- Email: [email protected]
Effect of intramolecular fatty acid distribution on aspects of triacylglycerol digestion and absorption studied in vitro
257
Biochimico et Biophysico Acta, 531 (1978) 257- 265 0 Elsevier/North-Holland Biomedical Press
BBA 57283
EFFECT OF INTRAMOLECULAR FATTY ACID DISTRIBUTION ON ASPECTS OF TRIACYLGLYCEROL DIGESTION AND ABSORPTION STUDIED IN VITRO
TAK YEE AW and MURRAY
R. GRIGOR
Department
University of Otago, P.O. Box 56 Dunedin (New Zealand)
(Received
of Biochemistry, May 5th, 1978)
Summary
The triacylglycerols from several natural sources have been used as substrates for in vitro assays of three different aspects of fat digestion and absorption. The triacylglycerol mixtures were chosen because they differed widely both in total fatty acid composition and in the intramolecular distribution of fatty acids. The assays were designed to test the rate of hydrolysis with pancreatic lipase, the ability to form insoluble calcium soaps during the hydrolysis, and the ability to form mixed micelles with bile during the hydrolysis. Differences in the rate of hydrolysis appeared to be related to the structure and the triacylglycerols from lard and human milk, both of which have palmitic acid esterified in the sn-2 position, were hydrolysed most rapidly. Similarly the ability to form calcium soaps was related to structure and those triacylglycerols which released saturated fatty acids on hydrolysis favoured calcium soap formation. While there was a range of abilities to form mixed micelles with bile no obvious correlation with structure was detected.
Introduction
The digestion of triacylglycerols and the absorption of the digestion products from the small intestine is a complex but efficient process. In normal adults over 95% of the dietary fat is absorbed. Fat absorption in infants is a little less efficient and appears to be dependent on the diet. Van de Kamer and Weijers [ 1 J showed that fat absorption in infants fed bovine milk was depressed by some 5% compared with infants fed human milk, but the differential had disappeared after the age of 1 year. The difference in fat absorption in infants fed bovine milk and those fed human milk has been attributed to differences in the composition of the milk fats. In particular a marked difference in the intramolecular distribution of the fatty acids within the triacylglycerol molecule has
258
been noted. Human milk triacylglycerols are unusual in that greater than 70% of the saturated palmitic acid is confined to the sn-2 position of the molecule [2-41, whereas in bovine milk a much more random distribution is observed [ 5-81. Here the saturated fatty acids are about equally distributed between the sn-1 and sn-2 positions. Both these structures are likely to differ also from any vegetable fat which may be included in milk substitutes. In these fats the sn-2 position is occupied almost exclusively by unsaturated fatty acids. That the intramolecular distribution of fatty acids can affect the absorbability of a fat has been shown by experiments in which infants were fed lard (a fat with structure similar to human milk) and chemically randomised lard [9]. Other experiments with rats showed that fats with palmitic acid in the sn-2 position of the triacylglycerol molecule are better absorbed than those where palmitic acids occurs at the sn-1 and sn-3 positions [lo]. Bovine milk triacylglycerols also differ from those of human milk in that they contain 12-1570 of fatty acids with chain lengths less than C-12, and lower amounts of unsaturated fatty acids (20-2570 compared with over 50% in human milk) [ 5-81. The short chain acids, particularly butyric and hexoic acids, tend to be confined to position 3 of the bovine milk triacylglycerols [8]. Their presence also allows the bovine milk triacylglycerols to be separated into two classes of high and low molecular weight. Since in fat digestion the fatty acids are released from the sn-1 and sn-3 positions of the triacylglycerols to produce a mixture of free fatty acids and sn-2-monoacylglycerols, the nature of the fatty acids released could affect the subsequent micelle formation and absorption process. A complicating feature is the formation of insoluble calcium soaps. These form more readily with saturated long chain acids rather than with short chain acids or unsaturated acids. This study describes in vitro experiments in which the effect of triacylglycerol structure on three phases related to lipid digestion and absorption have been investigated. These are: (1) the rate of hydrolysis by pancreatic lipase, (2) the ability to form calcium soaps during hydrolysis, and (3) the ability to form micelles with added bile during hydrolysis. Several different fats containing triacylglycerols in which the intramolecular distribution of fatty acids differs widely have been used. These include bovine milk, human milk and a vegetable oil, peanut oil. Materials and Methods Porcine pancreatic lipase and cholic acid were purchased from Sigma from Chemical Co., St. Louis, MO., U.S.A. Dried ox bile was purchased Hopkins and Williams, U.K. and gum arabic from May and Baker Ltd., Dagenham, U.K. Radioactive glycerol tri[ l-14C] oleate, glycerol tri[ l-14C]palmitate and 45CaC12 were purchased from the Radiochemical Centre, Amersham, U.K. Preparation of substrates. The total lipids from bovine and porcine adipose tissue samples and those from bovine and human milk were extracted with the solvent volumes of Bligh and Dyer [ll]. Commercial peanut oil was used. Each of the lipid samples was applied to a column of 20 g Florisil and the triacylglycerols were eluted with chloroform. The bovine milk triacylglycerols were separated into two fractions of high and low molecular weight by chromatog-
259
raphy on Florisil using a stepwise gradient of diethyl ether in hexane. Each of the triacylglycerol samples was analysed directly by gas-liquid chromatography using a column of 1% Dexsil 300 (Analabs, Inc., North Haven, Conn., U.S.A.) programmed from 280 to 350°C. The total fatty acids and those of the sn-2-monoacylglycerols and the fatty acids released from the sn-1 and sn-3 positions by pancreatic lipase digestion (see below) and recovered by preparative thin-layer chromatography were converted to methyl esters and analysed by gas-liquid chromatography using a column of 10% DEGS-F’S (Supelco, Inc., Bellefonte, Pa. U.S.A.). The fatty acid composition and intramolecular distribution of the fatty acids in samples used resembled those described for bovine milk [ 5-81, human milk [2-41, lard [ 12-151, bovine adipose [ 16 ] and peanut oil [ 171. Emulsions of each of the triacylglycerol samples containing 20 mglml were prepared by sonicating in 10% gum arabic (w/v) in 0.2 M Tris-HCI, pH 8.5, for 3 min. The samples were kept on ice during the sonication period. Pancreatic fipase digestion. The standard incubation for the digestion contained the emulsified substrate (10 mg), 0.27 M CaCl?, 0.1 M NaCl, 0.12 mM cholic acid and 0.75 mg pancreatic lipase in 1 ml 0.1 M Tris-HCl buffer, pH 8.5. Incubations were carried out with shaking at 37”C, and terminated by acidifying with HCI to release the fatty acids from the soaps and adding 2 ml chloroform/methanol (1 : 1, v/v). After centrifuging, the chloroform layer was removed and the fatty acids released on hydrolysis were titrated with methanolic NaOH [ 181. Calcium soap formation. For these studies incubation mixtures were set up as above with the addition of 20 nCi 4SCaC12. Incubations were carried out for 30 min after which the mixtures were immediately filtered through Whatman No. 54 paper into vials containing a drop of concentrated HCl to inactivate the enzyme. 0.3-ml aliquots were taken to determine the soluble 45CaC1,. Radioactivity was counted in a Packard liquid scintillation counter using 5 ml of a scintillant containing (0.6% w/v) 2-(4-tert-butylphenyl)-5-(4-biphenylyl)-1,3,4oxadiazole in toluene/Triton X-100 (2 : 1, v/v.). Micelle formation. To study micelle formation during the digestion of the triacylglycerol mixtures the standard incubation mixture was used except CaClz was removed and either glycerol tri[l-*4C]palmitate or glycerol tri[l14C]oleate (10 nCi) and ox bile (7.5 mg) were added. Digestions were carried out for 30 min and then diluted with an equal volume of bile solution (7.5 mg/ ml) and 1 ml of the mixture was immediately loaded onto a column (1 X 15 cm) of Sephadex G-200 to separate the emulsion from the micellar phases. The flow rate was adjusted to 14 ml/h and fractions were collected. The radioactivity in each fraction was determined after addition of the scintillant. Results The maximum rates of lipolysis using pancreatic lipase required the presence of 0.1 M NaCl, 0.27 M CaCI, and 0.12 mM cholic acid. In the absence of NaCl less than 25% of the maximal activity was observed. The requirements for CaCl, and cholic acid were less absolute and in their absence 60-80% of the maximal activity was observed.
260
For each tria~ylgly~erol sample two studies were carried out. The first involved a progress curve of the digestion using 10 mgfml of each substrate. The fats investigated fell into three groups depending on the nature of the progress curve. For the first group, which included lard and human milk fat, the reaction was essentially complete by 15 min. The fats of the second group which included the bovine adipose and milk fats were hydrolysed more slowly and the hydrolyses were not complete until 30-40 min. The third type, peanut oil, differed from the others in that there was an initial lag period after which hydrolysis was rapid and complete by 25 min. This lag was reproducible and was observed with another vegetable oil, corn oi1, aIthough it was not as pronounced as for peanut oil. The second experiment involved investigating the effect of altering the substrate concentration on the rate of hydrolysis. Normal saturation curves were obtained for each fat investigated although again the three groups of fat could be identified (Fig. 1). The first group had the highest V values, while peanut oil had the lowest. Double reciprocal plots of the data were drawn. Although differences in the apparent K, values were noted these have not been presented because of the difficulty in assessing the significance of the parameter for a lipolytic enzyme (see Discussion). The ability to form calcium soaps was investigated using digestions of 30 min, a time sufficient for the digestion to be close to compl&io& The removal
2.4
1.6
0.8
0
0
4 Substrate
12 concentration
16
18
fmg,/ml)
Fig. 1. Triacyiglyceroh emulsified *in gum arabic (20 mg/mf 10% (w/v) gum arabic solution) were digested in the presence of 0.1 M N&l, 0.27 M CaCiz. 0.12 mM cholic acid and 0.75 mgtml pancreatic fipase in 1 ml 0.1 Tris-HCI buffer (pH 8.5). The reactions were run for 10 or 15 min and stopped by extracting the lipids. The fatty acids released were determined by titration with methanolic KOH. Source of triacylgfycerols, Porcine adipose, 0; human milk, 0; bovine miik (high molecular weight fraction), 0; bovine milk (low molecular weight fraction), A: bovine adipose, X; and peanut oil, 0.
261 TABLE
I
FORMATION
OF
CALCIUM
SOAPS
DURING
PANCREATIC
LIPASE
DIGESTION
OF TRIACYL-
GLYCEROLS Values are mean f S.D. of 4-13
determinations.
Digestions
min using a triacylglycerol concentration of 10 me/ml ducts were filtered to obtain soluble calcium. Triac~~ycerol
Caa removed
source
were carried out as described
and in the presence
in Fig. 1 for 30
of 20 nCi 45CaC12.
The PrO-
as soap
Wmol)
Porcine
adipose
2.6 f 0.7 *
tissue
3.3 c 0.8
Human milk Bovine milk
3.6 f 0.8 6.4 + 1.2 1.8 * 0.03
unfractionated high molecular weight fraction low molecular weight fraction Bovine adipose
6.5 ?r 0.6
Peanut oil
3.9 f 0.6
1400-
1200 -
1000-
600 -
0
10
I
Fraction
number
Fig. 2. Products of digestion of trlacylglycerols from porcine adipose tissue in tbe presence of 10 nCi glycerol tri[ 1-l 4Clpalmitate and ox bile were diluted a-fold in a bile solution of the same concentration and eluted from a column (1 X 15 cm) of Sephadex G-200 with Trls-HCl buffer (0.1 M. pH 8.6) containing 0.1 M NaCl and 0.12 mM cholic acid. Ox bile concentration: zero. 0, and 10 mg/ml, 0. For the third curve. a, the bile concentration was 7.5 mg/ml, but the pancreatic lipase was omitted from the incubation. The emulsion phase precedes the micellar phase; V, = void volume.
262 TABLE
II
FORMATION OF MIXED TRIACYLGLYCEROLS
MICELLES
WITH BILE
DURING
PANCREATIC
LIPASE
DIGESTION
OF
Digestions were set up as described in Fig. 1 but in the absence of CaClz and containing ox bile (7.5 mg/ml) and labelled triacylglycerol (10 nCi). Reactions run for 30 min and the micellar phase separated from the emulsion phase by chromatography on Sephadex G-ZOO. Triacylglycerols
source
Micelle
formation
(%)
Glycerol trill-14C)palmitate _ ___ __~
Glycerol
42.7,
51.3,
57.5
44.9,
55.7
-
38.2. 86.4,
43.1 86.0
-
Bovine adipose
31.7,
35.5
36.0
Peanut oil
86.3,
89.8
85.9
-.______ Porcine
~ ~_~__ _-_ adipose
Human milk
_-._.
___ .__
48.1
tri[l-14C]oleate
Bovine milk high molecular low molecular
-___
weight fraction weight fraction
of calcium as the soap was proportional to the calcium concentration up to at least 0.27 M. Using this concentration marked differences in the extent of calcium soap formation were noted for the different fats (Table I). The triacylglycerols with the greatest tendency to form calcium soaps were those from bovine adipose and the high molecular fraction from bovine milk. The low molecular weight fraction from the bovine milk showed the least tendency for soap formation and, interestingly, the unfractionated bovine milk triacylglycerols were identical to those from human milk in this respect. Digestions of 30 min were also used to study the ability to form micelles. Gel chromatography has been used previously to separate the emulsion and micellar phases [19] and proved satisfactory for the present study (Fig. 2). Using lard as a test fat it was observed that the ability to form micelles was proportional to the bile concentration up to 10 mg/ml. All the triacylglycerols samples were tested using a bile concentration of 7.5 mg/ml. Similar results were observed when either glycerol tri[ l-14C]oleate or glycerol tri[ 1-14C]palmitate was used as the source of label to quantitate the emulsion and micellar phases. In each case thin-layer chromatography of the lipid in the emulsion phase showed that the digestion of the labeled triacylglycerols to fatty acids and sn-2-monoacylglycerols was essentially complete. The extent of mice& formation was also similar when digestions were carried out at pH 7.5 instead of the usual pH 8.5. The results for the different triacylglycerol samples are shown in Table II. For each sample the degree of micelle formation in our assay was reproducible. There were marked differences in the extent of micelle formation for the different fat samples used. In particular, peanut oil and the low molecular weight fraction of bovine milk favoured micelle formation, while the least extent of micelle formation occurred for the triacylglycerols from bovine adipose and the high molecular weight fraction from bovine milk. Discussion Our experiments were designed to investigate the effect of triacylglycerol structure on three aspects of lipid digestion and absorption independently so
263
that the resu!ts could be correlated with known differences in the absorbability of these fats. Significant differences were noted in the rates of hydrolysis by pancreatic lipase, the ability to form calcium soaps during the hydrolysis and the ability to form micelles during the hydrolysis. Lipolytic enzymes act at the interface of a two-phase system. In the case of pancreatic lipase the enzyme in the aqueous phase acts on the substrate in the oil phase. Several factors might affect the rate of hydrolysis. These include the ability of the enzyme to interact with the oil phase (the supersubstrate in Brockerhoff’s terminology [20]), the rate of removal of the products from the interface and the nature of the substrate used. The requirements for CaCl, and cholic acid for maximal activity are thought to be related to the removal of the products from the reaction interface [ 21,221. All the fats tested gave normal saturation curves when the substrate concentration was increased. From these curves obvious differences in the V values are apparent. Although double reciprocal plots of the same data were linear and did allow an apparent K, value for each substrate to be calculated, the significance of these values is hard to assess. For a lipolytic enzyme which requires a two-phase system the K, is a function not of the concentration of the substrate but rather of the surface area of the interface between the two phases. This in turn is a measure of the emulsification of the lipid. The conditions for forming the substrate emulsions in these experiments were uniform and no gross differences in particle size were obvious when the emulsions were studied by microscopy. For each triacylglycerol sample a range of particle sizes was noted. The differences in the V values, however, do appear to be a result of the differences in the compositions and structures of the lipids tested. For example, lard and human milk triacylglycerols have similar structures and both exhibit the high V values. For these fats the enzyme is removing unsaturated fatty acids to produce a saturated sn-2-monoacylglycerol. The curves for the bovine adipose triacylglycerols and the high molecular weight fraction of the bovine milk are also similar. These lipids have similar composition, and although the structures are not identical, a mixture of saturated and unsaturated fatty acids is released by the enzyme. It is possible that the V for the low molecular weight fraction from bovine milk is an underestimate resulting from incomplete extraction of the short chain fatty acids which would have been released by the enzyme. The low V value for the peanut oil triacylglycerols as welI as the unusual progress curve for their digestion, has no obvious explanation. This lipid is highly unsaturated (over 85% of the fatty acids are unsaturated) and both the fatty acids released and the sn-2-monoacylglycerols produced will be largely unsaturated. The ability to form calcium soaps appears to be related to the concentrations of long chain saturated fatty acids released during the hydrolysis by the lipase. Lard, human milk and peanut oil triacylglycerols all produce mainly unsaturated fatty acids and all have a similar capacity to produce calcium soaps. The bovine adipose triacylglycerols and the high molecular weight fraction from the, bovine milk both contain high proportions of saturated fatty acids in the sn-1 position and also have the greatest tendency to produce calcium soaps. The low degree of soap formation for the low molecular weight
264
fraction of bovine milk was unexpected. The ~~1-1 position of these triacylglycerols contains as high a proportion of saturated fatty acids as does the bovine adipose sample [5,16], but position 3 is occupied almost entirely by short chain fatty acids. These short chain acids are not likely to produce insoluble soaps. The tendency for the unfractionated bovine milk biacylglycerols to form calcium soaps appears to be the average of the two fractions and not different from that for human milk triacylglycerols. The ~corporation of the free fatty acids and the sn-2-monoacylglycerols into mixed micelles with cholesterol and bile acids is a necessary prerequisite for their absorption into the epithelial cells of the small intestine. Thus the results of the study of micelle formation are of interest. The only structural similarity of the two fats which show the greatest tendency to produce micelles, peanut oil and the low molecular weight fraction of bovine milk, is that the sn-2-monoacylglycerol produced will be unsaturated. However, the sn-2-monoacylglycerol produced from bovine adipose triacylglycerols will be equally unsaturated. This lipid, however, shows the least tendency to produce micelles. Calcium was omitted from this assay system and it was considered possible that hydrolysis of these triacylglycerols might be inhibited by accumulation of long chain saturated fatty acids on the reaction surface. However, removal of the calcium from an assay using this substrate reduced the lipase activity by only 14%. It is known that short chain fatty acids are much more readily incorporated into micelles than longer chain fatty acids [ 231. This could possibly explain the increased micelle formation observed with the low molecular weight fraction of bovine milk, The effect of different monoacylglyeerols on micelle formation does not appear to have been investigated. The extrapolation of our in vitro results to in vivo fat absorption requires caution. Although we have found differences in the rate of hydrolysis of the different fats under investigation, these differences may have limited physiological significance. The faecal lipids contain little esterified lipid and the fatty chains are present either as free acids or calcium soaps ]24,25]. Thus it is likely that the competition between soap formation and micelle formation will determine the absorbability of any one fatty acid. This in turn will determine the absorbability of the fat from which these fatty acids are released. Combining these considerations it would be predicted that the absorbability of bovine adipose lipid would be low. Furthermore, our results suggest that there should be little difference in the absorbability of unfractionated bovine milk, human milk and lard triacylglycerols. Penut oil with its greater tendency to form micelles may even be better absorbed. The difference in absorbability between bovine milk and human milk triacylglycerols in the newborn infant [l] may not be a function of the composition and structure of the fats but rather determined by some other factor. Such a factor may be the concentration of calcium in the milks. The concentration of calcium chosen for our experiments was high. Although it is close to that of bovine milk (260 mM) and several times that of human milk (70 mM) [26] most of the calcium in milk is not free but incorporated into casein micelles. Even so it is probable that the difference in absorbability of these two milk fats might well be determined by the difference in the calcium concentrations of the milks. However, the increase in the ability
265
of infants to absorb bovine milk with increasing age is important and may represent some modification in the ability to absorb fats. There is evidence that newborn animals have low lipase activities and may absorb t~acyl~ycerols directly 1271. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
20 21 22 23 24 25 26 27
Van de Kamer, J.H. and We&m, H.A. (1961) Fed. Proc. 20.335-344 Freeman, C.P., Jack, EL. and Smith, L.M. (1965) J. Dairy Sci. 48, 853-858 Breckenridge, W.C., Marai, L. and Kuksis. A. (1969) Can. J. Biochem. 47, 761-769 Grigor, M.R. (1977) Proc. Univ. Otago Med. Sch. 55. 23-24 Blank, M.L. and Privett, 08. (1964) J. Dairy Sci. 47, 481-488 Smith, L.M., Freeman C.P. and Jack, E.L. (1965) J. Dairy Sci. 48, 531-536 Pitas, R.E., Sampugna, J. and Jensen, R.G. (1967) J. Dairy Sci. 50, 1332-1336 Breckenridge, W.C. and Kuksis, A. (1968) J. Lipid Res. 9.388-393 Filer, Jr., L.J., Mattson, F.H. and Fomon, S.J. (1969) J. Nutr. 99, 293-298 TomareRi, R.M., Meyer, B.J.. Weaber, J.R.. and Bernhart. F.W. (1968) J. Nutr. 95.583-590 Bhgh. E.G. and Dyer, W.J. (1959) Can. J. Biochem. Physiol. 37,911-917 Blank. M.L. and Privett. O.S. (1966) Lipids 1, 27.-30 Anderson, R.E., Bottino, N.R. and Reiser. R. (1970) Lipids 5. 161-164 Christie, W.W. and Moore, J.H. (1970) Biochim. Biophys. Acta 210, 46-56 Grigor, M.R., Blank, M.L. and Snyder, F. (1971) Lipids 6.965-968 Brockerhoff, H., Hoyle, R.J. and W&mark. N. (1966) Biochim. Biophys. Acta 116. 67-72 Brockerhoff, H. and Yurkowski, M. (1966) I. Lipid Res. 7, 62-64 K&es, M. (1972) in Laboratory techniques in Biochemistry and Molecular Biology (Work, T.S. and Work. E., eds.), Vol. 3, P. 366, North-Holland Publishing Co., Amsterdam Feldman, E.B. and Borgstrom, B. (1966) Biochim. Biophys. Acta 126, 136-147 Brockerhoff, H. (1974) Bioorg. Chem. 3. 176-183 Benzonana, G. and Desnuelle, P. (1968) Biochim. Biophys. Acta 164,47-58 Brockerhoff, H. and Jensen, R.G. (1974) Lipolytic Enzymes, PP. 74-86, Academic Press, New York Hofmann, A.F. and Borgstram, B. (1962) Fed. Proe. 21,43--50 Holt, L.E., Courtney, A.M. and Fales, H.L. (1919) Am. J. Dis. Child. 17, 241-250 Ho& LE.. Courtney, A.M. and Fales, H.L. (1919) Am. J. Dis. Child. 17, 423-439 Gyorgy, p. 11971) Am. J. Clin. Nutr. 24,970-975 Rokos, J., Hann. P., Koldovsky, 0. and Prochiizka. P. (1963) Physiol. Bohemoslov. 12. 213-219
Biochimico et Biophysico Acta, 531 (1978) 257- 265 0 Elsevier/North-Holland Biomedical Press
BBA 57283
EFFECT OF INTRAMOLECULAR FATTY ACID DISTRIBUTION ON ASPECTS OF TRIACYLGLYCEROL DIGESTION AND ABSORPTION STUDIED IN VITRO
TAK YEE AW and MURRAY
R. GRIGOR
Department
University of Otago, P.O. Box 56 Dunedin (New Zealand)
(Received
of Biochemistry, May 5th, 1978)
Summary
The triacylglycerols from several natural sources have been used as substrates for in vitro assays of three different aspects of fat digestion and absorption. The triacylglycerol mixtures were chosen because they differed widely both in total fatty acid composition and in the intramolecular distribution of fatty acids. The assays were designed to test the rate of hydrolysis with pancreatic lipase, the ability to form insoluble calcium soaps during the hydrolysis, and the ability to form mixed micelles with bile during the hydrolysis. Differences in the rate of hydrolysis appeared to be related to the structure and the triacylglycerols from lard and human milk, both of which have palmitic acid esterified in the sn-2 position, were hydrolysed most rapidly. Similarly the ability to form calcium soaps was related to structure and those triacylglycerols which released saturated fatty acids on hydrolysis favoured calcium soap formation. While there was a range of abilities to form mixed micelles with bile no obvious correlation with structure was detected.
Introduction
The digestion of triacylglycerols and the absorption of the digestion products from the small intestine is a complex but efficient process. In normal adults over 95% of the dietary fat is absorbed. Fat absorption in infants is a little less efficient and appears to be dependent on the diet. Van de Kamer and Weijers [ 1 J showed that fat absorption in infants fed bovine milk was depressed by some 5% compared with infants fed human milk, but the differential had disappeared after the age of 1 year. The difference in fat absorption in infants fed bovine milk and those fed human milk has been attributed to differences in the composition of the milk fats. In particular a marked difference in the intramolecular distribution of the fatty acids within the triacylglycerol molecule has
258
been noted. Human milk triacylglycerols are unusual in that greater than 70% of the saturated palmitic acid is confined to the sn-2 position of the molecule [2-41, whereas in bovine milk a much more random distribution is observed [ 5-81. Here the saturated fatty acids are about equally distributed between the sn-1 and sn-2 positions. Both these structures are likely to differ also from any vegetable fat which may be included in milk substitutes. In these fats the sn-2 position is occupied almost exclusively by unsaturated fatty acids. That the intramolecular distribution of fatty acids can affect the absorbability of a fat has been shown by experiments in which infants were fed lard (a fat with structure similar to human milk) and chemically randomised lard [9]. Other experiments with rats showed that fats with palmitic acid in the sn-2 position of the triacylglycerol molecule are better absorbed than those where palmitic acids occurs at the sn-1 and sn-3 positions [lo]. Bovine milk triacylglycerols also differ from those of human milk in that they contain 12-1570 of fatty acids with chain lengths less than C-12, and lower amounts of unsaturated fatty acids (20-2570 compared with over 50% in human milk) [ 5-81. The short chain acids, particularly butyric and hexoic acids, tend to be confined to position 3 of the bovine milk triacylglycerols [8]. Their presence also allows the bovine milk triacylglycerols to be separated into two classes of high and low molecular weight. Since in fat digestion the fatty acids are released from the sn-1 and sn-3 positions of the triacylglycerols to produce a mixture of free fatty acids and sn-2-monoacylglycerols, the nature of the fatty acids released could affect the subsequent micelle formation and absorption process. A complicating feature is the formation of insoluble calcium soaps. These form more readily with saturated long chain acids rather than with short chain acids or unsaturated acids. This study describes in vitro experiments in which the effect of triacylglycerol structure on three phases related to lipid digestion and absorption have been investigated. These are: (1) the rate of hydrolysis by pancreatic lipase, (2) the ability to form calcium soaps during hydrolysis, and (3) the ability to form micelles with added bile during hydrolysis. Several different fats containing triacylglycerols in which the intramolecular distribution of fatty acids differs widely have been used. These include bovine milk, human milk and a vegetable oil, peanut oil. Materials and Methods Porcine pancreatic lipase and cholic acid were purchased from Sigma from Chemical Co., St. Louis, MO., U.S.A. Dried ox bile was purchased Hopkins and Williams, U.K. and gum arabic from May and Baker Ltd., Dagenham, U.K. Radioactive glycerol tri[ l-14C] oleate, glycerol tri[ l-14C]palmitate and 45CaC12 were purchased from the Radiochemical Centre, Amersham, U.K. Preparation of substrates. The total lipids from bovine and porcine adipose tissue samples and those from bovine and human milk were extracted with the solvent volumes of Bligh and Dyer [ll]. Commercial peanut oil was used. Each of the lipid samples was applied to a column of 20 g Florisil and the triacylglycerols were eluted with chloroform. The bovine milk triacylglycerols were separated into two fractions of high and low molecular weight by chromatog-
259
raphy on Florisil using a stepwise gradient of diethyl ether in hexane. Each of the triacylglycerol samples was analysed directly by gas-liquid chromatography using a column of 1% Dexsil 300 (Analabs, Inc., North Haven, Conn., U.S.A.) programmed from 280 to 350°C. The total fatty acids and those of the sn-2-monoacylglycerols and the fatty acids released from the sn-1 and sn-3 positions by pancreatic lipase digestion (see below) and recovered by preparative thin-layer chromatography were converted to methyl esters and analysed by gas-liquid chromatography using a column of 10% DEGS-F’S (Supelco, Inc., Bellefonte, Pa. U.S.A.). The fatty acid composition and intramolecular distribution of the fatty acids in samples used resembled those described for bovine milk [ 5-81, human milk [2-41, lard [ 12-151, bovine adipose [ 16 ] and peanut oil [ 171. Emulsions of each of the triacylglycerol samples containing 20 mglml were prepared by sonicating in 10% gum arabic (w/v) in 0.2 M Tris-HCI, pH 8.5, for 3 min. The samples were kept on ice during the sonication period. Pancreatic fipase digestion. The standard incubation for the digestion contained the emulsified substrate (10 mg), 0.27 M CaCl?, 0.1 M NaCl, 0.12 mM cholic acid and 0.75 mg pancreatic lipase in 1 ml 0.1 M Tris-HCl buffer, pH 8.5. Incubations were carried out with shaking at 37”C, and terminated by acidifying with HCI to release the fatty acids from the soaps and adding 2 ml chloroform/methanol (1 : 1, v/v). After centrifuging, the chloroform layer was removed and the fatty acids released on hydrolysis were titrated with methanolic NaOH [ 181. Calcium soap formation. For these studies incubation mixtures were set up as above with the addition of 20 nCi 4SCaC12. Incubations were carried out for 30 min after which the mixtures were immediately filtered through Whatman No. 54 paper into vials containing a drop of concentrated HCl to inactivate the enzyme. 0.3-ml aliquots were taken to determine the soluble 45CaC1,. Radioactivity was counted in a Packard liquid scintillation counter using 5 ml of a scintillant containing (0.6% w/v) 2-(4-tert-butylphenyl)-5-(4-biphenylyl)-1,3,4oxadiazole in toluene/Triton X-100 (2 : 1, v/v.). Micelle formation. To study micelle formation during the digestion of the triacylglycerol mixtures the standard incubation mixture was used except CaClz was removed and either glycerol tri[l-*4C]palmitate or glycerol tri[l14C]oleate (10 nCi) and ox bile (7.5 mg) were added. Digestions were carried out for 30 min and then diluted with an equal volume of bile solution (7.5 mg/ ml) and 1 ml of the mixture was immediately loaded onto a column (1 X 15 cm) of Sephadex G-200 to separate the emulsion from the micellar phases. The flow rate was adjusted to 14 ml/h and fractions were collected. The radioactivity in each fraction was determined after addition of the scintillant. Results The maximum rates of lipolysis using pancreatic lipase required the presence of 0.1 M NaCl, 0.27 M CaCI, and 0.12 mM cholic acid. In the absence of NaCl less than 25% of the maximal activity was observed. The requirements for CaCl, and cholic acid were less absolute and in their absence 60-80% of the maximal activity was observed.
260
For each tria~ylgly~erol sample two studies were carried out. The first involved a progress curve of the digestion using 10 mgfml of each substrate. The fats investigated fell into three groups depending on the nature of the progress curve. For the first group, which included lard and human milk fat, the reaction was essentially complete by 15 min. The fats of the second group which included the bovine adipose and milk fats were hydrolysed more slowly and the hydrolyses were not complete until 30-40 min. The third type, peanut oil, differed from the others in that there was an initial lag period after which hydrolysis was rapid and complete by 25 min. This lag was reproducible and was observed with another vegetable oil, corn oi1, aIthough it was not as pronounced as for peanut oil. The second experiment involved investigating the effect of altering the substrate concentration on the rate of hydrolysis. Normal saturation curves were obtained for each fat investigated although again the three groups of fat could be identified (Fig. 1). The first group had the highest V values, while peanut oil had the lowest. Double reciprocal plots of the data were drawn. Although differences in the apparent K, values were noted these have not been presented because of the difficulty in assessing the significance of the parameter for a lipolytic enzyme (see Discussion). The ability to form calcium soaps was investigated using digestions of 30 min, a time sufficient for the digestion to be close to compl&io& The removal
2.4
1.6
0.8
0
0
4 Substrate
12 concentration
16
18
fmg,/ml)
Fig. 1. Triacyiglyceroh emulsified *in gum arabic (20 mg/mf 10% (w/v) gum arabic solution) were digested in the presence of 0.1 M N&l, 0.27 M CaCiz. 0.12 mM cholic acid and 0.75 mgtml pancreatic fipase in 1 ml 0.1 Tris-HCI buffer (pH 8.5). The reactions were run for 10 or 15 min and stopped by extracting the lipids. The fatty acids released were determined by titration with methanolic KOH. Source of triacylgfycerols, Porcine adipose, 0; human milk, 0; bovine miik (high molecular weight fraction), 0; bovine milk (low molecular weight fraction), A: bovine adipose, X; and peanut oil, 0.
261 TABLE
I
FORMATION
OF
CALCIUM
SOAPS
DURING
PANCREATIC
LIPASE
DIGESTION
OF TRIACYL-
GLYCEROLS Values are mean f S.D. of 4-13
determinations.
Digestions
min using a triacylglycerol concentration of 10 me/ml ducts were filtered to obtain soluble calcium. Triac~~ycerol
Caa removed
source
were carried out as described
and in the presence
in Fig. 1 for 30
of 20 nCi 45CaC12.
The PrO-
as soap
Wmol)
Porcine
adipose
2.6 f 0.7 *
tissue
3.3 c 0.8
Human milk Bovine milk
3.6 f 0.8 6.4 + 1.2 1.8 * 0.03
unfractionated high molecular weight fraction low molecular weight fraction Bovine adipose
6.5 ?r 0.6
Peanut oil
3.9 f 0.6
1400-
1200 -
1000-
600 -
0
10
I
Fraction
number
Fig. 2. Products of digestion of trlacylglycerols from porcine adipose tissue in tbe presence of 10 nCi glycerol tri[ 1-l 4Clpalmitate and ox bile were diluted a-fold in a bile solution of the same concentration and eluted from a column (1 X 15 cm) of Sephadex G-200 with Trls-HCl buffer (0.1 M. pH 8.6) containing 0.1 M NaCl and 0.12 mM cholic acid. Ox bile concentration: zero. 0, and 10 mg/ml, 0. For the third curve. a, the bile concentration was 7.5 mg/ml, but the pancreatic lipase was omitted from the incubation. The emulsion phase precedes the micellar phase; V, = void volume.
262 TABLE
II
FORMATION OF MIXED TRIACYLGLYCEROLS
MICELLES
WITH BILE
DURING
PANCREATIC
LIPASE
DIGESTION
OF
Digestions were set up as described in Fig. 1 but in the absence of CaClz and containing ox bile (7.5 mg/ml) and labelled triacylglycerol (10 nCi). Reactions run for 30 min and the micellar phase separated from the emulsion phase by chromatography on Sephadex G-ZOO. Triacylglycerols
source
Micelle
formation
(%)
Glycerol trill-14C)palmitate _ ___ __~
Glycerol
42.7,
51.3,
57.5
44.9,
55.7
-
38.2. 86.4,
43.1 86.0
-
Bovine adipose
31.7,
35.5
36.0
Peanut oil
86.3,
89.8
85.9
-.______ Porcine
~ ~_~__ _-_ adipose
Human milk
_-._.
___ .__
48.1
tri[l-14C]oleate
Bovine milk high molecular low molecular
-___
weight fraction weight fraction
of calcium as the soap was proportional to the calcium concentration up to at least 0.27 M. Using this concentration marked differences in the extent of calcium soap formation were noted for the different fats (Table I). The triacylglycerols with the greatest tendency to form calcium soaps were those from bovine adipose and the high molecular fraction from bovine milk. The low molecular weight fraction from the bovine milk showed the least tendency for soap formation and, interestingly, the unfractionated bovine milk triacylglycerols were identical to those from human milk in this respect. Digestions of 30 min were also used to study the ability to form micelles. Gel chromatography has been used previously to separate the emulsion and micellar phases [19] and proved satisfactory for the present study (Fig. 2). Using lard as a test fat it was observed that the ability to form micelles was proportional to the bile concentration up to 10 mg/ml. All the triacylglycerols samples were tested using a bile concentration of 7.5 mg/ml. Similar results were observed when either glycerol tri[ l-14C]oleate or glycerol tri[ 1-14C]palmitate was used as the source of label to quantitate the emulsion and micellar phases. In each case thin-layer chromatography of the lipid in the emulsion phase showed that the digestion of the labeled triacylglycerols to fatty acids and sn-2-monoacylglycerols was essentially complete. The extent of mice& formation was also similar when digestions were carried out at pH 7.5 instead of the usual pH 8.5. The results for the different triacylglycerol samples are shown in Table II. For each sample the degree of micelle formation in our assay was reproducible. There were marked differences in the extent of micelle formation for the different fat samples used. In particular, peanut oil and the low molecular weight fraction of bovine milk favoured micelle formation, while the least extent of micelle formation occurred for the triacylglycerols from bovine adipose and the high molecular weight fraction from bovine milk. Discussion Our experiments were designed to investigate the effect of triacylglycerol structure on three aspects of lipid digestion and absorption independently so
263
that the resu!ts could be correlated with known differences in the absorbability of these fats. Significant differences were noted in the rates of hydrolysis by pancreatic lipase, the ability to form calcium soaps during the hydrolysis and the ability to form micelles during the hydrolysis. Lipolytic enzymes act at the interface of a two-phase system. In the case of pancreatic lipase the enzyme in the aqueous phase acts on the substrate in the oil phase. Several factors might affect the rate of hydrolysis. These include the ability of the enzyme to interact with the oil phase (the supersubstrate in Brockerhoff’s terminology [20]), the rate of removal of the products from the interface and the nature of the substrate used. The requirements for CaCl, and cholic acid for maximal activity are thought to be related to the removal of the products from the reaction interface [ 21,221. All the fats tested gave normal saturation curves when the substrate concentration was increased. From these curves obvious differences in the V values are apparent. Although double reciprocal plots of the same data were linear and did allow an apparent K, value for each substrate to be calculated, the significance of these values is hard to assess. For a lipolytic enzyme which requires a two-phase system the K, is a function not of the concentration of the substrate but rather of the surface area of the interface between the two phases. This in turn is a measure of the emulsification of the lipid. The conditions for forming the substrate emulsions in these experiments were uniform and no gross differences in particle size were obvious when the emulsions were studied by microscopy. For each triacylglycerol sample a range of particle sizes was noted. The differences in the V values, however, do appear to be a result of the differences in the compositions and structures of the lipids tested. For example, lard and human milk triacylglycerols have similar structures and both exhibit the high V values. For these fats the enzyme is removing unsaturated fatty acids to produce a saturated sn-2-monoacylglycerol. The curves for the bovine adipose triacylglycerols and the high molecular weight fraction of the bovine milk are also similar. These lipids have similar composition, and although the structures are not identical, a mixture of saturated and unsaturated fatty acids is released by the enzyme. It is possible that the V for the low molecular weight fraction from bovine milk is an underestimate resulting from incomplete extraction of the short chain fatty acids which would have been released by the enzyme. The low V value for the peanut oil triacylglycerols as welI as the unusual progress curve for their digestion, has no obvious explanation. This lipid is highly unsaturated (over 85% of the fatty acids are unsaturated) and both the fatty acids released and the sn-2-monoacylglycerols produced will be largely unsaturated. The ability to form calcium soaps appears to be related to the concentrations of long chain saturated fatty acids released during the hydrolysis by the lipase. Lard, human milk and peanut oil triacylglycerols all produce mainly unsaturated fatty acids and all have a similar capacity to produce calcium soaps. The bovine adipose triacylglycerols and the high molecular weight fraction from the, bovine milk both contain high proportions of saturated fatty acids in the sn-1 position and also have the greatest tendency to produce calcium soaps. The low degree of soap formation for the low molecular weight
264
fraction of bovine milk was unexpected. The ~~1-1 position of these triacylglycerols contains as high a proportion of saturated fatty acids as does the bovine adipose sample [5,16], but position 3 is occupied almost entirely by short chain fatty acids. These short chain acids are not likely to produce insoluble soaps. The tendency for the unfractionated bovine milk biacylglycerols to form calcium soaps appears to be the average of the two fractions and not different from that for human milk triacylglycerols. The ~corporation of the free fatty acids and the sn-2-monoacylglycerols into mixed micelles with cholesterol and bile acids is a necessary prerequisite for their absorption into the epithelial cells of the small intestine. Thus the results of the study of micelle formation are of interest. The only structural similarity of the two fats which show the greatest tendency to produce micelles, peanut oil and the low molecular weight fraction of bovine milk, is that the sn-2-monoacylglycerol produced will be unsaturated. However, the sn-2-monoacylglycerol produced from bovine adipose triacylglycerols will be equally unsaturated. This lipid, however, shows the least tendency to produce micelles. Calcium was omitted from this assay system and it was considered possible that hydrolysis of these triacylglycerols might be inhibited by accumulation of long chain saturated fatty acids on the reaction surface. However, removal of the calcium from an assay using this substrate reduced the lipase activity by only 14%. It is known that short chain fatty acids are much more readily incorporated into micelles than longer chain fatty acids [ 231. This could possibly explain the increased micelle formation observed with the low molecular weight fraction of bovine milk, The effect of different monoacylglyeerols on micelle formation does not appear to have been investigated. The extrapolation of our in vitro results to in vivo fat absorption requires caution. Although we have found differences in the rate of hydrolysis of the different fats under investigation, these differences may have limited physiological significance. The faecal lipids contain little esterified lipid and the fatty chains are present either as free acids or calcium soaps ]24,25]. Thus it is likely that the competition between soap formation and micelle formation will determine the absorbability of any one fatty acid. This in turn will determine the absorbability of the fat from which these fatty acids are released. Combining these considerations it would be predicted that the absorbability of bovine adipose lipid would be low. Furthermore, our results suggest that there should be little difference in the absorbability of unfractionated bovine milk, human milk and lard triacylglycerols. Penut oil with its greater tendency to form micelles may even be better absorbed. The difference in absorbability between bovine milk and human milk triacylglycerols in the newborn infant [l] may not be a function of the composition and structure of the fats but rather determined by some other factor. Such a factor may be the concentration of calcium in the milks. The concentration of calcium chosen for our experiments was high. Although it is close to that of bovine milk (260 mM) and several times that of human milk (70 mM) [26] most of the calcium in milk is not free but incorporated into casein micelles. Even so it is probable that the difference in absorbability of these two milk fats might well be determined by the difference in the calcium concentrations of the milks. However, the increase in the ability
265
of infants to absorb bovine milk with increasing age is important and may represent some modification in the ability to absorb fats. There is evidence that newborn animals have low lipase activities and may absorb t~acyl~ycerols directly 1271. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
20 21 22 23 24 25 26 27
Van de Kamer, J.H. and We&m, H.A. (1961) Fed. Proc. 20.335-344 Freeman, C.P., Jack, EL. and Smith, L.M. (1965) J. Dairy Sci. 48, 853-858 Breckenridge, W.C., Marai, L. and Kuksis. A. (1969) Can. J. Biochem. 47, 761-769 Grigor, M.R. (1977) Proc. Univ. Otago Med. Sch. 55. 23-24 Blank, M.L. and Privett, 08. (1964) J. Dairy Sci. 47, 481-488 Smith, L.M., Freeman C.P. and Jack, E.L. (1965) J. Dairy Sci. 48, 531-536 Pitas, R.E., Sampugna, J. and Jensen, R.G. (1967) J. Dairy Sci. 50, 1332-1336 Breckenridge, W.C. and Kuksis, A. (1968) J. Lipid Res. 9.388-393 Filer, Jr., L.J., Mattson, F.H. and Fomon, S.J. (1969) J. Nutr. 99, 293-298 TomareRi, R.M., Meyer, B.J.. Weaber, J.R.. and Bernhart. F.W. (1968) J. Nutr. 95.583-590 Bhgh. E.G. and Dyer, W.J. (1959) Can. J. Biochem. Physiol. 37,911-917 Blank. M.L. and Privett. O.S. (1966) Lipids 1, 27.-30 Anderson, R.E., Bottino, N.R. and Reiser. R. (1970) Lipids 5. 161-164 Christie, W.W. and Moore, J.H. (1970) Biochim. Biophys. Acta 210, 46-56 Grigor, M.R., Blank, M.L. and Snyder, F. (1971) Lipids 6.965-968 Brockerhoff, H., Hoyle, R.J. and W&mark. N. (1966) Biochim. Biophys. Acta 116. 67-72 Brockerhoff, H. and Yurkowski, M. (1966) I. Lipid Res. 7, 62-64 K&es, M. (1972) in Laboratory techniques in Biochemistry and Molecular Biology (Work, T.S. and Work. E., eds.), Vol. 3, P. 366, North-Holland Publishing Co., Amsterdam Feldman, E.B. and Borgstrom, B. (1966) Biochim. Biophys. Acta 126, 136-147 Brockerhoff, H. (1974) Bioorg. Chem. 3. 176-183 Benzonana, G. and Desnuelle, P. (1968) Biochim. Biophys. Acta 164,47-58 Brockerhoff, H. and Jensen, R.G. (1974) Lipolytic Enzymes, PP. 74-86, Academic Press, New York Hofmann, A.F. and Borgstram, B. (1962) Fed. Proe. 21,43--50 Holt, L.E., Courtney, A.M. and Fales, H.L. (1919) Am. J. Dis. Child. 17, 241-250 Ho& LE.. Courtney, A.M. and Fales, H.L. (1919) Am. J. Dis. Child. 17, 423-439 Gyorgy, p. 11971) Am. J. Clin. Nutr. 24,970-975 Rokos, J., Hann. P., Koldovsky, 0. and Prochiizka. P. (1963) Physiol. Bohemoslov. 12. 213-219