Increased Concentrations of Phosphatidate, Diacylglycerol and Ceramide in Ras- and Tyrosine Kinase (Fps)-Transformed ®Broblasts

Home , Ceramide, Diacylglycerol kinase

Increased concentrations of phosphatidate, diacylglycerol and ceramide in ras- and tyrosine kinase (fps)-transformed ®broblasts

Ashley Martin1,3, Patricia A Duy1,4, Christos Liossis1, Antonio Gomez-MunÄ oz1,5, Lori O'Brien1, James C Stone and David N Brindley1,2

1Signal Transduction Laboratories Lipid and Lipoprotein Research Group; 2Department of Biochemistry, 357 Heritage Medical Research Centre, University of Alberta, Edmonton, Alberta, T6G 2S2 Canada

Concentrations of the bioactive lipids, phosphatidate and cell division in certain circumstances. However, the full diacylglycerol, increased with time in culture in ras- and signi®cance of altered lipid pro®les in transformed cells tyrosine kinase (fps)-transformed ®broblasts but not in is not yet understood. control ®broblasts. On Day 3, diacylglycerol and Increased signaling by DAG in ras-transformed cells phosphatidate concentrations were about 3.3- and 5.5- could arise from increased DAG synthesis de novo fold higher respectively in the ras-transformed compared (Chiarugi et al., 1989), or increased phospholipid to control ®broblasts. These concentrations in fps- turnover. The latter possibility is supported by the transformed ®broblasts were increased about twofold. observation that injection of NIH3T3 ®broblasts with The changes in phosphatidate and diacylglycerol resulted antibodies that neutralize PLC-g blocks DNA synthesis from enhanced phospholipid turnover rather than from induced by co-injected p21ras (Smith et al., 1990). synthesis de novo. The increased ratio of phosphatidate Alternatively, it has been proposed that phosphatidyl- to diacylglycerol is explained by decreased activities of choline-(PC) speci®c phospholipase C (PLC) is two distinct phosphatidate phosphohydrolases and in- activated by p21ras (Lopez-Barahona et al., 1990; creased diacylglycerol kinase in ras-transformed ®bro- Diaz-Meco et al., 1991; Dominguez et al., 1991) and blasts. Ceramide concentrations were about 2.5- and increased phosphorylcholine concentrations in ras- threefold higher in the fps- and ras-transformed cells transformed cells are compatible with this hypothesis respectively on Day 3 compared to the controls. (Larrodera et al., 1990; Price et al., 1989). However, Incubating control ®broblasts from Days 1 to 3 with there is controversy as to the importance of PC-speci®c phosphatidylcholine-speci®c phospholipase C increased PLC in mammalian cells. Increased DAG concentra- diacylglycerol, phosphatidate and ceramide concentration in ras-transformed cells could result from tions, and decreased Mg2+-independent-phosphatidate activation of PC-phospholipase D (PLD) and subse- phosphohydrolase activity. 8-(4-chlorophenylthio)-cAMP quent conversion of phosphatidate (PA) to DAG by had a cytostatic eect in ras-transformed cells, it phosphatidate phosphohydrolase (PAP). Micro-injec- decreased the concentrations of phosphatidate and tion of Xenopus laevis oocytes with p21ras caused rapid diacylglycerol, but increased that of ceramide. The DAG production that was blocked by propranolol, an consequences of increased ceramide and phosphatidate inhibitor of PAP (Carnero et al., 1994a). Choline concentrations in ras-transformed cells are discussed in kinase activity is elevated in ras-transformed cells relation to signal transduction, cell division and the (Macara, 1989; Lacal, 1990; Teegarden et al., 1990), transformed phenotype. and CTP-phosphocholine cytidylyltransferase is lower (Teegarden et al., 1990). These observations could Keywords: ceramides; diacylglycerol kinase; phosphati- explain the increase in phosphorylcholine concentra- date phosphohydrolase; phospholipase D tions without the need for increased PC-PLC activity. Constitutive activation of basal PLD activity and the relative labeling of PA has been demonstrated in ras- The GTPase encoded by ras is a key component of transformed ®broblasts (Martin et al., 1993; Carnero et signal transduction involving growth factors. Somatic al., 1994b). This result is compatible with the activation mutations that constitutively activate p21ras occur of PLD by p21ras (Carnero et al., 1994a), which frequently in diverse human tumors (Bos, 1989; probably involves pp60src (Jiang et al., 1995a) and Wakelam et al., 1991; Satoh et al., 1992). Studies on Ral (Jiang et al., 1995b). In contrast, the activation of the metabolism of bioactive lipids in ras-transformed PLD by serum, platelet-derived growth factor, and cells have intensi®ed since the discovery of the elevated phorbol ester is attenuated in ras-transformed cells concentrations of diacylglycerol (DAG) in this model (Huang and Cabot, 1992; Fu et al., 1992; Martin et al., (Preiss et al., 1986). This lipid is important in 1993). This eect was attributed to the inability to activating of protein kinase C, which can stimulate activate PLD downstream of PLC-g (Alam et al., 1995). Martin et al. (1993) showed that the speci®c activity of plasma membrane PAP-2 was decreased in Correspondence: DN Brindley ras- and fps-transformed ®broblasts which produced Present address: 3Department of Clinical Oncology, Clinical Research more PA than DAG when stimulated with serum or Block, Queen Elizabeth Hospital, Birmingham, B15, UK; 4Group phorbol ester. 1800, ART Unit 1802, Crystal Mall One, Suite 11A16, Arlington, This latter result is signi®cant since PA is a putative 5 Virginia, 22202 USA; CSIC, Instituto de Investigaciones second messenger and agonist (Martin et al., 1994). PA Biomedicas, Arturo Duperier 4, 28029, Madrid, Spain Received 14 August 1996; revised 8 November 1996; accepted 28 stimulates superoxide generation in neutrophils (Rossi November 1996 et al., 1990; Bauldry et al., 1991), and increases the Phosphatidate and ceramide in transformed fibroblasts A Martin et al 1572 activities of cytosolic protein kinase(s) (Bocckino et al., which inhibit and stimulate cell division respectively, 1991), protein kinase C-z (Limotola et al., 1994), were also studied to see if these agents modify the phosphatidylinositol-4-phosphate kinase (Moritz et al., concentrations of PA, DAG, and ceramide. 1992), and PLC-g (Jones and Carpenter, 1993). Exogenous PA and lyso-PA inhibit adenylate cyclase (Murayama and Ui, 1987; van Corven et al., 1989; Results GoÂ mez-MunÄ oz et al., 1994) and increase arachidonate concentrations (Moolenaar, 1995), tyrosine kinase Eects of time in culture on lipid concentrations in activities (Moolenaar, 1995) and mitogen-activated control, ras- and fps-transformed ®broblasts protein (MAP) kinase activities in a variety of cells (Howe and Marshall, 1993; Moolenaar, 1995; GoÂ mez- The growth pro®le, in terms of DNA concentrations, MunÄ oz et al., 1995). Exogenous PA and lyso-PA also of control and transformed ®broblasts in culture is activate PLD (GoÂ mez-MunÄ oz et al., 1994; Moolenaar, illustrated in Figure 1 and this resembles the protein 1995) such that external PA is able to generate pro®le that was published previously (Martin et al., intracellular PA. Furthermore, activation of protein 1993). Control cells essentially reached con¯uence and kinase C by DAG also stimulates PLD activity and PA became growth arrested by Day 2 in culture whereas production (Eldar et al., 1993; Martin et al., 1994). In the ras- and fps-transformed ®broblasts continued to cell-free system PA inhibits the rasGTPase activating divide. By Day 4 the protein and DNA contents of the protein, and activates a protein that decreases the dishes were three to four times higher for the GTPase activity of p21ras (Tsai et al., 1989a,b; 1990). transformed cell lines compared to the controls. In all Thus, changes in lipid pro®les might modify p21ras of the present experiments these results were con®rmed signaling in addition to being in¯uenced by p21ras. by measuring either the protein or phospholipid These interactions might be important in cells content of the dishes and the results are expressed transformed with tyrosine kinases which are thought relative to these parameters, as appropriate. to activate p21ras chronically by regulating exchange Fibroblasts were preincubated for 2 h with factors in addition to having Ras-independent signaling [3H]myristate to label phospholipids, primarily PC. mechanisms. The subsequent redistribution of 3H in PA, DAG and GTP-bound p21ras binds directly to the serine/ triacylglycerol (TAG), relative to that in PC, was threonine protein kinase, p74raf (Leevers et al., 1994) determined under basal conditions to determine and recruits it to plasma membranes where Raf is dierences in phospholipid turnover between control activated. This attachment of Raf is facilitated by its and transformed cells. The incorporation of binding to PA and this process is stimulated by PLD [3H]myristate into PA was low on Day 1 and was not activation (Ghosh et al., 1996). Raf phosphorylates signi®cantly dierent among the cell types (Figure 2a). MEK which activates MAP kinase (Guan et al., 1994). The accumulation of [3H]PA increased in all cell lines This enzyme can activate numerous cellular targets when they were labeled after being cultured for 2 and 3

including phospholipase A2 (Lin et al., 1993) as well as transcription factors (Guan et al., 1994). p21ras also binds other targets such as phosphatidylinositol-3 kinase (Rodriguez-Viciana et al., 1994). Thus p21ras signaling might alter phospholipid metabolism in several ways: by interacting directly with phosphatidylinositol-3 kinase, by activating other phospholipid metabolizing enzymes via the kinase cascade, or by altering the transcription of genes which encode phospholipid metabolizing enzymes. In light of the importance of DAG and PA in signal transduction, this paper is concerned with changes in their relative masses in control, ras- and fps- transformed ®broblasts. We predicted that the transformed cells would contain high concentrations of PA because of the low Mg2+-independent PAP-2 activity. Furthermore, we hypothesized that the decreased agonist-induced stimulation of PLD might result from increased ceramide concentrations which can prevent PLD activation (GoÂ mez-MunÄ oz et al., 1994, 1995b; Venable et al., 1994; Nakamura et al., 1995). A major contribution of the present work is the demonstration that PA and DAG concentrations increase dramatically with time in culture in ras- transformed ®broblasts. These increases in PA and Figure 1 Eect of time in culture on the DNA content in DAG concentrations result primarily from increased cultures of control, ras- and fps-transformed ®broblasts. Cells phospholipid turnover rather than from synthesis de were maintained in culture for the number of days indicated when novo. The increased ratio of PA to DAG arises from they were washed and harvested. DNA content in control (&), increased DAG kinase activities and decreased activity ras-transformed (~) and fps-transformed ®broblasts (*) were determined. Results are from a representative experiment and of both the Mg2+-dependent and -independent PAP con®rm the growth pro®le expected from previous measurements activities. CPT-cAMP, and an exogenous PC-PLC, of protein content Phosphatidate and ceramide in transformed fibroblasts AMartinet al 1573 days. The increased accumulation of [3H]PA in the ras- lines compared to controls on all 3 days. These transformed cells became apparent after 2 days in increases for Day 3 were about six- and fourfold culture and on Day 3 was about tenfold higher than respectively for ras- and fps-transformed ®broblasts. for control ®broblasts. The equivalent increase was We also determined the mass of lipids present in only about twofold for the fps-transformed ®broblasts. the cells on the dierent days in culture. In the The relative labeling of DAG was not signi®cantly control cells, the mass of PA fell by about 28% dierent in control, ras-, and fps-transformed fibro- between Days 1 to 3 (Table 1). PA concentrations of blasts after 1 day in culture. On Day 3 there were in four ras-transformed lines was higher than in the about ®ve- and 2.5-fold increases in DAG labeling in control ®broblasts on Day 1. This concentration ras- and fps-transformed cells respectively compared to increased with time in culture such that on Day 3 control ®broblasts. Relative [3H]myristate incorpora- the PA concentration was ®ve- to sixfold higher than tion into TAG was higher in both transformed cell in control ®broblasts. In two fps-transformed lines, PA concentrations were similar to those in the controls on Day 1. There was a slight increase in fps-transformed cells on Days 2 and 3 and the concentrations were 1.7- and 1.8-fold higher than in controls. DAG concentrations in control and transformed cells were similar on Day 1. In control cells DAG concentrations remained relatively constant, but they increased in transformed lines with time in culture. On Day 3 DAG was three- to 3.6-fold higher in ras-transformed, and 1.8- to 2.4-fold higher in fps-transformed cells as compared to controls. Ceramide concentrations on Day 1 in culture were consistently higher in ras-transformed ®broblasts compared to controls and fps-transformed ®broblasts were intermediate. Ceramide concentrations remained higher in transformed cells with increasing time in culture. PC, phosphatidylethanolamine (PE), phosphatidylinositol and phosphatidylserine constituted about 50%, 30%, 8 ± 9% and 8 ± 9%, respectively, of the total glycerophospholipids in control and transformed cells and these values did not change signi®cantly with time in culture (results not shown).

Incorporation of [14C]glucose into lipids and activities of PAP-1 and DAG kinase in control and transformed ®broblasts The contribution of glycerolipid synthesis de novo to changes in PA and DAG concentrations was determined in ®broblasts cultured for 3 days. Initial experiments established that [14C]glucose incorporation was proportional to time over the 4 min incubation. The major incorporation was into PA and DAG. There was no signi®cant dierence in d.p.m. incorporated into total lipids in control and transformed ®broblast in four independent experiments (Table 2). However, the ratio of d.p.m. in DAG : PA was 1.03, 0.56 and 0.69 in control, ras- and fps-transformed ®broblasts respectively (Table 2). This was a consistent ®nding in all four experiments indicating a lower activity of PAP- 1 (which is involved in glycerolipid synthesis) in the transformed cells. Furthermore, the estimated ¯ux through PAP (the sum of d.p.m. in DAG, TAG, PC and PE) was consistently lower in transformed cells compared to controls. Total PAP-1 activity and Figure 2 Eect of time in culture on the incoporation of membrane-associated activity (which is the physiologi- 3 [ H]myristate into phosphatidate, diacylglycerol and cally active form) was higher on Day 3 in control triacylglycerol of control, ras- and fps-transformed ®broblasts. Fibroblasts were cultured for the time indicated and then relative to ras-transformed cells (Figure 3). This incubated for 2 h with [3H]myristate. Cells were washed with observation appeared to result from increased PAP-1 0.2% albumin solution and harvested 2 h later. [3H]myristate that activity between Days 1 and 3 in control cells was incorporated into PA, DAG and triacylglycerol was compared to no change in ras-transformed fibro- calculated relative to the corresponding incorporation into PC. blasts. The speci®c activity of dioctanoylglycerol The results are means+s.e. from three independent experiments. The labeling of PC constituted 85% of that incorporated into kinase was similar in control, ras- and fps-transformed total phospholipids in the control ®broblasts on Days 1 ± 3 ®broblasts after Day 1 of culture (Figure 4). On Day 3 Phosphatidate and ceramide in transformed fibroblasts A Martin et al 1574 Table 1 Effect of time in culture on the concentrations of phosphatidate, diacylglycerol and ceramide in control, ras- and fps-transformed fibroblasts Type of pmol of lipid/nmol of phospholipid Lipid fibroblasta Day 1 Day 2 Day 3 Phosphatidate Control 5.44+0.51b 4.96+1.76 3.91+1.4 ras 1 8.11+1.6 15.5+2.0 22.1+3.6 ras 2 12.8+0.21 not determined 24.2+1.6 ras 3 9.90+1.4 15.5+2.7 23.4+2.9 ras 4 8.42+1.2 15.6+2.4 19.6+3.5 fps 1 4.62+0.31 6.82+2.0 6.66+0.16 fps 2 5.39+1.2 7.39+3.7 7.22+2.1 Diacylglycerol Control 6.37+3.39 8.46+3.49 7.17+2.32 ras 1 6.11+2.68 12.1+5.19 23.2+7.04 ras 3 5.93+1.96 12.1+4.10 18.4+4.87 ras 4 5.49+2.04 14.7+4.56 24.5+8.01 fps 1 5.35+2.16 11.6+4.67 15.4+3.92 fps 2 7.26+3.19 12.2+7.22 15.6+6.54 Ceramide Control 3.68+0.30 3.19+0.80 2.29+0.82 ras 1 7.75+1.75 8.84+2.10 6.55+0.51 ras 3 9.24+1.31 10.2+2.47 6.65+1.26 ras 4 6.76+2.25 8.64+1.96 6.50+2.88 fps 1 5.38+1.68 6.90+2.05 5.90+0.91 fps 2 6.44+2.26 6.12+2.70 5.74+1.06 aThe control (R2) ®broblasts, four dierent ras clones and two fps clones were harvested at the time in culture that is indicated. The concentrations of the lipids are expressed relative to the total phospholipid of the cells. bMeans+s.e. for three independent experiments for PA and four for DAG and ceramide

Table 2 Incorporation of [U-14C]glucose into lipids by control, ras- and fps-transformed fibroblasts d.p.m. per mg of protein Lipidsa Control ras fps PA 11 175b 10 999 6 418 DAG 11 463 6 175 4 430 Triacylglycerol 2 185 1 229 1 223 Phosphatidylcholine 1 391 973 1 363 Flux through PAPc 16 176 9 253 7 578 aFibroblasts were cultured for 3 days and then incubated with 5 mCi/ ml of [14C]glucose for 4 min. bThe incorporation of glucose into individual lipids is given for a single experiment performed with duplicate dishes. The results were con®rmed in three further individual experiments. cThe estimated ¯ux through PAP was calculated as the sum of 14C in DAG, triacylglycerol, PC and phosphatidylethanolamine

and fps-transformed ®broblasts were 4.2- and 4.7-fold respectively. Therefore on Day 3 ras- and fps- transformed ®broblasts had about 1.7- and 1.8-fold greater DAG kinase activity than control cells. Similar results were obtained with stearoylarachidonoylglycerol as substrate (results not shown).

Eects of PC-PLC on the concentration of PA, DAG and ceramide, and on PAP-2 activity in control ®broblasts Figure 3 Eect of time in culture on the activity of the Mg2+- dependent PAP-1 in control and ras-transformed ®broblasts. Non-transformed rat 2 ®broblasts were treated with 10 Fibroblasts were cultured for the time indicated and then lysed units/ml PC-PLC from Days 1 ± 3 as a pharmacological with digitonin to separate the cytosol (black bars) from the tool to determine if an increased formation of DAG in membrane fraction (hatched bars). Total PAP-1 activity is control cells could mimic other aspects of the lipid indicated by the open bars. The speci®c activities of the N- ethylmaleimide sensitive and Mg2+-dependent PAP-1 is expressed pro®le of ras-transformed cells. Treatment of rat 2 relative to the total protein content of the ®broblasts. The results ®broblasts with PC-PLC increased DAG, PA and are mean+s.d. from three independent experiments ceramide concentrations by about 17.5-, 2- and 8.4-fold in two independent experiments (Figure 5). It is unlikely that the increased ceramide resulted predomi- the speci®c activity of DAG kinase in three nantly from contamination of PLC with sphingomye- independent experiments increased by an average of linase, since according to the Manufacturer, 2.7-fold in control cells. Equivalent increases for ras- sphingomyelinase activity was 50.05% of that of Phosphatidate and ceramide in transformed fibroblasts AMartinet al 1575 PLC. To check this we also incubated control ®broblasts with 5 m Units of sphingomyelinase/ml of medium from Days 1 to 3. Ceramide concentrations were increased by 2.6-fold and DAG concentrations were not signi®cantly changed. Therefore, contamination of PC-PLC with sphingomyelinase is unlikely to explain the 8.4-fold increase in ceramide (Figure 5). Treatment with PC-PLC also decreased the speci®c activities of Mg2+-independent PAP-2A and PAP-2B by about 23% (Table 3).

Eects of incubating ras-transformed ®broblasts with 8-(4-chlorophenylthio)cAMP (CPT-cAMP) on the concentrations of PA, DAG and ceramide Incubation of some ras-transformed cells with cAMP analogues for 2 days inhibits cell division and restores a more normal phenotype (Tagliaferri et al., 1988; Wakelam et al., 1991; Paterson et al., 1991). We Figure 4 Eect of time in culture on the activities of therefore used CPT-cAMP as a cytostatic agent to diacylglycerol kinase in control, ras- and fps-transformed study the changes in lipid concentrations in ras- ®broblasts. Control (open bars), ras-transformed (hatched bars) transformed cells. The phospholipid content per dish and fps-transformed (black bars) ®broblasts were cultured for the of cells was on average 59% lower when transformed time indicated and the speci®c activity of DAG kinase was ®broblasts were treated with CPT-cAMP for 2 days measured in cell sonicates by using dioctanoylglycerol. Results are from a representative experiment that was con®rmed in two (results not shown) which con®rms the decrease in cell further experiments division. PA and DAG concentrations were decreased by averages of 67% and 63% respectively in two experiments (Figure 6a). Therefore, the ratio of PA and DAG concentrations was not aected by CPTcAMP, and neither were the speci®c activities of PAP-2A and PAP-2B (results not shown). Treatment with CPT- cAMP increased ceramide concentrations by about 1.5- fold. We also measured the labeling of PA, DAG and TAG in ras-transformed cells treated for 2 days with CPTcAMP and then incubated with [3H]myristate for 2 h. There was about a 75% decrease in the relative labeling of PA, DAG and TAG (Figure 6b). However, no signi®cant decreases in PLD activity was obtained under basal incubation conditions which included 10% serum (results not shown).

Discussion

Previously we showed that PAP-2 activity, which converts PA to DAG in plasma membranes, is lower Figure 5 Eects of PC-PLC on the concentration of phospha- in ras- and fps-transformed ®broblasts than in control tidate, diacylglycerol and ceramide in control ®broblasts. Control cells over 4 days of culture (Martin et al., 1993). The ®broblasts were incubated from Days 1 to 3 with 10 units of PC- possible functional signi®cance of this lower PAP PLC/ml. The concentrations of the lipids are expressed relative to the protein content of the ®broblasts. The results were con®rmed activity was con®rmed by the higher ratio of PA to in a further independent experiment DAG that was generated by the transformed cells after stimulation with serum, or phorbol ester. The present work is concerned with dierences in the steady-state masses of bioactive lipids, rather than acute responses Table 3 Effect of incubating control (R2) fibroblasts with PC-PLC to agonist stimulation. The present work shows on the activity of PAP-2 substantial increases in PA concentrations in ras- Incubation nmol of DAG formed/min per mg protein transformed ®broblasts with time in culture. Changes conditiona PAP-2Ab PAP-2B in key enzymes that could regulate the relative Control 14.4+1.7c 15.0+1.5 concentrations of DAG and PA were determined to +PLC 11.2+1.9 11.6+1.5 provide an explanation for the dierences. P50.044 The ®ve- to sixfold increases in PA concentrations aControl ®broblasts were incubated from Days 1 to 3 with 10 units/ were detected in four independent ras-clones and these b ml of PC-PLC as described in Figure 4. PAP-2 activity was increases were about twice those for DAG and determined with liposomes of PA (PAP-2A), or with micelles of PA and PC with Triton X-100 (PAP-2B). cMeans+s.e. from three ceramide. The dierences in lipid concentration in the independent experiments. The signi®cance of the dierence compared ras- and two fps-clones were therefore not biased by to the appropriate control was calculated by using a paired t-test clonal variation. There were no signi®cant dierences Phosphatidate and ceramide in transformed fibroblasts A Martin et al 1576 blasts. Tyrosine kinase oncogenes activate products of endogenous cellular ras proto-oncogenes (Smith et al., 1986), and it is not surprising that fps-transformed ®broblasts resemble ras-transformed cells. The intermediate concentrations of lipids in the fps-transformation (Table 1) may mean that fps does not activate normal p21ras to levels obtained with v-ras in this model. Tyrosine kinases may also alter lipid pro®les through Ras-independent mechanisms. Figure 2 demonstrates gross time-dependent differ- ences in PA and DAG metabolism among control and transformed ®broblasts. These results con®rmed the mass determinations in Table 1, although this measurement depends upon the labeling of the lipids for 2 h with [3H]myristate followed by a 2 h chase. Myristate is incorporated preferentially into PC (Huang and Cabot, 1990; Martin et al., 1993, 1994) and this allowed us to detect dramatic dierences in turnover and the production of PA and DAG. However, we were not able to detect signi®cant increases in basal PLD activity with time in culture in ras-transformed ®broblasts (results not shown). To investigate the role of glycerolipid synthesis in modifying DAG and PA concentrations we incubated ®broblasts, that had been cultured for 3 days, with [14C]glucose for 4 min. There was little incorporation of 14C into PC and PE and therefore the labeled PA and DAG were derived almost exclusively from synthesis de novo and not from phospholipid turnover. There was no signi®cant increase in accumulation of label in PA plus DAG in ras-transformed ®broblasts compared to control cells. Therefore, the increased PA and DAG Figure 6 PA and DAG concentrations in ras-transformed mass in ras-transformed cells results from phospholipid ®broblasts are decreased by cAMP. Ras-®broblasts (clone 3 in turnover (see Introduction). An increased turnover of Table 1) were incubated from Days 1 to 3 with 0.5 mM CPT- cAMP. (a) shows the concentratios of phosphatidate, diacylgly- newly synthesized PC may explain why the labeling of cerol and ceramide which were calculated relative to the content DAG with [14C]glucose is increased after incubating of phospholipid in the cells. (b) shows the relative labeling of ras-transformed ®broblasts for 30 min (Chiarugi et al., phosphatidate, diacylglycerol and triacylglycerol with 14 [3H]myristate which was obtained as described in Figure 1. 1989). In our experiments, the ratio of CinPA Representative results for both experiments are shown and these relative to DAG was increased in transformed were reproduced in separate independent experiments ®broblasts. This eect was compatible with the low membrane-bound PAP-1 activity, which participates in glycerolipid synthesis (Brindley, 1987), in ras-trans- in DAG concentrations among control and transformed cells relative to control ®broblasts (Figure 2). formed cells on Day 1 in culture, although increases in Increased DAG production has been implicated in PA in ras-transformed ®broblasts were evident. There- the ras-transformation (see Introduction). We therefore fore increases in DAG and PA concentrations did not added puri®ed B cereus PC-PLC to control cells from result from the ras-transformation per se but were Days 1 to 3 in culture to see if this could mimic the dependent on the time of culture. We do not know dierences observed in the ras-transformed lines. Large whether these growth-dependent increases in PA and increases in DAG content occurred, PA mass was DAG concentrations depend upon the production of doubled, ceramide concentrations increased by about growth factors by the transformed cells. By contrast to 8.4-fold (Table 3). Ras-transformation also results in ras-transformed cells, PA concentrations in control lower speci®c activities of Mg2+-independent PAP-2 ®broblasts declined from Day 1 (Table 1), when the (Martin et al., 1993), and increased concentrations of cells were dividing, to Day 3, when they were PA, DAG and ceramide (Table 1). Both ras- essentially growth arrested. This decrease in PA transformed and PC-PLC-treated ®broblasts have concentration is compatible with increased PAP-2 decreased PAP-2 activity but how this might be activities between Days 1 and 3 in culture (Martin et related to increased DAG formation is not yet al., 1993). There was no increase in PAP-2 activity with known. Our observations with PC-PLC are consistent culture time in ras-transformed ®broblasts, the cells with proposals that increased DAG production mimics continued dividing and PA concentrations increased to some extent signaling through the p21ras pathway. (Table 1). The conversion of DAG to PA by DAG Most of the 8.4-fold increase in ceramide concentration kinases is increased in ras-transformed ®broblasts is probably accounted for by DAG activation of an relative to control cells on Day of culture (Figure 3). acidic sphingomyelinase (SchuÈ tze et al., 1992; Weig- Therefore changes in the balance of Mg2+-independent mann et al., 1994). In principle, increased DAG mass PAP-2 and DAG kinases activities could contribute to in transformed cells could also activate sphingomyeli- increased PA concentrations in ras-transformed fibro- nase and ceramide production (Table 1) which could Phosphatidate and ceramide in transformed fibroblasts AMartinet al 1577 contribute to the observed elevation in phosphorylcho- also increase the duration and magnitude of signaling line (Price et al., 1989). DAG is not the only regulator by the mitogenic lipids, lysoPA, ceramide 1-phosphate of sphingomyelinase, however, since ceramide concen- and sphingosine 1-phosphate in addition to PA since trations are high in transformed cells after one day in PAP-2 degrades all of these lipids (Waggoner et al., culture when DAG concentrations are similar to those 1996; Brindley and Waggoner, 1996). At present, the in control ®broblasts. Also, ras-transformed cells complex inter-relation that can exist between signaling treated with CPT-cAMP had decreased PA and DAG via cAMP and PA in terms of regulating cell division concentrations and increased ceramide concentrations and morphology is not fully understood. (Figure 5a). Ceramide concentrations can thus be We demonstrated for the ®rst time that PA dissociated from basal PA and DAG concentrations accumulation in ras-transformed ®broblast increased in the transformed cells. The decreased stimulation of markedly with time in culture. In relative terms, the PLD in ras- and fps-transformed ®broblasts in increase in PA concentrations exceeded those for response to serum, platelet-derived growth factor and DAG. The increase in PA concentrations in ras- phorbol ester (Huang and Cabot, 1992; Fu et al., 1992; transformed cells can be explained by the combination Martin et al., 1993; Alam et al., 1995) may result partly of the constitutively increased turnover of phospholi- from their increased concentrations of ceramides pids, decreased activities of Mg2+-dependent PAP-1 (Table 1) which can prevent PLD activation (GoÂ mez- and Mg2+-independent PAP-2 and increased DAG MunÄ oz et al., 1994, 1995b; Venable et al., 1994; kinase activity. The extent and mechanisms by which Nakamura et al., 1995). Furthermore, ceramides can PA can modify elements of the p21ras signaling pathway decrease cell division (Bielawska et al., 1992) and they and the transformed phenotype will require further could limit the increased cell division in ras-trans- study. formed ®broblasts. Speci®cally, ceramides can block the stimulation of DNA synthesis by PA and lyso-PA in rat ®broblasts (GoÂ mez-MunÄ oz et al., 1994). There appears to be a complex inter-relation Materials and methods between cAMP and PA metabolism. Ras-transformed Materials ®broblasts have decreased basal cAMP concentrations (Davies et al., 1989), and generate less cAMP in The source of materials and ®broblast lines have been response to b-adrenergic stimulation or forskolin described (Martin et al., 1993; GoÂ mez-MunÄ oz et al., 1994). compared to parental cells (Tagliaferri et al., 1988; Thin layer chromatography plates of silica gel G were from Wakelam et al., 1991). Stimulation of cAMP produc- Merck, or Macherey-Nagal. CPT-cAMP and PC-PLC (Grade 1 from B cereus)werefromBoehringerMann- tion by PGE is not changed 40 min after scrape- 1 heim, Laval, QueÂ bec. DAG kinase (Ecoli)wasfrom loading with active Val12 p21ras (Price et al., 1989) Calbiochem, San Diego, CA; cardiolipin, DAG, ceramide, indicating that any eect of p21ras on adenylyl cyclase is ATP, DETAPAC, and sphingomyelinase (B cereus)were probably a secondary long-term response. Incubation from Sigma Chemical Company, St Louis, MO; octylgluco- with cAMP analogues inhibits cell division (Tagliaferri side was from ICN, St Laurent, QueÂ bec. [32P]ATP and D- et al., 1988) with ras-transformed cells being more [U-14C] glucose were from Amersham Life Sciences, sensitive than the controls (Davies et al., 1989). cAMP Oakville, Ontario. can inhibit cell division by abrograting the activation of MAP kinase by Raf (Cook and McCormick, 1993; Culture of ®broblasts Hordijk et al., 1994). Our results show that the The characteristics of control and transformed ®broblasts pathway by which excess DAG and PA are produced and culture conditions were described previously (Martin is regulated partly by cAMP (Figure 5) and may result et al., 1993), except the medium was changed every 24 h from the eect of cAMP in taking the ®broblasts out when CPT-cAMP, or PC-PLC were employed. Cells were of the cell cycle. Long-term treatment of the ras- plated at 1.26106 cells per 10 cm tissue culture dish and transformed cells with CPTcAMP did not decrease incubated in DMEM containing 10% (v/v) fetal bovine basal PLD activity. However, increasing cAMP serum at 378Cin5%CO2 in air. DNA was measured as concentrations in ras-transformed ®broblast decreases described previously (Morgan and Pulleyblank, 1974). the elevated activity of phosphatidylinositol-speci®c PLC (Murphy et al., 1994) and this could limit DAG Incorporation of [3H]myrsitate and [14C]glucose into production. The decrease in PA concentrations caused glycerolipids by cAMP could inhibit cell division partly by Fibroblasts were washed twice in D-PBS and then decreasing Raf activation (Ghosh et al., 1996). maintained for 2 h in serum-free DMEM before labeling Conversely, diminished basal and stimulated cAMP for 2 h with [3H]myristate. Myristate was removed by concentrations in ras-transformed cells (Tagliaferri et washing with albumin solution, the cells were harvested 2 h al., 1988; Davies et al., 1989; Wakelam et al., 1991) later and lipids were analysed by thin-layer chromatogra- may result partly from the large increase in PA phy (Martin et al., 1993; GoÂ mez-MunÄ oz et al., 1994). For concentrations (Figure 2 and Table 1) which could the determination of glycerolipid synthesis de novo control inhibit adenylyl cyclase (Murayama and Ui, 1987; van and transformed ®broblasts that were cultured for 3 days Corven et al., 1989; GoÂ mez-MunÄ oz et al., 1994). in 60 mm dishes were washed twice with phosphate- Increased PA concentrations in ras-transformed fibro- buered saline. Cells were then incubated for 1 h in glucose-free DMEM containing 10% fetal calf serum that blasts could also alter morphology since PA, or its had been dialysed against 0.9 g NaCl/100 ml. The medium derivative lysoPA, can activate actin polymerization was replaced with 1.5 ml of identical medium containing involving Rho (Ha and Exton, 1993; Moolenaar, 1995; 5 mCi ml [U-14C]glucose (298 Ci/mol). After incubating for Cross et al., 1996). The decreased PAP-2 activity of the 4 min the medium was removed and cells were washed ras-transformed ®broblasts (Martin et al., 1993) could three times with ice-cold phosphate-buered saline. Cells Phosphatidate and ceramide in transformed fibroblasts A Martin et al 1578 were harvested and lipids were extracted and analysed by lipids and most zwitterionic phospholipids while PA thin layer chromatography. Plates (10610 cm) were remained near the origin. The plates were then developed developed in the ®rst dimension with chloroform/ in the reverse direction with chloroform/methanol/acetic

methanol/NH4OH (65 : 35 : 7.5, v/v/v). After drying the acid (8 : 2: 1, v/v/v) based upon a method of Kennerly plates were developed for 6 cm in the second dimen- (1987). Dried plates were then stained with 0.03% sion with chloroform/acetone/acetic acid/methanol/water Coomassie Blue R250 in 20% (v/v) methanol containing (50 : 20 : 15: 10 : 5, by vol) and then for the full length in the 100 mM NaCl (Nakamura and Handa, 1984). After 1 h same direction with hexane/diethyl ether/acetic acid plates were destained for 10 ± 15 min using 20% methanol

(60 : 40 : 1, v/v/v). Lipids were detected with I2 vapor, and then dried. The absorbancies of PA bands were identi®ed by co-chromatography with authentic standards measured with a re¯ectance densitometer at 580 nm and and the radioactivity was determined (Martin et al., 1993). the mass of PA calculated from the standard curves. More than 90% of the 14C was isolated in the glycerol Duplicate samples from each cell line were analysed on backbone after alkaline hydrolysis. Protein (Martin et al., separate plates, and average values calculated. PC, PE, 1993) and DNA concentrations (Morgan and Pulleyblank, phosphatidylinositol and phosphatidylserine were sepa- 1974) were also determined. rated from the lipid extract by thin layer chromatography (Kennerly, 1987) and organic phosphate concentrations were measured (Fiske and Subbarow, 1925) with a Measurement of PAP, PLD and DAG kinase activities minaturized assay. Activities of Mg2+-dependent PAP-1 was determined as the dierence in activity in the presence and absence of N- Measurement of DAG and ceramide mass ethylmaleimide (Martin et al., 1991; Jamal et al., 1991). Fibroblasts were lysed with digitonin (Martin et al., 1986) Samples of cell extract (25 ± 50 nmol of phospholipid) and the cytosol separated from cell ghosts to determine the together with standards of DAG and ceramide (50 ± functionally active PAP-1 which is membrane-bound 2000 pmol) were assayed (Preiss et al., 1986, 1987) with (Martin et al., 1986; Brindley, 1987). PAP-2 was assayed the following modi®cations: Reaction mixtures containing in two dierent ways: PAP-2A within its natural membrane 50 mM imidazole/HCl, pH 6.6, 1 mM DETAPAC, pH 6.6,

using liposomes of PA and PAP-2B in the presence of 50 mM NaCl, 12.5 mM MgCl2,1mM EGTA, 10 mM sucient Triton X-100 to solubilize PAP-2 and to form dithiothreitol, 1 mM ATP, DAG kinase (approximately mixed micelles with PA and PC. Homogenates were 15.5 munits/assay), 2.5% octyl-b-D-glucoside and 1 mM prepared from cells harvested in 400 mM vanadate and cardiolipin were incubated at 378C for 20 min before 1mM Zn2+ to maximize the activity of PAP-2A (Martin et 1 mCi/assay of [g-32P]ATP was added. Samples of this al., 1993). PLD activity was determined by the formation mixture (100 ml)wereaddedtoeachtubewhichwas of [3H]phosphatidylethanol in ®broblasts labeled in PC vortexed and incubated at 378C for 5 min. The tubes were with [3H]myristate (Martin et al., 1993). sonicated for 5 min and then incubated at 378Cfor5min. DAG kinase activity was measured with either diocta- Samples were resonicated for 5 min and incubated at 378C noylglycerol or stearoylarachidonoylglycerol and by follow- for 20 min. Exogenous DAG standards were quantita- ing the formation of [32P]PA from [g-32P]ATP (Lemaitre et al., tively converted to PA in the absence or presence of 1990). Thin layer plates were ®rst developed with chloroform/ 50 nmol phospholipid from the sample. Therefore the

methanol/NH4OH (65 : 35 : 7.5, v/v/v) for 3 cm above quantity of sample did not aect DAG kinase activity the origin. Air-dried plates were then developed for (Preiss et al., 1987; Paterson et al., 1991). Lipids were 19 cm with chloroform/acetone/acetic acid/methanol/water extracted (Paterson et al., 1991) with 470 mlofchloro- (50 : 20 : 15 : 10 : 5, by vol). Reaction rates were calculated by form/methanol/10 mM HCl (15 : 30 : 2, v/v/v) containing linear regression analysis of results for three protein 20 mg/ml of PA as carrier. Phases were separated with concentrations and on triplicate samples. The speci®city of 150 ml chloroform and 1 ml water. After centrifuging the the stearoylarachidonoyl kinase assay was established by the aqueous phase was aspirated and the chloroform phase failure of 1 mM dioctanoylglycerol to inhibit signi®cantly the washed twice with 1 ml portions of water, before drying production of stearoylarachidonoyl-PA. the lipid under vacuum. Products were separated on plastic thin layer plates of silica gel 60 (Merck) by developing with chloroform/methanol/NH OH Measurement of phospholipid mass 4 (65 : 35 : 7.5, v/v/v), drying and then developing in chloro- Cells were maintained in DMEM supplemented with 10% form/methanol/acetic acid/acetone/water (10 : 2 : 3 : 4 : 1 by (v/v) fetal calf serum. Thin layer plates were washed twice vol). 32P-labeled PA and ceramide-1-phosphate were with ice-cold D-PBS and extracted as before (Martin et al., identi®ed with authentic standards and a Bioscan Radio- 1993). Samples of the chloroform phases were analysed for imager, or by autoradiography. Radioactivity was organic phosphate (Fiske and Subbarow, 1925), and the determined by scintillation counting and lipid mass was remainder was dried under vacuum. For assays of cells calculated from standard curves obtained from ceramide after one day in culture, two 10 cm dishes were pooled to and DAG. provide a sucient mass of lipid. On Days 2 and 3 only one dish of cells was required. PA mass was measured basedonBocckinoet al. (1987). Lipid extracts were applied at 10 cm from the bottom of a 20 cm plastic thin-layer plate of silica gel 60 (Merck) that had been Acknowledgements developed in petroleum ether, b.p. 40 ± 608C/diethyl ether This work was supported by the Alberta Heritage (1 : 1, v/v). Standards of PA (0.5, 1.0, 2.5, and 10 mg) were Foundation for Medical Research, the Canadian National applied to every plate. The plates were developed with Cancer Institute, the Medical Research Council of Canada

chloroform/methanol/NH4OH (65 : 35 : 7.5, v/v/v), air dried and the Cancer Research Fund of the University of and then cut 1.5 cm above the origin. This removed neutral Alberta.

References

Alam S, Banno Y, Nakashima S and Nozawa Y. (1995). Bauldry SA, Bass DA, Cousart SL and McCall CE. (1991). J. Biochem. Biophys. Res. Commun., 207, 460 ± 466. Biol. Chem., 266, 4173 ± 4179. Phosphatidate and ceramide in transformed fibroblasts AMartinet al 1579 Bielawska A, Linardic CM and Hannun YA. (1992). FEBS Larrodera P, Cornet ME, Diaz-Meco MT, Lopez-Barahona Lett., 307, 211 ± 214. M, Diaz-Laviada I, Guddal PH, Johansen T and Moscat Bocckino SB, Blackmore PF, Wilson PB and Exton JH. J. (1990). Cell, 61, 1113 ± 1120. (1987). J. Biol. Chem., 262, 15309 ± 15315. Leevers SJ, Paterson HF and Marshall CJ. (1994). Nature, Bocckino SB, Wilson PB and Exton JH. (1991). Proc. Natl. 369, 411 ± 414. Acad. Sci. USA, 88, 6210 ± 6213. Lemaitre RN, King WC, MacDonald ML and Glomset JA. Bos JL. (1989). Cancer Res., 49, 4682 ± 4689. (1990). Biochem. J., 266, 291 ± 299. Brindley DN. (1987). Phosphatidate Phosphohydrolase. Limatola C, Schaap D, Moolenaar WH and van Blitterswijk Brindley DN. (ed) CRC Press: Boca Raton, vol. 1, WJ. (1994). Biochem. J., 304, 1001 ± 1008. pp. 1 ± 77. Lin LL, Wartmann M, Lin AY, Knopt JL, Seth A and Davis Brindley DN and Waggoner DW. (1996). Chem. Phys. RJ. (1993). Cell, 72, 269 ± 278. Lipids, 80, 45 ± 57. Lopez-Barahona M, Kaplan PL, Cornet ME, Diaz-Meco Carnero A, Dol® F and Lacal JC. (1994a). J. Cell Biochem., MT, Larrodera P, Diaz-Laviada I, Municio AM and 54, 478 ± 486. Moscat J. (1990). J. Biol. Chem., 265, 9022 ± 9026. Carnero A, Cuadrado A, del Paso L and Lacal JC. (1994b). Macara IG. (1989). Mol. Cell. Biol., 9, 325 ± 328. Oncogene, 9, 1387 ± 1395. Martin A, Hopewell R, Martin-Sanz P, Morgan JE and Chiarugi V, Bruni P, Pasquali F, Magnelli L, Basi G, Brindley DN. (1986). Biochem. Biophys. Acta, 876, 581 ± Ruggiero M and Farnararo M. (1989). Biochem. Biophys. 591. Res. Comm., 164, 816 ± 823. Martin A, GoÂ mez-MunÄ oz A, Jamal Z and Brindley DN. Cook SJ and McCormick FW. (1993). Science, 262, 1069 ± (1991). Methods Enzymol., 197, 553 ± 563. 1072. Martin A, GoÂ mez-MunÄ oz A, Waggoner DW, Stone JC and Cross MJ, Roberts S, Ridley AJ, Hodgkin MN, Stewart A, Brindley DN. (1993). J. Biol. Chem., 268, 23924 ± 23932. Claesson-Welsh L and Wakelam MJO. (1996). Curr. Biol., Martin A, GoÂ mez-MunÄ oz A, Duy PA and Brindley DN. 6, 588 ± 597. (1994). Signal-activated Phospholipases,LiscovitchM. Davies S-A, Houslay MD and Wakelam MJO. (1989). (ed). R.G. Landes Co.: Austin. pp. 139 ± 164. Biochim. Biophys. Acta, 1013, 173 ± 179. Moolenaar WH. (1995). J. Biol. Chem., 270, 12949 ± 12952. Diaz-Meco MT, QuinÄ ones S, Municio MM, Sanz L, Bernal Morgan AR and Pulleyblank DE. (1974). Biochem. Biophys. D, Cabrero E, Saus J and Moscat J. (1991). J. Biol. Chem., Res. Commun., 61, 396403. 266, 22597 ± 22602. Moritz A, DeGraan PNE, Gispen WH and Wirtz KWA. Dominguez I, Marshall MS, Gibbs JB, GarcõÂ a de Herreros (1992). J. Biol. Chem., 267, 7207 ± 7210. A, Cornet ME, Graziani G, Diaz-Meco MT, Johansen T, Murayama T and Ui M. (1987). J. Biol. Chem., 262, 5522 ± McCormick F and Moscat J. (1991). EMBO J., 10, 3215 ± 5529. 3220. Murphy SK, Lindberg JB and Prussack DA. (1994). Eldar H, Ben-Av P, Schmidt U-S, Livnch E and Liscovitch Biochem. Biophys. Res. Comm., 205, 1121 ± 1129. M. (1993). J. Biol. Chem., 268, 12560 ± 12562. Nakamura K and Handa S. (1984). Anal. Biochem., 142, Fiske CH and Subbarow I. (1925). J. Biol. Chem., 66, 375 ± 406 ± 410. 400. Nakamura T, Nakashima S, Ojio K, Banno Y, Miyata H and Fu T, Okano Y and Nozawa Y. (1992). Biochem. J., 283, Nozowa Y. (1995). J. Immuno., 156, 256 ± 262. 347 ± 354. Paterson A, Plevin R and Wakelam MJO. (1991). Biochem. GhoshS,StrumJC,SciorraVA,DanielLandBellRM. J., 280, 829 ± 836. (1996). J. Biol. Chem., 271, 8472 ± 8480. Preiss J, Loomis CR, Bishop WR, Stein R, Niedel JE and GoÂ mez-MunÄ oz A, Duy PA, Martin A, O'Brien L, Byun H- Bell RM. (1986). J. Biol. Chem., 261, 8597 ± 8600. A, Bittman R and Brindley DN. (1995). Mol. Pharmacol., Preiss JE, Loomis CR, Bell RM and Niedel JE. (1987). 47, 883 ± 889. Methods in Enzymol., 141, 294 ± 300. GoÂ mez-MunÄ oz A, Martin A, O'Brien L and Brindley DN. Price BD, Morris JDH, Marshall CJ and Hall A. (1989). J. (1994). J. Biol. Chem., 269, 8937 ± 8943. Biol. Chem., 264, 16638 ± 16643. GoÂ mez-MunÄ oz A, Waggoner DW, O'Brien L and Brindley Rodriguez-Viciana P, Warne PH, Dhand R, Vanhaesebroeck DN. (1995b). J. Biol. Chem., 270, 26318 ± 26325. B, Gout I, Fry MJ, Water®eld MD and Downward J. Guan KL. (1994). Cellular Signalling, 6, 581 ± 589. (1994). Nature, 370, 527 ± 532. Ha K-S and Exton JH. (1993). J. Cell. Biol., 123, 1789 ± 1796. Rossi F, Grzeskowiak M, Della Bianca V, Calzetti F and Hordijk PL, Verlaan I, Jalink K, Van Corven EJ and Gandini G. (1990). Biochem. Biophys. Res. Comm., 168, Moolenaar WH. (1994). J. Biol. Chem., 269, 3534 ± 3538. 320 ± 327. Howe LR and Marshall CJ. (1993). J. Biol. Chem., 268, Satoh T, Nakafuku M and Kaziro Y. (1992). J. Biol. Chem., 20717 ± 20720. 267, 24149 ± 24152. Huang C and Cabot MC. (1992). Biochem. Biophys. Acta, SchuÈ tze S, Pottho K, Machleidt T, Berkovic D, Wiegmann 1127, 242 ± 248. KandKroÈnke M. (1992). Cell, 71, 765 ± 776. Huang C and Cabot MC. (1990). J. Biol. Chem., 265, 17468 ± Smith MR, DeGudicibus SJ and Stacey DW. (1986). Nature, 17473. 320, 540 ± 543. Jamal Z, Martin A, GoÂ mez-MunÄ oz A and Brindley DN. Smith MR, Liu Y-L, Kim H, Rhee SG and Kung H-F. (1991). J. Biol. Chem., 266, 2988 ± 2996. (1990). Science, 247, 1076 ± 1077. Jiang H, Li Z, Luo J-Q, Wolfman A and Foster DA. (1995a). Tagliaferri P, Katsaros D, Clair T, Neckers L, Robins RK J. Biol. Chem., 270, 6006 ± 6009. and Cho-Chung YS. (1988). J. Biol. Chem., 263, 409 ± 416. Jiang H, Luo J-Q, Urano T, Frankel P, Lu Z, Foster DA and Teegarden D, Taparowsky EJ and Kent C. (1990). J. Biol. Feig LA. (1995b). Nature, 378, 409 ± 412. Chem., 265, 6042 ± 6047. Jones GA and Carpenter G. (1993). J. Biol. Chem., 268, Tsai M-H, Hall A and Stacey DW. (1989a). Mol. Cell. Biol., 20845 ± 24850. 9, 5260 ± 5264. Kennerly DA. (1987). J. Biol. Chem., 262, 16305 ± 16313. Tsai M-H, Yu C-L, Wei F-S and Stacey DW. (1989b). Lacal JC. (1990). Mol. Cell. Biol., 10, 333 ± 340. Science, 243, 522 ± 526. Lacal JC, Moscat J and Aaronson SA. (1987). Nature, 330, Tsai M-H, Yu C-L and Stacey DW. (1990). Science, 250, 269 ± 272. 982 ± 985. Phosphatidate and ceramide in transformed fibroblasts A Martin et al 1580 Venable ME, Blobe GC and Obeid L. (1994). J. Biol. Chem., Wakelam MJO, Black FM and Davies SA. (1991). Biochem. 269, 26040 ± 26049. Soc. Symp., 56, 103 ± 116. van Corven EJ, Groenink A, Jalink K, Eichholtz T and Wiegmann K, SchuÈ tze S, Machleidt T, Witte D and KroÈ nke Moolenaar WH. (1989). Cell, 59, 45 ± 54. M. (1994). Cell, 78, 1005 ± 1015. Waggoner DW, GoÂ mez-MunÄ oz A, Dewald J and Brindley DN. (1996). J. Biol. Chem., 271, 16506 ± 16509.