Turnover of phosphatidylcholine in Saccharomyces cerevisiae. The role of the CDP-choline pathway.

The regulation of phosphatidylcholine degradation as a function of the route of phosphatidylcholine (PC) synthesis and changing environmental conditions has been investigated in the yeast Saccharomyces cerevisiae. In the wild-type strains studied, deacylation of phosphatidylcholine to glycerophosphocholine is induced when choline is supplied to the culture medium and, also, when the culture temperature is raised from 30 to 37 degrees C. In strains bearing mutations in any of the genes encoding enzymes of the CDP-choline pathway for phosphatidylcholine biosynthesis (CKI1, choline kinase; CPT1, 1, 2-diacylglycerol choline phosphotransferase; PCT1, CTP:phosphocholine cytidylyltransferase), no induction of phosphatidylcholine turnover and glycerophosphocholine production is seen in response to choline availability or elevated temperature. In contrast, the induction of phosphatidylcholine deacylation does occur in a strain bearing mutations in genes encoding enzymes of the methylation pathway for phosphatidylcholine biosynthesis (i.e. CHO2/PEM1 and OPI3/PEM2). Whereas the synthesis of PC via CDP-choline is accelerated when shifted from 30 to 37 degrees C, synthesis of PC via the methylation pathway is largely unaffected by the temperature shift. These results suggest that the deacylation of PC to GroPC requires an active CDP-choline pathway for PC biosynthesis but not an active methylation pathway. Furthermore, the data indicate that the synthesis and turnover of CDP-choline-derived PC, but not methylation pathway-derived PC, are accelerated by the stress of elevated temperature.

The regulation of phosphatidylcholine degradation as a function of the route of phosphatidylcholine (PC) synthesis and changing environmental conditions has been investigated in the yeast Saccharomyces cerevisiae. In the wild-type strains studied, deacylation of phosphatidylcholine to glycerophosphocholine is induced when choline is supplied to the culture medium and, also, when the culture temperature is raised from 30 to 37°C. In strains bearing mutations in any of the genes encoding enzymes of the CDP-choline pathway for phosphatidylcholine biosynthesis (CKI1, choline kinase; CPT1, 1, 2-diacylglycerol choline phosphotransferase; PCT1, CTP:phosphocholine cytidylyltransferase), no induction of phosphatidylcholine turnover and glycerophosphocholine production is seen in response to choline availability or elevated temperature. In contrast, the induction of phosphatidylcholine deacylation does occur in a strain bearing mutations in genes encoding enzymes of the methylation pathway for phosphatidylcholine biosynthesis (i.e. CHO2/PEM1 and OPI3/PEM2). Whereas the synthesis of PC via CDP-choline is accelerated when shifted from 30 to 37°C, synthesis of PC via the methylation pathway is largely unaffected by the temperature shift. These results suggest that the deacylation of PC to GroPC requires an active CDP-choline pathway for PC biosynthesis but not an active methylation pathway. Furthermore, the data indicate that the synthesis and turnover of CDP-choline-derived PC, but not methylation pathway-derived PC, are accelerated by the stress of elevated temperature.
Two separate biosynthetic pathways lead to the production of phosphatidylcholine (PC) 1 in Saccharomyces cerevisiae. The stepwise methylation of phosphatidylethanolamine (PE) to PC occurs through the action of a pair of methyltransferases encoded by the CHO2/PEM1 and OPI3/PEM2 genes (Fig. 1B). The CDP-choline pathway for PC biosynthesis (in which CDP-choline is an intermediate) consists of three reactions catalyzed by choline kinase (Cki1p) as follows: CTP, phosphocholine cytidylyltransferase (Pct1p); and 1,2-diacylglycerol choline phosphotransferase (Cpt1p), respectively (Fig. 1B). The CDP-choline pathway requires free choline, which can be obtained through exogenous sources or as a result of the phospholipase D (PLD)mediated hydrolysis of PC. As in yeast, hepatocytes synthesize PC through both the methylation and the CDP-choline pathways, whereas other mammalian cell types produce only minor amounts of PC through the methylation of PE (1).
Although the regulation of PC biosynthesis, and phospholipid biosynthesis in general, has been well studied in S. cerevisiae (2), the regulation of phospholipid turnover has gathered much less attention. PC is known to be catabolized by B and D type phospholipases in S. cerevisiae. Phospholipase B (PLB) deacylates PC to produce glycerophosphocholine (GroPC) and free fatty acids, whereas PLD hydrolyzes PC to phosphatidic acid (PA) and choline (Fig. 1A). Three PLB genes (PLB1, PLB2, and PLB3) have been identified in Bakers' yeast (3,4). Although Plb1p appears to be primarily responsible for the GroPC released into the culture medium (3,4), the roles of the three PLB enzymes in intracellular GroPC production have not been analyzed previously. Only one PLD gene, PLD1 (SPO14), has been identified (5-7), although other PLD-encoding genes have been postulated (8,9). Pld1p has been shown to hydrolyze PC in vivo, to be activated when Sec14p (the phosphatidylinositol (PI)/phosphatidylcholine (PC) transfer protein) is inactivated (10,11), and to be required for sporulation (12).
The synthesis and turnover of PC are highly regulated in all eukaryotic cells. The importance of this regulation is illustrated by the many pathological and physiological events associated with alterations in PC metabolism. Blockage of the CDPcholine pathway, either genetically (44) or by using enzyme inhibitors (45), causes apoptosis in several mammalian cell types. Elevated levels of GroPC are found in Alzheimer's disease brain compared with normal controls (46,47). Oncogenic transformation results in alterations in the activity of enzymes involved in PC biosynthesis and turnover, thus affecting the cellular levels of Cho-P and GroPC (14,48,49). The existence of regulatory mechanisms that coordinate PC biosynthesis with PC hydrolysis has been suggested by a number of studies (13)(14)(15)(16)(17). In particular, overexpression of CTP:phosphocholine cytidylyltransferase (a key regulatory enzyme of the CDP-choline pathway) in COS cells (17), Chinese hamster ovary cells (16), or HeLa cells (15) results in increased PC biosynthesis and increased PC degradation to GroPC, thereby maintaining PC homeostasis. There is also evidence to suggest that GroPC produced through PC deacylation may inhibit lysophospholipid hydrolysis, providing a feedback mechanism for reducing the rate of phospholipid degradation (50).
Studies in S. cerevisiae indicate not only that phospholipid synthesis and turnover are highly integrated but that PC metabolism, in particular, plays a central role in the regulation of this metabolic network. Many of the phospholipid biosynthetic genes are transcriptionally repressed by the addition of inositol to the culture medium (38). Additionally, regulation in response to inositol availability requires ongoing PC biosynthesis (38), suggesting that a signal emanating from the metabolism itself is responsible for controlling the transcription of the coordinately regulated phospholipid biosynthetic genes. By using mutants blocked at various steps of PC biosynthesis and mutants exhibiting alterations in PC turnover, it has been shown that conditions that result in elevated PLD-mediated turnover of PC, and therefore, elevated levels of PA or a closely related metabolite cause constitutive expression of the phospholipid biosynthetic genes (10,11). These results and others indicate that delineation of the molecular mechanisms controlling phospholipid metabolism will require a detailed knowledge of the factors affecting all pathways for PC synthesis and degradation. In the present study we continue the analysis of PC metabolism, focusing on the deacylation of PC to GroPC. In particular, we have employed S. cerevisiae strains bearing mutations in the genes encoding enzymes of the CDP-choline and the methylation pathways for PC biosynthesis to examine how the route and extent of PC biosynthesis are coupled to the turnover of PC to GroPC.
Analysis of [ 14 C]Choline-labeled PC Turnover-Cells were grown to uniform labeling in medium containing 1 Ci/ml [ 14 C]choline (20 M), harvested in logarithmic phase, washed twice in fresh nonradioactive medium, and inoculated at A 600 ϭ 0.1 in medium containing 10 mM nonradioactive choline. At various times, 1-ml aliquots of the cultures were removed, and the cells were pelleted by centrifugation. The supernatant was saved as the "medium" fraction. The cell pellet was suspended in 0.5 ml of 5% trichloroacetic acid and incubated on ice for 20 min with occasional vortexing. Following centrifugation, the supernatant containing the acid extract was saved as the "intracellular water-soluble fraction" of the cell (20). The cell pellet was suspended in 0.5 ml of 1 M Tris buffer, pH 8, centrifuged, and the resulting supernatant was added to the intracellular water-soluble fraction, effectively neutralizing the acidic extract. The cell pellet was saved as the "membrane" fraction. The recovery of total counts (medium, membrane, and intracellular water-soluble fraction) at each time point was Ն90%.
Phosphatidylcholine Analysis-Phospholipids were extracted from the membrane fraction as described previously (21). Radiolabeled phosphatidylcholine was resolved by SG81 (Whatman) paper chromatography in chloroform:methanol:NH 4 OH (66:17:3) and the spot excised, and concentrations were determined by liquid scintillation counting. Alternatively, PC content was determined by subjecting the chromatogram to PhosphorImager (Molecular Dynamics) analysis.
Analysis of Choline-containing Water-soluble Metabolites-The choline-containing compounds (i.e. choline, choline-phosphate (Cho-P), and GroPC) in the medium and the intracellular water-soluble fractions were separated by either TLC (22) or ion-exchange chromatography (23), and concentrations were determined by liquid scintillation counting.

Determination of [ 14 C]Choline and [ 14 C]Methionine
Incorporation into PC-Cells were grown to mid-logarithmic phase in synthetic medium containing 1 mM choline. The cultures were split into two flasks; the A 600 of each was adjusted to 0.5, one flask incubated at 30°C and the other flask at 37°C. After 15 min of incubation for temperature adjustment, 1 Ci/ml of either [ 14 C]choline or [ 14 C]methionine was added to the cultures. At 20-, 40-, and 60-min intervals, 900-l aliquots of the cultures were removed to tubes containing 100 l of 50% trichloroacetic acid to kill the cells, and the tubes were incubated on ice for 30 min. Radiolabeled PC was extracted and quantified as described above.
Measurement of Water-soluble Metabolites by NMR-For cultures grown at 30°C in plus and minus 1.0 mM choline media, 1-liter cultures were grown overnight to mid-logarithmic phase (A 600 0.5-0.7). For cultures grown under temperature shift conditions, strains grown overnight at 30°C in 10 M choline were diluted into 500 ml of media (A 600 0.2) containing 10 mM choline at the requisite temperature of either 30 or 37°C and shaken for 6 h. To harvest the cells, the cultures were rapidly cooled in an ice bath, the cells collected by centrifugation, washed with water, and the wet paste placed at Ϫ70°C. The frozen cell paste was treated with 2 volumes of 1 N perchloric acid at 5°C, vortexed, and freeze-thawed 2 times (24). After centrifugation, the supernatant was neutralized with 2.5 N KOH, centrifuged to remove the precipitated KClO 4 , and the solution lyophilized. The dried samples were dissolved in 0.8 ml of 50 mM Tris, 10 mM EDTA, pH 8.0, plus 10% D 2 O as an internal lock source. 31 P NMR analysis was performed on 0.6 ml of the sample plus 20 l of triethylphosphate (TEP) as an internal standard for both chemical shift and calibration (final concentration, 0.16 mM). The 31 P NMR spectra were acquired using a 300-MHz Bruker DXR spectrometer with a multinuclear probe set to 121.5 MHz for phosphorus. The spectra were accumulated for 1 h to ensure sufficient signal and processed with integration using Bruker software. Metabolites were identified by comparison to known compounds and published chemical shifts (25).
Mass Spectrometry Analysis of PC-Strains grown to an A 600 of 0.5-0.7 in 25-50 ml of the indicated media were harvested, washed with water, and suspended in 5% trichloroacetic acid for 30 min on ice. The cells were harvested and washed two times with water, and 50 g of dimyristoylphosphatidylcholine was added to the cell pellet as a standard. Phospholipids were extracted as described (21), and the dried lipid extract was dissolved to a concentration of ϳ10 pmol of lipid species/l in CH 3 OH/CHCl 3 (2:1) containing 5 mM LiOH (26). Electrospray ionization-mass spectrometry (ESI-MS) (Finnigan LCQ) was used to determine the molecular species of the cellular PCs.

RESULTS
Previous reports have shown that PC in S. cerevisiae is hydrolyzed by PLD to produce PA and choline (10,11) and by PLB to produce GroPC and fatty acids (3,4,27). However, the factors controlling these routes of PC degradation are largely unstudied. Thus, we undertook a thorough analysis of the regulation of PC turnover in response to choline availability, temperature, and genetic perturbations of the biosynthetic routes for PC production.
PC Turnover in Wild-type Cells Grown at 30°C-Wild-type cells prelabeled with [ 14 C]choline, washed free of radioactivity, and chased with excess choline at 30°C ( Fig. 2A), exhibited a pattern of label movement indicative of PLD-mediated turnover. Radiolabel associated with PC slowly diminished, and a concomitant amount of label appeared in the medium, primarily as free choline (Table I, part B). The appearance of labeled choline in the medium suggested that the choline produced through PLD-mediated turnover exited the cell, as is known to occur in mammalian cells (28). Since S. cerevisiae rapidly incorporates free choline into PC via the CDP-choline pathway, the nonradioactive choline chase was necessary to dilute out the [ 14 C]choline produced through PC turnover and, thereby, obtain an accurate assessment of the turnover.
PC Turnover in Wild-Type Cells Grown at 37°C-In contrast to the PC turnover observed at 30°C, a wild-type strain uniformly labeled with [ 14 C]choline at 30°C and then shifted from 30 to 37°C in the presence of a cold choline chase exhibited massive PC degradation (Fig. 2B). The label lost from PC appeared in the intracellular fraction, primarily as GroPC (Table I, part A). Since there is no evidence for synthesis of GroPC de novo in any eukaryotic cell (29,30), the shift from 30 to 37°C is most likely to have induced the deacylation of PC to GroPC. These patterns of PC turnover occurred in all three wild-type strains that were tested (data not shown).
Effect of CDP-choline Pathway Mutations on PC Turnover at 37°C-A mutant lacking choline kinase (cki1⌬), the first enzyme of the CDP-choline pathway for PC biosynthesis, incorporates [ 14 C]choline into PC at 10% of the wild-type level (10). Even though the cki1⌬ mutant lacks choline kinase, the action of an ethanolamine kinase that has some specificity for choline (31) can explain the minor degree of incorporation found. This mutant, which was used to monitor the effect of decreased production of CDP-choline pathway-derived PC, exhibited a pattern of radiolabel distribution different than that of the wild type. Taken as a percentage of total choline-associated radio- label, the cki1⌬ mutant contained less intracellular GroPC than wild type at 30°C, and neither PC degradation nor GroPC levels increased at 37°C when 10 mM choline was present (Table II). One interpretation of this result is that Cki1p itself is required for cells to induce GroPC formation in response to temperature elevation. Alternatively, the CDP-choline pathway, in general, may be necessary for the response. To distinguish between these possibilities, a cpt1⌬ deletion mutant, which lacks the enzyme responsible for the final step of the CDP-choline pathway for PC biosynthesis (Fig. 1), but which can incorporate [ 14 C]choline label into PC due to the action of an ethanolamine phosphotransferase (31), was also analyzed. The cpt1⌬ mutant, like the cki1⌬ mutant, was unable to induce PC deacylation when shifted from 30 to 37°C (data not shown), suggesting that it is a functional CDP-choline pathway that is required to induce PC deacylation upon temperature elevation. The pct1⌬ mutant could not be analyzed in this manner, as it incorporated no [ 14 C]choline label into PC.
Effect of Choline Supplementation and Temperature Elevation on GroPC Levels in a Wild-type Strain as Analyzed by 31 P NMR Spectroscopy-To explore further the role of the CDPcholine pathway for PC biosynthesis in PC deacylation, 31 P NMR spectroscopy was employed, as it overcomes the drawbacks associated with radiolabeling cells with [ 14 C]choline.
[ 14 C]Choline labeling monitors only the PC derived from the CDP-choline pathway and not the PC derived from the methylation of PE. Furthermore, the labeling of CDP-choline pathway mutants with [ 14 C]choline is necessarily limited, as the incorporation of choline into PC depends upon the CDP-choline pathway. Finally, following [ 14 C]choline-labeled PC turnover requires a cold choline chase to prevent the labeled metabolites from being rapidly recycled into PC. 31 P NMR spectroscopy provides a quantitative method of measuring the intracellular levels of all phosphate-containing metabolites produced through PC turnover and does not require the presence of exogenous choline. A typical 31 P NMR spectrum is shown in Fig. 3.
By using 31 P NMR spectroscopy, the effects of both choline supplementation and temperature elevation on the intracellular concentration of GroPC were analyzed. As shown in Fig. 4, a wild-type strain grown in the presence of exogenous choline exhibited ϳ7-fold more intracellular GroPC than did the same strain grown in the absence of choline supplementation. In addition, a wild type strain shifted from 30 to 37°C in cholinecontaining medium exhibited a further 4-fold increase in intracellular GroPC (Fig. 5), in general agreement with the results obtained through radiolabeling (Fig. 2 and Tables I and II).
Effect of Choline Supplementation and Temperature Eleva-

Phosphatidylcholine Turnover in Yeast tion on GroPC Levels in CDP-choline Pathway Mutants as
Analyzed by 31 P NMR Spectroscopy-The mutants (cki1⌬, pct1⌬, or cpt1⌬) bearing deletions in genes encoding the CDPcholine pathway enzymes exhibited no increase in GroPC levels in response to choline supplementation when examined by 31 P NMR spectroscopy (Fig. 4). The pct1⌬ and cpt1⌬ strains did, however, display elevated levels of Cho-P, as might be expected for strains blocked in the reaction steps subsequent to the formation of Cho-P by choline kinase. Cho-P levels were highest in the presence of choline, where increased substrate for the CDP-choline pathway could lead to the build up of an intermediate, such as Cho-P, if Pct1p or Cpt1p were absent. In addition to choline supplementation, the CDP-choline pathway mutants were also largely unresponsive to temperature elevation. The cki1⌬ and pct1⌬ strains exhibited no increase in GroPC levels when shifted from 30 to 37°C in the presence of choline (Fig. 5). Although the cpt1⌬ strain did exhibit an increase in GroPC concentration upon shift to 37°C, the total amount produced under these circumstances was ϳ8-fold less than that seen in the wild-type strain (Fig. 5). The pct1⌬ and cpt1⌬ strains also exhibited increased levels of Cho-P at both 30 and 37°C, consistent with their being blocked in the reaction steps subsequent to the formation of Cho-P by choline kinase. 31 P NMR Spectroscopy-Because disruption of the CDP-choline pathway resulted in the inability of the cell to induce PC deacylation in response to choline supplementation or temperature elevation, we next investigated the effect of blocking the alternate route for PC biosynthesis, the stepwise methylation of PE. A cho2⌬ opi3⌬ mutant is blocked in all three methylation steps converting PE to PC, making it a choline auxotroph. Interestingly, a cho2⌬ opi3⌬ mutant contained more intracellular GroPC than the congenic wild type when both were grown in choline-containing medium at 30°C (Fig.  6A). The amount of GroPC in the cho2⌬ opi3⌬ mutant increased ϳ2-fold when the temperature was shifted to 37°C (Fig. 6B). Although the wild-type strain congenic to the cho2⌬ opi3⌬ mutant (Fig. 6A) had roughly half as much GroPC at 30°C as did the wild-type strain shown in Fig. 4, the pattern of PC deacylation in response to temperature shift was the same, with both strains exhibiting a 3-4-fold increase in GroPC when the temperature was raised to 37°C. Since wild-type strains with different genetic backgrounds can exhibit differences in lipid composition (32), it is not surprising that the levels of  phospholipid metabolites, such as GroPC, can also vary. The cho2⌬ opi3⌬ mutant exhibited roughly 10-fold higher levels of GroPE than any other strain used in these studies both at 30 and 37°C (Fig. 6). As expected for a strain that cannot carry out the methylation steps subsequent to PE formation (Fig. 1), a cho2opi3 mutant was shown previously to contain elevated levels of PE compared with wild type (33). Thus, the build up of GroPE seen in the cho2⌬ opi3⌬ mutant may be a reflection of increased levels of its catabolic precursor, PE.

Effect of Temperature Elevation on GroPC Levels in a Methylation Pathway Mutant Grown in Choline-containing Medium as Analyzed by
Increased PC Turnover May be Coupled to an Increase in PC Synthesis upon Temperature Elevation-The increased PC turnover observed in a wild-type strain upon shifting from 30 to 37°C may be a reflection of an increase in PC biosynthesis. To test this hypothesis, the incorporation of [ 14 C]choline and [ 14 C]methionine into the PC of wild-type cells was measured at 30 and 37°C (Fig. 7). Approximately 2-fold more [ 14 C]choline was found to be associated with PC at 37 than at 30°C following 20, 40, and 60 min of labeling. In contrast, the incorporation of [ 14 C]methionine into PC was relatively unaffected by temperature. This result suggests that the two pathways for PC biosynthesis are differentially affected by temperature elevation. Although Fig. 7 shows the data obtained for WT-A (the wild type congenic to cki1⌬, cpt1⌬, and pct1⌬), WT-C (the wild type congenic to cho2⌬ opi3⌬) was also analyzed with similar results (data not shown).
Analysis of PC Molecular Species by Mass Spectroscopy-Since the turnover and synthesis of CDP-choline-derived PC and methylation pathway-derived PC appeared to be differentially regulated, we hypothesized that their fatty acid contents might also differ. To test this possibility, the molecular species of PC were analyzed by mass spectrometry in pct1⌬ (which makes all of its PC via methylation of PE), cho2⌬ opi3⌬ (which makes all of its PC via the CDP-choline pathway), and their congenic wild-type strains (Table III). Although no drastic alteration in fatty acid composition among the strains was observed, the pct1⌬ strain appeared to have a somewhat higher percentage of C32:2 PC species (32%) at the expense of C32:1 PC species (26%) as compared with the wild-type strain (C32:2, 25%; C32:1, 35%).
Effect of PLB1, PLB2, and PLB3 Deletion Mutations on GroPC Production-The production of GroPC presumably occurs through the action of a PLB or a phospholipase A 1 or A 2 in conjunction with a lysophospholipase. Based on homology to a PLB from Penicillium notatum (34), three S. cerevisiae PLBencoding genes have been characterized, PLB1 (3), PLB2 (4), and PLB3 (4). The production of intracellular GroPC in a plb1⌬ mutant, a plb2⌬ mutant, a plb3⌬ mutant, and in a plb1⌬ plb2⌬ plb3⌬ triple mutant was indistinguishable from wild type when 14 C-labeled cells were chased with excess choline at both 30 and 37°C (data not shown). Thus, the PLB homologs identified to date (3,4) do not appear to be involved in the regulated PC deacylation reported here. Clearly other deacylating phospholipases must exist in yeast, although their corresponding genes have yet to be characterized. An example of potential phospholipase-encoding genes can be found using the Yeast Proteome Data Base (35), which identifies open reading frames of YDR444w, YJR107c, and YOR084w as having lipase/esterase motifs and ORF YNL040w as having a PLA 2 active site signature (36,37). DISCUSSION PC turnover in S. cerevisiae varies as a function of the environmental conditions in which the cells are grown and as a function of the route by which the PC is synthesized. We demonstrate that wild-type S. cerevisiae cells grown in the presence of exogenous choline exhibit severalfold greater concentration of GroPC than cells grown in the absence of exogenous choline (Fig. 4). Furthermore, we demonstrate that a strain bearing a deletion mutation in any gene encoding an enzyme of the CDP-choline pathway for PC biosynthesis (cki1⌬, pct1⌬, or cpt1⌬) is unable to induce PC deacylation and generate GroPC in response to choline availability (Fig. 4). In contrast, a disruption in the alternate route for PC biosynthesis, the stepwise methylation of PE, has the opposite effect. A cho2⌬ opi3⌬ mutant grown in the presence of choline exhibits roughly 4-fold more GroPC than does its congenic wild type (Fig. 6). Temperature also affects PC turnover, as wild-type cells shifted from 30 to 37°C in choline-containing medium exhibit rapid PC deacylation, resulting in intracellular GroPC levels roughly 4-fold greater than those observed at 30°C (Figs. 2 and 5 and Table I). However, cki1⌬ and pct1⌬ strains are unable to induce PC deacylation in response to temperature elevation (Table II and Fig. 5), and the cpt1⌬ strain exhibits only a modest increase. Disruption of the methylation pathway, again, leads to a different result, as shown by a cho2⌬ opi3⌬ mutant that exhibits roughly 2-fold more GroPC when shifted from 30 to 37°C (Fig. 6). An analysis of PC biosynthesis in response to temperature elevation indicates that although the CDP-choline pathway for PC biosynthesis is induced at 37°C, the methylation pathway is not (Fig. 7). Finally, mass spectrometry analysis (Table III) indicates that there are no major differences in the fatty acyl composition of PC generated via the CDP-choline and methylation pathways. These findings suggest two major conclusions as follows. (i) The catabolism of PC to GroPC is dependent upon an active CDP-choline pathway for PC biosynthesis but not an active methylation pathway. In fact, disruption of the methylation pathway for PC synthesis seems to enhance the deacylation of both PC and PE in cells grown at 30°C. (ii) The stress of temperature elevation accelerates the synthesis and turnover of CDP-choline derived PC but not methylation-derived PC.
One explanation for the observed increase in GroPC that occurs upon choline supplementation in S. cerevisiae (as well as in cultured rat hepatocytes (13)) is that having a functional CDP-choline pathway in the presence of exogenous choline results in excess biosynthesis of PC that must be counterbalanced by increased PC degradation. Since exogenous choline would only increase PC synthesis in strains having a functional CDP-choline pathway, the finding that choline supplementation does not increase PC deacylation in CDP-choline pathway mutants is consistent with this hypothesis (Fig. 4). Also in agreement with this interpretation is the finding that overexpression of CTP:phosphocholine cytidylyltransferase (the ratelimiting step of the CDP-choline pathway) in a number of mammalian cell types (15)(16)(17) results in increased PC biosynthesis and increased PC degradation to GroPC. In extending these studies to yeast, we demonstrate that the presence of exogenous choline only results in increased PC deacylation if it is successfully converted to PC. Mutants blocked at any step of the CDP-choline pathway (giving rise to the accumulation of pathway intermediates) do not exhibit increased PC catabolism when choline is added.
Temperature elevation from 30 to 37°C also causes wildtype yeast cells to undergo increased PC deacylation and concomitant GroPC production. Although strains bearing mutations in the CDP-choline pathway for PC biosynthesis do not produce increased GroPC upon temperature elevation, a strain blocked in the methylation pathway does. Furthermore, temperature elevation causes an increase in PC biosynthesis via the CDP-choline pathway but not through the methylation pathway. These results are also in agreement with the hypothesis that conditions that cause an increase in PC biosynthesis result in an increase in PC deacylation. According to this interpretation of the data, the CDP-choline pathway mutants do not experience an increase in PC deacylation upon shifting to 37°C, because there is no increase in PC biosynthesis upon shifting to 37°C. This interpretation also implies that the cell has a means of monitoring its PC levels and using that information to induce PC turnover when appropriate.
Changes in phospholipid metabolism associated with temperature elevation have been well documented (24,33,39,40). It has long been speculated that such changes are due to a need to alter the cellular phospholipid composition to one more favorable to higher temperature. If that hypothesis is correct, it is indeed surprising that only the CDP-choline pathway is affected by temperature, since the cell would presumably also need to alter its phospholipid composition when PC is being made primarily through the methylation pathway. Furthermore, if temperature elevation leads to the activation of a phospholipase(s) that catalyzes PC deacylation, our data would  III Analysis of PC species by mass spectrometry Strains were grown to mid-logarithmic phase in 1 mM choline-containing medium. Phospholipids were extracted (21), and the dried lipid extract was dissolved in CH 3 OH/CHCl 3 (2:1) containing 5 mM LiOH (26). Electrospray ionization mass spectrometry (Finnigan LCQ) was used in the positive ion full scan mode over the mass range, 450 to 1800 atomic mass units, to determine the PC species. Data are the average Ϯ S.D. of at least two independent experiments. The m/z [M ϩ Li] ϩ for the lithiated species observed were as follows: C32:2, 736.6; C32:1, 738.6; C34:2, 764.5; and C34:1, 766.5. Since the major acyl groups esterfied to yeast phospholipids are 16:0, 16:1, 18:0, and 18:1 (32), the likely fatty acyl compositions of each PC species can be deduced. For example, C32:1 ϭ 1 C16:1 and 1 C16:0. suggest that methylation pathway-derived PC is not a substrate for those phospholipases. The insensitivity of methylation pathway-derived PC to deacylation upon temperature shift could be explained if the PC generated by these two pathways differs in some regard, perhaps in their subcellular location or in the composition of their fatty acyl chains. Since neither the distribution of CDPcholine derived versus methylation pathway-derived PC is known nor the location of the phospholipase(s) involved, the importance of subcellular location is currently impossible to address. With regard to fatty acid composition, we report (Table III) that a small shift toward C32:2 PC species at the expense of C32:1 PC species occurs in the cpt1⌬ mutant (which makes its PC solely through the methylation pathway). It is unclear whether this rather minor change contributes to the drastic difference in deacylation sensitivity observed between PC generated by the two pathways. Waite and Vance (41) recently reported finding only negligible differences in the fatty acid composition of PC produced through the methylation and CDP-choline pathways in a Chinese hamster ovary cell line. In contrast, DeLong et al. (42) report that in rat primary hepatocytes, CDP-choline-derived PC is composed mainly of medium chain saturated species, whereas methylation pathway-derived PC is more diverse and contains more polyunsaturated species.
Previous studies have suggested that the pathways for PC biosynthesis in S. cerevisiae are not functionally identical. Mutations in the genes (CHO2 or OPI3) encoding the phospholipid methyltransferases that convert PE to PC, but not mutations in genes encoding enzymes of the CDP-choline pathway for PC biosynthesis, result in an overproduction of inositol (OpiϪ) phenotype. An OpiϪ phenotype indicates overexpression of INO1, which encodes inositol-1-phosphate synthase, the ratelimiting step of inositol biosynthesis (38). Next, mutations in the CDP-choline pathway, but not the methylation pathway, suppress the temperature-sensitive phenotype of a sec14 ts mutation (43).
PC is the major phospholipid found in eukaryotic cells and the precursor of various signaling molecules such as PA and diacylglycerol. The regulation of PC metabolism is crucial to cell function, as evidenced by the existence of physiological and pathological events (apoptosis (44,45) and oncogenic transformation (14,48,49), for example) associated with alterations in PC metabolism. Future studies will focus on the identification of the phospholipase(s) involved in PC deacylation and on the determination of how the cell senses the amount of its PC.