Mass Imaging and Identification of Biomolecules with MALDI-QIT-TOF

Anal. Chem. 2008, 80, 878-885

Technical Notes

Mass Imaging and Identification of Biomolecules with MALDI-QIT-TOF-Based System Shuichi Shimma,†,‡,| Yuki Sugiura,‡ Takahiro Hayasaka,† Nobuhiro Zaima,§ Mineo Matsumoto,§ and Mitsutoshi Setou*,†,‡,§

Okazaki Institute for Integrative Bioscience, National Institute of Natural Sciences, Okazaki, Aichi 444-8787, Japan, Department of Bioscience and Biotechnology, Tokyo Institute of Technology, Yokohama, Kanagawa 226-8501, Japan, and Molecular Gerontology Group, Mitsubishi Kagaku Institute of Life Sciences (MITILS), Machida, Tokyo 194-8511, Japan

Imaging mass spectrometry is becoming a popular visualization technique in the medical and biological sciences. For its continued development, the ability to both visualize and identify molecules directly on the tissue surface using tandem mass spectrometry (MSn) is essential. We established an imaging system based on a matrix-assisted laser/desorption ionization quadrupole ion trap time-offlight type instrument (AXIMA-QIT, Shimadzu, Kyoto, Japan), which was compatible with both imaging and highly sensitive MSn. In this paper, we present the operating conditions of the AXIMA-QIT as an imaging instrument and introduce the data converter we developed that is available free of charge. The converted data can be applied to Biomap, the commonly used visualization software. For the feasibility experiments, we demonstrated the visualization of phospholipids, glycolipid, and trypticdigested proteins in the mouse cerebellum. The visualized lipids were successfully identified by MSn directly on the tissue surface, with a strong ability to isolate precursor ions. In the analysis of tryptic-digested proteins, we compared the product ion spectra between AXIMA-QIT and a tandem TOF-type instrument. The results confirmed that AXIMA-QIT can provide a high quality of product ion spectra even on the tissue surface. Conventional morphological examination has been performed in part by the labeling technique, which is primarily immunohistochemistry at the level of either light1 or electron2 microscopy. The labeling technique also involves an overexpression of fusion * Corresponding author: (e-mail) [email protected]; (fax) +81-564-59-5291; (tel) +81-564-59-5267. † Okazaki Institute for Integrative Bioscience. ‡ Tokyo Institute of Technology. § Mitsubishi Kagaku Institute of Life Sciences (MITILS). | Present address: Department of Physics, Graduate School of Science, Osaka University, Osaka, Japan. (1) Ikegami, K.; Mukai, M.; Tsuchida, J. I.; Heier, R. L.; Macgregor, G. R.; Setou, M. J. Biol. Chem. 2006, 281, 30707-30716. (2) Setou, M.; Nakagawa, T.; Seog, D. H.; Hirokawa, N. Science 2000, 288, 1796-1802.

878 Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

protein of the concerned molecule and green fluorescent protein.3,4 However, the limitation of these techniques is a so-called selfcontradiction: we cannot observe what we have not intended. There is also a “nonlabeling” technique even at the electron microscopic level,5 and yet there is still a serious limitation on the object preference. Imaging mass spectrometry6 is an emerging technique that is expected to at once resolve these problems in conventional morphological examination. In imaging mass spectrometry, mass spectra associated with spatial information can be simultaneously recorded to obtain expression patterns of various molecules in specimens to be analyzed. This new generation of mass spectrometry has been used for analysis of biological compounds at either the tissue7-10 or single-cell level.11-14 Recent imaging mass spectrometry studies have been conducted on a variety of topics, including biological applications15-22 and patho(3) Setou, M.; Seog, D. H.; Tanaka, Y.; Kanai, Y.; Takei, Y.; Kawagishi, M.; Hirokawa, N. Nature 2002, 417, 83-87. (4) Fukuda, Y.; Kawano, Y.; Tanikawa, Y.; Oba, M.; Koyama, M.; Takagi, H.; Matsumoto, M.; Nagayama, K.; Setou, M. Neurosci. Lett. 2006, 400, 5357. (5) Setou, M.; Radostin, D.; Atsuzawa, K.; Yao, I.; Fukuda, Y.; Usuda, N.; Nagayama, K. Med. Mol. Morphol. 2006, 39, 176-180. (6) Luxembourg, S. L.; Mize, T. H.; McDonnell, L. A.; Heeren, R. M. Anal. Chem. 2004, 76, 5339-44. (7) Brotherton, H. O.; Yost, R. A. Anal. Chem. 1983, 55, 549-553. (8) Johnson, J. V.; Yost, R. A.; Faull, K. F. Anal. Chem. 1984, 56, 1655-1661. (9) Bernier, U. R.; Kline, D. L.; Barnard, D. R.; Schreck, C. E.; Yost, R. A. Anal. Chem. 2000, 72, 747-756. (10) Garrett, T. J.; Yost, R. A. Anal. Chem. 2006, 78, 2465-2469. (11) Rubakhin, S. S.; Churchill, J. D.; Greenough, W. T.; Sweedler, J. V. Anal. Chem. 2006, 78, 7267-7272. (12) Miao, H.; Rubakhin, S. S.; Sweedler, J. V. Anal. Chem. 2005, 77, 71907194. (13) Hatcher, N. G.; Richmond, T. A.; Rubakhin, S. S.; Sweedler, J. V. Anal. Chem. 2005, 77, 1580-1587. (14) Rubakhin, S. S.; Greenough, W. T.; Sweedler, J. V. Anal. Chem. 2003, 75, 5374-5380. (15) Sugiura, Y.; Shimma, S.; Setou, M. Anal. Chem. 2006, 78, 8227-8235. (16) Shimma, S.; Furuta, M.; Ichimura, K.; Yoshida, Y.; Setou, M. J. Mass Spectrom. Soc. Jpn. 2006, 54, 133-140. (17) Shimma, S.; Furuta, M.; Ichimura, K.; Yoshida, Y.; Setou, M. Surf. Int. Anal. 2006, 38, 1712-1714. (18) Chaurand, P.; Schwartz, S. A.; Reyzer, M. L.; Caprioli, R. M. Toxicol. Pathol. 2005, 33, 92-101. (19) Stoeckli, M.; Chaurand, P.; Hallahan, D. E.; Caprioli, R. M. Nat. Med. 2001, 7, 493-496. 10.1021/ac071301v CCC: $40.75

© 2008 American Chemical Society Published on Web 01/01/2008

logical applications.23,24 In addition to the analysis of protein described in the above references, direct lipid analysis in mammalian tissues25,26 has also been conducted, and histopathological materials27 and pharmacokinetics in the rat whole body section28 have been studied. Most of those studies used either a time-of-flight (TOF) mass analyzer coupled to either a matrix-assisted laser desorption ionization (MALDI) or secondary ion mass spectrometry (SIMS) ion source.29-31 Recently, other novel technologies for imaging mass spectrometry have been developed, such as a desorption electrospray ionization source32-34 and intermediate-pressure MALDI combined with a linear ion trap imaging system.35 To use a mass spectrometer as an imaging instrument, it is essential for the spectrometer to be equipped with an automatic rastering function, automatic data acquisition system, and visualization software. Recently, several manufacturers have released novel instruments having these features. Almost all manufacturers have developed in-house software for their instruments and have included a driver for instrument and image reconstruction. On the other hand, user-friendly visualization software, BioMap,36 is freely available; it is used commonly in mass imaging (MALDI MSI HP, http://www.maldi-msi.org/content). In a recent paper, Clerens et al. introduced a novel driver for imaging mass spectrometry and the data converter to generate a BioMapreadable file in Bruker’s MALDI-TOFMS instruments.37 In general, these imaging instruments were based mostly on MALDI-TOF, MALDI-TOF/TOF,38 or TOF-SIMS. MALDI-TOF imaging instruments are high throughput, and TOF-SIMS imaging instruments can provide submicrometer spatial resolution (∼500 (20) Sugiura, Y.; Shimma, S.; Moriyama, Y.; Setou, M. J. Mass Spectrom. Soc. Jpn. 2007, 55, 25-31. (21) Groseclose, M. R.; Andersson, M.; Hardesty, W. M.; Caprioli, R. M. J. Mass Spectrom. 2007, 42, 476-489. (22) Shimma, S.; Sugiura, Y.; Setou, M. J. Mass Spectrom. Soc. Jpn. 2006, 54, 210-211. (23) Schwartz, S. A.; Weil, R. J.; Johnson, M. D.; Toms, S. A.; Caprioli, R. M. Clin. Cancer Res. 2004, 10, 981-987. (24) Pierson, J.; Norris, J. L.; Aerni, H. R.; Svenningsson, P.; Caprioli, R. M.; Andren, P. E. J. Proteome Res. 2004, 3, 289-95. (25) Jackson, S. N.; Wang, H. Y.; Woods, A. S. J. Am. Soc. Mass Spectrom. 2005, 16, 2052-2056. (26) 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. (27) Touboul, D.; Piednoe¨l, H.; Voisin, V.; De La Porte, S.; Brunelle, A.; Halgand, F.; Lapre´vote, O. Eur. J. Mass Spectrom. 2004, 10, 657-64. (28) Khatib-Shahidi, S.; Andersson, M.; Herman, J. L.; Gillespie, T. A.; Caprioli, R. M. Anal. Chem. 2006, 78, 6448-6456. (29) Touboul, D.; Brunelle, A.; Halgand, F.; De La Porte, S.; Laprevote, O. J. Lipid Res. 2005, 46, 1388-1395. (30) Brunelle, A.; Touboul, D.; Laprevote, O. J. Mass Spectrom. 2005, 40, 985999. (31) McDonnell, L. A.; Mize, T. H.; Luxembourg, S. L.; Koster, S.; Eijkel, G. B.; Verpoorte, E.; de Rooij, N. F.; Heeren, R. M. Anal. Chem. 2003, 75, 437381. (32) Chen, H.; Pan, Z.; Talaty, N.; Raftery, D.; Cooks, R. G. Rapid Commun. Mass Spectrom. 2006, 20, 1577-1584. (33) Takats, Z.; Wiseman, J. M.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471-473. (34) Wiseman, J. M.; Ifa, D. R.; Song, Q.; Cooks, G. Angew. Chem., Int. Ed. 2006, 45, 7188-7192. (35) Garrett, T. J.; Yost, R. A. Anal. Chem. 2006, 78, 2465-2469. (36) Rohner, T. C.; Staab, D.; Stoeckli, M. Mech. Ageing Dev. 2005, 126, 177185. (37) Clerens, S.; Ceuppens, R.; Arckens, L. Rapid Commun. Mass Spectrom. 2006, 20, 3061-3066. (38) Shimma, S.; Sugiura, Y.; Hayasaka, T.; Hoshikawa, Y.; Noda, T.; Setou, M. J. Chromatogr., B 2007 (5-MAR).

nm).39 Both instruments have powerful advantages in imaging mass spectrometry.40 However, there is a potential problem with respect to the identification or structural analysis of molecules of interest. Ions generated on the tissue surface are extremely complicated, making it difficult to perform MSn. When the goal is to identify visualized molecules, it is essential to use a tandem mass spectrometer that can efficiently separate ions. With all of these issues in mind, we developed an imaging system based on a MALDI-quadrupole ion trap (QIT)-TOF (AXIMA-QIT, Shimadzu, Kyoto, Japan)43 mass spectrometer. The QIT acts as both an ion separator and a storage machine. Therefore, the instrument can provide molecular identification with highly sensitive MSn from mixture ions generated on the tissue surface. To ensure compatibility with imaging and structural analysis, we developed a data converter to provide BioMapreadable data from AXIMA data. This converter is available free of charge. In this paper, we describe the operating conditions under which AXIMA-QIT can be used for imaging. Imaging and molecular identification on mouse cerebellum are presented as demonstrations. EXPERIMENTAL SECTION AXIMA-QIT Operating Conditions for Mass Imaging Experiment. We wrote a manual describing how to operate the AXIMA-QIT software. For a detailed description, see the Supporting Information (SI; online), where the instrument’s operation windows are presented (Figures S-1-4). Mass Spectrometry. All imaging results shown in this paper were acquired using AXIMA-QIT. Acquisition was performed in the low-mass range mode (m/z 300-1000) by applying stage voltages of +14 (positive ion detection) and -10 V (negative ion detection). In the imaging experiment, a total of 20 laser shots per point were irradiated (4 s/point). In the MSn operation, the conditions of data acquisition (i.e., the laser power, collision energy, and number of laser irradiations) were changed in order to obtain product ion mass spectra that had high intensity and signal-to-noise ratios (S/N) of the fragment peaks. The calibration was performed with an external calibration method. Chemicals. See SI. Tissue Block Preparation. This study used 8-week-old male C57BL/6J mice purchased from Japan SLC (Shizuoka, Japan). The mice were sacrificed and dissected under diethyl ether anesthesia. Each tissue block was immediately frozen in liquid nitrogen to minimize degradation and kept at -80 °C. Sample Preparation for Lipid Analysis. Before sectioning, the tissue blocks were left for 30 min at -20 °C. Tissue sections were sliced to a thickness of 5 µm (brain)44 using a cryostat (CM 3050; Leica, Wetzler, Germany) and mounted onto an indium tin (39) Altelaar, A. F.; Klinkert, I.; Jalink, K., de Lange, R. P.; Adan, R. A.; Heeren, R. M.; Piersma, S. R. Anal. Chem. 2006, 78, 734-42. (40) Altelaar, A. F.; van Minnen, J.; Jimenez, C. R.; Heeren, R. M.; Piersma, S. R. Anal. Chem. 2005, 77, 735-41. (41) Touboul, D.; Brunelle, A.; Laprevote, O. Rapid Commun. Mass Spectrom. 2006, 20, 703-709. (42) Jackson, S. N.; Wang, H. Y.; Woods, A. S. J. Am. Soc. Mass Spectrom. 2006, in press. (43) Martin, R. L.; Brancia, F. L. Rapid Commun. Mass Spectrom. 2003, 17, 1358-1365. (44) Sugiura, Y.; Shimma, S.; Setou, M. J. Mass Spectrom. Soc. Jpn. 2006, 54, 45-48.

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

879

oxide (ITO) sheet purchased from Tobi (Osaka, Japan). This sheet has a thin ITO layer on a poly(ethylene terephthalate). The sheet was 125 µm thick, and its conductivity was 100 Ω. The transparency was 80% (λ ) 550 nm). On-Tissue Digestion. See SI. Matrix Application. A thin matrix layer was applied to the surface by a 0.2-mm nozzle caliber airbrush (Procon Boy FWA Platinum, Mr. Hobby, Tokyo, Japan). A total of 1 mL of 2,5dihydroxybenzoic acid (2,5-DHB) solution (30 mg/mL in 70% methanol/0.1% TFA) was sprayed. During spraying, the distance between the nozzle and the tissue surface was kept at 15 cm. After drying, the ITO sheet was attached to a MALDI-target plate by conductive tape to facilitate electrical conduction. The ITO sample attached to the MALDI-target plate is shown in Figure S-1. Conditions of Converter Test. See SI. How To Use the Converted Data in BioMap. The converted data file was loaded by “File f Load f Scan” in the BioMap menu bar. Detailed instructions are presented in the BioMap manual that comes with the BioMap software. Software Availability. The data converter is freely available for academic use at www.maldi-msi.org, or upon e-mail request ([email protected]). RESULTS AND DISCUSSION Data Converter from AXIMA Data to BioMap-Readable Data. An easy-to-use and high-throughput data converter was developed to generate a BioMap-readable data file (Analyze 7.5. file format) from AXIMA-QIT imaging data. The operation window is shown in Figure S-5. In this window, the operators select the locations and file name for the AXIMA data and the converted data. The extension of the AXIMA data was denoted by “.run”. There were two parameters: the m/z range to convert and the width of the bins. The time required for data conversion was mainly spent reading data. The data conversion time is proportional to the number of data points to be read. After the AXIMA data is read, the converted data file is generated in a few seconds. As a typical example, with the condition described in the Experimental Section, ∼30 s was required to generate a converted data file. After the conversion, two data files (“.img” and “.hdr”) were generated in the indicated locations. Imaging and Identification of Lipids in Positive Ion Detection Mode. To test and evaluate our IMS experiment system with QIT-TOF, we visualized lipid molecules on mouse brain sections. A photograph of the mouse cerebellum section with the imaging region (Figure 1a) and accumulated mass spectrum (Figure 1b) is presented. The mouse cerebellum was prepared on an ITO sheet and coated with 2,5-DHB. In this experiment, the time required for completing the scan at intervals of 100 µm in a 6 mm × 6 mm square was ∼4.0 h. Mass-selected images were recorded, and an m/z range of 600-900 was given in these images. In MS and MSn of a complex mixture of lipid molecules, QIT played an important role in obtaining high-quality MS (S/N, mass accuracy and mass resolution) and MSn data (S/N and number of fragment peaks). The most important feature of QIT is its high 880

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

capability of isolating the precursor ion (×B11 Da).7,45 This enables us to isolate one precursor ion of interest from extremely complex lipid profiles. Also, the storage of precursor ions of interest and dissociation with adequate collision-induced dissociation power were helpful for MSn analysis of less intense peaks. Another notable advantage of QIT is the avoidance of mass variance derived from the height variance of the sample surface. In AXIMAQIT, the TOF start signal is generated when ions are extracted from the QIT. As in TOF/TOF instruments, on the other hand, the TOF start signal is generated when the laser is irradiated. Especially on the irregular surface topography of a biological specimen, this feature of AXIMA-QIT improves mass resolution and mass accuracy without the need to optimize delayed extraction conditions or calibration methods at each data point. The cancellation of the ionization condition on the tissue surface by QIT and high mass resolution by the TOF analyzer are considered attractive for direct tissue MS and MSn. The scan detected a large number of peaks; in order to evaluate our system, we analyzed two peaks, at m/z 798.5 and 826.7, as an intense peak and less intense peak in the spectrum, respectively. As a result, we successfully made ion-distribution images and identified molecular species by performing MSn. Figure 1c shows that the ion at m/z 798.5 was distributed homogeneously on the whole section and that the ion at m/z 826.7 was localized in the white matter. After image reconstruction, we performed MSn for these peaks on the cerebellum surface and successfully identified the molecular species. The obtained MS2 product ion mass spectra (Figure 1d) indicates that these peaks were derived from the structure of alkali-metal adducted phosphatidylcholine (PC) according to its fragment ion patterns.46-48 Also, the 38 u loss of peaks observed in both MS3 spectra indicates the potassium adduct ion form. Furthermore, we detected the neutral losses of fatty acids and found that the difference in m/z came from the structure of the sn-1 fatty acid, i.e., m/z 798.5 was PC(16:0-18:1) and m/z 826.7 was PC(18:0-18:1). Imaging and Identification of Lipids in Negative Ion Detection Mode. In negative ion detection mode also, our experimental system provided good capabilities in both imaging and identification. The ion mode yielded information on negatively charged lipids, e.g., phosphatidylinositol (PI) and sulfatide (ST). We performed IMS and molecular identification with MSn in negative ion mode. Figure 2 shows the photograph of the imaging region in the mouse cerebellum section (Figure 2a) and the accumulated mass spectrum (Figure 2b). Here, we also selected two relatively minor mass peaks, which were at m/z 885.5 and 904.7. Images of these two m/z values were reconstructed (Figure 2c), and we found that these peaks were distributed complementarily; i.e., m/z 904.7 was in the white matter and medulla oblongata, and m/z 885.5 was in the gray matter. To identify these peaks, MSn was performed (Figure 2d). From the high-quality spectra obtained with QIT-TOF, characteristic peaks to determine molecular species were detected. The peak (45) Ikegami, K.; Heier, R. L.; Taruishi, M.; Takagi, H.; Mukai, M.; Shimma, S.; Taira, S.; Hatanaka, K.; Morone, N.; Yao, I.; Campbell, P. K.; Yuasa, S.; Janke, C.; MacGregor, G. R.; Setou, M. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 3213-3218. (46) Hsu, F. F.; Turk, J. J. Am. Soc. Mass Spectrom. 2003, 14, 352-363. (47) Ho, Y. P.; Huang, P. C. Rapid Commun. Mass Spectrom. 2002, 16, 15821589. (48) Han, X.; Gross, R. W. J. Am. Soc. Mass Spectrom. 1995, 6, 1202-1210.

Figure 1. Example of lipid imaging and precursor ion mass spectra obtained directly from the mouse cerebellum section with the positive ion detection mode. Scanning in a 6 mm × 6 mm region (a) revealed a large number of peaks (b). The imaging results indicated that m/z 798.5 was distributed homogenously and that m/z 826.7 was relatively localized in the white matter (c). These peaks were identified by direct MSn and were derived from phosphatidylcholine species (m/z 798.5; PC(16:0-18.1) and m/z 826.7; PC(18:0-18:1)) (d). Bars, 1.5 mm.

at m/z 241 was one deprotonated cyclic inositol monophosphate, and the peaks at m/z 283 and 303 were deprotonated fatty acids C18:0 and C20:4, respectively (Figure 2d, upper panel). The obtained product ion mass spectrum indicated that m/z 885.5 was derived from PI, and the composition of fatty acid was 1-stearoyl2-arachidonoyl (PI(18:0-20:4)). The structure information of the labeled peak is shown in Figure S-6. In another product mass spectrum of Figure 2d (lower panel), the peak at m/z 540 was derived from lyso-ST (Figure S-7). This lyso-ST was detected when ceramide contained hydroxy fatty acid.49 Therefore, the fatty acid

of ST could be calculated from the difference in m/z between the precursor ion and lyso-ST. We concluded that m/z 904.7 was ST(d18:1-C24h:1). We have shown that imaging with an AXIMA-QIT tandem mass spectrometer allows visualization of not only the differences in phospholipid classes but also the differences in fatty acid composition (species). The visualization of lipid distribution in situ is very important in molecular and functional biology, especially in relationship to diseases. Although some lipid imaging (49) Ohashi, Y. In Lipids III Glycolipids; Nihon Seikagakkai, Ed.; Tokyo Kagaku Dojin: Tokyo, Japan, 1990; pp 189-192.

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

881

Figure 2. Example of lipid imaging and precursor ion mass spectra with negative ion detection mode. From the imaging region (a), negatively charged lipids based on the head group structure were detected (b). The molecules of m/z 885.5 and 904.7 were distributed in the gray matter and white matter, respectively (c). The precursor ion mass spectrometry showed that m/z 885.5 was phosphatidylinositol (18:0-20:4) and m/z 904.7 was sulfatide (d18:1-C24h:1). Abbreviations: cIP, cyclic inositol phosphate; IP, inositol phosphate; FA, fatty acid; LPA, lysophosphatidic acid; LPI, lysophosphatidylinositol (d). Bars, 2.0 mm.

methods have been conducted by the use of tags50 and interaction between lipids and other biomolecules,51,52 these procedures provide visualization only of specific classes of lipids. In this context, IMS is the only method to visualize the lipid structure. (50) Kuerschner, L.; Ejsing, C. S.; Ekroos, K.; Shevchenko, A.; Anderson, K. I.; Thiele, C. Nat. Methods 2005, 2, 39-45. (51) Emoto, K.; Kobayashi, T.; Yamaji, A.; Aizawa, H.; Yahara, I.; Inoue, K.; Umeda, M. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 12867-12872. (52) Ishitsuka, R.; Yamaji-Hasegawa, A.; Makino, A.; Hirabayashi, Y.; Kobayashi, T. Biophys. J. 2004, 86, 296-307.

882

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

As shown in Figures 1c and 2c, visualization of the phospholipids reveals a quite different distribution on the mouse cerebellum section among their classes and even their species. So far, only IMS can visualize the distinct tissue distribution of phospholipid species that have different lipid moieties (sn-1 or sn-2), e.g., C16:0 and C18:0. Therefore, visualization and direct identification using IMS and direct tandem MS, especially with a high-performance instrument such as AXIMA-QIT, will soon become an effective tool for studies on lipids and glycolipids.

Figure 3. Example of tryptic-digested protein imaging and precursor ion mass spectra with positive ion detection mode. (a) Optical image of imaging region and (b) accumulated mass spectrum from imaging region. (c) and (d) are imaging results of m/z 1460.8 and 1743.9, which are labeled by asterisks in (b). Merged image (red, m/z 1460.8; and green, m/z 1743.9) is shown in (e). These peaks were identified by direct MSn and were identified as the fragment ions of myelin basic protein (MBP) (f) and histone H2B (g).

Imaging and Identification of Peptides in Positive Ion Detection Mode. Next, we tested the QIT-TOF instrument’s ability to analyze peptide ions directly on the tissue sections. We utilized trypsin-digested mouse brain sections16,17 to evaluate the instrument’s performance for both visualization and molecular identification of peptides. We performed IMS and MSn in positive ion mode. Figure 3 shows the results of the IMS experiment, including a photograph of a mouse cerebellum section (Figure 3a) and accumulated mass spectrum (Figure 3b). Images of two m/z values, 1460.8 and 1743.9, were reconstructed (Figure 3c,d), and we found these peaks distributed complementarily: m/z 1460.8 was in the white matter and medulla oblongata, and m/z 1743.9 was in the gray matter (Figure 3e). To identify these peaks, MSn was performed directly on the tissue sections, and they were

identified as fragment ions of myelin basic protein (MBP) and histone H2B, respectively (Figure 3f,g). To perform proper molecular identification with tandem mass spectrometry on tissue sections that have extremely complex mixtures of biomolecules, precursor ion selection with high resolution is a very important issue. In Figure 4, we demonstrate that QIT-TOF can perform efficient molecular identification of peptide ions on the tissue surface. We obtained product ion spectra and then compared the precursor-ion isolation ability and the quality of the spectra, i.e., S/Ns and mass resolution of product ion peaks, between QIT-TOF and TOF/TOF. Figure 4 shows the results of the product ion spectra of m/z 1460.8. Figure 4a shows that the ability of QIT-TOF to isolate a single monoisotopic peak with high resolution (1 Da).7,45 Employing an extra high separation Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

883

Figure 4. Comparison of product ion spectrum of m/z 1460.8 between QIT-TOF and TOF/TOF. Product ion spectra of (a) QIT-TOF and (b) TOF/TOF by changing the mass window of precursor ion selection. Comparison of (c) S/N and (d) mass resolution of fragment peaks between QIT-TOF and TOF/TOF.

mode, the instrument successfully isolated the monoisotopic peak (labeled with *) from the complex isotopic envelope, which presumably contained ions derived from other biomolecules (labeled with †). Moreover, at such a high-resolution setting, we 884 Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

were able to identify the peptide as MBP (sequence position, 237248) with a significant MASCOT score. On the other hand, the TOF/TOF instrument exhibited little ability to select precursor ions (Figure 4b). Figure 4c, which shows the S/N and mass

resolution of product ions (b9, b10, and y4 ions), indicates that the S/N and peak resolution were dramatically improved in the QITTOF instrument. We attribute this result to two instrumental features of AXIMA-QIT: the ions generated on the tissue surface do not enter the TOF analyzer directly, and before entering the TOF analyzer, ions are stored and energy is cooled in the QIT. Therefore, QIT can cancel the difference in ionization position and initial energy at ionization, thus degrading the peak shape. Furthermore, even if high-resolution precursor isolation was performed, the S/N and mass resolution tended to be unchanged by the QIT equipment. On the other hand, precursor ion selection at higher resolution (smaller PCIS width range) resulted in the degradation of spectra, presumably because of the loss of ions at the precursor selection stage in the TOF/TOF instrument.

pholipids, glycolipids, and tryptic-digested proteins. Data converter software is available at www.maldi-msi.org or upon request by e-mail ([email protected]).

CONCLUSIONS Two-dimensional imaging and molecular identification from tissue surfaces using AXIMA-QIT, which is a MALDI mass spectrometer in the QIT-TOF instrument, was presented for the first time. The methodology presented here covers methods for converting the data into mass-selective images, the use of MSn data to identify specific compounds seen in these images, and examples of the application of these capabilities including phos-

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

ACKNOWLEDGMENT We thank our colleagues Y. Kakude, M. Furuta, and K. Ichimura (Life Science Laboratory, Shimadzu, Kyoto, Japan) for their cooperation with us in developing the data converter. We also thank M. Furuta for the data of AXIMA-QIT basic performance. This work was supported by a Grant-in-Aid under the SENTAN program of the Japan Science and Technology Agency to M.S. S.S. and Y.S. contributed equally to this work. Parts of this shimadzu system spec and data were announced in commercial news letters.

Received for review June 20, 2007. Accepted November 5, 2007. AC071301V

Analytical Chemistry, Vol. 80, No. 3, February 1, 2008

885