The importance of the lysophosphatidylcholine and choline moiety of bile phosphatidylcholine in lymphatic transport of fat

364

Biochimica et Biophysics Acta, 528 (1978) 364-372 @ Elsevier/North-Holland Biomedical Press

BBA 57153

THE IMPORTANCE OF THE LYSOPHOSPHATIDYLCHOLINE AND CHOLINE MOIETY OF BILE PHOSPHATIDYLCHOLINE IN LYMPHATIC TRANSPORT OF FAT

PATRICK TSO, JOSEPH LAM and WILFRED J. SIMMONDS Department of Physiology, 6009 (Australia)

University of Western Australia, Nedlands, Western Australia,

(Received October 17th, 1977)

Summary A luminal supply of biliary phosphatidylcholine is important in the translocation of absorbed fat into lymph and in the amount and composition of phosphatidylcholine concurrently synthesized. This study was undertaken to determine whether the effect was due to absorbed lysophosphatidylcholine, to a specific (1-palmitoyl) biliary lysophosphatidylcholine or to extra choline supplied by lysophosphatidylcholine. Rats with bile fistulae and thoracic duct lymph fistulae were given test meals of oleic acid and monoolein (molar ratio 2 : 1) infused duodenally for 8 h. Addition of choline chloride to the test meal increased lymphatic output of triglyceride and phospholipid but not to values found previously in rats with supplements of bile phosphatidylcholine or with bile ducts intact. Addition of dioleoyl phosphatidylcholine increased triglyceride and phospholipid output to values found in rats with intact bile ducts. Since dioleoyl phosphatidylcholine was as efficient as biliary phosphatidylcholine it was concluded that a luminal supply of l-palmitoyl lysophosphatidylcholine was not essential. It seemed likely from the smaller effect of supplemented choline and from the fatty acid composition of lymph phosphatidylcholine that the essential requirement was a supply of absorbed lysophosphatidylcholine for rapid reacylation to phosphatidylcholine.

Introduction Luminal phosphatidylcholine may have an important role in fat transport by providing the necessary surfactant for chylomicron coating. O’Doherty et al. [l] found that the addition of egg phosphatidylcholine to a gastric test meal containing lipid and bile salts markedly reduced the mucosal accumulation of absorbed lipid in bile fistula rats. Employing a similar technique but measuring both gastric emptying and mucosal uptake, Tso and Simmonds [2] found no

365

differences in mucosal lipid content in bile fistula rats with or without phosphatidylcholine supplemented, nor did Rogers during steady infusions of lipid in bile salt when phosphatidylethanolamine replaced phosphatidylcholine [ 3J. However, none of the investigators monitored the transport of absorbed lipid in lymph. Using lymph- and bile-fistula rats receiving duodenal infusions of micellar monoolein and oleic acid we found that intraluminal biliary phosphatidylcholine enhanced the translocation of high doses of absorbed fat into lymph and influenced the amount and type of phosphatidylcholine synthesized [4]. Although there is some dissent [5-S], it is most likely that bile phosphatidylcholine is only absorbed after hydrolysis by pancreatic phospholipase AZ to 1-acyl lysophosphatidylcholine, rich in palmitate (16 : 0) and some stearate (18 : 0) [9]. This study was conducted to test whether the role of bile phosphatidylcholine in restoring normal fat transport in bile fistula rats is due to the lysophosphatidylcholine backbone, to a specific lysophosphatidylcholine with saturated fatty acids at the cy-position or to the choline moiety of the phosphatidylcholine molecule. By replacing bile phosphatidylcholine with a synthetic dioleoyl phosphatidylcholine, the importance of a specific lysophosphatidylcholine with saturated fatty acid was tested in bile fistula rats. The relative importance of the lysophosphatidylcholine backbone and of choline from the phosphatidylcholine molecule was tested by replacing bile phosphatidylcholine with choline chloride. Materials and Methods Animals. Male adult albino rats of the Wistar strain (200-250 g) were fasted overnight before operation. Under ether anaesthesia, thoracic duct and common bile duct were cannulated and a tube introduced into the duodenum for infusion. The thoracic duct was cannulated according to the method of Bollman et al. [lo]. Silicone tubing (outer diameter 1.6 mm, tipped inside with clear vinyl tubing, outer diameter 1.0 mm) was introduced about 2 cm down the duodenum through the fundus of the stomach. The tubing was secured in the duodenum through a transmural suture and the fundal incision was closed by a purse string suture. The common bile duct was cannulated just below the two hepatic ducts with silicone tubing (outer diameter 1.0 mm) tipped with polyethylene tubing (outer diameter 0.61 mm). Post-operatively, the animals were infused via the duodenal tube at a rate of 2.9 ml/h with saline (145 mM NaC1/4 mM KCl). The operated animals were allowed to recover for at least 36 h in restraint cages kept in a warm room (approx. 3O”C), before lipid infusions were given. Experimental plan. On the day of the experiment, lipid was infused in the same volume of fluid as post-operatively. There were three groups of rats: (1) bile fistula, group A; (2) bile fistula with dioleoyl phosphatidylcholine supplemented, group B; (3) bile fistula with choline chloride, group C. The bile fistula group received a mixture of oleic acid and monoolein in molar ratio 2 : 1, total fatty acid infused per h was 173 mol. In group B, the lipid dose was the same as the rats in group A except that dioleoyl phosphatidylcholine, 10 pmol/h, replaced an equivalent amount of fatty acids, 20 pmol/h, in the oleic acid and

366

monoolein mixture. In group C, the lipid dose was the same as in group A and choline chloride 10 pmol/h was added to the infusate. All groups received sodium taurocholate, 55 pmol/h, to ensure an adequate, continuous luminal supply of bile salts. Preparation of infusate. In all experiments, the lipid was labelled with Ci/mol of fatty acid). On the day of experi[l-14C]oleic acid (approx. 6 + 10e4 ment, stock lipid solutions were mixed and the solvent evaporated under a stream of nitrogen. Sodium taurocholate (19 t.lmol/ml, dissolved in phosphatebuffered saline; 6.75 mM Na~HPO~/~6,5 mM NaH~PO~/ll5 mM NaC1/4 mM KCl) was then added to make up to the required lipid concentration and the mixture sonicated. Both infusates formed a stable, slightly milky mixture; the pH of the infusate was adjusted to 6.4. Samples were analyzed for glyceride ester and radioactivity at the beginning and end of infusion (reproducibility +5%). ~~per~~enta~ procedure. Lymph was collected into precooled tubes containing 2 ml methanol for 2 h before lipid infusion (methanol was added to prevent any possible enzymatic degradation of lipids). The lymph sample was analysed as control. Lymph samples were collected as above between O-2 and 2-4 h and each further hour during the 8 h infusion. At the end of the infusion, the animal. was anaesthetized and killed by exsanguination. In some animals, the stomach, upper small intestine, lower small intestine and the colon were tied off separately and the contents of each eluted with 2 X 5-ml aliquots of 5 mM sodium taurocholate. Aliquots were taken from these washings and radioactivity determined in a water-miscible scintillant. Mucosa from the upper and lower intestinal wall was scraped off with a glass slide and lipid extracted with toluene/ethanol (2 : 1, v/v) [ 111. Aliquots were taken for both radioactivity determination and separation into lipid classes by thin-layer chromatography using the solvent system light petroleum/diethyl ether/acetic acid (75 : 15 : 0.6, v/v). ~~teriu~. [1-‘4C]Oleic acid was purchased from the ~adiochemical Centre, Amersham and was 97% pure. Unlabelled oleic acid was from May and Baker, Victoria, Australia and monoolein from Calbiochem, Calif. U.S.A. The samples were chromato~aphically pure, The fatty acid composition of the oleic acid as determined by gas-liquid chromatography was: 14 : 0, 3%; 16.0, 4.6%; 16 : 1, 8.1%; 18.1,76.2%; 18.2. 8.1%. Monoolein was found to contain more than 86% oleic acid; the rest was palmitate and palmitoleate. Dioleoyl phosphatidylcholine was synthesized by direct acylation of Liu-glyceryl phosphoryl choline with a mixture of oleic anhydride and pota~ium oleate [ 12 3. Chromato~apbically pure glyceryl phosphoryl choline was obtained by deacylation of purified egg yolk lecithin with a methanolic tetrabutylammonium hydroxide solution 1133, Oleic anhydride was freshly prepared by reacting oleic acid (Koch Light and Co., London, England, more than 98% pure) and dicyclohexylcarbodiimide [14]. Dioleoyl phosphatidylcholine was purified by silicic acid column chromato~aphy [15]. The purified dioleoyl phosphatidylcholine was found to be chromatographically pure using chloroform/methanol/water (65 : 25 : 4, v/v) on silica gel G. Fatty acid composition of the dioleoyl phosphatidylchol~e was shown by gas-liquid chromatography to be more than 98% oleic acid.

367

Sodium taurocholate was synthesized by the method of Lack et al. [ 161, and checked using propionic acidfisoamyl acetate~water/~-propanol (15 : 20 : 5 : 10, v/v) on silica gel G. Choline chloride was supplied by ~~biochem Calif. U.S.A. All other reagents used were of analytical grade; ethanol was redistilled before use. Analytical techniques. Lymph lipid was extracted by the method of Bligh and Dyer [ 171 and dissolved in chloroform with antioxidant. Aliquots were also taken for the separation of phosphatidyl~holine from other lipids using chloroform/methanol/water (65 : 25 : 4, v/v) on silica gel G plates, 0.5 mm thick [ZO]. Phosphatidylcholine was identified against a standard (supplied by Serdary Research Laboratories, Ontario, Canada) after spraying with DittmerLester spray [ 211. After the phosphatidylcholine spot was scraped off the plate, the remaining gel was scraped into a column and the lipid eluted by approx. 20 ml ~hloroform/methanol (2 : 1, v/v). Both the phosphatidylcholine spot and the rest of the eluted lipids were assayed for phosphorus content [ 191. Lymph phosphatidylcholine was isolated and methyl esters of fatty acids were prepared [22]. The fatty acid methyl esters were then analysed in a column of 5% EGSS-X on 100-120 mesh Aeropak 30 in a Varian Aerograph 1400 gas-liquid ~hromato~aphy machine. Radioactivity determination. Radioactivity of extracted lymph and mucosal lipid was determined after the solvent was evaporated under N2 and 10 ml scintilland (4 g 2,5diphenyloxazole/O.O5 g 4-bis-2-(4-methyl-5-phenyloxazolyl)benzene in 1 1 toluene) was added. The radioactivity of unextracted lymph and intestinal contents was measured by adding to a water-miscible scintillant, “Aquasolv” (Beckman, Fullerton, U.S.A.). Counting was done for 10 min in a Nuclear Chicago Isocap 300 counter (counting efficiency 92%). Statistics. Where applicable, the results from different groups of experiments were tested by Student’s t-test for independent variables [ 231; differences were considered significant if the probability of the difference occurring by chance was less than 5 in 100 (P < 0.05). ResuIts

Lymph flow The lymph flow was not different in the three groups of rats, increasing from a fasting value approx, 2 ml/h to a steady rate of approx. 3 ml/h after 5-6 h of lipid infusion.

Lipid output Phospholipid.

The fasting phospholipid output in lymph was low (approx. 0.35 pmolfh in all groups of rats, Fig. 1B). During lipid infusion, the lymph phospholipid increased markedly and reached a steady output after 5 h. Taking the 7- and 8-h values as the steady-state output, the corresponding values in pmol/h (mean f S.E.) were group A (no supplement), 1.5 f 0.1; group C (choline supplement), 2.3 it 0.1; group B (dioleoyl phosphatidylcholine supplement), 3.0 + 0.2. The difference in phospholipid output was significant with P < 0.001 for both group A versus group C and group C versus group B.

T-

0

2

4 HOURS

6

8 HOURS

Fig. 1. Triglyceride and phospholipid output before and during lipid infusion. A lipid test meal was infused into the three groups of bile fist& rats. l, without supplement (A); A, dioleoyl phosphatidylcholine supplemented (B); and A, choline supplemented (C) (see Materials and Methods for test meal composition). Both the triglyceride (A) and phospholipid (B) outputs in lymph were measured. Both outputs were measured 2 h preinfusion and also during the 8 h infusion period. All values are expressed as mean i- SE.. for 8, 5 and 6 rats in groups A, B and C, respectively.

~~glyceride. Prior to lipid infusion, triglyceride outputs were low in all groups. The output increased after lipid infusion and reached a steady level at 6 h. The steady-state triglyceride output followed a trend similar to phospholipid output in the three groups (Fig. IA). The rats from group B transported significantly more triglyceride than the group C rats (P < 0.02) and the group C rats transported more than the group A rats (P< 0.001). The steady-state triglyceride outputs in pmol/h (mean + SE.) were group A, 15.7 rt 1.1; group C, 24.7 rf:1.4; group B, 30.7 -i-1.4. Lymph phospha~idylcholi~e and its fatty acid composition. Phosphatidylcholine made up above 64% of the fasting lymph phospholipid. During lipid infusion, the percentage of phospholipid as phosphatidylcholine gradually increased to a steady-state value of 74.4 + 2.5% (mean ? S.E.) in group A, 82.7 F 2.0% in group C and 81.3 + 2.6% in group B. The fatty acid composition of phosphatidylcholine in lymph is summarised in Table I. The fasting lymph phosphatidylcholine was rich in palmitate (16 : 0), stearate (18 : 0) and linoleate (18 : 2). The fatty acid pattern of lymph phosphatidylcholine changed markedly during fat absorption. During steady-state fat absorption (7 and 8 h), although choline supplementation, group C, significantly increased phospholipid output, the fatty acid pattern of lymph phosphatidylcholine was remarkably similar to that in group A. However, in group B, supplements with dioleoyl phosphatidylcholine, the palmitate percentage was lower, and oleate percentage was higher, than in the other groups.

369 TABLE

I

FATTY

ACID COMPOSITION

A lipid test meal (see Materials

OF STEADY-STATE and Methods

LYMPH

PHOSPHATIDYLCHOLINE

for composition)

was infused

intraduodenally

to the three

groups of bile fistula rats: without supplement (A). dioleoyl phosphatidylcholine supplemented (B), and choline supplemented (C). Lymph was collected from the thoracic duct cannula and lipid extracted. Phos phatidylchohne was isolated from the lipid extract of the 7 and 8 lymph by thin-layer chromatography and the methyl esters of the fatty acids were then prepared and anslysed by gasliquid chromatography. Only a trace of pahnitoleate (16 : 1) was detected and thus it was excluded from this table. Results are expressed in percent by mass. Values expressed as mean + S.E. N indicates number of animals used. n.s. = not significant. 16 (I) Preinfusion: Steady state: (II) Group A Group C Group B BvA BvC

:0

0

18:

:

18

1

18

:2

20

:4

7)

34.0

r 4.4

22.9 f 1.9

16.3 2 3.2

19.6 + 2.2

7.2 f 1.2

(N= 8) (N=6) (N = 5)

15.0 15.8 10.2 P < P <

? 0.6 + 0.8 + 1.0 0.01 0.01

10.2 ? 0.8 9.0 f 0.8 7.7 + 1.1 n.s. n.s.

50.5 53.2 62.8 P < P <

20.2 f 0.8 18.8 * 0.5 17.6 ? 1.6 us. n.s.

3.5 + 0.9 3.3 f 0.8 2.2 r 0.6 n. s.

(N=

r 1.8 + 1.6 f 2.6 0.01 0.02

KS.

Luminal and mucosal recovery of labelled fatty acid. By using labelled fatty acid, the amount of radioactive lipid was measured both in the gut lumen and mucosa at the end of 8 h (Table II). No marked difference was observed between the luminal recoveries from the different parts of the gut in these three groups. This implied that any difference in lymphatic output of absorbed lipid cannot be attributed to different rates of absorption. Mucosal recoveries from the upper intestine were similar in all groups. The apparently higher mucosal radioactive lipid in the lower intestinal mucosa of bile fistula + choline rats was not statistically siginificant (P < 0.05).

TABLE

II

LABELLED OLEIC ACID AGE DOSE INFUSED

RECOVERED

FROM

LUMEN

AND

MUCOSA

EXPRESSED

AS PERCENT-

A lipid test meal (see Materials and Methods for composition) was infused intraduodenally to the three groups of bile fistula rats. At the end of the 8 h infusion period, the animals were killed and luminal recoveries from different parts of the gut were collected after washing with 5 mM sodium taurocholate solution. An ahquot of the luminal washing was taken and mixed with a water-miscible scintiIlant for radioactivity determination. Mucosa from the upper and lower small intestine was scraped off and lipid extracted by a toluene/ethanol (2 : 1, v/v) solvent system. The radioactivity from the mucosal lipid was then measured (see Materials and Methods for extraction and counting procedures). Results are expressed as mean f S.E. N = number of animals used. A(N= LuminaI

MucosaI

Stomach Upper intestine Lower intestine Caecum Upper Lower

1.1 5.8 4.7 1.7 9.2 5.0

+ f f f: * +

5) 0.5 2.8 2.3 0.5 2.3 1.0

B(N= 0.5 7.1 2.6 2.2 11.7 4.7

+ * + f f f

5) 0.2 4.2 0.7 0.8 3.5 1.9

C (N =5) 1.0 7.3 2.9 3.5 8.3 9.0

* + f. ? r +

0.3 1.7 0.7 1.9 1.7 2.0

370

Discussion The lymphatic outputs of phospholipid and triglyceride in bile fistula rats in which bile salts, but not phosphatidylcholine, was replaced were almost identical, in the present series, with those in a previous series given the same infusate 143. For the present series, the steady-state outputs were 1.5 Errno and 15.7 ,umol/h for phospholipid and triglyceride, respectively; whereas, in the earlier series, they were 1.5 and 14 pmol/h, respectively. Substitution of for an equivalent amount of total 10 pmol/h dioleoyl phosphatidylcholine, fatty acid in the infusate, restored phospholipid and triglyceride outputs to values previously found in bile fistula rats given supplements of 10 iumollh biliary phosphatidylcholine or in rats with intact bile ducts. The gross steadystate outputs for these groups, studied previously, were 3 and 3.4 pmol/h, respectively, for phospholipid output and 33.3 and 37 pmol/h for triglyceride output. The net outputs, after subtraction of basal lipid output (which is higher with bile ducts intact) were 2.7 and 2.1 pmol/h for phospholipid output and 32.5 and 33.7 pmol/h for triglyceride output. With dioleoyl phosphatidylcholine supplementation, the net outputs were 2.6 and 30.3 pmol/h for phospholipid and triglyceride, respectively. Since dioleoyl phosphatidylcholine was a fully effective substitute for an intact biliary supply and was no less effective than purified phosphatidylcholine, it does not seem that high palmitic acid content of biliary phosphatidylcholine or its 1-acyl iysophosphatidylcholine is important in chylomicron production. The question of whether the subsequent metabolism of chylomicra is influenced by the composition of the phosphatidylcholine coating is not answered. Bihar-y phosphatidylcholine is derived from a different hepatic pool [24] and is different in composition from phosphatidylcholine exported to the bloodstream. The significance of the specific fatty acid composition of bile phosphatidylcho~ne remains obscure. An equivalent supplementation with 10 @mol/h choline chloride was only partly effective. This suggested that the phospholipid supplement was more important in supplying lysophosphatidylcholine to the intestinal absorptive cells than as a source of choline for phospholipid synthesis. The results on lipid output are not decisive on this point since the choline chloride molecules absorbed intact and rapidly transported to the bloodstream may be less available as a substrate for phospholipid biosynthesis [25] than choline liberated in situ by hydrolysis of absorbed lysophosphatidylcholine. However, in two rats, no further improvement in lipid output was obtained when the choline supplement was increased to 20 ymol/h. The results of this series of experiments and those previously reported [4] would be consistent with the hypothesis that de novo synthesis of phosphatidylcholine (involving CDPcholine [25]) does not fully support the high phospholipid turnover which accompanies rapid chylomicron formation and release. Under such circumstances, acylation of absorbed lysophosphatidylcholine makes an important contribution. In contrast, as previously reported [4] luminal phosphatidylcholine is not essential to sustain a low rate of chylomicron formation and release. Some support for this hypothesis is given by the data on the fatty acid composition of lymph phosphatidylcholine from the two

371

series of experiments. Table I shows that partial improvement in output of triglyceride and phospholipid with supplementary choline was not accompanied by a significant alteration in the fatty acid composition of phosphatidylcholine suggesting an increase in the novo synthesis. However, when output of lipid was restored to normal by a supplement of dioleoyl phosphatidylcholine, the palmitic acid content of lymph phosphatidylcholine decreased and the oleic acid content increased. This would be consistent with the utilization of l-oleoyl lysophosphatidylcholine for increased phospholipid synthesis. In our previous experiments in which the luminal phosphatidylcholine was rich in palmitate in the fu-position, due to the presence of endogenous bile, or to supplementation with biliary phosphatidylcholine in bile fistula rats, the lymph phosphatidylcholine was correspondingly rich in palmitate [4]. It was recently reported that the novo synthesis alone can provide enough phosphatidylcholine for chylomicron coating and subsequent transport from the intestinal mucosa [ 261. The differences in experimental conditions may be significant. If mesenteric fistula provided complete recovery of intestinal lymph, chylomicron production in the other study [26] was somewhat slow (the steady-state triglyceride output was about 24 E.tmol/h at 24 h after operation; whereas in our rats, it was about 3’7 pmol/h on a smaller absorptive load, 48 h after operation). It has been found that bile fistula rats absorb lipid more effectively at 48 h than 24 h after operation 2271. Also, in the study of Mansbath [26] an emulsified triglyceride infusate was used, whereas we used a micellar monoolein/oleic acid mixture. Acknowledgements The authors are grateful to Professor John Balint for his interest and suggestions in this study and to Mrs. Heather Kendrick for excellent research assistance. This work was supported by grant from the National Health and Medical Research Council and the University of Western Australia. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

O’Doherty, P.J.A., Kakis. G, and Kuksis, A. (1973) Lipids 8, 249-256 Tso. P. and Simmonds, W.J. (1976) Austr. J. Exp. Biol. Med. Sci. 55, 355-357 Rodgers, J.B. (1975) Bioehem. Biophys. Acta 398,92-100 Tso, P., Balint. J.A. and Simmonds, W.J. (1976) Gastroenterology 13, x362-1367 Scow, R.O., Stein, Y. and Stein, 0. (1967) J. Biol. Chem. 242, 4919-4924 AmesjS, B., Nilsson, A., Barrowman, J. and Borgstrijm, B. (1969) &and. J. Gastroenterol. 4, 653-665 Parthasarathy, S.. Subbaiah. P.V. and Ganguly, J. (1974) Biochem. J. 140, 503-508 Boucrot, P. and Clement, J.R. (1971) Lipids, 6. 652-656 Possmayer, G., Scherphof, G.L., Dubbelman, T.M.A.R., van Golde, L.M.G. and van Deenen, L.L.M. (1969) Bioehim. Biophys. Acta 176,95-110 BolIman, J.L., Cain, J.C. and Grindby, J.H. (1948) J. Lab. CIin. Med. 33, 1349-1352 Clark, S.B. (1971) J. Lipid Res. 12, 43-55 Cuber0 Robles, E. and van den Berg, D. (1969) Biochim. Biophys. Acta 187, 52@-526 Brockerhoff, H. and Yurkowski, M. (1965) Can. J. Biochem. 43. 1777 Selinger. 2. and Lapidot, Y. (1966) J. Lipid Res. 7. 174-175 Hanahan, D.J., Dittmer, J.C. and Warashina, E. (1957) J. Biol. Chem. 228, 685-700 Lack, L., Dorrity, Jr., F.O.. Walker. T. and Singletary. G.D. (19731 J. Lipid Res. 14, 367-370 BIigh, E.G. and Dyer, W.J. (1959) Can. J. Biochem. Physiol. 37.911-917 Stem, I. and Shapiro. B. (1953) J. CIin. Pathol. 6, 158-160

372 19 20 21 22 23

Parker, F. and Peterson, N.F. (1965) J. Lipid Res. 6, 455-460 Slomiany, B.L. and Horowitz, M.I. (1970) J. Chromatogr. 49, 455-461 Vaskovsky, V.E. and Kostetsky, E.Y. (1968) J. Lipid Res, 9, 396 Balint, J.A.. Nyban, W.L., Lietman. P. and Turner, D.A. (1961) J. Lab. Clin. Med. 58, 548-558 Snedecor, G.W. and Coehran, W.G. (1967) Statistical Methods, 6th edn., PP. 59-62, Iowa UniverSitY

24 25 26 27

Press, Iowa Balint, J.A., Beeler, D.A., Treble, D.H. and Spitzer, H.L. (1967) Kennedy, E.P. (1961) Fed. Proe. 20, 934-940 Mansbach, C.M. (1977) J. Clin. Invest. 60, 411-420 Mor@m, R.G.H. (1966) Q. J. Expt. Physiol. 51, 33-41

J. Lipid Res. 8. 486-493