Thermolabile CDP-Choline Synthetase in an Animal Cell Mutant Defective in Lecithin Formation*

In previous studies, we have reported the isolation of a Chinese hamster ovary cell mutant (strain 58), which is temperature-sensitive for growth and defective in the biosynthesis of phosphatidylcholine (Esko, J. D., and Raetz, C. R. H. (1980) Proc. Nutl. Acad Sci U. S. A. 77, 5192-5196). We now present a detailed biochemical and enzymatic analysis of strain 58 and several spontaneous temperature-resistant revertants. When mutant 58 is shifted from 33 to 40 “C, the level of phosphatidylcholine declines immediately, while all other major lipid species, including sphingomyelin, continue to be made for about 20 h. The selective inhibition of phosphatidylcholine synthesis is accompanied by a 5-10fold drop in the intracellular “CDP-choline fraction,” which in Chinese hamster ovary cells consists of about 70% CDP-choline and 30% dCDP-choline. In cell extracts, the specific activity of the CDP-choline synthetase (phosphocholine cytidylyltransferase) is reduced 15-100-fold, and the residual activity is thermolabile, arguing in favor of a structural gene mutation. Since the formation of CDP-choline and dCDP-choline is defective both in vitro and in uiuo, the same enzyme appears to be responsible for the synthesis of both metabolites. Two other enzymes of phosphatidylcholine synthesis (choline kinase and choline phosphotransferase) are present with virtually the same specific activity in mutant as in the parental line. Most spontaneous temperature-resistant revertants of mutant 58 regain a nearly normal phospholipid composition at 40 “C and have CDP-choline synthetase levels comparable to the parental cells. Taken together, the present work provides strong evidence that a single mutation, most likely in a structural gene, is responsible for the biochemical and phenotypic properties of strain 58.

1 Recipient of a Camille and Henry Dreyfus Foundation teacherscholar award. defined mutants altered in the enzymatic assembly of the phospholipid bilayer is beginning to provide some insight into these problems. Similarly, the use of somatic cell mutants to probe phospholipid metabolism in higher eukaryotic cells should prove rewarding.
Recently, we described a rapid colony screening assay for detecting Chinese hamster ovary cell mutants altered in phosphatidylcholine synthesis (3), the major phospholipid of mammalian cells (2). Without using enrichment techniques, one temperature-sensitive mutant (strain 58), conditionally defective in phosphatidylcholine formation, was identified among 20,000 colonies derived from a stock of mutagen-treated cells (3). In preliminary studies, this variant was shown to be defective in CDP-choline synthetase (3), an enzyme which may regulate phosphatidylcholine formation in some mammalian systems (4)(5)(6)(7)(8). Unlike yeasts or fungi (9-ll), the biosynthesis of phosphatidylcholine by methylation of phosphatidylethanolamine (12) does not appear to be quantitatively significant in many mammalian cells cultured in vitro (13)(14)(15).
In the present work, aimed at a precise biochemical characterization of mutant 58, we have studied the CDP-choline synthetase deficiency in more detail and have obtained evidence for a structural gene alteration. Analysis of the "CDPcholine fraction" extracted from CHO' cells has revealed that it consists of both the rib0 and the deoxyribo derivatives, and that the same enzyme is responsible for the formation of both substances. Finally, the isolation of spontaneous, temperatureresistant revertants demonstrates conclusively that a single genetic alteration accounts for the conditional lethality and the enzymatic defect of strain 58.

Cell Lines and Crouch Conditions.
Hlfant 58 was isolated from the CHO-K1 The abbreviation used is: CHO, Chinese hamster ovary.

RESULTS
Temperature Sensitivity and Viability of Mutant 58 a t 40 "C-Under permissive conditions (33 "C), mutant 58 grew at about 80% of the parental rate, doubling every 24-30 h. Upon being shifted to 40 "C, the cell mass increased 2-4-fold, depending on the lot of fetal calf serum used, and thereafter further growth was inhibited. Although cells of mutant 58 that were shifted to 40 "C retained greater than 80% viability for at least 30 h, visible cell lysis begin to occur after 36 h. Continuous incubation at 40 f 0.2 "C resulted in the complete inhibition of colony formation.
Rapid Selective Inhibition of Phosphatidylcholine Synthesis in Mutant 58 Shifted to 40 "C-As shown previously, the relative amount (expressed as a percentage of the total phospholipid) of phosphatidylcholine dropped 2-fold when mutant 58 was shifted to 40 "C for 20 h (3). To determine the absolute amount of each lipid and the time course of the change in lipid content, the experiments in Figs. 1 and 2 were carried out. For cells maintained at 33 "C, the amount of ' ' P in the total phospholipid fraction, as well as in each of the individual, separated species, rose in parallel with the growth curve (Figs. 1 and 2, solid lines). In those cultures shifted to 40 "C (the broken lines in Figs. 1 and 2), the amount of total phospholipid and of all individual species (except phosphatidylcholine) continued to increase at (or above) the rate observed at 33 "C for approximately 20 h. In contrast, the amount of phosphatidylcholine recovered per dish (Fig. 1, right panel) declined immediately upon shift to 40 "C, indicating a selective block in phosphatidylcholine formation. Beyond 20 h at 40 "C, the accumulation of all phospholipids ceased in mutant 58 (Figs. 1 and 2), despite the high cell viability observed between 20 and 30 h. At still longer times (greater than 30 h), the recovery of all phospholipids per dish began to decline at a rate consistent with normal phospholipid turnover, previously observed in dividing parental cells (19). The phospholipid composition of parental cells was not markedly altered by a shift from 33 to 40 "C, and the rate of accumulation of each phospholipid was identical with the rate of growth at the corresponding temperature (data not shown).
Since the phospholipids of mutant 58 continued to turn over during prolonged incubation at 40 "C, the chemically determined ratio of phospholipid to protein fell considerably. Close inspection of the radioactivity profiie in the CDPcholine region (located by the absorbance at 280 nm of the chemical standard added to the extract) revealed an excess of radioactivity in the leading shoulder of the peak (Fig. 3, both panels). A similar distribution was also observed for parental and mutant cells grown at 33 "C (data not shown). To determine the nature of this anomaly, all fractions containing radioactivity in the vicinity of the CDP-choline peak were pooled from parental cells grown at 33 "C, mixed with chemically synthesized [methyl-'HIdCDP-choline, and subjected to a second ion exchange chromatography on AG 1-X2 in the presence of borate, as described in the legend to Fig. 4. In this case, the previously added CDP-choline standard was again detected by absorbance at 280 nm (Fig. 4, broken lines), while the chemically synthesized [methyl-"HIdCDP-choline was located both by its absorbance (Fig. 4, broken lines) and its "H radiolabel (Fig. 4, open circles). Each fraction was also analyzed for the presence of I4C (from the choline-labeled cells). About 30% of the I4C migrated with the dCDP-choline standard, while the rest was recovered in the CDP-choline region ( Fig. 4, closed circles). The I4C and AZRO were exactly coincident in Fig. 4, strongly suggesting that the radioactivity in the leading edge of the CDP-choline peak of Fig. 3 represented dCDP-choline. This metabolite is apparently present in CHO cells at levels considerably higher than in rat liver (29). These results demonstrate that mutant 58 is defective both in the Phospholipid, phosphatidylcholine, and sphingomyelin contents of mutant 58 at 33 "C and after a shift to 40 "C. The mutant was incubated at 33 "C in the presence of 32Pi (2 pCi/ml) for several generations to label the phospholipids to constant specific radioactivity (19). The cells were then harvested and dispensed at 3 X lo5 cells/100-mm diameter dish in medium containing 32P1 at the same specific radioactivity. After incubation at 33 "C for 1 day, some cultures were shifted to 40 "C. At the indicated times, duplicate cultures were harvested with trypsin and centrifuged at 600 X g . , for 5 min to remove residual medium. The cell pellets were resuspended,  Table I, there was no significant difference between mutant and wild type in the specific activities of choline kinase and choline phosphotransferase, while CDP-choline synthesis was reduced over 100-fold at 40 "C in this experiment. The fact that the CDP-choline synthetase has the highest specific activity in the parental cells (Table I) is of no special significance, since the assay conditions for the choline kinase and phosphotransferase have not been optimized for CHO cells.
Further enzymatic studies revealed that the residual CDPcholine synthetase activity present in extracts of mutant 58 was thermolabile relative to that of wild-type cells. The Qlo of the parental enzyme was approximately 2.5 under the assay conditions employed, while the specific activity of the mutant and a portion was extracted by the Bligh-Dyer method with carrier phospholipid (18,19). The lipid extracts were analyzed by two-dimensional thin layer chromatography as described previously (19), and the radioactivity in each spot was quantitated by liquid scintillation spectrometry. The relative content of each phospholipid was used to calculate the total amount of each radioactive phospholipid in each culture dish. Solid lines refer to the mutant at 33 "C; broken lines to the mutant at 40 "C. Left panel: total phospholipid content; right panel content of phosphatidylcholine ( P C ) and sphingomyelin (SPH).

( C L ) , and phosphatidylglycerol ( P G ) ;
rightpanel content of phosphatidylethanolamine ( P E ) and phosphatidylserine (PS). IO0 was considerably higher at 33 "C than at 40 "C. Attempts to stabilize this residual enzymatic activity by addition of 10% glycerol, 10-30% sucrose, 0.1-0.3 M KCl, or 0.6 mg/ml of lysophosphatidylethanolamine were not successful. Alternative methods of cell lysis (Dounce homogenization or hypotonic lysis) resulted in the same pattern shown in Table I. Under the best conditions, the level of synthetase in the mutant was not higher than 7% of wild type at 33 "C and only 14% at 40 "C. In addition, the conversion of dCTP to dCDPcholine was similarly defective in such extracts. The parental enzyme utilized CTP and dCTP with approximately equal efficiency (data not shown).
The Effect of Cytidine Derivatives on the Mutant-In initial studies of mutant 58, we observed that the temperature-sensitive phenotype was accentuated somewhat when dialyzed fetal bovine serum was employed (data not shown). When we tested the ability of various compounds involved in phosphatidylcholine synthesis (choline, phosphorylcholine, CDP-choline, and dCDP-choline) to suppress the temperature-sensitive phenotype of the mutant, we discovered that CDP-choline at 30-100 ,UM permitted the cells to grow considerably longer at 40 "C. Similar results were obtained by supplementing the  Fig. 3) were pooled and lyophilized to remove residual formic acid. The sample was redissolved in water and adjusted to pH 9 with ddute sodium hydroxide. Approximately 10' cpm of authentic [methyl-'HIdCDP-choline (see under "Experimental Procedures") was added and the sample was chromatographed as described in the legend to Fig. 3, except that the eluting gradient consisted of 0.2 M ammonium formate (pH 9.4) and uniformly contained 5 mM sodium tetraborate. After chromatography, 1.0 ml of each fraction was analyzed by liquid scintillation spectrometry, and the amounts of 'H (0) and I4C (0) were quantitated. Next, 0.1 ml of 10 N HC1 was added to the remainder of each fraction, and the absorbance at 280 nm (pH 2) was measured (X). The ultraviolet light absorbing material which migrated with the authentic [methyZ-3H]dCDP-choline standard resulted from the low specific radioactivity of this material (see under "Experimental Procedures"). The recovery of radioactivity and carrier was greater than 95%. mutant with CTP, CDP, CMP, and cytidine. These findings suggested that the accentuation of the phenotype of the mutant in the presence of dialyzed serum may have been caused by the removal of cytidine (or cytidine nucleotides) from the serum. Interestingly, when experiments like those shown in Figs. 3 and 4 were carried out with cells growing in the presence of 30 PM cytidine, the levels of CDP-choline (plus dCDP-choline) were 3-4-fold higher when normalized to the cell density both in CHO-K1 and mutant 58. However, in the parental CHO-K1 line, this did not significantly alter the phospholipid composition, while in mutant 58, there was a partial restoration of the phosphatidylcholine content (data not shown). These results suggested that cytidine might be acting to raise the intracellular pool of CTP, which has been suggested to limit the rate of phosphatidylcholine synthesis in  and several temperature-resistant revertants Extracts were made from cells growing in mid-late exponential phase at 33 "C. The CDP-choline synthetase assay was previously optimized for CHO cell extracts (3). The choline kinase (26) and the choline phosphotransferase (28) were assayed under conditions described for rat liver. The error in duplicate determinations was approximately *lo%.  (8). Higher levels of CTP in the in vitro assay of CDP-choline synthetase (5-50 mM) did not elevate the specific activity of the mutant relative to that of the wild type, arguing against reduction in the affinity of the enzyme for CTP in the mutant.
Biochemical Properties of Spontaneous, Temperature-resistant Revertants of Strain 58-In order to determine whether the conditional lethality of mutant 58 and the reduced CDP-choline synthetase were the result of separate unrelated mutations, spontaneous temperature-resistant revertants were isolated. This was done by placing 1-3 x lo6 cells in a 100-mm diameter tissue culture dish in the presence of 10% dialyzed fetal bovine serum and incubating them at 40 "C for 9 days. The growth medium was replaced every 3-4 days, and under these conditions, the incidence of revertants that grew at approximately the parental rate was about 1 in lo6. Altogether, 6 colonies were repurified and shown to grow normally at 40 "C. Revertants 1-5, isolated in this way, regained nearly normal levels of CDP-choline synthetase in vitro, as shown for strains 1 and 4 in Table I. Revertant 6 had only 2-3 times more enzymatic activity than mutant 58, presumably because it was not a true revertant or represented a bypass mutation. Indeed, the Q l o of the synthetase in revertant 6 between 33 and 40 "C was negative, like that of mutant 58.
The phospholipid compositions of CHO-K1, mutant 58, and the three revertants are presented in Table 11. Revertants 1 and 4 are virtually the same as CHO-K1 in this regard, consistent with restoration of a nearly normal CDP-choline After incubation at 40 "C for 20 h, the phospholipids were extracted and analyzed as described previously (19). The abbreviations used are: PC, phosphatidylcholine; SPH, sphingomyelin; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PS, phosphatidylserine; PG, phosphatidylglycerol.

Strain
"Other" consists of phosphatidic acid, cardiolipin, and less than 1% lysophosphatidylcholine and lysophosphatidylethanolamine. synthetase (Table I). Revertant 6 contains an intermediate level of phosphatidylcholine, which presumably is sufficient for normal cell proliferation.

DISCUSSION
In this report, we have presented evidence that a single gene mutation can account for the temperature-sensitive phenotype, the conditional alteration of phosphatidylcholine metabolism, and the striking reduction of CDP-choline synthetase in mutant 58. The best evidence for a single genetic alteration is provided by the isolation of spontaneous temperature-resistant revertants that regain the biochemical characteristics of the parent. Since the residual CDP-choline synthetase activity in strain 58 is itself thermolabile in vitro, it is likely that mutant 58 is somehow altered in the structural gene for the enzyme. However, other possibilities, such as abnormal posttranslational modification, cannot be excluded completely, and will have to await the purification of the mutant enzyme. The fact that some temperature-resistant revertants, like revertant 6, regain higher levels of what still appears to be a mutant enzyme suggests that there may be regulatory as well as structural genes in this system.
The resolution of the choline-nucleotide fraction into rib0 and deoxyribo CDP-cholines in this system is of interest, since the levels of dCDP-choline in CHO cells appear to be much higher than in rat liver (29), in which they account at most for 5% of the total. Since the synthesis of both rib0 and deoxy CDP-cholines is defective in mutant 58, it is likely that the same enzyme is responsible for the formation of both metabolites. The biological significance of the deoxy compound is uncertain. It may simply reflect the presence of high levels of dCTP in the cytoplasm of rapidly metabolizing cells (29). Nonetheless, the selective use of dCDP-choline for the synthesis of a topographically distinct pool of phosphatidylcholine or for the formation of certain molecular species cannot be eliminated. Since sphingomyelin synthesis continues under conditions of CDP-choline and dCDP-choline depletion, it is very unlikely that the deoxy derivative has a special role in sphingolipid metabolism.
When parental CHO cells are supplemented with 30 p~ cytidine, the levels of CDP-choline and dCDP-choline rise 3-4-fold. In this setting, the phospholipid composition of parental cells is not altered. This observation is superficially incompatible with the suggestion that the availability of CTP is limiting for phosphatidylcholine synthesis (4, 8), at least in this system. However, we cannot exclude the possibility that the increased CDP-choline pool leads to an increased rate of phosphatidylcholine synthesis, which is offset by an increased rate of phosphatidylcholine degradation. The converse situation, i.e. increased phosphatidylcholine synthesis in response to increased phosphatidylcholine degradation, has been observed in muscle cell cultures treated with exogenous phospholipase C (5,30). Further analysis of the CHO system in this regard is clearly warranted.
An intriguing aspect of lipid metabolism in mutant 58 is shown in the results in Figs. 1 and 2. Although the inhibition of phosphatidylcholine synthesis in this mutant is selective and immediate, the synthesis of all other phospholipids does not cease until after 20 h of incubation under nonpermissive conditions. This occurs at a time when macromolecular synthesis is still continuing (3) and viability is high. One possibility is that phosphatidylcholine depletion leads to the inhibition of total phospholipid synthesis, perhaps at the level of the glycerol-3-phosphate acyltransferase or at the level of de novo fatty acid synthesis. This possibility could be examined in extracts prepared from mutant 58 after incubation under nonpermissive conditions for about 20 h. Some uses of strains like mutant 58 deserve mention. In conjunction with existing cytogenetic techniques, mutant 58 should facilitate the mapping of the CDP-choline synthetase gene on human chromosomes (31). Although mutant 58 is probably defective in a structural gene, the possibility of altered regulation or posttranslational modification cannot be entirely eliminated. This can be explored effectively through the isolation of additional mutants and the purification of the mutant enzyme. Furthermore, the well defined temperaturesensitive phenotype of mutant 58 may allow for the isolation of genomic DNA capable of restoring temperature resistance using the recently developed techniques of animal cell transformation (32,33). Finally, it will be of great interest to determine the relationship of the CDP-choline synthetase gene to other genes involved in phosphatidylcholine and phospholipid metabolism, as well as genes for sterol, triglyceride, and lipoprotein biogenesis, since control mechanisms unique to membrane assembly may emerge.