Binary Matrix for MALDI Imaging Mass Spectrometry of

Suki Kang , Heysun Maeng , Baek Gil Kim , Gao Ming Qing , Yoon Pyo Choi , Hak Yong ... Karin A. Zemski Berry , Joseph A. Hankin , Robert M. Barkley , ...
0 downloads 0 Views 3MB Size
Download PDF
ARTICLE pubs.acs.org/ac

Binary Matrix for MALDI Imaging Mass Spectrometry of Phospholipids in Both Ion Modes Selina Rahman Shanta,† Li-Hua Zhou,†,‡ Young Seung Park,§ Young Hwan Kim,§,z Youngjun Kim,^ and Kwang Pyo Kim*,† †

Department of Molecular Biotechnology, WCU Program, Konkuk University, Seoul 143-701, Korea Faculty of Chemical Engineering and Light Industry, Guangdong University of Technology, Guangzhou 510006, China § Division of Mass Spectrometry Research, Korea Basic Science Institute, Ochang 863-883, Korea z Graduate School of Analytical Science and Technology, Chungnam National University, Daejeon 305-764, Korea ^ Department of Applied Biochemistry, Konkuk University, Chungju, Chungbuk, Korea ‡

bS Supporting Information ABSTRACT: Phospholipids (PLs) are the major building block molecules of cellular membranes. Their composition varies depending on cell types and cellular compartments. Thus, the information regarding PL distribution in tissue has important physiological and pathological significance. Recent developments in imaging mass spectrometry (IMS) have allowed complete mapping of the PL species on tissue. The IMS technique can detect different classes of PLs as well as their location information directly from tissue sections. PL head groups carry either positive and/or negative charges; therefore, IMS experiments must be conducted in both positive- and negative-ion mode to detect all types of phospholipids. Several conventional matrixes were applied on tissue for better identification. This study was conducted to enable appropriate matrix selection and optimized matrix preparation for IMS experiments in both ion modes that maximize PL identification from a single brain tissue section. The optimized matrix 2,5dihydroxybenzoic acid (DHB) and R-cyano-4-hydroxycinnamic acid (CHCA) with a mixture of trifluoroacetic acid (TFA) and piperidine as ion pairing agents showed improved stability and consistency during both ion mode experiments and successfully identified >100 peaks of PLs determined by parent ion m/z value. Further tandem mass spectrometric analysis (MS/MS) was performed to those PLs that are anatomically important according to their distribution on rat brain tissue section.

L

ipids, especially phospholipids (PLs), are the building block components of cellular membrane. These compounds have diverse chemical structures and cellular functions such as neurotransmission, signal transduction, vesicular trafficking, apoptosis, and post-translational modifications.1,2 Lipids are enriched in the central nervous system, especially the brain, in which half of the dry weight consists of lipids.3 Lipids can be further divided into glycerophospholipids,4 as complex lipids,5-7 or cholesterols8 and fatty acids as simple lipids according to their structures and functions. Although little is known about the specific localization of lipids in different brain regions, changes in the phospholipid distribution in specific brain regions have been found to be associated with development and disease.9 Biosynthetic pathways for phospholipid production appear to operate in different cellular mechanisms; thus, the composition of PLs may reflect different cellular states. Because metabolic disorders of PLs are directly related to various neurological r 2011 American Chemical Society

diseases such as Alzheimer’s disease, Niemann-Pick disease, and Parkinson’s disease, information regarding their location on tissues is useful for diagnosis and treatment.10,11 Recent progress in imaging mass spectrometry (IMS) has made it possible to identify important cellular components such as PLs, peptides, drugs, and other endogenous molecules directly on tissue sections.12-14 IMS has been proven to be a powerful tool for visualizing the molecular distributions of biological molecules in both healthy and diseased tissues.15 Similarly, IMS provides superior information regarding the distinct arrangement of different cellular molecules on tissue sections. Unlike other techniques for phospholipid analysis, with IMS, it is possible to Received: July 28, 2010 Accepted: December 23, 2010 Published: January 18, 2011 1252

dx.doi.org/10.1021/ac1029659 | Anal. Chem. 2011, 83, 1252–1259

Analytical Chemistry elucidate the spatial localization of PLs throughout the entire tissue region as well as their structural characterization. Due to the presence of amine or hydroxyl groups, several types of phospholipids prefer to carry positive or negative charges on their head groups. To detect all types of PLs on a tissue section, both positive- and negative-ion IMS detection modes must be conducted. However, difficulty always arises from the utilization of some common matrixes with low consistency on tissue. Due to their low vacuum stability, most of the commonly used matrixes are vaporized after one IMS experiment. There are several different matrixes that can be used for phospholipid identification in tissue sections or extracted samples, including 2,5-dihydroxybenzoic acid (DHB), R-cyano-4-hydroxycinnamic acid (CHCA), 9-aminoacriaine (9-AA), and 2-mercaptobenzothiazole (2-MBT).1,7,16,17 However, some drawbacks of these matrixes prevent them from being applied to IMS of PLs directly on tissue. For example, some matrixes only perform well in a particular mode of analysis, some are highly influenced by the solvent system, and some show high matrix-related background peaks. Moreover, some matrixes have relatively low stability under vacuum.18 Recently, ionic matrixes have been successfully introduced for analysis of biological molecules such as peptides and proteins directly from tissue sections.19 Meriaux et al. introduced 2,5 DHB based new ionic matrix for phospholipid imaging on tissue. The newly developed ionic matrix made a very homogeneous spot on the tissue and also was stable in vacuum system; however, the number of identified phospholipids from the rat brain tissue section was ∼40 according to the m/z value in both ionization modes as the referred paper reported.20 In this study, we attempted to overcome the drawbacks of different types of matrixes mentioned above by combining two known and frequently used matrixes: DHB and CHCA. Additionally, trifluoroacetic acid (TFA)21 and piperidine22 at a molar ratio of 4:1 were added to give the binary matrix ionic liquid properties. The binary matrix provided a homogeneous deposition on tissue, resulting in higher reproducibility and signal intensities when compared to any single traditional matrixes.23-27 The formation of liquid salt by adding a defined amount of acid and base can help the binary matrix to be stable under the high vacuum of the MALDI system and improve the consistency rate of the matrix on tissue. The advantages of the binary matrix over the single matrix were demonstrated by IMS of phospholipids on the same rat brain tissue section with sequential acquisitions in both polarities followed by tandem mass spectrometry (MS/MS) for further structural validation. Here, we present the successful application of the binary matrix in IMS of PLs on rat brain tissue sections by identification of the above 100 peaks of PLs according to their m/z value with structural information confirmed by MS/MS.

’ EXPERIMENTAL SECTION Materials. 2,5-Dihydroxybenzoic acid (DHB) and R-cyano4- hydroxycinnamic acid (CHCA) were obtained from Bruker Daltonics (Germany) and used without purification. All solvents were HPLC grade. Trifluoroacetic acid (TFA) and piperidine (spectrometric grade) were purchased from Sigma-aldrich. A mixture of seven peptide calibration standards was used for external calibration. Peptide calibration standard was purchased from Bruker Daltonics (Billerica, MA, USA). Tissue Sectioning. All experiments with rats were conducted under standard conditions. Animals were sacrificed and immediately

ARTICLE

dissected to remove the brain, after which they were stored at -80 C. The frozen tissue was then sectioned to a thickness of 12 μm at -20 C using a Cryo-microtome (Leica). The optimum cutting temperature polymer (OCT) can deteriorate signals from IMS;28 special care should be taken to avoid OCT fixing on sections when the brain tissue was stabilized by it. The frozen tissue sections were then thaw-mounted on indium-tin-oxide (ITO) coated glass slides (Bruker Daltonics), after which they were dried in desiccators for about 20 min and then stored at -80 C until use. Sample Preparation for MALDI-IMS. The binary matrix solution was prepared by dissolving 7 mg of each of DHB and CHCA in 1 mL of 70% methanol plus 0.1% TFA and 1% piperidine. An ImagePrep instrument (Bruker Daltonics) was used to spray a total of 3 mL of matrix solution on one tissue section. The optimal parameters (dry time, incubation time, and thickness) of the ImagePrep instrument were set to obtain homogeneous matrix crystal on tissue. After matrix application, the homogeneity of matrix on the tissue was checked by the imaging function of Chip-1000 (Shimadzu Biotech). The MTP slide adapter (Bruker Daltonics) mounted with the tissue section ITO slide was directly transferred to the MALDI mass spectrometer. MALDI-TOF MS and MS/MS. IMS spectra were acquired in both positive- and negative-ion reflector modes with the aid of a Bruker Daltonics Autoflex III MALDI time-of-flight mass spectrometer equipped with smart beam laser technology (Bremen, Germany). MS data were acquired in the m/z range between 0 and 3000 by averaging signals from 500 consecutive laser shots with a frequency of 200. All experiments conducted in this study were independently repeated three times. The spatial resolution of all of the mass images shown was 200 μm. Prior to each data acquisition, external calibration was conducted using a peptidemixed calibration standard with a m/z range of 800-3200. All obtained spectra were baseline subtracted. Data analysis software (FlexImaging ver. 2.0, FlexAnalysis ver. 3.0, and Clinprotools ver. 2.1) was employed to obtain the density maps of the species of interest and to check the differences in the intensity of each spectrum. The MALDI LIFT (MS/MS) analysis was directly conducted in the reflection on the tissue section after MALDI IMS. The LIFT data and some lipid databases (http://lipidsearch.jp or www.lipidmaps.org) were utilized to facilitate and confirm the assignment of phospholipid species.

’ RESULTS AND DISCUSSION Recently, IMS has been highlighted as a promising tool for visualization and identification of biomolecules on tissue sections. In addition to genomics or proteomics, the study of lipidomics has been considered to be of high importance in systems biology. IMS of lipids can provide a profile of PLs as well as spatial distribution of PLs on tissue sections. The major scheme of IMS experiments for PLs in this study is shown in Scheme 1. The initial work was conducted to determine the optimum matrix and matrix concentration for homogeneity, vacuum stability (Figure S1 in the Supporting Information), and improved signal intensities of PLs on rat brain tissue sections. To accomplish this, we optimized our IMS experiments with 2,5-DHB, SA, CHCA, or combinations of matrixes by the addition of peperidine and TFA as ion pairing agents. Data acquisition in both negative- and positiveion modes are required to increase the number of phospholipids identified and to cover both anionic and cationic phospholipid species in single IMS on the same tissue sections. 1253

dx.doi.org/10.1021/ac1029659 |Anal. Chem. 2011, 83, 1252–1259

Analytical Chemistry

ARTICLE

Scheme 1. Workflow of Imaging Mass Spectrometry

We used a 1:1 (w/w) mixture of 2,5-DHB and CHCA in 70% ethanol with 0.1% TFA and 1% piperidine. As shown in Figure 1, compared with DHB or CHCA, the binary matrix showed greatly enhanced performance when compared to other common matrixes. Specifically, it provided more peaks of phospholipids with increased S/N ratios. In the mentioned m/z region in Figure 1, we clearly identified some peaks which were totally absent or had very low intensities (indicated with asterisk) with single matrix DHB or CHCA. For example, in positive mode, more peaks were detected with higher intensities with the binary matrix than DHB or CHCA in especially m/z regions between 730 and 760 and 780 and 830. Some important phospholipids such as m/z 734 (PC 32:0), 760 (PC 34:1), 782 (PC 34:1 Na), 788 (PC 36:1), and 810 (PC 36:1 Na) showed at least 3 times higher intensities with binary matrix than DHB or CHCA matrix in positive mode. Similarly in negative mode, binary matrix gives more peaks with higher intensities (showed with asterisk) including m/z 784 (PS 36:3), 806 (ST 18:0), 850 (ST-OH 20:0), 878 (ST-OH 40:1), 906 (ST-OH 42:1), and 920 (ST 44:0). Another most important advantage of this binary matrix is its stability under vacuum during IMS. Due to the presence of different charged groups, some phospholipids can be identified in positive ionization mode, while others can be identified in negative mode or sometimes both modes. In the present study, we applied the binary matrix solution once and performed IMS on the same tissue section in both positive- and negative-ion modes, which made it possible to identify more than 100 peaks of phospholipids including phosphotidylcholine (PC), sphingomyeline (SM), phosphatic acid (PA), phosphotidylethanolamine (PE), phosphatidylinositol (PI), and sulfatides (ST) from one tissue section (Tables S6 and S7 in the Supporting Information). Most of the spectra appeared on m/z 500-1000; however, only a few peaks were detected above m/z 1000 with low intensity. Furthermore, the

binary matrix was stable enough to conduct MS/MS analysis to characterize most of the identified phospholipids with the same tissue section following IMS data acquisition in both ionization modes. Within all identified PL groups, PC and SM generate prominent peaks in positive mode. This may be because PCs and SMs are major constituents of cell membranes.29 In addition, the presence of the positively charged quaternary ammonium group in PC enhances ion formation in positive mode. Moreover, previous studies have shown that PCs can inhibit the detection of other phospholipids in positive mode due to their easy ionization properties.8 Phosphotidylinositol (PI) is an acidic phospholipid that acts as an important component in the EGFR pathway. Due to the presence of the hydroxyl group on PI, it can always be identified in negative ionization mode. However, a large amount of sulfatides (ST), which are sphingolipids with an additional sulfate group at the 30 position of the galactose moiety in galactocerebroside, also appear in negative ionization mode of IMS experiments. Because STs constitute ∼6% of the total lipids in the adult brain, it is obvious that signals from ST suppress ionization of PI in negative ionization mode. Additionally, the efficient ionization tendency of the sulfate group facilitates ionization of STs.8 Major groups of phospholipids and glycolipids were previously identified in different experiments such as electrospray ionization (ESI)-tandem mass analysis.2,30-32 Phospholipids matched with other experiments6,11,33-36 were proposed in Tables S6 (matched in negative ionization mode) and S7 (matched in positive ionization mode), Supporting Information. The current identification of the matched phospholipids was further validated by MS/MS spectra (Figure S5i,ii in the Supporting Information). MS/MS analysis was performed for a few phospholipids that showed differential distribution according to brain anatomy. Also, in positive ionization mode, MS/MS analysis was performed for the Hþ, Naþ, and Kþ form of phospholipids. For example, 1254

dx.doi.org/10.1021/ac1029659 |Anal. Chem. 2011, 83, 1252–1259

Analytical Chemistry

ARTICLE

Figure 1. Comparison of MS spectra between DHB, CHCA, and binary matrix in (A) positive ionization mode and (B) negative ionization mode.

MS/MS analysis was conducted for PC 34:1 (m/z 760), PC 34:1 Na (m/z 782), and PC 34:1 K (m/z 798) to confirm their structural information as shown in Figure S5ii(A-C) (Supporting Information), respectively. Furthermore, PC (24:3) was identified at m/z 616 that is shown in Figure S5iiF (Supporting Information). The structural information was confirmed by MS/MS analysis where fragment appeared at m/z 557 due to neutral loss of 59 u corresponding to trimethylamine, and identical PC headgroup fragment was found at m/z 184. For additional confirmation, we used the lipid search database Web site (http:// lipidsearch.jp or www.lipidmaps.org).

Figure 2 shows the MS/MS spectra of two representative phospholipids, phosphotidylinositol (PI) and salfatide (ST), obtained in negative ionization mode and one phosphotidylcholine (PC) acquired in positive ionization mode. MS/MS analysis of the precursor ion peak at m/z 885.64 was clearly assigned as phosphotidylinositol (PI) 38:4. The fragment ion peak at m/z 580.9 corresponds to the neutral loss of arachidonic acids, which further fragmented the resulting peak at m/z 418.9 corresponding to the neutral loss of 162 (inositol unit-H2O). Two characterized fragments of the polar headgroup appear at m/z 241 (inositolphosphate-H2O) and m/z 97, which is from the [H2PO4]1255

dx.doi.org/10.1021/ac1029659 |Anal. Chem. 2011, 83, 1252–1259

Analytical Chemistry

ARTICLE

Figure 2. MS/MS analysis of major phospholipids in negative ionization mode (A and B) and in positive ionization mode (C), respectively. (A) PI 38:4 was identified at m/z 885.64, (B) ST-OH 42:1 was identified at m/z 906.60 and (C) PC 32:0 K was identified at m/z 772.7 1256

dx.doi.org/10.1021/ac1029659 |Anal. Chem. 2011, 83, 1252–1259

Analytical Chemistry

ARTICLE

Figure 3. Differential distribution of phospholipid species in rat brain tissue. (a) 12 μm thick rat brain section, (b) after matrix application to the rat brain tissue section, distribution of (c) PI 38:4, (d) ST-OH 42:1, (e) PI 38:3, (g) PC 32:0K, (h) PS 36:1K, and (i) PC 36:1K on brain tissue. (f and j) Represent differential distribution of two PI and PC species according to their m/z value. Cbc, cerebral cortex; cc, corpus callosum; LV, lateral ventricle; aca, anterior commisure; CPU, caudate putamen.

anion. Two fatty acid fragments can be found at m/z 283.1 (C18H35O2), which corresponds to stearic acid 18:0, and at m/z 303 (C20H31O2), which corresponds to aracidonic acid 20:4. Another characteristic sphingolipid, salfatide (ST), was identified at m/z 906.6. ST is one of the most abundant sphingolipids in brain tissue and can primarily be detected in negative ionization mode. When we conducted MS/MS analysis of the peak at m/z 906.6, the fragment ion showed strong signals at m/z 96.9 (HSO4)- and m/z 240.8, which indicates that it is a dehydrated galactose-sulfate moiety. The product ions detected at m/z 521.8, m/z 539.9, and m/z 567.9 confirm the presence of a hydroxylated component at m/z 906.6. The addition of 1% piperidine in the MALDI matrix was reported to improve the ion intensity in negative mode data acquisition.20 It has been suggested that the addition of piperidine improves the ion intensity, especially in negative mode. As expected, the addition of piperidine made it possible to detect two rare PIs at m/z 891.5 and m/z 893.5, which were assigned to PI 38:1 and PI 38:0, respectively. Identification of these two phospholipids was further confirmed by MS/MS analysis (Figure S2 in the Supporting Information). IMS of brain tissue in positive ionization mode resulted in a prominent peak at m/z 772.7, shown in Figure 2C. When we conducted MS/MS analysis of the major precursor ion, a fragment ion of m/z 183.9 (structure shown in Figure 2C inset) arose from the phosphatidylcholine headgroup. In addition, two fragment ion peaks of m/z 713.4 and m/z 588.1 arose from neutral loss of 59 u corresponding to trimethylamine and further neutral loss of 124 u corresponding to cyclophosphate; which are the diagnostic fragment ions for PC under MS/MS. As shown in the Figure 2C inset, another two prominent peaks appeared at m/z 162.9 (assigned as potassiated headgroup of PC) and m/z 146.9 (assigned as sodiated headgroup of PC), which indicates that multiple ion adducts are merged on a single ionized PC.37 This feature is common in MALDI-MS experiments, especially with lipids from tissue sections. Because the confined intracellular concentration of

Naþ and Kþ is highly abundant on tissue slices, these alkali adducts are often observed prominently as [M þ H] þ. MALDI-IMS spectra of the brain section can be used to concurrently identify the heterogeneous distribution of several phospholipids. Figure 3 shows the result of MALDI-IMS obtained with the binary matrix from a rat brain tissue section in both positive and negative ionization modes. Figure 3a shows the rat brain tissue section with a thickness of 12 μm, while Figure 3b shows the tissue section after binary matrix application. Figure 3c-e displays the differential distribution of PI 38:4 (m/z 885.5), ST-OH 42:1 (m/z 906.6), and PI 38:3 (m/z 887.5), respectively, in negative ionization mode. On the basis of the images obtained from IMS analysis of the brain tissue section, we can easily determine that the distribution of PI 38:4 is high in CPU (caudate putamen), LV (lateral ventricle), and Cbc (cerebral cortex) regions but relatively low in aca (anterior commisure) and cc (corpus callosum) regions. In the case of PI 38:3, we found the opposite result. We checked the merged image of these two PI species, which concur our judgment as shown in Figure 3f. We also found differential distribution of ST 42:1 and PI 38:4 and SM 18:0 K and PC 36:1 in negative and positive ion modes, respectively (Figure S3 in the Supporting Information). Similarly, we distinguished distribution of several phospholipids on the same tissue section in positive ionization mode. Figure 3g-i demonstrates the distribution of PC 32:0K (m/z 772.5), PS 36:1K (m/z 850.6), and PC 36:1K (m/z 826.5), respectively. Comparison of the image of PC 32:0K and PC 36:1K strongly suggests that they have a dissimilar distribution on brain tissue as shown in Figure 3j. Principal component analysis (PCA), an unsupervised multivariate data analysis technique,38 was conducted to identify phospholipids that showed diverse distributional prototypes on brain tissue. The result of PCA illustrated possible ions of interest in three different plots that showed inconsistency in spectral fluctuation in different portions of brain tissue. The variance between groups of loading data (not shown) revealed that m/z 1257

dx.doi.org/10.1021/ac1029659 |Anal. Chem. 2011, 83, 1252–1259

Analytical Chemistry

ARTICLE

Figure 4. Images of phospholipids that were found to be distinctly distributed on brain tissue sections by PCA analysis.

806, m/z 886, m/z 890, and m/z 906 had diverse distributions on different parts of the brain tissue in negative ionization mode. These m/z values stand for ST 36:0, ST 42:3, ST 42:1, and ST-OH 42:1, respectively. In positive ionization mode, dissimilar distribution was found at m/z 798 (PC 34:1 K), m/z 734 (PC 32:0), m/z 772 (PC 32:0 K), and m/z 826 (PC 36:1 K). Figure 4 shows the ion density map produced from imaging ions with the aforementioned m/z values. In negative ionization mode, peaks at m/z 806, m/z 890, and m/z 906 showed higher densities in the cc and aca regions than CPU and very low density in the LV region. However, peaks at m/z 886 showed higher density in almost all regions when compared to LV. In positive ionization mode, although most of the presented phospholipids are PC, m/z 772 and m/z 734 have nearly identical spatial distribution. Additionally, these peaks showed higher intensities in all regions except cc, but phospholipid with m/z 826 was shown to be elevated in all regions except cc. The peak at m/z 798 showed mainly scattered distribution. Jackson et al. found that certain phospholipids showed different intensities between white and gray matter regions in brain tissue.8 The results of the present study are similar to those of previous studies. Specifically, PCA analysis of our MALDI image data generated a list of phospholipids distributed differently on brain tissue (Figure S4 in the Supporting Information). The table shown in Figure S4A in the Supporting Information summarizes the selected phospholipids representing different regions of the brain. The images generated with these selected PLs confirmed their unique distribution in the brain (Figure S4B in the Supporting Information). These findings demonstrate that PC 32:0 shows higher intensity in the cerebral cortex and thalamus, which are part of the gray matter; however, PC 36:1 was upregulated in the corpus callosum, which is present in white matter. PI 36:4 and PI 38:4 were also present in higher concentrations in the cerebral cortex and stratum in negative ionization mode but in lower concentrations in the corpus callosum. Conversely, sulfatides showed higher signal intensity in the corpus callosum. With the acquired images, MS/MS analyses were conducted on the localized domains to allocate lipid species.

’ CONCLUSIONS In this study, we developed a new matrix system for phospholipids with a composition that could identify a greater number

of molecules, was stable under vacuum, and had a greater signal intensity and reproducibility than commonly used matrixes. The developed binary matrix showed superior performance to conventional matrixes. This improved property makes the matrix extremely valuable for IMS experiments. Here, with single application of the binary matrix to the target tissue, it was possible to conduct an IMS experiment in both positive and negative ionization modes and to conduct MS/MS analysis. This facilitated the detection of more than 100 peaks of glycerophospholipids and sphingolipids from a single tissue section in both ionization modes. Additionally, the experiments were conducted in triplicate to ensure the reproducibility of PL IMS with the binary matrix. Recent progress has suggested that other factors such as cellular localization and environment might play an important role in secondary ion emission in MALDI.39 One major drawback of our matrix system is the vast fragmentation below ∼500 Da. The binary matrix itself generated a number of peaks below 500 Da regions. In the case of very lower molecular weight sample identification, binary matrix can affect the sample peak intensity. However, this has little effect on lipid identification because most of the glycerophospholipids and sphingolipids appear between m/z 500 and 1000. To the best of our knowledge, the binary matrix developed here can identify the highest number of phospholipids from a single tissue section with structural information and provide unique distributional information regarding dissimilar brain regions.

’ ASSOCIATED CONTENT

bS

Supporting Information. Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Fax: 82-2-450-3395.

’ ACKNOWLEDGMENT S.R.S. and L.-.Z. contributed equally to this work. This research was supported by the Converging Research Center Program (2009-0093622 and 2010K001121) and WCU (World Class University) program (Project No. R33-10128) through the 1258

dx.doi.org/10.1021/ac1029659 |Anal. Chem. 2011, 83, 1252–1259

Analytical Chemistry National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science, and Technology.

’ REFERENCES (1) Astigarraga, E.; Barreda-Gomez, G.; Lombardero, L.; Fresnedo, O.; Castano, F.; Giralt, M. T.; Ochoa, B.; Rodriguez-Puertas, R.; Fernandez, J. A. Anal. Chem. 2008, 80, 9105–9114. (2) Ivanova, P. T.; Cerda, B. A.; Horn, D. M.; Cohen, J. S.; McLafferty, F. W.; Brown, H. A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 7152–7157. (3) Woods, A. S.; Jackson, S. N. AAPS J. 2006, 8, E391–395. (4) Hayasaka, T.; Goto-Inoue, N.; Sugiura, Y.; Zaima, N.; Nakanishi, H.; Ohishi, K.; Nakanishi, S.; Naito, T.; Taguchi, R.; Setou, M. Rapid Commun. Mass Spectrom. 2008, 22, 3415–3426. (5) Chen, Y.; Allegood, J.; Liu, Y.; Wang, E.; Cachon-Gonzalez, B.; Cox, T. M.; Merrill, A. H., Jr.; Sullards, M. C. Anal. Chem. 2008, 80, 2780–2788. (6) Jackson, S. N.; Wang, H. Y.; Woods, A. S. J. Am. Soc. Mass Spectrom. 2007, 18, 17–26. (7) Sugiura, Y.; Konishi, Y.; Zaima, N.; Kajihara, S.; Nakanishi, H.; Taguchi, R.; Setou, M. J. Lipid Res. 2009, 50, 1776–1788. (8) Jackson, S. N.; Wang, H. Y.; Woods, A. S. Anal. Chem. 2005, 77, 4523–4527. (9) Adibhatla, R. M.; Hatcher, J. F.; Dempsey, R. J. AAPS J. 2006, 8, E314–321. (10) Irizarry, M. C. Ann. Neurol. 2003, 54, 7–8. (11) Wenk, M. R. Nat. Rev. Drug Discovery 2005, 4, 594–610. (12) Chaurand, P.; Schwartz, S. A.; Reyzer, M. L.; Caprioli, R. M. Toxicol. Pathol. 2005, 33, 92–101. (13) Seeley, E. H.; Caprioli, R. M. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 18126–18131. (14) Zaima, N.; Hayasaka, T.; Goto-Inoue, N.; Setou, M. J. Oleo Sci. 2009, 58, 415–419. (15) Shimma, S.; Sugiura, Y.; Hayasaka, T.; Hoshikawa, Y.; Noda, T.; Setou, M. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2007, 855, 98–103. (16) Fuchs, B.; Bischoff, A.; Suss, R.; Teuber, K.; Schurenberg, M.; Suckau, D.; Schiller, J. Anal. Bioanal. Chem. 2009, 395, 2479–2487. (17) Hayasaka, T.; Goto-Inoue, N.; Zaima, N.; Kimura, Y.; Setou, M. Lipids 2009, 44, 837–848. (18) Fuchs, B.; Schiller, J. Eur. J. Lipid Sci. Technol. 2009, 111, 83–98. (19) Lemaire, R.; Tabet, J. C.; Ducoroy, P.; Hendra, J. B.; Salzet, M.; Fournier, I. Anal. Chem. 2006, 78, 809–819. (20) Meriaux, C.; Franck, J.; Wisztorski, M.; Salzet, M.; Fournier, I. J. Proteomics 2010, 73, 1204-1218. (21) Schiller, J.; Arnhold, J.; Benard, S.; Muller, M.; Reichl, S.; Arnold, K. Anal. Biochem. 1999, 267, 46–56. (22) Lytle, C. A.; Gan, Y. D.; White, D. C. J. Microbiol. Methods 2000, 41, 227–234. (23) Anderson, J. L.; Ding, J.; Welton, T.; Armstrong, D. W. J. Am. Chem. Soc. 2002, 124, 14247–14254. (24) Armstrong, D. W.; He, L.; Liu, Y. S. Anal. Chem. 1999, 71, 3873–3876. (25) Armstrong, D. W.; Zhang, L. K.; He, L.; Gross, M. L. Anal. Chem. 2001, 73, 3679–3686. (26) Carda-Broch, S.; Berthod, A.; Armstrong, D. W. Rapid Commun. Mass Spectrom. 2003, 17, 553–560. (27) Fukuyama, Y.; Nakaya, S.; Yamazaki, Y.; Tanaka, K. Anal. Chem. 2008, 80, 2171–2179. (28) Schwartz, S. A.; Reyzer, M. L.; Caprioli, R. M. J. Mass Spectrom. 2003, 38, 699–708. (29) Petkovic, M.; Schiller, J.; Muller, M.; Benard, S.; Reichl, S.; Arnold, K.; Arnhold, J. Anal. Biochem. 2001, 289, 202–216. (30) Brugger, B.; Erben, G.; Sandhoff, R.; Wieland, F. T.; Lehmann, W. D. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 2339–2344. (31) Ivanova, P. T.; Milne, S. B.; Myers, D. S.; Brown, H. A. Curr. Opin. Chem. Biol. 2009, 13, 526–531.

ARTICLE

(32) Yang, K.; Cheng, H.; Gross, R. W.; Han, X. Anal. Chem. 2009, 81, 4356–4368. (33) Dreisewerd, K.; Lemaire, R.; Pohlentz, G.; Salzet, M.; Wisztorski, M.; Berkenkamp, S.; Fournier, I. Anal. Chem. 2007, 79, 2463–2471. (34) Estrada, R.; Yappert, M. C. J. Mass Spectrom. 2004, 39, 412–422. (35) Jackson, S. N.; Wang, H. Y.; Woods, A. S.; Ugarov, M.; Egan, T.; Schultz, J. A. J. Am. Soc. Mass Spectrom. 2005, 16, 133–138. (36) Norris, C.; Fong, B.; MacGibbon, A.; McJarrow, P. Lipids 2009, 44, 1047–1054. (37) Sugiura, Y.; Setou, M. Rapid Commun. Mass Spectrom. 2009, 23, 3269–3278. (38) Zaima, N.; Matsuyama, Y.; Setou, M. J. Oleo Sci. 2009, 58, 267–273. (39) Murphy, R. C.; Hankin, J. A.; Barkley, R. M. J. Lipid Res. 2009, 50 (Suppl), S317–322.

1259

dx.doi.org/10.1021/ac1029659 |Anal. Chem. 2011, 83, 1252–1259