Regulation of Phospholipid Biosynthesis in Isolated Rat Hepatocytes EFFECT OF DIFFERENT SUBSTRATES*

SUMMARY The effects of choline, ethanolamine and its N-methyl analogs, different fatty acids, and L-methionine on phospholipid biosynthesis via the CDP-ester pathways and the methylation pathway were studied in rat hepatocytes. Phosphatidylethanolamine synthesis was stimulated severalfold by 0.02 to 0.1 mrd ethanolamine, especially in the presence of long chain unsaturated fatty acids. At higher concentrations of ethanolamine, phosphorylethanolamine accumulated but the level of CDP-ethanolamine and the rate of phosphatidylethanolamine synthesis did not increase further. The rate of phosphatidylcholine synthesis via the CDP-ester pathway responded in a way analogous to that of phosphatidylethanolamine synthesis upon the addition of choline and fatty acid, except that a lo- to ZO-fold higher concentration of choline was required for maximal stimulation, probably due to the rapid oxidation of choline to betaine. Phospholipids containing N-monomethyl-or N,N-dimethylethanolamine were efficiently formed from the corresponding free bases in the absence of ethanolamine and choline. Ethanolamine, but not other bases, inhibited completely phospholipid formation from N-monomethylethanolamine, probably as a result of competition at the level of CDP-ester formation. The data indicate that the cytidylyltransferase reactions are rate-lim-iting steps in the synthesis of phosphatidylethanolamine


SUMMARY
The effects of choline, ethanolamine and its N-methyl analogs, different fatty acids, and L-methionine on phospholipid biosynthesis via the CDP-ester pathways and the methylation pathway were studied in rat hepatocytes. Phosphatidylethanolamine synthesis was stimulated severalfold by 0.02 to 0.1 mrd ethanolamine, especially in the presence of long chain unsaturated fatty acids. At higher concentrations of ethanolamine, phosphorylethanolamine accumulated but the level of CDP-ethanolamine and the rate of phosphatidylethanolamine synthesis did not increase further. The rate of phosphatidylcholine synthesis via the CDP-ester pathway responded in a way analogous to that of phosphatidylethanolamine synthesis upon the addition of choline and fatty acid, except that a lo-to ZO-fold higher concentration of choline was required for maximal stimulation, probably due to the rapid oxidation of choline to betaine. Phospholipids containing N-monomethyl-or N,N-dimethylethanolamine were efficiently formed from the corresponding free bases in the absence of ethanolamine and choline. Ethanolamine, but not other bases, inhibited completely phospholipid formation from N-monomethylethanolamine, probably as a result of competition at the level of CDP-ester formation. The data indicate that the cytidylyltransferase reactions are rate-limiting steps in the synthesis of phosphatidylethanolamine and probably also phosphatidylcholine. In addition, the availability of diacylglycerol and its fatty acid composition may significantly affect the rate of phospholipid synthesis.
The rate of phosphatidylcholine formation via phospholipid N-methylation approximately doubled when L-methionine was added at concentrations similar to that in rat plasma. Under these conditions the rate of phosphatidylcholine synthesis via this pathway was 20 to 40% of that via diacylglycerols and CDP-choline. The methylation of phosphatidylethanolamine to phosphatidylcholine remained essentially constant when the rate of phosphatidylethanolamine synthesis was varied &fold, but was significantly reduced when the * This work was supported by grants from the Swedish Medical Research Council (Project 03X-3968), from the H. Jeanssons Foundation, from the A. PBhlssons Foundation, and from the Medical Faculty, University of Lund. An account of this work was presented at the 17th International Conference on the Biochemistry of Lipids, Milan, September 1974. formation of N-monomethyl-or N, N-dimethylphospholipid was stimulated by addition of the corresponding base. These phospholipids not only replaced phosphatidylethanolamine as the substrate for methylation but also increased the rate of phosphatidylcholine formation via this pathway.
A method for the determination of nanomole amounts of merent ethanolamine compounds is described.
Phospholipids synthesized in liver participate in membrane formation within the organ or are transferred into bile or the lipoproteins of blood plasma (1). Although little is known about physiological variation in the rates of these processes certain experimental situations, such as the induction of microsomal enzymes by phenobarbital, seem to be accompanied by an increased rate of phospholipid biosynthesis (2, 3). The mechanisms behind such changes, and even the means by which the biosynthesis of phospholipids is controlled under normal conditions, are, however, largely unknown.
We have therefore studied the effects of different substrates on the rate of phosphatidylethanolamine and phosphatidylcholine synthesis via the Kennedy pathway (4), on phosphatidylcholine synthesis via phospholipid N-methylation (5), and on the level of some intermediary metabolites in rat hepatocytes in order to identify critical steps in the regulation of these pathways. They were purified by ion exchange chromatography (7) and determined ae Dnsl-ethanolamine (see below).
Preparation of Hepatocytes-Adult male Sprague-Dawley rats 3360 (Anticimex AB, Stockholm) weighing 200 to 300 g were fed a balanced diet ad lib&m.
Hepatocytes were prepared by the method of Berrv and Friend (11) aa described elsewhere (12). In later experiments a modification similar to that of Ingebretsen and Wagle (13) was used. Then hyaluronidase was omitted, the amount of collagenase (type 1, Sigma, St. Louis) was lowered (0.25 mg/ml) and 1% (w/v) delipidated bovine serum albumin was included in the perfusate. Vital staining of the cells with trypan blue showed 80 to 95yc of them to be intact.
Incubation Conditions-Incubations were performed in 25-ml Erlenmeyer flasks in a total volume of 0.5 ml of Hanks' solution (14) buffered with 10 mM phosphate (pH 7.4) and containing 2y0 (w/v) delipidated bovine serum albumin as previously described (15). The amount of cell protein ranged between 1.5 and 4.0 mg per incubation in different experiments. In experiments where metabolites were determined hepatocytes (29 to 107 mg of protein) were incubated in 250-ml flasks in a final volume of 8 ml. Different substrates and radioactive precursors were added m indicated.

Extraction and Sevaration of Labeled
Products-Linids were extracted with chlordform-metdanol (l:l), washed free of labeled precursors, and separated by thin layer chromatography on Silica Gel H (15). For the isolation of uhosnhatidvl dimethvlethanolamine, ihi; layer plates were devilopei in cgloroform-"methanolconcentrated NH3 (60:30:5 by volume) (16). A reference sample of this compound was prepared by reductive methylation of phosphatidylethanolamine with formaldehyde and formic acid (17). For the isolation of phosphatidyl methylethanolamine together with phosphatidylserine and phosphatidylinositol, the developing solvent 1-butanol-acetic acid-water (60:20:20 by volume) was used (16).
Incubations with [methyl-aH]choline were terminated by the addition of 0.5 ml of ice-cold 6% (w/v) trichloroacetic acid and were transferred into test tubes together with a rinse of 1 ml of 3% trichloroacetic acid. After shaking on a Vortex mixer for 1 min, followed by centrifugation, the supernatant was aspirated off and the sediment was washed twice with 1 ml of 3% trichloroacetic acid. The sediment was then neutralized with 1 drop of 4 M NHaOH and lipids were extracted with 3 ml of chloroform-methanol (1: 1). The acid-soluble extract was shaken 5 times with several volumes of diethyl ether to remove tricbloroacetic acid and the remaining water phase was subjected to thin layer chromatography (Silica Gel H; solvent, methanol-0.5% NaCl-concentrated NHa, 50:50:5 by volume). This system separated betaine plus CD@-choline (RF N_ 0.6). Dhosnhorvlcholine (RF N 0.3). and choline (RF N 0.1). +he compounds aere"visualize4 dy expoiure to iodine tipor. Fdr r+dioactivity determination the gel was scraped off into scintillation vials containing 1 ml of water. After shaking, 1 ml of methanol and 5 ml of Instagel (Packard) were added. The recovery of radioactivity from the plate was better than 90%. The isotope content of CDP-choline was considered negligible compared to that of betaine.

Determination of Ethanolamine
Compounds-The incubations were terminated with 2 ml of ice-cold 16% (w/v) trichloroacetic acid. Phosphoryl[IH]ethanolamine (0.07 &i) and CDP-[aH]ethanolamine (0.15 pCi) were then added. After cooling on ice the mixture was transferred to a test tube, shaken on a Vortex mixer for 1 min. left for 15 min. and then centrifuged. The sediment was washed t&e with 2 ml'of 3% trichloroac&c acid and from the combined supernatants the acid was extracted several times with diethyl ether. The water phase was adjusted to pH 8 to 9 with KOH, and phosphoryl-and CDP-ethanolamine were separated on a column (1.5 X 15 cm) of AG l-X4 (HCOO-) (Bio-Rad) as previously described (7). Fractions containing phosphoryl-or CDPethanolamine (1 to 200 nmol) were pooled and evaporated to dryness. CDP-Ethanolamine samples were hydrolyzed to CMP and phosphorylethanolamine by treatment with 1 M HCl at 100" for 90 min (18). Ethanolamine was then liberated from the hydrolyzed CDP-ethanolamine samples and from phosphorylethanolamine samples by treatment with acid phosphatase (grade I, Boehringer Mannheim) as follows. The samples were dissolved in 1 ml of 0.1 M sodium citrate buffer (pH 5.6), and 10 ~1 of enzyme suspension were added. After incubation at 37" for 90 min at least 98yo of the phosphorylethanolamine had been degraded. The enzymatic hydrolysis was stopped and liberated ethanolamine converted into Dns-ethanolamine by the addition of 2 ml of Dns-Cl in acetone (1 mg/ml) and an excess of solid Na&03 (19). The reaction with Dns-Cl was allowed to proceed for at least 10 hours to allow for hydrolysis of excess reagent. Samples containing free ethanolamine were directly treated with Dns-Cl.
After dansylation the samples were transferred into test tubes, acetone was evaporated under a stream of nitrogen, and nonpolar Dns-derivatives were extracted into 5 volumes of ethyl acetate. Dns-ethanolamine was then isolated by thin layer chromatography on Silica Gel H (solvent, ethylacetate-benzene (70:30, v/v)). Rechromatography was sometimes necessary for complete separation from byproducts. Dns-ethanolamine (RF N 0.4) was quantitatively eluted from the gel with 4 ml of benzene-methanol (90: 10, v/v). When chromatographed in any of three solvent systems: 1, benzene-triethylamine-methanol (80:15:5 by volume); 2, chloroform-ethyl acetate-concentrated ammonia (50: 48: 2) ; 3, diethyl ether-methanol-acetic acid (98:2:0.5), one single spot cochromatographing with Dns-ethanolamine was seen under ultraviolet light. Dns-ethanolamine was determined (Jobin Yvon spectrofluorimeter) in the eluate of benzene-methanol (90: lo), and an aliquot thereof was also taken for radioactivity determination. Excitation maximum was at 360 nm and the emission maximum at 510 nm. The fluorescence response was linear to the concentration of Dns-ethanolamine in the range 0.05 to 5 by (deviation <5yc). The fluorescence of the samples was corrected for that of blanks containing formic acid eluate from the AG-1 resin, acid phosphatase, and citrate buffer, which were carried through the same procedure. The fluorescence of blanks after elution with 4 ml of benzene-methanol (90: 10) corresponded to 0.03 to 0.06 PM Dnsethanolamine. Determination of Dia ylglycerol-Diacylglycerol' was isolated from the lipid extract of incubations prepared as described above. Blank incubations lacking hepatocytes were carried through the same procedure. After thin layer chromatography on Silica Gel H (developing solvent, toluene-chloroform-methanol (85:12:3 by volume)) only reference substances running beside the sample were sprayed with 0.2% dichlorofluorescein in ethanol. The diacylglycerol zone was eluted with 12 ml of chloroform-methanolacetic acid-water (50:39:1:10 by volume) and the extract was washed once with 4 M NH,OH and twice with methanol-water (1:l). Aliquots (2 to 20 nmol) were taken for quantitative determination by a modification of Laurell's procedure (20). The samples were taken to dryness and 100 ~1 of 0.015 M KOH in ethanol were added. After 15 min at 65" the samples were again taken to dryness and 100 ~1 of the following solution were added: 1 mM ATP, 1.5 mM NAD, 10 mM cysteine (pH 9), 0.5 mM MgC&, 0.2 M hydrazine-HCl (pH 9.4), crystalline-defatted bovine serum albumin (5 mg/ml), glycerokinase from Cundida mycoderma (Boehringer Mannheim) (0.5 pg/ml), and glycerol-3-phosphate dehydrogenase from rabbit muscle (Boehringer Mannheim) (25 &ml). After 1 hour at room temperature, 0.4ml of 0.01 M N&H ioitainina EDTA (0.4 a/liter) was added. The samnles were read in a Jo&n Yvon'spe&oflubrimeter at an excitini wavelength of 355 nm and an emission wavelength of 474 nm. Trioleoylglycerol (Hormel Institute, Austin) was used as a standard. The fluorescence of blank incubations corresponded to 0.1 to 0.5 nmol of diacylglycerol per mg of protein in different experiments. The determined values of diacylglycerol were corrected for losses during isolation from the recovery of isotope.

Phosphatidylethanolamine
SynthesisThe addition of ethanolamine to rat hepatocytes incubated with [aH]glycerol increased the incorporation of isotope into phosphatidylethanolamine approximately a-fold at 0.02 to 0.04 mM concentration (Fig. l), which is only slightly higher than the concentration of ethanolamine in rat blood plasma ( Table 1). The increased labeling of phosphatidylethanolamine was not accompanied by any significant change in total [aH]glycerol incorporation into lipids but was compensated for by a decreased incorporation into phosphatidylcholine and, above all, the neutral glycerides (Fig. 1). Since this indicates that the enzymes catalyzing triacylglycerol, phosphatidylcholine, and phosphatidylethanolamine synthesis may compete for diacylglycerol we also studied the effect of ethanolamine after stimula-* Diacylglycerol denotes in this paper 1,2-diacyl-sn-glycerol.   of ethanolamine was more pronounced in the presence of oleic acid while in the presence of lauric acid ethanolamine was without effect. Other experiments showed that the effects of ethanolamine and fatty acids on ]aH]glycerol incorporation into phosphatidylethanolamine paralleled those on the incorporation of [azP]phosphate and therefore probably reflect true changes of the rate of phosphatidylethanolamine synthesis. Several studies have shown that in the intact liver cell diacylglycerol containing docosahexaenoic acid is preferentially utilized for the synthesis of phosphatidylethanolamine (15,(21)(22)(23)(24). In agreement with such a specificity, docosahexaenoic acid was found to stimulate phosphatidylethanolamine synthesis at lower concentrations than did oleic acid, both in the absence and presence of a saturating concentration of ethanolamine (Fig. 3). The results so far indicate that the rate of phosphatidylethanolamine synthesis depends on the exogenous supply of ethanolamine and of fatty acid and in addition on the fatty acid structure. However, the strong inhibition by lauric acid and the fact that the rate of phosphatidylethanolamine synthesis leveled off at a relatively low concentration of ethanolamine, both in the presence and absence of fatty acids, required further elucidation. The effects of the different substrates on the level of phosphorylethanolamine, CDP-ethanolamine, and diacylglycerol were therefore determined. When the concentration of added ethanolamine was increased, phosphorylethanolamine accumulated whether a fatty acid was present or not, while no parallel change in the concentration of CDP-ethanolamine was observed (Fig. 4). In the absence of exogenous fatty acids the level of CDP-ethanolamine rose from 8.1 nmol per incubation to 16.2, 15.8, and 18.3 nmol at 0.02, 0.05, and 0.1 mM ethanolamine, respectively, which was confirmed in other experiments. Thus, in the absence of fatty acid, the effect of ethanolamine on the level of CDP-ethanolamine paralleled its effect on the incorporation of [*HIglycerol into phosphatidylethanolamine ( Figs. 1 and 2). The presence of oleic acid, on the other hand, prevented any significant rise in the level of CDP-ethanolamine.
When lauric acid was added, the changes in pool sizes were similar to those seen in the absence of fatty acid, indicating that the inhibition of phosphatidylethanolamine synthesis by lauric acid (Figs. 2 and 3) is not the result of interference with CDP-ethanolamine formation. The addition of long chain fatty acids to hepatocytes incubated with [sH]glycerol results in a severalfold increase in the isotope content of diacylglycerol (15). However, this accumulation was reduced by the simultaneous presence of ethanolamine or choline (Table II) (Table II) and the level determined chemically (Fig. 5) declined. It is therefore probable that the inhibition of phosphatidylethanolamine synthesis by lauric acid (Figs. 2 and 3) results from a shortage of diacylglycerols suitable for phospholipid formation, since the decline in total diacylglycerol is accompanied by the formation of an increasing proportion of dilaurylglycerol, which is insignificantly utilized for phosphatidylethanolamine syfithesis (15) tion of isotope into phosphatidylcholine (Fig. 6). Compared to the corresponding effect of ethanolamine on phosphatidylethanolamine synthesis (Fig. 1) the effect of choline was smaller and occurred only at a lo-to 20-fold higher concentration. The presence of oleic acid enhanced the stimulatory effect of choline on the incorporation of both [aH]glycerol and [azP]phosphate (Table  III) but the effect on azl' was always more pronounced. Although other factors might give rise to such a difference it is noticeable that since the cholinephosphotransferase (EC 2.7.8.2 .) reaction is reversible (25-28), the rate of diacylglycerol incorporation into phosphatidylcholine would be higher than the net synthesis of phosphatidylcholine from phosphorylcholine. One would therefore expect a greater relative increase of the latter reaction (and 3363 A reason for the high concentration of choline required for stimulation of phosphatidylcholine synthesis (Fig. 6) was found by studying the metabolism of [*HIcholine (Fig. 7). The oxidation of choline to betaine under the present conditions exceeded choline incorporation into phosphorylcholine and phosphatidylcholine by a factor which increased from 5 to 16 at 2 mM choline. A rapid oxidation of choline to betaine has previously been observed in the rat liver in tivo (29, 30) and in liver slices (31).
Biosynthesis of Phosphatidyl Monomethyl-and Dimethylethanolamine via CDP-es& Pathways-Exposure of the hepatocytes to N-methylethanolamine or N , N-dimethylethanolamine resulted in considerable formation of phosphatidyl methylethanolamine or phosphatidyl dimethylethanolamine from labeled glycerol, although the formation of the dimethyl phospholipid started more slowly and required preincubation with the corresponding base to become prominent (Fig. 8). Like the addition of ethanolamine or choline, N-methylethanolamine or N , N-dimethylethanolamine did not significantly affect total incorporation of [aH]glycerol into lipids but only its distribution among lipid classes. Incorporation of labeled N, N-dimethylethanolamine into liver phospho- 14 .) and phosphorylcholine cytidylyltransferase (EC 2.7.7.15.), respectively, have also been presented (35). We therefore studied the competition between pairs of such bases, since competition between ethanolamine and N-methylethanolamine and between choline and N , N-dimethylethanolamine would be expected to occur at the steps catalyzed by the cytidylyltransferases. All different combinations of bases inhibited the incorporation of [aH]glycerol into the corresponding phospholipid classes in a mutual way but only the addition of ethanolamine resulted in complete inhibition of phospholipid formation from N-methylethanolamine (Table IV). This is further evidence, although indirect, for a limited capacity in the step catalyzed by phosphorylethanolamine cytidylyltransferase. The less pronounced inhibition of phosphatidyl methylethanolamine formation by other bases most probably results from competition for a common substrate, such as diacylglycerol. A corresponding competition between choline and N , N-dimethylethanolamine at the phosphorylcholine cytidylyltransferase level could not be demonstrated. The formation of phosphatidyl dimethylethanolamine fell to 40,62, and 72% of the initial value (mean from four experiments) when ethanolamine,. N-methylethanolamine, and choline, respectively, were included in addition to N , N-dimethylethanolamine. Phosphatidylcholine Synthesis via Phospholipid N-methylatkn-The relative rate of phosphatidylethanolamine methylation was determined by following the conversion of phosphatidyl[aH]ethanolamine into N-methylated phospholipids during incubation with [aH]ethanolamine, while the total rate of phospholipid N-3364 Hepatocytes were preincubated for 30 min without baee or with ethanolamine, N-methylethenolamine (MME), N,N-dimethylethanolamine (DME) (0.4 mM) and/or choline (2.0 mM) as indicated; then [aH]glycerol (2.0 maa) wae added and incubation continued for another 60 min. PMME, phosphatidyl methylethanolamine. 2.9 MME 18.9 MME + ethanolamine..
13.2 D Including phosphatidylserine and phosphatidylinositol.
methylation was assessed with n-[methyl-W]methionine under the assumption that L-methionine via S-adenosyhnethionine is the only donor of methyl groups in this pathway (5).
The addition of methionine stimulated the conversion of phosphatidyl [aH]ethanolamine into phosphatidylcholine about a-fold over at least 2 hours (Fig. 9a), but no further increase occurred when the concentration of added methionine was raised above 0.1 mru (Fig. 9b). In accordance, the incorporation of [methyP4C]methionine was linear for about 2 hours and leveled off at about 0.1 mM concentration (Fig. 9c). The values reported for the concentration of L-methionine in rat plasma (55 to 90 PM, Refs. 36, 37) are only slightly lower than the concentration which gave maximal stimulation of phosphatidylcholine synthesis via Nmethylation.
The addition of N-methylethanolamine or N, N-dimethylethanolamine, which results in appreciable synthesis of the corresponding phospholipid (Fig. 8) increased the incorporation of labeled methyl groups at a saturating concentration of L-[methyl-W]methionine (Table V). In contrast, the conversion of phosphatidyl[aH]ethanolamine into phosphatidylcholine was inhibited by these bases. This was not due to trapping of isotope in partially methylated phospholipids, as a result of their formation from the added N-methylethanolamine or N, N-dimethylethanolamine. lnstead the rate of phosphatidylethanolamine methylation decreased. Since also the total incorporation of [aH]ethanolamine into phospholipids was depressed by the addition of N-methyl-or N, N-dimethylethanolamine, we thought that inhibition of phosphatidylethanolamine synthesis by the added bases might, secondarily, result in a lowered rate of phosphatidylethanolamine methylation.
However, the addition of lauric acid, which lowers the rate of phosphatidylethanolamine synthesis (Figs. 2 and 3), did not inhibit its conversion into N-methylated phospholipids; neither was it stimulated by the addition of ethanolamine, alone or together with oleic acid (Table V). Thus, we found no evidence for an immediate co-regulation of the rates of phosphatidylethanolamine synthesis and N-methylation, but the latter is apparently inhibited by an increased formation and methylation of phosphatidyl methylethanolamine or phospha-tidy1 dimethylethanolamine.
The total rate of phosphatidylcholine formation via phospholipid N-methylation at a saturating concentration of methionine was found to be 20 to 40yo of the rate of phosphatidylcholine synthesis from labeled glycerol, depending on the concentration of fatty acid and choline present (Table VI).

DISCUSSION
Albumin-bound long chain fatty acids stimulate glycerolipid synthesis over a wide concentration range in liver slices (38) and in isolated hepatocytes (15,39). Characteristically the formation of triacylglycerol accounts for most of this increase while the synthesis of phosphatidylcholine, and especially phosphatidylethanolamine, is much less affected. Thus the rate of phosphatidylcholine and phosphatidylethanolamine synthesis appears to be limited at the steps catalyzed by cholinephosphotransferase and ethanolaminephosphotransferase (EC 2.7.8.1.) converting diacylglycerols into phosphatidylcholine and phosphatidylethanolamine, respectively. However, this does not necessarily mean that the capacity of these ensymes is the limiting factor. The present study demonstrates that the rate of phosphatidylethanolamine synthesis and probably also phosphatidylcholine synthesis is (PC) in the controls wss 9.6 ZIZ 1.6 @I Z!Z S.E., n = 4) y. of lipid "H. Phosphatidyl mono-and dimethylethanolamine together contained less than 2% of lipid 'H, regardless of additions. When N-methylethanolamine (MME) was added, 1% in phosphatidyl dimethylethanolamine rose from 0.5 to 0.9 up to 2.5 to 4.1yo of lipid radioactivity. DME, N,N-dimethylethanolamine. a Absolute values are given in Table VI.
mainly determined by the availability of CDP-ester and diacylglycerol, substrates for the phosphotransferase enzymes.
The formation of CDP-ethanolamine is limited. by the supply of ethanolamine, unless exogenous ethanolamine is present at a concentration of about 0.05 mM. At higher concentrations phosphorylethanolamine accumulates but the rate of phosphatidylethanolamine synthesis and the level of CDP-ethanolamine re-3365 main unchanged, indicating that the step catalyzed by phosphorylethanolamine cytidylyltransferase becomes rate-limiting. The finding that the formation of phosphatidyl methylethanolamine from N-methylethanolamine is completely inhibited by ethanolamine but not by other phospholipid bases (Table IV) also suggests that the capacity at the cytidylyltransferase step is rate-limiting when exogenous bases are provided, since studies in vitro have indicated that the CDP-ester of N-methylethanolamine, but not of N ,N-dimethylethanolamine or choline, is formed by the phosphorylethanolamine cytidylyltransferase (35), although less efficiently than CDP-ethanolamine (34). In agreement with our conclusion, a study on the incorporation of [14C]ethanolamine into phosphatidylethanolamine in rat liver slices (40) indicated that the rate of this incorporation was determined at a step beyond the formation of phosphorylethanolamine. When the ratio between the pool sizes of phosphorylethanolamine and CDP-ethanolamine in freeze-clamped liver (16: 1, Ref . 7), the serum concentration of ethanolamine (Table I), and the data in Fig. 4 are considered, the phosphorylethanolamine cytidylyltransferase reaction appears to be a rate-limiting step in phosphatidylethanolamine synthesis under physiological conditions.
In addition, both the total availability of diacylglycerol and its fatty acid composition may significantly affect the rate of phosphatidylethanolamine synthesis, as illustrated by the effects of different fatty acids. Undoubtedly the regulation of diacylglycerol availability for phosphatidylethanolamine synthesis is complex, since it is influenced not only by factors governing the rate of diacylglycerol formation but also by the utilization of diacylglycerol for triacylglycerol and phosphatidylcholine synthesis. Measurements of the level of diacylglycerol and CDP-ethanolamine indicate that the stimulatory effect of unsaturated fatty acids on phosphatidylethanolamine synthesis and the inhibitory effect of lauric acid are both exerted via changes in the availability of diacylglycerol rather than via changes in CDP ethanolamine formation. However, the fatty acid supplements tested in this study are extreme in being composed of single fatty acids; to what extent physiological changes in total amount and composition of the fatty acid supply may affect the rate of phosphatidylethanolamine synthesis is therefore difficult to assess.
The Kennedy pathway of phosphatidylcholine synthesis (4) is analogous to that of phosphatidylethanolamine synthesis in most respects and our data on the effects of different substrates on phosphatidylcholine synthesis are also, with few exceptions, analogous to those on phosphatidylethanolamine synthesis. One such exception is that a much higher concentration of choline than of ethanolamine is required for maximal stimulation of phospholipid synthesis which might well be due to the rapid oxidation of choline to betaine. Although no direct experimental evidence for a rate-limiting role of the phosphorylcholine cytidylyltransferase reaction (analogous to that of phosphorylethanolamine cytidylyltransferase) is provided by the present study, several findings make such a role likely. First, the analogy between the effects of substrates on phosphatidylcholine and phosphatidylethanolamine synthesis noted in this study; second, the ratio between the pool sizes in liver of phosphorylcholine and CDP-choline which is very high (approximately 150: 1,Ref. 30); third, that the cholinephosphotransferase reaction utilizing CDP-choline appears to be reversible (25-28) and operate at near equilibrium, which is not expected for a rate-limiting reaction. Furthermore, recent studies on liver (41) and lung (42) tissue, where the incorporation of labeled choline into phosphatidylcholine in tissue slices and the activity of the enzymes of the Kennedy pathway have been followed during develop-