Article pubs.acs.org/ac
Quantitative Characterization of Tissue Globotetraosylceramides in a Rat Model of Polycystic Kidney Disease by PrimaDrop Sample Preparation and Indirect High-Performance Thin Layer Chromatography−Matrix-Assisted Laser Desorption/Ionization-Timeof-Flight-Mass Spectrometry with Automated Data Acquisition Hermelindis Ruh,†,‡,§,∥ Roger Sandhoff,‡,⊥ Björn Meyer,†,‡,§ Norbert Gretz,‡,§,∥ and Carsten Hopf*,†,‡,§ †
Instrumental Analysis and Bioanalysis, Department of Biotechnology, Mannheim University of Applied Sciences, Paul-Wittsack-Str. 10, 68163 Mannheim, Germany ‡ Applied Research Center in Biomedical Mass Spectrometry (ABIMAS), Mannheim University of Applied Sciences, Paul-Wittsack-Str. 10, 68163 Mannheim, Germany § Institute of Medical Technology, University of Heidelberg and Mannheim University of Applied Sciences, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany ∥ Medical Research Center, Theodor-Kutzer-Ufer 1-3, 68167 Mannheim, Germany ⊥ Lipid Pathobiochemistry, German Cancer Research Center, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany S Supporting Information *
ABSTRACT: Glycosphingolipids (GSL) have been associated with a variety of diseases, including cancer and autosomal dominant polycystic kidney disease (ADPKD). In contrast to glucosylceramide and gangliosides, alterations in complex neutral GSLs such as globotetraosylceramide (Gb4Cer) have not been investigated in ADPKD yet, and mass spectrometry analysis of Gb4Cer from tissue extracts remains challenging. To this end, we introduce PrimaDrop as an improved and widely applicable sample preparation method for automated matrix-assisted laser desorption/ionization-mass spectrometry (MALDI-MS) analysis of lipid extracts, which promotes homogeneous cocrystallization and enables relative quantification by indirect thin layer chromatography (TLC)−MALDItime-of-flight (TOF)-MS against an internal bradykinin standard. Application of the method for detailed investigation of Gb4Cer isoforms in kidneys of an ADPKD rat model revealed increased levels of sphingoid base-containing isoforms in cystic kidneys, whereas changes were subtle for Gb4Cer-containing phytosphingoid bases. We furthermore established an absolute LC−ESIMS/MS quantification method and demonstrate that absolute quantities of Gb4Cer correlate well with relative quantities obtained by indirect TLC−MALDI-TOF-MS. Taken together, our study proposes an effective sample preparation method for automated analysis of lipid extracts and TLC eluates and suggests that indirect high-performace (HP)TLC−MALDI-TOF-MS with automated data acquisition is a viable option for analysis of neutral glycosphingolipids and that Gb4Cer may play a role in the pathogenesis of ADPKD.
G
cysts and ultimately by destruction of kidney architecture, resulting in end-stage renal failure.8,9 The PKD rat is a widely used animal model of ADPKD, and several characteristics of human disease are manifest in heterozygous PKD rats.10 Levels of gangliosides and some neutral GSLs are elevated in kidneys of patients with ADPKD as well as in some animal models of the disease, but complex neutral GSLs such as Gb4Cer have not yet been investigated.11,12
lobotetraosylceramides (Gb4Cer) are members of the globo series of complex neutral glycosphingolipids (GSL) (see Figure S1A of the Supporting Information for structural details).1−3 Neutral GSLs play pivotal roles in a number of human diseases, including cancer, metabolic syndrome, insulin resistance, and Alzheimer’s disease.4−7 Moreover, defects in GSL metabolism cause inherited disorders (e.g., some lysosomal storage disorders), where GSL degradation is blocked and accumulation of nondegradable GSLs, including Gb4Cer, results in destruction of affected tissues.2 Autosomal dominant polycystic kidney disease (ADPKD) is a common hereditary disorder with a prevalence of about 1:1000. It is characterized by the formation of countless fluid-filled renal © XXXX American Chemical Society
Received: November 18, 2012 Accepted: May 23, 2013
A
dx.doi.org/10.1021/ac400931u | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
MS/MS absolute quantification method for targeted tissueGb4Cer analysis using self-made internal Gg4Cer standards and (ii) relative quantification by indirect TLC−MALDI-TOF-MS, utilizing an improved solvent mixture for effective HPTLC separation of Gb4Cer from highly abundant SM. Furthermore, we present an improved sample preparation method for lipid extracts, which promotes homogeneous analyte/matrix cocrystallization on a MALDI target and makes automated MS data acquisition possible.
Methods for quantification of the most simple and complex neutral GSLs in body fluids by liquid chromatographyelectrospray ionization-tandem mass spectrometry [LC− (ESI)-MS/MS] are well-established.13,14 Recently, bioanalysis of complex neutral GSLs in body fluids and in cells by matrixassisted laser desorption/ionization-mass spectrometry (MALDI-MS)-based workflows has been advanced, often in LC-MALDI configurations.15−20 Little progress, however, has been made in MS-based characterization of complex neutral GSL, such as Gb4Cer from tissue extracts. Major reasons for this are the high complexity and diversity of lipophilic tissue constituents: highly abundant, positively charged membrane components such as sphingomyelin (SM) and phosphatidylcholine can be detected with high sensitivity in the positive-ion mode,15 whereas acidic GSLs such as sulfatides can easily be analyzed in the negative-ion mode.21,22 Ionization of neutral GSLs is the most delicate. Although LC-ESI-MS methods are applied for quantification of simple GSL,12,23 valid quantitative analytical methods are lacking for complex globo series GSLs in tissue. Nevertheless, efforts have been made to investigate complex neutral GSL in equine kidney by MALDI-TOF-MS.24 Proposed work flows, however, require additional sample preparation steps, including derivatization such as permethylation and per- and deacetylation. Insensitive direct coupling of high-performance thin layer chromatography (HPTLC) with MALDI-MS without lipid extraction from the silica gel layer19 has been used for analysis of neutral GSLs from mouse organs by a MALDI source coupled to a quadrupole ion trap.25 TLC overlay assays using carbohydrate-binding antibodies and bacterial toxins have been combined with direct IR-MALDITOF-MS.15,26−28 Although densitometric evaluation enables relative quantification of total lipids present in the visualized TLC band, this method does not allow for quantification of distinct isoforms. We therefore set out to establish relative quantification by indirect TLC−MALDI-TOF-MS, in which lipid-containing bands are extracted from the silica gel layer prior to MS analysis. Automated acquisition of MALDI-TOFMS data is a prerequisite for reliable and rapid analysis of samples, including clinical ones. Therefore, easy-to-use sample preparation methods for analysis of lipid extracts and fractions by automated MALDI-TOF-MS are required. Inhomogeneous cocrystallization of matrix and analytes and formation of “sweet spots” in standard dried-droplet sample preparation is arguably the biggest hurdle for development of automated MALDITOF-MS applications. Alternative crystallization methods, such as fast evaporation,29,30 two-layer, or sandwich crystallization,31,32 have been successfully described for proteomic studies and intact cell profiling,33−36 but homogeneous lipid-sample crystallization methods for automated MS work flows have not been assessed so far. Careful matrix optimization often leads to significant improvements in sensitivity of lipid detection.37 MALDI matrices, 2,5-dihydroxybenzoic acid (DHB), 6-aza-2thiothymine (ATT), and 2′,4′,6′-trihydroxyacetophenone (THAP), are often used in analysis of diverse lipid classes by MALDI-TOF-MS.38−40 Enhanced desorption/ionization of lipids by the use of salt additives16,39,41 has been noted; for example, for ATT in combination with lithium citrate for detection of Gb4Cer in human erythrocytes.16 However, neither matrices nor salt additives have been evaluated for lipid analytes in automated MS. Here, we developed and compared two new quantitative analytical workflows for investigation of complex neutral GSL from tissue extracts on different MS platforms: (i) an LC-ESI-
■
EXPERIMENTAL SECTION See additional online Supporting Information for further experimental details. HPTLC Analysis. Prior to MS analysis, neutral kidney extracts corresponding to 5 mg dry weight were applied to silica HPTLC plates and separated in chloroform/methanol/0.2% aqueous CaCl2 (50/35/8, v/v/v). The plates were sprayed with primulin in a uniform manner until a barely visible film emerged. UV−visible Gb4Cer bands were outlined with a pencil and scraped off. Gb4Cer was eluted with 200 μL methanol by sonification in a water bath for 5 min and incubation overnight at room temperature. Samples were then centrifuged at 14000g for 10 min, and the supernatant was removed. The pellet was resuspended again in 200 μL methanol and incubated at 4 °C for 4−5 days. After centrifugation, the supernatants were pooled. Finally, the Gb4Cer extracts were dried under an air stream (45 °C). PrimaDrop (pd) Sample Preparation and MALDI-TOF/ TOF Mass Spectrometry. Neutral rat kidney extract was dissolved in 2 μL ACN/H20 (1/1, v/v)/mg dry weight and diluted 1/25, 1/50, and 1/100. Twenty-five micrograms per microliters Gb4Cer standard was prepared in the following solvents: ACN/H20 (1/1, v/v), ACN/H20 (8/2, v/v), MeOH/ H20 (1/1, v/v), Aceton/H20 (1/1, v/v), or CHCl3/MeOH (2/ 1, v/v). One microliter of each of these mixtures was dispensed on a ground steel target and then dried with a hair dryer. Dissolved in 100% MeOH were 66.7 mg/mL ATT, 2,5-DHB, or THAP, and then chloroform was added to a final matrix concentration of 20 mg/mL in chloroform/MeOH (7/3, v/v) and sonicated for 5 min. Dried spots were coated with 0.8 μL of matrix solution. Evaporation of the solvent and cocrystallization with the analyte took place within a few seconds. For evaluation of salt additives, the Gb4Cer standard [50 μg/mL in ACN/H20 (1/1, v/v)] was diluted (1:1) with either 5 mM NaCl in H2O, 10 mM lithium citrate (LiCit) in H2O, or ACN/H20 (1/1, v/ v). For comparison of these MALDI preparations, seven spots were analyzed for each combination. Gb4Cer-containing extracts from scraped silica plates were suspended in 30 μL of MeOH. Relative quantitation was done by mixing extracted Gb4Cer (1:2) with 5 mM NaCl (in H2O) containing bradykinin (final concentration of 330 fmol/μL). A peptide as an internal standard was used, as synthetic GSL standards in an appropriate mass range are not commercially available. Afterward, samples were processed as described above. For evaluation of MALDI preparation methods and matrix/ additive mixtures, an AutoXecute method was defined to sum up 4000 laser shots at 40 different positions of the sample. Only spectra with a resolution >2500 were accumulated and the acquisition was quit after 50 consecutive failed judgments. For relative quantitation of Gb4Cer from PKD kidneys, six spots were prepared for each sample. Spectra with resolution >2000 were summed up and 3500 laser shots at 35 different positions B
dx.doi.org/10.1021/ac400931u | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
Figure 1. An optimized solvent system for identification of Gb4Cer isoforms from PKD rat kidney extracts by combined HPTLC and MALDI-TOFMS. (A) Successful separation of SM and Gb4Cer from rat kidney extracts. A mixture of SM, Gb4Cer, and Gb3Cer lipid standards as well as a neutral lipid extract from PKD kidneys were separated in chloroform/methanol/0.2% aqueous CaCl2 [solvent system 1: 60/35/8, v/v/v (left panel) or optimized solvent system 2:50/35/8, v/v/v (right panel)]. Bands were visualized with primulin and scanned at 366 nm in the fluorescence mode. Red boxes indicate similar Rf values, the blue box in the right panel specifies the scraped band subjected to MALDI-TOF-MS analysis. (B) MALDITOF-MS spectrum of lipids comigrating with the Gb4Cer standard. (C) Representative MS/MS spectrum of Gb4Cer (d18:1,16:0 + Na+) (m/z 1250.0) confirms the identity of Gb4Cer isoforms. Complete series of B-type ions were derived, starting with m/z 226.06 representing sodiated HexNac and ending up with m/z 712.2 as the B4 ion. C-type ions were also detected, dependent on the cleavage site of the glycosidic bond from C2 to C4. Neutral loss of a HexNAc moiety at m/z 1046.9 with subsequent loss of two hexoses results in Y-type ions (m/z 884.8 and m/z 722.8). The fragment ions are assigned according to the following:46 blue circle = Glc, yellow circle = Galß, green circle = Galα, and yellow square = GalNac.
were accumulated with 4 × 25 shots per position in a random walk. UHPLC-ESI-QqQ-MS/MS. Quantification of globosides was adapted from an LC−ESI-MS/MS method described for other sphingolipids23 with the following modifications: the gradient for elution of tetraosylceramides was adjusted (see the Supporting Information). An internal Galβ1-3GalNacβ14Galβ1-4Glcβ1-Cer (Gg4Cer) standard [Gg4Cer(d18:1;19:0)] and Gg4Cer(d20:1;19:0) were synthesized in-house and used for quantification of globosides. The internal standard was calibrated with an external Gb4Cer standard from porcine red blood cells (Matreya LLC; Pleasant Gap, PA). Under identical conditions, the Gg4Cer standard was detected with 2.46 ± 0.03 times more sensitivity than corresponding concentrations of the
Gb4Cer standard. This correction factor was used when renal globosides were quantified. Selected reaction monitoring (SRM) of sphingolipids used the following transitions of protonated molecular ions to product ions: [HexNAc − OH]+ (m/z 204.1, 55 eV collision energy) for globosides and [HexHexNAc − OH]+ (m/z 366.1, 20 eV collision energy) for Gg4Cer. Globosides with identical nominal mass but a different ceramide backbone were distinguished by their retention times (Figure S5 of the Supporting Information).
■
RESULTS AND DISCUSSION Optimized TLC Solvent Mixture Enables Indirect TLC− MALDI-TOF Analysis of Gb4Cer. Pursuing our goal of developing a method for characterization of the complex
C
dx.doi.org/10.1021/ac400931u | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
with DHB and ATT by manual exploration of “sweet spots”, inhomogeneous distribution of the analyte/matrix cocrystals resulted in aborted measurements in automated data acquisition and could not be significantly improved with other crystallization methods, such as fast evaporation, twolayer, or sandwich crystallization. We therefore developed a method for lipid analytes that we term PrimaDrop: first, the lipid sample is deposited on a ground steel target in a solvent of choice. After drying, a concentrated matrix solution in a fast evaporating solvent without the sample is applied (Figure S2A of the Supporting Information). As the organic solvents chloroform and methanol are widely used for extraction of lipophilic compounds from tissue or body fluids, we found a chloroform/methanol mixture (7/3, v/v) as the matrix solvent to be preferred over acetone or methanol, as it remains constrained to a single spot on the target. With the use of PrimaDrop sample preparation, nearly homogeneous crystal distribution of the matrix only (Figure S2B of the Supporting Information), neutral kidney lipid extract, Gb4Cer standard and silica-extracted GSLs were achieved with ATT, DHB, and THAP and homogeneity was not influenced by addition of NaCl to the lipid sample (Figure S2, panels C−E, of the Supporting Information). With routine clinical analytical work in mind that requires the ease and reproducibility of automated data acquisition, we next sought to evaluate the performance of the PrimaDrop (pd) versus the dried droplet (dd) method with 1/25, 1/50, and 1/ 100 diluted neutral kidney lipid extract and matrices ATT and DHB. Parameters for evaluation of crystallization as well as of the matrix best suited for detection of neutral lipids were the number of peaks detected in the mass range of 700−2000 Da (Figure S3, panels A−C, of the Supporting Information) and S/ N values of three representative SM isoforms: SM(d18:1,h16:0) as the most prominent ion in the rat kidney extract, SM(d18:1,h24:0) with medium intensity, and SM(d18:24:0) as a minor signal (all as [M + Na+]+) (Figure S3, panels D−F, of the Supporting Information). As expected, Gb4Cer could not be detected in the kidney extract without TLC separation. The number of lipid peaks increased slightly for ATT with PrimaDrop instead of dried droplet (especially with a 1/100 diluted extract) and was considerably higher for the DHB PrimaDrop compared to the dried droplet (Table 1 and Table S1, sections A and B, and Figure S3, panels A−C, of the Supporting Information). Improved ionization efficiency of PrimaDrop over dried droplet preparation was furthermore reflected by significantly higher S/N values for all evaluated SM isoforms (Figure S3, panels D−F, of the Supporting Information). S/N fold-changes of SM species were calculated for the combinations ATTpd/ ATTdd and DHBpd/DHBdd. With dependence on the matrix and concentration of the lipid analyte, S/N values were found to be unchanged or up to 12-fold higher in PrimaDrop preparations (Table 1 and Table S1, sections A and B, of the Supporting Information). ATT matrix was slightly beneficial over DHB in PrimaDrop preparation (ATTpd/DHBpd), with fold changes in the range of 1.3 to 3.1. Finally, the overall improvement with ATT with PrimaDrop crystallization over the standard method, DHB in dried droplet preparation (ATTpd/DHBdd), was assessed: Of note, S/N values for SM(d18:1,h16:0) were 3.8−5.2, for SM(d18:1,h24:0) 7.6− 14.9, and for SM(d18:1,24:0) 8- to 16-fold increased when the optimized sample preparation method was used (Table 1 and Table S1, sections A and B, of the Supporting Information).
neutral GSL Gb4Cer (Figure S1A of the Supporting Information) using a combination of HPTLC and MALDITOF-MS, we separated the neutral GSL extract from the PKD rat kidney by HTPLC in a solvent mixture typically used for TLC analysis of GSLs.42 Total lipids were stained with primulin and glycolipids specifically with orcinol (Figure S1B of the Supporting Information and Figure 1A). A primulin-stained band comigrating with Gb4Cer standard was scraped off the silica gel, extracted with methanol, and analyzed by MALDITOF-MS/MS in dried droplet preparation with manual measurement. Despite strong staining of the scraped-off band indicating relevant amounts of Gb4Cer in the eluate, only a few ions could be assigned to Gb4Cer (Figure S1C of the Supporting Information). MALDI-TOF-MS/MS analysis of m/z 835.90 identified SM as the predominant ion species (Figure S1D of the Supporting Information), suggesting that it was either more abundant or suppressed ionization of Gb4Cer in the silica eluate. Therefore, we aimed to improve the TLC solvent mixture. When compared to the nonresolving standard solvent system, a chloroform/methanol/CaCl2 (50/35/8) mixture enabled baseline separation of all lipid standards (Figure 1A). Applied to neutral PKD rat kidney GSL extract, the optimized solvent mixture baseline-separated a peak comigrating with the Gb4Cer standard from the major peak corresponding to vastly abundant lipids comigrating with the SM standard (Figure 1A). To examine whether baseline separation of peaks comigrating with SM and Gb4Cer resulted in improved spectral quality for Gb4Cer isoforms, we scraped off the primulin-stained (presumably Gb4Cer-containing) band and subjected the eluate to MALDI-TOF-MS analysis (Figure 1, panels A and B). In the m/z range of expected globoside ions, we now detected intense signals, all representing sodium adducts of globosides. Ion intensities in the m/z range from 1250 to 1400 were from 3- to 10-fold higher than in the eluate obtained from the extract separated in the solvent system 1 (Figure S1 of the Supporting Information). The most prominent ion was m/z 1249.79, likely corresponding to Gb4Cer (d18:1,16:0) (Figure 1B). To confirm the proposed structures of Gb4Cer isoforms, we further analyzed metastable fragmentation of the most intense MALDI-derived ions, as shown representatively for m/z 1249.80 {[Gb4Cer (d18:1,16:0) + Na+]+} (Figure 1C). Neutral loss of an Nacetyl hexosamine (HexNAc) moiety followed by sequential loss of two hexoses resulted in a characteristic series of Y ions (Y3 toY1). In addition, a corresponding series of B ions [B1 (sodium adduct of HexNAc) to B4] was detected. This fragmentation pattern is strongly indicative of Gb4Cer (Figure 1C). Neutral tetraosylceramides, which are members of the ganglio-, lacto-, or neolacto series, would be detected with an undistinguishable m/z value for the parent ion, but in contrast to globosides, the fragmentation pattern would be different. For instance, Galβ1-3GalNAcβ1-4Galβ1-4Glcβ-Cer (asialo-ganglioside GM1) would be observed with the neutral loss of a hexose with m/z 162.0 for Y3 and the further neutral loss with m/z 203 for Y2 ions. PrimaDrop Sample Preparation As an Effective Method for Automated MALDI-TOF Analysis of Neutral Lipid Extracts and of Gb4Cer from Extracted TLC Silica Bands. Dried droplet preparations of lipid analytes for MALDI-TOF-MS analysis typically formed rings of large crystal aggregates around the edge of the spots for all matrices tested (Figure S2, panels B−E of the Supporting Information). Although quality spectra of silica-eluted Gb4Cer were obtained D
dx.doi.org/10.1021/ac400931u | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
dried droplet preparation failed with all of the tested matrices and could only be achieved with PrimaDrop (Table 2). With
Table 1. Evaluation of PrimaDrop (pd) vs Dried Droplet (dd) Sample Preparation of Neutral Kidney Lipid Extract with MALDI Matrices ATT and DHB by Automated MALDI-TOF-MS Data Acquisitiona dilution factor
1/25
1/50
number of peaks ATT pd 81 ± 7 75 ± 10 ATT dd 73 ± 11 65 ± 11 DHB pd 58 ± 11 71 ± 22 DHB dd 27 ± 1 32 ± 2 p value 0.95) between HPTLC−MALDI-TOF-MS and UPLC−ESI-MS/MS for fold-change values (healthy vs diseased kidney extracts) measured for the most ion-intense Gb4Cer isoforms (Figure 4C). Despite the striking correlation between relative quantification in HPTLC-MALDI-TOF MS and absolute quantification by UPLC−MS/MS, we observed a trend toward higher fold-change values when measured by UPLC−ESI-MS/MS (Figure 4C).
REFERENCES
(1) Gault, C. R.; Obeid, L. M.; Hannun, Y. A. Adv. Exp. Med. Biol. 2010, 688, 1−23. (2) Kolter, T.; Sandhoff, K. Biochim. Biophys. Acta 2006, 1758, 2057− 79. (3) Fujitani, N.; Takegawa, Y.; Ishibashi, Y.; Araki, K.; Furukawa, J.; Mitsutake, S.; Shinohara, Y. J. Biol. Chem. 2011, 286, 41669−79. (4) Mather, A. R.; Siskind, L. J. Adv. Exp. Med. Biol. 2011, 721, 121− 38. (5) Xu, Y. H.; Barnes, S.; Sun, Y.; Grabowski, G. A. J. Lipid Res. 2010, 51, 1643−75. (6) Summers, S. A. Curr. Opin. Lipidol. 2010, 21, 128−35. (7) Marks, N.; Berg, M. J.; Saito, M. Brain Res. 2008, 1191, 136−47. (8) Gabow, P. A.; Johnson, A. M.; Kaehny, W. D.; Kimberling, W. J.; Lezotte, D. C.; Duley, I. T.; Jones, R. H. Kidney Int. 1992, 41, 1311−9. (9) Iglesias, C. G.; Torres, V. E.; Offord, K. P.; Holley, K. E.; Beard, C. M.; Kurland, L. T. Am. J. Kidney Dis. 1983, 2, 630−9. (10) Neudecker, S.; Walz, R.; Menon, K.; Maier, E.; Bihoreau, M. T.; Obermuller, N.; Hoffmann, S. C. Am. J. Pathol. 2010, 177, 3000−9. (11) Chatterjee, S.; Shi, W. Y.; Wilson, P.; Mazumdar, A. J. Lipid Res. 1996, 37, 1334−44. (12) Natoli, T. A.; Smith, L. A.; Rogers, K. A.; Wang, B.; Komarnitsky, S.; Budman, Y.; Ibraghimov-Beskrovnaya, O. Nat. Med. 2010, 16, 788−92. (13) Kruger, R.; Bruns, K.; Grunhage, S.; Rossmann, H.; Reinke, J.; Beck, M.; Lackner, K. J. Clin. Chem. Lab. Med. 2010, 48, 189−98. (14) Murphy, R. C.; Axelsen, P. H. Mass Spectrom. Rev. 2011, 30, 579−99. (15) Fuchs, B.; Suss, R.; Schiller, J. Prog. Lipid Res. 2010, 49, 450−75. (16) Zarei, M.; Kirsch, S.; Muthing, J.; Bindila, L.; Peter-Katalinic, J. Anal. Bioanal. Chem. 2008, 391, 289−297.
■
CONCLUSION Little is known about the function of globo series GSLs after 25 years of research. For challenging complex neutral GSLs, a broad range of relative and absolute quantitative analytical workflows on different types of MS instruments is therefore required. To this end, we introduce indirect TLC−MALDITOF-MS for relative quantification of lipid extracts and absolute LC−ESI-MS quantification for tissue Gb4Cer. Normalizing to an internal standard is an established method for the quantitative analysis of small molecules in MALDITOF-MS, as it overcomes some of the variability (crystallization irregularities, variations in desorption, gas-phase effects) inherent in this approach.43 However, few studies have dealt with quantitative MALDI data of lipids, all investigating less complex mixtures or lipid classes that are ionized more easily.44,45 Here, we show that relative quantities of lipids determined by indirect TLC−MALDI-TOF-MS correlate well with absolute quantified UPLC−ESI-MS/MS data for pooled G
dx.doi.org/10.1021/ac400931u | Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
(17) Suchanowska, A.; Kaczmarek, R.; Duk, M.; Lukasiewicz, J.; Smolarek, D.; Majorczyk, E.; Jaskiewicz, E.; Laskowska, A.; Wasniowska, K.; Grodecka, M.; Lisowska, E.; Czerwinski, M. J. Biol. Chem. 2012, 287, 38220−38230. (18) Okuda, T.; Nakakita, S.; Nakayama, K. Glycoconjugate J. 2010, 27, 287−96. (19) Fuchs, B.; Schiller, J.; Suss, R.; Zscharnack, M.; Bader, A.; Muller, P.; Suckau, D. Anal. Bioanal. Chem. 2008, 392, 849−60. (20) Liang, Y. J.; Yang, B. C.; Chen, J. M.; Lin, Y. H.; Huang, C. L.; Cheng, Y. Y.; Yu, J. Stem Cells 2011, 29, 1995−2004. (21) Marsching, C.; Eckhardt, M.; Grone, H. J.; Sandhoff, R.; Hopf, C. Anal. Bioanal. Chem. 2011, 401, 53−64. (22) Cheng, H.; Sun, G.; Yang, K.; Gross, R. W.; Han, X. J. Lipid Res. 2010, 51, 1599−609. (23) Jennemann, R.; Kaden, S.; Sandhoff, R.; Nordstrom, V.; Wang, S.; Volz, M.; Grone, H. J. J. Biol. Chem. 2012, 287, 32598−616. (24) Tanaka, K.; Yamada, M.; Tamiya-Koizumi, K.; Kannagi, R.; Aoyama, T.; Hara, A.; Kyogashima, M. Glycoconjugate J. 2011, 28, 67− 87. (25) Suzuki, A.; Miyazaki, M.; Matsuda, J.; Yoneshige, A. Biochim. Biophys. Acta 2011, 1811, 861−74. (26) Souady, J.; Soltwisch, J.; Dreisewerd, K.; Haier, J.; PeterKatalinic, J.; Muthing, J. Anal. Chem. 2009, 81, 9481−92. (27) Distler, U.; Hulsewig, M.; Souady, J.; Dreisewerd, K.; Haier, J.; Senninger, N.; Muthing, J. Anal. Chem. 2008, 80, 1835−46. (28) Dreisewerd, K.; Muthing, J.; Rohlfing, A.; Meisen, I.; Vukelic, Z.; Peter-Katalinic, J.; Berkenkamp, S. Anal. Chem. 2005, 77, 4098−107. (29) Fuchs, B.; Suss, R.; Schiller, J. Prog. Lipid Res. 2011, 50, 132. (30) Vorm, O.; Roepstorff, P.; Mann, M. Anal. Chem. 1994, 66, 3281−3287. (31) Dai, Y.; Whittal, R. M.; Li, L. Anal. Chem. 1999, 71, 1087−1091. (32) Onnerfjord, P.; Ekstrom, S.; Bergquist, J.; Nilsson, J.; Laurell, T.; Marko-Varga, G. Rapid Commun. Mass Spectrom. 1999, 13, 315−322. (33) Munteanu, B.; von Reitzenstein, C.; Hansch, G. M.; Meyer, B.; Hopf, C. Anal. Bioanal. Chem. 2012, 404, 2277−86. (34) Garaguso, I.; Borlak, J. Proteomics 2008, 8, 2583−2595. (35) Puddick, J.; Prinsep, M. R.; Wood, S. A.; Craig, C. S.; Hamilton, D. P. Phytochem. Anal. 2012, 23, 285−91. (36) Schumaker, S.; Borror, C. M.; Sandrin, T. R. Rapid Commun. Mass Spectrom. 2012, 26, 243−53. (37) Teuber, K.; Schiller, J.; Fuchs, B.; Karas, M.; Jaskolla, T. W. Chem. Phys. Lipids 2010, 163, 552−60. (38) Papac, D. I.; Wong, A.; Jones, A. J. Anal. Chem. 1996, 68, 3215− 23. (39) Stübiger, G.; Belgacem, O. Anal. Chem. 2007, 79, 3206−13. (40) Stübiger, G.; Belgacem, O.; Rehulka, P.; Bicker, W.; Binder, B. R.; Bochkov, V. Anal. Chem. 2010, 82, 5502−10. (41) Cerruti, C. D.; Touboul, D.; Guerineau, V.; Petit, V. W.; Laprevote, O.; Brunelle, A. Anal. Bioanal. Chem. 2011, 401, 75−87. (42) Meisen, I.; Mormann, M.; Muthing, J. Biochim. Biophys. Acta 2011, 1811, 875−96. (43) van Kampen, J. J.; Burgers, P. C.; de Groot, R.; Gruters, R. A.; Luider, T. M. Mass Spectrom. Rev. 2011, 30, 101−20. (44) Fujiwaki, T.; Tasaka, M.; Takahashi, N.; Kobayashi, H.; Murakami, Y.; Shimada, T.; Yamaguchi, S. J. Chromatogr., B 2006, 832, 97−102. (45) Fujiwaki, T.; Tasaka, M.; Yamaguchi, S. J. Chromatogr., B 2008, 870, 170−6. (46) Domon, B.; Costello, C. E. Biochemistry 1988, 27, 1534−43.
H
dx.doi.org/10.1021/ac400931u | Anal. Chem. XXXX, XXX, XXX−XXX
