Regulation of phosphatidylcholine biosynthesis in mammalian cells. I. Effects of phospholipase C treatment on phosphatidylcholine metabolism in Chinese hamster ovary cells and LM mouse fibroblasts.

Addition of phospholipase C from Clostridium perfringens to cultures of Chinese hamster ovary (CHO) cells resulted in rapid degradation of cellular phosphatidylcholine with concomitant release of phosphocholine. The rate of incorporation of radiolabeled choline into lipids was increased 2-fold in phospholipase C-treated CHO cells as compared to untreated controls. The only enzyme in the pathway of phosphatidylcholine biosynthesis with increased activity in phospholipase C-treated cells was CTP:phosphocholine cytidylyltransferase, indicating that the cytidylyltransferase plays an important role in the stimulation of phosphatidylcholine biosynthesis. The phospholipase treatment was toxic to a CHO mutant cell line with abnormally low cytidylyltransferase activity. Mouse LM fibroblasts were resistant to enzymatic attack by phospholipase C, and cytidylyltransferase activity in LM cells did not change upon phospholipase C treatment.

Addition of phospholipase C from Clostridium perfringens to cultures of Chinese hamster ovary (CHO) cells resulted in rapid degradation of cellular phosphatidylcholine with concomitant release of phosphocholine. The rate of incorporation of radiolabeled choline into lipids was increased 2-fold in phospholipase Ctreated CHO cells as compared to untreated controls. The only enzyme in the pathway of phosphatidylcholine biosynthesis with increased activity in phospholipase C-treated cells was CTP:phosphocholine cytidylyltransferase, indicating that the cytidylyltransferase plays an important role in the stimulation of phosphatidylcholine biosynthesis. The phospholipase treatment was toxic to a CHO mutant cell line with abnormally low cytidylyltransferase activity.
Mouse LM fibroblasts were resistant to enzymatic attack by phospholipase C, and cytidylyltransferase activity in LM cells did not change upon phospholipase C treatment.
The regulation of the synthesis of phosphatidylcholine, the principal mammalian phospholipid, is currently being studied under a variety of developmental and physiological conditions (1)(2)(3)(4) and by genetic manipulations (5, 6). Using a system in which cultured embryonic chick muscle cells were treated with phospholipase C from Clostridium perfringens, we discovered that phosphatidylcholine synthesis in these cells was subject to a kind of feedback regulation (7, 8). The phospholipase treatment increased the rate of degradation of cellular phosphatidylcholine and concomitantly increased the rate of phosphatidylcholine synthesis. Based on analyses of the activities of enzymes in phosphatidylcholine synthesis as well as levels of metabolic intermediates, it was concluded that the CTP:phosphocholine cytidylyltransferase was regulatory under the experimental conditions. We proposed a model for the role of the cytidylyltransferase in regulating phosphatidylcholine synthesis in which the phospholipid composition of cellular membranes can determine the activity of the cytidylyltransferase (8).
To determine if the cytidylyltransferase plays such a regulatory role in cells other than embryonic chick muscle, it was desirable to extend these studies to mammalian cell lines that were already commonly used for studies on lipid metabolism. CHO' cells have been frequently used for studies on fatty acid, phospholipid, and cholesterol metabolism, and have been widely used for genetic studies.
The fist mammalian cell variants with specific lesions in phospholipid biosynthesis have been isolated from CHO cells (5). LM cells, a mouse fibroblast cell line, also appeared attractive for these studies both because of the common use of this cell line in lipid research and because LM cells can be grown in a chemically defined medium (9)(10)(11). In this paper, the effects of phospholipase C treatment on phosphatidylcholine metabolism in both CHO and LM cells are reported. Subsequent papers in this series describe the effects of phospholipase C treatment and large alterations in phospholipid composition on the activity and subcellular distribution of the cytidylyltransferase.

RESULTS
Degradation of Phosphatidylcholine in Phospholipase Ctreated CHO cells-Phospholipase C-mediated degradation of phosphatidylcholine was determined by measuring the depletion of prelabeled phospholipid during a chase in nonradioactive medium. The loss of [32P]phosphatidylcholine from phospholipase C-treated CHO cells was rapid (Fig. 1). Approximately 50% of the total phosphatidylcholine was degraded within 4 to 8 h of phospholipase treatment. To determine if phospholipase C treatment resulted in degradation of phosphatidylcholine or simply release of the lipid into the culture medium, the chase medium was examined for 32Plabeled compounds 3 h after phospholipase addition. Less than 0.5% of the radioactivity lost from the cells was recoverable in the medium as phosphatidylcholine, but greater than 70% of the label lost from the cells could be accounted for by [32P]phosphocholine in the medium. These data strongly suggest that addition of phospholipase C to the medium of CHO cell cultures results in degradation of cellular phosphatidylcholine with concomitant release of phosphocholine into the culture medium. This effect is similar to that seen with primary muscle cells ( 7 ) .
The abbreviation used is: CHO, Chinese hamster ovary. Portions of this paper (including "Materials and Methods," Figs. 4,5, and 9, and   The total phospholipid content of CHO cells incubated with phospholipase C for 3 h was somewhat lower than that of untreated cells (2.6 nmol/pg of DNA versus 3.6 nmol/pg of DNA). If, however, the phospholipase treatment was extended for 24 h there was no difference in phospholipid content between control and phospholipase-treated cells. The phospholipid composition of control and phospholipase C-treated cells was nearly identical (Fig. 2). These data suggest that the rates of synthesis of phosphatidylcholine must be increased in the phospholipase C-treated cells to compensate for the increased rate of phosphatidylcholine degradation.

Stimulation of Phosphatidylcholine Synthesis in CHO
Cells by Treatment with Phospholipase C-The effect of treatment with phospholipase C on the synthesis of phosphatidylcholine was investigated by measuring incorporation of isotopic precursors into phosphatidylcholine. Methylation of phosphatidylethanolamine did not contribute significantly to the synthesis of phosphatidylcholine since, after incubation of CHO cell cultures with ['Hlethanolamine for either 3 or 24 h, radioactivity was incorporated only into phosphatidylethanolamine as detected by two-dimensional thin layer chromatography and fluorography. When CHO cells were incubated with ['Hlcholine for 5 h, greater than 95% of the lipid-soluble radioactivity was associated with phosphatidylcholine and the remainder with sphingomyelin.
Time courses for incorporation of radiolabeled choline into phospholipid are shown after a short term (Fig. 3A ) or long term ( Fig. 4) incubation with phospholipase C. The rates of proximately twice the rates in the absence of the enzyme whether the isotope was added at the beginning of the incubation with the phospholipase (Fig. 3A ) or 24 h later (Fig. 4). The slow initial incorporation presumably reflects the time needed to completely label the intracellular pools of cholinecontaining precursors. The maximum stimulation of incorporation of either ['Hlcholine or '*Pi into phospholipids was observed with about 0.03 units of phospholipase C/ml of culture medium (Fig. 3B). Thus, the rate of production of phosphatidylcholine, as measured by the incorporation of choline into lipid, is increased in CHO cells treated with phospholipase C. The increased synthesis begins within 30 min after addition of the phospholipase and continues for at least 24 h.

Enzymes of Phosphatidylcholine Synthesis in Extracts of
Phospholipase C-treated CHO Cells-The enzymes involved in the synthesis of phosphatidylcholine were assayed to determine if any were enhanced in cells treated with phospholipase C ( Table I). The activities of enzymes that participate in the synthesis of the diacylglycerol moiety were not significantly altered in phospholipase C-treated cultures. Of the enzymes that participate in synthesis of the polar head group, choline kinase exhibited the lowest activity in vitro. The activity of choline kinase was not altered by phospholipase C treatment, suggesting that this enzyme was not controlling the stimulated incorporation of choline into phosphatidylcholine in these cells. Choline phosphotransferase activity also was not affected by phospholipase treatment. The only enzyme with significantly altered activity in phospholipase Ctreated cultures was CTP:phosphocholine cytidylyltransferase, the activity of which was increased 2-fold by phospholipase treatment. The increased activity of cytidylyltransferase suggests that this enzyme was responsible for the increased production of phosphatidylcholine in phospholipase C-treated CHO cells.
Effect of Phospholipase Treatment on Sizes of Zntermediate Pools-Regulatory enzymes can at times be identified by determining the levels of pathway intermediates as the rate of flux through the pathway is changed. That is, in certain cases a crossover point can be demonstrated where the level of substrate for the regulatory reaction decreases as the pathway flux increases (23). In this manner cytidylyltransferase was shown to regulate the increased incorporation of choline into phosphatidylcholine in phospholipase C-treated chick muscle cells (8). To determine if the levels of any metabolite in CHO

Relative concentrations of water-soluble choline-containing compounds and derivatives found in CHO cells
After passage, a stock of CHO cells was grown to confluency in medium containing 0.7 pCi/ml of ["C]choline. The cells were trypsinized from the flask, replated in the radiolabeled medium, allowed to grow for 2 days, and then harvested. At the appropriate times prior to harvest, phospholipase C at a final concentration of 0.050 unit/ml was added to some of the dishes. Choline, phosphocholine, CDPcholine, and betaine were separated as described under "Materials and Methods." cells changed as a result of phospholipase C treatment, cytosolic pools were labeled to a constant specific radioactivity and the amount of label in choline, phosphocholine, and CDPcholine was determined. The oxidation product betaine was also labeled under these conditions so it was also rlonitored. As shown in Table 11, treatment with phospholipase C for either 3 or 24 h resulted in a decrease in the pool sizes of both choline and phosphocholine as well as betaine. Phospholipasetreated cells contained approximately the same level of CDP- choline as untreated controls. The reason for the decreased pool sizes is currently unknown, but the decrease did not appear to be due to damage to the cells, as discussed below.
Viability of CHO Cells Treated with Phospholipase C-Several lines of evidence indicate that CHO cells treated with phospholipase C were healthy, functioning cells. 1) Greater then 98% of both phospholipase-treated and control cells excluded the vital stains trypan blue and eosin B. 2) Cells treated with phospholipase C for 3 h released no more lactate dehydrogenase activity into the medium than did control cells cpm/pg of DNA, respectively, were recovered in trichloroacetic acid-insoluble material. Protein synthesis, therefore, proceeded normally after phospholipase C treatment. 5) The growth rate of phospholipase C-treated cells was only slightly lower than control cells (Fig. 5) and CHO cells could be grown in the presence of 0.05 unit/ml of phospholipase C for several passages. Therefore, by several criteria, CHO cells were able to meet the challenge of the phospholipase-catalyzed destruction of cellular phosphatidylcholine while otherwise functioning normally.

Effect of Phospholipase C Treatment on a CHO Variant with a Defective Cytidylyltransferase-Esko and Raetz (5)
have isolated a temperature-sensitive CHO variant, strain 58, which is defective in phosphatidylcholine production. Evidence has been presented suggesting that strain 58 contains a mutation in the structural gene for CTP:phosphocholine cytidylyltransferase (6). We have assayed the cytidylyltransferase activities of both parental and strain 58 cells grown under our culture conditions (Table 111) and the results are in agreement with those of Esko et al. (6) in that the cytidylyltransferase activity of strain 58 was found to be lower than that of parental cells even at the permissive temperature, 33 "C. Moreover, while activity of the parental cytidylyltrans-  Lipids were extracted from the total homogenate and analyzed as described (13).

Radioactivity per dish in phosphatidylcholine (O), phosphatidylethanolamine (A), and in the remaining phospholipids (0) is presented. Data points are the averages of duplicate determinations.
ferase assayed a t 40 "C was considerably higher than that assayed at 33 "C, the strain 58 cytidylyltransferase activity was not increased at the higher temperature, suggesting a thermolabile enzyme (6).
T o test whether strain 58 cells were able to synthesize phosphatidylcholine faster in response to phospholipase C, parental and mutant cells were incubated at both 33 and 40 "C in the presence and absence of phospholipase C (Fig. 6). The phospholipase C treatment caused increased incorporation of ["Hlcholine into lipid of parental cells at both temperatures, as expected. Choline incorporation in strain 58 was increased in response to phospholipase C at 33 C. At 40 "C, however, choline incorporation in the mutant was reduced in the presence of phospholipase C, probably due to the rapid toxicity of the phospholipase treatment to the mutant cells a t 40 "C. While the morphology of the parental strain was unaffected by phospholipase C treatment (Fig. 7), the morphology of strain 58 was dramatically affected at both temperatures. Within 5 h after addition of phospholipase C at 40 C the strain 58 cells became rounded and began to detach from the dish. The same effect was observed with strain 58 cells a t 33 "C, but only after a more prolonged incubation with the phospholipase. This is consistent with the higher mutant cytidylyltransferase activity at the permissive temperature which would allow strain 58 to survive phospholipase C treatment for a longer period than at 40 "C. However, because the activity of strain 58 cytidylyltransferase at 33 "C was much lower than that of the parental strain, the mutant could not meet the continued demand for high levels of phosphatidylcholine synthesis and eventually died.
Effect of Phospholipase C on Phosphatidylcholine Metabolism in LM Cells-Phosphatidylcholine of LM cells was apparently inaccessible to phospholipase C action as demonstrated by the inability of this enzyme to degrade labeled phosphatidylcholine in LM cell cultures (Fig. 8). Moreover, phospholipase C treatment did not stimulate phosphatidylcholine synthesis in these cells (Fig. 9). The activity of CTP:phosphocholine cytidylyltransferase was measured in homogenates of LM cells grown in the presence and absence of phospholipase C. Specific activities of 0.56 k 0.07 and 0.61 * 0.12 nmol/min/mg of protein were found in cells grown in the absence and presence of the enzyme, respectively. As expected, when phospholipase C was unable to degrade cellular phospholipids there was no resulting stimulation of phosphatidylcholine synthesis and no activation of the cytidylyltransferase.

DISCUSSION
The results presented in this paper demonstrate that the mammalian cell line, CHO, is capable of increasing phosphatidylcholine synthesis in response to phospholipase C treatment as has been previously shown for chicken embryonic muscle cells (7). Phospholipase C stimulated degradation of CHO cellular phosphatidylcholine as shown by both a decrease in prelabeled cellular [32P]phosphatidylcholine and a concomitant increase in [32P]phosphocholine in the medium. Moreover, the biosynthesis of phosphatidylcholine was stimulated in cultures treated with the phospholipase C, as determined by the increased incorporation of radioactive choline into phospholipid. A possible reason for the faster rate of incorporation of labeled choline into lipid in cells treated with phospholipase C is that these cells contained a lower level of cytosolic choline-containing precursors and that the label was therefore not diluted as much as in control cells. This dilution effect could not have been the sole cause of the observed stimulated choline incorporation for two reasons. 1) When phospholipase C and radioactive choline were added simultaneously, stimulated choline incorporation was observed (Fig. 3). Because the pool sizes were the same in the control and phospholipase C-treated cells at the start of the experiment, the increased incorporation could not have been due to differential dilution of cytosolic pools. 2) The lag in incorporation of radioactive choline seen in Fig. 3 and 4 was presumably due to the time required to equilibrate the cytosolic pools to a constant specific radioactivity, after which the rate of incorporation would reflect the true rates of incorporation of choline into phosphatidylcholine. If one extrapolates the data in Fig. 4 back to zero choline incorporation, one can see that the cytosolic pools were, in fact, equilibrated faster after cells had been pretreated with phospholipase C. The stimulated incorporation, however, continued long after the equilibration phase, indicating that the rate of incorporation was truly stimulated in the phospholipase C-treated cells.
The only enzyme in the pathway of phosphatidylcholine biosynthesis affected by phospholipase C treatment of CHO cells was the CTP:phosphocholine cytidylyltransferase. An accompanying paper (24) demonstrates that the subcellular distribution of the cytidylyltransferase was altered by phospholipase C treatment. With embryonic chick muscle cells it was possible to demonstrate that the cytidylyltransferase was regulatory in phospholipase C-treated cells by showing that levels of phosphocholine decreased and CDP-choline increased in phospholipase C-treated cells, while the choline pool remained constant (8). This type of analysis has proved unsuccessful with CHO cells. The phosphocholine pool was the largest soluble choline-containing pool in CHO cells, which is consistent with the cytidylyltransferase being rate-limiting, and the level of phosphocholine decreased with phospholipase treatment. However, phospholipase treatment also resulted in decreased pools of choline and betaine. The reason for the decreased sizes of all three pools is presently unknown. A possible explanation is that when phosphatidylcholine synthesis is stimulated by phospholipase C treatment, the entry of choline into the cells becomes limiting.
The experiments with CHO strain 58 also provide additional evidence that the cytidylyltransferase activity is critical for the regulation of production of phosphatidylcholine synthesis in phospholipase C-treated cells. Strain 58 cells, which had reduced activity of cytidylyltransferase even at the permissive temperature for growth, could not withstand continued treatment with phospholipase C, indicating this enzyme must not be in great excess in untreated cells and is certainly insufficient in phospholipase C-treated mutant cells. A comparison of the choline incorporation data with cytidylyltransferase activities of parental and strain 58 cells reveals that choline incorporation is not strictly proportional to the activity of cytidylyltransferase measured in uitro. At 33 "C the ratio of parental to strain 58 cytidylyltransferase activity was 20, but the ratio of the amount of [3H]choline incorporated in 5 h in the absence of phospholipase C (Fig. 6) was only 3.5. In addition, parental cytidylyltransferase activity a t 40 "C was about 2fold higher than at 33 "C, but choline incorporation was only 20% higher at the higher temperature. It is possible, of course, that activity of cytidylyltransferase measured under maximal conditions in vitro does not reflect the activity of the enzyme under cellular conditions. It is also possible that the cytidylyltransferase reaction is not the only slow step in the pathway of phosphatidylcholine biosynthesis. The relative rates of choline kinase and cytidylyltransferase measured in uitro are quite different in muscle cells and CHO cells. In embryonic muscle cultures the cytidylyltransferase was the slower enzyme (8). In CHO cells, however, choline kinase activity was much lower than in muscle, making choline kinase the slower enzyme in uitro. Increased choline kinase levels have been shown to be responsible for increased rates of phosphatidylcholine synthesis in estrogen-stimulated rooster liver (27), indicating that choline kinase can influence the rate of the pathway, at least under certain conditions. It has also been predicted from measurements of intracellular metabolites that both choline kinase and cytidylyltransferase are rate-limiting in rat liver (28). Phosphatidylcholine synthesis may therefore have two slow steps, choline kinase and cytidylyltransferase. Which enzyme is the principal regulatory enzyme may depend on cell type and environmental conditions. It is possible that the cytidylyltransferase may be a "repair" enzyme and regulate increased phosphatidylcholine synthesis in response to membrane damage. Further evidence must be obtained to determine if the cytidylyltransferase is the sole rate-limiting enzyme in phosphatidylcholine biosynthesis or if both choline kinase and cytidylyltransferase may each be rate-limiting under different conditions. It was somewhat surprising the LM cells were resistant to phospholipase C digestion in that phosphatidylcholine in the LM plasma membrane is externally disposed (29). An obvious difference between the LM and CHO experiments was the presence of serum in the CHO cultures. This was probably not a critical factor, however, because serum does not influence the ability of phospholipase C to degrade HeLa cell phospholipids.3 It is also possible that an extensive glycocalyx surrounds the LM cells and prevents access of phospholipase C to the membrane phospholipids, but this possibility remains to be examined.