Anal. Chem. 2008, 80, 1664-1678
MALDI Seamless Postsource Decay Fragment Ion Analysis of Sodiated and Lithiated Phospholipids Gerald Stu 1 biger, Ernst Pittenauer, and Gu 1 nter Allmaier*
Institute of Chemical Technologies and Analytics, Vienna University of Technology, Vienna, Austria
In this paper, we present the results of a detailed study using MALDI seamless postsource decay (sPSD) fragment ion analysis of all major glycerophospholipid (GPL) classes, cardiolipin (bisphosphatidylglycerol), and sphingomyelin, respectively. The matrix compound 2,4,6trihydroxyacetophenon recently introduced for lipid analysis (Stu1 biger, G.; Belgacem O. Anal. Chem. 2007, 79, 3206-3213) was applied in conjunction with a highresolution curved field reflectron analyzer allowing detection of the fragment ions without stepping the reflectron voltages (seamless PSD). This instrumental feature helps to define in a fast way the polar headgroups of the different GPL classes and gives information about the constituent fatty acid residues dependent on the type of adduct ion used. Our experiments demonstrate that fragment ions specifying the fatty acid composition of the lipid molecules could only be generated from cationized molecular ions (sodiated or lithiated). Additionally, information about the stereospecificity of the fatty acid residues on the glycerol backbone (sn-1, and -2 position) of particular GPLs could be obtained during sPSD analysis. In the case of phosphatidylcholine, significant fragmentation related to the loss of the acyl groups could only be observed from [M + Li]+ ions. Generally, alkali adduction (sodium and lithium) enhanced fragmentation of most lipid classes, particularly favoring fragment ions associated with the polar headgroups. Glycerophospholipids (GPL) are the most important class of structural lipids in eucaryotic cells. The outer membranes and the subcellular organelles are composed of a bilayer of various GPL classes. The fatty acid composition of these molecules plays an essential role in normal cell functions and is mainly influenced by environmental and endogenous factors such as infection,1,2 oxidative stress,3 physiological disorders, and nutrition.4 They also play an important role in cell growth and development by interaction with different enzymes, hormones, and neurotrans* To whom correspondence should be addressed. E-mail: guenter.allmaier@ tuwien.ac.at. Telephone: +43 1 58801 15160. Fax: +43 1 58801 15199. (1) Panjwani, N.; Zhao, Z.; Raizman, M. B.; Jungalwala, F. Infect. Immun. 1996, 64, 1819-1825. (2) Abrahamsen, L. H.; Clay, M. J.; Lyle, J. M.; Zink, J. M.; Fredrikson, L. J.; DeSiervo, A. J.; Jerkofsky, M. A. Intervirology 1997, 39, 223-229. (3) Cross, C. E.; Halliwell, B.; Borish, E. T.; Pryor, W. A.; Ames, B. N.; Saul, R. L.; McCord, J. M.; Harmann, D. Ann. Intern. Med. 1987, 107, 526-545. (4) Zeisel, S. H.; Blusztajn, J. K. Choline and Human Nutrition. Annu. Rev. Nutr. 1994, 14, 269-296.
1664 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
mitters. Hydrolysis by phospholipases produces different important secondary messengers such as 1,2-diacylglycerol, inositol 1,4,5-trisphosphate, or sphingosine and ceramide derived from sphingolipids.5,6 Due to these facts, the analysis of phospholipids is an important subject of lipid analysis in general. GPL consists of a lipid part and polar headgroups linked to the glycerol backbone (Figure 1). The smallest member of this group is phosphatidic acid (PA) where the headgroup only consists of the phosphate ester, while the largest is cardiolipin representing a dimeric form of PG consisting of two diacylglycerolphosphate moieties linked together by glycerol. In contrast, sphingomyelin (SM) contains the same headgroup as PC (cholinephosphate) but linked to ceramide instead of glycerol (sphingolipids). SM contains only one fatty acid residue bound to sphingosine (a long-chain base, usually containing 18 carbon atoms) as an amide (Figure 1). Chemically, the different GPL classes are defined by their polar headgroups, whereas the individual PL species are defined only by their content of the fatty acid residues esterified at the sn-1, and -2 hydroxyl groups of glycerol (Table 1). In biological samples, GPL exhibit a characteristic stereospecificity showing often a preferred location of unsaturated fatty acid residues at the sn-2 position (e.g., human plasma phospholipids). Mass spectrometry can be considered as an universal tool in the analysis of the various lipid classes in biological samples and is becoming a indispensable tool in the emerging field of “lipidomics”.7,8 Several so-called “soft” desorption/ionization techniques have been used extensively such as desorption chemical ionization and fast atom bombardment,9,10 plasma desorption,11 and electrospray ionization (ESI) in positive and negative ionization modus.12-16 Matrix-assisted-laser-desorption/ionization mass spectrometry (MALDI-MS) combined with a time-of-flight (TOF) (5) Ashcroft, J. S. Adv. Exp. Med. Biol. 1997, 426, 73-80. (6) Hunnan, Y. A. Adv. Exp. Med. Biol. 1997, 400A, 305-312. (7) Han, X.; Gross, R. W. J. Lipid Res. 2003, 44, 1071-79. (8) Schiller, J.; Suess, R.; Fuchs, B.; Mu ¨ ller, M.; Zschornig, O.; Arnold, K. Front. Biosci. 2007, 12, 2568-2579. (9) Ayanoglu, E.; Wegmann, A.; Pilet, O.; Marbur, G. D.; Hass, J. R.; Djerassi, C. J. Am. Chem. Soc. 1984, 106, 5246-5251. (10) Murphy, R. C.; Fiedler, J.; Hevko, J. Chem. Rev. 2001, 101, 479-526. (11) Pittenauer, E.; Schmid, E. R.; Allmaier, G.; Puchinger, L.; Kienzl, E. Eur. Mass Spectrom. 1996, 2, 247-262. (12) Kerwin, J. L.; Tuininga, A. R.; Ericsson, L. H. J. Lipid Res. 1994, 35, 11021114. (13) Smith, P. B. W.; Snyder, A. P.; Harden, C. S. Anal. Chem. 1995, 67, 18241830. (14) Byrdwell, W. C. Rapid Commun. Mass Spectrom. 1998, 12, 256-272. (15) Koivusalo, M.; Haimi, P.; Heikinheimo, L.; Kostiainen, R.; Somerharju, P. J. Lipid Res. 2001, 42, 663-672. 10.1021/ac7018766 CCC: $40.75
© 2008 American Chemical Society Published on Web 01/30/2008
Figure 1. Structures of the main phospholipid classes incuded in this study. R1, R2 are fatty acid residues bound to the sn-1 and -2 hydroxyl groups of the glycerol. R3 represents the polar headgroups esterified to the glycerol phosphate backbone (sn-3 hydroxyl). In the case of SM, the fatty acid (R1) is linked via an amide of the sphingosine moiety to form a ceramide.
analyzer has been used to a lesser extent in the field of lipid analysis.17-24 MALDI offers a number of advantages compared to other techniques, especially the sample preparation can be easily performed, exhibits a relative high tolerance to impurities (e.g., salts and detergents) particularly compared to ESI, and shows an excellent sensitivity in absolute terms.25 An important advantage of MALDI in lipid analysis is that the lipid sample and most of the common matrices are readily soluble in organic solvents resulting in a homogeneous crystallization of the analyte/matrix mixture on the MALDI target. This leads to reproducible and wellresolved MALDI mass spectra.20 In the case of using a reflectron (RTOF) mass spectrometer, detailed structural information of lipids can be obtained by postsource decay (PSD) fragment ion analysis26,27 or collision-induced dissociation (CID) using the latest generation of MALDI-TOF/RTOF mass spectrometers.28 It turned out that MALDI-PSD analysis provides an excellent tool for (16) Ekroos, K.; Ejsing, C. S.; Bahr, U.; Karas, M.; Simons, K.; Shevchenko, A. J. Lipid Res. 2003, 44, 2181-2192. (17) Marto, J. A.; White, F. M.; Seldomridge, S.; Marshall, A. G. Anal. Chem. 1995, 67, 3979-3984. (18) Zo¨llner, P.; Schmid, E. R.; Allmaier, G. Rapid Commun. Mass Spectrom. 1996, 10, 1278-1282. (19) Zabrouskov, V.; Al-Saad, K. A.; Siems, W. F.; Hill, H. H., Jr.; Knowles, N. R. Rapid Commun. Mass Spectrom. 2001, 15, 935-940. (20) Schiller, J.; Arnhold, J.; Benard, S.; Muller, M.; Reichl, S.; Arnold, K. Anal. Biochem. 1999, 267, 46-56. (21) Li, Y. L.; Gross, M. L.; Hsu, F. F. J. Am. Soc. Mass Spectrom. 2005, 16, 679-82. (22) Estrada, R.; Yappert, M. C. J. Mass Spectrom. 2004, 39, 412-22. (23) Stu ¨ biger, G.; Belgacem, O. Anal. Chem. 2007, 79, 3206-3213. (24) Stu ¨ biger, G.; Pittenauer, E.; Allmaier, G. Phytochem. Anal. 2003, 14, 337346. (25) Schiller, J.; Su ¨ β, R.; Arnhold, J.; Fuchs, B.; Leβig, J.; Mu ¨ ller, M.; Petkovic, M.; Spalteholz, H.; Zscho ¨rnig, O.; Arnold, K. Prog. Lipid Res. 2004, 43, 449-488. (26) Al-Saad, K. A.; Siems, W. F.; Hill, H. H.; Zabrouskov, V.; Knowles, N. R. J. Am. Soc. Mass Spectrom 2003, 14, 373-382. (27) Fuchs, B.; Schober, C.; Richter, G.; Su ¨ β, R.; Schiller, J. J. Biochem. Biophys. Methods 2007, 70, 689-692.
structural elucidation of the diverse lipid classes after isolation from biological samples.29 In this paper, we describe the results of seamless PSD (sPSD) fragment ion analysis of different GPL classes obtained by means of a vacuum MALDI-TOF instrument. The application of a curved field reflectron (CFR) analyzer30 for PSD experiments allows the simultaneous detection of all generated fragment ions during one experiment without the need for scanning or stepping the reflectron voltage (sPSD). Obviously, reservation for applying PSD analysis (mentioned in the literature) can be attributed to the need of an unusual high number of laser pulses and thus time as well as sample consumption by the use of classical reflectrons using particular types of instruments.27,31,32 In contrast, seamless PSD makes the structural elucidation quite easy and moreover improves sensitivity of detection and quality of spectra. In our case, this particular instrumental feature was exploited for the analysis of positional isomers of various GPL classes. We demonstrate that isomeric GPL could be characterized based on relative intensities of the fragment ions resulting from the loss of the acyl groups from the sn-1 and -2 position of the glycerol backbone, respectively. Additionally, sPSD analysis revealed quite different fragmentation pathways of differentially cationized GPL species (sodiated or lithiated). EXPERIMENTAL SECTION Chemicals and Standards. Phospholipid standards (Table 1) were all purchased from Sigma (St. Louis, MO) and were used without further purification. All solvents and sodium acetate were obtained from Merck (Darmstadt, Germany). Lithium acetate was obtained from Aldrich (Milwaukee, WI), whereas 2,4,6-trihydroxyacetophenon (THAP) and diammonium hydrogen citrate were obtained from Fluka (Buchs, Switzerland). Ultrapure water (18 MΩ-1 cm-1) was prepared by an EASY pure LF apparatus (Barnstead/Thermolyne, Dubuque, IO). All chemicals and lipid standards were obtained in the highest available purity. Sample Preparation. THAP matrix was prepared as a methanolic solution with a final concentration of 20 mM. Lipid samples were dissolved in chloroform/ methanol (1:1, v/v), pure methanol, or a mixture of methanol/water (4:1, v/v) in the case of PI. Usually 0.3-0.5 µL of sample (100 pmol/µL) directly followed by an equal volume of matrix solution was applied on the target (“dried droplet”). In the case of the aqueous PI solution, a thin-layer preparation was applied. The matrix solution was deposited on the target first, and after evaporation of the matrix solvent, a droplet (0.3 µL) of sample solution was put onto the matrix layer, followed by drying under a stream of warm air. The crystallization occurs immediately after evaporation of the solvents, leading to a very fine analyte/matrix film on the target surface. For propagating the generation of certain types of adduct ions, solutions of alkali or ammonium salts were mixed with THAP matrix solution to get final salt concentrations between 10 and 50 mM (“salt doping technique”).23 (28) Jackson, S. N.; Wang, H. Y.; Woods, A. S. J. Am. Soc. Mass Spectrom. 2007, 18, 17-26. (29) Ham, B. M.; Jacob, J. T.; Cole, R. B. Anal. Chem. 2005, 77, 4439-47. (30) Belgacem, O.; Bowdler, A.; Brookhouse, I.; Brancia, F. L.; Raptakis, E. Rapid Commun. Mass Spectrom. 2006, 20, 1653-60. (31) Spengler, B. J. Mass Spectrom. 1997, 32, 1019-1036. (32) Cordero, M. M.; Cornish, T. J.; Cotter, R. J.; Lys, I. A. Rapid Commun. Mass Spectrom. 1995, 9, 1356-1361.
Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
1665
Table 1. Nomenclature of the Used Phospholipids Standardsa nomenclature (abbreviation)
Mr
R1
1-palmitoyl-2-oleoyl-glycero-3phosphatidylethanolamine (POPE) N-monomethyl-1,2-dipalmitoyl-glycero-3phosphatidylethanolamine (MMPE) N,N-dimethyl-1,2-dipalmitoyl-glycero-3phosphatidylethanolamine (DMPE) 1,2-dipalmitoyl-glycero-3phosphatidylcholine (DPPC) 1-stearoyl-2-docosahexaenoyl-glycero-3phosphatidylcholine (SDPC) 1,2-dipalmitoyl-glycero-3-phosphatidylserine (DPPS) 1,2-dipalmitoyl-glycero-3-phosphatidylglycerol (DPPG) 1-palmitoyl-2-oleoyl-glycero-3phosphatidylglycerol (POPG) 1-stearoyl-2-arachidonyl-glycero-3phosphatidicacid (SAPA) phosphatidylinositol (PI) (from bovine liver) 1,3-di-(3-phosphatidyl)-glycerol (cardiolipin) (from bovine heart) N-palmitoyl-sphingomyelin (SM) (from chicken egg yolk)
717.5
16:0
18:1
C2H4NH2 (ethanolamine)
705.5
16:0
16:0
C2H4NHCH3 (methylethanolamine)
719.5
16:0
16:0
C2H4N(CH3)2 (dimethylethanolamine)
733.6
16:0
16:0
C2H4N(CH3)3+ (choline)
833.6
18:0
22:6
735.5
16:0
16:0
CH2CH(NH2)COOH (serine)
722.5
16:0
16:0
CH2CH(OH)CH2OH (glycerol)
748.5
16:0
18:1
724.5
18:0
20:4
886.6
R2
mainly 18:0 and 20:4
1449.0
18:2
18:2
702.6
16:0
sphingosine (C18)
R3 (headgroup)
H (hydrogen) C6H5O5 (inositol) phosphatidylglycerol C2H4N(CH3)3+ (choline)
a R and R (fatty acid residues esterified at the sn-1, and -2 hydroxyl groups of glycerol), R headgroup esterified at the sn-3 phosphate group; 1 2 3 fatty acid residues (number of carbon atoms and of double bonds): palmitic acid (16:0), stearic acid (18:0), oleic acid (18:1), linoleic acid (18:2), arachidonic acid (20:4), and docosahexaenoic acid (22:6). Mr calculated molecular mass values of the neutral molecules.
Mass Spectrometry. MALDI mass spectra were obtained on an Axima CFR (Shimadzu Biotech, Manchester, UK) reflectron time-of-flight mass spectrometer equipped with a nitrogen laser (337 nm, 3-ns pulse width) in positive ion mode using delayed extraction. The delay time was set according to the analyzed precursor ion range of the samples (in general, 166 ns for m/z values between 500 and 1000). The optimal gate width for precursor ion selection for the seamless postsource decay (sPSD) experiments in the necessary m/z range was (5 Da. The ion acceleration voltage was set at 20 kV, and during all sPSD experiments, the ion source/analyzer pressure was typically 9.4 × 10-7 mbar or less. An integrated 1-GHz transient recorder was used, and sPSD spectra were acquired by averaging 150-400 laser profiles (each profile is corresponding to two laser pulses). Mass spectra were calibrated externally by means of a standard lipid mixture. No signal smoothing or baseline subtraction was performed except for the sPSD spectrum of cardiolipin. RESULTS AND DISCUSSION Our intention was (1) the identification of headgroup-related fragment ions, which can be used for mass spectrometric characterization of the different GPL classes (“diagnostic ions”), (2) the generation of fragment ions to identify the fatty acid composition of the individual molecules (GPL species), and (3) the differentiation of positional isomers. According to previous investigations, loss of the acyl groups was found to be dependent on the number of alkali cations attached to the lipid molecules (“degree of alkali ion saturation”).23 Therefore, during the actual study, it was our intention to use certain types of alkali adducts to serve as precursor ions for the subsequent sPSD analysis. In the following, the results of our investigation based on description of MALDI-sPSD spectra of selected GPL standards (see Table 1) are presented. An overview of characteristic fragment ions for 1666 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
identification of the different GPL classes (diagnostic ions) is given in Table 2. Additionally, a detailed comparison of fragment ions derived from sodiated and lithiated adduct ions is provided in the Supporting Information. Phosphatidylethanolamine. In the first set of experiments, sodiated molecules were selected as precursor ions for sPSD analysis. In Figure 2, the sPSD spectra of [M + Na]+ and [M H + 2Na]+ ions derived from POPE are shown. The spectrum of the precursor ion [M + Na]+ (m/z 740) shows four major fragment ions at m/z 697, 599, 577, and 164, which are all members of the headgroup (Figure 2a). Especially the loss of ethanolamine (possibly as vinylamine, ∆m ) 43u)33,34 as well as ethanolaminephosphate (∆m ) 141u) and the presence of sodiated ethanolaminephosphate (m/z 164) were found to be characteristic for the class of PE (Table 2). In contrast, the precursor ion [M - H + 2Na]+ (m/z 762) provides an even more complex sPSD spectrum (Figure 2b). The most characteristic fragments related to the headgroup were in this case the loss of ethanolamine (∆m ) 43u) and the detection of disodiated ethanolaminephosphate (m/z 186). Additionally, fragments derived from the headgroup at m/z 168 and 125 (most probably as amidophosphane structures) could be observed with low intensity. In contrast to [M + Na]+, no neutral loss of ethanolaminephosphate could be observed indicating another fragmentation pathway. Besides these headgroup-associated ions, in the range between m/z 400 and 550 fragment ions could be attributed to the loss of the acyl groups (m/z 480 and 506) and the acyl groups together with ethanolamine (m/z 415 and 441), respectively (see inset). These directly related fragment ion pairs show a significant intensity difference (roughly 1:2) between the losses of palmitic acid (16:0) bound at the sn-1 position to those (33) Han, X.; Gross, R. W. J. Am. Soc. Mass Spectrom. 1995, 6, 1202-1210. (34) Hsu, F.-F.; Turk, J. J. Mass Spectrom. 2000, 35, 596-606.
Table 2. Overview of Diagnostic Fragmentations (1) and Diagnostic Ions (2) of the Different GPL Classesa lipid class
[M+H]+
[M+Na]+
[M+Li]+
[M-H+2Na]+
[M-H+2Li]+
[M-2H+3Na]+
PE
141 (1)
n/a
43 (1) 186 (2) 168 (2)
no n/a
43 (1) 129 (1) (154) (2) 136 (2) yes (3 : 1) n/a
n/o
acyl loss MMPE
43 (1) 141 (1) (163) (1) 164 (2) no n/a
n/a
n/a
n/a
n/a
acyl loss DMPE
acyl loss PC
184 (2)
59 (1) 183 (1) (205) (1) 147 (2) (86) (2)
acyl loss PG
no n/a
no (154) (1) 172 (1) 194 (1) 195 (2)
acyl loss PS
n/a
acyl loss PA acyl loss PI
acyl loss cardiolipin
no 87 (1) (185) (1) 207 (1) 208 (2) no
(59) (1) 183 (1) 189 (1) (200) (2) (131) (2) (86) (2) yes (2 : 1) n/a
n/a
n/o
n/a
n/a
n/a
260 (1) (282) (1) 283 (2)
n/a
n/a
acyl loss
no 598 (1) 599 (2) (275) (2)
n/a
no
SM
184 (2)
acyl loss
no
59 (1) 183 (1) (147) (2) (86) (2) no
59 (1) 183 (1)
yes (2 : 1) 57 (1) 182 (2) 143 (2) (125) (2) yes (position not testable) 71 (1) 214 (2) (174) (2) (125) (2) 72 (2) yes (position not testable) n/o
74 (1) (217) (2) 199 (1) (177) (2) (125) (2) yes (2.5 : 1) 87 (1) (229) (1) (230) (2)
yes (position not testable) n/o
(74) (1) (154) (1) (160) (1) (185) (2) 167 (2) yes (5 : 1) n/a
yes (position not testable) no headgroup loss yes (1.5 : 1) 162 (1) (305) (2) 287 (2) 269 (2) yes (2.5 : 1) 598 (1) 617 (1) (740) (2) 599 (2) (297) (2) (278) (2) no n/o
71 (1) (151) (1) (182) (2) 72 (2)
no headgroup loss yes (2 : 1) n/a
n/a
n/o
n/o
n/o
n/o
n/o
87 (1) (109) (1) (253) (2) (235) (2) yes (position not testable) no headgroup loss yes (1 : 2) n/o
616 (1) (797) (1) (619) (2) 597 (2) (461) (2) yes (position not testable) n/o
no
a n/o, not observed; n/a, not analyzed. Low abundant signals are indicated in parentheses. It should be noted that diagnostic ions always show fixed m/z values while diagnostic fragmentations are nominal losses dependent on the mass of the precursor ion. Acyl group fragmentation and the approximate ratio of loss from sn-1: sn-2 positions (in parentheses) are indicated for each type of adduct ion, individually. In the case of identical fatty acid residues, the stereospecificity was not testable (adapted from ref 23).
of oleic acid (18:1) at the sn-2 position of the glycerol. These results clearly demonstrate that a differentiation of positional isomers of PE is possible by MALDI-sPSD fragment ion analysis. Additionally, two prominent signals in the lower mass range (m/z 327 and 301) could be observed corresponding to disodiated carboxylate cations from oleic and palmitic acid, respectively. These ions directly reflect the fatty acid composition of the corresponding PE species.
Such types of fragment ions were observed also in the sPSD spectra of MMPE, DMPE, and PG (see below). Comparison with actual results obtained by Fuchs et al.27 revealed striking similarities to our observations using another type of MALDI instrumentation (classical reflectron analyzer). In a second experiment, lithium acetate was used as additive of THAP matrix. Figure 2c shows the sPSD spectrum of the [M Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
1667
Figure 2. sPSD spectrum of POPE (16:0/18:1) derived from (a) the precursor ion [M + Na]+ (m/z 740.5), (b) [M - H + 2Na]+ (m/z 762.5), and (c) [M - H + 2Li]+ (m/z 730.6). Most prominent diagnostic fragmentations are related to the direct loss of ethanolamine (43u) and ethanolaminephosphate (141u) in the case of the mono- and disodiated molecules and lithiated amidophosphane (129u) using the dilithiated molecules, respectively. Fragment ions related to the loss of the fatty acid residues are only observed using disodiated or dilithiated molecules (insets). Thereby, relative signal intensity ratios show a preferable loss of 16:0 (sn-1) over 18:1 (sn-2). Moreover peaks at m/z 301 and 327 represent disodiated carboxylates (for details, see text). 1668 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
- H + 2Li]+ precursor ion (m/z 730). In contrast to the sodiated molecules, very intense fragments due to headgroup losses could be observed. The fragment ion at m/z 687 corresponds to the loss of vinylamine, and the most abundant ion at m/z 601 results from the neutral loss of lithiated ethanolaminephosphate (as amidophosphane, ∆m ) 129u). No corresponding type of ion was observed using [M - H + 2Na]+ as precursor. In contrast to the headgroup, very similar fragmentation behavior was observed for the loss of the fatty acids residues. The [M - H + 2Li - (RCOOLi) - 43]+ ions exhibit a two times more intense signal from the loss of the acyl moiety in sn-1 (m/z 425) than sn-2 position (m/z 399) similar to the sPSD spectrum originating from a disodiated precursor molecule (see Figure 2b). Interestingly, only very low abundant signals from the loss of the acyl groups as free fatty acids were found. The weak signal at m/z 474 corresponds to the [M - H + 2Li (RCOOH)]+ ion related to the loss of palmitic acid (not shown). These data indicate a different fragmentation behavior of the sodiated and lithiated precursor ions related to loss of the fatty acid residues. In the case of dilithiated molecules, the fragmentation reaction seems to involve the headgroup, preferentially. Two very low intensity signals could be detected at m/z 339 and 314. According to previous investigations,34 these ions were identified as [M - H + 2Li - 129 - (RCOOLi)]+, related to the consecutive decomposition of the fragment ion m/z 601 (neutral loss of lithiated amidophosphane) by release of lithiated carboxylates of the fatty acids. This fragmentation pathway seems to be only relevant for dilithiated molecules. In contrast, the fragment ions corresponding to dilithiated amidophosphane (m/z 136) as well as ethanolaminephosphate (m/z 154) were in agreement with the sPSD spectra using the disodiated precursors (see Supporting Information, Table S-1). Mono- and Dimethylphosphatidylethanolamine. The sPSD spectra from the [M - H + 2Na]+ ions of MMPE (m/z 750) show two characteristic headgroup ions, at m/z 693 a predominant signal from the neutral loss of monomethylethanolamine (possibly as methylvinylamine structure, ∆m ) 57u). Further headgrouprelated fragments were detected at m/z 200 (disodiated monomethylethanolaminephosphate) and m/z 182 (cyclophosphane structure) (Figure 3a). The loss of the fatty acid generates three different types of fragment ions [M - H + 2Na - (RCOOH)]+, [M - H + 2Na - (RCOOH) - 57]+ and [M - H + 2Na (RCOONa) - 57]+, respectively, with considerably equal signal intensities. At m/z 301, a signal corresponding to disodiated palmitate can be detected Figure 3b shows the sPSD spectrum of [M - H + 2Na]+ ions derived from DMPE (m/z 764) as precursor. In this case, the most prominent fragments ions from the headgroup at m/z 693 result from the neutral loss of dimethylethanolamine (possibly as dimethylvinylamine, ∆m ) 71u). Corresponding ions can be detected and at m/z 72 and at m/z 214 (disodiated dimethylethanolaminephosphate ions). The predominant signal in the spectrum corresponds to the [M - H + 2Na - (RCOOH)]+ ions from the loss of the acyl groups (m/z 508). An approximately four times less intense [M - H + 2Na - (RCOOH) - 71]+ signal at m/z 437 results from a combined loss of a fatty acid and dimethylethanolamine. In agreement with the results using PE and MMPE, the peak at m/z 301 corresponds to disodiated palmitate (see Supporting Information, Table S-2).
Phosphatidylcholine. sPSD experiments were performed using either protonated or sodiated precursor ions derived from DPPC, respectively. The spectra of [M + H]+ (m/z 734) show only one significant fragment ion at m/z 184 corresponding to cholinephosphate26,35,36 (Figure 4a). In contrast, the [M + Na]+ (m/z 756) produced three fragment ions resulting from the neutral loss of trimethylamine (∆m ) 59u), cholinephosphate (∆m ) 183u), and sodiated cholinephosphate (∆m ) 205u) as well as m/z 147 (sodiated cyclophosphane) and with very low intensity m/z 86 (“dehydrocholine”), respectively. These ions are characteristic for the choline headgroup, but no characteristic fragment ions from the loss of the fatty acid residues could be observed from the sodiated molecules at all (Figure 4b). This observation was in good agreement with recently published data from the literature.36 According to results of ESI-MS/MS experiments,35 we repeated our experiments using lithium acetate as THAP matrix additive. In Figure 4c, the sPSD spectrum of the [M + Li]+ ions (m/z 840) derived from SDPC is shown. The most abundant fragment ions arise from the neutral loss of cholinphosphate (m/z 657) and the corresponding lithium adduct (m/z 651). The small peak at m/z 781 indicates the loss of trimethylamine. In the mass range between m/z 450 and 600, three corresponding pairs of fragment ions resulting from the loss of the acyl groups could be detected. The most abundant signals result from the [M + Li (RCOOH) - 59]+ ions, indicating a combined loss of the acyl groups together with trimethylamine, appeared as the main fragmentation pathway of the acyl groups. A comparison of the relative abundances of the signals from the loss of stearic acid (m/z 497) shows an intensity twice as those from docosahexaenoic acid (m/z 453), indicating a preferential loss from fatty acids at the sn-1 over those at sn-2 position also in this case. The same intensity ratio was observed for the neutral loss of stearic acid (m/z 556) and docosahexaenoic acid (m/z 512) as protonated species, respectively. In contrast, loss of the corresponding lithium carboxylates of the fatty acid residues (m/z 506 and 550) shows approximately equal intensities. Further peaks corresponding to fragment ions of the choline headgroup were observed at m/z 131 (lithiated cyclophosphane) and 86 (see Supporting Information, Table S-3). Phosphatidylserine. Normally, three types of adduct ions can be detected corresponding to [M + Na]+, [M - H + 2Na]+, and [M - 2H + 3Na]+, respectively. In the sPSD spectrum of the [M + Na]+ ions from DPPS (m/z 758), four types of fragment ions resulting from losses out of the headgroup were generated. The signals at m/z 671, 573, and 551 represents the neutral loss of serine (∆m ) 87u), serine phosphate, and a sodiated serine phosphate (∆m ) 207u), respectively. A dominating fragment ion signal could be observed at m/z 208 representing sodiated serine phosphate (Figure 5a). The spectrum of the [M - H + 2Na]+ ions (m/z 780) show two signals at m/z 693 and 551, resulting from the neutral loss of serine and disodiated serinephosphate (Figure 5b) In contrast to the sodium adduct, an intense signal can be found at m/z 437 representing [M - H + 2Na - (RCOOH) - 87]+ ions from the (35) Hsu, F.-F.; Bohrer, A.; Turk, J. J. Am. Soc. Mass Spectrom. 1998, 9, 516526. (36) Jackson, S. N.; Wang, H. Y.; Woods, A. S. J. Am. Soc. Mass Spectrom. 2005, 16, 2052-2056.
Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
1669
Figure 3. sPSD spectrum of [M - H + 2Na]+ ions derived from (a) MMPE (m/z 750.5) and (b) DMPE (m/z 764.5). Diagnostic fragmentations are the loss of mono- and dimetylethanolamine (57u and 71u), respectively, along with low mass headgroup specific ions. In the case of MMPE, prominent fragment ions were found to be related to the loss of the fatty acid (m/z 494) and fatty acid together with the headgroup (m/z 437 and 415), respectively. The latter ions are absent or underrepresented in case of DMPE. Peaks at m/z 301 represent disodiated carboxylates (for details, see text).
combined loss of the acyl group (palmitate) and serine. The peak at m/z 230 corresponds to disodiated serinephosphate. The sPSD spectra of the [M - 2H + 3Na]+ ions (m/z 802) show the most complex fragmentation pattern (Figure 5c). The most abundant headgroup fragment ion results from neutral loss of serine (m/z 715) and the corresponding loss of sodiated serine (∆m ) 109 u) at m/z 693. The signals at m/z 546 and 459 represent [M -2H + 3Na - RCOOH]+ and [M - 2H + 3Na (RCOOH) - 87]+ ions resulting from the loss of palmitic acid and the combined loss of palmitic acid with the serine headgroup, respectively. The peak at m/z 437 corresponding to [M - 2H + 3Na - (RCOOH) - 109]+ represents the predominant signal from loss of the acyl groups. This peak might be explained by the neutral loss of palmitic acid and sodiated serine, but may also result from the combined loss of sodium palmitate and serine. In the lower mass range, the signal at m/z 252 represents trisodiated serinephosphate. The peak at m/z 234 most probably corresponds to the same ion species but as amidophosphane structure. Corresponding signals could be found in the sPSD spectra using lithiated precursor ions (see Supporting Information, Table S-4). 1670 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
Phosphatidylglycerol. In MALDI mass spectra, PG was detected as [M + Na]+ and [M - H + 2Na]+ ions, respectively. In Figure 6a, the sPSD spectrum of the [M + Na]+ ions of DPPG (m/z 745) is shown. Three abundant headgroup-related fragment ions resulting from the neutral loss of glycerolphosphate (∆m ) 172u, m/z 573), the sodium adduct of glycerolphosphate (∆m ) 194u, m/z 551) and the corresponding ion (m/z 195) could be observed, respectively. In Figure 6b, the sPSD spectrum of the [M - H + 2Na]+ ions of POPG (m/z 793) is shown. The fragment ion at m/z 719 results from the neutral loss of glycerol (∆m ) 74u), whereas the very weak signal at m/z 217 represents disodiated glycerolphosphate. The predominating signal in the spectra results from loss of glycerolphosphate (m/z 199), according to literature probably as a six-membered cyclophosphane ring.37 At m/z 177 and 125, two signals of headgroup-related ions could be detected with very low intensity. In contrast to the [M + Na]+ ions, no neutral loss of glycerolphosphate could be observed in the case of [M - H + (37) Hsu, F.-F.; Turk, J. J. Am. Soc. Mass Spectrom. 2001, 12, 1036-1043.
Figure 4. sPSD spectrum of (a) [M + H]+ (m/z 734.6), (b) [M + Na]+ (m/z 756.5) derived from DPPC (16:0/16:0), and (c) [M + Li]+ (m/z 840.6) derived from SDPC (18:0/22:6). The only diagnostic ion of the protonated molecules is cholinephosphate (m/z 184). In contrast, sodiated molecules show prominent loss of trimethylamine (59u) and (sodiated) cholinephosphate (205u and 183u). Lithiated molecules show predominant loss of cholinephosphate (183u) along with fragment ions related to the fatty acid residues (inset). Relative signal intensity ratios indicates the preferable loss of 18:0 (sn-1) over 22:6 (sn-2).
2Na]+. Moreover, in the mass range between m/z 400 and 600, different peaks resulting from the loss of the acyl groups could
be detected. The predominating peaks represent [M - H + 2Na - (RCOOH)]+ ions from the loss of palmitic acid (m/z 537) and Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
1671
Figure 5. sPSD spectrum of DPPS (16:0/16:0) obtained from (a) [M + Na]+ (m/z 758.5), (b) [M - H + 2Na]+ (m/z 780.5), and (c) [M - 2H + 3Na]+ (m/z 802.6) as precursor ions. Sodiated molecules show the preferential loss of sodiated serine phosphate (207u), while loss of serine (87u) is the most prominent diagnostic fragmentation of di- and trisodiated molecules, respectively. Most prominent fragment ions related to the acyl groups are loss of the fatty acid residues (16:0) together with the serine headgroup, respectively (details see text).
oleic acid (m/z 511) of the glycerol backbone, respectively. The signal from the loss of the sn-1 fatty acid (16:0) shows a relative intensity ratio twice as abundant as those from the sn-2 position (18:1). Moreover, peaks from the corresponding losses of sodium 1672 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
oleate (m/z 489) and sodium palmitate (m/z 515) could be observed. Additionally, two signals from the combined loss of palmitic acid (m/z 463) and oleic acid (m/z 437) together with the glycerol headgroup could be detected with lower intensity.
Figure 6. sPSD spectrum of (a) [M + Na]+ (m/z 745.5) derived from DPPG (16:0/16:0) and (b) [M - H + 2Na]+ (m/z 793.3) derived from POPG (16:0/18:1). Sodiated molecules show diagnostic fragmentation predominantly by loss of sodiated glycerol phosphate (194u). Disodiated molecules show characteristic loss of glycerol (74u) and diagnostic glycerol phosphate ions (m/z 199) together with fragment ions related to the fatty acid residues (inset). A preferable loss of 16:0 (sn-1) over 18:1 (sn-2) based on the relative signal intensity ratios is observed. Peaks at m/z 301 and 327 represent disodiated carboxylates (for details, see text).
The signals at m/z 327 and 301 represent disodiated carboxylates, [(RCOONa) + Na]+ of oleic acid and palmitic acid, respectively. A comparison with [M - H + 2Li]+ precursor ions using the same PG standard shows a quite similar fragmentation behavior. In this case, the most intense signals resulting from the loss of the acyl groups were the [M - H + 2Li - (RCOOLi)]+ and [M - H + 2Li - (RCOOH) - 74]+ ions, respectively. Additionally, two signals from the glycerol headgroup represening the neutral loss of glycerolphosphate and the corresponding lithium adduct (m/z 606 and 601), respectively. In the case of the [M - H + 2Na]+ ions, no corresponding signals were obtained (see Supporting Information, Table S-5). Phosphatidic Acid. Generally, the MALDI mass spectra of PA are composed of [M + Na]+, [M - H + 2Na]+, and [M - 2H + 3Na]+ ions, respectively. Using sodium acetate (10-25 mM) as matrix additive, [M - H + 2Na]+ and [M - 2H + 3Na]+ were observed the predominant ion types. In the case of the [M - H + 2Na]+ ions from SAPA (m/z 769), the neutral loss of stearic acid (m/z 485) and arachidonic
acid (m/z 465) in the protonated form was observed, whereas the [M - 2H + 3Na]+ ions (m/z 791) showed a preferential loss of the corresponding sodium carboxylates, respectively. In the case of [M - H + 2Na]+ ions (Figure 7a) the relative signal intensity ratios of the [M - H + 2Na - RCOOH]+ ions was nearly 3:2 between those from the sn-1 and sn-2 positions, respectively. Interestingly, when using [M - 2H + 3Na]+ as precursor ions (Figure 7b), an inversed intensity relation was observed between the [M - 2H + 3Na - RCOONa]+ ions showing an enhanced acyl group fragmentation from the sn-2 position. Moreover, loss of the fatty acid residues in protonated form could be observed only with very low abundance. A similar fragmentation behavior was observed using [M - H + 2Li]+ and [M - 2H + 3Li]+ as precursor ions (data not shown). Phosphatidylinositol. MALDI mass spectra of PI from bovine brain show a complex signal distribution, which is based on the inhomogeneous fatty acid composition and impurities of the sample. Signals at m/z 909 and 931 are correlated to [M + Na]+ and [M - H + 2Na]+ ions representing the most prominent Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
1673
Figure 7. sPSD spectrum of SAPA (18:0/20:4) obtained from (a) [M - H + 2Na]+ (m/z 769.6) and (b) [M - 2H + 3Na]+ (m/z 791.6) as precursor ions. Fragment ions correspond solely to the loss of the acyl groups. Signal abundance of loss from 18:0 (sn-1) and 20:4 (sn-2) was dependent on the precursor ion type. Disodiated molecules show preferable loss from sn-1, while loss from sn-2 was favored in the case of trisodiated molecules.
molecular ions. sPSD spectra of the [M + Na]+ ions (Figure 8a) exhibit three fragment ions resulting from the loss of the headgroup. The peaks at m/z 649 and 627 represent the neutral loss of inositol 1-phosphate (∆m ) 260u) and the corresponding sodium adduct (∆m ) 282u), respectively. The signal at m/z 283 results from sodiated inositol 1-phosphate ions. In contrast, the [M - H + 2Na]+ ions (Figure 8b) show a more complex fragmentation pattern. The signal at m/z 769 represents the neutral loss of inositol (as C6H10O5, ∆m ) 162u), while in the range between m/z 400 and 700 fragment ions reflecting the loss of the acyl groups could be detected. The peaks at m/z 647 and 627 represent [M - H + 2Na - (RCOOH)]+ ions corresponding to the neutral losses of stearic (∆m ) 284u) and arachidonic acid (∆m ) 304u). The signal intesity ratio of ∼3:1 between both peaks indicates for the localization of arachidonic acid at the sn-2 position. A corresponding signal from the loss of sodium stearate could be detected at m/z 625. Additionally, [M - H + 2Na - (RCOOH) - 162]+ ions indicating a combined loss of the acyl groups and 1674 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
inositol were detected at m/z 485 and 465, respectively. A characteristic headgroup specific ion was observed at m/z 287. This represents disodiated inositol 1-phosphate after elimination of a water molecule.38 Cardiolipin (Bisphosphatidylglycerol). MALDI spectra using THAP matrix exhibit three types of molecular ions [M + Na]+ (m/z 1472), [M - H + 2Na]+ (m/z 1493), and [M - 2H + 3Na]+ (m/z 1515), respectively. sPSD spectra of the [M + Na]+ ions (Figure 9a) show only two dominant fragment ions, one of them (m/z 873) indicates the neutral loss of dilinoleoylglycerol (∆m ) 598u), and the other represents a dilinoleoylglycerol cation (m/z 599). A small signal was observed at m/z 275 representing sodiated glycerol 1,3-bisphosphate. The most abundant signals in sPSD spectra of the [M - H + 2Na]+ ions (Figure 9b) result from loss of dilinoleoylglycerol (m/z 895) and the additional loss of water molecule (m/z 877). The peaks at m/z 741 and 599 (38) Hsu, F.-F.; Turk, J. J. Am. Soc. Mass Spectrom. 2000, 11, 986-999.
Figure 8. sPSD spectrum of PI (18:0/20:4) obtained from (a) [M + Na]+ (m/z 909.5) and (b) the [M - H + 2Na]+ (m/z 931.5) as precursor ions. Loss of inositol 1-phosphate (260u) and sodiated inositol 1-phosphate (m/z 283) are the characteristic fragmentations of sodiated molecules. Fragment ions related to the fatty acid residues are observed only in the case of disodiated molecules (inset). Relative signal intensity ratios indicate the preferable loss of 18:0 (sn-1) over 20:4 (sn-2).
correspond to disodiated dilinoleoylphosphatidic acid and again the dilinoleoylglycerol cation, respectively. The latter may also result from disodium phosphate loss of m/z 741. Low mass signals represent disodiated glycerol 1,3-bisphosphate (m/z 297) and the dehydrated form of this ion species (m/z 278), respectively. Similar fragment ions derived from cardiolipin were described in the literature recently.39 All relevant sPSD fragment ions from cardiolipin incorporate fatty acid moieties, but only the [M - 2H + 3Na]+ ions allowed direct assignments of the acyl groups (Figure 9c). The peaks at m/z 1235 and 1213 represent the loss of linoleic acid (18:2) from the glycerol backbone and the corresponding sodium carboxylate, respectively. Signals corresponding to another fatty acid residue could not be detected, indicating for the exclusive presence of linoleic acid in this sample as was also described previously.33 The most prominent fragment ion (m/z 899) results from the loss of one dilinoleoylglycerol moiety together with water (∆m ) 616u). The signal at m/z 797 corresponds to a dehydrated and (39) Hsu, F.-F.; Turk, J. J. Am. Soc. Mass Spectrom. 2006, 17, 1146-1157.
disodiated dilinoleoylphosphatidylglycerol. The signals at m/z 621 and 599 can be explained from the consecutive loss of dilinoleoylglycerol from m/z 1235 and 1213, respectively. In the lower mass range, disodiated lysophosphatidic acid (m/z 461) and dehydrated and disodiated glycerol 1, 3-bisphosphate (m/z 279) were observed. Sphingomyelin. MALDI mass spectra of SM using THAP as matrix show [M + H]+ and [M + Na]+ as the predominant molecular ions. sPSD spectra of the [M + H]+ ions (m/z 703) show only one fragment ion at m/z 184, corresponding to the neutral loss of cholinphosphate (data not shown). These spectra are completely identical with those from [M + H]+ ions of PC and can therefore not be used to define members of the SM class unambiguously. In Figure 10, the sPSD spectrum of the [M + Na]+ ions (m/z 725) from SM from chicken egg yolk is shown. Also, the sodiated precursor ions exhibit a fragmentation pattern nearly similar to those of PC, showing two dominant signals from the neutral loss of trimethylamine (m/z 666) and cholinphosphate (m/z 542). In the lower mass range, signals with very low intensities from sodiated cyclophosphane (m/z 147) and dehyAnalytical Chemistry, Vol. 80, No. 5, March 1, 2008
1675
Figure 9. sPSD spectrum of cardiolipin (18:2/18:2) derived from (a) [M + Na]+ (m/z 1472.0), (b) [M - H + 2Na]+ (m/z 1493.9), and (c) [M 2H + 3Na]+ (m/z 1515.9) as precursor ions. Most prominent diagnostic fragmentation corresponds to loss of one glycerol moiety containing the fatty acid residues (dilinoleoylglycerol) (598u and 616u). Only the trisodiated molecules show fragment ions indicating the loss of the fatty acid residues (details see text).
drocholine (m/z 86) were observed. In contrast to sodiated PC (see Figure 4b), SM showed no fragment ions corresponding to the neutral loss of sodiated cholinephosphate. This can be used 1676 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
as a differential diagnostic feature between both lipid classes. Even in the case of using lithiated molecules, neither diagnostic fragment ions allowing the characterization of the sphingosine
Figure 10. sPSD spectrum of SM (16:0) using [M + Na]+ (m/z 725.6) as precursor ions. Most abundant diagnostic fragmentation corresponds to loss of trimethylamine (59u) and choline phosphate (183u). No fragment ions indicating the fatty acid composition are observed.
moiety nor indication of the acyl group could be obtained (data not shown). CONCLUDING REMARKS A general finding during our investigation was that, independent of the GPL class, the fragmentation pattern related to the headgroup and particularly fragment ions resulting from the loss of the acyl groups was strongly dependent on the used precursor ion type (protonated or cationized). Thereby, fragmentation behavior was considerably enhanced in the following order: [M + H]+ < [M + Na/Li)]+ < [M - H + 2Na/2Li]+ or [M 2H + 3Na/3Li]+. Protonated molecules of most GPL classes show almost no significant fragment ions (e.g., in the case of PC and PE). In this case, the generation of alkali ion adduction obviously reduces the threshold energy of the molecules favoring their tendency for dissociation. This observation was in agreement with recent studies applying fundamentally different MALDI instrumentation.29,36 Based on our findings, in general a two-step competitive reaction mechanism between the two electronegative centers of GPL molecules (i.e., the phosphate group and the carboxyl ester of the fatty acid residues) depending on the alkali salt concentration of the sample/matrix preparation appears most plausible. First, under low salt concentrations, a preferred attachment of an alkali ion will take place at the phosphate ester, primarily promoting the generation of fragment ions from the polar headgroup. Second, a cleavage reaction of the acyl groups from the glycerol backbone will take place, if an alkali ion is located in the environment of the carboxyl ester groups of the fatty acids. This alkali attachment is obviously crucial to overcome the energy requirement for the acyl group fragmentation. Generally, the acyl groups were found to be lost as neutral fatty acids (protonated) and the corresponding alkali salts (sodiated or lithiated), respectively. Quite often loss of the acyl groups was accompanied by headgroup fragmentation, possibly in a concerted reaction. Nevertheless, it should be mentioned that the fragment ions observed in sPSD spectra are only based on net losses from the selected precursor ions and thus do not directly reflect all
consecutive steps of the underlaying elimination process. Some investigators proposed fragmentation pathways involving five- or six-membered ring products (e.g., cyclo- and amidophosphane structures) involving the phosphate ester group.40,41 Strong evidence is given for very similar fragmentation products under MALDI-PSD conditions based on the observed fragment ions in our experiments. The loss of dehydrated headgroup structures (e.g., vinylamine, dehydrocholine, etc.) can be deduced from McLafferty rearrangement reactions.42
CONCLUSION Positive ion MALDI-sPSD fragment ion analysis of alkali adduct ions represents a promising technique to obtain structural information of phospholipids. It was shown, that sPSD analysis made it possible to readily characterize the members of the different glycerophospholipid classes based on the generation of informative fragment ions related to the polar headgroups (diagnostic ions). The use of certain alkali metal adducts (sodium or lithium) as precursor ions was crucial to propagate the dissociation behavior (metastable decay) of the GPL molecules. Particularly, the generation of fragment ions reflecting the fatty acid constitution of the molecules (acyl group composition) was strongly dependent on the type of precursor ions (protonted or cationized) and the number of attached alkali ions. As was outlined comparing our actual results with those already obtained by ESI low-energy CID, striking similarities could be observed concerning the fragmentation behavior and degree of structural information available from most GPL classes. Thereby, both ESI low-energy CID and MALDI sPSD allow assignments of the fatty acid residues as well as their position on the glycerol backbone. The loss of the acyl groups from the sn-1 hydroxyl of glycerol was in most cases favored over the sn-2 hydroxyl, probably due to the formation of sterically more preferable intermediate structures. These facts outline that MALDI sPSD is an equally suited (40) Hsu, F.-F.; Turk, J. J. Am Soc. Mass Spectrom. 2000, 11, 892-899. (41) Hsu, F.-F.; Turk, J. J. Am Soc. Mass Spectrom. 2000, 11, 437-449. (42) McLafferty, F. W. Org. Mass Spectrom. 1980, 15, 114-121.
Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
1677
technique for structural analysis of phospholipids compared to ESI low-energy CID and thus it will be an interesting alternative approach for future lipidomics studies. ACKNOWLEDGMENT This project was partly supported by the Austrian Ministry of Agriculture (grant 1163 to G.A.). Furthermore, we thank O. Belgacem for continuous discussion and helpful feedback on the manuscript.
1678
Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
SUPPORTING INFORMATION AVAILABLE Additional information as noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org.
Received for review November 28, 2007. AC7018766
September
6,
2007.
Accepted
