Journal of Oleo Science Copyright 2017 by Japan Oil Chemists Society doi : 10.5650/jos.ess16180 Quantification of Triacylglycerol Molecular Species in Edible Fats and Oils by Gas Chromatography- Flame Ionization Detector Using Correction Factors Kazuaki Yoshinaga 1, Junji Obi 2, Toshiharu Nagai 1, Hiroyuki Iioka 1, Akihiko Yoshida 1, Fumiaki Beppu 2 and Naohiro Gotoh 2* 1 Tsukishima Foods Industry Co. Ltd. (3-17-9 Higashi Kasai, Edogawa-ku, Tokyo 134-8520, JAPAN) 2 Department of Food Science and Technology, Tokyo University of Marine Science and Technology (4-5-7 Konan, Minato-ku, Tokyo 108-8477, JAPAN) Abstract: In the present study, the resolution parameters and correction factors (CFs) of triacylglycerol (TAG) standards were estimated by gas chromatography-flame ionization detector (GC-FID) to achieve the precise quantification of the TAG composition in edible fats and oils. Forty seven TAG standards comprising capric acid, lauric acid, myristic acid, pentadecanoic acid, palmitic acid, palmitoleic acid, stearic acid, oleic acid, linoleic acid, and/or linolenic acid were analyzed, and the CFs of these TAGs were obtained against tripentadecanoyl glycerol as the internal standard. The capillary column was Ultra ALLOY + -65 (30 m 0.25 mm i.d., 0.10 μm thickness) and the column temperature was programmed to rise from 250 C to 360 C at 4 C/min and then hold for 25 min. The limit of detection (LOD) and limit of quantification (LOQ) values of the TAG standards were > 0.10 mg and > 0.32 mg per 100 mg fat and oil, respectively, except for LnLnLn, and the LOD and LOQ values of LnLnLn were 0.55 mg and 1.84 mg per 100 mg fat and oil, respectively. The CFs of TAG standards decreased with increasing total acyl carbon number and degree of desaturation of TAG molecules. Also, there were no remarkable differences in the CFs between TAG positional isomers such as 1-palmitoyl-2-oleoyl-3-stearoyl-rac-glycerol, 1-stearoyl-2-palmitoyl-3-oleoyl-rac-glycerol, and 1-palmitoyl-2-stearoyl-3-oleoyl-rac-glycerol, which cannot be separated by GC-FID. Furthermore, this method was able to predict the CFs of heterogeneous (AAB- and ABC-type) TAGs from the CFs of homogenous (AAA-, BBB-, and CCC-type) TAGs. In addition, the TAG composition in cocoa butter, palm oil, and canola oil was determined using CFs, and the results were found to be in good agreement with those reported in the literature. Therefore, the GC-FID method using CFs can be successfully used for the quantification of TAG molecular species in natural fats and oils. Key words: gas chromatography, molecular species, quantification, triacylglycerol 1 INTRODUCTION Triacylglycerol TAG is the main constituent of edible fats and oils, and consists of one glycerol molecule and three fatty acid molecules 1. Since many classes of fatty acids exist in nature, an almost infinite number of TAG molecular species exist. Furthermore, two different binding positions for the fatty acids are possible, defining the primary and secondary alcohol groups on glycerol as the sn-1/3 position and the β sn-2 position, respectively. In the case of TAGs comprising two types of fatty acids, A and B, with two A and one B moieties on the glycerol backbone, there are two types of TAG molecular species, i.e., a TAG with A located at the β position AAB and a TAG with B located at the β position ABA. Since these TAGs consist Abbreviations: AOCS, American Oil Chemists Society; C10, capric acid; C12, lauric acid; C14, myristic acid; C15, pentadecanoic acid; CF, correction factor; DB, total double bond number; ECN, equivalent carbon length; GC-FID, gas chromatography flame ionization detector; HPLC, high performance liquid chromatography; L, linoleic acid; Ln, linolenic acid; LOD, limit of detection; LOQ, limit of quantification; O, oleic acid; P, palmitic acid; PN, partition number; Po, palmitoleic acid; RP, reverse phase; S, stearic acid; s/n, signal-to-noise ratio; TAG, triacylglycerol; TCN, total acyl carbon number. * Correspondence to: Naohiro Gotoh, Department of Food Science and Technology, Tokyo University of Marine Science and Technology (4-5-7 Konan, Minato-ku, Tokyo 108-8477, JAPAN) E-mail: ngotoh@kaiyodai.ac.jp Accepted October 31, 2016 (received for review September 13, 2016) Journal of Oleo Science ISSN 1345-8957 print / ISSN 1347-3352 online http://www.jstage.jst.go.jp/browse/jos/ http://mc.manusriptcentral.com/jjocs 259
K. Yoshinaga, J. Obi and T. Nagai et al. of the same components and number of fatty acids, they are considered TAG positional isomers. The TAG structure, dependent on the TAG molecular species and TAG positional isomers, strongly affects the characteristics of TAGs regarding their digestion, absorption, and physiology 2 4. Therefore, a method to quantify the TAG content in fats and oils is needed to characterize certain foods. So far, several analytical methods using high performance liquid chromatography HPLC or gas chromatography GC have been developed for quantifying TAG in fats and oils. Among them, reverse phase RP -HPLC for the resolution of TAG molecular species has been well established by many organizations, such as the American Oil Chemists Society AOCS 5, 6, the Japan Oil Chemists Society JOCS 7, and the International Union of Pure and Applied Chemistry IUPAC 8. RP-HPLC can separate the respective TAG molecular species by their Partition Number PN 9 or Equivalent Carbon Number ECN 10. The PN or ECN is defined by the equation PN TCN 2 DB, where TCN is the total acyl carbon number and DB is the total double bond number in the three fatty acids esterified on the TAG molecule. TAG molecular species having the same PN are called critical pairs and are difficult to separate. For example, the following TAG molecular species consisting of palmitic acid P and/or oleic acid O tend to elute together: PPP, POP, PPO, OPO, POO, and OOO. The resolution of these TAG critical pairs, including TAG positional isomers such as POP and PPO, using RP-HPLC has been reported by a few research groups 11 13. However, their elution pattern is complicated, and thus lipid analysts are required to possess extensive expertise on TAG resolution for RP-HPLC analysis. GC methods, on the other hand, result in resolution patterns of TAG molecular species that are simple and easy to understand. The resolution patterns of TAGs are governed by the polarity of the stationary phase with which the capillary column is coated. In general, nonpolar polysiloxane stationary phases allow separation according only to the TCN of TAG molecular species 14. For example, when TAGs consisting of P, stearic acid S, O, and/or linoleic acid L are analyzed using a nonpolar capillary column, the TCN52 peak contains PSS, POS, POO, and POL. Improvements in the separation of these TAGs have been achieved using more polar polysiloxane phases with high contents of phenyl groups 50 60 15, 16. These capillary columns can separate TAGs by their TCN in addition to their degree of unsaturation. Thus, when the TCN52 peak described above is analyzed with a polar capillary column, the fully saturated PPS is eluted first, followed by POS, POO, and finally POL. However, TAG positional isomers, such as POS, SPO, and PSO, cannot be resolved under typical analytical conditions. The official methods by which TAG molecular species in fats and oils are quantified have been accredited by the AOCS 17, JOCS 18, and IUPAC 19. These official methods employ GC with a flame ionization detector FID equipped with a nonpolar stationary phase. In contrast, the AOCS official methods Ce 11-05 20 and Ce 11a-07 21 employ a polar stationary phase and can separate POP, POS, and SOS. However, the aim of these methods is to detect and quantify cocoa butter equivalents in cocoa butter or chocolate, and other fats and oils cannot be determined. Thus, an analytical method for the quantification of the large number of TAG molecular species that exist in natural fats and oils is needed. In terms of TAG analysis by GC-FID, some points to consider are the discrimination during sample injection 22, 23 and the thermal degradation during analysis 24. Discrimination occurs when carrying out a split injection of a mixture of volatile compounds with a wide range of volatile points; the peaks of high-molecular-weight compounds are smaller than those of low-molecular-weight compounds because the high-molecular-weight compounds volatilize and enter the column at a slower pace. On the other hand, thermal degradation occurs when analyzing thermally unstable compounds such as highly unsaturated components. Therefore, discrimination and thermal degradation are to be avoided in analysis, and the peak areas of TAGs need to be corrected using factors calculated from TAG standards to obtain high accuracy and precision. However, there are many classes of TAGs in natural fats and oils, and it is difficult to obtain the necessary TAG standards because of their cost and limited number of suppliers. In particular, TAG standards consisting of two or three different fatty acids AAB- and ABC-type TAGs are more expensive and difficult to obtain than homogenous AAA-, BBB- and CCCtype TAGs. Therefore, a new method capable of predicting the correction factors CFs of AAB- and ABC-type TAGs from CFs of homogenous TAGs is highly desirable from an analytical point of view. The aim of the present study was to develop a precise and simple quantification method for TAG molecular species. For this purpose, a series of TAG standards were synthesized, allowing the detailed investigation of their separation behavior and CFs. In addition, cocoa butter, palm oil, and canola oil were selected as representative fats and oils on the market for further investigation, and their TAG content was determined. Furthermore, a simple method capable of predicting the CFs of complicated TAGs from a few simpler TAG standards was developed as an alternative to CF calculation methods that require large numbers of TAG standards. 2 EXPERIMENTAL 2.1 Chemicals and materials The 47 TAG standards used in this study, including capric acid C10, lauric acid C12, myristic acid C14, 260
Quantification of triacylglycerol by GC-FID Table 1 Classification of 47 TAG standards used in this study according to their TCN and DB. No. TCN DB TAG No. TCN DB TAG 1 30 0 C10C10C10 15 52 0 SPS, SSP 2 32 0 C10C12C10, C10C10C12 16 52 1 POS, SPO, PSO 3 34 0 C12C10C12, C12C12C10 17 52 2 OPO, OOP 4 36 0 C12C12C12, C10C12C14 18 52 3 POL, PLO, OPL 5 38 0 C12C14C12, C12C12C14 19 52 4 LPL, LLP 6 40 0 C14C12C14, C14C14C12 20 54 0 SSS 7 42 0 C14C14C14 21 54 1 SOS, SSO 8 44 0 PC12P, PPC12 22 54 2 OSO, OOS 9 45 0 C15C15C15 23 54 3 OOO 10 48 0 PPP 24 54 4 OLO, OOL 11 48 3 PoPoPo 25 54 5 LOL, LLO 12 50 0 PSP, PPS 26 54 6 LLL 13 50 1 POP, PPO 27 54 9 LnLnLn 14 50 2 PLP, PPL pentadecanoic acid C15, P, palmitoleic acid Po, S, O, L, and/or linolenic acid Ln, are summarized in Table 1. For example, POS, SPO, and PSO mean 1-palmitoyl-2-oleoyl- 3-stearoyl-rac-glycerol, 1-stearoyl-2-palmitoyl-3-oleoylrac-glycerol, and 1-palmitoyl-2-stearoyl-3-oleoyl-rac-glycerol, respectively. They were synthesized according to the method reported by Lísa and Holčapek 25 with 99 purity. Any other regents were obtained from Wako Pure Chemical Industries, Ltd. Osaka, Japan. Cocoa butter, palm oil, and canola oil were manufactured in-house or purified products Tsukishi ma Foods Industry Co., Ltd., Tokyo, Japan. 2.2 Analytical conditions The analysis was carried out using a GC-FID system TRACE GC Ultra, Thermo Fisher Scientific, Waltham, MA, USA equipped with a capillary column Ultra ALLOY -65, 30 m 0.25 mm i.d., 0.10 μm thickness, Frontier Laboratories Ltd., Fukushima, Japan. The column temperature was programmed to rise from 250 C to 360 C at 4 C/min and then hold for 25 min. Both the injector and detector temperatures were set at 350 C. The flow rate of the carrier gas helium was 1.0 ml/min. The split ratio was 50:1 and the injection volume was 1 μl. All analyses were carried out five times, and each result has been given as the mean value. 2.3 Estimation of separation behavior and CFs The calibration curves for the respective TAG standards were obtained from the GC-FID chromatogram peak areas. Tripentadecanoyl glycerol C15C15C15 was used as the internal standard. TAG standards were dissolved in dichloromethane at concentrations of 60, 40, 20, 10, 5, 1, and 0.5 mg/10 ml. All sample solutions were also mixed with C15C15C15 at a concentration of 5 mg/10 ml. The samples were injected into the GC-FID system and the peak areas were used to create the calibration curves. The concentration ratio TAG/C15C15C15 was plotted on the x-axis and the chromatogram peak area ratio TAG/C15C15C15 was plotted on the y-axis. All the calibration curves were expressed as first-order equations to evaluate their linearity. The limit of detection LOD and limit of quantification LOQ were evaluated using the TAG standard solution and calculated using the signal-to-noise ratio s/n. The LOD and LOQ were based on s/n 3 and 10, respectively 26. The CFs of the respective TAG standards were examined using a standard solution prepared by dissolving the 47 TAG standards in dichloromethane. The standard solution also contained the C15C15C15 internal standard. The CFs were calculated from the following equation: CF A IS A TAG C TAG C IS where A IS and A TAG are the peak areas and C IS and C TAG are the concentrations of the internal standard and TAG, respectively. 2.4 Quantification of TAG molecular species in cocoa butter, palm oil, and canola oil The TAG molecular species in cocoa butter, palm oil, and canola oil were quantified using the obtained CFs. A 100 mg sample of cocoa butter, palm oil, or canola oil was transferred into a 10 ml volumetric flask and mixed with 1 ml of a 5 mg/ml solution of C15C15C15 in dichloromethane as the internal standard. The volumetric flask was then filled with dichloromethane and the concentration of the sample and C15C15C15 was thus set to 100 mg/10 ml and 261
K. Yoshinaga, J. Obi and T. Nagai et al. 5 mg/10 ml, respectively. Identification of the individual TAG molecular species was carried out using the retention times of the pure TAG standards. The TAG content g/100 g oil was calculated from the following equation: TAG content A TAG W IS A IS W S CF where W IS and W S represent the weight of the internal standard and the sample, respectively. Furthermore, the accuracy of this method was evaluated using a recovery test. Correctly quantified TAG standards and C15C15C15 were added to cocoa butter, palm oil, or canola oil. The fat or oil samples with or without added TAG standards were analyzed by GC-FID, and the chromatogram peak areas were used for quantification. The amount of TAG molecular species with or without additional TAG standards was obtained using the above equation, and the recovery rate was calculated. 2.5 Prediction of CFs of TAG molecular species A simple CF prediction method was developed as an alternative to the methods requiring many TAG standards. The predicted correction factor PCF ABC of TAGs comprising two or three different fatty acids was calculated from the following equation: PCF ABC 1/3CF AAA 1/3CF BBB 1/3CF CCC where PCF ABC is the CF of the TAG consisting of fatty acids A, B, and C, and CF AAA, CF BBB, and CF CCC are the CFs obtained by analyzing the homogenous TAGs consisting of fatty acids A, B, or C, respectively. 3 RESULTS AND DISCUSSION The analysis of TAG molecular species by GC-FID equipped with a polar capillary column has been described by many researchers. These techniques have been used to characterize cocoa butter and cocoa butter equivalents because they are suitable for quantifying their main TAGs POP, POS, and SOS 20, 21. However, the quantification of many other types of TAG molecular species in natural fats and oils has not been established. This is caused by the difficulty in obtaining the corresponding TAG standards required for the calculation of these CFs. In the present study, we synthesized 47 TAG standards Table 1, and investigated their separation behavior and CFs. Figure 1 shows the chromatogram of the TAG standards. The peak numbers in Fig. 1 indicate the TAG combinations described in Table 1. TAG standards with the same TCN and DN, namely the same peak number, eluted together. For example, TAG positional isomers such as POP and PPO could not be resolved. In contrast, TAG standards with different peak numbers were successfully separated. The TAG standards eluted in the order of increasing TCN and DB, in agreement with previous reports 15, 16. The peak areas of the TAG standards decreased with the increasing TCN and DN, and a small peak for trilinolenoyl glycerol LnLnLn was observed. Table 2 shows the correlation coefficients R 2 of the calibration curves, LODs, and LOQs of the TAG standards. The calibration curves of the TAG standards were plotted between 60 and 0.5 mg/10 ml as first-order equations R 2 0.992. In the sample preparation procedure used in this study, 100 mg of fat or oil was diluted with 10 ml of dichloromethane. Therefore, the units mg/10 ml of the LOD and LOQ in Table 2 mean the detectable and quantifiable TAG amount mg in 100 mg of fat and oil. The LOD and LOQ values of the TAG standards were 0.10 mg and 0.32 mg per 100 mg fat and oil, respectively, except for LnLnLn, and the LOD or LOQ values between the TAG standards having the same TCN and DB were almost the same. However, the LOD and LOQ values of LnLnLn were higher than those of other TAG standards, at least 2 mg of LnLnLn in 100 mg of fat and oil could be quantified. These results suggest that this method can detect and quantify TAGs over a wide concentration range with high sensibility. The CFs of the TAG standards against the internal standard are shown in Fig. 2. The CFs of TAG standards decreased with the increasing TCN and DN. The highest value 1.38 was observed for C10C10C10 and the lowest value 0.05 was observed for LnLnLn. CFs are known to be affected by many factors in GC-FID analysis: column length, polarity of stationary phase, analytical time, column temperature, flow rate of carrier gas, and the injection technique 22, 27 30. The optimization of these factors can improve the variation of CFs, but cannot eliminate all the negative effects completely. Conventional analytical methods employ nonpolar stationary phases, consider all TAG molecular species with the same TCN as a single one regardless of their DB, and quantify them using their CFs, which are calculated from TAG standards. The present study revealed that the CFs of TAG molecular species with the same TCN decreased with the increasing DB. For example, the CFs of PSP, POP, and PLP were 0.78, 0.77, and 0.72, respectively. These findings suggest that conventional methods could lead to miscalculation of the total TAG content in fats and oils when the analytical conditions are not optimized. The CFs of TAG positional isomers were compared in the present study since these isomers are usually eluted together in GC analysis. Table 3 shows the comparison of CFs between TAG positional isomers. The CFs of POS, SPO, and PSO were 0.64, 0.67, and 0.65, respectively. Although the values of CFs were somewhat different, the difference between TAG positional isomers was found to be below 5. Therefore, inseparable TAG positional isomers could be quantified using the same CFs under the analytical conditions used in this study. 262
Quantification of triacylglycerol by GC-FID Fig. 1 Gas chromatogram of TAG standards. GC column: Ultra ALLOY -65 30 m 0.25 mm i.d., 0.10 μm thickness. Peak assignment: 1 C10C10C10; 2 C10C12C10; 3 C12C10C12; 4 C12C12C12; 5 C12C14C12; 6 C14C12C14; 7 C14C14C14; 8 PC12P; 9 C15C15C15, internal standard IS ; 10 PPP; 11 PoPoPo; 12 PSP; 13 POP; 14 PLP; 15 SPS; 16 POS; 17 OPO; 18 POL; 19 LPL; 20 SSS; 21 SOS; 22 OSO; 23 OOO; 24 OLO; 25 LOL; 26 LLL; 27 LnLnLn. Table 2 Correlation coefficient R 2 of calibration curves and LOD and LOQ of TAG standards. No. TAG TCN:DB R 2 LOD LOQ No. TAG TCN:DB R 2 LOD LOQ 1 C10C10C10 30:0 0.997 0.06 0.20 15 SPS 52:0 0.994 0.03 0.11 2 C10C12C10 32:0 0.992 0.10 0.32 16 POS 52:1 0.993 0.03 0.11 3 C12C10C12 34:0 0.992 0.09 0.29 17 OPO 52:2 0.999 0.05 0.16 4 C12C12C12 36:0 0.995 0.09 0.29 18 POL 52:3 0.999 0.05 0.16 5 C12C14C12 38:0 0.998 0.08 0.26 19 LPL 52:4 0.996 0.05 0.15 6 C14C12C14 40:0 0.993 0.09 0.29 20 SSS 54:0 0.997 0.04 0.13 7 C14C14C14 42:0 0.993 0.08 0.25 21 SOS 54:1 0.994 0.07 0.22 8 PC12P 44:0 0.998 0.06 0.20 22 OSO 54:2 0.996 0.05 0.18 9 C15C15C15 45:0 0.999 0.06 0.22 23 OOO 54:3 0.997 0.04 0.15 10 PPP 48:0 1.000 0.06 0.21 24 OLO 54:4 0.998 0.05 0.17 11 PoPoPo 48:3 1.000 0.03 0.10 25 LOL 54:5 0.996 0.03 0.09 12 PSP 50:0 0.999 0.04 0.12 26 LLL 54:6 0.999 0.03 0.10 13 POP 50:1 0.999 0.07 0.23 27 LnLnLn 54:9 0.997 0.55 1.84 14 PLP 50:2 0.998 0.07 0.23 263
K. Yoshinaga, J. Obi and T. Nagai et al. Fig. 2 CFs of TAG standards. Table 3 Comparison of CFs between TAG positional isomers. TCN:DB TAG CF Difference TCN:DB TAG CF Difference 32:0 C10C12C10 1.33 POS 0.64 0.03 (POS vs. SPO) C10C10C12 1.32 52:1 SPO 0.67 0.02 (PLO vs. PSO) C12C10C12 1.27 PSO 0.65 (POS vs. PSO) 34:0 0.00 C12C12C10 1.27 OPO 0.62 52:2 C12C12C12 1.22 OOP 0.63 36:0 0.05 C10C12C14 1.27 POL 0.57 0.03 (POL vs. PLO) 38:0 C12C14C12 1.15 52:3 PLO 0.60 0.03 (PLO vs. OPL) C12C12C14 1.15 OPL 0.57 0.00 (POL vs. OPL) 40:0 C14C12C14 1.12 LPL 0.52 0.00 52:4 C14C14C12 1.12 LLP 0.54 0.02 44:0 PC12P 1.03 SOS 0.50 0.03 54:1 PPC12 1.07 SSO 0.52 0.02 50:0 PSP 0.78 OSO 0.47 0.02 54:2 PPS 0.80 OOS 0.48 0.02 50:1 POP 0.77 OLO 0.40 0.04 54:4 PPO 0.81 OOL 0.41 50:2 PLP 0.72 LOL 0.35 54:5 PPL 0.74 LLO 0.36 52:0 SPS 0.65 SSP 0.66 264
Quantification of triacylglycerol by GC-FID Fig. 3 Gas chromatograms of TAG molecular species in cocoa butter A, palm oil B, and canola oil C. Peak numbers are summarized in Table 1. The chromatograms of TAG molecular species of cocoa butter, palm oil, and canola oil are shown in Fig. 3. Cocoa butter contains three main TAG molecular species: POS peak 16, POP peak 13, and SOS peak 21. These peaks were clearly separated from other peaks. Additionally, critical pairs such as POO peak 17 and PLS were separated. POO was eluted ahead of PLS, suggesting that TAGs having one diunsaturated fatty acid are retained more effectively than TAGs having two monounsaturated acids on the column. Similarly, the main TAG molecular species in palm oil and canola oil were well separated. The TAG molecular species content in cocoa butter, palm oil, and canola oil is shown in Table 4. There are many reports on the content of POP, POS, and SOS in cocoa butter in the literature. For example, Beppu et al. reported that cocoa butter contains 17.3, 39.7, and 27.0 of POP, POS, and SOS, respectively 31. As quantified by GC-FID, the POP, POS, and SOS content in cocoa butter was obtained as 14.5, 37.8, and 28.1, respectively Table 4. On the other hand, the majority of the TAG molecular species in palm oil consists of POP PPO 26.2, followed by OPO OOP 18.6 and POL PLO OPL 8.7. Additionally, the main TAG molecular species in canola oil is OOO 24.0, followed by OLO OOL 20.3. The TAG amounts quantified in this study are in agreement with the previously reported data on the distribution of TAG molecular species in palm oil and canola oil 32, 33. In order to evaluate the accuracy of this method, recovery tests of TAG standards spiked into cocoa butter, palm oil, and canola oil samples were carried out and the results are summarized in Table 5. The recovery rates of this method were in the range of 85 115. Although further research is required to improve the accuracy of our analysis, the recovery rates obtained in this study are satisfactory for the quantification of TAGs in fats and oils. Moreover, we observed that the CFs of TAG standards were different when using another GC instrument Agilent 7890A, Agilent Technologies, Wilmington, DE, USA. However, there were no remarkable differences in the TAG contents in cocoa butter, palm oil, and canola oil between the two GC instruments provided by Agilent Technologies and Thermo Fisher Scientific data not shown. These results indicate that quantification of TAG molecular species in natural fats and oils is possible using our method. Methods that use many different TAG standards to calculate CFs are very complicated. Therefore, we developed a CF prediction method and the results are shown in Table 6. The measured and predicted CFs of POS, which is the 265
K. Yoshinaga, J. Obi and T. Nagai et al. Table 4 TCN:DB TAG composition in cocoa butter, palm oil, and canola oil. TAG (For example) Cocoa butter Palm oil Canola oil 48:0 PPP 0.1 5.6 0.0 50:0 PSP, PPS 0.4 1.0 0.1 50:1 POP, PPO 14.5 26.2 1.0 50:2 PLP, PPL 1.6 7.7 0.8 52:0 SPS, SSP 0.4 0.1 0.3 52:1 POS, SPO, PSO 37.8 4.5 0.2 52:2 OPO, OOP 2.0 18.6 4.1 52:3 POL, PLO, OPL 0.5 8.7 3.5 52:4 LPL, LLP 0.2 1.4 1.6 54:0 SSS 0.3 0.0 0.2 54:1 SOS, SSO 28.1 0.5 0.1 54:2 OSO, OOS 2.3 2.1 1.7 54:3 OOO 1.8 2.8 24.0 54:4 OLO, OOL 0.0 1.3 20.3 54:5 LOL, LLO 0.0 0.2 11.1 54:6 LLL 0.0 0.0 2.6 main component of cocoa butter, were 0.64 and 0.62, respectively. Similarly, the measured and predicted CFs of POP, which is the main TAG in palm oil, were 0.77 and 0.74, respectively. The difference between measured and predicted CF values is below 5. Therefore, our CF prediction method could be used for quantifying TAG molecular species in natural fats and oils. The AOCS 17, JOCS 18, and IUPAC 19 official methods cannot separate TAG molecular species having the same TCN, and TAG molecular species having different degrees of unsaturation. These methods are applicable to TAG molecular species in the range from TCN24 to TCN54. Our method, on the other hand, can separate TAGs by their TCN in addition to their degree of unsaturation. Therefore, our method could have a wider application range of TAG molecular species as compared with official methods. The quantification of TAG having medium chain fatty acids such as coconut oil should be possible through optimizing the analytical conditions, which is our next target. Table 5 Spiked TAG Recovery test of TAG standards spiked into fat and oil. TCN:DB Cocoa butter Palm oil Canola oil C10C10C10 30:0 95.9 97.3 99.4 C10C12C10 32:0 99.8 105.3 104.4 C12C10C12 34:0 87.8 103.1 101.5 C12C12C12 36:0 100.0 101.5 101.0 C12C14C12 38:0 98.8 100.3 94.3 C14C12C14 40:0 99.9 104.4 98.1 C14C14C14 42:0 101.1 100.5 100.9 PC12P 44:0 102.3 100.7 102.7 PPP 48:0 100.3 103.6 100.5 PoPoPo 48:3 102.2 98.6 101.6 PSP 50:0 99.3 98.9 96.6 POP 50:1 93.6 87.9 89.1 PLP 50:2 92.4 87.7 90.9 SPS 52:0 102.2 87.0 90.5 POS 52:1 92.7 88.6 97.2 OPO 52:2 89.9 107.7 103.3 POL 52:3 98.6 91.5 93.9 LPL 52:4 95.1 98.5 95.3 SSS 54:0 104.3 94.2 90.6 SOS 54:1 88.8 94.2 87.7 OSO 54:2 93.7 96.5 88.5 OOO 54:3 87.9 96.4 104.4 OLO 54:4 100.4 99.0 90.8 LOL 54:5 96.4 97.9 88.0 LLL 54:6 110.6 95.5 90.9 LnLnLn 54:9 85.9 86.8 87.8 the present method using CFs was found suitable for the quantification of TAG molecular species in natural fats and oils. Furthermore, we developed a simple CF prediction method for complex TAGs using easily obtainable TAG standards. We strongly believe that our method represents a step forward toward the quantification of TAG molecular species. 4 CONCLUSION In the present study, the separation behavior and CFs of TAG standards were investigated. It was found that the respective CFs were different depending on the TCN and DB of the TAG standards. It was also observed that the CFs of TAG positional isomers were almost the same. In addition, Conflict of interest The authors declare no conflicts of interest. 266
Quantification of triacylglycerol by GC-FID Table 6 Comparison of the measured and predicted CFs. TAG TCN:DB Measured Predicted TAG TCN:DB Measured Predicted C10C10C10 30:0 1.38 SPS 52:0 0.65 0.63 C10C12C10 32:0 1.33 1.32 POS 52:1 0.64 0.62 C12C10C12 34:0 1.27 1.27 OPO 52:2 0.62 0.60 C12C12C12 36:0 1.22 POL 52:3 0.57 0.55 C12C14C12 38:0 1.15 1.10 LPL 52:4 0.52 0.50 C14C12C14 40:0 1.12 1.13 SSS 54:0 0.51 C14C14C14 42:0 1.09 SOS 54:1 0.50 0.49 PC12P 44:0 1.03 0.99 OSO 54:2 0.47 0.48 C15C15C15 45:0 1.00 OOO 54:3 0.46 PPP 48:0 0.88 OLO 54:4 0.40 0.41 PoPoPo 48:3 0.83 LOL 54:5 0.35 0.36 PSP 50:0 0.78 0.76 LLL 54:6 0.31 POP 50:1 0.77 0.74 LnLnLn 54:9 0.05 PLP 50:2 0.72 0.69 References 1 Murray, R.K.; Granner, D.K.; Rodwell, V.W. Harper s Illustrated Biochemistry. 27th ed. McGraw-Hill Companies, New York, p. 124 2006. 2 Hunter, J.E. Studies on effect of dietary fatty acids as related to their position on triglycerides. Lipids 36, 655-668 2001. 3 Kubow, S. The influence of positional distribution of fatty acids in native, interesterified and structure-specific lipids on lipoprotein metabolism and atherogenesis. J. Nutr. Biochem. 7, 530-541 1996. 4 Sato, K. Crystallization behavior of fats and lipids a review. Chem. Eng. Sci. 56, 2255-2265 2001. 5 Triglycerides in vegetable oils by HPLC, AOCS Official Method, Ce 5b-89. 6 Individual triglycerides in oils and fats by HPLC, AOCS Official Method, Ce 5c-93. 7 Triacylglycerol composition high performance liquid chromatography, JOCS Official Method, 2.4.6.2-2013. 8 Wolff, J.P.; Mordret, F.X.; Dieffenbacher, A. Determination of triglycerides in vegetable oils in terms of their partition numbers by high performance liquid chromatography. Pure Appl. Chem. 63, 1173-1182 1991. 9 Wada, S.; Koizumi, C.; Nonaka, J. Analysis of triacylglycerides of soybean oil by high-performance liquid chromatography in combination with gas liquid chromatography. Yukagaku 26, 95-99 1977. 10 Plattner, R.D.; Spencer, G.F.; Kleiman, R. Triglycerides separation by reverse phase high performance liquid chromatography. J. Am. Oil Chem. Soc. 54, 511-515 1977. 11 Momchilova, S.; Tsuji, K.; Itabashi, Y.; Nikolova-Damyanova, B.; Kuksis, A. Resolution of triacylglycerol positional isomers by reversed-phase high performance liquid chromatography. J. Sep. Sci. 27, 1033-1036 2004. 12 Kuroda, I.; Nagai, T.; Mizobe, H.; Yoshimura, N.; Gotoh, N.; Wada, S. HPLC separation of triacylglycerol positional isomers on a polymeric ODS column. Anal. Sci. 24, 865-869 2008. 13 Nagai, T.; Gotoh, N.; Mizobe, H.; Yoshinaga, K.; Kojima, K.; Matsumoto, Y.; Wada, S. Rapid separation of triacylglycerol positional isomers binding two saturated fatty acids using octacocyl silylation column. J. Oleo Sci. 60, 345-350 2011. 14 Litchfield, C.; Harlow, R.D.; Reiser, R. Gas-liquid chromatography of triglyceride mixtures containing both odd and even carbon number fatty acids. Lipids 2, 363-370 1967. 15 Geeraert, E.; Sandra, P. Capillary GC of triglycerides in fats and oils using a high temperature phenylmethylsilicone stationary phase, part I. J. Sep. Sci. 8, 415-422 1985. 16 Geeraert, E.; Sandra, P. Capillary GC of triglycerides in fats and oils using a high temperature phenylmethylsilicone stationary phase. part II. The analysis of chocolate fats. J. Am. Oil Chem. Soc. 64, 100-105 1987. 17 Triglycerides by gas chromatography, AOCS Official Method, Ce 5-86. 18 Triacylglycerol composition Gas chromatography, JOCS Official Method, 2.4.6.1-2013. 19 Pocklington, W.D.; Hautfenne, A. Determination of triglycerides in fats and oils. Pure Appl. Chem. 57, 1515-1522 1985. 20 Determination of cocoa butter equivalents in cocoa 267
K. Yoshinaga, J. Obi and T. Nagai et al. butter and plain chocolate, AOCS Official Method, Ce 11-05. 21 Detection and quantification of cocoa butter equivalents in milk chocolate, AOCS Official Method, Ce 11a-07. 22 Mareš, P.; Hušek, P. Quantitative capillary gas-liquid chromatography of triglycerides on a fused-silica column with a chemically bonded stationary phase. J. Chromatogr. 350, 87-103 1985. 23 Mayer, B.X.; Lorbeer, E.A. Fused silica capillary column coated with a medium polar stationary phase for HRGC. J. Chromatogr. A, 758, 235-242 1997. 24 Andrikopoulos, N.K. Triglyceride species compositions of common edible vegetable oils and methods used for their identification and quantification. Food Rev. Int. 18, 71-102 2002. 25 Lísa, M.; Holčapek, M. Characterization of triacylglycerol enantiomers using chiral HPLC/APCI-MS and synthesis of enantiomeric triacylglycerols. Anal. Chem. 85, 1852-1859 2013. 26 Hernandez, M.J.; Garcia-Moreno, M.V.; Duran, E.; Guillen, D.; Barroso, C.G. Validation of two analytical methods for the detection of ochratoxin A by reversed-phased high-performance liquid chromatography coupled to fluorescence detection in musts and sweet wines from Andalusia. Anal. Chin. Acta 566, 117-121 2006. 27 Watts, R.; Dils, R. Separation of triglycerides by gasliquid chromatography. J. Lipid Res. 9, 40-51 1968. 28 Litchfield, C.; Harlow, R.D.; Reiser, R. Quantitative gas-liquid chromatography of triglycerides. J. Am. Oil Chem. Soc. 42, 849-857 1965. 29 Hinshaw, J.V.; Seferovic, W. Programmed-temperature split-splitless injection of triglycerides: comparison to cold on-column injection. J. Sep. Sci. 9, 69-72 1986. 30 Buchgraber, M.; Ulberth, F.; Anklam, E. Interlaboratory evaluation of injection techniques for triglyceride analysis of cocoa butter by capillary gas chromatography. J. Chromatogr. A 1036, 197-203 2004. 31 Beppu, F.; Nagai, T.; Yoshinaga, K.; Mizobe, H.; Kojima, K.; Gotoh, N. Quantification of triacylglycerol molecular species in cocoa butter using high-performance liquid chromatography equipped with nano quantity analyte detector. J. Oleo Sci. 62, 789-794 2013. 32 Deffense, E. Fractionation of palm oil. J. Am. Oil Chem. Soc. 62, 376-385 1985. 33 Neff, W.E.; Mounts, T.L.; Rinsch, W.M.; Konishi, H.; El- Agaimy, M.A. Oxidative stability of purified canola oil triacylglycerols with altered fatty acid compositions affected by triacylglycerol composition and structure. J. Am. Oil Chem. Soc. 71, 1101-1109 1994. 268