Effects of dietary conditions on the pool sizes of precursors of phosphatidylcholine and phosphatidylethanolamine synthesis in rat liver

Biochimica et Biophysics Elsevier

Acta 959 (1988) 1-8

BBA 52164

Effects of dietary conditions on the pool sizes of precursors of phosphatidylcholine and phosphatidylethanolamine synthesis in rat liver Lilian B.M. Tijburg, Martin Houweling, Math J.H. Geelen and Lambert M.G. van Golde Laboratory of Veterinaty Biochemistry, (Received

Key words:

Phosphatidylcholine;

University of Utrecht, Vtrecht (The Netherlands) 30 September

1987)

Phosphatidylethanolamine; HPLC; (Rat liver)

Choline

metabohte;

Ethanolamine

metabolite;

We developed a new method for the determination of choline- and ethanolamine-containing precursors of phosphatidylcholine and phosphatidylethanolamine after their separation by HPLC and we have studied the effects of different dietary conditions on the pool sizes of these metabolites in rat liver. Fasting for 48 h induced only a small decrease in the amounts of phosphatidylethanolamine and its water-soluble precursors. Upon refeeding with a high-sucrose, fat-free diet for 72 h, the levels of ethanolamine-containing compounds were only slowly restored. The effects of various dietary conditions on the amounts of phosphatidylcholine and its water-soluble precursors were much more pronounced. Fasting induced a sharp decrease, especially of the amount of cholinephosphate. However, the levels of phosphatidylcholine and the choline-containing precursors were rapidly restored upon refeeding for 24 h. Continued refeeding for an additional 48 h enhanced the cholinephosphate pool size to a level more than double that found in normally fed rats. The latter effect was accompanied by an inhibition of the enzyme CTF’: choline-phosphate cytidylyltransferase. The results are discussed in view of a possible regulatory mechanism that may balance the amounts of phosphatidylcholine and phosphatidylethanolamine.

Introduction Although the pathways involved in the biosynthesis of phosphatidylcholine (PC) and phosphatidylethanolamine (PE) have long since been established [l], many problems remain regarding the regulation of the biosynthesis of these compounds. Investigations are usually restricted to the determination of the velocity of the incorporation

Abbreviations: anolamine.

PC, phosphatidylcholine;

PE, phosphatidyleth-

Correspondence: L.B.M. Tijburg, Laboratory of Veterinary Biochemistry, P.O. Box 80.177, 3508 TD Utrecht, The Netherlands.

0005-2760/88/$03.50

0 1988 Elsevier Science Publishers

of radiolabeled precursors into intermediates and endproducts or to the measurement of enzyme activities in vitro. Determination of the pool sizes of water-soluble phospholipid precursors might be an important extension of the experimental procedures and form a fruitful contribution to a more complete understanding of the regulation of PC and PE synthesis. A method for determination of these precursors should involve a fast and complete separation of these compounds in combination with a sensitive and specific quantitative determination. Separation of choline, cholinephosphate, CDPcholine, ethanolamine, ethanolaminephosphate and CDPethanolamine has been accomplished in part or completely by ion-exchange

B.V. (Biomedical

Division)

2

chromatography [2-51 or paper chromatography [4]. These separations are time consuming and the number of samples that can be handled is limited. Employment of high-performance liquid chromatography (HPLC) does not have these disadvantages and might, therefore, be very suitable for determination of PC and PE precursors. Recently, Mazzola and Kent [6] introduced a rapid separation of choline and ethanolamine metabolites by ion-exchange HPLC. However, this procedure was only applied to tracer studies with labeled choline and ethanolamine. In this paper, we present a rapid separation of choline- and ethanolamine-containing precursors of PC and PE synthesis by HPLC in combination with a sensitive, quantitative determination of the pool sizes of these compounds. Earlier experiments from this laboratory and others [7,8] have demonstrated that the effects of fasting followed by refeeding on the synthesis of hepatic PC were quite different from those on PE synthesis. Upon fasting, the level of PC decreased faster and to a larger extent than that of PE. Moreover, refeeding induced a rapid restoration of the amount of PC, whereas the amount of PE was restored only slowly [8]. In an attempt to explain the changes in the amounts of PC and PE under these different dietary conditions, we applied the above-mentioned, new procedure to investigate the effects of fasting and fasting followed by refeeding on the pool sizes of choline and ethanolamine metabolites in rat liver. Materials and Methods Materials. [ methyl-14C]Choline, [methylI4Clcholinephosp hate and CDP[ methyl-‘4C]cho-

line were purchased from Amersham International (Amersham, U.K.). [1,2-i4C]Ethanolamine and [v“P]ATP were obtained from New England Nuclear (Dreieichenhain, F.R.G.). CDP[1,2-14C]ethanolamine was supplied by ICN Radiochemicals (Irvine, U.K.) and [1,2-‘4C]ethanolaminephosphate was prepared as described by Sundler [9]. Choline kinase from yeast was obtained from Boehringer (Mannheim, F.R.G.). Dansylchloride was supplied by Janssen Chimica (Beerse, Belgium) and AGl-X8 anionic resin (200-400 mesh,

formate form) by Bio-Rad (Richmond, CA). All other chemicals were of analytical grade and obtained from Baker Chemical Co. (Deventer, The Netherlands). Animals and diets. Male Wistar rats, weighing 180-200 g, which had free access to water, were divided into four groups of different nutritional states: (a) rats fed, ad libitum, a regular chow pellet diet, (b) rats fasted for 48 h, (c) rats fasted for 48 h and then refed a high-sucrose, fat-free diet for 24 h and (d) rats fasted for 48 h and subsequently refed a high-sucrose, fat-free diet for 72 h. The composition of the regular chow pellet diet and the high-sucrose, fat-free diet have been described in detail previously [lo]. After perfusion of the livers in situ in order to free them of blood, a part of the liver was homogenized in 4 ~01s. of ice-cold water and used immediately as starting material for determination of the pool sizes of PC, PE and their precursors. Another part was homogeniz,ed in 4 ~01s. homogenization buffer containing 0.145 M NaCl/lO mM Tris-HCl (pH 7.4)/l mM EDTA/lO mM NaF for determination of the activity of CTP : choline-phosphate cytidylyltransferase (EC 2.7.7.15). Extraction

of lipids and water-soluble precursors.

An aliquot of the homogenate in water was extracted with 3.75 ~01s. chloroform/methanol (1: 2) [ll]. Trace amounts of labeled choline and ethanolamine intermediates were routinely added to the extraction mixture to calculate recoveries. Samples were separated in an aqueous and an organic phase by the addition of 1 vol. water and 1 vol. chloroform. The lower chloroform phase was washed twice with methanol/water (5 : 4). The upper phases, containing the water-soluble intermediates, were combined and evaporated to dryness. The residue was dissolved in 500 ~1 25% (w/v) trichloroacetic acid, vortexed thoroughly and subsequently extracted with 625 yl of a rnixture of 1,1,2-trichlorotrifluoroethane/ tri-n-octylamine (1: 1) to remove the excess trichloroacetic acid. This neutralized solution was used for HPLC of the water-soluble precursors of PC and PE as described below. The chloroform phase was evaporated to dryness and used for determination of the amounts of phospholipid [12] and diacylglycerol as described earlier [13].

3

Separation of phospholipid precursors by high-performance liquid chromatography. For the chromatographic separation of the intermediates of PC and PE synthesis an LKB 2150 liquid chromatograph was employed, which was equipped with a 200 ~1 loop injector. Detection of separated compounds was carried out using either an Iomess radioactivity monitor (Ramona-LS) equipped with a solid scintillation flow cell, or an LKB fixedwavelength Uvicord (model 2158). An aliquot of the neutralized solution, containing water-soluble precursors, was applied to an anion-exchange column (LKB TSK 545 DEAE, 7.5 X 150 mm) and eluted with 0.01 M sodium phosphate buffer (pH 5.6) at a flow rate of 0.9 ml/mm This step resulted in a separation into three groups consisting of choline plus ethanolamine, cholinephosphate plus ethanolaminephosphate and CDPcholine plus CDPethanolamine, respectively. The fractions containing free choline plus ethanolamine were pooled, evaporated to dryness and resuspended in water. Separation of choline and ethanolamine was, subsequently, accomplished on a cation-exchange column (Partisil 10 SCX, 4.6 x 250 mm) that was eluted with 0.3 M potassium phosphate (pH 4.3) at a flow rate of 1 ml/mm. Choline and ethanolamine content were, subsequently, assayed as described below. The fractions containing cholinephosphate plus ethanolaminephosphate were combined, evaporated to dryness and hydrolyzed in 500 ~1 distilled 6 M HCl at 110°C for 18 h. This treatment resulted in almost complete hydrolysis of both phosphate esters. After evaporation of the acid under a stream of N,, the residues were dissolved in water and choline was separated from ethanolamine as described above. The fractions containing the CDP derivatives of choline and ethanolamine were evaporated to dryness and redissolved in 500 ~1 water. An aliquot of this sample was applied to a Lichrosorb RP-18 column (4 x 250 mm) that was eluted with 0.1% phosphoric acid, adjusted to pH 2.5 with NaOH at a flow rate of 0.6 ml/mm. CDPcholine and CDPethanolamine, which were clearly separated in this system, were detected spectrophotometritally at a wavelength of 280 nm. The CDPethanolamine pool size could be determined directly by

comparison with a standard curve of CDPethanolamine. The detection limit was 100 pmol and the absorbance was linear with the amount of CDPethanolamine from 0 to 10 nmol. However, the CDPcholine peak obtained in this separation contained one or more impurities. Therefore, these CDPcholine-containing fractions were pooled, evaporated to dryness, hydrolyzed in 6 M HCl as described above and assayed for choline content. Assay of choline and ethanolamine. The cholinerepresenting either free containing fractions, choline or choline originating from cholinephosphate and CDPcholine, were evaporated to dryness and dissolved in a known amount of water. Aliquots of this solution, containing up to 10 nmol of choline, were assayed enzymatically for choline content as described by Post et al. [3]. Briefly, the incubation mixture comprised, in a total vol. of 150 ~1, 100 mM glycylglycine (pH 9.2)/4 mM MgC1,/6 mM ATP/l PCi [Y-~~P]ATP/0.5 U/ml choline kinase. The samples were incubated for 1 h at 37” C and the reaction was terminated with an equal amount of ethanol. Labeled cholinephosphate was separated from [y32P]ATP exactly as described by Post et al. [3]. The radioactivity in cholinephosphate was determined and choline content was calculated by comparison with a standard curve prepared with known amounts of choline. Ethanolamine-containing fractions (free ethanolamine or ethanolamine originating from ethanolaminephosphate) were dried and the residues were treated with dansyl chloride. The dansyl derivatives were extracted, separaied and quantified as described by Sundler and Akesson [2]. Other methoa!s. A 20% homogenate in 0.145 M NaCl/lO mM Tris-HCl (pH 7.4)/l mM EDTA/lO mM NaF was separated in a cytosolic and a microsomal fraction by centrifugation as described by Pelech et al. [14]. The determination of the activity of CTP : choline-phosphate cytidylyltransferase was performed exactly as described in Ref. 14. Protein content was determined by the method of Lowry et al. [15] using bovine serum albumin as standard. Results are shown as mean f S.D. of three different animals. Significance was calculated using the Student’s t-test.

4

RCdtS Separation and de~errni~~~~onof ~hos~hoiip~d precursors

The protocol that we used for the separation of choline, ethanolamine, cholinephosphate, ethanolaminephosphate, CDPcholine and CDPethanolamine is outlined in Fig. 1. The first step involved chromatography on an HPLC anion-exchange column. Choline and ethanolamine eluted as a single peak from this column. They were, however, clearly separated from their phosphate esters as well as from their CDP derivatives. Although cholinephosphate and ethanolaminephosphate were partially resolved, we decided to combine these fractions. CDPcholine and CDPethanolamine were the last to emerge from the column. As we did not always obtain a satisfactory separation between these nucleotides, particularly when larger amounts of sample had to be applied on the column, the fractions containing CDPcholine and CDPeth~ola~ne were also pooled. As shown in Fig. 1, the time of one run is less than 30 min. Free choline and free ethanolamine could easily be separated within 25 min by a second HPLC step using a cation-exchange column. The pooled fractions choline- and ethanolaminephosphate were hydrolyzed by heating in 6 M HCl, followed

HPLC Rfw mPemmine hydoiysis ,

I uv

(28Omn)

Fig. 1. Outline of the procedure for the separation and quantification of water-soluble precursors of PC and PE.

i--y---A,

0

) 4

6

TIME

,

t 6

L

(

,

10

tminf

Fig. 2. Elution pattern of the separation of C~Peth~o~a~ne and CDPcholine by reversed-phase HPLC. The fractions containing CDPcholine and CDPethanolamine (fractions 5 and 6, Fig. 1) were combined and applied to a Lichrosorb RP-18 column and eluted with 0.1% phosphoric acid, adjusted to pH 2.5 with NaOH, at a flow rate of 0.6 mf/min. (1) and (4), u~denti~ed; (2), CDPethanoi~ne; (3), CDPcholine.

by separation of the liberated choline and ethanolamine on the cation-exchange column. The acid hydrolysis of the phosphate esters could not be replaced by enzymatic hydrolysis with either alkaline or acid phosphatase, as breakdown products of these enzymes appeared to interfere with the subsequent determination of choline with choline kinase. An excellent separation between CDPcho~ne and CDPethanolamine was achieved by chromatography on a reversed-phase Cl8 column (Fig. 2). The direct spectrophotometric detection and quantification of CDPeth~ol~ne that we employed is faster and more sensitive, with a detection limit of 100 pmol, than the Odansylation method described by Sundler and Akesson [2]. Unfortunately, the CDPcholine fraction contained one or more contaminating compounds (Fig. 2). Therefore, CDPcholine was hydrolyzed in 6 M HCl and determined enzymatically as choline. Enzymatic determination of choline with choline kinase is a very sensitive method. In our hands, the detection limit was 0.5 nmol and the reaction was linear from 0 to 10 nmol. In contrast to earlier studies 13,161, we found that the commercially available choline kinase from yeast was not completely specific for choline, but also showed some activity towards ethanolamine (results not shown). A similar finding was recently

5

reported by Warden and Friedkin [17]. These observations are in line with a recent report by Ishidate et al. [18] of a complete co-purification of choline kinase and ethanolamine kinase from rat kidney. Betaine did not interfere with the determination of choline (not shown). While the assay of choline by choline kinase requires the removal of ethanolamine, the latter metabolite can be determined by the dansylation method without the necessity of removing the choline. The overall recoveries of the entire procedure were 55% for choline and cholinephosphate, 40% for CDPcholine, 75% for ethanolamine and 65% for ethanolaminephosphate and CDPethanolamine. The effects on PE and its precursors of fasting and fasting followed by refeeding In liver, the CDPethanolamine pathway is an important route for the synthesis of PE. This pathway includes phosphorylation of ethanolamine and subsequent conversion of ethanolaminephosphate to CDPethanolamine, which then reacts with diacylglycerol to form PE [I]. In agreement with an earlier study from this laboratory [8], we found that fasting induced a sharp decrease of the amount of diacylglycerols in liver (7.3 + 0.5 pmol/liver in normally fed rats; 3.1 + 0.5 pmol/liver in rats fasted for 48 h). Refeeding with a high-sucrose, fat-free diet induced an immediate increase of the amount of diacylglycerols (14.1 + 1.3 pmol/liver after 24 h

TABLE

refeeding and 16.9 + 3.7 pmol/liver after 72 h refeeding). In the present study we focused our attention on the effects of fasting and refeeding on the level of PE, PC and, particularly, their watersoluble precursors. The effects of different dietary conditions on the amounts of PE and its watersoluble precursors are shown in Table I. Fasting for 48 h induced a decrease in the amount of PE to 69% of that in normally fed rats. The increase upon refeeding a high-sucrose, fat-free diet for 24 h was only 14%. Even after 72 h refeeding, the amount of PE was still significantly lower than that in normally fed rats (n = 3, P < 0.02). Upon fasting for 48 h, the amount of liver ethanolamine tended to decrease (Table I). It is striking that refeeding a high-sucrose, fat-free diet for 24 h hardly changed the ethanolamine pool size. However, after continued refeeding for another 48 h, the amount of ethanolamine was indeed restored. Table I also shows that deprivation of food resulted in a small, albeit non-significant decline of the amount of ethanolaminephosphate. Refeeding for 24 h induced a further decrease to 36% of the amount in normally fed rats. After 72 h this effect was reversed and the ethanolaminephosphate pool size was slightly enhanced, although still smaller than in control animals. The effects of different dietary states on the CDPethanolamine pool size are different from those on ethanolamine and ethanolaminephosphate. Fasting induced a considerable reduction of the CDPethanolamine pool size to 44% of the value found in normally fed rats. After 3 days refeeding, the amount of

I

THE EFFECTS OF DIFFERENT ETHANOLAMINE

NUTRITIONAL

STATES

ON THE AMOUNTS

OF PRECURSORS

OF PHOSPHATIDYL-

Ethanolamine and ethanolamine metabolites were separated and quantified as described in Materials and Methods. Each value represents the mean+ SD. in pmol per liver of three different rats. The weights of the livers were lO.O+ 0.1 g (normally fed), 5.9kO.l g (fasted), 11.8 f 0.4 g (fasted and 24 h refed) and 10.3 + 0.6 g (fasted and 72 h refed). a Significantly different from normally fed rats, P < 0.05; b Significantly different from fasted rats, P < 0.05.

Ethanolamine Ethanolaminephosphate CDPethanoIamine PE

Normally fed

48 h fasted

48 h fasted, 24 h refed

48 h fasted, 72 h refed

2.23 f 0.53 7.11+ 2.27 0.62 f 0.08 117.9 *5.1

1.30+0.57 4.90 f 0.90 0.27 f 0.02 a 81.0 k2.6 a

1.48 f 0.06 2.53 + 0.48 a*b 0.70*0.01 b 92.1 f 1.6 aVb

2.76 4.09 0.85 105.3

f 0.54 b * 0.37 f 0.04 a,b *0.9 a.b

6 TABLE II THE EFFECTS OF DIFFERENT YLCHOLINE

NUTRITIONAL

STATES ON THE AMOUNTS OF PRECURSORS

OF PHOSPHATID-

Choline and choline metabolites were separated and quantified as described under Materials and Methods. Each value represents the mean* S.D. in pmol per liver of three different rats. The weights of the livers are listed in Table I. The ratio PC/PE was calculated from the results in Table I and this table. a Significantly different from normally fed rats, P < 0.05; b Significantly different from fasted rats, P c 0.05.

Choline Cho~ephosphate CDPcholine PC PC/PE

Normally fed

48 h fasted

48 h fasted, 24 h refed

48 h fasted, 72 h refed

2.11 kO.26 16.2 f3.2 0.50 * 0.09 202.0 f1.3 1.7

1.58+0.11 3.19 & 0.97 a 0.28 + 0.03 = 124.4 +4.0 a 1.5

2.26+ 11.3 + 0.33 * 186.4 + 2.0

3.55 f 35.7 i 0.47* 231.1 + 2.2

CDPethanolamine was 37% higher than in normally fed rats (Table I). The effects on PC and its water-~ol~le precursors of fasting and jmting followed by refeeding In mammalian liver, PC is synthesized mainly via the CDPcholine pathway. The effects of various nutritional states on the pool sizes of PC and its water-soluble precursors are shown in Table II. Fasting for 48 h decreased the amount of hepatic PC to 62% of that in normally fed rats. Upon refeeding the amount of PC was enhanced by 49%. After continued refeeding for 72 h the PC pool size was sig~ficantly higher than in normally fed rats (Table II). These observations (Tables I and II) demonstrate that the effects of various nutritional conditions on the amount of liver PC are more pronounced than those on the amount of PE. Table II shows that upon fasting, the ratio PC/PE decreased from 1.7 to 1.5. Refeeding induced an immediate reversal of this effect and after 3 days of refeeding the ratio PC/PE was 2.2. The effects of different dietary conditions on the pool sizes of the intermediates of PC synthesis are shown in Table II. Choline is, in contrast to ethanolamine, an essential dietary component. It is obvious from the results in Table II that deprivation of food and, consequently, choline resulted in a decrease of all choline metabolites. The pool size of cholinephosphate especially showed a tremendous decrease. Readministration of food

0.39 1.7 b 0.03 a 12.6 b

0.34 .kb 1.9 &b 0.07 b 17.2 qb

induced a rapid enhancement of the amounts of choline and cholinephosphate, while the CDPcholine pool was replenished more slowly. Continued refeeding induced an enormous accusation of cholinephosphate. After 3 days refeeding, the amount of cholinephosphate was more than double that in normally fed rats, whereas the level of CDPchohne was approximately the same as in control animals. The accumulation of cholinephosphate might be an indication of a diminished conversion of cholinephosphate to CDPcholine,

TABLE III ACTIVITY OF CTP : CHOLINE-PHOSPHA~ CYTIDYLYLTRANSFERASE IN LIVERS OF RATS OF DIFFERENT NUTRITIONAL STATES Livers were homogeuized and subsequently separated by centrifugation in a microsomal and a cytosolic fraction as described under Materials and Methods. Values are the mean *S.D. of livers of three rats, each assayed in duplicate. a Significantly different from normally fed rats, P < 0.01; b Significantly different from fasted rats, P < 0.01. Nut~tion~ state

Spec. act. (nmol-min-‘erng-’ microsomes

cytosol

fed 48 h fasted 48 h fasted, 24 h refed 48 h fasted, 72 h refed

0.34f0.01 0.82 f 0.08 *

0.11*0.02 0.24 rt 0.03 a

0.41 f 0.08 b

0.06 f 0.01 4b

0.20 f 0.03 a,b

0.07 * 0.02 b

Normally

protein)

i.e., an inhibition of the key regulatory enzyme of the pathway, CTP : choline-phosphate cytidylyltransferase [19]. There is accruing evidence from studies on various cell types that the active form of cholinephosphate cytidylyltransferase is located in the endoplasmic reticulum [19-211. Table III shows the activity of choline-phosphate cytidylyltransferase in the microsomal as well as the cytosolic fractions of livers from animals in different nutritional states. Deprivation of food resulted in an enhancement of the microsomal activity, an effect that was in line with the decrease of the level of cholinephosphate (Table II). Refeeding with a high-sucrose, fat-free diet resulted in an immediate decline of the choline-phosphate cytidylyltransferase activity in the microsomes. After 3 days refeeding, the microsomal cytidylyltransferase activity was only 59% of the value found in normally fed rats. This decline is in line with the enormous enhancement of the amount of cholinephosphate (Table II). The changes in cytosolic cytidylyltransferase paralleled those of the microsomal activity (Table III). Discussion In this paper we present a procedure for the complete separation of precursors of PE and PC synthesis by HPLC, in combination with a sensitive method for measuring the chemical amounts of these compounds. The procedures described earlier by other investigators [3-5,22,23] suffered from various disadvantages and were rather time consuming. To our knowledge, this is the first report of a procedure to determine the chemical amounts of PC and PE precursors, following a complete separation of these compounds by HPLC. The employment of HPLC to separate choline and ethanolamine metabolites offers several advantages. It is much faster than conventional column chromatography, the recovery of the various compounds in each HPLC step is between 90 and 95% and, as we used an isocratic separation method, frequent regeneration of the column is not necessary. The increased sensitivity of the methods of determination of choline, ethanolamine and their metabolites, described in the pre-

sent study, implies that this procedure might not only be applied to tissue homogenates, but may also be very suitable for all kinds of isolated cell types. The concentration of free choline and ethanolamine found in the present study are in good agreement with those reported by Sundler et al. [22,23] in freeze-clamped liver. Our results with respect to ethanolamineand choline-containing metabolites (Tables I and II, respectively) are comparable to those reported by Korniat and Beeler [24] and Sundler [22,23] for freeze-clamped rat liver, except that Sundler et al. [22] found CDPcholine content slightly lower than that found by Komiat and Beeler [24] and in the present study. The results in Table I show that alterations in dietary conditions provoke parallel changes in the amounts of ethanolamine and ethanolaminephosphate, i.e., decrease by fasting followed by a slow replenishment of the pools by refeeding. The observation that the amount of ethanolaminephosphate after 24 h of refeeding was even lower than that in liver of a fasted rat, and after 72 h refeeding was not restored to the level found in control rats, can be explained by a shortage of ethanolamine or, more likely, by the relatively low activity of ethanolamine kinase compared to that of ethanolamine-phosphate cytidylyltransferase [8]. It is unlikely that the delayed recovery of PE upon refeeding is due to a limited amount of CDPethanolamine, as the pool size of this intermediate after 72 h refeeding is even higher than that in control livers (Table I). Although the amount of diacylglycerols is a factor of two higher than under normal feeding condition, it should be pointed out that the diacylglycerols in livers of refed animals consist primarily of monoenoic species [8]. As it was also demonstrated that ethanolamine-phosphotransferase shows a preference for highly unsaturated diacylglycerol species [25], it is plausible that the available diacylglycerols upon refeeding are an unsuitable substrate for this enzyme. This might explain the delayed recovery of the amount of PE despite the enhanced level of CDPethanolamine. It should be mentioned, however, that decarboxylation of phosphatidylserine may also contribute to the levels of PE [26,27], although the relative importance of this pathway

8

in liver has not yet been established. Table II demonstrates that the amount of hepatic PC after 24 h of refeeding is almost as high as that in normally fed rats. The formation of PC is obviously not disturbed by the nature of the available diacylglycerols. On the contrary, in spite of a diminished activity of microsomal cholinephosphate cytidylyltransferase after 72 h refeeding, this did not result in an inhibition of PC synthesis, since the amount of PC increased upon refeeding for 72 h. (Tables II and III). There might be two possible explanations for the discrepancy between the increasing amount of PC and the inhibition of the rate-determining enzyme of the pathway, choline-phosphate cytidylyltransferase. Activation of phosphatidylethanolamine-N-methyltransferase might result in an enhanced conversion of PE to PC. This might explain the delayed recovery of the amount of PE, but does not explain why cholinephosphate cytidylyltransferase is inhibited concomitantly. An alternative and more attractive explanation was recently suggested by Kennedy and collaborators [28]. These investigators demonstrated that the composition of phospholipids in Escherichia coli was subject to regulation by a mechanism which involved independent feed-back regulation and, therefore was remarkably constant. If extrapolated to phospholipid synthesis in mammalian liver, this model might imply that the ratio of PE and PC, as indicated in Table II, would be kept as constant as possible. It might be speculated that the delayed increase of PE upon refeeding (Table I), possibly due to a lack of suitable diacylglycerols, is a signal for slowing down PC synthesis by inhibition of choline-phosphate cytidylyltransferase. It will be of great interest to investigate whether such feedback inhibition may be a general phenomenon of the synthesis of PC and PE in rat liver. Acknowledgements

These investigations were supported by the Netherlands Foundation for Chemical Research (SON) with financial aid from the Netherlands Organization for the Advancement of Pure Research (ZWO).

References 1 Kennedy, E.P. (1991) Fed. Proc. 20, 934-940. 2 Sundler, R. and Akesson, B. (1975) J. Biol. Chem. 250, 3359-3367. 3 Post, M., Batenburg, J.J., Smith, B.T. and Van Golde, L.M.G. (1984) B&him. Biophys. Acta 795, 552-557. 4 Salerno, D.M. and Beeler, D.A. (1973) B&him. Biophys. Acta 326, 325-338. 5 Haines, D.S.M. and Rose, C.J. (1970) Can. J. Biochem. 48, 885-892. 6 Mazola, G. and Kent, C. (1984) Anal. Biochem. 141, 137-142. 7 Park, C.E., Marai, E. and Mookerjea, S. (1972) B&him. Biophys. Acta 270, 50-59. 8 Groener, J.E.M., Klein, W. and Van Golde, L.M.G. (1979) Arch. B&hem. Biophys. 198, 287-295. 9 Sundler, R. (1975) J. Biol. Chem. 250, 8585-8590. 10 Groener, J.E.M. and Van Golde, L.M.G. (1977) B&him. Biophys. Acta 487, 105-114. 11 Bligh, E.G. and Dyer, W.J. (1959) Can. J. B&hem. Physiol. 37, 911-917. 12 Bartlett, G.R. (1959) J. Biol. Chem. 234, 466-468. L.B.M., Schuurmans, E.A.J.M., Geelen, M.J.H. 13 Tijburg, and Van Golde, L.M.G. (1987) Biochim. Biophys. Acta 919,49-57. 14 Pelech, S.L., Pritchard, P.H. and Vance, D.E. (1981) J. Biol. Chem. 256, 8283-8286. N.J., Farr, A.L. and Randall, 15 Lowry, O.H., Rosebrough, R.J. (1951) J. Biol. Chem. 193, 265-275. D.R., Gerber, N., Pflueger, A.B. and Zweig, M. 16 Haubrich, (1981) J. Neurochem. 36, 1409-1417. Biophys. 17 Warden, C.H. and Friedkin, M. (1984) B&him. Acta 792, 270-280. K. and Nakazawa, Y. (1985) Bio18 Ishidate, K., Furusawa, chim. Biophys. Acta 836, 119-124. P.H., Brindley, D.N. and Vance, 19 Pelech, S.L., Pritchard, D.E. (1983) J. Biol. Chem. 258, 6782-6788. 20 Sleight, R. and Kent, C. (1983) J. Biol. Chem. 258,831-835. 21 Weinhold, P.A., Feldman, D.A., Quade, M.M., Miller, J.C. and Brooks, R.L. (1981) B&him. Biophys. Acta 665, 134-144. 22 Sundler, R., Arvidson, G. and Akesson, B. (1972) Biochim. Biophys. Acta 280, 559-568. 23 Sundler, R. (1973) Biochim. Biophys. Acta 306, 218-226. 69, 24 Komiat, E.K. and Beeler, D.A. (1975) Anal. B&hem. 300-305. Biophys. Acta 218, 249-258. 25 Kanoh, H. (1970) B&him. Biophys. Acta 833, 3966405. 26 Bjerve, K.S. (1985) B&him. 27 Vance, J.E. and Vance, D.E. (1986) J. Biol. Chem. 261, 448664491. 28 Jackson, B.J., Gennity, J.M. and Kennedy, E.P. (1986) J. Biol. Chem. 261, 13464-13468.