Anal. Chem. 2008, 80, 5788–5793
Regioisomeric Structure Determination of r- and γ-Linolenoyldilinoleoylglycerol in Blackcurrant Seed Oil by Silver Ion High-Performance Liquid Chromatography and Mass Spectrometry Heidi Leskinen,* Jukka-Pekka Suomela, Janne Pinta, and Heikki Kallio Department of Biochemistry and Food Chemistry, University of Turku, Vatselankatu 2, 20014 Turku, Finland γ-Linolenic acid (Gla) and oils containing Gla have evident positive effects on a variety of disorders, and therefore, the structure of triacylglycerols (TAGs) containing Gla is of special interest. The regioisomeric structures of TAGs 18:3(n-3)/18:2/18:2 (Ala/L/L) and 18:3(n-6)/18:2/ 18:2 (Gla/L/L) in blackcurrant seed oil were determined by Ag-HPLC/APCI-MS and Ag-HPLC/ESI-MS/MS techniques. In the latter, silver ion adducts were prepared by adding silver nitrate to the postcolumn flow. A new AgHPLC solvent system containing nonchlorinated solvents for the separation of Gla- and Ala-containing TAGs is introduced. Ag-HPLC separation of Ala/L/L and Gla/L/L was sufficient and regioisomers sn-LnLL + sn-LLLn and sn-LLnL (Ln, linolenic acid) could be differentiated well with the MS methods used. No discrimination was made between the sn-1 and sn-3 positions. The results show that the methods used are suitable to discriminate and determine the regioisomeric structure of TAGs. The regioisomeric structure of TAG with the fatty acid combination Gla/L/L in blackcurrant seed oil was considered to be practically random (32.7-37.8% of sn-LGlaL). In the fatty acid combination Ala/L/L, the regioisomeric structure is nonrandom (7.3-12.1% of sn-LAlaL) with Ala preferentially in a primary position. It can be concluded that the positional distribution of Ala and Gla is different in Ln/L/L TAGs of blackcurrant seed oil. Oils containing γ-linolenic acid (Gla) have been under investigation for a long time due to the possible positive effects of Gla on, for example, skin diseases such as atopic eczema. In 1992, Horrobin reviewed the nutritional and medicinal properties of Gla.1 In more recent publications, the therapeutic efficacy of Gla and oils containing Gla on a variety of disorders has been more inconsistent. Evening primrose oil, borage oil, and blackcurrant seed oil are the main sources of Gla used for nutritional purposes. Blackcurrant seed oil contains 11-19% Gla and 14-18% R-linolenic acid (Ala).2–6 Other abundant fatty acids (FAs) in * To whom correspondence should be addressed. E-mail:
[email protected]. Fax: +358 2 333 6860. (1) Horrobin, D. F. Prog. Lipid Res. 1992, 31, 163–194. (2) Johansson, A.; Laakso, P.; Kallio, H. Z. Lebensm.-Unters.-Forsch. A 1997, 204, 300–307. (3) Tahvonen, R. L.; Schwab, U. S.; Linderborg, K. M.; Mykka¨nen, H. M.; Kallio, H. P. J. Nutr. Biochem. 2005, 16, 353–359.
5788
Analytical Chemistry, Vol. 80, No. 15, August 1, 2008
blackcurrant seed oil are 18:1(n-9) (10-13%) and 18:2(n-6) (48-49%).2,3,5,6 The most abundant triacylglycerol (TAG) species in blackcurrant seed oil are 54:7 and 54:6 (ACN:DB, acyl carbon number/number of double bonds) comprising ∼40% of the total amount of TAG.7 It is not known whether and how enzymes involved in TAG synthesis in blackcurrant seeds discriminate between Ala and Gla. Regioisomerism of TAGs in blackcurrant seed oil ingested may also influence TAG metabolism in man. Thus, the structure of TAGs containing Ala and Gla is an interesting issue. Mass spectrometry (MS) is a powerful tool in analyzing TAGs because it is rapid and reasonably quantitative information about the positional distribution of FAs in individual TAGs can be obtained directly without any enzymatic treatment of the sample. However, liquid chromatographic separation of TAGs containing different FA isomers (double-bond positional isomerism, cis/transisomerism) prior to MS is essential, because these cannot be differentiated with MS as such. Kallio et al.8 developed a direct inlet ammonia negative-ion chemical ionization tandem mass spectrometric (MS/MS) method to be applied in TAG regioisomerism analysis, but successful connection of the method in a proper HPLC system is a challenge. Recently, several reports of TAG regioisomerism studies based on atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI) combined with high-performance liquid chromatography (HPLC) have been published. Byrdwell and Emken9 were the first to report the mass spectrometric analysis of pure TAGs using an APCI interface and HPLC. APCI ionization of TAGs produces mostly protonated [M + H]+ molecules and diacylglycerol (DAG) fragment ions [M + H - RCOOH]+. Laakso and Voutilainen,10 Mottram and Evershed,11 and Mottram et al.12 were the first to show dependence of the intensities of DAG fragments and positional distribution of FAs in an HPLC-MS (4) Manninen, P.; Laakso, P.; Kallio, H. Lipids 1995, 30, 665–671. (5) Traitler, H.; Winter, H.; Richli, U.; Ingenbleek, Y. Lipids 1984, 19, 923– 928. (6) Ruiz del Castillo, M. L.; Dobson, G.; Brennan, R.; Gordon, S. J. Agric. Food Chem. 2002, 50, 332–335. (7) Johansson, A.; Laakso, P.; Kallio, H. Z. Lebensm.-Unters.-Forsch. A 1997, 204, 308–315. (8) Kallio, H.; Currie, G. Lipids 1993, 28, 207–215. (9) Byrdwell, W. C.; Emken, E. A. Lipids 1995, 30, 173–175. (10) Laakso, P.; Voutilainen, P. Lipids 1996, 31, 1311–1322. (11) Mottram, H. R.; Evershed, R. P. Tetrahedron Lett. 1996, 37, 8593–8596. (12) Mottram, H. R.; Woodbury, S. E.; Evershed, R. P. Rapid Commun. Mass Spectrom. 1997, 11, 1240–1252. 10.1021/ac8004132 CCC: $40.75 2008 American Chemical Society Published on Web 06/26/2008
analysis. They found that the abundance of DAG fragment ions formed by a loss of a fatty acyl residue from the sn-2 position was less than that formed by a loss of a fatty acyl residue from the primary glycerol positions sn-1/3. In recent years, APCI-MS has been used in the regioisomer analysis of various TAGs in different fats and oils.13–16 The same features in the formation of DAG product ions [M + NH4 - RCOONH4]+ are reported with ESIMS/MS analysis of ammoniated TAGs [M + NH4]+. Hvattum17 combined RP-HPLC (RP, reversed phase) with ESI-MS/MS detection, and he also found that loss of sn-2 fatty acid was energetically less favorable than loss of fatty acid from sn-1/3 positions, and since then, ESI-MS/MS has been also used for the regioisomeric structure analysis of various TAGs in fats and oils.16,18–20 In addition to [M + NH4]+ ions, also the use of [M + Na]+21 and [M + Li]+22,23 ions in the analysis of TAG regioisomerism has been studied. Sandra et al.15 produced [M + Ag]+ ions of TAGs by coordination ion spray (CIS) in supercritical fluid chromatography and ESI-MS system, but they were not able to elucidate positional isomers. Silver nitrate was introduced in methanol as a makeup solvent. Silver ion chromatography has been proved to be applicable in TAG analysis, and the method has also been applied for the separation of different molecular species of lipids. More detailed reviews of silver ion chromatography are available in the articles of Dobson et al.24 and Nikolova-Damyanova.25 Preparation of an Ag-HPLC column from ion-exchange columns was optimized by Christie.26 Laakso and Voutilainen10 were the first to use the AgHPLC method of Christie26 in separating Ala- and Gla-containing TAGs in seed oils. The gradient was composed of dichloromethane/dichloroethane, acetone, and acetone/acetonitrile. Use of mobile phases containing chlorinated solvents has been common in Ag-HPLC, although they display carcinogenic and mutagenic properties, and a risk of solvent traces in the laboratory air is evident. Also acetonitrile, which is used in mobile phases of Ag-HPLC, has harmful properties. However, analyses performed with nonchlorinated solvents have been published. Adlof et al. have studied extensively acyl glycerols and FAs with Ag-HPLC using commercially available columns. They have used an isocratic system of appropriate proportion of acetonitrile in n-hexane in (13) Fauconnot, L.; Hau, J.; Aeschlimann, J. M.; Fay, L. B.; Dionisi, F. Rapid Commun. Mass Spectrom. 2004, 18, 218–224. (14) Jakab, A.; Jablonkai, I.; Forgacs, E. Rapid Commun. Mass Spectrom. 2003, 17, 2295–2302. (15) Sandra, P.; Medvedovici, A.; Zhao, Y.; David, F. J.Chromatogr., A 2002, 974, 231–241. (16) Leskinen, H.; Suomela, J.; Kallio, H. Rapid Commun. Mass Spectrom. 2007, 21, 2361–2373. (17) Hvattum, E. Rapid Commun. Mass Spectrom. 2001, 15, 187–190. (18) Kalo, P.; Kemppinen, A.; Ollilainen, V.; Kuksis, A. Lipids 2004, 39, 915– 928. (19) Kalo, P.; Kemppinen, A.; Ollilainen, V.; Kuksis, A. Int. J. Mass Spectrom. 2003, 229, 167–180. (20) Malone, M.; Evans, J. J. Lipids 2004, 39, 273–284. (21) Cheng, C.; Gross, M. L.; Pittenauer, E. Anal. Chem. 1998, 70, 4417–4426. (22) Han, X.; Gross, R. W. Anal. Biochem. 2001, 295, 88–100. (23) Hsu, F. F.; Turk, J. J. Am. Soc. Mass Spectrom. 1999, 10, 587–599. (24) Dobson, G.; Christie, W. W.; Nikolova-Damyanova, B. J. Chromatogr., B: Biomed. Appl. 1995, 671, 197–222. (25) Nikolova-Damyanova, B. In Advances in lipid methodology; 5th ed.; Adlof, R. O. Ed.; The Oily Press: Bridgwater, England, 2003; pp 43-123. (26) Christie, W. W. J. High Resolut. Chromatogr. Chromatogr. Commun. 1987, 10, 148–150.
order to separate different FAs as methyl esters (FAMEs),27–30 TAG positional isomers,31,32 TAGs enriched with different conjugated linoleic acid isomers,33 and acetate derivatives of 1- and 2-monoacylglycerols and 1,2- and 1,3-DAGs34 using different amounts of Ag-HPLC columns connected in series. The advantages of this isocratic solvent system are the lack of solvent equilibration between the runs and the use of nonchlorinated solvents, but in complicated TAG analysis, this system may not be sufficient. Elfman-Bo¨rjesson et al.35 presented analysis of cis and trans isomers of FAME, mixtures of FAMEs and fatty alcohols, and TAGs with gradients of n-heptane and acetonitrile in the Ag-HPLC system. Fe´vrier et al.36 used one commercial silver ion column and mobile phases containing n-hexane, 2-propanol, ethyl acetate, and acetonitrile to separate positional isomers of TAGs and DAGs. To our knowledge, Ag-HPLC separation of Glaand Ala-containing TAGs has not been reported with nonchlorinated solvents. The aim of this study was to identify the regioisomeric structure of TAGs 18:3(n-3)/18:2/18:2 (Ala/L/L) and 18:3(n-6)/ 18:2/18:2 (Gla/L/L) in blackcurrant seed oil by Ag-HPLC and different MS methods. A new ESI-MS/MS method for the regioisomeric structure determination of TAGs using silver ion adducts is introduced. This study provides a comparison of LC/ APCI-MS and LC-ESI-MS/MS methods in the determination of TAG regioisomerism and introduces a new modified Ag-HPLC solvent system containing nonchlorinated solvents for the separation of Gla and Ala containing TAGs. EXPERIMENTAL SECTION Abbreviations and Nomenclature. Ala is defined as R-linolenic acid [18:3(n-3)], Gla as γ-linolenic acid [18:3(n-6)], and L as linoleic acid 18:2(n-6). Ln is defined as linolenic acid, when doublebond positional isomerism is not considered. A/A/B denotes a TAG containing two different fatty acids A and B. Regioisomers are written as sn-ABA and sn-AAB + sn-BAA. No distinction is made between the sn-1 and sn-3 positions. DAG ions [M + H RCOOH]+ and [M + Ag - RCOOH]+ are defined also as [AA]+ and [AB]+, where A and B denote different fatty acids. Materials. Seed oil of blackcurrant (Ribes nigrum L.) was extracted from the ground seeds by supercritical CO2 (Aromtech Ltd., Tornio, Finland). Reference TAGs sn-18:2(n-6)-18:2(n-6)-18: 3(n-6) + sn-18:3(n-6)-18:2(n-6)-18:2(n-6), sn-18:2(n-6)-18:3(n6)-18:2(n-6), sn-18:2(n-6)-18:2(n-6)-18:3(n-3) + sn-18:3(n-3)-18: 2(n-6)-18:2(n-6), and sn-18:2(n-6)-18:3(n-3)-18:2(n-6) were purchased from Larodan Fine Chemicals (Malmo¨, Sweden). All solvents were HPLC grade and were used without further purification. Ammonium nitrate (p.a. grade) was purchased from Merck (Darmstadt, Germany) and silver nitrate (>99%) from Sigma-Aldrich GmbH. (Steinheim, Germany). Adlof, R. O.; Emken, E. A. J. Chromatogr., A 1994, 685, 178–181. Adlof, R. O. J. Chromatogr. 1994, 659, 95–99. Adlof, R.; Lamm, T. J. Chromatogr., A 1998, 799, 329–332. Adlof, R. O. J. Chromatogr., A 2004, 1033, 369–371. Adlof, R. O. J. High Resolut. Chromatogr. 1995, 18, 105–107. Adlof, R. O.; List, G. R. J. Agric. Food Chem. 2003, 51, 2096–2099. Adlof, R. O.; Menzel, A.; Dorovska-Taran, V. J. Chromatogr., A 2002, 953, 293–297. (34) Adlof, R. O. J. Chromatogr.,, A 1996, 741, 135–138. (35) Elfman-Bo ¨rjesson, I.; Van Den Hark, S.; Ha¨rro ¨d, M. J. Am. Oil Chem. Soc. 1997, 74, 1177–1180. (36) Fevrier, P.; Binet, A.; Dufosse, L.; Gree, R.; Yvergnaux, F. J. Chromatogr., A 2001, 923, 53–57. (27) (28) (29) (30) (31) (32) (33)
Analytical Chemistry, Vol. 80, No. 15, August 1, 2008
5789
Fractionation of Blackcurrant Seed Oil. Blackcurrant seed oil was fractionated with RP-HPLC system using Discovery HS C18 guard column (20 mm × 4.0 mm, 5 µm particle size; Supelco Inc., Bellefonte, PA), Ascentis C18 column (250 mm × 4.6 mm i.d., 5 µm particle size; Supelco Inc.), and Hitachi L-6200 Intelligent Pump (Tokyo, Japan). Elution pattern of peaks was assured by evaporative light scattering detector (ELSD) (Sedex 75, S.E.D.E.R.E., Alfortville, France). The elution rate was 0.85 mL/min and the linear gradient used was as follows: acetone in acetonitrile was increased from 40% to 90% in 70 min. Eight fractions were collected of which fraction 4 (elution time 23-29 min) contained TAGs of ACN:DB species 54:7. In addition to TAGs 54:7, small amounts of 54:6, 54:8, 52:6, and 52:7 were eluted in fraction 4. Silver Ion High-Performance Liquid Chromatography. The Ag-HPLC column was prepared according to the procedure introduced by Christie.26,37 The Ag-HPLC system consisted of silver ion bonded EC 250/4.6 Nucleosil 100-5 SA column (250 mm × 4.6 mm i.d.; Macherey-Nagel GmbH & Co. KG, Du¨ren, Germany) and Acquity Ultra Performance LC equipment (Waters Corp., Milford, MA). During the analyses, the flow rate was 0.8 mL/min of which ∼0.3 mL was introduced to MS. A binary solvent gradient, modified from the method of Christie37 and Laakso and Voutilainen,10 consisted of (A) acetone and (B) acetone/acetonitrile (4:1, v/v). The gradient program was as follows: initial A/B (100:0), held isocratic for 5 min; linear from 5 to 30 min to A/B (70:30); linear from 30 to 40 min to A/B (47:53); linear from 40 to 41 min to A/B (0:100); held isocratic for 8 min. The column was kept at room temperature 21 °C. (The atmosphere and the temperature of the laboratory was controlled and checked daily.) Analyses of reference TAGs Ala/L/L and Gla/L/L with Ag-HPLC/ APCI-MS, Ag-HPLC/ESI-MS, and Ag-HPLC/ESI-MS/MS were conducted in quadruplicate as pure regioisomers and three regioisomer mixtures. The three mixtures were prepared at the following sn-ABA/(sn-AAB + sn-BAA) ratios: 25/75, 50/50, and 75/25. Both reference TAGs, Ala/L/L and Gla/L/L, were analyzed and injected onto the column simultaneously. Also, duplicate analyses of sn-LAlaL, sn-AlaLL + sn-LLAla, sn-LGlaL, and sn-GlaLL + sn-LLGla injected separately were performed in order to check the possible differences in the elution times and DAG ion intensities. Mass Spectrometry. Ag-HPLC/APCI-MS, Ag-HPLC/ESI-MS, and Ag-HPLC/ESI-MS/MS analyses were performed with a Micromass Quattro Premier tandem quadrupole mass spectrometer (Waters Corp., Milford, MA). When using APCI, the probe and source temperatures were set to 600 and 150 °C, respectively, and corona current was set to 4 µA. Nitrogen was used as the desolvation and cone gases (150 and 50 L/h, respectively). Fullscan MS spectra (m/z 500-1000) were collected in positive ionization mode. In ESI mode, the capillary was set to 5 kV and the source and desolvation temperatures to 100 and 150 °C, respectively. Desolvation and cone gas flows were set to 550 and 50 L/h, respectively. Full-scan mass spectra (m/z 500-1000) were collected in positive ionization mode by ESI-MS. In ESI-MS/MS analyses, collision gas flow was set to 0.35 mL/min and collision energy to 39 eV. AgNO3 (4 µg/mL in water) was introduced 6 µL/min by syringe pump to postcolumn flow prior to MS in order to produce silver ion adducts of TAGs [M + Ag]+ by CIS. The (37) Christie, W. W. J. Chromatogr. 1988, 454, 273–284.
5790
Analytical Chemistry, Vol. 80, No. 15, August 1, 2008
Figure 1. Ag-HPLC/APCI-MS total ion chromatograms of the reference triacylglycerols (A) sn-LGlaL (peak 1) and sn-LAlaL (peak 2) and (B) sn-GlaLL + sn-LLGla (peak 1) and sn-AlaLL + sn-LLAla (peak 2). Triacylglycerols were injected separately, and the chromatograms are displayed as overlaid. Abbreviations: Gla, γ-linolenic acid; Ala, R-linolenic acid; and L, linoleic acid.
daughter ion spectra (m/z 500-1000) of the two precursor ions [M + 107Ag]+ and [M + 109Ag]+ were collected. The natural proportions of isotopes [M + 107Ag]+ and [M + 109Ag]+ are 51.84 and 48.16%, respectively. The proportions of the DAG ions [AA]+ and [AB]+ in ESIMS, ESI-MS/MS, and APCI-MS analyses were determined from each pure regioisomer and regioisomer mixture, and calibration curves were plotted. In the ESI-MS mass spectra, the m/z ratios of [LL + 107Ag] and [LLn + 109Ag]+ were overlapping (m/z 705.4), and the M + 2 isotopes of the DAG ions complicated the interpretation. In order to make the calculations easier, the DAG ion intensity ratios were calculated in ESI-MS as follows: m/z 703.4/705.4, 705.4/707.4, and 703.4/707.4. In ESI-MS/MS, two calibration curves were produced: one based on DAG ions [LLn + 107Ag]+ and [LL + 107Ag] (m/z 703.4 and 705.4) and one based on [LLn + 109Ag]+ and [LL + 109Ag] (m/z 705.4 and 707.4). In APCI-MS, the M + 2 isotope of DAG fragment [LnLn]+ formed from TAG 16:0/18:3/18:3 was overlapping with DAG fragment [LnL]+ (m/z 597.5) and this was taken into account. The relative proportions of regioisomer sn-ABA (100 * sn-ABA/(sn-ABA + snAAB + sn-BAA), %) of the analyzed TAGs in blackcurrant seed oil were calculated based on the corresponding ratios of the DAG ion intensities and calibration curves. Statistical Methods. SPSS 14.0 for Windows (Chicago, IL) was used for data analysis. Normal distribution of the of data was tested with the Shapiro-Wilk test, and statistical differences between the regioisomeric structures were compared using oneway ANOVA and Tukey’s HSD-test as post hoc test. RESULTS AND DISCUSSION Silver Ion High-Performance Liquid Chromatography. Figure 1 shows the overlaid Ag-HPLC/APCI-MS total ion chromatograms (TICs) of the reference TAGs Gla/L/L and Ala/L/L, which were injected separately. Figure 2 shows the Ag-HPLC/ APCI-MS TIC of blackcurrant seed oil. The identifications of the peaks are given in Table 1. TAGs were identified according to the [M + H]+ and DAG ions in APCI-MS mode and according to
Figure 2. Ag-HPLC/APCI-MS total ion chromatogram of blackcurrant seed oil. Peak numbers refer to Table 1. Table 1. Identified Triacylglycerols in Blackcurrant Seed Oil by Ag-HPLC/APCI-MSa peak number ACN:DBb 1 2 3 4 5 6 7 8 9, 10 11 12 13 14 15 16 b
52:2 54:3 52:3, 54:3 54:4 52:4, 54:4 54:5, 56:5 54:5, 54:6 54:6, 52:5 54:6, 56:6, 54:7 54:7, 52:6 54:7, 52:6 54:7 54:8 54:9 54:10
triacylglycerol 16:0/18:1/18:1, 16:0/18:2/18:0 18:1/18:1/18:1 16:0/18:2/18:1, 18:2/18:1/18:0 18:2/18:1/18:1 16:0/18:2/18:2, 18:2/18:2/18:0 18:2/18:2/18:1, 18:2/18:2/20:1 18:3/18:1/18:1, 18:3/18:2/18:1 18:2/18:2/18:2, 16:0/18:3/18:2 18:3/18:2/18:1, 18:3/18:2/20:1, 18:3/18:2/18:2 18:3(n-6)/18:2/18:2, 16:0/18:3/18:3 18:3(n-3)/18:2/18:2, 16:0/18:3/18:3 18:3/18:3/18:1, 18:3/18:3/18:2 18:3/18:3/18:2 18:4/18:3/18:2 18:4/18:3/18:3
a The most abundant triacylglycerols in a peak are in boldface type. ACN, acyl carbon number; DB, number of double bonds.
the [M + Ag]+ and DAG ions in ESI-MS and ESI-MS/MS modes. The elution pattern of blackcurrant seed oil was similar to that of Laakso and Voutilainen.10 The most abundant TAGs in blackcurrant seed oil were 18:2/18:2/18:1 (peak 6 in Figure 2), 18:2/18: 2/18:2 (peak 8), 18:3/18:2/18:1 (peaks 9 and 10), 18:3(n-6)/18: 2/18:2 (peak 11), 18:3(n-3)/18:2/18:2 (peak 12) and 18:3/18:3/ 18:2 (peaks 13 and 14), which is consistent with the results of Laakso and Voutilainen10 and Kallio et al.38 The elution times of Ala/L/L and GLA/L/L obtained by our Ag-HPLC system were comparable to those obtained by the AgHPLC system of Laakso and Voutilainen.10 Gla/L/L was less retained on the column than Ala/L/L. Also, Christie37 noted that TAG fractions containing Gla tended to elute before those containing Ala, and Nikolova-Damyanova et al.39 reported that Gla eluted before Ala as methyl esters. The elution order was confirmed by injecting the reference compounds Ala/L/L and Gla/L/L separately. Gla/L/L and Ala/L/L were not baseline separated, but the separation was enough to perform the regioisomeric structure determination. The separation of Gla/L/L and Ala/L/L appears to be slightly better with the chromatographic system of Laakso and Voutilainen10 compared to ours. Adlof and List40 found that temperature affected the elution times of TAGs in Ag-HPLC (slower elution at higher temperatures). We did not have the (38) Kallio, H.; Currie, G.; Gibson, R.; Kallio, S. Ann. Chim. 1997, 87, 187– 198. (39) Nikolova-Damyanova, B.; Herslof, B. G.; Christie, W. W. J. Chromatogr. 1992, 609, 133–140. (40) Adlof, R.; List, G. J. Chromatogr., A 2004, 1046, 109–113.
opportunity to study the effects of different column temperatures to elution times and separation. sn-GlaLL + sn-LLGla was retained slightly longer in the column and eluted ∼15 s later than sn-LGlaL, but eluted still in the same chromatographic peak. This was confirmed when regioisomers sn-LGlaL and sn-GlaLL + sn-LLGla were injected separately in consecutive runs. The elution patterns of these runs can be seen in Figure 1. The first peak sn-GlaLL + sn-LLGla in Figure 1B elutes later than peak sn-LGlaL in Figure 1A and overlaps sligthly more with the second peak (Ala/L/L). The phenomenon could be seen in all of the reference TAG runs. It did not pose a problem to the regioisomeric structure determination of Gla/L/L and Ala/L/L, because when preparing the calibration curves, the reference TAGs Gla/L/L and Ala/L/L were injected simultaneously and the area of the peak was selected in the same manner in reference TAGs and blackcurrant seed oil TAGs. The separation of regioisomers is possible by Ag-HPLC,10,31,32,36 and Adlof31 concluded that the location of the FA in the TAG is also a factor when considering the elution order of the TAGs. The DBs in FAs located on the sn-1/3 position interact with silver ions more efficiently than the DBs in secondary FAs. This phenomenon is perceived also in this study with regioisomers of Gla/L/L, although separation is minor and not seen with Ala/L/L. In the study of Adlof,31 where a dual-column system was used, sn-LLO + sn-OLL and sn-LOL (where O denotes oleic acid) were separated incompletely with retention times of over 40 min, and the separation of sn-LnLnL + LLnLn and sn-LnLLn with retention times near 70 min was even poorer. Adlof31 admitted that TAG isomer pairs become more difficult to separate when the total number of DBs in the TAG exceeds 4 and if the number of DBs in FA A and FA B differs by only one (e.g., LLO and LOL). Also, Fe´vrier et al.36 failed to separate CLL and LCL (where C denotes caprylic acid) with one silver ion column and mobile phases containing n-hexane, 2-propanol, ethyl acetate, and acetonitrile. ESI-MS and ESI-MS/MS. The column did not leach silver ions to a significant extent, and therefore, in order to produce silver ion adducts in CIS, aqueous silver nitrate was pumped by syringe pump to the LC flow prior to MS. The ESI-MS and ESIMS/MS mass spectra of TAGs exhibited [M + 107Ag]+ and [M + 109Ag]+ ions as well as [M + 107Ag - RCOOH]+, [M + 109Ag - RCOOH]+, and [M - RCOO]+ ions. Intensities of the DAG ions [M + Ag - RCOOH]+ were dependent on the positional distribution of FAs in TAGs. The Ag-HPLC/ESI-MS/MS spectra of regioisomers sn-AlaLL + sn-LLAla and sn-LAlaL are shown in Figure 3. The spectra comprised mainly DAG product ions [M + Ag - RCOOH]+, but in ESI-MS, the amounts of [M + Ag RCOOH]+ ions were less abundant than [M + Ag]+ ions. The regioisomeric structure determination of TAGs in ESI-MS/MS was Analytical Chemistry, Vol. 80, No. 15, August 1, 2008
5791
Figure 3. Ag-HPLC/ESI-MS/MS spectra of regioisomers (A) snLAlaL and (B) sn-AlaLL + sn-LLAla using parent ion m/z 985.5 ([M + 109Ag]+). The ions corresponding to m/z 705.3 and 707.3 are diacylglycerol product ions [M + 109Ag - L]+ and [M + 109Ag - Ala]+, respectively. The ions corresponding to m/z 597.7 and 599.7 are diacylglycerol product ions [M + H - L]+ and [M + H - Ala]+, respectively. See abbreviations in Figure 1.
Figure 4. Ag-HPLC/APCI-MS spectra of regioisomers (A) sn-GlaLL + sn-LLGla and (B) sn-LGlaL. The ions corresponding to m/z 597.7 and 599.7 are diacylglycerol fragment ions [M + H - L]+ and [M + H - Gla]+. The ion corresponding to m/z 877.9 is [M + H]+. See abbreviations in Figure 1.
Figure 5. Calibration curves obtained by Ag-HPLC/ESI-MS/MS using [M + 107Ag]+ as the parent ion. The ratio of diacylglycerol ions [LnL]+ and [LL]+ is determined in the mixtures of sn-LLnL and sn-LnLL + sn-LLLn. sn-LLnL content is calculated as follows: 100 × sn-LLnL/ (sn-LLnL + sn-LnLL + sn-LLLn). (A) Ala/L/L, y ) 0.0172x + 1.6812 and (B) Gla/L/L, y ) 0.0099x + 1.4627. Four parallel analyses were conducted from each regioisomeric mixture. Ln, linolenic acid; see other in Figure 1.
calculated based on the [M + Ag - RCOOH]+ ions, because the intensities of those were higher than those of [M - RCOO]+ ions. The regioisomeric structure determination by ESI-MS was complicated because the overlapping DAG ions in spectra (e.g., [LL + 107Ag]+ and [LnL + 109Ag]+) made the calculations more difficult and prone to errors. Also TAG 16:0/18:3/18:3 with DAG ion [LnLn + 109Ag]+ was overlapping with [LLn + 107Ag]+, and M + 2 ion of [LnLn + 109Ag]+ was overlapping with [LLn + 109Ag]+ and [LL + 5792
Analytical Chemistry, Vol. 80, No. 15, August 1, 2008
Figure 6. Calibration curves obtained by Ag-HPLC/ESI-MS/MS using [M + 109Ag]+ as the parent ion. The ratio of diacylglycerol ions [LnL]+ and [LL]+ is determined in the mixtures of sn-LLnL and sn-LnLL + sn-LLLn. sn-LLnL content is calculated as follows: 100 × sn-LLnL/ (sn-LLnL + sn-LnLL + sn-LLLn). (A) Ala/L/L, y ) 0.0174x + 1.7584 and (B) Gla/L/L, y ) 0.0102x + 1.5481. Four parallel analyses were conducted from each regioisomeric mixture. Ln, linolenic acid; see other in Figure 1.
Figure 7. Calibration curves obtained by Ag-HPLC/APCI-MS. The ratio of diacylglycerol ions [LnL]+ and [LL]+ is determined in the mixtures of sn-LLnL and sn-LnLL + sn-LLLn. sn-LLnL content is calculated as follows: 100 × sn-LLnL/(sn-LLnL + sn-LnLL + sn-LLLn). (A) Ala/L/L, y ) 0.0065x + 1.2797 and (B) Gla/L/L, y ) 0.0044x + 0.6476. Four parallel analyses were conducted from each regioisomeric mixture. Ln, linolenic acid; see other in Figure 1.
Ag]+ in the ESI-MS spectra. It is questionable whether the coeluting TAGs affect the fragmentation of particular TAGs in ESI ionization or not and whether this has any significant effect on the proportions of the DAG fragment ions. Because of the considerable dispersion of the results of parallel analyses, the ESIMS results were not considered to be as reliable as the results of ESI-MS/MS and APCI-MS. ESI-MS/MS is also more appropriate to use than ESI-MS, because TAGs with different m/z values can be analyzed separately by tandem mass spectrometry. APCI-MS. The Ag-HPLC/APCI-MS spectra of regioisomers sn-GlaLL + sn-LLGla and sn-LGlaL are shown in Figure 4. The mass spectra exhibited [M + H]+ ions and DAG fragment ions [M + H - RCOOH]+, which were formed depending on the positional distribution of FAs in TAG as reported previously.11,14,16 The small amount of M + 2 isotope of DAG ion [LnLn]+ from TAG 16:0/18:3/18:3 had to be taken into account. As in ESI-MS, it is not known whether the coeluting TAGs affect the fragmentation of particular TAGs or not. Calibration Curves. Discrimination of regioisomers was obvious with all the methods applied, which can be observed in different intensities of DAG ions. The calibration curves of Ala/ L/L and Gla/L/L determined by Ag-HPLC/ESI-MS/MS using precursor ions [M + 107Ag]+ and [M + 109Ag]+ are plotted in Figures 5 and 6, respectively. The Ag-HPLC/APCI-MS calibration curves are plotted in Figure 7. The calibration curves of ESI-MS are not shown. 107
Table 2. Regioisomerism of Ala/L/L and Gla/L/L in Blackcurrant Seed Oil Determined by Ag-HPLC/APCI-MS and Ag-HPLC/ESI-MS/MSa TAG regioisomer
APCI-MS
ESI-MS/MSb
ESI-MS/MSc
n ) 3d
n ) 12e
sn-LAlaL sn-AlaLL + sn-LLAla sn-LGlaL sn-GlaLL + sn-LLGla
12.1 ± 1.8 * 87.9 ± 1.8 * 37.8 ± 4.7 62.2 ± 4.7
7.3 ± 1.3 * 92.7 ± 1.3 * 32.7 ± 1.6 67.3 ± 1.6
10.0 ± 1.0 90.0 ± 1.0 34.2 ± 1.9 65.8 ± 1.9
9.8 ± 2.4 90.2 ± 2.4 34.9 ± 2.6 65.1 ± 2.6
9.8 ± 2.4 90.2 ± 2.4 34.9 ± 3.6 65.1 ± 3.6
a Results in a row marked with asterisk * are statistically significantly different p
