High-Reactivity Matrices Increase the Sensitivity of Matrix Enhanced

Sep 22, 2011 - ... a raw text data format by saving as “.asc” files, which were imported into MATLAB (The ... Analyte labels at the top of Figure ...
0 downloads 0 Views 3MB Size
ARTICLE pubs.acs.org/ac

High-Reactivity Matrices Increase the Sensitivity of Matrix Enhanced Secondary Ion Mass Spectrometry Fabian N. Svara,† Andras Kiss,†,‡ Thorsten W. Jaskolla,§,|| Michael Karas,§ and Ron M. A. Heeren*,†,‡ †

FOM Institute for Atomic and Molecular Physics (AMOLF), Science Park 104, 1098 XG Amsterdam, The Netherlands The Netherlands Proteomics Centre, Padualaan 8, 3584 CH Utrecht, The Netherlands § Cluster of Excellence Macromolecular Complexes, Institute of Pharmaceutical Chemistry, Goethe-University Frankfurt, Max-von-Laue-Strasse 9, 60438 Frankfurt, Germany Institute of Medical Physics and Biophysics, University of Muenster, Robert-Koch-Strasse 31, 48149 Muenster, Germany

)



bS Supporting Information ABSTRACT: Secondary ion mass spectrometry (SIMS) is a desorption/ionization method in which ions are generated by the impact of a primary ion beam on a sample. Classic matrix assisted laser desorption and ionization (MALDI) matrices can be used to increase secondary ion yields and decrease fragmentation in a SIMS experiment, which is referred to as matrix enhanced SIMS (ME-SIMS). Contrary to MALDI, the choice of matrices for ME-SIMS is not constrained by their photon absorption characteristics. This implies that matrix compounds that exhibit an insufficient photon absorption coefficient have the potential of working well with ME-SIMS. Here, we evaluate a set of novel derivatives of the classical MALDI matrices α-cyano-4hydroxycinnamic acid (CHCA) and 2,5-dihydroxybenzoic acid (DHB) for usability in ME-SIMS. This evaluation was carried out using peptide mixtures of different complexity and demonstrates significant improvements in signal intensity for several compounds with insufficient UV absorption at the standard MALDI laser wavelengths. Our study confirms that the gas-phase proton affinity of a matrix compound is a key physicochemical characteristic that determines its performance in a ME-SIMS experiment. As a result, these novel matrices improve the performance of matrix enhanced secondary ion mass spectrometry experiments on complex peptide mixtures.

S

econdary ion mass spectrometry (SIMS) is a method in which material from a solid sample surface is desorbed and ionized by the impact of primary ions under ultrahigh vacuum conditions. The secondary ions that are thus generated are introduced into a mass spectrometer for mass separation and detection. SIMS is a promising approach in mass spectrometric imaging, since it allows for high spatial resolution, with pixel sizes routinely in the range of hundreds of nanometers. The ion yield, however, is much lower than in the competing desorption and ionization technique, matrix assisted laser desorption and ionization (MALDI), and drops steeply with increasing analyte masses.1 6 One approach to address this problem is to use matrices in SIMS that enhance analyte ionization efficiency and lower analyte fragmentation, which is referred to as matrix enhanced-SIMS (ME-SIMS).7,8 A wide variety of matrix compounds has been investigated in the past. Frozen noble gases,9 ammonium chloride,10 carbon,11 nitrocellulose,12 frozen organic solvents,13 ice,14 trehalose,15 and classical MALDI matrices, such as αcyano-4-hydroxycinnamic acid (CHCA) and 2,5-dihydroxybenzoic acid (DHB)8 have been tested. Nowadays, the term MESIMS mainly refers to SIMS experiments that use typical MALDI matrices and sample preparation protocols. The underlying mechanism by which the matrices improve the performance of r 2011 American Chemical Society

SIMS experiments is not precisely known. It has been suggested that, upon primary ion impact, matrix and analyte molecules and ions are sputtered in clusters. Dissociation of these clusters cools and stabilizes the incorporated analyte species, yielding lower internal energy secondary analyte ions and thereby reducing fragmentation.16 Another contributing mechanism might be that polymeric chains of larger analytes are disentangled by the matrix and that the matrix leads to a lower overall cohesiveness of the sample, possibly also leading to a separation of reactive molecules from each other.17 We recently experimentally verified that protonated matrix molecules are the species necessary for analyte ionization in MALDI.18 It was already assumed before that alleviated proton transfer reactions from protonated matrix molecules enable higher analyte ion yields19 which led to the halogenated α-cyanocinnamic acid derivatives 4-chloro-α-cyanocinnamic acid (ClCCA)19,20 and α-cyano-2,4-difluorocinnamic acid (DiFCCA).21 This was recently shown for the compounds α-cyano-4-trifluoromethylcinnamic acid (F3C-CCA) and α-cyano-2,3,4,5,6-pentafluorocinnamic acid (PentaF-CCA) which allow Received: August 23, 2011 Accepted: September 22, 2011 Published: September 22, 2011 8308

dx.doi.org/10.1021/ac202222h | Anal. Chem. 2011, 83, 8308–8313

Analytical Chemistry

ARTICLE

Figure 1. Structures and abbreviations of the α-cyanocinnamic acid derivatives and 2,5-disubstituted benzoic acid derivatives investigated as ME-SIMS matrices.

for very high analyte sensitivities at MALDI-MS using optimized irradiation wavelengths but are completely unsuited as MALDI matrices at the standard wavelengths.22 Since no such constraint exists in ME-SIMS, a larger number of potentially reactive derivatives of the classical MALDI matrices may prove beneficial in ME-SIMS. Consequently, the concept of increasing the matrix sensitivity by lowering their PA with the help of electron-withdrawing substituents19 was extendable to even more reactive compounds. In addition to the highly reactive halogenated α-cyanocinnamic acid derivatives mentioned above, we characterize a set of 14 novel matrices, all derivatives of CHCA or DHB with assumed low PAs, with respect to their performance in ME-SIMS of peptide samples. We demonstrate that some of these compounds afford significant increases in signal intensity. These findings support the hypothesis that the gas-phase basicity is an important parameter in the SIMS desorption and ionization process and may lead to further structural optimizations of ME-SIMS matrixes in the future.

’ MATERIALS AND METHODS Materials. The library of experimental matrices was synthesized by Dr. Thorsten Jaskolla in the group of Prof. Michael Karas. The tryptic digest of casein was obtained as Bacto Tryptone for microbiological growth media (BD, Franklin Lakes, NJ). Bruker Peptide Calibration Standard II was from Bruker Daltonics (Billerica, MA). All other chemicals were purchased from Sigma-Aldrich (Taufkirchen, German) and used as received. Light Microscopy. Reflected light microscopy was performed on a Leica DMRX microscope. Objectives used were Leitz Wetzlar PL Fluotar 20x 0.45D and 50x 0.85D. Digital images

were acquired using a Nikon DXM1200 camera and the Nikon ACT-1 software version 2.20. Dried-Droplet Sample Preparation. Matrices were dissolved in mixtures of acetonitrile and water, with 0.1% TFA. Mixture ratios were 50:50, 70:30, and 80:20 acetonitrile/water, where DHB-derived matrices were dissolved in high-water content (50:50) mixtures and highly halogenated CHCA analogues (PentaBr-CCA and pentaFCCA) were dissolved in high-acetonitrile (80:20) mixtures. Matrix solutions were prepared to concentrations of 50 mM, except for PentaBr-CCA, which fully dissolved at a maximum concentration of 5 mM, which was used for these experiments. Pure matrix solutions (0.5 μL) as well as matrix analyte solutions (0.5 μL each) were spotted on metal target plates (Waters Corporation, Milford, MA), originally produced for the Synapt HDMS instrument and air-dried. ME-SIMS-TOF. Samples were measured on a TRIFT II (Physical Electronics, Eden Prairie, MN) SIMS-time-of-flight (TOF) system, equipped with an AuGe primary ion gun (FEI, Hillsboro, OR) tuned for the selection of Au+ primary ions, with raster sizes set to values between 50 and 200 μm. The gun was operated at an extractor current of 2.25 μA. A postaccelerator setting of 7 kV was typically used. The mass range from 0 to 3500 m/z was typically selected for profiling experiments. Mass spectra were acquired and analyzed using the WinCadence software, version 4.4 (Physical Electronics). MALDI-TOF. MALDI-TOF experiments were performed on an Ultraflex III instrument (Bruker Daltonics, Bremen, Germany) equipped with a 355 nm diode pumped Nd:YAG laser operating at a repetition rate of 50 Hz with a pulse energy up to 150 μJ. Data Processing and Analysis. TOF-SIMS profiling experiments were exported from the WinCadence software to a raw text 8309

dx.doi.org/10.1021/ac202222h |Anal. Chem. 2011, 83, 8308–8313

Analytical Chemistry

ARTICLE

Figure 2. ME-SIMS ion images (left) compared to optical images (right) of pure matrix crystals of the investigated compounds prepared by standard DD preparations. The numbers in the ion image scale bars refer to the maximal counts per pixel.

data format by saving as “.asc” files, which were imported into MATLAB (The MathWorks, South Natick, MA) for plotting, visual inspection, and automated analysis using scripts developed for the purpose of this work. Peak picking was performed using functions provided by the Bioinformatics Toolbox. Spectra were smoothed using the least-squares polynomial method provided by the mssgolay function. Picking was then performed on the smoothed spectra using the wavelet-based approach implemented by mspeaks.

’ RESULTS Characterization of the Pure Matrix Compounds. We evaluated a set of 21 matrices including 19 novel compounds and the classic matrices CHCA and DHB for ME-SIMS. All of the novel compounds were expected to have reduced proton affinities compared to their classic counterparts, because of the introduction of electron withdrawing substituents. The complete set is shown in Figure 1.

We initially characterized the pure matrix compounds by observing their dried droplet (DD) crystallization patterns by reflected light microscopy and subsequent positive ion mode SIMS imaging of dried droplets. The comparison of optical and ion images obtained is shown in Figure 2. Ion images are based on characteristic peaks. If the peaks corresponding to the singly protonated m/z value could not be detected, sodiated or water loss peaks were used instead, if available. The numbers in the ion image scale bars refer to the maximal counts per pixel. The characteristic crystal shapes can, in most cases, clearly be recognized in the ion images. Crystal shapes varied, ranging from homogenously distributed crystals of sizes below 1 μm to needlelike structures of tens of micrometers length. Signal intensities for characteristic peaks also showed considerable variability, with some compounds not being detected at all. Standards. The matrix compounds were evaluated by DD preparations of matrix solutions mixed with a commercial calibration standard (Bruker Peptide Calibration Standard II) and a tryptic digest of casein (Bacto Tryptone). ME-SIMS 8310

dx.doi.org/10.1021/ac202222h |Anal. Chem. 2011, 83, 8308–8313

Analytical Chemistry

ARTICLE

Figure 3. ME-SIMS spectra of DD preparation of the Bruker Peptide Calibration Standard II measured with the compounds of interest. Stars next to compound names indicate that some peptide detection was also achieved from the same sample on a MALDI instrument. Areas labeled in gray indicate intense interfering matrix-related peak distributions.

spectra were recorded in positive ion mode, using a postacceleration voltage of 9 kV to detect higher mass analytes more efficiently. Three replicates were measured per matrix. To reduce hot spot effects in the dried droplet sample, the stage was manually moved to different regions in which crystals were clearly visible on the sample camera display during the acquisition. Figure 3 shows representative ME-SIMS spectra for the calibration standard. The area in the PentaBr-CCA spectrum labeled in gray highlights matrix clusters interfering with the detection of angiotensin II. Asterisks next to compound abbreviations indicate that it was possible to obtain some peptide signal from the same sample on an Ultraflex III MALDI instrument, with only weak analyte ion intensities in most cases. Analyte labels at the top of Figure 3 refer to the m/z ratios of singly protonated species. Significant differences in performance between the matrices were observed, with many compounds yielding higher analyte intensities than the classic compounds CHCA and DHB in ME-SIMS. Additionally, certain matrices favor certain analytes, for example, PentaF-CCA afforded relatively high signal intensities for the singly protonated bradykinin 1-7 peak but performed less well than many other matrices for the remaining peptides in the sample. The casein digest spectra, shown in Figure 4, confirm these results. Matrix performance was not always consistent between different analytes investigated. For that reason we derived overall rankings by ranking by intensity for single analytes first, followed by counting the occurrences among the top six matrices for every

analyte, weighing the top two with 3, the next two with 2, and the last two with 1. Weighing all compounds equally lead to largely comparable results, but analyte intensity differences between the first few rank positions can be big and we thus expect the weighing to better reflect the actual matrix performance. The results are shown in Table 1. For the peptide calibration standard, the six lowest mass singly protonated peptide peaks were used for the ranking. The casein digest was ranked using a selection of six singly protonated peaks. The complete rankings by analyte are included in Supplementary Table 1 in the Supporting Information for the calibration standard and Supplementary Table 2 in the Supporting Information for the casein digest.

’ DISCUSSION AND CONCLUSIONS We have demonstrated that new derivatives of classic MALDI matrices can yield improved signal intensities in ME-SIMS measurements of peptide samples. 4-NO2-CCA exhibited outstanding results among all tested experimental and classic matrix compounds. Other promising compounds for ME-SIMS were PentaBr-CCA, 2,4-DiF-CCA, PentaF-CCA, 2,4-DiCl-CCA, 2,4,6-TriF-CCA, and 4-CN-CCA. PentaBr-CCA with five electron-withdrawing bromine substituents enables intense secondary ion yields for most analytes. However, the high number of bromines also causes restricted practical usability due to matrix ion signals with broad isotopic distributions virtually over the entire mass range covered in these experiments. 8311

dx.doi.org/10.1021/ac202222h |Anal. Chem. 2011, 83, 8308–8313

Analytical Chemistry

ARTICLE

Figure 4. ME-SIMS spectra of a DD preparation of Bacto Tryptone (tryptic digest of casein) obtained using the experimental matrices in positive ion mode. Areas labeled in gray indicate intense interfering matrix-related peak distributions. Analyte labels at the top are based on a list of known peptide components of the sample obtained from an LC ESI experiment courtesy of Dr. Maarten Altelaar, Netherlands Proteomics Center, expanded by m/z values yielded by sodiation and water loss.

Table 1. Relative Rankings of the Experimental Matrices Derived from Peptide Samplesa ranking

casein digest

standard peptides

combined

1

4-NO2-CCA

PentaBr-CCA

4-NO2-CCA

2

PentaF-CCA

4-NO2-CCA

PentaBr-CCA

3

2,4,6-TriF-CCA 4-CN-CCA

2,4-DiCl-CCA

4

2,4-DiCl-CCA

2,4-DiCl-CCA

CHCA

5

CHCA

2,4-DiF-CCAamide

PentaF-CCA

6 7

3,5-DiCl-CCA 4-F3C-CCA

CHCA DHB

4-CN-CCA 2,4-DiF-CCAamide

8

PentaBr-CCA

4-(2,2-DiCNvinyl)-BA 4-F3C-CCA

9

2,5-DiCN-BA

PentaF-CCA

2,4,6-TriF-CCA

DiFCCA

4-Br-CCA

4-(2,2-DiCNvinyl)-BA

10 a

Rank positions are based on the number of occurrences among the top six performers for a selection of eight clearly detected analytes for each sample.

The rankings were complicated by the fact that matrix performance varied between analytes in the same sample. While top-performing matrices, such as 4-NO2-CCA, consistently provided good analyte sensitivity for ME-SIMS, the relative rank positions could vary significantly between individual peaks. This can, to a degree, be attributed to experimental variability, especially as SIMS spectra become increasingly noisy with increasing m/z. Hot-spot effects likely caused most of the variability in our experiments, as crystallization patterns were very different between matrices and we even occasionally observed different analyte signals from different crystals in the same sample.

The assumed correlation between lowered matrix proton affinity and increased proton transfer efficiency for efficient formation of analyte ions in ME-SIMS was the main hypothesis guiding the synthesis of the compound library investigated here. Thus, all synthesized compounds of this work are more acidic than their classic matrix counterparts due to electron withdrawing halogens or cyano- or nitro-groups in the cases of 4-CNCCA, 5-CN-2-OH-BA, 2,5-DiCN-BA, or 4-NO2-CCA, respectively. Most of the investigated matrix derivatives proved to exhibit increased analyte sensitivities compared to the classic structures and therefore support this correlation. The matrix that performed best, 4-NO2-CCA, was one of the most acidic matrices in the set exhibiting a proton affinity of about 50 kJ/mol lower than that of CHCA (calculation data not shown). Even though this very reactive compound strongly absorbs at the standard UV-laser wavelengths, it exhibits only poor MALDI-MS performance. This most probably can be explained by fragmentation of the matrix structure upon photoionization with electron loss from the zwitterionic nitro group leading to excessive matrix fragmentation in MALDI-MS which obviously is not a problem in ME-SIMS with ions for energy influx. Additionally, the presented results clearly show that a low proton affinity alone is not sufficient for high analyte sensitivities For example, 2,4-DiCl-CCA performed significantly better than 2,4-DiF-CCA. As discussed earlier, it can be assumed that appropriate matrix morphology is an additional necessary requirement for efficient analyte ionization.18 The rationale guiding the synthesis of the compounds investigated here can be extended to further substances with assumed even higher sensitivities. On the basis of the outstanding performance of 4-NO2-CCA, it might be promising to investigate 8312

dx.doi.org/10.1021/ac202222h |Anal. Chem. 2011, 83, 8308–8313

Analytical Chemistry its higher substituted counterparts 2,4- and 3,5-DiNO2-CCA, and 2,4,6-TriNO2-CCA. However, it can be assumed that the maximum number of nitro-groups will be limited by the requirement of sufficient matrix stability. A further clarification of the physicochemical variables that determine the performance of a matrix, by screening larger and more representative compound collections, will likely be beneficial to the applicability of ME-SIMS as an analytical technique.

’ ASSOCIATED CONTENT

bS

Supporting Information. Detailed analyte rankings of the investigated standard peptide mixture used for Table 1 (Supplementary Table 1) and detailed analyte rankings of the investigated tryptic casein digest used for Table 1 (Supplementary Table 2). This material is available free of charge via the Internet at http:// pubs.acs.org.

ARTICLE

(17) Delcorte, A.; Garrison, B. J. J. Phys. Chem. B 2003, 107, 2297–2310. (18) Jaskolla, T. W.; Karas., M. J. Am. Soc. Mass Spectrom. 2011, 22, 976–988. (19) Jaskolla, T. W.; Lehmann, W.-D.; Karas, M. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 12200–12205. (20) Jaskolla, T. W.; Papasotiriou, D. G.; Karas, M. J. Proteome Res. 2009, 8, 3588–3597. (21) Teuber, K.; Schiller, J.; Fuchs, B.; Karas, M.; Jaskolla, T. W. Chem. Phys. Lipids 2010, 163, 552–560. (22) Jaskolla, T. W., Soltwisch, J.; Hillenkamp, F.; Karas, M.; , Dreisewerd, K. 59th ASMS Conference on Mass Spectrometry and Allied Topics, Denver, Colorado, June 5 9, 2011.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT This work is part of the research program of the “Stichting voor Fundamenteel Onderzoek der Materie (FOM)”, which is financially supported by the “Nederlandse organisatie voor Wetenschappelijk Onderzoek (NWO)”. We thank Dr. Maarten Altelaar from The Netherlands Proteomics Center for LC ESI analysis of the used casein digest. This work was partially supported by the WI Bank for economic and public infrastructure development and the German Research Council (Grant DFG JA 2127/1-1). ’ REFERENCES (1) Altelaar, A. F. M.; Heeren, R. M. A. Methods Mol. Biol. (Clifton, N.J.) 2009, 492, 295–308. (2) Amstalden van Hove, E. R.; Smith, D. F.; Heeren, R. M. A. J. Chromatogr., A 2010, 25, 3946–3954. (3) Chughtai, K.; Heeren, R. M. A. Chem. Rev. 2010, 110, 3237–3277. (4) Heeren, R.; McDonnell, L.; Amstalden, E.; Luxembourg, S.; Altelaar, A.; Piersma, S. Appl. Surf. Sci. 2006, 252, 6827–6835. (5) Heeren, R. M. A.; Smith, D. F.; Stauber, J.; K€ukrer-Kaletas, B.; MacAleese, L. J. Am. Soc. Mass Spectrom. 2009, 20, 1006–1014. (6) McDonnell, L. A.; Heeren, R. M. A. Mass Spectrom. Rev. 2007, 26, 606–643. (7) Delcorte, A. Appl. Surf. Sci. 2006, 252, 6582–6587. (8) Wu, K. J.; Odom, R. W. Anal. Chem. 1996, 68, 873–882. (9) Jonkman, H. T.; Michl, J.; King, R. N.; Andrade, J. D. Anal. Chem. 1978, 50, 2078–2082. (10) Liu, L. K.; Busch, K. L.; Cooks, R. G. Anal. Chem. 1981, 53, 109–113. (11) Ross, M. M.; Colton, R. J. Anal. Chem. 1983, 55, 150–153. (12) Touboul, D.; Halgand, F.; Brunelle, A.; Kersting, R.; Tallarek, E.; Hagenhoff, B.; Laprevote, O. Anal. Chem. 2004, 76, 1550–1559. (13) Huang, M.-W.; Chei, H.-L.; Huang, J.-P.; Shiea, J. Anal. Chem. 1999, 71, 2901–2907. (14) Wucher, A.; Sun, S.; Szakal, C.; Winograd, N. Anal. Chem. 2004, 76, 7234–7242. (15) Cheng, J.; Winograd, N. Anal. Chem. 2005, 77, 3651–3659. (16) Cooks, R. G.; Busch, K. L. Int. J. Mass Spectrom. Ion Phys. 1983, 53, 111–124. 8313

dx.doi.org/10.1021/ac202222h |Anal. Chem. 2011, 83, 8308–8313