We recently reported that DAG (diacylglycerol) generated during sphingomyelin synthesis plays an important role in protein kinase C activation and cell proliferation in Madin–Darby canine kidney cells [Cerbon and Lopez-Sanchez (2003) Biochem. J. 373, 917–924]. In yeast cells, IPC (inositol phosphoceramide) synthase catalyses the transfer of phosphoinositol from phosphatidylinositol to ceramide to form IPC and generates DAG. In the present study, we found that, during the G1 to S transition after N2-starvation, there was a significant increase in the synthesis of IPC accompanied by a progressive increase (up to 6-fold) in the level of DAG. The increased DAG levels coincided with decrements in ceramide and sphingoid base levels, conditions that are adequate for the activation of putative protein kinase C required for the G1 to S transition and proliferation of yeast cells. To separate the role of DAG generated during IPC synthesis from that originating from other sources, we utilized β-chloroalanine and myriocin, inhibitors of serine:palmitoyl-CoA transferase, the first committed step in sphingolipid synthesis, to avoid accumulation of sphingolipid intermediates. When the synthesis of sphingolipids was inhibited, DAG accumulation was significantly decreased and the G1 to S transition was blocked; such blockage was avoided by metabolic complementation with phytosphingosine. The DAG/ceramide ratio was 0.27 and it changed to 2.0 during growth re-initiation, suggesting that the synthesis of phosphosphingolipids could act to switch growth arrest (increased ceramide) to a mitogenic signal (increased DAG), and that this signalling process is preserved in yeast and mammalian cells.
- diacylglycerol (DAG) generation
- G1 to S transition
- inositol phosphoceramide (IPC) synthesis
- nitrogen starvation
Lipids play a central role in signal transduction and in yeast cells, among the signalling lipid molecules, ceramide and DAG (diacylglycerol) have been found to play a role in the regulation of cell growth and viability. A ceramide-activated protein phosphatase mediates ceramide-induced G1 arrest in Saccharomyces cerevisiae , and the addition of C2 ceramide to these yeasts resulted in growth arrest . Similarly, ceramide accumulation is accompanied by cell death in IPC (inositol phosphoceramide) synthesis-deficient yeast mutants . Elevation of DAG levels has been observed before or at the start of transition, the presumptive step of PKC (protein kinase C) activation in response to Cdc28 . The best-studied pathway of DAG generation is through phosphoinositide breakdown with concomitant increases in DAG and intracellular Ca2+ [secondary to IP3 (inositol trisphosphate) generation], and activation of PKC . The addition of ammonium sulphate to starved yeast (G1) cells results in a rapid 3–4-fold increase in IP3 and DAG; however, DAG levels exceeded those of extractable IP3 by 10–15-fold, indicating activation of other phospholipases . In addition, when inositol lipid metabolism was studied in a series of Cdc (cell division cycle) mutants of S. cerevisiae, no evidence for the formation of inositol polyphosphates was found . Other mechanisms of DAG production are mediated by the consecutive action of PC (phosphatidylcholine)-specific PLD (phospholipase D) and phosphatidate phosphohydrolase and by PC hydrolysis through a PC-specific phospholipase C. In S. cerevisiae, homozygous diploid disruption mutants of PLD/SPO 14 have no apparent growth phenotype. They are viable and grow but they are incapable of sporulation, and a second PLD activity has been detected , which in contrast to pld1p/spo14p, is Ca2+-dependent but independent of phosphatidylinositol 4,5-bisphosphate. The genes encoding this second PLD have not yet been identified. Alternatively, phosphatidate phosphatase (Lpp1p) and DAG pyrophosphatase (Dpp1p) appear to be involved in DAG formation. However, strains with deletions in DPP1, LPP1 and the dpp1/lpp1 double mutant are viable and do not exhibit any obvious growth defects . Also, mutants have been isolated which synthesized virtually no PC , suggesting that the DAG necessary for signalling may arise from other sources.
In contrast, yeast mutants incapable of synthesizing sphingolipids (due to impaired synthesis of sphingoid bases), show inability to grow and lose viability if not supplemented, suggesting that there are vital functions for IPCs in yeast . IPC synthase transfers inositol phosphate from PI (phosphatidylinositol) to ceramide to form IPC  and to generate DAG. IPC synthase therefore has the ability to directly regulate, in opposite directions, ceramide and DAG levels, potentially controlling opposite cellular processes such as cell proliferation and growth arrest. In this study, we investigated the synthesis of IPCs during growth re-initiation in S. cerevisiae G0/G1 (stationary phase) cells and in G1 (N2-starved) cells. We analysed the consequences of inactivation of SPT (serine:palmitoyl-CoA transferase) and IPC synthase in S. cerevisiae, which are considered to be rate-limiting enzymes in the sphingolipid biosynthesis pathway . β-Chloroalanine and myriocin , which are inhibitors of SPT (EC 126.96.36.199), the first committed step of sphingolipid synthesis, were used. Blocking the first step of this pathway would result in progressive inhibition of the incorporation of labelled precursors [3H]palmitate, [32P]Pi and [3H]inositol into IPC, MIPC (mannose IPC) and M(IP)2C [mannose (inositol phosphate)2 ceramide].
Furthermore, since S. cerevisiae utilizes exogenous long-chain bases in the synthesis of sphingolipids , supplementation of the growth medium with PHS (phytosphingosine) might serve to restore the synthesis of sphingolipids. We also tested Aur-A (aureobasidin A), an inhibitor of IPC synthase . We provided evidence showing that IPC synthesis was required for the generation of DAG during the presumptive step of PKC activation and for the normal transit through the cell cycle. We found that PHS restored IPC synthesis and cell cycle progression in cells in which SPT was inhibited, and that IPC synthesis was required for the generation of DAG. We discuss the interesting possibility that IPC synthesis could account not only for the control of the ceramide levels, but also in DAG generation. During the synthesis of IPCs, DAG was generated, and the amounts of ceramide and sphingoid bases were decreased. These are adequate conditions for the activation of the putative PKC.
MATERIALS AND METHODS
32P (carrier-free; 10 mCi/ml), [γ-32P]ATP (PB-108; >5000 Ci/mmol), myo-[2-3H]inositol (10–20 Ci/mmol), [9,10(n)-3H]palmitic acid (40–60 Ci/mmol) and DAG assay system were purchased from Amersham Biosciences. Peptone and yeast extract were purchased from Difco Laboratories (Detroit, MI, U.S.A.), the synthetic medium was from Q-biogene (Vista, CA, U.S.A.), β-chloroalanine, o-phthalaldehyde, myriocin and other chemicals were from Sigma, Aur-A was from PanVera (Madison, WI, U.S.A.), organic solvents and Silica gel 60 thin layer plates were obtained from Merck (México) and scintillation liquid from New England Nuclear (Perkin Elmer, México). Phospholipids and neutral lipids were purchased from Avanti Polar Lipids (Alabaster, AL, U.S.A.) and Sigma.
Strains and culture conditions
S. cerevisiae A 364 A (Mata, ade1, ade2, ura1, his7, lys2, tyr1, gal1) obtained from the Department of Genetics (CINVESTAV-IPN, Mexico) and S. cerevisiae ATCC 9080 (A.T.C.C., Manassas, VA, U.S.A.), a natural inositol auxothroph, were used in the experiments. Cultures were maintained on YPD medium [1% (w/v) yeast extract, 2% (w/v) bactopeptone, 2% (w/v) glucose, which was supplemented with 50 mM succinic acid] containing 1.5% (w/v) bacto-agar. Strains were propagated in CSM (complete synthetic medium) and Molecular Biology Certified yeast growth media (Q-biogene), containing 2% glucose, 6.7 g/l nitrogen base, 0.8 g/l supplement mixture (nucleotides and amino acids), 5 g/l ammonium sulphate and 2 μg/ml inositol. Cultures were performed in a rotary shaker (160 rev./min at 30 °C) and harvested in the stationary phase of growth. These cells were used in turn as the inoculum for each experiment (G0/G1 cells). For nutrient induction experiments, cells were grown in CSM to stationary phase (24 h), and transferred to starvation media by centrifugation and resuspension in N2-starvation medium MM-N (minimal medium without nitrogen), which contained yeast nitrogen base, 2.0% glucose and 2 μg/ml inositol, without ammonium sulphate and amino acids at A600 and incubated for 16–20 h at 160 rev./min at 30 °C. The cell concentration and percentage of budded cells were determined. At least 200 cells were scored for their bud morphology. Microscopic examination and plating experiments indicated an arrest of 90–95% unbudded (G1) viable cells. These cells were washed twice with MM-N and were used as the inoculum for the study of the G1 to S transition.
Cells were incubated in CSM at 30 °C in the absence and presence of the indicated concentrations of the inhibitors of sphingolipid synthesis. Myriocin and Aur-A were dissolved in ethanol and added to CSM to the final concentrations as indicated in the Results section. The concentration of ethanol in CSM was 0.5% β-chloroalanine dissolved in water. Proliferation was determined as the absorbance A at 600 nm. For viability measurement, an yeast culture was mixed with an equal volume of a Methylene Blue solution (0.2 mg/ml in 0.1 M phosphate buffer, pH 4.6). The mixture was left for 5 min at room temperature (23 °C) and total cells and stained cells (dead) were counted under a microscope. Viability was expressed as the mean of the percentage of stained cells.
Labelling and analysis of sphingolipids, phospholipids and neutral lipids
Pulse and steady-state labelling of lipids with myo-[2-3H]inositol, [9,10(n)-3H]palmitic acid and [32P]Pi were performed to quantify the effects of inhibitors on total sphingolipid synthesis. Washed G0/G1 cells were resuspended to an initial density of approx. 106 cells/ml in CSM in the absence or presence of the inhibitor as indicated; cells were then incubated with the radiolabelled precursor for the indicated time periods to pulse-label sphingolipids, phospholipids and neutral lipids. Lipids were extracted from labelled cells using a method described by Hanson and Lester . Cells were treated with 5% (w/v) (final concentration) trichloroacetic acid for 20 min at 4 °C. Cells were centrifuged, washed twice with 1.0 ml of 5% trichloroacetic acid, and finally with water. Cell pellets were extracted with 1.0 ml of 95% ethanol/water/diethyl ether/pyridine (15:15:5:1, by vol.) for 60 min at 60 °C. Phospholipids were analysed by one-dimensional TLC on silica gel plates activated at 110 °C for 15 min and impregnated with 1% potassium oxalate and 2 mM NaEDTA in methanol/water (2:3, v/v). TLC was developed with chloroform/acetone/methanol/acetic acid/water (40:15:13:12:8, by vol.). In some experiments, sphingolipids were analysed after alkaline hydrolysis of glycerophospholipids. Lipids were deacylated by the addition of 1 ml of 0.2 M NaOH in methanol and on incubation at 30 °C for 45–60 min. After the addition of 1 ml of 5% EDTA, samples were neutralized by the addition of 0.2 ml of 1 M acetic acid (60 g/l). Lipids were then extracted with 1 ml of chloroform, dried under N2, resuspended in 250 μl of chloroform and 50 μl samples were analysed by TLC. Neutral lipids were separated by one-dimensional TLC on silica gel G plates, developed with light petroleum (b.p. 30–75 °C)/ethyl ether/acetic acid (90:10:1, by vol.). After separation, lipids were visualized by exposure of the plates to iodine vapour or after autoradiography of plates, where the corresponding standards were included. The individual phospholipids were removed from the plate and quantified by scintillation counting. The phosphorus in phospholipids (nmol of Pi present in the chloroform phase) was determined by the method of Ames and Dubin . The resultant phosphate value was used to normalize samples against each other (see ). Radioactivity was expressed in terms of c.p.m. or d.p.m. in 107 cells. Inositol-containing sphingolipids were identified based on their non-saponifiable characteristic. When the [myo-2-3H]inositol- and [32P]Pi-labelled non-deacylable lipid fractions were subjected to TLC, 4–5 bands were found and RF values were 0.5, 0.44, 0.40, 0.37 and 0.28. Two of these lipids (RF 0.4 and 0.28) gave a positive test when sprayed with the orcinol-sulphuric acid reagent and were labelled with [14C]mannose (60 min pulse), indicating that they were MIPC and M(IP)2C. The upper one (RF 0.4) does not separate well by TLC and labelling of IPC+MIPC was evaluated. The lower one (RF 0.28), in addition to being more polar, was identified as M(IP)2C, since pulse labelling shows that the RF 0.28 band was labelled in relation to their precursors (IPC and MIPC). Quantification was performed in the same TLC plate. The restricted number of manipulations avoids loss of the sample and inaccuracies. The chromatographic mobilities observed compare well with the results reported by Smith and Lester . Also, after alkaline hydrolysis, the amide bond of ceramides was detected by infrared spectroscopy . Glycerophospholipids were identified by co-migration with standards from Sigma and Avanti Polar. Also, phosphatidylethanolamine and PS (phosphatidylserine) were identified by trinitrobenzylation, PI by incorporation of [3H]inositol and PC by incorporation of [3H]methyl choline as described previously .
DAG mass measurements
A radioenzymatic assay employing Escherichia coli DAG kinase (Amersham Biosciences) was used for a quantitative determination of DAG under defined mixed micelle conditions to solubilize the DAG. The enzyme converts DAG to [32P]phosphatidic acid with [γ-32P]ATP (PB 108; >5000 Ci/mmol). The assay was performed according to the manufacturer's instructions. [32P]Phosphatidic acid was extracted and subsequently separated from other lipids and residual [γ-32P]ATP by TLC, using phosphatidic acid (Sigma) as a reference. The radioactivity of [32P]phosphatidic acid was estimated by liquid scintillation counting. DAG/phospholipid ratios were calculated and expressed as pmol of DAG/nmol of phosphate. The use of DAG/phospholipid ratios corrects for the differences in cell number and recoveries. Ceramide levels were evaluated also using the DAG kinase method using ceramide phosphate (Sigma) as described previously .
Measurement of levels of sphingoid bases
Sphingoid bases from unlabelled cells were quantified using HPLC as described by Merrill et al. . Sphingoid base standards were used for the identification and quantification. Initial studies indicated that a minimum of 1×107 yeast cells were required for sufficient cellular sphingoid bases to be measured by this method.
Measurement of phospholipid turnover
G0/G1 cells were subcultured in CSM containing [3H]palmitic acid (5 μCi/ml) in the absence or presence of 5 mM β-chloroalanine. At time zero, the average cell number was (1.0–1.2)×106 cells/ml. After 20 min, the cells were washed with PBS, and fresh medium in the absence or presence of 5 mM β-chloroalanine was added. At the indicated times, up until 4 h, duplicate samples were treated with trichloroacetic acid, and lipids were extracted, separated by TLC and quantified. At the 20 min pulse time, there was (146±10)×103 d.p.m./107 cells in the control cells and (140±12)×103 d.p.m./107 cells in the β-chloroalanine-treated cells. At the indicated times, duplicate samples were harvested, and lipids extracted and analysed as described above.
Effects of β-chloroalanine on yeast cell viability and proliferation
Inhibitors of specific steps in yeast and mammalian sphingolipid biosynthesis have been identified and proved to be useful in understanding sphingolipid functions. β-Chloroalanine and myriocin (ISP-1) inhibit SPT, and it has been suggested that the first step in sphingolipid biosynthesis is probably regulated and found to be important for controlling the rate of ceramide synthesis de novo and the rate of complex sphingolipid synthesis . Figure 1 shows that there was a concentration-dependent inhibition of cell proliferation and that 10 mM β-chloroalanine resulted in 90% inhibition. No lysis or other gross changes in cell morphology were observed in drug-treated cells and cells were able to grow after washing out the β-chloroalanine even after 10 h of exposure. There was no significant loss of cell viability (5–7.5% dead cells). This means that β-chloroalanine-treated cells were prevented from normal cell division for an extended period of time. The response to β-chloroalanine was decreased by increasing cell number; thus all subsequent experiments were conducted at starting cell concentrations of 0.2–0.5 absorbance unit/ml.
Effects of β-chloroalanine on phospholipid synthesis
Exogenously added [32P]Pi was used to estimate the synthesis of phospholipids. G0/G1 cells were washed twice with 25 ml of water and then were resuspended to an A600 of 0.4 in 20 ml of CSM [(4.0–4.4)×107 cells] and labelled with 5 μCi/ml [32P]Pi for 60 min in the absence or presence of 5 mM β-chloroalanine. In Figure 2, the results of TLC and autoradiography (top panel) and the radioactivity incorporated into lipids (bottom panel) show that β-chloroalanine decreased the amount of 32P in sphingolipids to 53% of the control value (P<0.001), whereas the amount of label incorporated into the precursor PI was increased slightly. Although there was a 24% decrement in the amount of 32P incorporated into PC (P<0.05), the incorporation of radiolabel into total lipid was not altered by exposure to β-chloroalanine during this period of time [(298±15)×103 compared with (263±21)×103 c.p.m./107 cells]. Similar results were obtained with the inositol auxothroph strain (A.T.C.C. 9080). After 60 min of incubation in the presence of 5 mM β-chloroalanine, sphingolipid labelling was decreased to 58% of control [from (99±10)×103 to (53±15)×103 c.p.m./107 cells; P<0.01], whereas the amount of labelled PI was increased slightly [from (98±6)×103 to (116±17)×103 c.p.m./107 cells]. Also, the label incorporated into PC was decreased by 23%.
Effects of time of incubation and β-chloroalanine on cellular DAG generation
G0/G1 cells were pre-labelled with [3H]palmitic acid (1 μCi/ml) to steady state (during 24 h), washed twice with water and resuspended to an absorbance of 0.2 unit/ml in CSM without radiolabel in the absence or presence of 5 mM β-chloroalanine. At the indicated times, lipids were extracted, neutral lipids were separated by TLC and the label incorporated in DAG was determined. Figure 3 shows that labelled DAG was increased 2.2 times during the first 40 min of incubation, and then decreased towards the initial values after 60 min of incubation. In the cells exposed to β-chloroalanine, after a transitory increment during the first 10–20 min, levels of labelled DAG were decreased to 0.5–0.3 times the control levels at 40 min of incubation (P<0.01). The inhibition of both the generation of DAG and sphingolipid synthesis by β-chloroalanine suggested that these two processes are probably related.
Effects of β-chloroalanine on phospholipid turnover
Previous studies on the dynamics of inositol-containing phospholipids in S. cerevisiae during growth , showed that during the lag phase (first 2 h of incubation) there was a significant increment of 9.0–21.6% in labelled IPCs (IPC+MIPC) accompanied by a 16% decrement in PI in [3H]inositol pre-labelled yeast cells at steady state. Therefore we examined the effects of 5 mM β-chloroalanine on inositol phospholipid turnover using a 20 min pulse with [3H]palmitic acid (5 μCi/ml) in the absence or presence of 5 mM β-chloroalanine, and a 1–4 h chase. After the 20 min pulse, the incorporation of radiolabel into total phospholipids was similar in both the control and β-chloroalanine-treated cells [(146±10)×103 and (140±12)×103 d.p.m./107 cells respectively]. Figure 4 shows that labelled PI was decreased to 54% after a 60 min chase, whereas labelled IPCs were increased 1.44 times. After a 120–240 min chase, labelled PI was further decreased (75%) whereas labelled IPCs were increased 2.2 times. In the β-chloroalanine-treated cells, there were no significant changes in labelled IPCs and labelled PI was not decreased during the first 2 h of chase. These results suggest that the slightly increased labelling of PI observed in the β-chloroalanine-treated cells (Figure 2) was due to the inhibition of IPC synthesis and that there was no detectable degradation of sphingolipids.
Changes in endogenous sphingoid bases during growth initiation
To ensure that β-chloroalanine addition had effectively inhibited the activity of SPT, we determined the levels of DHS (D-erythrosphingosine) and PHS during growth re-initiation. It can be seen (Table 1) that after 2 h of incubation in CSM (lag phase), the levels of sphingoid bases were decreased to 25% of the levels observed at 0 h. After 8 h of incubation (exponential phase), only 2.5% of sphingoid base levels determined at 0 h remained, indicating an efficient metabolic transformation following growth re-initiation. During the late-exponential to stationary phase of growth (12–24 h), the levels of DHS and PHS were increased towards the initial values. In the cells exposed to β-chloroalanine, after 2 h of incubation, DHS and PHS were not detected, which is consistent with inhibition of the synthesis of sphingoid bases. The inhibition by long-chain bases of PKC and other enzymes, including CTP:phosphocholine transferase , IPC synthase, PS synthase and phosphatidate phosphohydrolase, was observed in vitro, and suggested to occur in vivo . However, this possibility seems unlikely during the lag–exponential phases of growth, considering the decreased concentrations of DHS and PHS determined. These results also suggested that the decrease in the synthesis of PC observed in the β-chloroalanine-treated cells could be related to the decreased generation of DAG and not to the inhibition by sphingoid bases of the CTP:phosphocholine transferase.
Effects of inhibiting the synthesis of sphingolipids on the G1 to S transition
It has been reported that there was a 2–3-fold increase in DAG levels prior to budding in cells re-entering the cell cycle following recovery from G1 arrest, induced by mating pheromone as well as in early G1 cells obtained by centrifugal elutriation . Having found that β-chloroalanine inhibited sphingolipid synthesis, decreased the DAG levels and inhibited proliferation, we tested the effect of β-chloroalanine on the G1 to S transition in G1 cells synchronized by nitrogen arrest. Since β-chloroalanine appears to inhibit various pyridoxal phosphate-dependent enzymes, in addition to SPT, we examined if myriocin, another specific inhibitor of SPT  added to the incubation medium, inhibits the G1 to S transition. We also tested the effects of Aur-A, an inhibitor of IPC synthase, considering that DAG is generated during the synthesis of IPC. The results in Table 2 show the effects of inhibiting the synthesis of sphingolipids on the G1 to S transition. After release of N2 starvation (180 min of incubation time), the G1 to S transition (budding) was inhibited to 70% with 5 mM β-chloroalanine, 55–69% with myriocin (10–20 μM) and 51–75% with Aur-A (4.5–20 μM). Inhibition of the G1 to S transition with 20 μM myriocin was restored by metabolic complementation with 50 μM PHS, indicating that the inhibition of SPT was specific. When the synthesis of sphingolipids was determined after N2 starvation, it was found that [3H]palmitic acid (3 μCi/ml) labelling of sphingolipids during a 60 min pulse was decreased to 29% of control with 5 mM β-chloroalanine, to 39% with 20 μM myriocin and to 34% with 20 μM Aur-A.
Effects of Aur-A on cellular concentrations of ceramide and DAG mass levels during the G1 to S transition
IPC synthase transfers the phosphoinositol head group from PI to ceramide, forming IPC and generating DAG. This activity places IPC synthase at an important regulatory position in signalling pathways that utilize ceramide and DAG as second messengers, because it could act to switch a growth arrest to a mitogenic signal. Therefore we decided to determine the balance between mitogenic DAG and anti-mitogenic ceramide during the G1 to S transition and the effect of the inhibition of IPC synthase with Aur-A. Figure 5 shows that during the first 40 min of incubation in CSM after release of N2 starvation, the amount of DAG in the control non-exposed cells was increased 6 times and decreased towards the initial values after 60 min of incubation and ceramide levels were decreased by 1.5 times. During the first 40 min of incubation in the cells exposed to Aur-A, the amount of DAG was decreased to 33% of control values (P<0.01) and during the first 60 min of incubation, ceramide levels were increased 1.4 times (P<0.01). The DAG/ceramide ratio was high in proliferating cells (2.0 versus 0.4 in the cells arrested in G1 with Aur-A), indicating that IPC synthesis, DAG generation and decreased levels of ceramide were required for the normal transit through the cell cycle.
Changes in sphingoid base levels during the G1 to S transition
To determine whether the increments in DAG and the decrements in ceramide coincided with decrements in DHS and PHS, the levels of sphingoid bases were determined. Table 3 shows that, after the release of N2-starvation, the levels of sphingoid bases were decreased to 53% after 20 min and to 21% after 40 min; after 60 min, only 0.75% of the levels remained compared with those observed at 0 h. These results, together with those of Table 1, indicate that sphingolipid metabolism was efficiently co-ordinated, and that preformed ceramide, DHS and PHS were metabolized during growth re-initiation. To know if these precursors were used for the synthesis of IPCs, yeast cells prelabelled with [3H]inositol (1 μCi/ml) during a 24 h period were N2-starved, and G1 cells (90–95% unbudded cells) were washed and resuspended in CSM at 30 °C. At the indicated times, lipids were extracted, separated by TLC and radioactivity incorporated in IPC/MIPC and M(IP)2C was determined. Figure 6 shows that there was a significant increment of [3H]inositol radioactivity incorporated in sphingolipids. During the first 60 min, a 2.4-fold increment was observed [76% in IPC/MIPC and 24% in M(IP)2C], and between 60 and 120 min the radioactivity incorporated was increased 1.5 times more [61% in IPC/MIPC and 39% in M(IP)2C]. Radioactivity of the prelabelled [3H]PI was decreased 1.8 times between 0 and 2 h, indicating that in addition to synthesis de novo, the preformed precursors DHS, PHS, ceramide and PI were used for the synthesis of sphingolipids. Therefore the conditions of increased DAG levels, together with decreased levels of ceramide and sphingoid bases, were adequate for the activation of the putative PKC necessary for the G1 to S transition and cell proliferation.
Rate of phospholipid synthesis
Having found that during the first 20–40 min of incubation, the amount of DAG was increased (Figure 3), and in order to determine if all the DAG was generated during the synthesis of IPCs, the rate of phospholipid synthesis and DAG mass levels were determined. G0/G1 yeast cells were incubated in CSM containing 20 μCi/ml [32P]Pi for 30–60 min and phospholipid labelling was determined. From the knowledge of the specific radioactivity of [32P]Pi in the culture medium and from the absorbance of the culture at the end of the pulse period, the rate of incorporation was calculated. It was found that during the first 30 min of incubation, the rate of phospholipid synthesis was 1450±100 pmol/107 cells and after 60 min it was 2636±300 pmol/107 cells. Considering that 20–23% of the radioactivity was in the sphingolipids, 290–323 and 535–606 pmol of IPC/107 cells was synthesized during the first 30 and 60 min respectively of incubation. The DAG mass levels were 300±30 and 350±50 pmol/107 cells respectively, and considering that 1 mol of DAG was generated for every 1 mol of IPC synthesized, the synthesis of sphingolipids was sufficient to account for the DAG generated during growth re-initiation. A similar amount of DAG (409±48 pmol/107 cells) was determined during the first 40 min of incubation after the release of N2 arrest (Figure 5).
During growth re-initiation (G0/G1cells), a clear correlation between sphingolipid synthesis (Figure 2) and DAG generation (Figure 3) was observed in coincidence with a decrement in DHS and PHS (Table 1). We also found that the inhibition of sphingolipid synthesis with β-chloroalanine decreased the generation of DAG, suggesting that these two processes were related. The study of G1 cells showed that all these processes take place during the G1 to S transition. The decrease in sphingolipid synthesis with β-chloroalanine and myriocin, inhibitors of SPT, as well as the inhibition of IPC synthase with Aur-A, blocked the G1 to S transition (Table 2) and DAG generation (Figure 5). The restoration of the G1 to S transition by the addition of PHS indicated that SPT inhibition by myriocin was specific. We previously observed that addition of DHS to β-chloroalanine-exposed Madin–Darby canine kidney cells restored both the generation of DAG and cell proliferation  and it has been reported that PHS restored growth of myriocin-exposed S. cerevisiae cells . In S. cerevisiae cells during growth re-initiation, long-chain bases and ceramide levels were decreased (Tables 1 and 3 and Figure 5), whereas the generation of DAG was increased (Figures 3 and 5), and these are adequate conditions for the activation of the putative PKC. The role of DAG in the activation of Pkc-1p remains a controversial point. Thus, in contrast to the mammalian system, in vitro Pkc-1p was completely inactive, and a partially purified Pkc-1p was independent of cofactors (Ca2+, DAG and PS) . Nevertheless, it has been found that the sites for binding of DAG are highly conserved, and the study of mutants affected in the putative DAG-binding site of S. cerevisiae PKC indicates that DAG may regulate Pkc-1p activity . In addition, increased DAG levels have been found in a yeast Cdc 28 mutant during the G1 to S transition  and we found that the DAG generated during sphingomyelin synthesis is involved in PKC activation in Madin–Darby canine kidney cells . Also, human PKC-ε was able to complement the growth defect caused by deletion of PKC-1 in S. cerevisiae, and PKC-ε is DAG-dependent . Therefore it is possible that suitable in vitro conditions that determine PKC-1 activity in S. cerevisiae extracts have not yet been found. Considering all the above-mentioned results, it seems that DAG generation during the synthesis of sphingolipids is a prerequisite for the transit of yeast cells through the cell cycle and probably of PKC-1 activation and proliferation. The contribution of IPC synthesis to the generation of DAG was also shown when the synthesis was inhibited at the site of IPC synthase with Aur-A (Figure 5). S. cerevisiae cells have been shown to display a cell cycle-dependent pathway involving the hydrolysis of PC to DAG and choline phosphate during the G1 to S transition . However, it is unlikely that the concentrations of β-chloroalanine, myriocin and Aur-A that decreased the levels of DAG were inhibiting PC-phospholipase C. Considering that the increments observed in choline phosphate coincided with decrements in labelled glycerophosphocholine  and that PC must be turned over for the synthesis of PC in the absence of external choline , it seems that the increase in labelled choline phosphate might be related to PC turnover.
The accumulation of DHS and PHS during the late exponential phase and ceramide during the stationary phase of growth may participate in the processes that cause yeast cells to enter and remain in the stationary phase of growth, and deserve further investigation. All the above-mentioned results suggested that sphingolipids act as signalling molecules in S. cerevisiae during unstressed conditions also, and this signalling process might be preserved from yeast to mammals. The observations reported in the present and the previous studies  indicate that sphingolipid and glycerolipid metabolism and signalling are coupled or interrelated. A perturbation in the DAG/ceramide ratio may lead to changes in cell growth and viability. Under growth-inhibited conditions (N2-arrested cells), the DAG/ceramide ratio was 0.26, and it changed to 2.0 during growth re-initiation (first 40 min). To our knowledge, this is the first evidence indicating that DAG generated during IPC synthesis participates in signalling and probably PKC activation. The pool of DAG analysed is derived from PI but not for activation of catabolic pathways. Considering that there is no evidence for degradation of IPC, MIPC and M(IP)2C in S. cerevisiae  (for a review see ), the observed accumulation of bioactive intermediates can be considered as being due to sphingolipid synthesis de novo. The role of ceramide, sphingoid bases and their derivatives under stressed conditions has been reviewed [23,33–35]. A paper that appeared just as the present study was completed showed that down-regulation of IPC synthase decreased DAG levels, and that the recombinant Pkc-1 protein of Cryptococcus neoformans transformed in S. cerevisiae was activated by DAG .
We thank Dr A. Garcia-Sainz and Dr J. L. Reyes for critically reviewing the manuscript. This work was supported by a Consejo Nacional de Ciencia y Tecnologia (CONACYT) grant 211085-5-26404 and fellowships to D.S.-C. and A.F.
Abbreviations: Aur-A, aureobasidin A; CSM, complete synthetic medium; DAG, diacylglycerol; DHS, D-erythrosphingosine; IP3, inositol trisphosphate; IPC, inositol phosphoceramide; M(IP)2C, mannose (inositol phosphate)2 ceramide; MIPC, mannose inositol phosphoceramide; PC, phosphatidylcholine; PHS, phytosphingosine; PI, phosphatidylinositol; PKC, protein kinase C; PLD, phospholipase D; PS, phosphatidylserine; SPT, serine:palmitoyl-CoA transferase
- The Biochemical Society, London