Direct separation of regioisomers and enantiomers of monoacylglycerols by tandem column high-performance liquid chromatography

Journal of Chromatography A, 1165 (2007) 93–99

Direct separation of regioisomers and enantiomers of monoacylglycerols by tandem column high-performance liquid chromatography Li Deng a,b , Hideo Nakano a , Yugo Iwasaki a,∗ a

Laboratory of Molecular Biotechnology, Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8610, Japan b Department of Bioengineering, College of Life Sciences and Technology, Beijing University of Chemical Technology, Beijing 100029, China Received 7 May 2007; received in revised form 23 July 2007; accepted 25 July 2007 Available online 2 August 2007

Abstract An HPLC-based method for direct separation of the regioisomers and enantiomers of monoacylglycerols (MAGs), i.e. sn-1-MAG, sn-2-MAG and sn-3-MAG, has been established. The method employs a tandem column system, in which two different columns (a conventional silica gel column and an enantioselective column) are connected in series. Three isomers of monooleoylglycerols (MOGs) and monolinoleoylglycerols (MLGs) were resolved on the system with resolution factor (Rs ) of more than 1.1 between adjacent peaks. In addition, all types of oleoylglycerols, i.e. trioleoylglycerol (TOG), sn-1,2-dioleoylglycerol (DOG), sn-2,3-DOG, sn-1,3-DOG, sn-1-MOG, sn-3-MOG and sn-2-MOG, were successfully separated on the tandem column system, although baseline separation of the enantiomers was not achieved. By means of the established analytical method, the reaction course of Candida antarctica lipase B (CALB)-mediated esterification of glycerol with oleic acid was monitored. It was found that sn-1-MOG and sn-2,3-DOG were preferably generated over sn-3-MOG and sn-1,2-DOG, respectively, in the early stage of the reaction, and the maximal enantiomer excess (%ee) of sn-1-MOG and sn-2,3-DOG were 32 and 53%, respectively, at 2 h. The enantiomeric purities of these chiral acylglycerols decreased after prolonged reaction. The mechanisms for the formation of these chiral acylglycerols are discussed. © 2007 Elsevier B.V. All rights reserved. Keywords: Regioisomeric and enantiomeric separation; Monoacylglycerols; Enantioselective HPLC; Esterification; Lipase selectivity

1. Introduction Monoacylglycerols (MAGs) are widely used as emulsifiers in food, cosmetic and pharmaceutical industries. MAGs consist of three isomers, which are regioisomeric (sn-2) and enantiomeric (sn-1 and -3) MAGs. Pure isomers of MAGs, especially enantiomeric sn-1(3)-MAGs, have potential use as intermediates for organic synthesis of structured lipids, phospholipids, glycolipids and the preparation of enzyme agonists and antagonists. Conjugates of these molecules with various pharmaceuticals are potentially attractive as prodrugs or for their controlled release [1,2].

∗

Corresponding author. Tel.: +81 52 789 4143; fax: +81 52 789 4145. E-mail address: [email protected] (Y. Iwasaki).

0021-9673/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2007.07.073

Regioisomerically and enantiomerically pure MAGs are difficult to synthesize by traditional chemical reactions, and lipase-catalyzed regio- and stereoselective reactions would provide alternative routes. The interest in enzymatic production of MAGs is significantly increasing because enzymatic approach could be performed under milder conditions and produces desired stereo-specific MAG products [3]. Enzymatic production of sn-2-MAGs by ethanolysis of triacylglycerols (TAGs) with an excess of ethanol or hydrolysis of TAGs, and production of racemic sn-1(3)-MAGs by esterification of glycerol using sn1,3-specific lipases are the most common approaches reported [1,4]. In addition, glycerols with specifically protected positions were employed for enzymatic esterification of fatty acids (FAs) followed by removal of the protecting group, producing enantiomerically or regioisomerically pure MAGs [3]. A bottleneck for the establishment of synthetic methods for enantiomerically pure sn-1- or sn-3-MAGs and investigating

94

L. Deng et al. / J. Chromatogr. A 1165 (2007) 93–99

stereoselectivity of lipases is the difficulty in rapid separation and quantification of regioisomeric sn-2-MAGs and enantiomeric sn-1(3)-MAGs. Indeed, it is reported that most 1,3specific lipases discriminate sn-1 position from sn-3 position of TAGs in hydrolysis [5–7]. A direct separation of enantiomeric MAGs has not been achieved by chromatographic means, although it would be useful as a fundamental technology in the stereochemical study of acylglycreols. At present, sn1(3)- and sn-2-MAGs can be separated by boric acid thin-layer chromatography (TLC) [8] or by reversed-phase (RP)-HPLC [9]. Enantioselective columns (Sumichiral OA-4100, OA-4000 and/or YMC-A-KO3) or RP-HPLC are available for separation of the regioisomeric and enantiomeric sn-1-, sn-2- and sn-3-MAGs following the derivatization of MAGs to the corresponding bis-(3,5-dinitrophenylurethanes) [bis-(3,5-DNPUs)] [10–12]. However, these methods, especially for enantiomeric pairs of MAGs, involve laborious multi-step derivatization of the lipid samples including separation of bis-(3,5-DNPUs) on TLC and therefore require a relatively large amount of samples. In addition, the possibility of the spontaneous acyl migration of sn-1(3) to sn-2 position during the derivatization procedure can not be excluded. Hence, a simple rapid analytical method without any derivatization steps is greatly preferable. In our previous works, we demonstrated a direct method for analyzing chiral TAGs by enantioseletive HPLC [13] and enantiomeric sn-1,2-/2,3-diacylglycerols (DAGs) by tandem column HPLC system with a combination of a normal phase silica gel column and an enantioselective column [14]. Chiralcel OD and OF are suitable for the separation of the enantiomeric acylglycerols. The purpose of the present work is to establish an analytical method in which regioisomeric and enantiomeric MAGs are directly separated without any derivatization steps. This is the first report of the direct enantioseparation of MAG enantiomers. The separations of the regioisomeric and/or enantiomeric monostearoylglycerols (MSGs), monopalmitoylglycerols (MPGs), monoloeoylglycerols (MOGs) and monolinoleoylglycerols (MLGs) were demonstrated. Using the methods established, the stereoselectivity of Candida antarctica lipase B (CALB) during esterification of oleic acid and glycerol was investigated.

2. Experimental 2.1. Chemicals and enzymes Trioleoylglycerol (TOG), oleic acid, glycerol, ethanol (99.5%), boric acid, HPLC-grade n-hexane and 2-propanol were purchased from Wako (Osaka, Japan). Trilinoleoylglycerol (TLG), sn-1,2-isopropylidene glycerol and sn-2,3isopropylidene glycerol were from Tokyo Chemical Industries (Tokyo, Japan). Novozym 435 [immobilized C. antarctica lipase B (CALB), non-regio-specific] was a gift from Novozymes Japan. (Chiba, Japan). All other chemicals were of analytical quality or better.

2.2. Preparation of authentic MAG standards sn-1 and sn-3-MAGs were synthesized by esterification reaction of enantiomerically pure isopropylidene glycerols and FAs, followed by hydrolysis for removal of the isopropylidene group [13]. 2,3-O-Isopropylidene-sn-glycerol (396.5 mg, 3 mmol) and 847.4 mg oleic acid (3 mmol) were reacted in the presence of 125.8 mg CALB (10% based on the reactant weight) and 3% water. The mixture was incubated with stirring at 40 ◦ C for 1 h at atmospheric pressure, and then under reduced pressure (933 Pa) for 23 h to remove the condensed water. After the reaction, the lipase was removed by filtration, giving 1-oleoyl-2,3-O-isopropylidene-sn-glycerol in 95% yield. The isopropylidene group of the resultant ester was deprotected with 10 ml of 90% trifluoroacetic acid at −20 ◦ C for 30 min, neutralized with approximately 50 ml of ice-cold 2 M NaOH solution, and then the acylglycerols were extracted with 50 ml of chloroform/methanol (4:1, v/v). The resultant lipid solution was evaporated to remove the solvent under reduced pressure, giving sn-1-MOG in 84% yield. By means of the same reaction approach, sn-3-MOG, sn-1 and 3-MLGs, sn-1 and 3-MSGs, and sn-1 and 3-MPGs were prepared using appropriate protected glycerols and FAs. sn-2-MAGs were synthesized by lipase-catalyzed ethanolysis reaction of homogeneous TAG in excess ethanol [4]. About 200 mg TOG and 800 mg absolute ethanol were reacted in the presence of 100 mg CALB. The mixture was incubated at 25 ◦ C with stirring for 4 h. After the reaction, the lipase was removed by filtration, and then the excess ethanol was evaporated from the reaction mixture at 25 ◦ C. The concentrate was dissolved in 3.8 ml of acetonitrile/water (95:5, v/v) and washed three times with 3.8 ml n-hexane for ethyl ester removal. The solvent was evaporated from the acetonitrile phase, and the concentrate was dissolved in 4 ml chloroform, which was extracted with 4 ml water/ethanol (9:1, v/v) for the extraction of glycerol generated as by-product. The layers were separated, and the water layer was washed with chloroform as above. The two chloroform layers were combined, and the solvent was evaporated, yielding the target sn-2-MOG. Similarly, sn-2-MLG was prepared using TLG. 2.3. Boric acid-impregnated thin-layer chromatography (TLC)/flame-ionization detection (FID) TLC/FID analyzer (Iatroscan MK-5; Iatron Laboratories, Tokyo. Japan) enables the separation of acylglycerols, and FAs. The rods (Chromarod-SIII) were subjected to a blankscan, and then dipped into 3% aqueous boric acid solution for 5 min, and then dried in an oven for 5 min at 120 ◦ C. The rods were blank-scanned again, and the analytes were spotted on the rods. The rods were eluted with benzene/chloroform/acetic acid (70:30:2, v/v/v) for 8 cm, and dried. Then, the rods were developed again with the same solvent for 10 cm, dried, and scanned to detect the lipids. By this analytical method, TAG, sn-1,2(2,3)-DAGs, sn-1,3-DAG, sn-1(3)-MAGs, sn-2MAG and FA are separated from each other. The lipid composition was calculated from the peak areas. The other

L. Deng et al. / J. Chromatogr. A 1165 (2007) 93–99

operation parameters for TLC/FID were described elsewhere [15]. 2.4. Enantioselective HPLC We employed a tandem column HPLC system, in which two analytical columns, a silica gel column (Wakosil 5SIL, 100 mm × 4.6 mm, Wako, Osaka, Japan) and a enantioselective column, Chiralcel OD [cellulose-tris(3,5-dimethylphenylcarbamate)-impregnated silica, 250 mm × 4.6 mm, Daicel Chemical, Tokyo, Japan [16]], were connected in series. The mobile phase was n-hexane/2-propanol (93:7–97:3, v/v) (specified later) at flow rate of 1.0 ml/min at room temperature. A 5 ␮L aliquot of the sample (authentic standards or lipase-catalyzed reaction mixtures containing 10–60 ␮g of sn-1, sn-3 and/or sn-2-MAGs dissolved in n-hexane) was injected using a 20 ␮L injection loop. The peaks were detected using an evaporative light scattering detection (ELSD) system (Model ELSD-LT, Shimadzu Corporation, Kyoto, Japan). The drift tube temperature was 45 ◦ C. Nitrogen was used as evaporation gas at 360 kPa. All separations were conducted at least three times, and the relative standard deviation (RSD) of retention times of each peak was less than 1%. 2.5. Lipase-catalyzed reaction Esterification of oleic acid and glycerol was conducted using CALB as a catalyst (Fig. 1). In a typical esterification reaction, mixtures consisting of oleic acid (1000 mg, 3.54 mmol) and glycerol (326 mg, 3.54 mmol) were incubated at 40 ◦ C with stirring for 10 min to emulsify the reaction mixtures. Then, the reaction was started by addition of 66 mg (5% of the reactants)

95

of the lipase. The reaction was conducted at atmospheric pressure (0.1 MPa) for 1 h, and then at 400 Pa for 23 h. Portion of the reaction mixtures (20–50 ␮l) was withdrawn at intervals, mixed with 500 ␮l of n-hexane. Then, the sample was mixed vigorously and centrifuged at 15000 rpm for 5 min. Finally, the organic phase was collected as the sample for TLC/FID and HPLC analysis. The lipid composition in the resultant solution such as TOG, sn-1,3-DOG, sn-1,2(2,3)-DOGs, sn-2-MOG, sn-1(3)MOGs and FA were analyzed with TLC/FID analyzer as described in the Section 2.3. The amount of each enantiomeric MOGs (sn-1-MOG and sn-3-MOG) and enantiomeric DOGs (sn-1,2-DOG and sn-2,3-DOG) were calculated by means of the TLC/FID results and the enantioselective HPLC analysis to be described below. Duplicate measurements were made and the average was taken for all the analysis results, with RSD of less than 1%. For investigating enantiomeric purities of sn-1(3)-MOGs and sn-1,2(2,3)-DOGs, 5 ␮l of the sample was analyzed by the tandem column HPLC system. The weight-based amount (in ␮g) of each enantiomer was calculated from the corresponding peak area using standard curves drawn with known amounts of sn-1(3)-MOGs and sn-1,2(2,3)-DOGs. The enantiomer excess (%ee) values of sn-1-MOG and sn-2,3-DOG were calculated using the weight-based amount of each isomer as follows: %ee(sn-1-MOG) =

[sn-1-MOG] − [sn-3-MOG] × 100% [sn-1-MOG] + [sn-3-MOG]

%ee (sn-2, 3-DOG)=

[sn-2, 3-DOG]−[sn-1, 2-DOG] × 100% [sn-2, 3-DOG]+[sn-1, 2-DOG]

Fig. 1. Lipase-catalyzed esterification between glycerol and oleic acid. Glycerol is reacted with oleic acid (RCOOH) by lipase, giving initially sn-1-MOG or sn-3MOG, which are further converted to sn-1,2-DOG, sn-1,3-DOG and sn-2,3-DOG, which possibly are further converted to TOG depending on the lipase’s selectivity and the reaction conditions. The carbons in the glycerol moiety are numbered stereospecifically. Asterisks indicate the chiral centers at the sn-2-position.

96

L. Deng et al. / J. Chromatogr. A 1165 (2007) 93–99

Table 1 Retention time of enantiomeric MAGs on Chiralcel ODa MAGs

MSG MPG MOG MLG a b

ECNb

Retention time (min)

18 16 16 14

sn-1-MAGs

sn-3-MAGs

19.8 20.4 20.5 22.4

22.2 22.7 22.7 24.7

Mobile phase was n-hexane:2-propanol = (96:4, v/v). Equivalent carbon number.

3. Results and discussion 3.1. Separation of the authentic standards on Chiralcel OD column Fig. 2 shows the separation of enantiomeric pairs of MAGs on Chiralcel OD column. The authentic enantiomeric pairs were eluted with 96:4 (v/v) of n-hexane/2-propanol as mobile phase with a resolution factor (Rs ) of more than 1.10 and a separation factor (α) of more than 1.12 for all the four molecular species of MAGs. The peaks were identified by injecting each of the authentic enantiomers separately (data not shown). The peak elution order of molecular species with the same configuration followed the equivalent carbon number (ECN) rule as shown in Table 1. The bigger the ECN value was, the shorter the retention time became. The MPGs and MOGs were eluted at almost the same retention time because both have the same ECN. The optimization of mobile phase composition was conducted with the separation of enantiomeric MOG pair as shown in Table 2. Because the polymeric chiral selector (cellulose derivatives) of Chiralcel OD is physically coated onto silica support, solvents should be carefully used and limited in nhexane/2-propanol and n-hexane/ethanol by reference to the column instruction manual. In the present study, we tested several mixtures of n-hexane and 2-propanol with different composition as the mobile phase. Table 2 displays the effect of mobile phase composition on separation of enantiomeric MOGs on Chiralcel OD. The α and Rs decreased as the solvent polarity increased (increasing 2-propanol content from 3 to 6%). Taking into account the separation efficiency and the total analysis time, the mobile phase of n-hexane to 2-propanol at the ratio of 96:4 was taken for the optimum to be used in the subsequent experiments. Table 2 Effect of mobile phase on separation of enantiomeric MOG on Chiralcel OD Mobile phase composition (n-hexane/2-propanol)

97:3 96:4 95:5 94:6 a b

Retention time (min) sn-1-MOG

sn-3-MOG

26.1 20.5 14.4 11.9

28.4 22.7 15.6 12.8

Separation factor between sn-1-MOG and sn-3-MOG. Resolution factor between sn-1-MOG and sn-3-MOG.

αa

1.13 1.12 1.11 1.11

Rs b

1.19 1.17 1.1 1.0

Fig. 2. Enantioseparation of several molecular species of MAGs on a Chiralcel OD column. The mobile phase was n-hexane/2-propanol (96:4, v/v). (A) sn-1and sn-3-MSG, (B) sn-1- and sn-3-MPG, (C) sn-1- and sn-3-MOG and (D) sn-1and sn-3-MLG.

L. Deng et al. / J. Chromatogr. A 1165 (2007) 93–99

97

Fig. 3. Separation of regioisomers and enantiomers of MAGs on the tandem column HPLC system. Chromatograms of authentic isomers of MOGs (A and C) and MLG (B and D) in isocratic elution (A and C), and in stepwise elution (C and D) are shown. The isocratic elution was with n-hexane/2-propanol (96:4, v/v). The stepwise elution was with n-hexane/2-propanol of 96:4 (v/v) (0–10 min) then 94:6 (v/v) (10 to 40 min).

Unfortunately, use of only Chiralcel OD was not enough for separating three isomers of MAGs, because the peaks of sn-2-MAGs completely overlapped with those of sn-3-MAGs [retention time of sn-2-MOG and sn-2-MLG are 22.7 and 24.7 min, respectively with n-hexane/2-propanol (96:4 v/v)]. Therefore, we employed the tandem column system, in which a conventional silica gel column was connected upstream of Chiralcel OD so that sn-2-MAGs were separated from sn-1(3)MAGs before they entered into Chiralcel OD. Fig. 3A and B show the separation of the three authentic regioisomers and enantiomers of MOG and MLG on the tandem column system with n-hexane/2-propanol (96:4, v/v). The three isomers were resolved within 60 min. In both cases, sn-1(3)-MAGs eluted faster than sn-2-MAGs. Applying a stepwise elution resulted in a more satisfactory separation: all isomers were resolved within 40 min, as shown in Fig. 3C and D. Our previous work [14] demonstrated a direct separation of regioisomers and enantiomers of DAGs on the same tandem column system with n-hexane/2-propanol (300:7, v/v). Therefore, an attempt was made to resolve a mixture of all types of oleoylglycerols, (i.e. TOG, sn-1,3-DOG, sn-2,3-DOG, sn-1,2-DOG, sn-1-MOG, sn-3-MOG, and sn-2-MOG) on the tandem column system. Stepwise elution enabled separation of the all the seven oleoylglycerols as shown in Fig. 4. The elution order depends on their polarity, and the retention times are 5.3 min for TOG, 14.2, 16.9 and 18.2 min, respectively, for sn-1,3-DOG, sn-2,3-DOG and sn-1,2-DOG, and 59.5, 61.2 and 70.2 min, respectively, for sn-1-MOG, sn-3-MOG and sn-2-MOG. To draw calibration lines for sn-1-MOG and sn-3-MOG, different concentrations of sn-1- or sn-3-MOG (2–10 ␮g/ml) were injected onto the tandem column system with 5 ␮l injection. The

Fig. 4. Separation of all types of oleoylglycerols on the tandem column HPLC system. Stepwise elution was with n-hexane/2-propanol of 98:2 (v/v, 0 to 13 min), 97:3 (v/v, 13 to 19 min), 96:4 (v/v, 19 to 53 min), 96:5 (v/v, 53 to 63 min), and 94:6 (v/v, 63 to 73 min).

98

L. Deng et al. / J. Chromatogr. A 1165 (2007) 93–99

double logarithmic plots of the peak areas and the weights of the analytes gave excellent linear fits as follows: log(peak area) = 1.6443 × log(weight) + 4.4712, with R2 of 0.9968 for sn-1-MOG,

and

log(peak area) = 1.6486 × log(weight) + 4.4297, with R2 of 0.9968 for sn-3-MOG. A drawback of the established analytical method is that the enatiomers are not baseline-separated. This drawback will be improved by lowering the polarity of the mobile phase at the sacrifice of the analytical time. 3.2. Monitoring of lipase-catalyzed reaction In the lipase-catalyzed esterification of oleic acid and glycerol, the sn-1(3)-MOGs are initially synthesized as intermediates, which are further converted to sn-1,3- and/or sn-1,2(2,3)-DOGs. The DOGs are possibly further acylated to TOG as shown in Fig. 1. Some lipases are known to have stereoselectivity toward the sn-1 or sn-3 position in hydrolysis reaction [5], but no information was available in esterification reaction and for MAGs synthesis. Investigation of stereoselectivity of some lipases in esterification is one of our main objectives in this analysis. The reaction course was monitored using TLC/FID and the tandem column HPLC established here. The TLC/FID was for estimating the lipid composition, whereas the HPLC was for estimating the ratios of sn-1-MOG vs. sn3-MOG and sn-1,2-DOG vs. sn-2,3-DOG. The amounts of sn-1-MOG and sn-3-MOG were calculated using the standard line obtained in the Section 3.1, while those of sn-1,2-DOG and sn-2,3-DOG were calculated using the line reported previously [14]. None of these analytical methods require any derivatization steps, allowing rapid analysis with minimal amounts of samples. Table 3 shows the time course changes in lipid composition and in enantiomer excess of sn-1-MOG and sn-2,3-DOG in the CALB-mediated reaction. 80% of oleic acid was consumed rapidly in 7 h and almost exhausted at 23 h. Accordingly, sn-1-MOG, sn-3-MOG and sn-1,3-DOG were rapidly accumulated to 20, 14 and 34% at 7 h, respectively, and amounted to

25, 21 and 31% at 30 h, respectively. In contrast, the accumulation of sn-1,2-DOG and sn-2,3-DOG was slower; (3.1 and 5.4%, respectively, at 7 h and 7.4 and 7.5% at 30 h). The amounts of sn-2-MOG and TOG in the reaction mixture were smaller than 3.1 and 4.5% throughout, respectively. The above results indicate that the CALB has regiopreference towards sn1 and/or sn-3 positions over sn-2 position of glycerol under the conditions tested. The %ee of the sn-1-MOG was higher than 30% in the early stage of the reaction (2–5 h), and then gradually decreased to 8.7% as the reaction proceeded. On the other hand, the %ee of the sn-2,3-DOG was 53% at 2 h, and gradually decreased to 27% at 7 h then to almost zero at 23 h. The enantiomeric purity of the sn-1(3)-MOGs may be affected by two different types of “selectivities” of the lipase. One is stereoselectivity towards either sn-1 or sn-3 position of the prochiral glycerol during its acylation, and the other is enantioselectivity towards either sn-1-MOG or sn-3-MOG during its further acylation into DOGs. Similarly, the enantiomeric purity of sn-1,2(2,3)-DOG may be also affected by two kinds of enantioselectivities, which are (i) enantioselectivity against sn-1-MOG or sn-3-MOG during the acylation of them to sn-1,2-DOG or sn-2,3-DOG, respectively, and, (ii) enantioselectivity against sn-1,2-DOG or sn-2,3-DOG during the further (complete) acylation of them into TOG. The amount of sn-1-MOG was larger than that of sn-3-MOG throughout the reaction (Table 3). This may be explained by (i) faster formation of sn-1-MOG than sn-3-MOG (by stereopreference towards sn-1 position over sn-3 position), or (ii) faster consumption of sn-3-MOG than sn-1-MOG (i.e. sn-3-MOG being a better substrate than sn-1-MOG). As for the enantiomeric purity of sn-2,3-DOG (Table 3), the amount of sn-2,3-DOG was larger than sn-1,2-DOG in the initial stage of the reaction, suggesting again that sn-3-MOG may be a better substrate than sn-1-MOG. The %ee of sn2,3-DOG decreased to almost zero after prolonged reaction. Since the formation of TOG (the completely acylated product) was very low, the decrease in the %ee of sn-2,3-DOG was not because of the faster acylation (consumption) of sn-2,3-DOG, but because of the formation of sn-1,2-DOG. In the later stage, the accumulation rate of sn-1-MOG was slower than that of

Table 3 Lipid composition and enantiomer excess (%ee) in CALB-catalyzed esterification between oleic acid and glycerol Time (h)

Lipid composition (mol%) FA

0 1 2 5 7 23 30 a b c

100 93 82 42 20 0.78 0.84

sn-1-MOG

sn-3-MOG

0 2.1a 3.1 16 20 26 25

1.6 8.6 14 21 21

sn-2-MOG 0 0 0 0.59 1.1 3.1 2.6

sn-1,2-DOG

0.17 1.6 3.1 7.5 7.4

sn-2,3-DOG

0 0.3b 0.56 3.3 5.4 7.7 7.5

The sum of sn-1-MOG and sn-3-MOG analyzed by TLC/FID is shown. The sum of sn-1,2-DOG and sn-2,3-DOG analyzed by TLC/FID is shown. Not determined.

sn-1,3-DOG

TOG

0 4.4 11 26 34 31 31

0 0 1.8 1.7 1.9 3.6 4.5

%ee (sn-1-MOG) (%)

%ee (sn-2,3-DOG) (%)

– NDc 32 30 18 11 8.7

– NDc 53 35 27 1.3 0.67

L. Deng et al. / J. Chromatogr. A 1165 (2007) 93–99

sn-3-MOG. This may be because the further acylation of sn-1MOG was accelerated due to its high content in the later stage. The acylation of sn-1-MOG increased the sn-1,2-DOG content, resulting in the decrease in the %ee of sn-1-MOG and sn-2, 3-DOG. 4. Conclusions The present study provides a simple method for separation of regioisomeric and enantiomeric sn-1, sn-2 and sn-3-MAGs. The method will be useful for evaluation of selectivity of lipases and optimization of lipase-catalyzed lipids reactions. In addition, it may be applied for preparative purposes as it is a non-destructive method.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]

Acknowledgments [14]

L.D. is grateful to China Scholarship Council and Ministry of Education, Culture, Sports, Science and Technology (MEXT, Japan) for financial support.

99

[15] [16]

M. Berger, M.P. Schneider, J. Am. Oil Chem. Soc. 69 (1992) 961. J.N. Jacob, G.W. Hesse, V.E. Shashoua, J. Med. Chem. 33 (1990) 733. U.T. Bornscheuer, Enzyme Microb. Technol. 17 (1995) 578. R. Irimescu, Y. Iwasaki, C.T. Hou, J. Am. Oil Chem. Soc. 79 (2002) 879. E. Rogalska, C. Cudrey, F. Ferrato, R. Berger, Chirality 5 (1993) 24. E. Rogalska, S. Nury, I. Douchet, R. Verger, Biochem. Soc. Trans. 25 (1997) 161. U.T. Bornscheuer, R.J. Kazlauskas (Eds.), Hydrolases in Organic Synthesis, Wiley–VCH, Weinheim, 1999, p. 164. A.E. Thomas, J.E. Sharoun, H. Ralston, J. Am. Oil Chem. Soc. 42 (1965) 789. L.Y. Yang, A. Kuksis, J.J. Myher, G. Steiner, J. Lipid Res. 36 (1995) 125. T. Takagi, Y. Ando, Lipids 26 (1991) 542. A. Kuksis, Y. Itabashi, Methods 36 (2005) 172. Y. Ando, M. Satake, Y. Takahashi, Lipids 35 (2000) 579. Y. Iwasaki, M. Yasui, T. Ishikawa, R. Irimescu, K. Hata, T. Yamane, J. Chromatogr. A 905 (2001) 111. W. Piyatheerawong, Y. Iwasaki, T. Yamane, J. Chromatogr. A 1068 (2005) 243. W. Piyatheerawong, Y. Iwasaki, X. Xu, T. Yamane, J. Mol. Catal. B: Enzym. 28 (2004) 19. Y. Okamoto, M. Kawashima, K. Hatada, J. Chromatogr. 363 (1986) 173.