Anal. Chem. 2007, 79, 3206-3213
Technical Notes
Analysis of Lipids Using 2,4,6-Trihydroxyacetophenone as a Matrix for MALDI Mass Spectrometry Gerald Stu 1 biger*,† and Omar Belgacem‡
Institute of Chemical Technologies and Analytics, University of Technology, Getreidemarkt 9/164, A-1060 Vienna, Austria, and Shimadzu Biotech, Wharfside, Trafford Wharf Road, Manchester M17 1GP, U.K.
Lipids exhibit a broad range of chemical properties that make their analysis quite demanding. Today, matrixassisted laser desorption/ionization mass spectrometry (MALDI-MS) represents a versatile tool in the field of lipid analysis, also offering the possibility for molecular structural identification using novel MALDI tandem time-offlight (TOF/TOF) instrumentation. In this study, we evaluated 2,4,6-trihydroxyacetophenone (THAP) for the analysis of various lipid classes including neutral storage lipids (triacylglycerols), polar membrane lipids (glycerophospho- and sphingolipids), and glycosphingolipids. THAP proved to be a versatile matrix for the routine analysis of various lipids from biological samples (“lipidomics”). A sample preparation methodology was established using selective alkali salt doping for subsequent MS/MS experiments. Sodiated and lithiated molecules provided superior structural information on lipids (i.e., acyl group identification); thus, following this approach, both selective peak detection with high sensitivity and more reliable structural information were obtained simultaneously. Lipids are one of the most important classes of biomolecules that exhibit a variety of biological functions. Triacylglycerols (TAGs) (Figure 1) belong to the group of lipid storage molecules in plants, animals, and humans. Edible fats and oils consist of over 90% TAGs, providing the main source of essential fatty acids in our nutrition. Evidence exists that TAG composition plays a major role in human lipid metabolism and the development of obesity or even cardiovascular diseases.1,2 Glycerophospholipids (GPLs) and sphingolipids (SLs) (Figure 1) are important structural lipids (e.g., constituents of cell membranes) and play a regulatory role in metabolism through their actions as secondary messengers3,4 * Corresponding author. E-mail:
[email protected]. † University of Technology. ‡ Shimadzu Biotech. (1) Ooi, T. C.; Ooi, D. S. Crit. Rev. Clin. Lab. Sci. 1998, 35, 489-516. (2) Murase, T.; Mizuno, T.; Omachi, T.; Onizawa, K.; Komine, Y.; Kondo, H.; Hase, T.; Tokimitsu, I. J. Lipid Res. 2001, 42, 372-378. (3) Maxfield, F. R.; Tabas, I. Nature 2005, 438, 612-621. (4) Zeisel, S. H.; Blusztajn, J. K. Annu. Rev. Nutr. 1994, 14, 269-296.
3206 Analytical Chemistry, Vol. 79, No. 8, April 15, 2007
(e.g., phosphatidic acid, sphingomyelin, and ceramides). Glycosphingolipids (GSLs) belong to the structurally most complex class of lipid molecules consisting of a lipid part (ceramide) together with a neutral (e.g., cerebrosides) or acidic (e.g., gangliosides) oligosaccharide moiety (Figure 1).5-7 Disorders in GSL metabolism are associated with several known neuronal diseases8 (e.g., Tay-Sachs, Gaucher’s, and Fabry’s diseases). During the past decade, various matrix systems and sample preparation strategies have been developed for lipid analysis by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS).9-12 In practice, however, many matrix substances have proven to be inadequate because of poor sensitivity, inhomogeneous crystallization (“sweet-spot phenomenon”), and/or extended fragmentation especially in the case of “hot” matrices.13 Furthermore, intense matrix ions can hamper the analysis of lowmolecular-mass compounds (20 mM)
+ sodium acetate (10-20 mM)
+ lithium acetate (5-10 mM)
[M + Na]+ [M + K]+ [M + Na]+ [M + K]+
[M + Na]+
[M + Na]+
[M + Li]+
[M + Na]+
[M - H + 2Na]+
[M - H + 2Li]+
[M + H]+ [M + Na]+ [M + K]+ [M + Na]+ [M - H + 2Na]+ [M - 2H + 3Na]+ [M + Na]+ [M - H + 2Na]+ [M - 2H + 3Na]+ [M - 3H + 4Na]+ [M + Na]+ [M - H + 2Na]+ [M + Na]+ [M - H + 2Na]+ [M - 2H + 3Na]+ [M + Na]+ [M - H + 2Na]+ [M - 2H + 3Na]+ [M + H]+ [M + Na]+ [M + K]+ [M + Na]+ [M + K]+ [M + Na]+ [M - H + 2Na]+ [M - H + Na + K]+ [M - 2H + 3Na]+ [M + Na]+ [M + K]+
[M + H]+
[M + Na]+
[M + Li]+
[M + Na]+
[M - 2H + 3Na]+
[M - 2H + 3Li]+
[M + Na]+
[M - H + 2Na]+ [M - 2H + 3Na]+ [M - 3H + 4Na]+
b
[M + Na]+
[M - H + 2Na]+
[M - H + 2Li]+
[M + Na]+
[M - H + 2Na]+ [M - H + 3Na]+
b
[M + Na]+ [M - H + 2Na]+
[M - 2H + 3Na]+
[M - 2H + 3Li]+
[M + H]+
[M + Na]+
[M + Li]+
[M + Na]+
[M + Na]+
b
[M + Na]+
c
c
[M + Na]+
b
b
THAP (pure methanolic solution)
a TAG, triacylglycerol; PE, phosphatidylethanolamine; MMPE, monomethyl phosphatidylethanolamine; DMPE, dimethylphosphatidylethanolamine; PC, phosphatidylcholine; PS, phosphatidylserine; PI, phosphatidylinositol; PIP, phosphatidylinositol monophosphate; PIP2, phosphatidylinositol bisphosphate; PG, phosphatidylglycerol; PA, phosphatidic acid; cardiolipin, bisphosphatidylglycerol; SM, sphingomyelin; gangliosides, acidic glycolipids; cerebrosides, neutral glycolipids. b Not investigated. c Deteriorating effect (glycan fragmentation).
In the second approach, sodium was added to reduce the relative concentration of potassium, and in the third, lithium was added to displace both sodium and potassium. The third approach has two main advantages: On one hand, the unique isotopic signature of lithium can be exploited for unambiguous peak assignments in lipid mixture analysis.53 On the other hand, lithiation was observed to be advantageous for the structural analysis of lipids by MALDI-MS/MS experiments (see below). Obviously, the diverse functional sites of complex lipid molecules, such as the carboxyl groups of the fatty acid residues (1), the phosphate ester (2), and the headgroups (3) or the oligosaccharide moiety (4), contribute together to their alkali ion attachment behavior (Figure 1). Actually, the number of accessible sites determines the predominant adduct ion types observed in the MALDI mass spectra. The degree of alkali ion attachment was found between (0)1-3(4) in most cases (Table 2). The influence of the salt doping technique using THAP was examined in greater detail with PE and PC standards. Using measurements of signal intensity ratios of the adduct ions, the influence of different concentrations of ammonium citrate, sodium (53) Cvacka, J.; Svatos, A. Rapid Commun. Mass Spectrom. 2003, 17, 22032207.
3210
Analytical Chemistry, Vol. 79, No. 8, April 15, 2007
acetate, and lithium acetate was evaluated (Figure 2). Addition of citrate above 25 mM favored protonated PC molecules exclusively (Figure 2a). In contrast, only [M + Na]+ ions were observed after addition of sodium between 10 and 20 mM, whereas [M + Li]+ ions were already observed following the addition of 5 mM lithium acetate to the matrix solution (Figure 2b). [M + 2Na - H]+ ions of PE could be observed predominantly in the concentration range between 5 and 10 mM sodium acetate (Figure 2c). The same effect was observed in the case of lithium acetate addition, where the resulting mass spectra of PE were dominated by [M + 2Li - H]+ ions (Table 2). Based on these experiments, the logarithm of the molar quotient of the concentrations (in mM) of the analyte (sample) and salt additive (saltDot) could be calculated as
(
Y ) -log
)
[sample] [saltDot]
(1)
According to this equation, the critical value of predominant alkali adduct formation (contribution of counter adduct ions below 20%) in most of the GPL classes was found to be at an optimum for Y > 2. Without knowing the alkali salt contamination of the sam-
Figure 3. Signal intensity variation of PC and PE species from soybean lecithin dependent on the (a) sodium and (b) lithium concentration. The vertical axis represents the relative signal intensity of all PC/PE adduct ions normalized to the most intense peak (designated as 100%) in the range of m/z 680-880. PC was recorded as [M + K]+, [M + H]+, and either [M + Na]+ or [M + Li]+, respectively. PE was detected as [M - H + 2Na]+ or [M - H + 2Li]+ depending on the salt additive used.
Figure 2. Influence of salt additives on the adduct ion signal intensity of (a) PC using citrate, (b) PC using lithium, and (c) PE using sodium. A sample of 30 pmol was applied to THAP matrix solution in each case. The vertical axis represents the percentage of the sum of the peaks of interest (adduct ions) at each concentration of salt additive.
ples ab initio, for most lipid classes, an optimum concentration between 5 and 20 mM salt additive within the matrix solution was found to enhance exclusively one abundant alkali adduct ion type (Table 2). Application to a Complex Phospholipid Mixture. Soybean lecithin is a lipid mixture and a very rich source of dietary choline (lecithin) that plays an essential role in PC metabolism and
synthesis of cell membranes.54 It is known that PC suppresses the detection of other GPL classes in mixtures, particularly when it is the major component.55 Thus, this phenomenon was studied in more detail using increasing concentrations (0.01-500 mM) of sodium and lithium acetate in THAP solution. The results of these experiments are illustrated in Figure 3. In the range between 10 and 25 mM sodium acetate, [M + Na]+ ions of PC became the predominant ions of the MALDI mass spectra. Surprisingly, increasing the sodium concentration in the matrix solution above 50 mM led to the detection of [M - H + 2Na]+ ions of two PE species that were not present in the original mass spectrum (Figure 3a). Through subsequent MS/MS analysis, these signals were identified as [M - H + 2Na]+ ions of two PE species containing either one palmitic (16:0) and one linoleic (18:2) or two linoleic (18:2) acid residues (see Supporting Information). The same scenario was observed in the case of an increasing concentration of lithium acetate; interestingly, in contrast to the observations using sodium acetate, the [M - H + 2Li]+ ions corresponding to PE were only slightly increased, even at high (>100 mM) lithium acetate concentrations (Figure 3b). These experiments clearly show that alkali ion affinity differences between PC and PE can be exploited for their selective detection from lipid mixtures using THAP matrix without the need for changing the experimental conditions.29 A relatively low critical salt (sodium or lithium) concentration was sufficient for the exclusive detection of PC, whereas PE could be favorably detected only above a sodium concentration of 50 mM. The observed (54) Zeisel, S. H. J. Am. Coll. Nutr. 2000, 19, 528S-31S. (55) Petkovic, M.; Schiller, J.; Muller, M.; Benard, S.; Reichl, S.; Arnold, K.; Arnhold, J. Anal. Biochem. 2001, 289, 202-216.
Analytical Chemistry, Vol. 79, No. 8, April 15, 2007
3211
Figure 4. MS/MS spectrum of the [M + Li]+ ions of sn-1-stearoyl-2-docosahexaenoyl-PC (22:6/18:0). Fragment ions are indicated in the schematic view of the molecule. Fragmentation of the fatty acid residues (R1, R2) was observed as neutral loss of the acyl groups either alone or together with trimethylamine (59 amu) from the choline headgroup.
phenomenon was interesting and could be very useful in practice. Undoubtedly, the concentration of the lipids in the sample will affect the outcome of the doping experiment. In the soybean lecithin sample used, PC and PE were present in quite similar amounts, as was verified by thin-layer chromatography (data not shown). An overabundance of PC in other samples might hamper the simultaneous detection of PE using this methodology. Nevertheless, following this approach, crude lipid mixtures can be qualitatively investigated by MALDI analysis, thus obviating the need for time-consuming chromatographic sample preparation. Considerations in the Structural Analysis of Lipids. In the following experiments, the salt doping method was evaluated for the structural analysis of lipids by MALDI-MS/MS experiments. Thus, formation of distinct alkali adduct ions by the addition of specific salt additives was exploited to favor distinct fragmentation pathways of the molecules. In particular, our aim was to obtain information about the fatty acid composition of the molecules, which is a crucial structural and functional feature of lipids in general. It is known from several studies23,30,56,57 that, depending on the investigated lipid class, tandem mass spectra show either preferential losses of the acyl groups in the case of neutral lipids (e.g., TAG) or fragment ions mainly resulting from the headgroups in the case of the polar membrane lipids (e.g., GPLs). Mechanistic studies58,59 have shown that fragmentation pathways of phospholipids strongly depend on the precursor ion types (protonated or alkaliated) and the size (atomic diameter) of the attached alkali metals (K+ > Na+ > Li+). In this context, it is known that sodium (56) Domingues, M. R.; Marques, M. G.; Vale, C. A.; Neves, M. G.; Cavaleiro, J. A.; Ferrer-Correia, A. J.; Nemirovskiy, O. V.; Gross, M. L. J. Am. Soc. Mass Spectrom. 1999, 10, 217-223. (57) Stubiger, G.; Pittenauer, E.; Ralphson, K.; Allmaier, G. In Proceedings of the 49th ASMS Conference on Mass Spectrometry and Allied Topics; American Society for Mass Spectrometry (ASMS): Santa Fe, NM, 2001. (58) Hsu, F. F.; Turk, J. J. Am. Soc. Mass Spectrom. 2000, 11, 892-899. (59) Hsu, F. F.; Turk, J. J. Am. Soc. Mass Spectrom. 2000, 11, 437-449.
3212 Analytical Chemistry, Vol. 79, No. 8, April 15, 2007
adducts provide a much higher degree of structural information than protonated species.60 In our case, the MS/MS spectra of [M + H]+ ions from PC or SM consisted only of cholinephosphate (m/z 184), whereas those of PE were dominated by ethanolaminephosphate (m/z 141) exclusively. In contrast, a much richer fragmentation spectrum could be observed from [M + Na]+ ions, which allowed further differentiation based on additional diagnostic ions. A very interesting finding during our study was that the loss of the acyl groups from most lipids is strongly enhanced depending on their degree of alkali ion saturation (see Supporting Information). Interestingly, applying the same collision energy, we observed that, in contrast to the addition of sodium, addition of lithium allowed the direct assignment of the fatty acid substituents (e.g., in the case of PC). As an example, the MS/MS spectrum of [M + Li]+ from sn-1-stearoyl-2-docosahexaenoyl-PC (m/z 840) is shown (Figure 4). In addition to the abundant peaks from the loss of cholinephosphate (m/z 651 and 657), breakdown of the acyl groups could be observed in the range of m/z 400-600. Fragment ions resulting from the loss of stearic acid (18:0) from the sn-1 position (m/z 497) were about twice as abundant as those from the loss of docosahexaenoic acid (22:6) esterified at the sn-2 hydroxyl group (m/z 453). Along with the acyl fragmentation, the concomitant loss of trimethylamine (59 amu) could be observed in both cases. The preferential loss of the sn-1 acyl group seems to be a crucial phenomenon for GPL molecules in general, given that it has been observed during several studies actually independent of the instrumentation employed.61,62 (60) Al Saad, K. A.; Siems, W. F.; Hill, H. H.; Zabrouskov, V.; Knowles, N. R. J. Am. Soc. Mass Spectrom. 2003, 14, 373-382. (61) Hsu, F. F.; Bohrer, A.; Turk, J. J. Am. Soc. Mass Spectrom. 1998, 9, 516526. (62) Jackson, S. N.; Wang, H. Y.; Woods, A. S. J. Am. Soc. Mass Spectrom. 2005, 16, 2052-2056.
A detailed summary of our results from MALDI-MS/MS experiments of the various lipid classes is available in the Supporting Information. CONCLUSIONS As a consequence of the post-genomic era and the establishment of proteomics, we are now entering the area of lipidomics,63,64 representing a link to the understanding of whole metabolic processes. Thus, approved as well as new analytical techniques are increasingly demanded. MALDI-MS has proved to be a very promising tool in the field of lipid analysis. According to our evaluation of various matrix substances, THAP appears to be an outstanding matrix for the analysis of most lipid classes. Based on the establishment of the salt doping technique, the complexity of MALDI mass spectra could be reduced with respect to one predominant adduct ion type for each lipid class. This methodology is advantageous for the structural analysis of lipid molecules, as each type of alkali adduct provides different structural informa(63) Ivanova, P. T.; Milne, S. B.; Forrester, J. S.; Brown, H. A. Mol. Interv. 2004, 4, 86-96. (64) Han, X.; Gross, R. W. J. Lipid Res. 2003, 44, 1071-1079.
tion during MALDI-MS/MS experiments. In the analysis of lipid mixtures, members of different lipid classes (e.g., PC and PE) could be differentiated based on defined salt additives to the THAP matrix (selective alkali salt doping). However, the variability in affinity to alkali ions has to be considered carefully in the case of any quantitative analysis of complex lipid mixtures by MALDIMS. ACKNOWLEDGMENT The authors thank Professor Guenter Allmaier for his support and Dr. Ernst Pittenauer for helpful discussions. They also thank Professor Raymond March for his feedback on the manuscript. 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 26, 2006. Accepted January 19, 2007. AC062236C
Analytical Chemistry, Vol. 79, No. 8, April 15, 2007
3213
