Graduate School of Agricultural Science, Tohoku University (1-1, Tsutsumi-dori, Amamiya-machi, Aoba-ku, Sendai , JAPAN)

Transcription

1 Journal of Oleo Science Copyright 2015 by Japan Oil Chemists Society doi : /jos.ess15040 Abuances of Triacylglycerol Positional Isomers a Enantiomers Comprised of a Dipalmitoylglycerol Backbone a Short- or Medium-chain Fatty Acids in Bovine Milk Fat Toshiharu Nagai 1, Natsuko Watanabe 2, Kazuaki Yoshinaga 1, Hoyo Mizobe 1, Koichi Kojima 1, Ikuma Kuroda 3, Yuki Odanaka 4, Tadao Saito 5, Fumiaki Beppu 2 a Naohiro Gotoh 2* 1 Tsukishima Foods Iustry Co. Ltd. (3-17-9, Higashi Kasai, Edogawa-ku, Tokyo , JAPAN) 2 Graduate School of Marine Science a Technology, Tokyo University of Marine Science a Technology (4-5-7, Konan, Minato-ku, Tokyo , JAPAN) 3 GL Sciences Inc. (6-22-1, Nishi Shinjuku, Shinjuku-ku, Tokyo , JAPAN) 4 School of Pharmacy, Showa University (1-5-8 Hatanodai, Shinagawa-ku, Tokyo, , JAPAN) 5 Graduate School of Agricultural Science, Tohoku University (1-1, Tsutsumi-dori, Amamiya-machi, Aoba-ku, Seai , JAPAN) Abstract: Bovine milk fat (BMF) is composed of triacylglycerols (TAG) rich in palmitic acid (P), oleic acid (O), a short-chain or medium-chain fatty acids (SCFAs or MCFAs). The composition a biing positions of the fatty acids on the glycerol backbone determine their physical a nutritional properties. SCFAs a MCFAs are known to characteristically bi to the sn-3 position of the TAGs in BMF; however, there are very few non-destructive analyses of TAG enantiomers biing the fatty acids at this position. We previously reported a method to resolve the enantiomers of TAGs, biing both long-chain saturated fatty acid a unsaturated fatty acid at the sn-1 a 3 positions, in palm oil, fish oil, a marine mammal oil using chiral HPLC. Here, we further developed a method to resolve several TAG enantiomers containing a dipalmitoyl (PP) glycerol backbone a one SCFA (or MCFA) in BMF. We revealed that the predominant TAG structure in BMF was homochiral, such as 1,2-dipalmitoyl-3-butyroyl-sn-glycerol. This is the first quantitative determination of many TAG enantiomers, which bi to a SCFA or MCFA, in BMF was evaluated simultaneously. Furthermore, the results iicated that the amount ratios of the positional isomers a enantiomers of TAGs consisting of a dipalmitoyl (PP) glycerol backbone a SCFA (or MCFA), resembled the whole TAG structures containing the other diacylglycerol backbones consisting of P, O, myristic acid, a/or stearic acid in BMF. Key words: chiral separation, enantiomers, milk fat, triacylglycerol 1 INTRODUCTION Triacylglycerols TAGs, which are esters composed of three fatty acids a one glycerol molecule, are an important energy source 1. The structure of TAGs a the biing positions of the fatty acids on the glycerol backbone are shown in Fig. 1 using a Fischer projection 2. To better uersta the structure of edible oils a fats, it is important to examine both the fatty acid composition a their biing positions on the glycerol backbone. Brockerhoff developed a stereospecific analytical method using phospholipase A 2 that enabled the analysis of the fatty acid composition at the sn-1, 2, a 3 positions 3 5. Itabashi et al. developed another stereospecific method using chiral HPLC 6, 7. These analytical techniques have been used in many studies on the positional distributions of fatty acids in TAGs from various organisms in nature Bovine milk contains about 3.5 to 5 total lipid, a 98 or more of the lipid is TAG 11. Bovine milk fat BMF is rich in palmitic acid P a oleic acid O. Long-chain saturated fatty acids, such as myristic acid M, P, a stearic * Correspoence to: Naohiro Gotoh, Graduate School of Marine Science a Technology, Tokyo University of Marine Science a Technology, 4-5-7, Konan, Minato-ku, Tokyo , JAPAN ngotoh@kaiyodai.ac.jp Accepted May 28, 2015 (received for review February 17, 2015) Journal of Oleo Science ISSN print / ISSN online

2 T. Nagai, N. Watanabe a K. Yoshinaga et al. Fig. 1 Structures of the triacylglycerol (TAGs) staards. A) Symmetric TAGs, B) asymmetric TAGs (single enantiomer), C) asymmetric TAG (racemate) comprised of two palmitic acids (Ps) a one short-chain or medium-chain fatty acid (SCFA or MCFA). acid S, are mainly distributed at the sn-1 a 2 positions 12. In addition, the milk fat of ruminants such as BMF distinctively contains short- a medium-chain fatty acids SCFAs a MCFAs, including butyric acid C 4, caproic acid C 6, a caprylic acid C 8 ; these fatty acids are esterified predominantly at the sn-3 position 10, 12. Similar tres are observed in rat milk fat RMF a human milk fat HMF 13, 14. Although RMF a HMF do not contain C 4 C 6 a C 4 C 8, respectively, the C 8 in RMF, a the capric acid C 10 a lauric acid C 12 in HMF, which are shorter fatty acids in their respective milk fats, are mainly bou to the sn-3 positions. These SCFAs a MCFAs at the sn-3 position are selectively hydrolyzed by lingual lipase or gastric lipase in the stomach, a then absorbed into gastric mucosa 15, 16. The energy obtained by the action of these stereoselective lipases is essential for newborns, who require an efficient energy supply from milk fat. The positional distribution of fatty acids in BMF 12 iicates that the TAGs of BMF are likely to be homochiral. However, studies using the positional distribution of fatty acids do not show concrete TAG molecular species with intact structures, where TAG molecular species only refers to the combination of three fatty acids on the glycerol backbone without consideration of their biing positions. The physical a nutritional properties of oils a fats depe upon the TAG structure 17, which consists of both the constituent fatty acids a their biing positions. To investigate the structure of TAGs, their positional isomers a enantiomers should be analyzed. Herein, a TAG positional isomer refers to TAGs where the composition of the fatty acids is identical but only the fatty acid at the β-position sn-2 position is fixed, e.g., β-pop a β-ppo are positional isomers. The three letters refer to the combination of fatty acids on the glycerol backbone a β is a prefix that signifies that the fatty acid at the β-position is fixed 2. In β-ppo, whether O is bou at the sn-1 or 3 position is unclear. However, an asymmetric TAG could be comprised of enantiomers because the carbon atom at the sn-2 position forms a stereogenic center. When enantiomers are considered, β-ppo is recognized as a mixture of two enantiomers, sn-ppo a sn-opp. A TAG enantiomer designated as sn-ppo is a TAG biing two Ps a one 944

3 Triacylglycerol stereoisomers in bovine milk fat O at the sn-1, 2, a 3 positions, in that order. A mixture of an equal amount of sn-ppo a sn-opp is designated as rac-ppo or rac-opp, because it is racemic. Recently, we developed two separation methods for TAG positional isomers a enantiomers. We showed that pairs of TAG positional isomers, which consisted of two longchain saturated fatty acids a one other fatty acid long chain unsaturated fatty acid, SCFA, or MCFA, such as β-pop/β-ppo, β-pc 4 P/β-PPC 4, or β-pc 10 P/β-PPC 10, were resolved using reversed-phase HPLC on an octacosyl C28 column We also showed that racemic mixtures of asymmetric TAGs biing both long-chain saturated fatty acid a unsaturated fatty acid at the sn-1 a 3 positions, such as rac-ppo, rac-ppl, a rac-poo, were resolved using chiral HPLC on a polysaccharide-based chiral column 21. In the separation of TAG positional isomers using a C28 column, we showed that both the difference between saturated a unsaturated fatty acids, a the difference in the length of the acyl chains, contributed to the resolution of the TAG positional isomers. However, there is no data on the chiral separation of asymmetric TAGs consisting of only saturated fatty acids, where the acyl chain lengths are different in the sn-1 a 3 positions, such as rac-ppc 4. Here, we examined whether our chiral HPLC method could separate racemic mixtures of asymmetric TAGs that consisted of two palmitic acids a one SCFA or MCFA in BMF. We then applied our two separation methods to the quantification of BMF TAG positional isomers a enantiomers consisting of two palmitic acids a one of C 4, C 6, C 8, C 10, or C 12, because P is a major fatty acid in BMF 12, a the structure of whole TAGs with SCFA or MCFA in BMF is presumably represented by the profile of TAG isomers comprised of dipalmitoyl glycerol backbone a SCFA or MCFA. 2 EXPERIMENTAL PROCEDURES 2.1 Chemicals a Materials The TAG isomers shown in Fig. 1 were in-house products Tsukishima Foods Iustry Co., Ltd., Tokyo, Japan. All of the other reagents were purchased from Wako Pure Chemical Iustries, Ltd. Osaka, Japan. The butter was purchased at a supermarket in Tokyo, Japan. 2.2 Lipid Extraction from Butter Butter 1 g was diluted in 9 ml of 1 saline solution. The diluted butter solution was then added to 20 ml of a chloroform a methanol 2/1, v/v mixture in a test tube equipped with a screw cap, which was mixed vigorously using a vortex mixer, a centrifuged at 1,500 g for 10 min. The aqueous phase was extracted two more times using the same method. The combined bottom layers were collected a were concentrated using a rotary evaporator uer vacuum. The BMF thus obtained was stored in an argon-purged screw-capped vial at 30 until analysis. 2.3 Analysis of extracted ion chromatograms correspoing to diacylglycerol ions in BMF BMF 8 mg was dissolved in 1 ml of acetone. A 20 µl aliquot of the sample solution was then injected into an HPLC Alliance e2695, Waters Corporation, Milford, MA with a C28 reversed phase column Sunrise C28, 4.6 mm i.d. 250 mm, 5 μm, ChromaNik Technologies Inc., Osaka, Japan with a mass spectrometer MS detector using an atmospheric pressure chemical ionization APCI probe Quattro micro API, Waters Corporation. The column temperature, a flow rate were 35 a 1.0 ml/min, respectively. The composition of the mobile phase was varied during analysis using a linear gradient elution mode. The initial composition of the mobile phase was acetone/acetonitrile 50/50, v/v ; the ratio of acetone was then linearly increased to acetone/acetonitrile 100/0, v/v over a period of 30 min, a then held constant until all components were eluted. The corona current, cone voltage, source block temperature, desolvation temperature, desolvation gas flow rate, a cone gas flow rate of the APCI ion source were 3.0 μa, 20 V, 120, 450, 200 L/h, a 50 L/h, respectively. A total ion current chromatogram TICC was obtained using a full scan mode over the range of m/z The extracted ion chromatograms EICs of TAG RCOO, which correspoed to diacylglycerol DAG ions produced from the in-source collision-iuced dissociation, were obtained by data processing of the TICC. 2.4 Quantification of TAG positional isomers Five pairs of TAG positional isomer staards, sn-pc n P/ rac-ppc n where n 4, 6, 8, 10, a 12 were mixed together a dissolved in 2-propanol. The concentration of each TAG isomer was adjusted to either 250, 200, 150, 100, 50, or 5 μg/ml a 100 μg/ml of triuecanoin was added as an internal staard IS to obtain calibration curves. A sample solution of 50 mg of the BMF obtained in Section 2.2 a 1.0 mg of triuecanoin as the IS were weighed into a 10 ml volumetric flask, which was filled with 2-propanol. For a recovery test, 0.50 mg 50 μg/ml of each of the TAG isomer staards, i.e., sn-pc n P a rac-ppc n where n 4, 6, 8, 10, a 12, were added to the sample solution. The ammonium adduct ion of each TAG a the dipalmitoyl glycerol ion were selected as the precursor a product ions, respectively, for the selected reaction monitoring SRM channels. The SRM transitions for PPC 4, PPC 6, PPC 8, PPC 10, a PPC 12 were m/z , , , , a , respectively Table 1. The cone voltage a collision energy were 35 V a 20 ev, respectively, for both PPC 4 a PPC 6 ; 40 V a 20 ev, respectively, for PPC 8 ; a 40 V a 25 ev, respectively, for 945

4 T. Nagai, N. Watanabe a K. Yoshinaga et al. Table 1 Calibration curves, multiple correlation coefficients (r), a recoveries of TAG isomers obtained by SRM chromatograms for each TAG positional isomer. TAG SRM transition Equation r Recovery (%) β-pc 4 P y = x β-ppc y = x β-pc 6 P y = x β-ppc y = x β-pc 8 P y = x β-ppc y = x β-pc 10 P y = x β-ppc y = x β-pc 12 P y = x β-ppc y = x y: area ratio (analyte/is), x: concentration (μg/ml) Limited of detection: 1.7 μg/ml (S/N>3), Limited of quantification: 5.6 μg/ml (S/ N>10) both PPC 10 a PPC 12. The SRM transition for the triuecanoin IS was m/z The cone voltage a collision energy for triuecanoin were 30 V a 25 ev, respectively. The setup of LC-MS/MS a the C28 column were identical to that in Section 2.3. The column temperature, flow rate, a injection volume were 15, 1.0 ml/min, a 20 μl, respectively. The composition of the mobile phase was varied during analysis using a linear gradient elution mode. The initial composition of the mobile phase was acetone/acetonitrile 80/20, v/v ; the ratio of acetone was then linearly increased up to acetone/acetonitrile 100/0, v/ v over 40 min, a then held constant until all components were eluted. The operating coitions of the APCI ion source were identical to those listed in Section 2.3. The calibration curves were obtained using the peak areas of the SRM chromatograms, a the sample solutions were analyzed in the same way. The abuance ratios a recoveries of each of the TAG positional isomers in the sample solutions were calculated using equations 1, 2, a 3, a each analysis was performed in triplicate. Here, the staards of the TAG positional isomers were expressed as β-pc n P/β-PPC n where n 4, 6, 8, 10, a 12 instead of sn-pc n P/rac-PPC n where n 4, 6, 8, 10, a 12 to identify the TAG positional isomers in the staards a BMF in the same way. Conc. analyte μg/ml Area analyte /Area IS /Slope 1 Abuance ratio wt Conc. analyte /Conc. BMF Recovery Conc. spiked Conc. not spiked /Conc. calculated Fractionation of β-ppc n where n 4, 6, 8, 10, a 12 from BMF Collection of the asymmetric TAG positional isomers from the BMF was performed using the method described in Section 2.4. The concentration of the BMF was adjusted to be approximately 8 mg/ml using acetone, a 20 μl of the sample was injected into the C28 column. Each β-ppc n peak where n 4, 6, 8, 10, a 12 was detected by the SRM chromatograms using their retention times, a eight fractions of each were collected. The identical fractions were combined a concentrated uer vacuum. 2.6 Chiral separation of staards of TAG enantiomers using a recycle HPLC system with a chiral column The asymmetric TAG staards, rac-ppc 4, rac-ppc 6, rac-ppc 8, rac-ppc 10, a rac-ppc 12, were dissolved in 2-propanol a their concentrations were adjusted to 1000 μg/ml. 20 μl of the sample solution was injected into a recycle HPLC system, which consisted of a recycle pump PU712R, GL Sciences Inc., Tokyo, Japan, an autosampler GL-7420, GL Sciences Inc., a column oven CO705C, GL Sciences Inc., a UV-VIS detector UV702, GL Sciences Inc., a two automatic valves VALVE UNIT 401, FLOM Co., Ltd., Tokyo, Japan, which allowed the analytes to pass through the same column repeatedly a reduced the extra-column volume of the autosampler during additional passes, a a chiral column with a guard cartridge CHI- RALCEL OD-3R, 4.6 mm i.d. 150 mm, 3 μm, a 4.0 mm i.d. 10 mm, 3 μm, respectively, Daicel Corporation, Tokyo, Japan. The analytes were detected by MS using an APCI probe Quattro micro API, Waters Corporation. The column temperature, flow rate, a the mobile phase were 25, 0.5 ml/min, a methanol, respectively. The recycle 946

5 Triacylglycerol stereoisomers in bovine milk fat HPLC system a MS system were controlled using their operational software, EZChrom Elite, Agilent Technologies, Inc., Santa Clara, CA a MassLynx Ver. 4.1, Waters Corporation, respectively. The operating coitions of the APCI source were identical to those listed in Section 2.3. A protonated molecule of each TAG a the dipalmitoyl glycerol ion were selected as the precursor a product ions, respectively, for the selected reaction monitoring SRM channels. The SRM transitions for PPC 4, PPC 6, PPC 8, PPC 10, a PPC 12 were m/z , , , , a , respectively. The cone voltage a collision energy were 15 V a 10 ev, respectively. After each racemate was resolved into its iividual enantiomers by passing through the column five times, their peaks were detected using MS. The elution order of the enantiomers was confirmed by comparing the retention times of a racemate with one of its enantiomers e.g., rac- PPC 4 a sn-ppc 4 data not shown. 2.7 Enantiomeric ratios of asymmetric TAGs in BMF The asymmetric TAG fraction obtained in Section 2.5, which contained β-ppc 4, β-ppc 6, β-ppc 8, β-ppc 10, a β-ppc 12, was dissolved in 2-propanol a its concentration was adjusted to approximately 1 mg/ml. The sample solution was analyzed using the setup a analytical coitions of the LC-MS/MS system that were described Section 2.6. The enantiomeric ratios were determined using the peak area ratios of the SRM chromatograms. The abuance ratios of each TAG enantiomer of β-ppc 4, β-ppc 6, β-ppc 8, β-ppc 10, a β-ppc 12 in the BMF were calculated using equations 2 a 4. sn-ppc n β-ppc n Area sn-ppcn / Area sn-ppcn Area sn-cnpp 4 3 RESULTS AND DISCUSSION 3.1 Analysis of extracted ion chromatograms correspoing to diacylglycerol ions in BMF We examined the TAG molecular species in BMF using a C28 column at 40. Uer these coitions, TAGs are eluted in asceing order of Partition number PN 22, where PN carbon number of three acyl chains 2 number of double bos, a the TAG positional isomers were not separated. The mass spectrum of each peak in the total ion current chromatogram TICC was difficult to analyze because several TAG peaks overlapped Fig. 2F ; however, the use of extracted ion chromatograms EICs facilitated the estimation of the fatty acid composition of the TAGs. Figure 2A E show the EICs of m/z 495, 523, 577, 551, a, 579, which were DAG ions i.e., TAG RCOO, generated by the in-source collision-iuced dissociation CID. Because BMF contains greater than 10 of M, P, a O 12, the m/z 495, 523, 577, 551, a 579 peaks most likely correspoed to MM, MP, PO, PP, a PS, respectively. It was reported that BMF TAGs consist of two long-chain fatty acids LCFAs at the sn-1 a sn-2 positions, a one SCFA or MCFA at the sn-3 position 10, 12, 23. The patterns of these chromatograms were similar to each other, which meant that the SCFA a MCFA might be bou to the glycerol backbones with two LCFAs in this way, though the biing positions of the fatty acids on the glycerol backbone could not be distinguished using these HPLC coitions. For example, in case of an EIC correspoing to PP Fig. 2D, the first five peaks correspoed to PPC 4, PPC 6, PPC 8, PPC 10, a PPC 12. The fatty acid composition of each TAG was identified using the APCI/MS spectra data not shown. The next peaks presumably correspoed to PPPo Po: palmtoleic acid, PPM, PPO, PPP, a PPS according to PN though it was difficult to estimate fatty acid composing TAG molecular species because three or more DAG fragment ions were detected from coeluting peaks. The other EICs probably show that TAGs comprised of two LCFAs a one SCFA or MCFA as is the case with EIC of PP. Accordingly, we assumed that the actual amount ratio of the TAG positional isomers a enantiomers, which consisted of a dipalmitoyl glycerol backbone, a one SCFA or MCFA, represented the whole TAG structure, which contained other DAG backbones present in BMF. We analyzed the amount ratios of these TAG isomers because we could separate both the TAG positional isomers a enantiomers that contained a dipalmitoyl glycerol backbone. 3.2 Quantification of TAG positional isomers The objective of the separation a quantification of TAG positional isomers in this section is not only to examine the ratio of TAG positional isomers but also to fractionate asymmetric TAGs for chiral separation in a later section Section 3.4. We previously reported that the separation of β-pc 10 P/β-PPC 10 a β-pc 4 P/β-PPC 4 was achieved using a C28 column at 15 uer isocratic coitions with either acetone or mixture of acetone a acetonitrile Although we have already reported that the ratio of β-pc 10 P/β-PPC 10 in the milk a cheese of ruminants including cow 20, the data iicated only the ratio of β-pc 10 P/ β-ppc 10 but not their amounts in milk fat. Meanwhile, we have already quantified β-pc 4 P/β-PPC 4 in BMF 19. In both cases, each enantiomer was not separated a quantified. To separate the TAG positional isomers, β-pc n P/β-PPC n where n 4, 6, 8, 10, a 12, we used a C28 reversedphase column at 15 a a solvent gradient of acetone a acetonitrile for the simultaneous separation of β-pc n P/ β-ppc n where n 4, 6, 8, 10, a 12. To selectively detect these TAGs, we used the SRM mode of the LC/MS/MS, a selected PPC n NH 4 a PP as the precursor a product ions, respectively. Ammonium adduct ions are pre- 947

6 T. Nagai, N. Watanabe a K. Yoshinaga et al. Fig. 2 Extracted ion chromatograms (EICs) a a total ion chromatogram (TICC) of the TAG molecular species in bovine milk fat (BMF). A) m/z 495 correspoing to the dimyristoylglycerol ion, [MM] + ; B) m/z 523 correspoing to the myristoylpalmitoylglycerol ion, [MP] + ; C) m/z 577 correspoing to the palmitoyloleoylglycerol ion, [PO] + ; D) m/z 551 correspoing to the dipalmitoylglycerol ion, [PP] + ; E) m/z 579 correspoing to the palmitoylstearoylglycerol ion, [PS] + ; a F) TICC. Column: Sunrise C28 (4.6 mm i.d. 250 mm, 5 μm), column temperature: 35, mobile phase: acetone/ acetonitrile (v/v) 50/50 (0 min) 100/0 (30 min) 100/0 (45 min), flow rate: 1.0 ml/min, ion source of MS: APCI positive, scan range: m/z The biing positions of fatty acids on the glycerol backbone could not be distinguished uer these coitions. sumably caused by ammonium contained in acetonitrile as trace impurity. As a result, all the staards of TAG positional isomer pairs were resolved Fig. 3A E. Retention times of each pair of symmetric a asymmetric TAGs increased with carbon number of three acyl chains, a symmetric TAGs were eluted earlier. The linear calibration curves for each of the staard solutions of the TAG positional isomers were obtained over a range of μg/ml, a all of the correlation coefficients were greater than 0.97 Table 1. The values of slopes of β-ppc n were greater than those of β-pc n P because formation of PP from β-ppc n in collision cell is energetically more favorable than that from β-pc n P where n 4, 6, 8, 10, a In BMF Fig. 3F J, no symmetric TAG positional isomers containing C 4, C 6, C 8 or C 10 were detected, but a small amount of that containing C 12 was detected Table 2. The result that only β-ppc 4 was detected in BMF was consistent with our previous paper 19. The detected ratio of β-pc 10 P/β-PPC 10 conflicts with our previous report 20. The reason for this is unclear, but this inconsistency might be caused by differences in the origin a season in which the milk was produced 25, 26. Some of the positional distributions of C 10 a C 12, which were reported by several researchers, are also inconsistent. We suspect that the amount ratios of these TAG positional isomers might be affected by seasonal a regional variations in the positional distributions of fatty acids 12, 27. However, at least we fou that most TAG positional isomers that consisted of two palmitic acids a either C 4, C 6, or C 8 existed in the asymmetric form. 948

7 Triacylglycerol stereoisomers in bovine milk fat Fig. 3 TAG positional isomer separation of bovine milk fat (BMF) TAGs comprised of two palmitic acids a one SCFA (or MCFA). A) β-pc 4 P/β-PPC 4, B) β-pc 6 P/β-PPC 6, C) β-pc 8 P/β-PPC 8, D) β-pc 10 P/β-PPC 10, a E) β-pc 12 P/β-PPC 12 are selected reaction monitoring (SRM) chromatograms of the staard solutions; a F) β-pc 4 P/β-PPC 4, G) β-pc 6 P/β-PPC 6, H) β-pc 8 P/β-PPC 8, I) β-pc 10 P/β-PPC 10, a J) β-pc 12 P/β-PPC 12 are SRM chromatograms of BMF. Column: Sunrise C28 (4.6 mm i.d. 250 mm, 5 μm), column temperature: 15, mobile phase: acetone/ acetonitrile (v/v) 80/20 (0 min) 100/0 (40 min), flow rate: 1.0 ml/min, ion source of MS: APCI positive, SRM transitions: m/z for PPC 4 ; m/z for PPC 6 ; m/z for PPC 8 ; m/z for PPC 10, a m/z for PPC 12. Table 2 sn-ppc 4 4.5±0.2 sn-ppc 6 2.3±0.2 sn-ppc 8 0.8±0.1 sn-ppc ±0.1 sn-ppc ±0.1 Mean±SD (n=3) Abuance ratios of triacylglycerol positional isomers a enantiomers comprised of dipalmitoylglycerol backbone with short- or medium-chain fatty acid in bovine milk fat (wt%). sn-pc 4 P sn-pc 6 P sn-pc 8 P sn-pc 10 P sn-pc 12 P 0.12±0.01 sn-c 4 PP sn-c 6 PP sn-c 8 PP sn-c 10 PP sn-c 12 PP : not detected (less than 0.1wt%) 3.3 Chiral separation of staards of TAG enantiomers using a recycle HPLC system with a chiral column Next, we examined the ratios of the TAG enantiomers in β-ppc 4, β-ppc 6, β-ppc 8, β-ppc 10, a β-ppc 12 using chiral HPLC equipped with a recycling system. First, we confirmed whether the staard solutions of racemic mixtures of the asymmetric TAGs rac-ppc 4, rac-ppc 6, rac-ppc 8, rac-ppc 10, a rac-ppc 12 were separated into their enantiomers on the CHIRALCEL OD-3R chiral stationary phase CSP, which was the same as the CHIRALCEL OD-RH we used previously 21. As a result, rac-ppc n where n 4, 6, 8, 10, a 12 was resolved into their enantiomers using the CHIRALCEL OD-3R column a the recycle HPLC system which makes the analyte pass through the column five times Fig. 4A E. This is the first chiral separation of TAG enantiomers comprised of two Ps a one SCFA or MCFA. Retention time of each pair of asymmetric TAGs increased with carbon number of three acyl chains, a 949

8 T. Nagai, N. Watanabe a K. Yoshinaga et al. Fig. 4 TAG enantiomer separation of bovine milk fat (BMF) asymmetric TAGs comprised of two palmitic acids a one SCFA (or MCFA). A) rac-ppc 4, B) rac-ppc 6, C) rac-ppc 8, D) rac-ppc 10, a E) rac-ppc 12 are selected reaction monitoring (SRM) chromatograms of the staard solutions, a F) β-ppc 4, G) β-ppc 6, H) β-ppc 8, I) β-ppc 10, a J) β-ppc 12 are SRM chromatograms of BMF. Column: CHIRALCEL OD-3R (4.6 mm i.d. 150 mm, 3 μm) with a guard cartridge (CHIRALCEL OD-3R, 4.0 mm i.d. 10 mm, 3 μm), column temperature: 25, mobile phase: methanol, flow rate: 0.5 ml/min, the number of passages through the column by the recycle HPLC system: five passes for each TAG, ion source of MS: APCI positive, SRM transitions: m/z for PPC 4 ; m/z for PPC 6 ; m/z for PPC 8 ; m/z for PPC 10, a m/z for PPC 12. sn-ppc n type enantiomer where n 4, 6, 8, 10, a 12 was eluted later. Although we reported that both saturated a unsaturated fatty acids at the sn-1 a 3 positions of the asymmetric TAGs, such as rac-ppo a rac-oop, were necessary for the chiral separation, these results showed that the difference in the length of the saturated-acyl chains at the sn-1 a 3 positions was recognized by the CSP of the CHIRALCEL OD-3R. This chiral recognition is probably due to the CH π interaction between the saturated-acyl chains a the CSP 28. Therefore, the performance of chiral HPLC using CHIRALCEL OD-3R with a recycle HPLC system was sufficient to analyze the enantiomeric ratios of β-ppc n where n 4, 6, 8, 10, a 12 in BMF. 3.4 Enantiomeric ratios of asymmetric TAGs in BMF To analyze the enantiomeric ratios of these asymmetric TAGs in BMF, the fractionation of asymmetric TAGs using reversed-phase HPLC was required before the chiral HPLC analysis because the asymmetric TAGs were iistinguishable from the symmetric TAGs when they overlapped on the SRM chromatogram. The chiral analysis of the asymmetric TAG fractions, β-ppc n where n 4, 6, 8, 10, a 12, showed that each of the asymmetric TAGs was composed of mostly one enantiomer with SCFA or MCFA at the sn-3 position, namely, sn-ppc 4, sn-ppc 6, sn-ppc 8, sn- PPC 10, a sn-ppc 12 Fig. 4F J. If the positional distribution 12 of C 10 a C 12 was considered, more sn-c 10 PP a sn-c 12 PP should be detected in the asymmetric TAGs fraction. TAGs other than those with dipalmitoyl glycerol back- 950

9 Triacylglycerol stereoisomers in bovine milk fat bones would bi C 10 a C 12 at their sn-1 a 2 positions. The result that most of the TAG isomers with a dipalmitoylglycerol backbone containing C 4, C 6, C 8, a C 10 existed as sn-ppc 4, sn-ppc 6, sn-ppc 8, a sn-ppc 10, was consistent with the positional distribution of C 4, C 6, C 8, C 10, a P. The contents of sn-ppc 4, sn-ppc 6, sn-ppc 8, sn-ppc 10, a sn-ppc 12 in BMF were 4.5, 2.3, 0.8, 1.3, a 0.8 wt, respectively, with a total of 9.7 wt Table 2. Furthermore, we fou that the composition of the sn-ppc 4, sn-ppc 6, sn-ppc 8, a sn-ppc 10 was similar to those of C 4, C 6, C 8, a C 10 at the sn-3 position in BMF 19.2, 8.9, 3.0, a 6.1 wt, respectively, which were converted into weight percent using the positional distribution of fatty acids in BMF 12. Therefore, the composition of the TAG positional isomers a enantiomers, which were comprised of dipalmitoyl glycerol backbone a SCFA or MCFA, sufficiently represented the structure of all of the TAGs with SCFA or MCFA in BMF. These data iicate TAGs in BMF are homochiral. SCFA or MCFA at the sn-3 position of BMF plays an important role in nutrition in calf. Pregastric lipase of calf shows a specificity for C 4, C 6, C 8, a C 10 a preferentially hydrolyzes ester bos at the α positions sn-1,3 positions 15, 29. These SCFA a MCFA would be efficient energy source for calf because they are selectively hydrolyzed by pregastric lipase in the stomach a then absorbed into gastric mucosa as mentioned above. TAG in BMF is synthesized by acyltransferases in glycerol phosphate pathway 30 at mammary gla 7. Bovine DAG acyltransferase probably bis SCFA or MCFA stereoselectively to the sn-3 position of 1,2-DAG 31. In this way, homochiral TAGs in BMF are likely synthesized by DAG acyltransferase in mammary gla with consideration for calf nutrition. In addition to nutritional properties, BMF might have specific physical properties due to its homochiral structure. We recently reported a difference in the physical properties of rac-ppo a sn-ppo or sn-opp 32, a the fact that the physical properties of a racemic TAG a one side of its enantiomers were different is of interest to researchers a workers in lipid science a the oil a fat iustry. Future studies should focus on the nutritional a physical properties of the iividual TAGs in BMF. 4 CONCLUSION TAG positional isomers a enantiomers comprised of dipalmitoylglycerol backbone a SCFA or MCFA in BMF were resolved a quantified using C28 reversed-phase column a CHIRALCEL OD-3R chiral column. This is the first chiral separation of TAG enantiomers biing only saturated fatty acids at sn-1 a sn-3 positions, a the result iicates that the structure of BMF TAG is homochiral, which means SCFAs or MCFAs are bou predominantly to the sn-3 position of dipalmitoyl glycerol backbone. Our result is consistent with previous researches on the positional distribution of fatty acids in BMF. References 1 Botham, K. M.; Mayes, P. A. Lipids of physiologic significance. in Harper s Illustrated Biochemistry 29th ed. Murray, R. K.; Beer, D. A.; Botham, K. M.; Kennelly, P. J.; Rodwell, V. W.; Weil P. A. ed., McGraw-Hill Companies, New York, pp Nawar, W. W. Lipid. in Food chemistry. 3rd ed. Owen, R. F. ed. Marcel Dekker Inc., New York, pp Brockerhoff, H. A stereospecific analysis of triglycerides. J. Lipid Res. 6, Brockerhoff, H.; Yurkowski, M. Stereospecific analyses of several vegetable fats. J. Lipid Res. 7, Brockerhoff, H.; Hoyle, R. J.; Wolmark, N. Positional distribution of fatty acids in triglycerides of animal depot fats. Biochim. Biophys. Acta. 116, Itabashi, Y.; Takagi, T. High performance liquid chromatographic separation of monoacylglycerol enantiomers on a chiral stationary phase. Lipids 21, Itabashi, Y.; Kuksis, A.; Myher, J. J. Determination of molecular species of enantiomeric diacylglycerols by chiral phase high performance liquid chromatography a polar capillary gas-liquid chromatography. J. Lipid Res. 31, Takagi, T.; Ao, Y. Stereospecific analysis of triacylsn-glycerols by chiral high-performance liquid chromatography. Lipids 26, Ao, Y.; Nishimura, K.; Aoyanagi, N.; Takagi, T. Stereospecific analysis of fish oil triacyl-sn-glycerols. J. Am. Oil Chem. Soc. 69, Itabashi, Y.; Myher, J. J.; Kuksis, A. Determination of positional distribution of short-chain fatty acids in bovine milk fat on chiral columns. J. Am. Oil Chem. Soc. 70, Jensen, R. G.; Ferris, A. M.; Lammi-Keefe, C. J.; Heerson, R. A. Lipids of bovine a human milks: A comparison. J. Dairy Sci. 73, Christie, W. W.; Clapperton, J. L. Structures of the triglycerides of cows milk, fortified milks including. infant formulae, a human milk. Int. J. Dairy Technol. 35, Christie, W. W. Structure of the triacyl-sn-glycerols in plasma a milk of the rat a rabit. J. Dairy Res. 52, Breckenridge, W. C.; Marai, L.; Kuksis, A. Triglyceride structure of human milk fat. Canad. J. Biochem. 47, Jensen, R. G.; DeJong, F. A.; Clark, R. M. Determina- 951

10 T. Nagai, N. Watanabe a K. Yoshinaga et al. tion of lipase specificity. Lipids 18, Rogalska, E.; Ransac, S.; Verger, R. Stereoselectivity of lipases. II. Stereoselective hydrolysis of triglycerides by gastric a pancreatic lipases. J. Biol. Chem. 265, Bracco, U. Effect of triglyceride structure on fat absorption. Am. J. Clin. Nutr Suppl, 1002S- 1009S Nagai, T.; Gotoh, N.; Mizobe, H.; Yoshinaga, K.; Kojima, K.; Matsumoto, Y.; Wada, S. Rapid separation of triacylglycerol positional isomers biing two saturated fatty acids using octacocyl silylation column. J. Oleo Sci. 60, Yoshinaga, K.; Nagai, T.; Mizobe, H.; Kojima, K.; Gotoh, N. Simple method for the quantification of milk fat content in foods by LC-APCI-MS/MS using 1,2-dipalmitoyl-3-butyroyl-glycerol as an iicator. J. Oleo Sci. 62, Gotoh, N.; Matsumoto, Y.; Nagai, T.; Mizobe, H.; Yoshinaga, K.; Kojima, K.; Kuroda, I.; Kitamura Y.; Shimizu, T.; Ishida, H.; Wada, S. Actual ratio of triacylglycerol positional isomers in milk a cheese. J. Oleo Sci. 61, Nagai, T.; Mizobe, H.; Otake, I.; Ichioka, K.; Kojima, K.; Matsumoto, Y.; Gotoh, N.; Kuroda, I.; Wada, S. Enantiomeric separation of asymmetric triacylglycerol by recycle high-performance liquid chromatography with chiral column. J. Chromatogr. A. 1218, Wada, S.; Koizumi, C.; Nonaka, J. Analysis of triglycerides of soybean oil by high-performance liquid chromatography in combination with gas liquid Chromatography. J. Jpn. Oil Chem. Soc. Yukagaku 26, Breckenridge, W. C.; Kuksis, A. Specific distribution of short-chain fatty acids in molecular distillates of bovine milk fat. J. Lipid Res. 9, Mottram, H. R.; Evershed, R. P. Structure analysis of triacylglycerol positional isomers using atmospheric pressure chemical ionization mass spectrometry. Tetrahedron Lett. 37, Parodi, P. W. Fatty acid composition of Australian butter a milk fats. Aust. J. Dairy Technol. 25, Jensen, R. G. The composition of bovine milk lipids: January 1995 to December J. Dairy Sci. 85, Parodi, P. W. Positional distribution of fatty acids in triglycerides from milk of several species of mammals. Lipids 17, Nishio, M.; Umezawa, Y.; Fantini, J.; Weiss, M. S.; Chakrabarti, P. CH π hydrogen bos in biological macromolecules. Phys. Chem. Chem. Phys. 16, Nelson, J. H.; Jensen, R. G.; Pitas, R. E. Pregastric esterase a other oral lipases a review. J. Dairy Sci. 60, Botham, K. M.; Mayes, P. A. Metabolism of acylglycerols. in Harper s Illustrated Biochemistry 29th ed. Murray, R. K.; Beer, D. A.; Botham, K. M.; Kennelly, P. J.; Rodwell, V. W.; Weil P. A. ed., McGraw-Hill Companies, New York, pp Marshall, M. O.; Knudsen, J. Biosynthesis of triacylglycerols containing short-chain fatty acids in lactating cow mammary gla. Eur. J. Biochem. 81, Mizobe, H.; Tanaka, T.; Hatakeyama, N.; Nagai, T.; Ichioka, K.; Hooh, H.; Ueno, S.; Sato, K. Structures a binary mixing characteristics of enantiomers of 1-oleoyl-2,3-dipalmitoyl-sn-glycerol S-OPP a 1,2-dipalmitoyl-3-oleoyl-sn-glycerol R-PPO. J. Am. Oil Chem. Soc. 90,