Synthesis of the structured lipid 1,3-dioleoyl-2-palmitoylglycerol from palm oil

Article abstract of DOI:10.1007/s11746-006-0935-2

Human milk fat contains 20-25% palmitic acid (16:0) and 30-35% oleic acid (18:1). More than 60% of the palmitic acid occurs at the sn-2 position of the glycerol backbone. Palm oil is a rich source of both palmitic and oleic acids. The structured lipid 1,3-dioleoyl-2-palmitoylglycerol (OPO) is an important ingredient in infant formula. OPO was synthesized from palm oil by a three-step method. In the first step, low-temperature fractionation was applied to palm oil FA, yielding a palmitic acid-rich fraction (87.8%) and an oleic acid-rich fraction (96%). The palmitic acid content was further increased to 98.3% by transforming palmitic acid into ethyl palmitate. In the second step, esterificalion of ethyl palmitate and glycerol catalyzed by lipase Novozym 435 under vacuum (40 mm Hg) was employed for the synthesis of tripalmitin. Finally, OPO was obtained by the reaction of tripalmitin with oleic acid catalyzed by Lipase IM 60. In this final step, the TAG content in the product acylglycerol mixture was 97%, and 66.1 % oleic acid was incorporated into TAG. Analysis of the FA composition at the sn-2 position of TAG showed 90.7 mol% of palmitic acid and 9.3 mol% of oleic acid. OPO content in the product TAG was ca. 74 mol%. Thus, an efficient method was developed for the synthesis of OPO from palm oil.

Full text of DOI:10.1007/s11746-006-0935-2

Synthesis of the Structured Lipid
,3-Dioleoyl-2-palmitoylglycerol from Palm Oil
Ming-Lung Chen, Shaik Ramjan Vali, Jih-Yao Lin, and Yi-Hsu Ju*
1
Department of Chemical Engineering, National Taiwan University of Science and Technology, Taipei 106-07, Taiwan
ABSTRACT: Human milk fat contains 20–25% palmitic acid
Earlier studies used pure tripalmitin (PPP) and oleic acid for
16:0) and 30–35% oleic acid (18:1). More than 60% of the the synthesis of the structured lipid OPO (4,5). However, large-
(
palmitic acid occurs at the sn-2 position of the glycerol backbone.
Palm oil is a rich source of both palmitic and oleic acids. The
structured lipid 1,3-dioleoyl-2-palmitoylglycerol (OPO) is an im-
portant ingredient in infant formula. OPO was synthesized from
palm oil by a three-step method. In the first step, low-temperature
fractionation was applied to palm oil FA, yielding a palmitic acid-
rich fraction (87.8%) and an oleic acid-rich fraction (96%). The
palmitic acid content was further increased to 98.3% by transform-
ing palmitic acid into ethyl palmitate. In the second step, esterifi-
scale applications of these methods are expensive because of
the high cost of the substrates. To achieve an economically fea-
sible process, it is important that starting substrates such as
palmitic acid and oleic acid be isolated and purified from an in-
expensive source. Palm oil is such a source for both palmitic
and oleic acids, which constitute 44.0 and 39.2%, respectively,
of the total palm oil FA content. Therefore, isolating and puri-
fying these FA from palm oil may offer a considerable reduc-
cation of ethyl palmitate and glycerol catalyzed by lipase tion in the cost of the substrates used to synthesize OPO.
Novozym 435 under vacuum (40 mm Hg) was employed for the
The synthesis of structured TAG is usually carried out by
synthesis of tripalmitin. Finally, OPO was obtained by the reac- the transesterification of different TAG or by the acidolysis
tion of tripalmitin with oleic acid catalyzed by Lipase IM 60. In
this final step, the TAG content in the product acylglycerol mix-
ture was 97%, and 66.1% oleic acid was incorporated into TAG.
Analysis of the FA composition at the sn-2 position of TAG showed
reaction between a TAG and a FA. Several analytical meth-
ods, such as TLC (6), enzymatic hydrolysis (7), or chemical
degradation using a Grignard reagent (8) and C NMR (9),
have been used in the past for the determination and quantifi-
cation of TAG and their positional isomers. None of the above
methods is capable of identifying the molecular species in the
TAG, and only positional distributions of the FA are deter-
mined (10).
13
90.7 mol% of palmitic acid and 9.3 mol% of oleic acid. OPO con-
tent in the product TAG was ca. 74 mol%. Thus, an efficient
method was developed for the synthesis of OPO from palm oil.
Paper no. J10743 in JAOCS 81, 525–532 (June 2004).
KEY WORDS: Esterification, lipases, low-temperature crystal-
Silver-ion HPLC is an extremely powerful tool for the sep-
lization, oleic acid, 1,3-oleoyl-2-palmitoylglycerol, palmitic aration and quantification of positional and geometrical iso-
acid, palm oil, structured lipids, tripalmitin.
mers of FA and TAG. The separation takes place according to
the number, geometrical configuration, and position of double
bonds in the molecule (6). RP-HPLC is an alternative to sil-
Structured lipids, such as 1,3-dioleoyl-2-palmitoylglycerol ver-ion HPLC. With the former technique, the separation of
OPO), have attracted attention recently owing to their poten- individual TAG molecules is by both chain-length and degree
(
tial pharmaceutical and nutraceutical applications. Breast of unsaturation of FA. However, in many cases the order of
milk fat is one of the main sources of nutrients and energy for elution of the TAG species with silver-ion HPLC is quite dis-
infants. About 50–60% of the dietary energy that an infant tinctive and easy to understand, and, depending on the degree
needs is provided by human milk fat (HMF) (1). Palmitic and of unsaturation, much better resolution is possible, in contrast
oleic acids are the two most abundant FA present in all HMF to HPLC in the reversed-phase mode (6,11). Thus, the silver-
and are the most interesting from the standpoint of human nu- ion technique does not replace RP-HPLC but complements it.
trition. Palmitic acid represents about 25% of the total milk Various detection methods have been used in the analysis of
FA and accounts for more than 10% of the infant’s total en- TAG separated by HPLC. Nagao et al. (5) used a refractive
ergy intake. The stereospecific distribution of FA on TAG index (RI) detector to analyze structured lipids containing
structures in HMF plays a valuable role and influences the oleic and palmitic acids that had been separated by silver-ion
fate of fat absorption in infants. In HMF the saturated FA, pre- HPLC. ELSD offers advantages over RI and UV detection in
dominantly palmitic acid, are located primarily at the sn-2 po- that no baseline drift occurs, and there are no limitations in the
sition (>60%) of the glycerol backbone (2,3), whereas the sn- use of mobile phase solvents. Furthermore, ELSD is stable and
1
,3 positions are mainly occupied by monounsaturated FA. sensitive when elution gradients are used, as is required to re-
Thus, the main component of the milk dienoic TAG is OPO.
solve TAG mixtures. The use of gradient elution is an alterna-
tive approach for reducing the retention times of higher M.W.
saturated TAG and for improving the chromatographic resolu-
tion (12). In this paper, we report an efficient method for the
*
To whom correspondence should be addressed at Department of Chemical
Engineering, National Taiwan University of Science and Technology, 43
Sec. 4, Keelung Road, Taipei 106-07, Taiwan. E-mail: ju@ch.ntust.edu.tw
Copyright © 2004 by AOCS Press
525
JAOCS, Vol. 81, no. 6 (2004)

526
M.-L. CHEN ET AL.
purification of palmitic and oleic acids from saponified palm vessel containing 20 mL of n-hexane. To this was added 0.4 g
oil; the synthesis of OPO from palmitic acid, oleic acid, and of immobilized PS-30 lipase (10% based on FFA weight), and
glycerol; and the separation and quantification of OPO by sil- the reaction was carried out at 37°C in a thermostated water
ver-ion HPLC with detection by ELSD.
bath with stirring at 150 rpm. At various reaction times, a 100-
µL sample was taken for the analysis of palmitic acid content
in the unesterified FFA. To this sample, 100 µL of n-hexane
containing 0.05 mmol tridecanoic acid was added as an inter-
EXPERIMENTAL PROCEDURES
Materials. Refined palm oil was purchased from Blessing nal standard. After adding 200 µL of 1 N KOH solution, the
Brothers Industrial Company (Taipei, Taiwan). Lipase from content was vortexed vigorously, and the FFA soap and its
Pseudomonas cepacia (Lipase PS “Amano” 30, powder) was ester were separated by centrifuging (5,000 × g) for 5 min.
a product of Amano Enzyme Inc. (Nagoya, Japan). Lipozyme FFA in the aqueous phase were extracted with 300 µL of n-
IM 60 (Rhizomucor miehei lipase immobilized on a macro- hexane after returning it to an acidic pH (pH <2) with 200 µL
porous anion exchange resin) and Novozym 435 (Candida of aqueous 2 N HCl. After removing the solvent by nitrogen
antarctica lipase immobilized on a macroporous acrylic resin) purging, the resulting FFA were transformed into methyl es-
were supplied by Novo Nordisk A/S (Bagsvaerd, Denmark) ters by heating with BF /methanol (20%) and analyzed by GC.
3
as gift samples. Microporous polypropylene powder was ob- The degree of conversion of palmitic acid to ethyl palmitate
tained from Akzo Nobel Faser AG (Obernburg, Germany). was expressed as the initial weight of palmitic acid minus the
Silica gel 60M (230–400 mesh) was purchased from final weight of palmitic acid divided by the initial weight of
Macherey-Nagel GmbH & Co., KG (Düren, Germany). Stan- palmitic acid multiplied by 100.
dard TAG samples were obtained from Sigma Chemical Com-
(v) Lipase-catalyzed synthesis of PPP. Novozym 435 and
pany (St. Louis, MO). The internal standard pentadecanoic Lipozyme IM 60 were used as biocatalysts for the synthesis
acid (>99%) was obtained from Aldrich Chemical Co. (Mil- of PPP from ethyl palmitate and glycerol. The reaction was
waukee, WI). All solvents were either of HPLC or analytical performed in a solvent-free system at 60°C under vacuum in
reagent grade and were obtained from commercial sources.
a two-necked test tube. One neck was connected to a vacuum
Methods. (i) Enzyme immobilization. PS-30 lipase was im- pump, and the other was closed with a rubber septum. A typi-
mobilized onto a microporous polypropylene solid support cal reaction was carried out with 0.58 mmol of glycerol, 1.76
following the method of Huang et al. (13). Activity of the im- mmol of ethyl palmitate, and 50 mg of lipase (10% based on
mobilized lipase was assayed by the olive oil emulsion method ethyl palmitate weight). After the FA ester was melted, the
(
13). The immobilized Lipase PS-30 used in this study had a reaction was initiated by adding lipase, and the reactor pres-
specific activity of 8 units/(mg immobilized lipase).
ii) Preparation of FFA from palm oil. FFA, obtained by
sure was maintained at 40 mm Hg.
(
(vi) Analysis of the reaction product by TLC-FID. The con-
the saponification of palm oil, were prepared according to the tents of TAG, DAG, MAG, and ethyl esters in the reaction
method described by Haagsma et al. (14). product were measured with a thin-layer chromatograph/FID
iii) Separation of palmitic and oleic acids. The separation analyzer (Iatroscan MK-5; Iatron Co., Tokyo, Japan) after devel-
(
of palmitic and oleic acids from the FFA mixture of saponi- opment in 1,2-dichloroethane/chloroform/acetic acid (92:8:0.1,
fied palm oil was carried out by low-temperature solvent crys- by vol) (15). Sample (100 µL) was dissolved in 1 mL of chloro-
tallization according to the flow chart shown in Figure 1. Typ- form and filtered through a cotton plug. The volume of the solu-
ically, 400 mL of acetone was added to a glass vial contain- tion was reduced to 200 µL by passing N gas through the solu-
2
ing 10 g of FFA. This mixture was stirred at 40°C until all FA tion. One microliter of this lipid solution was spotted on silica
were dissolved. The solution was allowed to cool to room gel Chromarods-SIII (Iatron Co.) and developed in the above
temperature and was then stored in an ultra-low-temperature solvent system. The Iatroscan was operated with hydrogen and
freezer chamber (MDF-192; Sanyo Gunma, Japan) at −25°C air flow rates of 160 and 2000 mL/min, respectively.
for 24 h. The content was filtered under suction using an Ace
(vii) Purification of TAG from the reaction mixture by sil-
Buchner funnel (25–50 µm). The solid fraction with palmitic ica gel column chromatography. Two grams of reaction mix-
acid as the major FA was designated as FFA-I and was fur- ture was dissolved in 12 mL of hexane and then applied to the
ther purified by selective enzymatic esterification, giving a head of a column (32 × 2.4 cm i.d.) of silica gel (60 g, 60–200
mixture of FA ethyl esters with 98.3% ethyl palmitate. The mesh; Merck, Darmstadt, Germany). Unreacted ethyl palmi-
liquid fraction (designated as FFA-II) was further purified by tate was eluted first with a mixture of n-hexane/ethyl acetate
successive low-temperature crystallizations, as shown in Fig- (200 mL of 99:1, vol/vol), followed by the elution of TAG
ure 1, to yield a solid phase with 96% oleic acid.
with n-hexane/ethyl acetate (500 mL of 97:3,vol/vol) at a flow
iv) Lipase-catalyzed selective esterification. The enrich- rate of about 2 mL/min; eluents were collected in 20-mL frac-
(
ment of palmitic acid in FFA-I was accomplished by the ester- tions. The purity of each fraction was analyzed on a silica
ification of FFA-I with ethanol catalyzed by immobilized Li- TLC plate (6 × 2 cm, 250 µm) using authentic standards.
pase PS 30 in n-hexane. Typically, FFA-I (4 g, average M.W. Plates were developed with a solvent system composed of n-
=
259.4) and absolute ethanol (0.9 mL, purity 99.5%) with a hexane/ethyl acetate/acetic acid (90:10:0.5, by vol) as devel-
molar ratio of 1:1 were put into a 50-mL screw-capped glass oping solvent. After air-drying, the spots were visualized by
JAOCS, Vol. 81, no. 6 (2004)

SYNTHESIS OF STRUCTURED LIPIDS FROM PALM OIL
527
FIG. 1. Flow chart for low-temperature and enzymatic fractionation of mixed FA of palm oil.
spraying with methanolic sulfuric acid (10%), followed by
(x) Purification of the acidolysis reaction product. At the
heating at 150°C. Fractions containing only TAG were com- end of incubation, the enzyme was removed from the reaction
bined, and the solvent was removed with a rotary evaporator. product by passing it through a bed of anhydrous sodium sul-
PPP with a purity greater than 99% and a corresponding re- fate. The FA composition of the TAG product was determined
covery of 92.5% was obtained by silica gel chromatographic by GC analysis after separation by TLC. Twenty milligrams of
purification.
reaction product was dissolved in 0.2 mL of hexane, and TAG
(
viii) Acidolysis of PPP. PPP (80.7 mg, 0.1 mmol), oleic were separated by TLC on precoated silica gel G 60 F 254
acid (225.6 mg, 0.8 mmol), 20 mg of Lipase IM 60, and 3 mL plates (1-mm thick; Merck), developed with n-hexane/diethyl
of hexane were placed in a 10-mL glass vial, which was then ether/acetic acid (90:10:0.1, by vol). The bands were visual-
sealed. The mixture was incubated for 24 h in a thermostated ized under UV light after spraying with 0.2% 2,7-dichlorofluo-
water bath at 50°C and magnetically stirred at 150 rpm. rescein in ethanol. Bands corresponding to TAG were scraped
Aliquots of the mixture (100 µL) were withdrawn at definite off and methylated with 6% methanolic HCl at 75°C for 2 h
intervals for the analysis of acylglycerides.
ix) Analysis of the acidolysis reaction product. The extent and 1 mL of 0.1 M KCl, centrifuged, and the top hexane layer
of the reaction was determined by taking samples at regular recovered and its FA composition determined by GC.
time intervals and analyzing them by TLC-FID. The contents (xi) Regiospecific analysis of TAG. The positional distribu-
(16). FAME were separated by the addition of 2 mL of hexane
(
of TAG, DAG, MAG, and FFA in the reaction mixture were tion of FA in TAG was determined by the method of Liu et al.
analyzed by a TLC-FID analyzer after development in 1,2- (17). Typically, 100 mg of TAG was dissolved in 0.2 mL of
dichloroethane/chloroform/acetic acid (92:8:0.1, by vol) (15). DMSO and mixed with 3 mL of Tris-HCl buffer (pH 7.6), 0.1
JAOCS, Vol. 81, no. 6 (2004)

528
M.-L. CHEN ET AL.
mL of 0.2% (wt/vol) of bile salts, 0.2 mL of 22% (wt/vol) cal- RESULTS AND DISCUSSION
cium chloride, and 100 mg pancreatic lipase in a 10-mL test
tube. The mixture was incubated in a water bath at 37°C for 5 Separation and purification of palmitic and oleic acids from
min. Thereafter, the reaction mixture was extracted with di- palm oil. Table 2 shows the FA compositions of palm oil be-
ethyl ether and 2-MAG was separated by a silica gel G TLC fore and after purification from saponified palm oil. The re-
plate (Merck) after developing in n-hexane/ethyl acetate/acetic sults show that the crystallization of FA in acetone at −25°C
acid (50:50:1, by vol). The band corresponding to 2-MAG was yielded two distinct fractions: a solid fraction containing
scraped off and extracted with ethyl acetate, methylated, and mainly saturated FA (mostly palmitic acid, ~88 wt%) and a
then analyzed by GC.
xii) GC analysis of the FA composition. FFA were con- oleic acid). Stearic acid, which accounted for ca. 8.7%, was
verted into their corresponding FAME by heating with the other major acid in the solid fraction.
BF /methanol (20%). TAG and MAG were methylated with
Further enrichment of palmitic acid by the removal of
% methanolic HCl at 75°C for 2 h. The FA composition was stearic acid via solvent crystallization was difficult because of
liquid fraction containing mainly unsaturated FA (mostly
(
3
6
analyzed by a model 8700F China Chromatographic gas chro- their similarity in structure, polarity, and solubility. On the
matograph (Taipei, Taiwan) equipped with an FID. The col- other hand, removal of the remaining saturated acids in the liq-
umn used was an SP-2330 (30 × 0.25 mm i.d.; Supelco, Belle- uid fraction was carried out in the second stage by low-temper-
fonte, PA). The injector and detector temperatures were set at ature solvent crystallization. The resulting crystals of saturated
2
50 and 270°C, respectively. The column was held at 180°C FA were removed by filtration. As shown in Table 2, this step
for 10 min, increased to 235°C at a constant rate of 15°C/min, removed 78% of the saturated FA, raising the oleic acid con-
and then held at 235°C for 3 min. One microliter of sample tent from 69.5 to 75.8 wt% with a yield of 97.4%. The liquid
was injected at a split ratio of 1:50.
xiii) Separation and quantification of TAG of the acidoly- 21.3% linoleic acid. Previous reports from this laboratory sug-
fraction from the second-stage crystallization still contained
(
sis product by silver-ion HPLC-ELSD. A JASCO HPLC sys- gest the use of acetonitrile for the enrichment of PUFA (18,19).
tem with a ChromSpher 5 Lipids silver-ion chromatography When crystallization was performed at −20°C and an acetoni-
column (250 × 4.6 mm × 1/4″; Chrompack, Middelburg, The trile-to-FFA ratio of 60 mL/g, after 24 h the oleic acid content
Netherlands), and an ELSD 2000 (Alltech, Deerfield, IL) was was raised from 75.83 to 96 wt%, with a yield of 81.6%. The
used for the TAG class composition analysis. The separation linoleic acid content was reduced from 21.3 to 1.0 wt%.
of TAG was carried out using a binary solvent gradient pro-
Lipase-catalyzed selective esterification of FFA-I and al-
gram of A (acetone) and B (dichloromethane). The gradient cohol. Our preliminary results on the selective esterification
program of these two mobile phases is shown in Table 1. The of different FA with alcohols showed that lipase from P. cepa-
mobile-phase flow rate was 0.8 mL/min. The column was cia exhibited substantial selectivity for palmitic acid over
held at ambient temperature. ELSD was used at a 50°C drift stearic acid (data not shown). The palmitic acid content in the
tube temperature. Air at a flow rate of 1.6 mL/min was used solid fraction of the first solvent crystallization (FFA-I) was
to evaporate the solvent in the heated chamber. The lipid sam- further enriched by the selective esterification of FA with
ples were dissolved in dichloromethane (20 mg/mL), and 20 ethanol catalyzed by immobilized P. cepacia lipase. Figure 2
µL of solution was injected.
ix) Standard calibration curves for each individual TAG. and the degree of esterification. As expected, P. cepacia
Lipid standards such as PPP, 1,3-dipalmitoyl-2-oleoylglycerol showed high selectivity for palmitic acid and converted it into
POP), 1,2-dipalmitoyl-3-oleoylglycerol (PPO), 1,3-dioleoyl- ethyl esters, leaving more stearic acid in the residual FA.
shows a time-course plot of the formation of ethyl palmitate
(
(
2
-palmitoylglycerol (OPO), 1,2-dioleoyl-3-palmitoylglycerol After 12 h of reaction, the degree of esterification was 89.5%,
(
(
OOP), and triolein (OOO) with purity greater than 99% and the product contained 98.3 wt% ethyl palmitate. It is im-
Sigma Chemical) were dissolved in dichloromethane at dif- portant to use a stoichiometric amount of ethanol to achieve a
ferent concentrations and analyzed by HPLC-ELSD under the high degree of esterification.
previously mentioned conditions. Relationships between log
Lipase-catalyzed synthesis of PPP. Figure 3 illustrates a
peak area and log sample concentration were plotted to obtain time-course plot of the PPP content in the acylglycerol frac-
standard calibration curves for each of the TAG standards.
tion and the extent of esterification catalyzed by Lipase IM 60
and Novozym 435. Novozym 435 showed a higher degree of
esterification and a higher TAG content in the acylglycerols
after 36 h of reaction and was therefore selected for the syn-
thesis of PPP from ethyl palmitate and glycerol. Table 3 shows
time-course variations in the acylglycerol composition (wt%)
of the reaction product and the degree of esterification. At the
early stage of the reaction, the contents of MAG and DAG
were higher than that of TAG. As the reaction proceeded, the
conversion of these partial glycerides led to a gradual increase
in TAG. About 88% conversion with 91 mol% of PPP, along
TABLE 1
Gradient Program of Mobile Phase in the HPLC analysis of TAG
Time (min)
Acetone (%)
Dichloromethane (%)
0
8
2
2
16
28
98
2
98
98
84
72
2
1
2
2
3
3
0
2
8
2
5
98
98
2
JAOCS, Vol. 81, no. 6 (2004)

SYNTHESIS OF STRUCTURED LIPIDS FROM PALM OIL
529
TABLE 2
a
FA Compositions (wt%) of Palm Oil Before and After Various Stages of Solvent Crystallization
Stage #1b
Stage #2c
Stage #3d
FA
Initial
Solid fraction (FFA-I)
0.72 ± 0.09
Liquid fraction (FFA-II)
Liquid fraction (FFA-III)
Liquid fraction (FFA-IV)
14:0
0.43 ± 0.3
39.98 ± 0.1
0.32 ± 0.08
9.69 ± 0.11
0.34 ± 0.02
2.22 ± 0.08
0.49 ± 0.02
2.46 ± 0.03
16:0
87.82 ± 0.21
e
(
85.35%)
16:1
18:0
18:1
0.32 ± 0.5
4.67 ± 0.3
42.49 ± 0.1
NDf
0.46 ± 0.08
0.87 ± 0.16
ND
ND
ND
8.72 ± 0.1
2.03 ± 0.12
0.29 ± 0.03
75.83 ± 0.11
69.52 ± 0.09
96.04 ± 0.12
g
g
g
(98.1%)
(97.41%)
(81.61%)
18:2
12.1 ± 0.3
0.71 ± 0.08
19.14 ± 0.11
21.32 ± 0.08
1.01 ± 0.04
a
Mean absolute deviation of three independent determinations.
b
c
Solvent = acetone, solvent/FFA = 40 mL/g, temperature = −25°C, storage time = 24 h.
Solvent = acetone, solvent/FFA = 20 mL/g, temperature = −25°C, storage time = 24 h.
d
Solvent = acetonitrile, solvent/FFA = 60 mL/g, temperature = −20°C, storage time = 24 h.
e
f
Yield of palmitic acid (16:0) in the solid fraction.
ND, not detected.
g
Yield of oleic acid (18:1) in the liquid fractions.
with a minor amount of partial glycerides, was accomplished
after 36 h of reaction. PPP with a purity greater than 99% (and
a corresponding recovery of 92.5%) was obtained by silica gel
chromatographic purification.
Acidolysis of PPP. Figure 4 shows the effects of various
parameters on the time course of oleic acid incorporation in
the acidolysis of PPP and oleic acid. When the reaction mix-
ture contained 0.1 mmol of PPP, 0.8 mmol of oleic acid, 20
mg of Lipase IM 60 (6.5% based on total substrate weight),
and 3 mL of n-hexane, the reaction product, after 24 h of re-
action at 50°C, contained 97% TAG with an oleic acid con-
tent of 66.1 mol%.
Quantification of TAG by HPLC. The distribution of oleic
acid in TAG positions (sn-1,3 and sn-2) was determined by sil-
FIG. 2. Time-course plot of the formation of ethyl palmitate and the de- ver-ion complexation HPLC-ELSD. The standard lipids PPP,
gree of esterification. Reaction conditions: Lipase PS 30 (Amano En-
zyme Inc., Nagoya, Japan) 0.4 g (10% based on FA weight), molar ratio
were injected separately into the high-performance liquid chro-
of FA to ethyl alcohol = 1:1 (4 g FFA-I), 20 mL of n-hexane, 37°C, stirrer
POP, PPO, OPO, OOP, and OOO at different concentrations
matograph. The standard calibration equations for the six TAG
speed of 150 rpm.
were as follows (X = log mg/L, and Y = log peak area):
2
PPP: Y = 1.02956 + 1.25793 * X (R = 0.99787)
[1]
[2]
2
POP: Y = 0.45319 + 1.26938 * X (R = 0.99759)
TABLE 3
Time-Course Variations in Composition of Products (%) During the
a
Synthesis of Tripalmitin Under Vacuum with Novozym 435
Composition of the acylglycerols (wt%)
Degree of
Time (h) esterification
TAG
DAG
MAG
1
3
6
2
8
4
12.8 ± ±2.2
28.8 ± ±2.0
50.1 ± ±1.6
69.3 ± ±1.6
80.1 ± ±1.4
85.5 ± ±1.7
88.1 ± ±1.6
88.3 ± ±1.1
6.37 ± ±0.5
17.53 ± ±0.7
34.88 ± ±0.6
54.88 ± ±0.5
65.50 ± ±0.7
78.24 ± ±1.1
89.0 ± ±1.1
90.9 ± ±0.8
26.26 ± ±0.7
30.00 ± ±0.5
37.40 ± ±0.5
28.62 ± ±0.5
22.36 ± ±0.5
14.32 ± ±0.3
7.56 ± ±0.3
6.21 ± ±0.2
67.37 ± ±1.0
52.47 ± ±0.8
27.72 ± ±0.5
16.51 ± ±0.6
12.14 ± ±0.2
7.44 ± ±0.3
3.48 ± ±0.2
2.89 ± ±0.1
1
1
2
FIG. 3. Time-course plot of tripalmitin (PPP) formation and the degree 30
of esterification. Reaction conditions: molar ratio of glycerol to ethyl
palmitate = 1:3 (0.58 mmol glycerol), lipase load of 50 mg (10% based
on ethyl palmitate weight), 60°C, reactor pressure of 40 mm Hg, speed
of 300 rpm. (▲,■) Degree of esterification; (▲,■) PPP content in acyl-
glycerols; (▲,▲) Novozym 435; (■,■) Lipase IM 60 (both from Novo
Nordisk, Bagsvaerd, Denmark).
36
aMean absolute deviation of three independent determinations. Reaction
conditions: 50 mg (10 wt% of ethyl palmitate) Novozym 435 (Novo Nordisk
A/S, Bagsvaerd, Denmark), 0.58 mmol glycerol, and 1.76 mmol ethyl palmi-
tate were incubated at 60°C under 40 mm Hg with a stirrer speed of 300 rpm
for 36 h.
JAOCS, Vol. 81, no. 6 (2004)

530
M.-L. CHEN ET AL.
FIG. 4. Effects of various parameters on the incorporation of oleic acid (OA) in the acidolysis of PPP with lipase IM
6
(
8
0. (A) Effect of the molar ratio of OA/PPP. Reaction conditions: 0.1 mmol PPP in 3 mL hexane, 20 mg of lipase
6.5% substrate weight), 50°C, and 150 rpm magnetic stirring. (B) Effect of enzyme load. Molar ratio of OA/PPP =
:1; other conditions are the same as in A except for the enzyme load. (C) Effect of temperature. Molar ratio of
OA/PPP = 8:1; other conditions are the same as in A except for the temperature. (D) Effect of water content. Molar
ratio of OA/PPP = 8:1; other conditions are the same as in A except for the water content. (E) Effect of PPP concen-
tration. Molar ratio of OA/PPP = 8:1; other conditions are same as in A except for the PPP concentration. For en-
zyme supplier, see Figure 2. For other abbreviation, see Figure 3.
2
PPO: Y = 0.6207 + 1.44453 * X (R = 0.99578)
[3] the molecule. The separation of regioisomeric pairs such as
[4] POP, PPO, OPO, and OOP were the most important in quanti-
[5] fying individual structured TAG in the reaction product. Good
[6] separation of TAG was obtained on a ChromSpher 5 Lipids
silver-ion chromatography column using a binary solvent gra-
2
OPO: Y = 0.54614 + 1.33721 * X (R = 0.99948)
OOP: Y = 1.16442 + 1.12684 * X (R = 0.99892)
OOO: Y = 1.14749 + 1.17615 * X (R = 0.99775)
2
2
The separation of TAG isomers was based on the number and dient program of A (acetone) and B (dichloromethane) as elu-
position of the double bonds and the chain length of the FA in ents. Figure 5A shows a chromatogram in which all six TAG
JAOCS, Vol. 81, no. 6 (2004)

SYNTHESIS OF STRUCTURED LIPIDS FROM PALM OIL
531
FIG. 5. HPLC separation of (A) a standard TAG lipid mixture, PPP; POP (1,3-dipalmitoyl-2-
oleoylglycerol); PPO (1,2-dipalmitoyl-3-oleoylglycerol); OPO (1,3-dioleoyl-2-palmitoylglyc-
erol); OOP (1,2-dioleoyl-3-palmitoylglycerol); and OOO (triolein). (B) Structured lipids ob-
tained after the acidolysis of PPP. Reaction conditions: 0.1 mmol of PPP, molar ratio of PPP/OA
=
1:8, 3 mL of hexane, 20 mg of Lipase IM 60 (6.5% based on substrate weight), 50°C, 150
rpm magnetic stirring for 24 h. For enzyme supplier, see Figure 2. For other abbreviation, see
Figure 3.
components from a standard lipid mixture were baseline re- min were attributed to impurities present in the substrate oleic
solved. Linear responses for all six TAG were obtained in the acid (purity 96.7%). The OPO content in TAG produced by
concentration range of 0.12 to 10 µg per injection. Contents of the acidolysis reaction was 74 ± 1.2 mol%, as illustrated in
each individual TAG were quantified using the corresponding Figure 5B.
calibration curves.
When the TAG fraction of the acidolysis product at a reac-
tion time of 24 h was analyzed, only five peaks were ob-
ACKNOWLEDGMENTS
served, corresponding to PPP, PPO, OPO, OOP, and OOO Financial support from the National Science Council of Taiwan
under grant NSC90-2214-E011-004 is gratefully acknowledged.
(
Fig. 5B). At short reaction times (<8 h), in which the incor-
poration of oleic acid in TAG was less than 35%, only PPP,
PPO, and OPO were observed (data not shown). The absence
of oleic acid at the sn-2 position of TAG implied that the R.
miehei lipase exhibited strict sn-1,3 specificity and that there
was no acyl exchange involved at the sn-2 position with a
short reaction time. However, at longer reaction times with a
higher incorporation of oleic acid, small quantities of OOP
and OOO were also detected, as shown in Figure 5B. Pancre-
atic lipase digestion of the TAG fraction of the acidolysis
product showed that the palmitic and oleic acid contents at
the sn-2 position were 90.7 and 9.3 mol%, respectively. The
minor unknown peaks at retention times of 18.23 and 30.89
REFERENCES
1
2
3
4
. Jensen, R.G., and G.L. Jensen, Specialty Lipids for Infant Nutri-
tion. 1. Milks and Formulas, J. Pediatr. Gastroenterol. Nutr.
1
5:232–245 (1992).
. Martin, J.-C., P. Bougnoux, J.-M. Antoine, M. Lanson, and C.
Couet, Triacylglycerol Structure of Human Colostrum and Ma-
ture Milk, Lipids 28:637–643 (1993).
. Jensen, C., N.R. Buist, and T. Wilson, Absorption of Individual
Fatty Acids from Long-Chain or Medium-Chain Triglycerides
in Very Small Infants, Am. J. Clin. Nutr. 43:745–751 (1986).
. Schmid, U., U.T. Bornscheuer, M.M. Soumanou, G.P. McNeil,
and R.D. Schmid, Highly Selective Synthesis of 1,3-Oleoyl-2-
JAOCS, Vol. 81, no. 6 (2004)

532
M.-L. CHEN ET AL.
palmitoylglycerol by Lipase Catalysis, Biotechnol. Bioeng. 12. Hierro, M.T.G., A.I. Najera, and G. Santa-Maria, Analysis of
6
4–66:678–684 (1999).
Triglycerides by Reversed-Phase HPLC with Gradient Elution
Using a Light-Scattering Detector, Rev. Esp. Cienc. Tecnol.
32:635–651 (1992).
5
. Nagao, T., Y. Shimada, A. Sugihara, A. Murata, S. Komemushi,
and Y. Tominaga, Use of Thermostable Fusarium heterosporum
Lipase for Production of Structured Lipid Containing Oleic and 13. Huang, F.C., Y.H. Ju, and C.W. Huang, Enrichment of γ-
Palmitic Acids in Organic Solvent-Free System, J. Am. Oil
Chem. Soc. 78:167–172 (2001).
. Dobson, G., W.W. Christie, and B. Nikolova-Damyanova, Sil- 14. Haagsma, N., C.M. van Gent, J.B. Luten, R.W. de Jong, and E.
ver Ion Chromatography of Lipids and Fatty Acids, J. Chro-
matogr. B 671:197–222 (1995).
. Foglia, T.A., E.J. Conkerton, and P.E. Sonnet, Regioselective 15. Christie, W.W., and M.L. Hunter, Separation of Neutral Lipids
Analysis of Triacylglycerols by Lipase Hydrolysis, J. Am. Oil
Chem. Soc. 72:1275–1279 (1995).
. Becker, C.C., A. Rosenquist, and G. Holmer, Regiospecific
Analysis of Triacylglycerols Using Allyl Magnesium Bromide,
Lipids 28:147–149 (1993).
Linolenic Acid from Borage Oil via Lipase-Catalyzed Reac-
tions, J. Am. Oil Chem. Soc. 74:977–981 (1997).
6
7
8
9
van Doorn, Preparation of an ω-3 Fatty Acid Concentrate from
Cod Liver Oil, Ibid. 59:117–118 (1982).
on Chromarods, J. Chromatogr. 171:517–518 (1979).
16. Akoh, C.C., and C.O Moussata, Lipase-Catalyzed Modification
of Borage Oil: Incorporation of Capric and Eicosapentaenoic
Acids to Form Structured Lipids, J. Am. Oil Chem. Soc.
75:697–701 (1998).
. Bergana, M.M., and T.W. Lee, Structure Determination of 17. Liu, J.W., E.G. Bobik, Jr., and Y.S. Huang, In vitro Hydrolysis
Long-Chain Polyunsaturated Triacylglycerols by High-Resolu-
of Fungal Oils: Distribution of Arachidonic Acid-Containing
Triacylglycerol Molecular Species, Ibid. 75:507–510 (1998).
18. Chen, T.C., and Y.H. Ju, Polyunsaturated Fatty Acid Concen-
trates from Borage and Linseed Oil Fatty Acids, Ibid.
78:485–488 (2001).
1
3
tion C Nuclear Magnetic Resonance, J. Am. Oil Chem. Soc.
3:551–556 (1996).
7
1
1
0. Han, J.J., Y. Iwasaki, and T. Yamane, Monitoring of Lipase-
Catalyzed Transeterification Between Eicosapentaenoic Acid
Ethyl Ester and Tricaprylin by Silver Ion High-Performance 19. Chen, T.C., and Y.H. Ju, Concentration of n-3 PUFA in Saponi-
Liquid Chromatography and High-Temperature Gas Chro-
matography, Ibid. 76:31–39 (1999).
fied Menhaden Oil by Fractional Solvent Crystallization, J.
Chin. Inst. Chem. Engr. 33:143–147 (2002).
1. Nikolova-Damyanova, B., Silver Ion Chromatography and
Lipids, in Advances in Lipid Methodology—One, edited by
W.W. Christie, The Oily Press, Dundee, United Kingdom, 1992,
pp. 181–237.
[Received October 23, 2003; accepted April 12, 2004]
JAOCS, Vol. 81, no. 6 (2004)