seits wird auch uber eine unveranderte Aktivitat 37 oder iiber einen Anstieg 38, 39 der Succinatdehydrogenase be- richtet. Diese z. T. uneinheitlichen Aussagen der Lite- ratur, verglichen mit unseren Beobachtungen, konnen eventuell auf unterschiedliche Versuchsbedingungen zu- riickgefuhrt werden. So ist vor allem bekannt, dai3 die Tonicitat des Untersuchungsmediums und die Vorbe- handlung der Mitochondrien, z. B. mit Ultraschall, Ver- iinderungen in der Permeabilitat der Membran hervor- rufen und somit die Enzymaktivitaten beeinflussen.

Auch bei der Messung der Cytochrom-c-oxidase nach der Methode von J . Schole 40, modifiziert nach V. Dzapo 41?

konnten wir einen Aktivitatsabfall feststellen (Abb. 7) . Es bestand auch hier ein deutlicher Unterschied zwischen Gruppe I1 und den Gruppen mit Linolsaurezulagen zum

37 H . 0. Kunkel u. /. N . Williams Jr., The effects of fat deficiency upon enzyme activity in the rat, J. biol. Che- mistry 189, 755 [1951]

38 E. Levin, R. M. lohnson u. S. AZbert, Mitochondria1 changes associated with essential fatty acid deficiency in rats, J. biol. Chemistry 228, 15 [1957].

39 T. Hayashida u. 0. W . Portman, Changes in succinic de- hydrogenase activity and fatty acid composition of rat liver mitochondria in essential fatty acid deficiency, J. Nutrit. 81, 103 [1963].

40 1. Schole, Beschleunigung der Cytochrom-c-Oxidation durch Thyroxin, Hoppe-Seyler's Z. Physiol. Chem. 317, 281 [ 19591.

41 V . Dzafio, Institut fur Tierzucht und Haustiergenetik der Justus-Liebig-Universitat GieBen, personliche Mitteilung.

Futter. Die in der Literatur z.T. abweichenden Ergeb- nisse 3 6 1 42 sind ebenfalls mit den unterschiedlichen Ver- suchsbedingungen zu begriinden.

S c h 1 u 8 b e t r a c h t u n g

Wir konnten zeigen,' dai3 eine Abhangigkeit der Struk- tur von Membranen von der oralen Zufuhr an Linol- saure besteht, und zwar sowohl bei Plasmamembranen, d. h. den Erythrozyten, als auch bei intrazellularen Mem- branen wie den Mitochondrien. Dies machen die Mes- sungen der osmotischen Resistenz und des Cholesterin/ Phospholipid-Quotienten der Erythrozyten und die Ver- anderungen der Phospholipidfraktionen in den Mito- chondrien deutlich.

Wahrend in den Erythrozyten die Aktivitat der Acetylcholinesterase nicht und die Aktivitat der (Na+ + K+)-ATPase erst relativ spat durch Ausschlug der Linolsaure vom Futter beeinfluat wurden, konnten wir in den Untersuchungen an den Mitochondrien beobach- ten, daB die Aktivitaten der Atmungskettenenzyme sen- sitiver auf die Linolsaurezufuhr reagierten als die mole- kulare Zusammensetzung der Membran. Eingegangen am 17. September 1981.

42 P. G. Tulfiule u. /. N . Will iams, Study of the role of essential fatty acids in liver metabolism, J. biol. Chemistry 217, 229 [1955].

The Triglyceride Composition of Mango (Mangifera indica) Kernel Fat

B y W. v a n P e e , M . F o m a and L. Bani'>

The triglyceride composition of the kernel fat of 9 different mango varieties has been determined. Stearic and oleic acids represent respectively from 32.7 to 44.0 O/O and from 43.7 to 53.40/0 of the total fatty acids. The remaining fatty acids were palmitic (6.7-9.7 @/o), linoleic (3.66.9 o / o ) , arachidic (1.1-2.5 O/o)

and linolenic (0.3-1.0 O / o ) acids. The triglyceride components were determined by separating the triglycerides according their degree of unsaturation by means of thin-layer chromato- graphy on silica gel impregnated with silver nitrate. The fatty acid composition of the different triglyceride fractions and of the fatty acids incorporated at the sn-2-position of each triglyceride fraction was determined. Moreover, the triglyceri- des were separated according to their carbon number by gas liquid chromatography using an open-tubular glass column, wall-coated with CP-Sil 5. The triglyceride compositions obtained by thin-layer chromatography on silica gel impreg- nated with silver nitrate were in agreement with the compo- sitions predicted by the 1,3-random-Z-random distribution hypothesis.

1. I n t r o d u c t i o n

Mango, the fruit of Mangifera indica, is the most cultivated fruit in tropical and subtropical regions. About 9.6 million tons are produced each year'. The dried kernels of the seeds, which are left as waste after the consumption of the fruit, may contain up to 12% of fat.

'F Authors' address: Prof. Dr. W. van Pee, M . Foma and L. Boni, Tropical Food Production, Catholic University of Leuven, Kardinaal Mercierlaan 92, B-3030 Heverlee.

1 FA0 Production Yearbook, Series No. 22, Vol. 32, p. 177, Ed. Food and Agriculture Organization of the United Nations, FA0 Statistics, Rome 1979.

83. Jahrgang Nr. 10 1981 F E T T E . S E I F E N . A N S T R I C H M I T T E L

Die Triglycerid-Zusammensetzung von Mangokernfett (Mangifera indica]

Es wurde die Zusammensetzung der Triglyceride der Kern- fette neun verschiedener Mangoarten analysiert. 32.7 bis 44.0 O / o bzw. 43.7 bis 53.4 o/o aller Fettsauren sind Stearinsaure bzw. Ulsaure. Der Rest bestand aus Palmitinsaure (6.7-9.7 "/a], Linol- saure (3.6-6.9 O/o), Arachinsaure (1.1-2.5 O/O] und Linolensaure (0.3-1.0 O/O]. Durch Trennung der Triglyceride nach ihrem Grad der Ungesattigtheit mittels Diinnschichtchromatographie auf mit Silbernitrat impragnierten Kieselgelplatten wurden die Trigly- cerid-Bestandteile bestimmt. Es wurde die Fettsaure-Zusam- mensetzung verschiedener Triglycerid-Fraktionen und der in sn- Z-Stellung jeder Triglycerid-Fraktion gebundenen Fettsauren er- mittelt. Dariiber hinaus wurden die Triglyceride nach ihrer Kohlenstoffzahl mittels Gaschromatographie an einer offenen Glasrohrsaule belegt mit CP-Sil 5 getrennt. Die durch Diinn- schichtchromatographie an mit Silbernitrat impragniertem Kieselgel erhaltene Triglycerid-Zusammensetzung stimmte mil der iiberein, die sich nach der 1,3-Random-2-Random-Verteilung ergibt.

The fat extracted from the kernels has a yellow color and melts at 35-38OC. Stearic and oleic acids are the main fatty acids 3.

With regard to the fatty acid composition, mango kernel fat belongs to a small group of tropical seed fats characterized by a high content of stearic and oleic acids. This group of fats received considerable attention for their particular physical properties, which make them

2 M . Y . Ikramul Haq and A . F. M . Ehteshamuddin, Sci. Ind. 8, 207 [1971].

Rao, J. Amer. Oil Chemists' SOC. 54, 494 [1977]. 3 B. L. Narasimha Char, B. R. Reddy and S. D. Thirumala

383

very suitable for the fabrication of confectionary. The triglycerides of these fats consist mainly of both sn-l,3- disaturated-2-oleoylglycerols and sn-1(3)-saturated-2.3- (1,2)-dioleoylglycerols. The most known fat of this group is cocoa butter 4.

Studies on the recuperation of mango seeds have shown that the kernels represent a potential source of oil. It has been estimated that for India alone 30000 tons of this fat can be obtained each yearS. Till now the triglyceride composition of mango kernel fat has not been determined. Therefore the kernel fat of several mango varieties has been analysed and their triglyceride composition has been compared.

2. E x p e r i m e n t a l

2.1. Mango seeds The seeds from 9 different mango varieties were

obtained from the Agricultural Research Center of M’Vuazi (Republic of Zaire). In the laboratory the kernels were drawn out of the stones and lyophilized. The variety M’Vuazi I11 was derived by selection from the variety Haden. The other varieties were obtained from local African centers and from Asiatic selection centers.

2.2. Isolation of the triglycerides The dried kernels were extracted with chloroform :

methanol @:I, v/v) as described by /. Folch, M. Lees and G. H . Slonne Stanley5. The triglycerides were isolated by preparative thin-layer chromatography (TLC) on silica gel G using the solvent mixture pe- troleum ether (40-60° C b. p.) : diethyl ether : acetic acid (80:20:1, v/v/v). The triglyceride band was recovered and the triglycerides were eluted from the silica gel with chloroform.

2.3. Silver nitrate thin-layer chromatography The thin-layer plates (20 X 20 cm) were prepared with

a slurry containing 10 g of AgN03 per 40 g of silica gel G. Twenty mg of triglycerides were applied on the plate and the plate was developed with chloroform for 60 min. The plate was dried and developed with the solvent mixture chloroform : methanol (99:1, v/v) for 30 min. Individual bands were visualized under ultra-violet light after spraying with 2’,7’-dichlorofluorescein. The bands were recovered and extracted first with chloro- form : methanol (90:10, v/v) followed by hexane 6. A known amount of triheptadecanoin, as internal standard, was added after the extraction of the silica gel. The recovered triglycerides were further analysed for their fatty acid composition and for the fatty acids in- corporated at the sn-2-position.

2.4. Lipolysis of the triglycerides About 10 mg of fat were incubated with 20 pl of

hexane and 200 p1 of Tris buffer (pH 8.0, 1 M) contain- ing 1 mg of pancreatic lipase (Steapsin, Sigma Chemical

T. P. Hilditch and P. N . Will iams, The Chemical Consti- tution of Natural Fats, 4th edn., Chapman and Hall, Lon- don 1964. J. Folch, M. Lees and G. H. Sloane Stanley, J. biol. Che- mistry 226, 497 [1957]. J . Bezard and M . Bugaut, J. Chromatogr. Sci. 10, 451 [1972].

Co., Saint Louis, Mo., USA) and 200 pg of CaCl?‘. After visible flocculation the reaction was stopped with 1 N H,S04 and the incubation mixture was extracted with diethyl ether. The hydrolysis products were se- parated by TLC on silica gel G using petroleum ether (40-60°C b. p.) : diethyl ether : acetic acid (70:30:1, v/v/v). The sn-2-monoglycerides which represent the fatty acids incorporated at the sn-2-position were re- covered and converted to methylesters.

2.5. Gas liquid chromatograbhy (GLC) of methyl esters Methylesters were prepared using the reagent mixture

methanol : benzene : sulfuric acid (2:1:0.06, v/v/v) at 7OOC. After 2 hr the tubes were cooled and water was added. The methyl esters were extracted with petroleum ether (40-60° C b. p.) and analysed by GLC. A Hewlett- Packard 5830A gas chromatograph equipped with a dual flame ionization detector and stainless steel columns (length: 6 ft, inner diameter: 3 mm) packed with 10Vo SP-2330 on 100/200 mesh Chromosorb W A W (Supelco Inc., Crans, Switzerland) was used for the determination of the fatty acid methyl esters. The following conditions were used: nitrogen: 15 ml/min; injection port and detector temperature: 270° C; initial oven temperature: 190° C; linear temperature programming: 8 O C/min up to 24OOC starting 1 min after the injection. Peak areas and retention times were determined by a Hewlett- Packard 18850A GC terminal.

2.6. Calculation of the individual triglyceride components The hypothetical triglyceride components of the oils

were calculated by the 1,3-random-2-randorn distribution hypothesis and the l,2,3-random distribution hypo- thesis. The former hypothesis assumes two different fatty acid pools of which the first pool represents the fatty acids esterified at the sn-2-hydroxyl group of the glycerol and of which the second pool represents the fatty acids esterified at the sn-1- and sn-3-hydroxyl group of the glycerol. At each position the fatty acids are in- corporated at random. The amount of each triglyceride component can be calculated from the equation:

mole o/o mole O/o mole O/o

O/o ABC = A at 1,3- x B at 2- X C at 1,3- X position position position

In the same way the hypothetical triglyceride com- ponents were calculated by the 1,2,3-random hypothesis. This hypothesis states that one pool of fatty acids is distributed at random between the three hydroxyl groups of the glycerol backbone of the triglycerides. The fatty acids were represented as: S: the sum of palmitic, stearic and arachidic acids; 0: oleic acids; L: linoleic acids. The triglyceride components are represented as racemic mixtures.

The experimental triglyceride composition was derived from the results of the fractionation of the triglycerides on silver nitrate-TLC and from the composition of the

H. Brockerhof, Determination o f the positional distribution of fatty acids in glycerolipids, in: Methods in Enzymology, Vol. 35, Part. B, pp. 315-325, Ed. J. M. Lowenstein, Aca- demic Press, New York 1975. R. /. Vander W a l , J . Amer. Oil Chemists’ SOC. 37, 18 [1960].

@ M . H . Coleman, The structural investigation of natural fats, in: Advances in Lipid Research, Vol. 1 , pp. 2-64, Ed. R. Paoletti and D. Kritchevsky, Academic Press, New York 1963.

F E T T E . S E I F E N . A N S T R I C H M I T T E L 83. Jahrgang Nr. 10 1981 384

fatty acids esterified at the sn-2-position of the trigly- cerides of these fractions lo.

2.7. Gas liquid chromatography of triglycerides The triglycerides were separated by GLC using an

open-tubular glass column (length: 7 meter, inner dia- meter: 0.49 mm) wall-coated with CP-Sil 5 (Chrompack, Middelburg, The Netherlands). This type of phase allows a separation of the triglycerides according to their number of acyl carbons. A Hewlett-Packard 5830A gas chromatograph, equipped with a flame ionization detector was adapted with an all glass solid injector (Chrompack) and a stainless steel make-up T (Chrom- pack). The following conditions were used: carrier gas: nitrogen, 3 ml/min; auxiliary gas: nitrogen, 15 ml/min; injection port and detector temperature: 360° C; oven temperature: 350° C. Peak areas and retention times were determined by a Hewlett-Packard 18850A GC terminal.

3. R e s u l t s

3.1. Fatty acid composition of mango kernel triglycerides Stearic and oleic acids constituted about 85 O / o of the

fatty acids of mango kernel triglycerides. The remaining fatty acids were palmitic, linoleic, arachidic and linolenic acids (Table 1). Depending on the variety the stearic and oleic acid contents varied from 32.7 to 44.0°/o and from 43.7 to 53.4”/0 respectively. The contents of the stearic (x) and oleic (y) acids were inversely related (y = - 0 . 7 2 ~ + 74.7; r = -0.8603). Oleic and linoleic acids were the main fatty acids incorporated at the sn-2- position of the glycerol backbone of the triglycerides.

Table 1

Fatty acid composition o f mango kernel triglycerides and o f the f a t t y acids incorporated a t the sn-2-position o f the tri-

glycerides

Fatty acid (mole O/o)

Variety 16:O 18:O 18:l 18:2 18:3 20:O

A m i n i sn-2-position

A roenanis sn-2-position

Cecil sn-2-position

Gedong Gilletii sn-2-position

Goleck sn-2-position

H a d e n sn-2-position

M’Vuazi I sn-2-position

M’Vuazi III sn-2-position

M’Vuazi I V sn-2-position

6.7 44.0 43.7 3.6 0.6 1.1 89.8 8.0 8.7 33.1 51.0 4.2 0.3 0.6 88.0 9.2 8.1 40.5 44.8 3.8 0.3 0.9 88.5 9.4 7.8 39.8 46.5 4.1

8.2 34.5 49.1 5.6 - 0.4 86.7 11.7 7.8 36.2 49.5 4.9 - 0.7 86.8 11.7

9.7 36.2 45.0 6.3 0.1 0.4 86.8 12.1 7.6 32.7 53.4 4.6 0.4 1.5 88.3 9.2 7.5 34.5 48.9 6.9 - 0.5 84.5 14.1

,0.1 0.6 87.8 10.7

0.5 0.5 1 .o 1.8 0.4 0.9 0.3 0.8 0.8 1.2 0.3 0.8 0.3 0.6 0.6 0.6 0.6 0.9

1.5

2.0

2.4

1.5

1.8

1.3

2.5

1.1

1.6

-

-

-

-

-

-

-

-

-

3.2. Silver ion thin-layer chromatography Four bands were obtained by chromatography on silica

gel impregnated with AgNO,. Three of them were well separated. The fourth band, i. e. the slowest migrating band, was not homogeneous and may have contained at least two fractions. This pattern was obtained for

Table 2

Fatty acid composition of the triglycerides of mango kernel (M’Vuazi I V ) fat and of i ts fractions isolated b y silver nitrate thin-layer chromatograbhy

Fraction I Fraction I1 Fraction I11 Fraction IV Trig1 ycerides 39.2 37.8 15.8 7.2 calculated experimental

Fatty acids

16:O 11.6 0.1 6.6 0.1 3.9 0.1 5.7 0.2 8.1 0.1 7.5 - 18:O 52.4 0.6 29.7 0.2 12.1 0.2 12.2 0.4 34.5 0.4 34.5 0.5 18:l 34.5 99.3 57.3 86.5 67.5 69.5 43.8 41.3 49.0 85.6 48.9 84.5 18:2 - - 4.8 13.2 16.1 30.5 32.5 47.9 6.7 14.8 6.9 14.1

- 5.8 10.2 0.5 0.7 0.6 0.9 18:3 - - - - 1.6 - 20:o 1.5 - 1.6

saturated 65.5 0.7 37.9 0.3 16.0 0.3 17.9 0.6 mono-unsaturated 34.5 99.3 57.3 86.5 67.5 69.5 43.8 41.3 di-unsaturated 4.8 13.2 16.1 30.5 32.5 47.9 tri-unsaturated 0.4 5.8 10.2 double bonds

- 0.4 - 1.2 - - - -

per molecule 1.04 2.01 3.03 3.79

The amounts of oleic and linoleic acids esterified at the sn-2-position were nearly constant for the different mango varieties.

each of the examined varieties. The fatty acid com- position of the bands and of the fatty acids esterified at the sn-2-position of the triglycerides of the variety M’Vuazi IV is given in Table 2. Assuming that each

F E T T E S E I F E N * A N S T R I C H M I T T E L 83. Jahrgang Nr. 10 1981 385

composition that the triglycerides of the fractions I-IV have on the average 1.04, 2.01, 3.03 and 3.79 double bonds per triglyceride molecule. From the proportion of each band and the fatty acid composition of each fraction, the fatty acid composition of the triglycerides was calculated and compared to the experimental com- position. The figures of the calculated compositions and the experimentally determined compositions are shown in Table 2. The calculated values agree very well with the experimental values, indicating that no gross errors have been made.

3.3. T h e triglyceride composition The contents of the individual triglycerides as cal-

culated by the 1,3-randorn-Z-random distribution theory (A) and as measured experimentally using silver nitrate thin-layer chromatography (B) are presented in Table 3.

00- - M

d I I

Table 3

Triglyceride composition o f mango kernel f a t as calculated by the I,S-random-2-random theory ( A ) and as found by

silver nitrate thin-layer chromatography ( R )

Triglyceride components (mole " l o )

.MONO- 0 DI-

Variety

A m i n i A 53.9 28.7 4.8 3.8 2.6 2.0 4.2 B 53.1 28.3 5.1 4.5 2.3 1.6 5.1

Aroenanis A 37.5 37.3 3.9 9.3 3.9 2.0 6.1 B 39.5 36.4 3.7 9.6 3.2 2.6 5.0

Cecil A 51.0 30.8 5.4 4.7 3.3 1.3 3.5 B 51.6 29.7 5.2 5.5 2.2 1.5 4.8

Gedong gilletii A 47.2 33.3 5.7 5.9 4.0 1.0 2.9 B 47.3 31.5 5.0 7.0 3.1 1.4 4.7

Goleck A 38.4 35.0 5.2 8.0 4.7 3.0 5.7 B 41.2 33.6 4.2 7.9 4.0 2.7 6.4

Haden A 39.7 36.2 5.4 8.3 4.9 1.8 3.7 B 42.1 35.0 4.5 8.7 3.1 1.4 5.2

M V u a z i I A 45.4 30.2 6.3 5.0 4.2 4.3 4.6 B 48.0 29.8 4.1 5.3 3.4 4.0 5.4

Af'Vuazi 111 A 33.0 38.8 3.4 11.4 4.0 2.5 6.9 B 36.7 38.4 3.7 11.4 3.5 1.6 4.7

M'Vuazi I V A 35.9 34.3 6.0 8.2 5.7 3.6 6.3 B 38.9 32.7 5.0 9.6 4.8 2.8 6.2

Only the triglyceride components which contain 1, 2 and 3 double bonds and which represent more than 0.5OIo are presented. The triglycerides containing more than 3 double bonds are not presented, because they could not be isolated as homogeneous fractions. The experi- mental triglyceride composition agrees very well with the expected composition, calculated as predicted by the hypothesis of R. J . Vander W a l s . On the other hand, when the triglyceride composition of the different mango varieties was calculated supposing a complete 1,2,3- random distribution, the following contents were ob- tained: SSS: 7.0-13.3 V o , SOS: 26.9-31.6O/o, SOO: 29.4-34.8 O / o , SLS: 2.2-4.4 O/o, 000: 9.1-14.7 O/o and SLO: 5.0-8.2 "/a. The content of the triglyceride com- ponents SSS, SOS and 000 are quite different from the experimental values.

For each variety the respective sum of the mono- saturated and disaturated triglyceride contents was plottet against the saturated fatty acid content of the oil

(Fig. 1). It can be seen that, as the saturated fatty acid content increases, the monosaturated triglyceride content decreases whereas the content of the disaturated

9 40- + 3 l-

cn

a a

a

// i o i s 50 55 SATURATED FATTY ACIDS IN TOTAL TRIGLYCERIDES (MOLE %)

Fig. 1. Relationship between the content of the triglyceride components containing saturated fatty acids (x) and the saturated fatty acid content of the total triglycerides (y) of

mango kernel fat Linear regressions: monosaturated glycerides: y = - 1 . 0 3 ~

+ 85.8; disaturated glycerides: y = 1.73~-331.3

triglyceride content increases. Linear relationships were also found between the mono-oleoyl-, di-oleoyl- and tri- oleoylglycerols and the oleic acids content of the oil (Fig. 2).

MONO- 0 DI- x TRI- \

45 50 5 7 OLEIC ACIDS IN TOTAL TRIGLYCERIDES (MOLE %)

Fig. 2. Relationship between the content of the triglyceride components containing oleic acids (x), and the oleic acid

content of the total triglycerides (y) of mango kernel fat Linear regressions: mono-oleoglycerides: y = - 1 . 7 2 ~ + 132.7; di-oleoglycerides: y = 1 . 0 4 ~ - 17.3; triolein: y =

0 . 7 7 ~ - 29.4

3.4. Gas liquid chromatography of the triglycerides A characteristic gas liquid chromatogram of the

separation of the triglycerides of the variety MVuazi IV is shown in Figure 3. Four peaks, corresponding to the carbon numbers CS0, C5*, C,, and C58, were obtained. The major fraction was Cj4 followed by C52. The results of the analysis of the different mango varieties and the

F E T T E . S E I F E N A N S T R I C H M I T T E L 83. Jahrgang Nr. 10 1981 386

Table 4

Triglyceride composition of mango kernel fa t as found by gas liquid chronzatograkhy and as calculated by the 1,3-randorn-Z-random theory

Gedong MVuazi MVuazi MVuazi Amini Aroenanis Cecil Gilletii Goleck Haden I III IV

c50 1.1 0.9 1.4 1.6 1.6 1.4 1.4 1.3 c5, 17 .1 16.8 19.4 21.5 19.5 19.9 19.9 20.0 c,, 77.2 78.4 74.8 71.9 73.7 72.8 75.0 74.9 c,, 4.6 3.9 4.4 5.0 5.2 5.9 3.7 3.8

5 = Symbols denote total acyl carbon number

52

0 ' 2 ' d min

Fig. 3. Gas liquid chromatography elution pattern of mango kerngl triglycerides, obtained using an open-tubular column

wall-coated with CP-SiI 5 The triglycerides are denoted by their carbon number (total

acyl carbon atoms)

values of the different triglyceride fractions, as cal- culated by the l,3-random-2-random distribution hypo- thesis are shown in Table 4. It can be seen that the ex- perimental values are in general in agreement with the expected values. It was also found that the difference between the expected composition as calculated using the 1 ,d-random-Z-randorn hypothesis and the 1,2,3- random hypothesis was very small.

From these results and from the fatty acids incor- porated at the sn-2-position of the triglycerides it can be deduced that the main triglyceride components of mango kernel fat are sn-2-oleo-1,3(3,1)-palmitostearin, sn-2-oleo-1,3(3,~)oleopalmitin, sn-2-oleo-1,3-distearin, triolein, sn-2-linoleo-l,3-distearin and sn-Z-linoleo-l,d- (3,l)-oleostearin.

4. D i s c u s s i o n

Different hypotheses describing the distribution of the fatty acids between the positions of the glycerols have been put forward to explain the heterogenity of the triglycerides of an oil. The I &%random distribution

1.4 1.5 1 .7 1.4 1.6 2.1 1.0 1.2 1 .1 1.3 18.3 20.7 20.2 20.3 20.4 23.5 16.5 19.0 17.9 19.4 75.6 73.2 74.9 74.9 72.5 68.3 80.0 76.9 77.4 75.2 4.7 4.6 3.2 3.4 5.5 6.1 2.5 2.9 3.6 4.1

theory states that one pool of fatty acids is incorporated at random a t the three different hydroxyl groups of the glycerol backbone of the triglycerides of the oil. This theory holds for animal depot fats, but proved incorrect for seed oils. R. 1. Vander W a l e and M. H . Coleman9 proposed independently for vegetable oils the 1,3- random-2-random distribution hypothesis. Support for the l,3-random-2-random theory has been based on the fractionation of triglycerides by silver nitrate thin-layer chromatography and on the pancreatic lipase hydrolysis of triglycerides11~i2. In fact, the analysis of the trigly- cerides of seed oils showed that palmitic and stearic acids are almost exclusively esterified at the sn-1- and sn-3-positions of seed oil triglycerides. In some oils with a high palmitic acid content palmitic acids may also be incorporated at the sn-2-position 13. The 1,3-random- 2-random theory takes the fatty acids, incorporated at the sn-2-position as base for the calculation of the individual triglyceride components. The incorporation of saturated fatty acids at the sn-2-position determines the presence of fully saturated triglycerides in the oil. However, in oils with a low saturated fatty acid content the triglyceride composition predicted by the 1,2,3- random theory and by the 1,3-random-Z-randorn theory may be indistinguishablei4. For mango seed oil which is characterized by a high saturated fatty acid content the difference between the values as calculated by both theories was obvious. The results of the silver nitrate thin-layer chromatography alone allowed to conclude that the fatty acids of mango seed oils were distributed according to the 1,3-random-2-random hypothesis.

Further analysis of the positional distribution of fatty acids within the triglycerides of vegetable oils showed that the composition of the fatty acids incorporated at the sn-1- and sn-3-position was different 15. Till yet no restriction in the incorporation of one of the common fatty acids, palmitic, stearic, oleic, linoleic acids, at the sn-1- or sn-3-position has been reported. Arachidic acids on the other hand show some preference for the sn-%position. Considering these observations S . Tsuda

11 R. J . Vander W a l , Triglyceride structure, in: Advances in Lipid Research, Vol. 2, pp. 1-16, Ed. P . Paoletti and D. Kritchevsky, Academic Press, New York 1964. C. B. Barrett, M . S. J . Dallas and F. B. Padley, J. Amer. Oil Chemists' SOC. 40, 580 [1963].

13 A . Seneubta and S. K . Rov Choudhury, .T. Amer. Oil Che- . I mists' go;. 55, 621 [1978]. - M . L. Blank, B. Verdino and 0. S. Privett, J. Amer. Oil Chemists' SOC. 42, 87 [ 19651.

15 H . Brockerhoff and M . Yurkowski, J. Lipid Res. 7, 62 [ 19661.

F E T T E . S E I F E N . A N S T R I C H M I T T E L 83. Jahrgang Nr. 10 1981 387

proposed the 1 -random-2-random-3-random distribution hypothesis, which states that at each hydroxyl group of the glycerol hackbone of the triglycerides a different pool of fatty acids is esterified at random 16. With regard to the 1,3-random-2-random theory, the 1 -random-2- random-3-random theory offers little advantage in the calculation of the different triglyceride components. Moreover, it is not possible to measure directly with the present analytical methods the content of the isomers of the racemic triglyceride components 17.

The results of the gas liquid separations of mango seed triglycerides show that the fatty acids were distri- buted at random within the triglycerides. But it was not possible to show that the fatty acids are distributed either according to the 1,2,3-random distribution theory or according to the 1,3-randorn-2-random distribution theory. The expected values, as calculated by both theories, were similar. Higher contents of palmitic acids, i. e. higher differences in the carbon number, are needed to obtain significant differences between the triglyceride contents as calculated by the 1,2,3-random theory and the 1,3-random-2-random theory. Nevertheless the use of open-tubular wall-coated columns may become a

16s. Tsuda, Yukagaku 11, 408 [1962]. l7 H. Brockerhofl, Lipids 6, 942 [1971].

quick and reliable method to study the heterogenity of triglycerides 18.

This study shows that the triglyceride composition of mango kernel fat is in agreement with the 1,3-random- 2-random distribution theory of Vander Wal and that the main triglycerides are sn-1,3-disaturated-2-oleoyl- glycerols and sn- 1 (3)-saturated-2,3( 1,2)-di-oleoylglyce- rols. The fatty acid composition of mango kernel fat differs from cocoa butter by its higher stearic acid con- tent and its lower palmitic acid content. With regard to its fatty acid composition and its triglyceride com- position, mango kernel fat is comparable to shea butter, the fat of the seeds of Butyrospermum par& le? lo. The stearin fraction of the latter fat, when incorporated in chocolate products, gives them an improved stability Mango kernel fat can in this regard also become an economically valuable fat.

Received 10th November 1980.

18 A. Monseigny, P. Y . Vigneron, M. Levacq and F. Zwoboda, Rev. franG. Corps gras 26, 107 [1979].

is G. Iurriens, Analysis of glycerides and composition of natural oils and fats, in: Analysis and characterization of oils, fats and fat products, Vol. 2, pp. 217-298, Ed. H. A. Boekenoogen, Interscience Publishers, London 1968. B. Andersen, Confect. Mf. Mktg. 11, 6 [1974].

Die Bestimmung des Fettgehaltes von Kartoffelfritierprodukten durch NMR * Von B. P u t z " : ?

Aus der Bundesforschungsanstalt f u r Getreide- und Kartoffelverarbeitung, Detmold

Nachdem sich die Messung des Fettgehaltes mittels NMR bei Raps und Mais gut bewahrt hat, wurde versucht, dieses Ver- fahren auch fur die Bestimmung des Fettgehaltes von Kar- toffelfritierprodukten (Chips und Pommes frites) einzusetzen. Zunachst zeigte sich, daO der hohe Wassergehalt der Poinmes frites eine Vortrocknung notig macht. Dieses kann bei Chips unterbleiben, sofern deren Restwassergehalt 3 o / o nicht iiber- steigt. Eine Messung mittels NMR ist nur dann moglich, wenn als Referenzprobe das gleiche Fett oder 01 wie bei der Pro- duktherstellung verwendet wird. Weiterhin wurde festgestellt, daD die sich wahrend des Fritierens laufend verschlechternde Fettqualitat zu erheblichen Abweichungen bei der NMR-Mes- sung fuhrt, was bedeutet, daO die zu messende Produktprobe immer zusammen mit einer Fettprobe gezogen werden muO. Werden diese Bedingungen alle eingehalten, so konnen im Ver- gleich zwischen NMR-Messung und Petrolather-Extraktion immer noch Differenzen von bis zu 4 O/O Fett auftreten.

1. E i n l e i t u n g

Bei der Herstellung von Kartoffelfritierprodukten wie Chips odor Pommes frites ist der Fettgehalt eines der wichtigsten Qualitattskriterien. Chips enthalten bei einem Restwassergehalt von ca. 2 O / o etwa 35-42 O/o

Fett, wahrend in vorfritierten Pommes frites bei Rest- wassergehalten von maximal 65-76 % die Hochstgrenze

* Vortrag anlai3lich der 36. DGF-Vortragstagung in Kiel am 9. September 1980. Veroffent1.-Nr. 4828 der Bundesforschungsanstalt fur Ge- treide- und Kartoffelverarbeitung, Detmold

"" Anschrift des Verfassers: Dr. B. Putz, Bundesforschungs- anstalt fur Getreide- und Kartoffelverarbeitung, Institut fur Starke- und Kartoffeltechnologie, Postfach 23, 4930 Det- mold.

388

Determination of Fat Content in Fried Potato Products by NMR Since determination of the fat content by NMR in rape seed

and mize is well proved it was tried to use this method for fat content determination in fried potato products (chips and pommes frites), too. At first it was found that the high water content of pommes frites requires predrying. This can remain undone for chips, if the residual water content does not exceed 3 O/O. Measurement by NMR is only possible if the same fat or oil of production is used as reference. Further it was found that the fat quality which deteriorates continuously during frying leads to considerable deviations in NMR measurement. This involves that the product sample, which is to investigate, has to be taken always together with the fat sample. If all these requirements are met, differences up to 4 0 / 0 fat can occur in comparison between NMR determina- tion and petrolether extraction.

des Fettgehaltes bei 8.5Vo liegtl. Bisher ist es immer noch iiblich, den Fettgehalt der Produkte mittels Petrol- atherextraktion zu ermitteln. Diese Methode hat jedoch den Nachteil, dai3 sie mehrere Stunden dauert und da- mit fur eine laufende Betriebskontrolle ungeeignet ist. Da der Fettgehalt wesentlich vom Fritiervorgang ab- hangig ist, bedarf es fur diese Produkte einer Schnell- methade, die es ermoglicht, die Produktionsadagen nach dem ermittelten Fettgehalt einzustellen. Nur so wisd es moglich, fruhzaitig zu hohe Gehalte festzustellen und Abhilfe zu schaffen. Hohe Fettgehalte irn Chips und

Bund fur Lebensmittelrecht und Lebensmittelkunde, Richt- linien fur Kartoffelverarbeitungserzeugnisse, Behrs Verlag, Hamburg 1975, Heft 82.

F E T T E . S E I F E N . A N S T R I C H M I T T E L 83. Jahrgang Nr. 10 1981