Regulation of Phosphatidylcholine Metabolism in Mammalian Cells ISOLATION AND CHARACTERIZATION OF A CHINESE HAMSTER OVARY CELL PLEIOTROPIC DEFECTIVE IN BOTH CHOLINE KINASE AND CHOLINE-EXCHANGE REACTION ACTIVITIES*

means of an in situ autoradiographic assay for the base-exchange reaction of phospholipids with L-serine in Chinese hamster ovary cell colonies immobilized activity of the 2-fold less than in the ATP:choline phosphotrans- (choline kinase) the enzyme catalyzes the base-exchange of phospholipids with choline (choline-exchange enzyme) 4-fold 10- respectively), the specific activities of other of phosphatidylcholine in a the

By means of an in situ autoradiographic assay for the base-exchange reaction of phospholipids with Lserine in Chinese hamster ovary cell colonies immobilized on filter paper (Esko, J. D. and Raetz, C. R. H. (1978) Proc. Nutl. Acad. Sei. U. S. A. 75,1190-1193), a mutant (designated 89.1) was isolated in which the specific activity of the serine-exchange enzyme was about 2-fold less than in the parent. Unexpectedly, it was demonstrated that in extracts of the mutant the specific activities of both ATP:choline phosphotransferase (choline kinase) (EC 2.7.1.32) and the enzyme that catalyzes the base-exchange of phospholipids with choline (choline-exchange enzyme) were strikingly reduced (3-to 4-fold and 10-to 15-fold, respectively), while the specific activities of other enzymes of phosphatidylcholine synthesis were normal. Several lines of evidence presented here suggested that the partial defect of serine-exchange activity in this mutant was due to a decrease of acceptor phospholipid(s) for the reaction. The growth rates and phospholipid compositions of the mutant and parent were quite similar. However, mutant 89.1 exhibited a significant defect in its ability in vivo to synthesize phosphatidylcholine. The fact that the mutant was also defective in phosphorylcholine biosynthesis in uiuo, together with the finding of an enzymatic lesion of the mutant in choline kinse i n vitro as described above, clearly demonstrated that with respect to the reduced phosphatidylcholine biosynthesis the primary defect was at the level of choline kinase. In addition to the decreased synthetic rate of phosphatidylcholine, the turnover rate of phosphatidylcholine was also reduced approximately 2-fold in this mutant. These decreased rates of both synthesis and degradation of phosphatidylcholine probably account for the identical phosphatidylcholine contents between the mutant and parent. As a conclusion, it may be given that strain 89.1 is a pleiotropic mutant which possesses several alterations in phosphatidylcholine metabolism, and such mammalian mutants have not been isolated previously.
The base-exchange enzymes which catalyze the Ca*+-dependent, non-energy-requiring exchanges of free L-serine, choline, and ethanolamine with the polar head groups of phospholipids to produce phosphatidylserine, phosphatidylcholine, and phosphatidylethanolamine, respectively, are known to occur in eukaryotic cells  for a review, see Ref. 8). Several biochemical studies (6,(9)(10)(11) and partial purifications (12,13) of the enzymes suggest that separate enzymes exist for the respective base-exchange reactions, and these enzymes appear to be localized in the microsomal subcellular fractions (9,14,15). It has been suggested that the serine-exchange enzyme might be responsible for phosphatidylserine biosynthesis in mammalian cells (9); however, no direct evidence is available. In addition, very little is known about the physiological roles of the choline-and ethanolamine-exchange reactions. Biochemical and genetic studies of animal cell mutants defective in the base-exchange reactions would be extremely useful to understand their physiological functions, but such mutants have not been obtained.
In this study, we have developed a rapid autoradiographic in situ enzymatic assay for detecting a base-exchange activity of phospholipids with serine in CHO cell colonies immobilized on filter paper and obtained a CHO cell mutant (designated 89.1) partially defective in the serine-exchange activity. We now report the isolation and initial biochemical characterization of the mutant 89.1 which also possesses several alterations in phosphatidylcholine metabolism. Cells growing as monolayers were mutagenized and then stored according to the method of Esko and Raetz (16,18). Under our conditions, the rate of occurrence of ouabain-resistant (at 1 mM) variants was approximately 5 X 10e5. About 5,000-10,006 mutagen-treated cells were grown in a plastic Petri dish (100 mm in diameter) at 33 "C overnight. and then a disc of filter paper (Whatman No. 56) was floated in the dish and weighed down with glass beads (16). After 20 days at 33 "C, the paper was removed from the master dish, washed with phosphate-buffered saline (25), and then placed in a fresh plastic dish as described by Esko and Raetz (16). After an additional 7 days at 33 "C, the paper was again removed and placed cell-side-up on top of a single, even layer of glass beads in a Petri dish (16). After the dish was incubated in a CO*incubator at 40 "C for 1 day, the filter paper was washed with phosphate-buffered saline, blotted on a paper towel, placed cell-sideup in a dish containing 2 ml of phosphate-buffered saline, and then stored at -70 "C (Revco freezer) until the assay. Other Enzyme Assays-Choline kinase and CDP-choline synthetase were assayed at 37 "C for 30 min essentially according to the methods of Weinhold and Rethy (26) and Esko and Raetz (18), respectively, except that radioactive reaction products were analyzed on thin layer plates as described below. Choline phosphotransferase was assayed at 37 "C for 15 min under the conditions described by Weiss et al. (27), except that radioactive phospholipids were extracted by the method of Bligh and Dyer (28), and separated and quantitated by one-dimensional thin layer chromatography as described below. Phospholipid methyltransferase(s) was assayed as described previously (29

Characterization
of Base-exchange Reactions of Phospholipids with Serine, Choline, and Ethanolamine in CHO Cell Extracts-In order to isolate mutants defective in base-exchange reactions, we first examined the properties of the reactions in CHO cells. Fig. 1 shows some properties of the base-exchange reaction of phospholipids with serine in CHO cell extracts. The activity was Ca2+-dependent and was almost completely inhibited by EGTA (Fig. lA), and the pH optimum was about 7.5 (Fig. 1B). The rate of L-[?3]serine incorporation was linear for 20 min (Fig. IC), and a linear increase in the L-[Ylserine incorporation was seen up to 200 pg of added protein (Fig. 1D). Similar results were obtained for [methyl-Wlcholine and [1,2-'*C]ethanolamine incorporation (data not shown). K,,, values for serine, choline, and ethanolamine were 0.027, 0.18, and 0.048 mM, respectively. The base-exchange activities for all three substrates were totally sedimented by centrifugation at 100,000 X g for 60 min.
The chloroform-soluble radioactive materials generated in vitro by base-exchange reactions of phospholipids with L-['~C] serine, [methyl-?3]choline, and [1,2-'4C]ethanolamine using CHO cell extracts were characterized by two-dimensional thin layer chromatography and found to consist of approximately 80% phosphatidylserine plus 20% phosphatidylethanolamine, more than 95% phosphatidylcholine, and more than 95% phosphatidylethanolamine, respectively. In addition, almost all of the radioactivity in 5% (w/v) trichloroacetic acid-insoluble materials produced in vitro was recovered in the chloro- form phase after extraction by the method of Bligh and Dyer (28).
Isolation of a Mutant (Strain 89.l)"utant strain 89.1 was originally isolated as a variant defective in the serine-exchange reaction as described here. An in situ autoradiographic assay (16) for the base-exchange of phospholipids with L-['~C] serine in CHO cells was developed, by employing the assay conditions and dependences of the serine-exchange reaction established in this study (see previous section). Approximately 10,000 individual mutagen-treated colonies immobilized on filter paper were screened, and one mutant was obtained. Fig. 24 shows the Coomassie blue-stained filter paper on which the mutant was found by visual comparison with the corresponding autoradiogram shown in B. The arrows indicate the position of the mutant which appeared to be defective in enzyme activity. This colony was retrieved from the master dish and subjected to a second cycle of cloning followed by autoradiography (16). Cells from one of the defective colonies, designated strain 89.1, were employed for all further biochemical studies.
Comparison of Base-exchange Activities in CHO-Kl and Mutant 89.1-Extracts of CHO-K1 and strain 89.1 grown a t 37 "C were assayed for base-exchange activities with L-['~C] serine, [methyl-'4C]choline, and [ 1,2-'4C]ethanolamine as substrates ( Table I). The specific activity of the serine-exchange reaction was reduced approximately 2-fold in extracts of the mutant, while there was no significant difference of the specific activity for the ethanolamine-exchange reaction between the two strains. Unexpectedly, the specific activity of the choline-exchange reaction in this mutant was only 5 to 10% of that in the parent. Similar results were obtained by using extracts of the mutant and parent cells grown at both 33 and 40 "C, with respect to the specific activities of the three baseexchange reactions. The membrane fraction of mutant 89.1 grown at 37 "C was also defective in the serine-and cholineexchange activities to the same extent as observed in crude extracts of the cells. Mixing experiments demonstrated that no inhibitor of the serine-and choline-exchange reactions was present in the mutant extracts.
The residual serine-exchange activity of strain 89.1 was not different from the wild type activity with respect to its K,,, value for serine and thermolability (data not shown). As shown in Table 11, Triton X-100 (0.1%, w/v) almost completely inhibited the serine-and choline-exchange reactions; however, the inhibition of the former reaction was reversed, even in the presence of Triton X-100, by adding either phosphatidylcholine (1.25 mM) or phosphatidylethanolamine (1.25 mM) to the reaction mixture whereas the activity of the latter reaction was not restored in this way. The specific activities of the serine-exchange reaction in the mutant and parent were found to be practically identical when assayed in the presence of Triton X-100 and either phosphatidylcholine or phosphatidylethanolamine (Table 11). Furthermore, when the serine-and choline-exchange enzymes were solubilized from the membranes of the mutant and parent cells and assayed in the presence of exogenous phospholipids (Asolectin) as substrates, the level of serine-exchange activity in the mutant was stored to the parental level; on the other hand, the choline-exchange activity was found to be still much weaker in the preparation from the mutant, although the cholineexchange enzyme from both strains appeared to be significantly inactivated during the solubilization (Table 111). These results, taken together, suggest that the mutant cells do not contain a mutation in the polypeptide of the serine-exchange enzyme, and the reduced activity of serine-exchange in the mutant can probably be ascribed to the decreased availability of acceptor phospholipid(s) for the reaction, and also suggest that the choline-exchange enzyme is distinct from the serineand ethanolamine-exchange enzymes as judged by genetic criteria.
Enzymes of Phosphatidylcholine Synthesis in Extracts of Mutant 89.1-As the rate of incorporation of [meth~l-'~C] choline into phosphatidylcholine in uiuo was demonstrated to be significantly decreased in mutant 89.1 as described below, activities of other phosphatidylcholine biosynthetic enzymes were also studied. As shown in Table IV, there were no significant differences between the mutant and wild type in the specific activities of CDP-choline synthetase in the homogenate, cytosolic and particulate fractions, choline phosphotransferase, and phospholipid methyltransferase(s) in cells grown in the presence of either choline, N,N'-dimethylethanolamine, or N-monomethylethanolamine. However, the specific activity of choline kinase was reduced 3-to 4-fold in the mutant. No significant differences were observed between the mutant and wild type in the properties of choline kinase such as K, values for choline and ATP (0.050 f 0.010 versus 0.040 k 0.011 mM and 1.5 f 0.1 versus 1.3 f 0.1 mM, respectively), pH optimum (pH 8.0-9.5), and thermolability (data not shown).
Studies on Growth, PhospholipidlProtein Ratio, Phospholipid Composition, and Karyotype-Mutant 89.1 was not temperature-sensitive for growth and grew exponentially in a monolayer culture in growth medium containing 10% newborn calf serum a t 33,37, and 40 "C, with doubling times quite similar to the parent (about 22, 16, and 12 h, respectively).

FIG. 2. Isolation of mutant 89.1.
Colonies derived from mutagen-treated cells were allowed to grow on a disc of filter paper and assayed for serine-exchange activity, as described under "Experimental Procedures." A, Coomassie blue-stained filter paper on which the mutant colony appeared B, autoradiogram made from this filter paper. The arrows mark the position of the mutant colony.

Enzymatic assays of base-exchange activities in extracts prepared from mutant and wild type cells
Extracts were made from cells at the mid-late exponential phase a t 37 "C and assayed, as described under "Experimental Procedures." The error in duplicate determinations was about +5%. Although the values of specific activities varied somewhat from experiment to experiment, the ratios of those between mutant 89.1 and CHO-K1 were almost constant (compare Tables I and 11).

Substrate
Specific activity

Effects of Triton X-100, phosphatidylcholine, and phosphatidylethanolamine on serine-and choline-exchange activities in extracts prepared from mutant and wild type cells
Extracts were made as described in Table I Less than 0.01.
When cells were analyzed for protein and phospholipid contents by the method of Lowry et al. (36) and by perchloric acid digestion and quantitation of the released phosphorus (32), respectively, the amount of phospholipid per mg of protein was virtually identical with that in the parent (180 uersus 191 nmol/mg of protein). As for the phospholipid compositions determined chemically, the mutant was also indistinguishable from the parent; similar results were obtained when phospholipid compositions were determined on long term ("2P)orthophosphate and [ l-14C]acetic acid labeling (data not shown). These results indicate that the steady state contents of phospholipid species, including phosphatidylcholine, in the mutant cells are comparable to those in the parental cells.
Chromosomal analysis showed no significant difference in the karyotype for mutant 89.1 as compared to the parent; both strains had a model chromosome number of 20.
Reduced Synthesis of Phosphatidylcholine in Mutant 89.1-In order to examine the synthesis of phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine and protein in uiuo, suspensions of mutant and parental cells harvested with a rubber policeman from monolayer cultures a t 37 "C were pulse-labeled with [methyl-'4C]choline, [1,2-'4C]ethanolamine, ~-['~C]serine, and ~-['~C]leucine, respectively, for 5,10, or 20 min at 37 "C. As shown in Fig. 3A, the incorporation of [methyl-'4C]choline into phosphatidylcholine was strikingly reduced (3to 5-fold) in the mutant cells, while there were no significant differences between the mutant and parent cells in the incorporation of [ 1,2-'4C]ethan~lamine into phosphatidylethanolamine (Fig. 3B), and ~-["C]leucine into protein (Fig. 3 0 )  phatidylserine at about 50% of the parental rate (Fig. 3C), corresponding to the partial defect in serine-exchange activity in vitro, but the difference between the mutant and parent disappeared when the cells were labeled for more than 1 h (data not shown).
The kinetics of phospholipid synthesis in vivo were also investigated by pulse-labeling of cells in monolayer cultures with (32P)orthophosphate. Again the mutant was deficient in the labeling of phosphatidylcholine in short pulses (Fig. 4A), suggesting that the reduced synthesis of phosphatidylcholine was due to a defect not in the choline-exchange or phospholipid methylation pathway, but in the CDP-choline pathway. The incorporation of (32P)orthophosphate into other phospholipids, including phosphatidylethanolamine (Fig. 4B) and phosphatidylinositol (Fig. 40), was not grossly affected. These results suggested that altered ATP, CTP, or diglyceride was not responsible for the reduction in phosphatidylcholine synthesis in mutant 89.1. As shown in Fig. 4C, the incorporation of (32P)orthophosphate into phosphatidylserine was reduced about Z-fold in the mutant, supporting the idea that phosphatidylcholine may serve as a precursor for phosphatidylserine.
Defective Synthesis of Phosphorylcholine and CDP-choline in Mutant 89.1 ---In order to determine at what step Imethyl-%]choline and (32P)orthophosphate incorporations into phosphatidylcholine were altered, mutant and parental cells were labeled with [methyl-'*C]choline under the same conditions as in Fig. 3A and the amount of radioactivity in watersoluble choline metabolites was examined. The choline-containing metabolites were extracted by the method of Bligh and Dyer (28), and the resulting methanolic water phase was analyzed by thin layer chromatography. As shown in Fig. 5, the incorporation of [methyl-Wlcholine into phosphorylcholine was reduced about 3-fold in the mutant. Because the amount of radioactivity incorporated into CDP-choline was very small in the experiment in leucine (342 mCi/mmol) (D) at 1 &i/ml, respectively. All these procedures were performed at 4 "C. The cell suspensions were incubated for 5, 10, or 20 min at 3'7 "C, and the reaction was terminated by adding an equal amount of 10% (w/v) trichloroacetic acid. The resulting cell pellets were washed twice more with 5% (w/v) trichloroacetic acid. In A, B, and C, phospholipids were extracted and separated by one-dimensional thin layer chromatography, and then spots of phosphatidylcholine (A), phosphatidylethanolamine (B), and phosphatidylserine plus phosphatidylethanolamine (C) were scraped off and counted, as described under "Experimental Procedures." In D, the washed cell pellets were solubilized in phosphate-buffered saline containing 2% (w/v) sodium dodecyl sulfate, and then a portion of the solubilized sample was directly counted, using ACS II (Amersham International) scintillation fluid. The results are expressed as counts per min/mg of protein., 0, CHO-Kl; l , mutant 89.1.
showed that the incorporation of [methyl-'4C]choline into CDP-choline in the mutant and parent was 174 cpm/lO min/ mg of protein and 1197 cpm/lO min/mg of protein, respectively. The results shown in Figs. 3A and 5 and Table IV, taken together, clearly showed that mutant 89.1 was defective in phosphorylcholine, CDP-choline, and phosphatidylcholine biosynthesis in vivo and that the primary defect was at the level of choline kinase.
Turnover of Phospholipids in CHO-Kl and Mutant 89.1-Because the reduced synthetic rate of phosphatidylcholine did not result in its net decrease, the turnover of phosphatidylcholine was investigated by measuring the loss of [methyl-'%]choline from pulse-labeled lipid after removing [methyl-'*C]choline from the growth medium. As shown in Fig. 6, the apparent turnover rate of phosphatidylcholine in the mutant was about Z-fold slower than in the parent; namely, the halflife time of phosphatidylcholine in the mutant and parent was approximately 30 and 15 h, respectively. The turnover of phospholipids was also examined by measuring the loss of 89.1 were grown at 37 "C in medium containing 10% (w/v) newborn calf serum. Cells were harvested with trypsin and used to inoculate a series of 60-mm diameter dishes containing the same medium. Cells were grown to a density of 4-5 X 106/dish at 37 "C and then pulselabeled at zero time by removing the medium by aspiration, washing the cells with 5 ml of fresh medium, and then adding medium containing 1 pCi/ml of (32P)orthophosphate. After incubation at 37 "C in a C02 incubator for 2, 4, 8, or 12 b, one dish of each strain was removed, and cells were harvested by scraping with a rubber policeman. Phospholipids were extracted, and individual phospholipids were quantitated, as described under "Experimental Procedures." A, phosphatidylcholine; B, phosphatidylethanolamine; C, phosphatidylserine; D, phosphatidylinositol. The results are expressed as counts per min/106 cells. 0, CHO-Kl; 0, mutant 89.1.
(32P)orthophosphate from continuously labeled lipid after removing (32P)orthophosphate (Fig. 7). Under these conditions, the turnover of phosphatidylcholine in this mutant was specifically depressed to a similar extent to that seen for [rnethyl-"Clcholine, whereas there was no significant difference in the turnover of other phospholipids. The decreased rates of both synthesis (Figs. 3A and 4) and degradation (Figs. 6 and 7) of phosphatidylcholine probably account for the identical phosphatidylcholine contents of the mutant and parent.

DISCUSSION
In the present study, mutant 89.1 was isolated by in situ autoradiographic enzyme screening for detecting a base-exchange activity of phospholipids with serine in CHO cell lysates immobilized on filter paper. The mutant was about 2fold defective in the serine-exchange activity. An unexpected finding was that in extracts of this mutant the specific activities of the choline-exchange enzyme and choline kinase were remarkably reduced (10-to 15-fold and 3-to 4-fold, respectively). The data presented here suggest that the reduced activity for serine-exchange in the mutant cells is not due to a mutation in the structural gene of the serine-exchange enzyme, but rather it is probably due to the decreased availability of acceptor phospholipid(s) for the reaction. Several workers (9,(37)(38)(39) have suggested that phosphatidylserine in mammalian cells may be synthesized in uivo by the baseexchange of serine with phosphatidylcholine. Our results in Fig. 4C also support this possibility. Because the phosphati- dylcholine content in bulk was identical between the mutant and parent, the amount of some particular phosphatidylcholine molecules which can be used as acceptors for the serineexchange reaction may be decreased in mutant 89.1; phosphatidylcholine molecules which are formed by the choline-exchange reaction may serve as acceptors for the serine-exchange reaction.
The phosphorylation of choline catalyzed by choline kinase is the first step in the de novo synthesis of phosphatidylcholine (40). From a theoretical point of view, Infante (41) has predicted that choline kinase is a rate-limiting enzyme for phosphatidylcholine biosynthesis in rat liver. On the other hand, Vance and Choy (42) have argued against this prediction with several lines of experimental evidence. We have shown here that mutant 89.1 is about %{old defective in choline kinase activity in vitro, and also correspondingly reduced in phosphorylcholine, CDP-choline, and phosphatidylcholine synthesis in uivo. These results strongly suggest that the choline kinase detected under our assay conditions is a rate-limiting enzyme for phosphatidylcholine synthesis in CHO cells. Mutant 89.1 is the first animal cell mutant defective in choline kinase and so should be useful to study the nature and function of this enzyme. Since the presence of four isoenzymes for choline kinase was observed in rat liver (43) but the function of each isoenzyme is unknown, it would be of interest to determine if a particular isoenzyme is missing in the mutant. Mutant 89.1 also showed another intriguing characteristic. We found a decreased turnover rate of phosphatidylcholine, which probably accounts for the normal level of phosphatidylcholine in the mutant in which the synthetic rate of in regular medium were harvested with trypsin, and replated at 2.5 X IO5 cells/60-mm diameter dish containing 5 ml of Ham's F-12 lacking choline, supplemented with 10% (v/v) dialyzed newborn calf serum and then incubated overnight in order to decrease the pool sizes of phosphatidylcholine precursors. The cells were then labeled for 1 h in 5 ml of Ham's F-12 lacking choline, supplemented with 10% (v/v) dialyzed newborn calf serum (16) and [methyl-"C]choline (0.5 pCi/ ml). After removing the radioactive medium (0 time), the cells were washed with phosphate-buffered saline and incubated in the medium lacking choline for 1.5 h, and then 1 mM choline was added to the medium. At the times indicated, the cells were harvested with a rubber policeman, and phospholipids were extracted and analyzed by one-dimensional thin layer chromatography, as described under "Experimental Procedures." Cells were cultured at 37 "C, and the growth of CHO-K1 and mutant 89.1 was virtually identical under the experimental conditions. The amount of radioactivity remaining in phosphatidylcholine per dish, expressed as percentage of original radioactivity, is plotted uersm time. 0, CHO-K1; 0, mutant 89.1. grown in regular growth medium containing (3zP)orthophosphate (10 pCi/ml) for 3 days in order to label the cellular phospholipids to constant radiospecific activity. The cells were harvested with trypsin and replated at 2 X lo5 cells/lOO-mm diameter dish in the absence of label in regular growth medium. At the times indicated, the cells were harvested with a rubber policeman and extracted, and phospholipids were analyzed by two-dimensional thin layer chromatography, as described under "Experimental Procedures." The amount of radioactivity remaining in each phospholipid per dish is plotted uersm time. Cells were cultured at 37 "C during this experiment, and cell growth was virtually identical between CHO-Kl and mutant 89.1 under the conditions in this experiment. PC, phosphatidylcholine; PE, phosphatidylethanolamine; SM, sphingomyelin; PS, phosphatidylserine; PI, phosphatidylinositol. A, turnover in mutant 89.1; E , turnover in CHO-K1.
phosphatidylcholine was remarkably reduced. These results suggest the presence of a control mechanism which coordinates the synthesis and degradation of phosphatidylcholine to maintain an adequate amount of phosphatidylcholine for cell function, Recently, Sleight and Kent (44-46) have shown that phospholipase C treatment enhanced the rates of both degradation and synthesis of phosphatidylcholine in CHO cells, but in this case the activity of CDP-choline synthetase was specifically enhanced. Although the molecular basis for the reduced rate of phosphatidylcholine turnover is unknown, the defect of the choline-exchange reaction in mutant 89.1 may be responsible for the phenotype. However, the involvement of an additional lesion which is responsible for the reduced phosphatidylcholine turnover cannot be excluded. Whatever the mechanism is, this mutant will be useful to study the participation of phosphatidylcholine turnover in various membrane functions such as endocytosis and exocytosis. Miura and Kanfer (12) succeeded in obtaining separate enzyme fractions which possessed principally the incorporation activity of either choline, serine or ethanolamine, by using column chromatographies. Our results provide the first genetic evidence that the choline-exchange enzyme is different from the serine-or ethanolamine-exchange enzyme. To understand the physiological function of the choline-exchange reaction more precisely, we are isolating additional mutants more grossly defective in the reaction by an in situ assay method similar to that for serine-exchange mutants.
Finally, at present it is unknown whether the phenotypic abnormalities seen in mutant 89.1 were caused by a single mutation or multiple mutations. We favor the former possibility, since the mutant could be isolated by screening of only about 10,000 mutagen-treated ceil colonies. Furthermore, the fact that the thermolabilities of both the choline-exchange enzyme2 and choline kinase were identical between the mutant and parent suggests that the mutation in mutant 89.1 was in a regulatory gene instead of structural gene(s). The gene expression of both choline-exchange enzyme and choline kinase may be controlled in a coordinate manner by a common factor or a common regulatory mechanism (47,48) which may be altered in this mutant. Obviously, more studies are needed to test these and other possibilities.