionophore A23187 (23). The neutrophils were prelabeled with 1-0-[3H]hexadecyl-2lyso-GPC as in the present study. In the present studies, the release of phosphocholine and the preponderance of diglyceride, rather than phosphatidic acid, at early time points supports a phospholipase C reaction rather than a pathway mediated by phospholipase D and a phosphohydrolase. However, it is difficult to rule out the possibility that a reversal of the cholinephosphotransferase reaction (27) may be responsible for our observations. Both 1-acyl-linked and 1-alkyl-linked diglycerides are formed in TPA-treated MDCK cells. The diacylglycerides, in particular l-oleoyl-2-acetyl-sn-glycerol, have been shown to stimulate protein kinase C (14) and mimic the actions of TPA in intact platelets (28) and mouse epidermal cells (29). However, this diacylglycerol does not mimic the effect of TPA in HL-60 cells, although it was reported to stimulate protein kinase C (30). In contrast, the 1-alkyl-linked glyceride, 1-0hexadecyl-2-acetyl-sn-glycerol, does not stimulate protein kinase C (15) but, like TPA, induces differentiation in HL-60 cells (16). Thus, a portion of the effects of TPA may require the generation of endogenous 1-alkyl-linked glycerides and appears to be independent of protein kinase C. The stimulated degradation of PC described herein results in transient increases in 1-alkyl-linked glycerides and these compounds may mediate bioactivities of TPA which do not involve protein kinase C.